Digitized by tine Internet Arciiive

in 2011 witin funding from

University of Toronto

littp://www.arcliive.org/details/pt1minutesofproce95inst

!^

MINUTES OF PKOCEEDINGS

THE INSTITUTION

CIVIL ENGINEERS;

WITH OTHER

SELECTED AND ABSTRACTED PAPERS.

Vol. XCV.

EDITED BY

JAMES FOREEST, Assoc. Inst. C.E., Secretary.

LONDON: Publtsl)el) bg tfje Snstituttcn,

25, GREAT GEORGE STREET, WESTMINSTER, S.W.

[Telegrams, "Institution, London." Telephone, "3051."]

1889.

[The right of Publication and of Translation is reserved.]

/A

ADVEETISEMENT.

The lastitution as a body is not responsible eitlier for tlie statements made, or for the opinions expressed, in the following page-s.

LONDOS; PltlNTED BY WILLIAM CLOWES Aim SONS, LmiTED, STAUFOnL SIEBET AND CIIABINC CBOaS.

CONTENTS.

Sect. I.— MINUTES OF PROCEEDINGS.

13 and 20 November, 1888.

PAGE

" Friction-Brake Dynamometers." By "W. W. Beaumont. (18 cuts) . . 1

Discussion on ditto. (14 cuts) 29

Correspondence ou ditto. (4 cuts) 70

27 November, 1888.

" The Witham New Outfall Channel and Improvement Works." By J. E.

Williams. (1 plate) 78

Appendixes to ditto. (1 cut) 87

Discussion on ditto. (1 cut) 92

Correspondence on ditto. (2 cuts) 105

4 and 11 December, 1888.

Transfer of Associate Members to the class of Members 112

Admission of Students 112

Election of Members, Associate Members, and Associates 11^

" On the Influence of Chemical Composition on the Strength of Bessemer- Steel Tires." By J. 0. Arnold. (11 cuts) 115

Discussion on ditto 131

Correspondence on ditto 160

18 December, 1888.

" The Friction of Locomotive Slide-Valves." J. A. F. Aspinall. (3 plates) 167

Discussion on ditto (11 cuts) 179

Correspondence on ditto 194

IV CONTENTS.

Sect. II.— OTHER SELECTED PAPERS.

PAGE

" Preliminary Survey in New Countries, as exemplified in the Survey of

Windward Hawaii." By T. G. Gribblb 195

Appendixes to ditto. (3 cuts) 202

" Rapid Surveying." By F. D. Topham 209

" The Practice of Survepng in the Australasian Colonies." S. K. Vickehy 211

Appendixes to ditto (1 cut) 215

" The Manufacture of Oil-Gas on the Pintsch System, and its Application to the Lighting of Railway Carriages." By G. M. Hunter. (1 plate.

Scuts) 218

Appendix to ditto 228

" Hurst's Triangtilar Prismatic Formula for Earthwork compared with the

Prismoidal Formula." By J. W. Smith. (11 cuts) 229

" Alpine Engineering." By L. F. Verxon-Haecourt. (2 plates, 1 cut) . 237

Appendix to ditto 278

" The River Clyde." By D. Macalister 279

" The Failure of the Kali Nadi Aqueduct on the Lower Ganges Canal."

Abstracted by W. H. Thelwall 283

" The Reparation of Betchworth Tunnel, Dorking, on the London, Brighton

and South Coast Railway." By G. Lopes. (5 cuts) 291

" The Permanent-Way of some Railways in Germany and in Austria- Hungary." Translated and Abstracted by W. B. Woethington . . . 303 " The Speed-Trials of the latest additions to the Admiral Class of British

War- Vessels." By D. S. Capper. (1 plate, 5 cuts) 325

Appendix to ditto 343

♦* On the Use of Heavier Rails for Safety and Economy in Railway TraflBc."

By C. P. Sandberg. (1 plate, 1 cut) 354

Obituary 360

Viscount Eversley, 360 ; John Brown, 361 ; William Armitage Brown, 363 ; Henry Carr, 364 ; Robert Denny, 369 ; James Easton, 370 ; John Fowler (of Stockton), 371 ; William Francis, 374 ; Frank Alexander Brown Geneste, 375 ; Charles Markham, 377 ; Julius Pazzani, 379 ; William Rogers, 380 ; George Hennet Ross, 382 ; Ranson Colecome Batterbee, 383; James John Alexander Flower, 384; Samuel Harpur, 385; Ernest Frederic Morant, 387; Robert Pinchin, 388 ; Charles Thomas Spencer, 391 ; John Trickett, 392 ; John Wakeford, 393 ; Herbert Francis Waring, 394 ; John Ashworth, 394 ; Major Augustus Samuel William Connor, 396 ; George HawMns, 397 ; Staif Commander Graham Hewctt Hills, 398.

CONTENTS.

Sect. III.— ABSTRACTS OF PAPEES IX FOREIGN TRANSACTIONS AND PERIODICALS.

PAGE

New Theory of Friction. N. Peteofp 407

On the Critical Extension of Bodies strained simultaneously in Several

Directions. — Wehage 410

Landslip at Zug, Switzerland, 5 July, 1887 411

A Folding Levelling-Stafif. H. Bentabol 415

The Alignment of a Tunnel at Stuttgart. — Widmann 415

Methods of Testing the Resistance of Stones, Cements, and other Building

Materials. L. Durand-Clate 416

On the Testing of Paper. N. Haselkoos 420

Yield of Hydraulic Mortars. — Bonnami 421

Pulverization of Clay and its Application at the Works of the Societe

Arnaud Etienne & Cie. C. Bidois 422

Stone-cutting and Quarrying by "Wire 424

The Theory of Jointed Bow-Girders. E.A.Werner 426

Experiments on a New Form of Strut. C. L. Strobel 428

Highway Bridges of Iron and Steel 430

Inspection and Maintenance of Railway Structures 432

The Garahit Viaduct. G. Eiffel 434

The Bridge over the Po at Casalmaggiore for the Parma-Brescia Railway . 438 Erection of the Large Girders of the Machinery Hall at the Paris Exhibi- tion of 1889. E. Henard 440

Reports of the French Delegates on the Proceedings of the Second Inter- national Inland Navigation Congress, held at Vienna in 1886 .... 441 Measurements of the Flow of the Elbe in Saxony, 1886 and 1887. A.

RiNGEL 444

Regulation of the Isar according to Wolf's Method. R. Iszkowski . . . 445 On the Improvement of the River Moldau at Prague, and the Construction

of a Port there 447

Special Plant for Blasting under Water at the Panama Canal Works. Max

DE Najjsouty 448

The Embankment of the Po at Turin. T. Prinetti 449

Jandin's Compressed-Air Dredger. M. Boulle 450

Renewal of the Water in the Hague Canals. M. R. von Pichlek . . . 450

The Cable Railway on the New York and Brooklyn Bridge. G. Leveeich 453

The New Harbour Works at La Rochelle 454

The Qualities of Potable Waters 457

Water-Supply in the Kingdom of Wurtemberg. J. R 458

Facts in Relation to Friction, Waste and Loss of Water in Mains. C. B.

Brush 459

The East Orange Sewage-Works 4G0

On the Disinfecting Action of a Current of Superheated Steam. Prof M.

Gruber . . 4G1

Elucidation of the Disinfccting-Power of Steam. A. Walz .... 462

Comparative Trials of various Gas-Burners. S. Lamansky 462

Wilmsmann's Smoke-Consuming Furnace. — Seiler 464

Raising the steamer " Forndale," sunk in the Entrance Channel of tlie Port

of St. Nazairc. • — Kerviler and — Preverez 465

VI CONTENTS.

TAGK

Consolidation of Earthworks on the Railway from Gicn to Atixcrre.

— Lethier and — Joyan 466

The Laon Steep-Gradient Railway. A. Braj^cher 468

Cost Prices on Railways. G. Ricovr 469

Signalling-Apparatus on the St. Gothard Railway. — Cox 470

Diminution of Earth Temperature in Deep Mines 471

An Apparatus for Measuring Earth-pressure Underground 471

Differences of Level in the Mines of Austria and Hungary. F. R. M. von

Friese 472

On the Relations between Seismic and Atmospheric Disturbances and the

Disengagement of Fire-Damp. G. Chesnau 473

Shaft-Sinking by Haase's Method 475

Cast-iron Tubbing for Lining Levels 476

A New Modification of the Bessemer Process 477

Electrolytic Reduction of Antimony from Ores. W. Borchees .... 478

Electrolytic Coi^per-Refining in Hungary. A. Soltz 479

The Smelting of Gold and Silver Ores in Eastern Hungary and

Transylvania. Dr. Schnabel 480

A Winding-Engine with Spiral Balance-Drum. K. Habermann and J. von

Hauer 484

On the Beer System of Wire Ropeways. C. Raovlt 485

Desrozier's New Disk-Dynamo. E. Meylan 488

Gadot Accumulators, pattern 1888. J. Laffaegue 489

Account of a Series of Experiments made on Hessner's Cell. W. Chukoloff 489

On the Measurement of the Resistance of Submarine Cables. A. Rovillard 490

Philippart's Electrical Tramcars in Paris 492

On the Telephone-Equation. C. L. Madsen 493

The Teleplione Line between Paris and Marseilles 494

The Regulation of Arc Lamps. E. Hospitalier 495

The Electric Lighting of the City of Geneva. R. Chavannes .... 496

Self-Regulating Electric Search-Light. W. E. Fein 497

Electric-Light Installation on the Armour-clad Cruiser "Admiral

Nakimofi'." Lieut. Kolokoltzoff 498

On Siemens and Halske's Electric Winding-Engine at Neu Stassfurt . . 500 On the Connecting of Lightning-Conductors with Water- and Gas-Pii3es.

L. Weber 501

The Decomposition of Salt by Electrolysis. N. N. Beketoff .... 504

The Wimshurst Machine. E. Dieudonne 504

The Glaser Influence Machine 505

Danger-Indicator for the Prevention of Collisions at Sea. P. Marcillac . 506 Studies on the Gas-Thermometer, and Comparison of the Mercury-Thenno-

meter therewith. P. Chappuis 507

The Poisonous Action of Water-Gas. H. Schiller 508

Heat of Combustion of the Coal of the North of France. — Schevrer-

ELestner 509

Index 510

CONTENTS. Vll

COEEIGENDA.

Vol. Ixxxi., p. 378, line 9 from bottom, for "millimetres " re id " metres." „ .xciv., p. 299, line 3 from bottom, /or " 1869 " read " 1859 "; and /or " Elizabeth "

read " Eliza." „ „ p. 381, line 6, for « 72 • 56 " read " 74 -68." „ „ „ „ 11, /or « 84-94 " rmJ" 87 -06."

„ „ „ „ 12,7or "15-06" r^wi « 12-94."

THE

INSTITUTION

CIVIL ENGINEERS.

SESSION 1888-89.— PART I.

Sect. I.— MINUTES OF PROCEEDINGS.

13 November, 1888.

Sir GEOEGE B. BRUCE, President, in the Chair.

The Telford and Watt Medals, the George Stephenson Medal, the Telford Premiums, the Manby Premium, and the Miller Prizes, awarded for the Session 1887-88 (vol. xciv. pp. 152, 153) were presented by the President to the various recipients.

{Paper No. 2328.)

"Friction-Brake Dynamometers."

By William Worby Beaumont, M. Inst. C.E.

Although the friction-brake dynamometer can never give a scientifically accurate measurement of the rate of absorption of the work done upon it by any motor, it has been much used for that purpose, with results not sufficiently incorrect to affect materially the truth of calculations based upon them.

The indications of any apparatus depending on friction must be as variable as the causes of friction and the conditions affecting it, and hence the value of the measurements, obtained by means of a friction-brake dynamometer, must depend on the completeness with which these causes and conditions are taken into considera- tion or are eliminated.

In a friction-brake the causes of friction are similar to those which generally obtain in other applications of materials when brought into rubbing contact ; but the conditions are more variable, and are chiefly due to variation in pressure and in lubrication, both these being affected by variation in the rate of work-absorption and temperature.

From time to time the accuracy of the measurements of power obtained with friction-dynamometers has been questioned ; and this Paper is written with the object of placing such facts and considerations before the Institution as will elicit the views which are supported by practice among engineers.

[the INST. C.E. VOL. XCV.] B

2 BEATTMONT ON FEICTION-BEAKE DYNAMOMETERS. [Minutes of

Friction-brakes in nse, and recognized as satisfactory instruments for the measurement of power, or as the best available, are very- few; but as the friction-dynamometer is an instrument of great importance to engineers and to the users of steam and other motors, some of them will be described, with a view to comment on those questions of principle of construction and of working, which are primary to confidence in their indications.

The friction - brake dynamometer in nearly all its forms is essentially that devised by Prony, an outline diagram of which is shown in Fig. 1. Here the friction between the surface of a wheel A, rotated by the motor whose power is to be measured, and

Fig. 1

SS^S^^5SK5^^!^SSSSS

the wood blocks W above and below the wheel, is employed in maintaining a weight P, suspended at a point on a horizontal line level with the centre of the wheel. The point is at some distance above the lever B, by which the load is supported, and, in order that the distance E from the centre of the wheel to the vertical centre of the load may remain constant, the load is suspended from a quadrant whose radius is E. This brake is capable of sufficiently exact determinations of power, ranging from 5 to 200 HP., for most practical purposes, where more than a mere check upon calculations is required ; but it presents some of the elements of inaccuracy which pertain to the friction type of absorption-

Proceedings.] BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. 3

dynamometers, especially when used for the measurements of variable powers. For measuring the power of a motor, capable of running at a uniform speed with a constant load, the inaccuracy of its indications may be very small and often insignificant ; but when the power varies, the inertia of the lever B, and the work done in moving it and the load P, count without record against the motor. Moreover, a small load due to the lever, though balanced while the lever is horizontal, constitutes a load against the motor when the lever rises ; and the work done when this load is lifted is also neither recorded nor measured. With a truly circular wheel and uniform turning power, this brake gives very nearly acciirate results, provided the lubrication between the rubbing surfaces be

Fig. 2.

K5S!SS!5^E5!5SSSSSrSSS3H!553Sl

uniformly maintained ; but slight variations in this respect, due to variation in quantity, quality, and temperature of the lubricant and rubbing surfaces, make it difficult to keep the tension of the strap E in strict accord with the total friction necessary to main- tain the load at a constant level. The frequent change of this tension by the screws C for the purpose of meeting these variations introduces further inaccuracy.

A simple form of friction-brake, much used as a dynamometer in portable engine-building establishments, for testing engines by running them against a known load before sending them out, consists simply of a thin iron or steel strap, or a pair of straps, E, Fig. 2, to which are attached a number of blocks W of wood. Sometimes a leather strap is used. At H a hook is fastened to the straps for the suspension of the load P, and at S the ends of the

B 2

4 BEAUMONT ON FRICTION -BRAKE DYNAMOMETERS. [Minutes of

strap are connected by a right- and left-handed screw for the adjustment of the tension on the strap or pressure of the blocks W upon the wheel A, so as to obtain the necessary frictional grip to carry the otherwise unsupported load P. With a truly turned wheel A, and with uniform lubrication, this brake will run for hours with a variation of but a few inches in the level of the load P, if the engine under test is of good design and has a fly-wheel A of the weight many makers adopt. The variations that do occur from several causes are, however, sufficient' to make adjustment by the screw S necessarj^ sometimes frequently. The errors in estimating the work done by the engine, which result from these causes, are generally small ; but it is desirable to remove them, if possible, when very accurate tests are required. It has been with this object that devices for automatically varying the tension in the belt E with the variation in the total friction have been introduced by numerous experimenters ; and at an early date a brake-wheel with an internal water channel was used to avoid the mechanical difficulties which resulted from the heating of the brake-wheel and the variations due to the heating of the lubricant. With either of the brakes mentioned, the work in HP. done in supporting the load P in lbs. at the distance E in feet from the centre of the wheel A, making N revolutions in time T in minutes will be —

H P = (I^x2)X7rxPxN 33,000 X T

or taking C = circumference of circle of radius E, and V = velocity in feet per minute of circumference, then V = C N, and

HP= ^^^ ^^

P =

33,000 T ~ 33,000 T

33,000 HP _ 33,000 H P

C"N V

^ 33,000 HP ,-^ 33,000 HP ^=— Np-'^"^^ = — CP--

One form of automatically-adjusting or compensating brake,

suitable for small powers, is due to Mr. Deprez, and is shown at

Fig. 3. In this ^ a brake-wheel A is attached to a disk B, and to

the disk are pivoted levers EE. These levers are connected by a link

F and lever G loaded by a weight Q. The weight of the levers E E

is balanced by the quadrant-shaped counterpoise C, so that the (

' See " Guide pour I'essai des machiues li vapour," par J. Buehetti.

Proceedings.] BEAUMONT ON PRICTION-BEAKE DYNAMOMETERS. 5

weight P is the representative of the work done by the motor, the weight Q having no varying effect on the work done by the motor, or on the position of the load P, so long as the lever G remains horizontal; but as it is by the influence of Q and the lever G that the friction-blocks W are brought to bear upon the brake-wheel A, and as that influence decreases as G leaves the horizontal and approaches the vertical, the weight which the frictional grip will support decreases. Hence the brake becomes automatic in its adjustment of the friction, the descent of the weight P increasing the frictional grip, and correspondingly the lifting of the weight causes a decrease in the friction, so that the weight again falls, a

Fig. 3.

position of equilibrium being ultimately taken, in which load and friction are equal. It will be readily seen, that with any of the causes enumerated, variation of speed of the motor, or oscillation of the load P, the latter must be raised through some distance, or the lever G must pass through a considerable angle before the frictional grip of the brake-blocks is modified. The work of lifting the load being lost work is an objection to this brake. The system of compensation does not, however, contain any element that will affect the load actually carried by the motor, and the only objec- tion to the brake as a dynamometer for small powers is that the compensation takes place so slowly that any of the errors attaching to the simple brake (Fig. 1) remain, and adjustment must fre-

6 BEAUMONT ON FEICTION-BEAKE DTNAMOMETEKS. [Minutes of

quently be made so that the point of suspension of Q remains truly at the centre of the brake-wheel.

A simple form of self-adjusting brake-dynamometer, due to Mr. J. Imray, M. Inst. C.E., is shown by Fig. 4. A quadrant Q, balanced by a weight w, receives the brake-strap carrying P. A smaller weight p is attached to the other end of the brake-strap. The compensating action is due to the increase in total friction which accompanies increase of circumferential surfaces in contact. If the motor lifts the weight P, the frictional grip is lessened as the arc BH is lessened, and P returns or finds a position of equilibrium.

Fig. 4.

What is spoken of as a modification of this brake consists ^ in using a spring-balance instead of the small weight p; and in dispensing with the quadrant Q, the wood blocks being continued on the P side of the band, and the lower end of the sjjring being fixed. This arrangement, also shown on Fig. 4, acts much in the same way as Mr. Imray's brake, but the range of movement necessary for self-adjustment will be less as the pull of the spring

' Variously attributed to Navier and to Messrs. Easton and Anderson, and to one of Messrs. Eansomes, Sims and Head's managers ; but due, the Author be- lieves, to Mr. H. A. Byng of the latter firm, who used it first in Paris in 1867 ; at Brussels in 1SG8 ; at Santiago, Chile, in ISGii ; and at Cairo in 1874, where it ■sas been by Sir Frederick BramwelL

Proceedings.] BEAUMONT ON FEICTION-BEAKE DYNAMOMETEKS. 7

is lessened or increased with any movement, while the weight p in the other case remains the same, and considerable movement must take place so as to make a sensible difference in the total friction due to difference of arc in contact. The spring arrange- ment is more accurate, as every rise or fall of the weight P decreases or increases the tension on the spring, and thus a very small range of movement of P is sufficient to satisfy small variations in turning moment or frictional grip.

One of the best known forms of friction-brake dynamometers, fitted with a compensating device, is that designed by Mr. C. E. Amos and Mr. Appold, and is the form used for the larger powers by the Eoyal Agricultural Society. It is similar to that shown by Fig. 2, but, besides the hand-adjusting screw S, it is provided

Fig. 5.

pj-n

with a compensating lever K as shown in Fig. 5, by means of which the rise or fall of the load P is supposed to be attended with a decrease, or increase, in tension on the brake-strap, so that a position of equilibrium is automatically attained without causing inaccuracy in the indications. With a given tension in the belt, and with the load P carried so that its point of suspension H is opposite the pointer T, the lever K takes a vertical position ; but as soon as the load P is lifted, as in Fig. 6, the lever pivoted at X moves with, and virtually increases the length of, the belt, and thus slackens it, allowing the load again to descend. If, on the other hand, the total friction decreases and is insufficient to carry the load in its normal position, the descent of the load carries the compensating lever to a position such as is shown by Fig. 7, thus

8

BEAUMONT ON FKICTION-BRAKE DYNAMOMETEBS. [Minutes of

tightening the belt and increasing the frictional grip until the conditions are again such as will enable the load to reassume the medial position. If the change in the position of the point of

Fig

Suspension of the load has been duo to a temporary cause, this automatic action may restore the balance without further adjust- ment ; butllif the departure from the medial position is not small,

Fig. 7.

then the adjustment by the hand-screw S must be resorted to. It will be seen that the compensating action cannot come into play except by the rise or fall of the weight from its proper position,

Proceedings.] BEAUMONT ON FRICTION-BBAKE DYNAMOMETERS.

9

and hence the value of the device is confined to its power of limiting that rise and fall. In practice, generally speaking, the adjustment required by means of the screw S is as necessary with the compensating lever as without it, and its value may therefore be questioned for this reason alone. A further reason, however, for questioning the value of this compensating lever is that it introduces an element of error, which may be small or consider- able, but which must be variable almost directly in proportion to the extent to which the lever comes into play. With a heavily-loaded brake the error must exist, and with a lubricant

Fig. 8.

Fig. 9

Vi"-^ a

which materially lessens the total friction of the wood blocks upon the wheel, it must amount to a considerable part of the whole indicated power, unless adequate allowance is made for it. This allowance has seldom been made, since it has not been necessary with the conditions under which these friction-brakes have gener- ally been used, namely, with the brake-strap slack as shown exaggerated at N, Fig. 9.

When the lever does act in fulfilling its purpose, the tension in the brake-strap causes it to press against the pivot X with a pressure proportionate to that tension. This will be more fully referred to hereafter.

10

BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

In a better form of compensating brake designed by Mr. Balk, and used by Messrs. Eansomes, Sims and Jefferies, Fig. 10, the compensating lever is outside, instead of ^dthin, the circumference of the brake-strap. It is connected at B and at C to the ends of the strap, and a fixed pin F passes through a slot at the outer end of the lever. To the latter is suspended a scale-pan and a weight p sufficient to keep the lever floating with the pin F midway between the two sides of the slot. This weight jp becomes a measure of the tension upon the belt, at least at the parts to which it is attached ; but it must be varied with change of condition of the brake-blocks, the lubricant and the temperature of the blocks and wheel, and as it must be taken as acting at the radius O F in favour of the weight P, these variations become troublesome by virtually making P a variant.

Fig. 10.

Fig. 11 shows the larger friction-brake dynamometer used by Messrs. Eansomes, Sims and Jefieries, and is precisely similar to the first one of the kind which Mr. Balk perfected. The Author is in- debted to Mr. John Jefferies for this drawing, and for the following tabulated particulars showing the gross load, the weight on the end of the tension or compensating lever, and the variation of the latter with different speeds and powers. These particulars have been taken from the records of numerous experiments made with Messrs. Eansomes' engines by Mr. H. A. Byng. The wheel is 6 feet in diameter, and 1 foot in width, the load being suspended from flat steel springs or tapes, forming a tangent to a circle exactly

Proceedings.] BEAUMONT ON FKICTION-BRAKE DYNAMOMETERS.

11

20 feet in circumference. A counter is attached to the side of the frame, and thrown in or out of gear by the clutch handle H. The great advantage of this brake over all others is that the experi- menter can always ascertain the actual load, although the tension- lever acts as a compensating lever. There is considerable variation in the scale-weights with the same gross load. This, Mr. Byng explains, is due to the heating of the brake-wheel. When much heated, more grease has to be used, and in consequence the co- efficient of friction is reduced, and the weights in the scale have to be augmented to give more tension in the brake-straps. The figures appear to show that, the larger the number of revolutions, the smaller the tension for a given total friction, and this infer-

FlG. 11.

7onciRCumFff£MCc

enco is supported by experience. Mr. Byng has found that with the higher speeds less weight is required on the scale for a given load, or, in other words, the higher the speed the less the necessary tension in the brake-belt. By means of the Balk brake the tension at the ends of the brake-belt is measured directly, and for accurate trials a higher speed is preferred with less gross load and with a scale- weight below 7 lbs.

The wood blocks, beech or plane-tree, of the brake above referred to, have been in use a long time, and are now semi-charred and saturated with grease ; the surface next the wheel has been several times coated with black-lead, and some of the blocks have become

12

BEAUMONT ON FKICTION -BRAKE DYNAMOMETEKS. [MinutoB of

Table I. — The Balk Friction-Brake Dynamometer.

The folio-wing Table shows the gross load and counterbalance weights used in trials, at various speeds and HP.

Gross Load.	On Scale.	Effec- tive Load.	Revolu- tions.	HP.	Gross Load.	On Scale.	Effec- tive Load.	Eevoln- tions.	HP.
Cvvt. qrs.lbs. 1 0 1	Lbs. 4	Lbs. 105	130	20	8-28,	Cwt 2	qrs.lbs. 1 14	Lbs. 17	Lbs. 232	134-42	18-90
0 3 22	2	104	152	10	9-40	2	1 22	11	252	125-20	19-12
1 1 8	6	136	128	50	10-59	2	2 20	20	260	123-50	19-46
1 0 24	5h	125	145	00	11-00	2	1 16	16	236	136-75	19-56
1 1 17	n	142	140	00	12 -Oo'	2	1 18	19	232	139-59	19-62
1 1 10	3	144	138	30	12-07	2	1 16	16	236	144-00 20-60 j
1 2 1	ej	156	131	25	12-40	2	1 8	12	246	150-00 21-45 1
1 2 14	7	168	136	92	13-94	2	1 16	16	236	160-00 22-93
1 2 4	3§	165	139	53	13-95	2	1 16	16	236	155-00 22-17
1 2 18	2	182	134	74	14-86	2	1 12	14	236	160-00 23-00
2 0 10	16	202	127	75	15-64	2	3 10	19	280	135-79 23 04
2 0 10	16	202	131	30	16-07	2	3 14	24	274	142-45 24-00
2 0 0	14	202	134	00	16-40	3	1 0	25	314	136-00 25-88
2 0 0	11	202	137	00	16-78	3	1 0	25	314	140 -40| 26-72
2 0 12	17	202	135	25	16-55	3	0 14	18	314	146-00 27-78
2 0 0	11	202	136	88	16-75	3	0 14	11	328	140-89 28-00
2 0 8	15	202	145	75	17-84	3	2 21	14	385	141-50 33 00
1 3 24	9	202	148	09	18-20	3	3 18	25	388	151 -22' 35-56
2 0 0	9^	205	146	22	18-16 1

almost metallic on the surface. In starting for a set of trials, the wheel and the blocks are thoroughly cleansed and left free of adherent grease. As soon as the wheel gets warm, their saturated condition supplies the necessary lubrication. For loads up to 20 HP. there is no difficulty, the Author is informed, in running this brake all day with perhaps two alterations of the scale weights, and by applying a very small quantity of cold tallow to the wheel and rubbing off the excess immediately with dry waste. Mr. Byng is of opinion that the blocks should be of metal, say Babbit, properly ground to fit the wheel and jointed together. With such a brake-strap he thinks some definite relation between load, tension and velocity might be obtained.

A water-cooled brake used by Messrs. Eichard Garrett and Sons,

Proceedings.] BEAUMONT ON FEICTION-BRAKE DYNAMOMETERS. 13

Leiston, is shown by Figs. 12, 13 and 14. The wheel is 5 feet in Fig. 12. Fig. 13.

Fig

diameter, and 11 inches wide between the flanges within which the wood blocks run. An annular trough is formed by internally projecting flanges 3 • 5 inches in depth. The straight form of the Appold compensating lever is employed, though under conditions which seldom, if ever, bring it into play sufiiciently to affect materially the accuracy of the indication of the brake. The Author is indebted to Mr. Frank Garrett, M. Inst. C.E., for some interesting particulars concerning the work- ing of this dynamometer. No lubricant is now used; but the blocks, which are of beech, are probably thoroughly saturated with years of previous use on a wheel not cooled by water. It has been ascertained that the brake-wheel with the water con- tained in it absorbs about 0*75 indicated HP. at 180 revolutions per minute. The circumference of the load circle is 1 7 • 5 feet, and the maximum load

M

BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

carried 418 lbs. Three experimental runs with this brake were recently made by Mr. Garrett with a view to obtain figures as to the quantity of water used, its rise in temperature, and the relation between the work mechanically done and thermo- dynamically accounted for by the heat conveyed to the water. Before starting the engine-water was put into the wheel to the full depth of the annular trough ; this water was weighed and called A. As soon as the engine was started water was added, which was also weighed. This second quantity, occupying a depth of 2 inches in the internal annular trough, w^as called B. The supply is kept up under ordinary circumstances by a pipe con- stantly passing a small stream into the wheel. During the experi- ments referred to, however, the water was all weighed and added by hand, the depth of 2 inches being as far as practicable main- tained. This quantity may be called C. When the engine is stopped at the end of the run all the water falls out except a quantity equal to the first quantity mentioned, namely, that which fills the lower part of the wheel to the full depth of the trough. During the runs the water was carefully maintained at the depth of 2 inches, so that the quantity added during the run may be taken as evaporated. The engine ran at 130 revolutions per minute.

Table II. — ExPERniEXTS avith Garrett's "Watee-Cooled Brake.

Trial No

Time .... minutes

Brake HP.

Water A

„ B

„ C evaporated . Temperature t at start

T at end . . Heat-units expended, raising!

A+B-f-CtoT . . . ./ Heat-units expended in eva-l

porating C from T . . . / Total heat-units .... Mechanical equivalent "i

foot-lbs./

HP. minutes

HP

Per cent, of brake HP.

1 150 31 24 lbs. 180 „ 108 „ 54° Fahr. 174° „

37,440

107,244

144,684

111,696,048

3,384-4 22-53 72-6

2

155

28

24 IbB.

180 „

90 „

58° Fahr.

162° „

30,576

90,090

120,666

93,054,152

2,819-8 18-29 65-0

3

154 18 24 Iba. 180 „ 64 „ 52° Fahr. 148° „

25,728

64,704

90,432

69,813,504

2,115-5 13-73

76-27

When the loss by radiation of the frictionally developed heat is considered, the heat represented by the water raised in tempera- ture and evaporated, a mean of 71 • 3 per cent., must be admitted to be a large proportion even if some deduction be made for slight loss of water by spray.

Proceedings.] BEAUMONT ON FEICTION-BRAKE DYNAMOMETERS. 15

The Author has since made an experimental run with this brake. The figures obtained show that the evaporation takes place at a mean temperature of about 160^, and that of the mechanical work done a mean of about 72 per cent, is accounted for by the water heated and evaporated in the wheel-trough.

A form of brake used by Messrs. J. and H. McLaren, of Leeds, as devised by Mr. Druitt Halpin, M. Inst. C.E., is shown in Figs. 15 and 16. The engine upon which it was used gave off about 20 HP.

Fig. 16.

Fig. 15.

The brake-wheel is 5 feet in diameter, and 7 inches in width. The radius of the load circle is 32 • 19 inches. Water is constantly supplied to the trough, and constantly taken away by a scoop pipe. In 1886 Mr. Halpin proposed to use the Appold compensating lever, but has since abandoned it, as introducing an error varying with the coefficient of friction of the brake-blocks on the wheel.

As a brake which within a limited range of power might be considered automatic, and without the errors introduced by some of the forms of compensating action, the form shown at Fig. 17

16

BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

may be suggested. In this the ends of the "brake-strap E E' instead of being connected are led over the rollers E E, which roll on a path T T, and are connected by links to a lever L pivoted and connected to the weight P. In the ordinary working of such a brake the lower blocks would touch the wheel with small pres- sure, and the weights p p' would be adapted to the total friction required. The whole of the parts would be put in balance by the movable weight w. With variations in the turning moment or the lubricant P would rise and fall, but the rise and fall would be of short range as the separation or approach of the rollers E E would rapidly change the extent of surface in contact.

i^sSMiSM^^Wsrf

Special reference should be made to a very simple form of friction- brake dynamometer, first proposed by Professor James Thomson, consisting of a cord or rope passed over the upper half-circumference once, or taking one complete turn round a smooth wheel, the one end carrying a weight P, and the other attached to a sj)ring-balance in a manner similar to that shown in Fig. 15, the rope bearing directly upon the wheel without the intervention of blocks. This brake works exceedingly well for small powers, and there seems to be no reason against its use for large powers if a number of separate roi:)es be used. All the ropes might be attached to a cross-head at either end, from which the weight P would hang, and to which the spring-balance could be attached. The pull on

Proceedings.] BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. 17

the spring-Lalance would be deducted from the weight P for the actual load carried, as in the case of Figs. 14 or 15. During the recent trials of gas- and steam-engines under the auspices of the Society of Arts, Professor Kennedy, M. Inst. C.E., used two hemp ropes, of about If inch circumference, passed once round the fly- wheel of the engine under trial, and these were found to be quite sufficient to absorb 20 HP., a water-trough wheel being employed and very little unguent.

Proportions and Dimensions of Brakes.

In English practice there has not been much diversity in the proportions given to brakes of these several forms for the absorp- tion of a given power or the measurement of a given quantity of work ; but, in the United States, a Prony brake has been used of dimensions supi^osed to have been sufficient to enable it to absorb about 500 HP. It was made with a wheel 5 feet in diameter, having a rim of trough-shaped section supplied with constantly renewed cooling water.

A comparison may now be made by an examination of the pro- portions of the brakes described, and by reference to the results of their working.

The brake used by several engineering firms is of the type Fig. 5, and of the size used by the Eoyal Agricultural Society at Newcastle in 1887, for a maximiim of about 20 HP. at 130 revolu- tions per minute, and fitted with a compensating device, Figs. 5, 6, and 7. The wheel is 5 feet in diameter and 7 inches in width, and the radius E of the point of suspension of the load about 2 feet 9 inches, C being 23-75 lbs. With a load P=270 lbs., and the number of revolutions per minute N = 150, this brake gives 21 • 17 HP. The lubricant used for the brake-blocks was tallow and water.

The similar brake used by Messrs. McLaren gave HP. = 19-1 with N = 148 • 5 and E = 2 feet 9 • 38 inches ; and another of similar dimensions. Fig. 12, but with water-cooled wheel, also used by Messrs. McLaren, gave 20-2 HP. with P = 271 lbs., N = 145-7 and E = 2 feet 8-19 inches ; but it can be used for much higher powers. Tallow was the lubricant employed.

The water-cooled brake of Messrs. Garrett, for powers up to a maximum of 40 HP., has a rim of the section shown in Fig. 14, capable of taking friction blocks of about 10-5 inches in width, the width between the exterior flanges of the wheel being 11 inches. With this brake, E =2*8 feet, maximum load P = 418 lbs., and C = 17-5 lbs., HP.= 40 with N = 180, Mr. Garrett finds that the

[the INST. C.E. VOL. XCV.] 0

18 BEAUMONT ON FRICTION-BKAKE DYNAMOMETEES. [TMinutes of

brake runs most smoothly at 180 revolutions per minute; but, as it is connected by universal joints to the engine to be tested, it must run at the speed of the engine. Taking X = 150 for this brake with the same load P = 418, the HP. = 33-4.

The Eoval Agricultural Society has a brake with three wheels of the diameter and width of the single brake already mentioned, and this with X = 150 is intended to measure a maximum of 100 HP., or about 33 • 3 HP. per wheel.

As a convenient means of comparing the relative capacity of these brakes, iudging by the amount of Avork for which they have been

WV

designed, or to which they have been put, a coefficient K =

may l)e employed, W being the width of the wheel in inches, and y = the velocity of the periphery of the wheel in feet per minute. This gives for the

Eoyal Agricultiiral Society's single brake K = 824

„ „ „ treble „ K = 495

Garrett's water-cooled brake . . K = 740

Eansomes', Balk's brake . . . K = 1,020

or a mean of 8G0, omitting the Eoyal Agricultural Society's treble brake.

Compared in this same manner the Prony brake described by Professor E. H. Thurston,^ as devised for measuring a maximum of 540 HP., the wheel being 5 feet in diameter and 2 feet wide, and N = 100, gives K = only 75. This brake is stated to have been freely luljricated with beef tallow, plumbago and lard oil, and, although designed for a maximum of 540 HP., it does not appear to have worked above 200 HP., and at this power K = 188.

The Eoyal Agricultural Society's brake, the similar one of Messrs. McLaren, and that of Messrs. Garrett, are all fitted with compensating-levers. Without this arrangement there is no doubt that, although the brakes are not cooled with water, they would work with accuracy to higher powers, and would probably give satisfactory results with a constant K = 850, although, for long- continued runs, K = 900 to K = 950 would be better.

Even with the high coefficient K = 1,020 obtained from Messrs. Eansomes', Sims and Jefferies' brake with maximum load, the wheel heats sufficiently to make much lubrication necessary during long runs. This is shown by the variation in the scale weights with the same gross load.

It would appear that the water-cooled brake may be made to

> Journal of the Franklin Institute, April 1886. 3rd Series. Vol. xci. p. 290.

Proceedings.] BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. 19

work with a very much lower coefficient ; probably a coefficient of 450 would be high enough for a water-cooled brake if not fitted with the compensating-levers.

The water-cooled brake described by Professor Thurston provides an exceptional case ; but, although this brake was designed for 540 HP. as a maximum, it does not appear to have been employed for more than 174*5 HP., and it is questionable whether its performance at the proposed maximum would have been at all satisfactory, or even at half the maximum, especially when it is remembered that only two bands 3 inches wide were used to carry the wood blocks, which though 24 inches in length were only 2 • 5 inches in thickness. The flexure of the blocks, under the maxi- mum tension of the belts, would have been sufficient to have made the pressure immediately under the belts excessive ; while the pressure between the belts and at the ends of the blocks would be very small, and the effect of this on the wood and lubricant coTild not have conduced to good running. Even with the load carried. Professor Thurston remarks that : " the friction between the brake-blocks and the face of the pulley was reduced to a mini- mum by effective lubrication." This seems to show that the brake was not strong enough for the power measured, for it is hardly a desirable thing to have to reduce the friction to a minimum. Taking 180 HP. as the greatest for which this brake was used, then the coefficient K = 209.

The determination of the diameter and width of a brake-wheel for a given power is often affected by prior fixed circumstances and conditions, but an appeal to experience leads to the conclusion that a somewhat greater width of wheel than has commonly been used for the higher powers is advisable. Taking the Eoyal Agri- cultural Society's single-wheel brake, for width, W = -7^^ =

824 X 20 — — —- = 7 inches, C being the circumference of the wheel ; but

if the compensating-levers were not used, a rather larger power might be measured on this brake, and using the proposed constant for long runs, K = 900, then with a diameter of wheel remaining the same, i.e., 5 feet, W would according to the above = 7 • 6 inches. For such powers, however, it may be questioned whether a wheel of lesser diameter and greater width would not be preferable, although the pressure upon the wood blocks would have to be greater. For 20 HP. a wheel 3 feet G inches in diameter would require a width of 10*9 inches.

For a water-cooled brake the proportions adopted in this respect

c 2

20 BEAUMONT ON FRICTION-BEAKE DYNAMOMETERS. [Minutes of

by Messrs. Eansomes', Sims and JefFeries for their brake of Balk's design, or by Messrs. Garrett and Sons for their brake as described, would seem to be satisfactory.

For several purposes in connection with these brakes, it is necessary to consider the tension in the brake-strap or straps, or belt, and the pressure upon the wood-blocks.

For a given power the total frictional resistance F at the face of the wheel must be proportional to the velocity V of the periphery

of the wheel of diameter D in feet, or F = =- — =-^ — =

HP. X 33,000 w T rp ,a . , . • X. 1 1 ^, , and taking 1, the maximum tension m the brake

strap,

t = the minimum tension.

a = the fraction of the circumference of the wheel embraced

by the brake-blocks, / = the coefficient of friction, then following Rankine,^ F = T — / and the ratio T : i is the number whose common

logarithm is 2-729/«,

T

or - = i0-"'25/« = n.

t

This may be put

T Log - = 2-729/rt = n,

and T = F^l + --^\ or T = F

T - t/ n-1

In these brakes a = 1, and by way of illustration/ may be taken as 0-2. Then, in the case of the large Prony brake already mentioned as described by Professor Thurston —

T Log - = 2-729/a = 0-5458,

and n = 3*5.

For the maximum HP. for which the lirake was calculated

540 X 33,000 ,.„,.„ ^ = - 1,570-8 = 1''^^^ ^^^• T = 11,345 X 1-4 = 15,833 lbs.

^ = ^-^'^^ = 3,241 lbs.; 6' o

but taking the maximum HP. at which the brake was worked = 200 HP., T = 5,902 lbs. and t = 1,204 lbs.

* " Machinery and Millwork," p. 403.

Proceedings.] BEAUMONT ON FRICTION-BKAKE DYNAMOMETERS. 21

If in the same way the tensions on the Eoyal Agricultural Society's single brake are taken as used at Newcastle last year with about 20 HP., F =^ 270 lbs., T = 378 lbs., and t = 77 lbs. If /is taken as 0'3, then with the same brake-load T = 333 "7 lbs., and t = 51-6 lbs.

For a comparative approximation to the maximum pressure p per square foot of wood block upon the wheel surface, W being the width of the wheel in feet, p may be taken

T + T = Dp W, or T = - ^^^

and p = -^^.

If this be allowed as giving figures which will permit the pressures on the wood blocks of the several brakes to be compared, or as an approximation to the pressure, then the greatest pressure per square foot of surface of Professor Thurston's Prony brake, if it had been used as proposed for 540 HP., would have been, assuming the coefficient / = 0*2,

15,883 X 2 i> = -5-;^^- 3,176 lbs.

Assuming the blocks to be only four-fifths of the total area of wheel-face, then p - 3,811 lbs. At 200 HP., which is more than the actual power exerted, p = 1,180 lbs.

In the same way, and with the same coefficient, the greatest pressure per square foot on the blocks of the Eoyal Agricultural Society's brake at Newcastle would hep = 260 lbs. with the blocks close, or 312 lbs. with the blocks covering 0*8 of the surface of the wheel.

With the brake used by Messrs. McLaren, it would be the same if the same coefficient obtained.

With Messrs. Garretts' brake, p = 269 lbs., or with the blocks covering 0*8 of the surface p = 323 lbs.

With Messrs. Eansomes', Sims and Jefferies' brake, the wheel of which is 6 feet in diameter and 1 foot wide, and where the wood blocks are close together. Fig. 11, p = 110 lbs.

This is only intended as a rough approximation to the greatest pressure per square foot of rubbing-surface of wood, biit it affords a means of comparison of the pressi;re with the different brakes. The mean pressures are probably somewhat less than half those given. From these figures the advantages of width of wheel will be seen. The pressure for a given load will, of course, be less for

22 BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

a given power absorption with increase in the diameter of the wheeL The limit of velocity has, however, here to be considered, especially with the wheels not cooled with water by the aid of an internal trough. What that limit is cannot now be said, but it is noteworthy that Messrs. Garrett find their brake runs more smoothly at 180 revolutions per minute than at any lower speed, presumably for the same power.

The Appold Friction-Brake Dynamometer.

It is necessary now to consider the eifect of the introduction of the Appold compensating-lever, at least to do so sufficiently to find whether it is a desirable feature or not.

In illustration of this point, it may be conceded that, if the friction on the one hand be conceived to be almost infijiitely small, the load P would have to be infinitely great, and the reaction at X would also be infinite. If, on the other hand, the frictional grip be increased so as to approach adhesion, then with any load P that could be put upon the brake, while it still acted as a dynamometer, the reaction at the pivot at the end X of the compensating lever may be infinitely small, supposing the lever to come into play. It may therefore be seen that the reaction at X for any given brake and given power, may vary with the frictional grip, as affected by lubri- cation, from zero to any maximum. The greater the difference between what may be called the capacity of the brake, and the power of the motor, the less will a given reduction in the friction affect the compensating-lever, because, with the belt ends working almost slack in their connections, the greater may be the increase in the tension in the belt, before it becomes sufficient to make the reaction at X a quantity sufficiently appreciable to affect the indications.

It may, however, be easily seen that under the ordinary working conditions, that is to say, with the size of engine usually tested on the brakes, such as those used by the Eoyal Agricultural Society, the effect of the lever on the accuracy of the registration may be very small.

If, for instance, the lever be of the proportions shown in Fig. 8 (p. 9), and the tension on the two ends A of the strap be assumed to be 100, then the reaction at X will be equal to the difference of the reactions at A and B, or about 11.

This will act in helping to lift the weight, its effective help in this direction being proportional to the relation between the distance from the centre of the wheel to the point X, and the distance from the centre of the wheel to the point of suspension

Proceedings.] BEAUMONT ON FKICTION-BRAKE DYNAMOMETERS. 23

of the weight, and, if this relation be as 1 to 2, then the reaction tending to raise the weight will be 5-5, and may be, with the ordinary relations between the weight P and the tensions in the ends of the brake-strap, an entirely negligible quantity, and of practically no numerical effect on the HP. indicated by the brake. If, however, a brake is used in such a way that the coefficient of friction is very small, and anything like the al)ove assumed tension in the strap-ends be allowed to arise, then the error may become an important quantity.

With the brake, for instance, of the type shown by Fig. 5, as used by the Eoyal Agricultural Society at Newcastle, with an engine exerting 20 HP., and a weight P of about 270 lbs., if the tension at the ends of the brake-strap were as much as 100 lbs., and the re- action as much as 1 1 lbs., reducing the effective weight to 204 • 5 lbs., the difference, 5 • 5, is 2 i^er cent, of the whole. Neglecting the effect of the compensating lever, and taking the gross load on the brake as 270 lbs., the circumference of the load-circle being 17-25 feet, and the revolutions 135 per minute, the engine would apparently exert 19-05 HP. If, however, the tensions at the ends of the strap, where connected to the lever, be, to take an extreme example, as much as 100 lbs., and assuming the reaction to be as above, 11 lbs., and the gross load thus reduced by 5-5 lbs., or to 264-5 lbs., then the engine would have exerted only 18 -GG HP., the difierence being 0-39 HP. This with an engine using 30 lbs. of steam per brake HP. per hour is equal to 11-7 lbs. of feed- water per hour, or, with a boiler evaporating 10 lbs. of water per 1 lb. of coal, it represents 1-17 lb. of coal per hour, or 3-8 lbs. in a run of four hours of an engine exerting 19 brake IIP.

Under the same circumstances and assumptions, but with the engine running at 150 revolutions, the HP. would be 21 -IG, neglecting the effect of the lever, and 20-73 when correction is made for it, making a difference of 0 - 43 HP.

If, however, the tension of the brake-strai) be taken, as calculated by the expression given on p. 22, and an assumed coefficient / = 0-2, there is obtained by approximation an error of 0-34 HP. instead of 0-43 HP. In the same way, for / = 0-3 the result is an error of 0-23 HP., i.e., 20-83 HP. instead of 21-16 HP.

It is, however, an objection that these calculations proceed on the assumption that the relation between the gross load P, and the tension T in the belt, remains the same for the velocity concerned, i.e., 2,355 feet per minute at the periphery of the wheel, as for the low velocity at which the l>rukc would just slip, when tightened up on the fixed wheel.

24 BEAUMONT ON FRICTION-BBAKE DYNAMOMETERS. [Minutes of

If the belt of blocks is set so that it will move at a low velocity under the influence of a weight suspended, as in Fig. 2, when the wheel is fixed, the tension in the belt and the load that will over- come the friction of rest being thus ascertained, a coeificient of friction for that speed may be at once deduced. It would, however, be useless to apply this coefficient to any calculations regarding the brake when at work under ordinary conditions, inasmuch as the coefficient so found will not apply at the high velocity of wheel-periphery commonly to be dealt with. This is shown by the facts given on p. 11. The tension in the belt, as controlled by the adjusting-screw S, miist be miich greater to support a given weight when the wheel is running than when the wheel is at rest. The velocity very materially affects the friction, and no demonstration of the dynamics of the compensating-lever of the Appold brake is possible without taking these facts into consideration. To be able to do this, however, experimental inquiry is necessary.

Some useful information on this point was, however, obtained by Messrs. J. and H. McLaren, who attached a spring-balance to the upper end X of the compensating-lever of the brake, which they made of the same dimensions as that of the Eoyal Agricultural Society. By this means they ascertained the varia- tion of the reaction at X with variation of velocity of wheel- periphery, and of lubrication of the wood-blocks. They found that the pull on the spring-balance increased as the speed of the engine fell, when the steam-pressure decreased during the clinkering of the grate. This is precisely what would be expected, from what has already been said as to the greater tension in the brake- strap with a very low velocity, and it is clear that with the falling speed of the engine the weight P descended a little and increased the tension in the brake-strap, and therefore the reaction at X. Their experiments conclusively show the direct relation between tension in the brake-strap and the lubrication. In some cases, when using tallow, they got 24 lbs. pull on the spring-balance at X, and when using water the pull was 258 lbs. When they got the 24 lbs. pull there was a considerable quantity of water getting in between the brake-blocks and the wheel, but in some of the tests when, to use their own words, they " fairly smothered the brake in tallow," so that the water could not get in between the blocks and the wheel-face, they measured 20 HP. with only 4 to 5 lbs. pull on the spring-balance. Their experiments lead them to con- clude that, with an internally cooled brake-wheel of the size employed, compensating-levers could be used without error up to 20 HP.

Proceedings.] BEAUMONT ON FRICTION-BKAKE DYNAMOMETERS.

25

So long as the Appold brake, like that of the Eoyal Agricul- tural Society, is not used for more than 15 HP., and is sufficiently, but still sparingly, lubricated with tallow or suet, the friction between the wood and iron is such that the weight of the brake- strap and blocks with the suspended load is sufficient, at the ordinary speeds of the engines tested, to carry the load without screwing up the belt so that there is more than a few lbs. tension at the compensating-lever. Under such conditions the lever does not affect results, and adjustment of the frictional grip and position at which the load is carried has to be made by the hand-screw S. The conditions are the same as, or very similar to,

Fig. 18.

those which would obtain if the brake were as shown by Fig. 2, that is to say, without compensating-lever, but with a belt so slack that the bottom blocks barely touch the wheel.

The following investigation on this subject,^ which has been kindly placed at the Author's disposal by the Consulting Engineers of the Eoyal Agricultural Society of England, supports the views herein expressed : —

Let W = load on brake-strap (Fig. 18) ;

T^ T4 = tensions just above and below the points of suspen- sion of W ;

' The Journal of the Royal Agricultural Society of England, pp. G72-78.

October 1888.

26 BEAUMONT ON FRICTION-BEAKE DYNAMOMETERS. [Minutes of

To T3 = tensions at two ends, C and D, of strap connected

to lower ends of compensating levers ; P = pull on upper ends of these levers ;

a, a' — radii of brake-strap and wheel respectively ; 0 D = c?, C D = 6, and 0 E = c ; Fj = friction of A B C, F2 = friction of A D, and F = total friction of brake-strap.

The portion A B C of the brake-strap is kept in equilibrium by the tensions Tj and T2 at its ends, the friction exerted on it by the wheel and by its own weight. This gives, taking moments about 0 : —

Ti a = T2 0 M 4- Fi a' + moment of weight of A B C. Considering the portion A D in the same way —

T3 0 N = T4 a -f F2 a' — moment of weight of A D.

Adding these two equations, then : —

Ti a + T3 0 N = T2 0 M + T4 a -h (Fi -f F^) a' -f- difference of moments of weight of A B C and A D.

But the moment of weight of A B C = moment of weight of A D because the strap is accurately balanced. Therefore (T^ - TJ a = T2 0 M - T3 0 N + F a. Now 0 M = 0 C sine 0 C M = (cZ-6) sine 0 C M,

and ON = 0D sine ODN = cZ sine ODN,

and Ti - T4 = W.

Substituting these values in the above equation : — ■ W a = T2 {d-l) sine 0 C M - T3 d sine 0 D N + Fa' ;

= d {T2 sine 0 C M - T3 sine 0 D N} - fcT., sine O C M -|- Fa'. But because the lever is in equilibriiim, considering the forces at right-angles to it : —

T2sineOCM = P + T3sineODN, or P = T2 sine 0 C M - T3 sine 0 D N ;

and by taking moments about D of all the forces acting on the levers,

D E X P = C D sine 0 C M T2

or ((Z-c) P = 6 T2 sine 0 C M.

Substituting these two relations in the above equation, then : —

Wa = d P - (f? - c) P + F a' ;

W« = c P 4- F a ;

or W-'^P=-F.

a a

Proceedings.] BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS.

27

that is to say, in order to get the effective value of W, the pull at E, diminished in the ratio of c to a, must be deducted.

This result agrees with the principle that the external forces acting on the system should balance. Since it is at rest, and the tensions are internal forces, the load, the pull at E, and the friction are the only external forces that have to be considered, as the point of support is in the centre of gravity of the strap. This gives at once, by taking moments about 0 : —

a W = c P + F a' ;

as before. Also because

P =

h T, sine 0 C M d — c

it follows that P depends upon the tension of the strap, and upon the proportion which C D bears to the whole length of the lever. Now the tension of the strap depends upon the lubrication ; hence, the more efficient that is, the greater will be the inaccuracy of the brake.

It is evideut that, in order to determine the probable amount of error in the Cardiif and Newcastle trials, the pull upon the upper ends of the levers must be ascertained when the brakes are running under exactly the same conditions as to power, speed, temperature of air and lubrication. When the lubrication is constant, the pull on the upper ends of the levers always bears a constant ratio to the load on the brake.

The following Table gives the results obtained by Messrs. McLaren, by means of the water-cooled brake, and by the Engineers to the Eoyal Agricultural Society with the same engine at Newcastle.

Table III
	Messrs. McLaren's Trial.	K. A. S.
	Halpin'a Brake.	Brake made like R. A. S.
	Newcastle.
Indicated HP	23-70	22-20	24-020
Brake HP	20-20	19-10	20-770
Coal per brake HP. jicr hour ....	2-11	2-14	2-267
Feed-water per brake HP. per hour	22-10	22-00	21-530
Mechanical efficiency	0-85	0-86	0-860

It will be seen that the figures agree very closely.

28 BEAUMONT ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

The Axithor has not here described more of the many forms of friction-brake dynamometers than are necessary to comprise most of those of a typical character. Eeference may be made to an essay,^ by Professor E. Brauer, on various other forms, includ- ing numerous devices for arriving at a compensating action, none of which, however, appear to be free from the objections herein referred to.

It may be suggested that a very useful dynamometer might be obtained by a combination of absorption and transmission dynamo- meters.

The Paper is accompanied by three sheets of drawings, from which the Figs, in the text have been reproduced.

I

' Zeitsclirift cles Vereincs deutsclier Ingenieure. Baud xxxii. Seite 5G 1888 ; aud Miuutes of Proceediugs lust. C.E., vol. Ixss. p. 266.

[Discussion.

Proceedings.] DISCUSSION ON FRICTION -BRAKE DYNAMOMETERS. 29

Discussion.

Mr. W. WoRBY Beaumont wislied to direct attention to a friction- Mr. Beaumont. brake dynamometer that had not been mentioned in the Paper. It had been designed, he believed, by Messrs. Ayrton and Perry. The wheel was grooved, and over it was placed a rope, of small diameter, which ran freely in the groove, bearing well upon the bottom of the groove. The small rope passed round the upper part of the wheel. At one end it carried a heavy load, and at the other end it was fastened to a piece of rope of much larger diameter. That rope did not fit in the groove so as to touch the bottom ; but the point of suspension of a smaller weight hung from it, when the rope was lifted sufficiently, would be at a greater distance from the centre of the wheel than the heavy load hung from the small part of the rope ; so that a difference in the value of the balancing weight was brought about by the rise and fall of the weight. Another point was that the rope of larger diameter gripped, or was gripped, more strongly than the smaller rope, as it did not touch the bottom of the groove ; it tended to jam between the two sides of the groove. He had not seen that brake in use, and if his descrijition of it were inaccurate perhaps Pro- fessors Ayrton and Perry would set him right. He wished further to state that the Paper had been written some time ago. If it had been written lately it would have been extended in various ways ; and, instead of the brief reference to rope dynamometers, a fuller description would have been given of them. When the Paper was written, early in the year, brakes with compensating levers were the subject of a great deal of controversy ; they were, in fact, a burning question, but since then the question had dropped, and therefore a great deal of what might have been of considerable interest then was much less so now.

Professor Archibald Barr said that the subject of friction- Professor Barr. brakes was an exceedingly important one for engineers, and was becoming more important every day, because the value of correct experiments on the economy of the steam-engine was now more appreciated than it had ever previously been. He would allude to the statement in the Paper, first, that the friction-brake could not give a scientifically accurate measurement of the amount of work being done by a motor. No doubt it was perfectly true that no instrument could give an absolutely correct record ; but he wished

30 DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

ofessor Barr. to express strongly his belief that the friction-brake dynamometer, properly constructed and used, would give a much more accurate determination of the amount of work being done by an engine than the steam-engine indicator would give of the work done by the steam in the cylinder; and, as between the two, he thought the dynamometer was the most important instrument that engineers possessed for the determination of what engines were capable of doing. In the second paragraph of the Paper the Author went on to speak of the variations of friction and the influence they had upon the accuracy of the brake results. In a properly designed brake the variations in friction should have no influence whatever upon the accuracy of the results. If the construction of the brake and the method of taking its indications were correct, the variations of friction were a matter of secondary importance, and need not be taken into account in any results deduced from the readings.

With regard to the different forms of dynamometer described in the Paper, the first principle to be observed in constructing a brake was to haA^e it as accurate as possible, and the second was to have it as simple as possible ; and he believed that, of all the arrangements shown in the figures, the simplest, and the one best calculated to give good results in most cases was that shown by Fig. 15. The still simpler brake on the same principle to which he would afterwards refer, in which the strap and blocks were replaced by a rope or ropes, could be used in many cases with advantage (Pig. 19). The Author had stated (p. 3) that the inertia of the lever of the Prony brake, and the work done in moving it and the load, counted without record against the motor. He would not enter into the question of the influence of the inertia of the load and that of other parts of the apparatus ; that would require a somewhat lengthy treatment ; but he might say that the eff'ect would, on the whole, be in favour of, or against, the motor, according to circumstances. If the weight moved up and down (as it did in all friction-brake dynamometers), the eff'ect upon the motor of such motion was not measured, in any case, by the work done in raising the weight. If the engine were made simply to wind a weight out of a mine, the whole of the work Avould be spent in raising the weight. In raising the weight in the friction-brake dynamometer, a certain amount of work was done, but neglecting inertia just that amount less was done against friction. If the engine were running at a uniform speed, and the weight went up and down slowly (or under other circumstances into which he need not then enter), the brake would give a correct result, independently of the distance the weight was lifted ; therefore that criticism which, if sound, would apply very

Proceedings.] DISCUSSION ON FKICTION-BRAKE DYNAMOMETERS. 31

strongly against some brakes, such as Thomson's (Fig. 21), did Professor Barr. not, he thought, hold.

The internal water-channel was an exceedingly valuable pro- vision ; but it was impossible to have a trough-brake constructed in many cases. He had recently had occasion to test a gas-engine, and had used the simple rope-brake. There was some trouble from the pulley overheating, and, when water was allowed to fall on the inner surface of the brake-pulley, some got on to the ropes, which

Fig. 19.

so reduced the friction, that the pull at the spring-balance altered suddenly from say 10 lbs. to 100 lljs., and thus prevented any trustworthy results being obtained. He overcame the difficulty completely by tying to the arms of the pulley a number of pieces of cloth, so that they lay inside the rim of the pulley, and keeping these soaked with water. In that way a trough-brake could virtually be improvised in a few minutes. At p. 7 the Author spoke of the arrangement with the spring-balance shown in Fig. 4, as being better than that with the weierht. Professor Barr did not think

32

DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. pHinutes of

rofessor Barr. that was the case as far as accuracy was concerned, because it was necessary in the one case to take constant readings of the spring- balance, and, as he had already pointed out, the rise and fall of the weights in the other case did not necessarily cause any error. If it were possible to get a simple, self-regulating brake, with a steady load at the tail end, he thought it would be better, and it would certainly be much more convenient to use, than a brake with a variable pull at that point. At p. 11 the Author stated that the brake shown by Fig. 11 had the great advantage over all others that the experimenter could always ascertain the actual load, although the tension lever acted as a compensating-lever. He did not think that that superiority could be claimed for the arrange- ment. It was a good brake, no doubt, but there were other brakes

Fig. 20.

which allowed the experimenter quite as well to determine the amount of pull at the adjusting point. He thought that the brake would be much improved if, instead of putting the lever half way round the wheel from the point of suspension of the load, the load were suspended by means of steel tapes, and the lever applied to the slackest part of the belt, just under the load. There would then be a much smaller force at the end of the lever, and, therefore, a much smaller variation of force than in the other arrangement. He was now constructing a brake on this principle, as shown by Fig. 20. With regard to some of the brakes described, such as those shown by Figs. 10 and 15 for example, he thought there was one point to which attention had not been sufficiently paid, namely, the thickness, or rather the thinness of the strap. In the case of such brakes, the strap used should be exceedingly thin ; it should.

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 33

in fact, be sometliing of the nature of a steel ribbon rather than Professor Barr.

a piece of stiff iron. With reference to the rope-brake mentioned

on p. 16, which the Author attributed to Professor James

Thomson, if it was due to one of the Thomsons, it was to Sir

William, who had used it before any one else, so far as he knew,

though not in quite the same form as that illustrated by Fig. 19.

The brake invented by Professor James Thomson was an entirely

different one (Fig. 21). If the friction was too great the loose pulley

was drawn round, unwinding a portion of the rolling cord from the

running pulley, and thus automatically adjusting the resistance.

Fig. 21.

There was no mention in the Paper of that brake, nor was there any mention of a brake which he thought was, though not the simplest, one of the best ever introduced. He referred to the modi- fication of Mr. Froude's brake, which had recently been adopted at the Owens College by Professor Reynolds. It worked with great nicety, with perfect regulation, and with no trouble or mess; it was a water-brake in which the work was absorbed by fluid friction instead of by solid friction. Another point, not mentioned in the Paper, was of much more importance than usually supj^osed, namely, the dashpot often introduced into brakes. The dashpot might

[the INST. C.E. VOL. XCV.] h

34 DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

rofessor Barr. effect a great deal of harm. It was sometimes said that the work done in it ought to be added to the work done against friction as calculated from the brake-load and speed. That, however, was a serious mistake, because the amount of work done in the dashpot was the force acting upon the dashpot piston, multiplied by the distance through which the dashpot piston moved ; but the effect upon the engine was the amount of force upon the dashpot piston multiplied by the distance through which the circuniference of the wheel, considered as having a radius equal to the arm at which the load acted, moved in the time, and that might be many thousand times as great as the amount of work done in the dash- pot itself. If the force on the dashpot piston were to vary in a certain way, and the engine to run at a uniform speed, the up and down motions of the dashpot would exactly counteract each other, so far as work done bj^ the engine was concerned ; but that would be best secured, not by the dashpot commonly used with water and a loose piston, but with one filled with a very viscid fluid, so arranged that the resistance would be proportional to the velocity : then, if the engine were running at a uniform speed, the resistance of the dashpot would be very great, and would give a great deal of regulation ; but still it would correct itself, and no error would be introduced. But the right place to introduce the dashpot, if one was to be used, was not always, or usually, at the load. In the case of the brake shown by Fig. 15, for instance, Professor Barr used a dashpot about Ij inch in diameter filled with oil, and attached to the spring-balance. If a dash-pot was applied to the loaded end of the strap, when the load was being lifted the dash-jjot resisted the lifting, virtually increased the load, and therefore tightened the strap and increased the tendency of the brake to lift the load. If, on the other hand, it was applied to the slack end, any lifting of the load slackened the strap on account of the dashpot resistance, independently of the reduction of the pull of the spring-balance, and therefore acted in the right direction. Besides this, a force of 1 lb. at the slack end would do as much in regulating the brake as a force of perhaps 20 lbs. at the tight end. For these two reasons the dashpot should be applied to the slack end. Professor James Thomson had jiro- posed to regulate the action of his brake (Fig. 21) by passing a cord from the light weight round a spindle carrying a small fly-wheel, so that, as the weight rose and fell, the fly-wheel would require to be rotated. This " inertia regulator " would introduce no error if the engine ran at a constant speed, or varied in speed only slowly. He had used such a regulator with success. He could not see the

Piwoedings.] DISCUSSION ON FPJCTION-ERAKE DYNAMOMETERS. 35

w v

use of the coefficient K = -,:r-^ given on p. 18. The formula might Professor Barr.

w

be simplilied to K = 33,000 -p where P was the load on the brake.

That was, K was a coefficient which expressed the width of wheel allowed for a certain amount of load. He did not think that the width of the wheel should be proportioned to the load upon the brake, giving a certain width of wheel per lb. load; but rather that a certain surface of wheel should be provided per HP., because similar brakes might be expected, roughly speaking, to absorb or dissipate a certain amount of heat per square inch of surface per minute. It would be a coefficient of that kind for each style of brake-wheel (water-trough wheel, &c.) which would be useful. It would be noticed that what was called a constant in the Paper was exceedingly variable. He should be glad if the Author would explain the observation (p. 19), "But if the compensating-levers were not Tised, a rather larger power might be measured on this brake." The difference between the amounts of work which could l^e recorded by a certain brake, with or without comjiensating levers, must simply be the amount of error which the compensating levers introduced. He thought that, p. 20, the Author had made a slip in speaking of frictional resistance being proportional to the velocity ; he probably meant inversely proportional to the velocity. The calculation, p. 20, appeared to him to be rather an abuse of mathematical processes. It was given as following Professor Rankine. Professor Eankine undoubtedly said that the formula was approximately true for a chain of blocks, and it would be so in the case of a large numljer of blocks, attached to the flexible strap, and running dry on the wheel ; but he felt certain that Professor Eankine would not have applied that formula to a brake with a small number of blocks such as that shown in Fig. 15 of the Paper, or the brake used by the Royal Agricultural Society. The formula was only applicable strictly in the case of an infinite number of blocks and a perfectly flexible strap. Further, it was inapplicable to most friction-brake dynamometers, because the friction in almost all brakes with which they had to do — and certainly those to which the Author of the Paper applied the formula — was not purely solid friction ; and did not follow the laws of solid friction uj^on which the formula was founded. He thought that the coefficients of friction assumed (0 • 2 and 0 • 3) were in no way near to what they generally were in brakes. He should wish to protest against such analytical methods being used to deduce quantitive results without the constants involved being obtained from experiments made

D 2

36 DISCUSSION ox FKICTION-BKAKE DYNAMOMETEKS. [Miuutcs of

i-ofessor Barr. under like conditions to those to which they were applied. With regard to the Appold compensa ting-levers, if they did not intro- duce any error, they did not introduce any compensation. That was the simple principle which should be borne in mind. Brakes could be made perfectly well without those levers, and he there- fore thought that a simpler brake should be used in jDlace of the " compensating-lever " one, which introduced errors. The Author had calculated what the error might be, and he had given as an extreme case a force of 11 lbs. at the end of the comi^ensating- lever ; but, where Messrs. McLarens' brake was referred to, it would be seen that the smallest force they got was 24 lbs., and the greatest 258 lbs.; therefore the force of 11 lbs. at the end of the compensating-lever could hardly be called an extreme case. He presumed that the value, 11 lbs., was what was too often falsely called a " theoretical result," being founded on some such erroneous assumption as those he had referred to. There was also, on pp. 25 and 26, a mathematical investigation which he thought was not by any means necessary. The jDrincijile was given in a few lines following the mathematics, and the result could be got at directly from the principle itself. The Paper stated, " Since it is at rest, and the tensions are internal forces, the load, the pull at E, and the friction, are the only external forces that have to be considered ; " therefore the force at the end of the compensating-lever, diminished in the ratio of the distances of the load and the end of the lever from the centre of the wheel, must be deducted from the load W to get the net load. A great deal had been said about the results obtained by Mr. Halpin and himself in testing Messrs. JMcLarens' engine, and they were referred to in the Paper as corroborating the results obtained by the Eoyal Agricultural Society. It was true that the engine tested by Mr. Halpin and himself was the engine tested by the Eoyal AgTicultural Society. He was not himself present at the Society's trial, but he had been informed that the engine was working as diiferently as possible on the occasion when they tested it from what it was doing on the occasion of the Society's trial ; he there- fore did not think that their results, corrected for the error intro- duced by the lever, could fairty be taken as showing that the Eoyal Agricultural Society's resiilts, uncorrected, were perfectly trust- worthy. Further, the most striking coincidence was in the mechanical efficiencies ; but he understood that the indicator- and brake-trials of the engine by the Eoyal Agricultural Society were made at different times, and therefore he did not think that the coincidence of the values got by the Eoyal Agricultural Society and

Procccdiugs.] DISCUSSION ON FUICTION-BRAKE DYNAMOMETEIiS. 37

liy Mr. Halpin and himself could beheld as proving that there was Professor Barr

no important error in the Eoyal Agricultural Society's trials. He

thought that in future the so-called " compensating-levers " should

be dispensed with, and he had no doubt that they would be. He

was glad that the Author had not entered into a subject on

which there had been a considerable amount of controversy, as to

the difference between the crooked and the straight lever. That

was only a very small matter of detail, and did not affect the

accuracy of any of the statements made with regard to the error

introduced by the Appold brake. He might be permitted to

express the conviction, and the hope, that the last had been heard

in this Paper of the Appold brake, except as a matter of historical

interest ; and if so he would rejoice at such a result.

Dr. Edward Hopkinson observed that Professor Barr had alluded ^'*^'- ^- Hniikin- to a type of brake not mentioned in the Paper, and had asked to have some description of it ; he referred to the hydraulic-brake originally introduced by Mr. Froude. Having had something to do with the more recent modifications of that brake, he might, ]ierhaps, be permitted to describe it. He believed that Mr. Froude introduced the brake about twelve years ago, and since its intro- duction it had been occasionally used, chiefly, he thought, by the Admiralty in connection with testing marine engines. Though exceedingly powerfiil it was very difficult to regulate its resistance, and it was not until it had been reconstructed by Professor Osborne Eeynolds that the brake had become a practical piece of mechanism. Mr. Froude's original apparatus consisted essentially of a wheel keyed on to the shaft, the torsional power of which it was desired to measure, surrounded by a casing supported by the shaft, but free to move round it. The wheel was formed of two bowls or hemispheres with their convex surfaces placed together and the flat surfaces outwards. In the two concave portions of the bowl there were a number of inclined vanes which divided the hollow space, raking forward in the direction of rotation. Corresponding with the vanes in the bowls, there was a similar series of vanes in the outer casing which raked in the opposite direction. If water was introduced into the wheel rotating with the shaft, it would be caught by the vanes of the casing, and rotational motion would be set up, absorbing the power of the shaft, and increasing imtil balanced by friction. The pressures parallel to the shaft on the two halves of the wheel would be balanced. The tangential force on the casing, tending to turn it round the shaft in the direction of rotation, could be balanced and measured by a weight at the end of a lever rigidly attached to the casing, as in the case of the

38 DISCUSSION ON FEICTION-BRAKE DYNAMOMETERS. [Minutes of

Hopkiu- Prony brake. In the original form of Mr. Froixcle's brake there were great difficulties in keeping the couple or tiirning moment on the casing constant, although there would be no difficulty in constructing a brake of that form which would absorb many thousand HP. The improvements originated by Professor Reynolds were in respect of the way in which water was introduced into the brake, and allowed to flow out of it. According to his method the water entered along the shaft into the wheel between the two convex surfaces. It was there forced outwards by centrifugal force to the peripherj^ of the wheel ; it then passed through passages or ducts cut through the metal of the vanes, and emerged with rotational motion on the faces of the wheel ; it was caught by the vanes of the casing, and finally it passed outwards between the wheel and the casing into an outer chamber formed in the latter, from which it escaped by a drain-pipe. There were also air passages in the vanes of the casing, which, as the water entered the brake, allowed the air to escape in front of it. Supposing a brake to be applied to an engine, and the water allowed to flow into it, it would gradually fill the brake, and the couple which the shaft of the engine exerted on the brake would increase, the air would be driven out through the air passages, and finally the water itself would emerge by the exit prepared for it. If the exit were closed, or partially closed, the pressure would increase, and more power would be absorbed by the brake. By regulating the exit of the water in proportion to the sxipply, the resistance of the brake could soon be adjusted to any desired amount, and so long as a constant current of water was flowing through, and the speed with which the shaft was rotated was kej^t constant, the turning moment on the brake would be precisely constant. That was a very great advantage over the Prony form of brake. Any one who had had anj^thing to do with the Prony brake knew that it was a difficult instniment to deal with ; it was very difficult to keep the friction constant, it needed con- tinual care and watching, and there were many sources of error which, however carefully guarded against, might creep in. The brake of Professor Eeynolds could be put on the engine-shaft, and, if the engine was properly governed so that the speed was constant, it might be left upon the shaft for hours without the least atten- tion, because, the flow of the water being constant, the work done in the brake would also be constant. The essential condition was that the engine should be properly governed, and the speed kept constant ; but Professor Eej-nolds had introduced a modification by which the resistance or the couple could be kept constant, even

riuccedingb.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 39

although the speed of the engine varied. This was effected hj Dr. E. Hnpkin- a simple arrangement connected with the lever which measured ^""' the tiirning moment on the casing, by which a subsidiary system of levers actuated and controlled the valve, allowing the water to flow into the brake. When the engine ran at a greater speed than the normal, and the resistance consequently increased, the weight lever wotild tilt up, the valve would be j^artially closed, diminishing the How of water through the brake, and consequently the resistance would be diminished down to its normal value ; and vice versa when the speed of the engine fell and the resistance diminished. Another advantage was that the brake required no lubrication whatever. The water passing through the brake sufficiently lubricated it. He had seen one of those brakes working for six or eight hours at the Owens College, Manchester, without any atten- tion whatever. All that it was required to know at the end of six hours' run was, what weight had been on the lever of the brake, and what was the number of revolutions through which the engine had turned during the time ; it could then be told accurately what work the engine had done in the interval. The work done in the brake in overcoming the resistance of the water apjieared in heating the water. The heat thus generated might be carried away by the steady stream of water through the brake without causing any great rise of temperature, or by restricting the stream the temperature might be raised until the water boiled, when the vapour or steam would pass away through the air passages without causing any considerable increase of pressure. A brake, 18 inches in diameter, Avould measure from 1 to 30 HP. at lOO revolutions jier minute, and for other speeds the power would vary as the cube of the speed. He might mention that his firm, Messrs. Mather and Piatt, had constructed for Professor Eeynolds three 18-inch brakes which had been working in connection with the triple- expansion experimental engines in the Whitworth Laboratory at the Owens College, and that the results had been very satisfactory.

Mr. R. E. Froude said he proposed to confine his remarks to two Mr. Fromle. of the subjects suggested by the Paper, and in the course of the discussion. The first was the question to what extent and in what way variations in the frictional resistance of a frictional dynamo- meter-brake might prejudice the accuracy of the record. Using general terms, he would treat a friction-brake dynamometer as consisting essentially of two parts, the rotating wheel driven by the motor, and the comparatively stationary resisting lever. The frictional couple between the wheel and the lever he should call the reaction, and the counterbalancing couple resident in the lever

40 DISCUSSION ox FMCTION-BRAKE DYNAMOMETERS. [Miuutes of]

Froude. he should call the preponderance of the lever. Thus by reaction he meant to describe equally the couple impressed by the lever upon the wheel, or by the wheel upon the lever, and by the pre- ponderance he meant the tendency of the lever to resist that couple, which might or might not be equal to it ; for there was no law in nature which compelled an equality between those forces, except that any difference between them must be satisfied by an angular acceleration of the lever, and consequently any prolonged difference would lead to an accumulation of angular velocity on the part of the lever which would be practically inadmissible. Therefore, although those two quantities were not naturally equal, they liad to be made equal. Hence it was found that in most of the instruments described in the Paper, there was a provision made for automatically varj'ing the preponderance of the lever to suit varia- tions in reaction, and it would be found, on a close examination, that that provision had in all cases the character which it conspicuously possessed in the instrument shown by Fig. 1, the original Prony brake — a character which he would term "statical stability" of the lever about the axis ; in other words, the preponderance of the lever must vary with the angular position of the lever much as did the preponderance of a pendiilum. The cause of the property which he termed stability was various in the different examples. In some, for instance, it was given hj the attachment of a spring- balance to some point of the lever or its appurtenances ; but, whatever the cause, it had that character of "stability" in so far as that the variation in the preponderance was essentially accom- panied by some variation, great or small, in the angular position of the lever. Since the preponderance of the lever was the only measure they had of the reaction, it was clearlj- necessary for real accuracy that the variations in the preponderance should be measured, and various provisions had been made for measuring it in some of the examples. His main purpose was to treat of the considerations which had to be kejit in view in devising such refinements in that mode of measurement as might be required by the circumstances of the case, and the degree of accuracy needed. It ■would be readily recognized that the ne plus ultra of refinement — the ideal result to which the refinements should tend — would be a continuous representation of the momentary reaction as the ordinate of a diagram, in which the scale of abscissas represented waits of circumferential travel of the motor. In such a case the area of the diagram would represent the total work delivered by the motor to the lever, on the same principle as the area of an ordinary indicator diagram measured the total work delivered by the steam upon the

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 41

piston. In the attempt to translate that ideal dia^am into practice, Mr. Froude. they were confronted with the circumstance that, whereas the ordinate of the ideal diagram represented momentary reaction, the only force of which they had a measure was the momentary pre- ponderance of the lever, which might differ from the momentary reaction by whatever force was used in the momentary acceleration of the lever ; because, as he had said, any variations in the prepon- derance must be accompanied by some angular motion on the part of the lever. Therefore, what they had to consider was how far, in sub- stituting what might be called a practicable diagram for the ideal diagram — substituting a diagram of which the ordinates represented preponderance instead of reaction — how far they were introducing error in virtue of the inertia of the lever. That was a question which at first sight seemed complicated, but there was one very simple cri- terion to which it might be submitted, in virtue of the circumstance, that the lever, in consequence of its stability about the axis, was capable of a free oscillation, which would be accomplished in a certain definite period of time depending upon the inertia of the lever, and upon the scale of its stability. If, by making the inertia small enough, or the scale of stability great enough, that period could be made much more rapid than the most rapid fluctuations in the reaction, then the difference between the momentary reaction and the momentary preponderance would practically be nil, and the difference between the ideal and the practicable diagram would he nil. If, on the other hand, that condition as to the period of the lever was not secured, then the difference between the ideal ordinate and the practicable ordinate would amount to some important quantity, and then there might be a condition of affairs something like that represented by Fig. 22. There was the ideal diagram of which the ordinates represented momentar}- re- action, and the actual diagram, of which the ordinates rejiresented momentary preponderance, the difference between the two being the forces momentarily emjiloyed in angular acceleration of the lever. This was an error in the diagram, a local ordinate error, but not necessarily an error in the area of the diagram, which was the measurement of the total work done. Whether there would, or would not, be an error in the area of the diagram depended upon the nature of the fluctuations of the speed of the motor. That could be best followed by first supposing that the speed of the motor was perfectly uniform. In that case the scale of abscissas of the diagram, which primarily represented units of circumferential travel of the motor, might be equally taken to represent units of time, and then it would be easily seen that in

42

DISCUSSIOX ox FEICTIOX-BEAKE DYNAMOMETERS. [Miuutes of

Mr. Froude. the long run the areas of the positive error must balance the areas of negative error, because, if not, the balance would represent a certain amount of momentum which had been imimrted to the lever, and which would have had to be satisfied by an accumulation of angular velocity, which they knew had not taken place. Let it be next supposed that the speed of the motor, instead of being uniform, was varj^ing in such a way as he had represented by the line AA on the diagram, of which the ordinates represented momentary speed, what would happen? The excess of speed would exjjand the longitudinal scale in the regions of positive error, and the defect of speed would contract the longitudinal scale in the regions of negative error ; consequently there would be an excess of the sum of the positive errors over that of the negative

Fig. 22.

IDEAL DIAGRAM

,' MCTUAl DlAGRAlA

OF DIAGRAM

ORDINATES SHEW MOMENTARY SPEED

CBDAC BDA

AA, Overloading Phjxee, B-B. CnxLcrlogqmn do, C. C cr D D, Neutrnl do.

errors, and an excess in the total area of the diagram, which would therefore over-log the power. Or if the fluctuation of the speed possessed the precisely opposite character, as B B, then for the same reasons the diagTam would under-log the power. That kind of conjunction of variation of speed and error of diagTam might be described in the technical language of harmonics by saying that the fluctuations of speed synchronized with the fluctuations of error, with such a j^base relation, as it was tenned, that the moments of maximum and minimum speed coincided respectively with the moments of maximum positive and negative error, or the precise reverse. If, on the other hand, the phase relation of the fluctuations of speed to the fluctuations of error was precisely midway between these' two, as C C or D D, then the error would be nil. Of course it was not to be exi)ected that the actual

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 43

fluctuations of speed should have precisely any one of these Mr. Fromle. characters ; they would jorobably be a compound of fluctuations of various kinds. What might be said in such a case was that the extent to which the initial ordinate error of diagram would be converted into error of area, and therefore into total error of record, depended upon the magnitude of such element in the total fluctuation of speed as bore precisely the character he had specified. He did not wish to be understood as implying that under all circiimstances perfect practical accuracy required the use of such a .diagram as he had supposed. Whether it did, or did not, depended entirely upon the nature and the degree of the variations in the reaction and in the momentary speed of the motor. The purposes of a continuous record by diagram might, in any case, be very conveniently fulfilled by the well-known expedient of an in- tegrating wheel working on the face of a disk rotated by the motor. This was a theoretically perfect substitute for such a diagram, and if carefully constructed and used would give ad- mirable results. He had hitherto confined himself to the as- sumption that for want of, or in spite of, apparatus for regulating the reaction to a fixed amount, there were sufficiently prolonged variations in reaction to necessitate the provision for counterpart variations in preponderance of lever which he termed " stability," an assumption which applied to most of the instruments described in the Paper, in which indeed the stability was generally fur- nished by forces brought into play by the very apparatus which approximately regulated the reaction. The instruments. Figs. 3 and 4, however, were theoretically exempt from this condition, and no doubt in these or other ways instruments might be made to so far exclude prolonged variations in reac- tion as to practically dispense with the property of stability of lever, without danger of excessive acciamulation of angular velocity, or, in other words, of the instrument being " thrown over." In such a case, the stability being nil, the " period " of lever would be infinite, and the preponderance constant instead of varying; but, taking due account of these conditions, the prin- ciples he had indicated would apf)ly, and the treatment would be somewhat simplified by the circumstance that the differences between the momentary reaction and the preponderance, which he had termed the " errors of the diagram," would be simply the variations in the reaction. And, although a continuous diagram to record the constant preponderance became unnecessary, it was useful to keep in mind the " ideal diagram," of which the ordinates represented momentary reaction, as a gra})hic expression of the

44 DISCUSSION ON FRICTION-BKAKE DYN.UIOMETEKS. [Mimites of

:\lr. Fronde, fact that the total work delivered by the wheel was the sum of the products of the momentary reactions into the corresponding successive units of travel. The treatment, as applied to this case,

might be instructively paraphrased thiis : —

Let P = the constant preponderance.

K = the momentary reaction throughout an infinitesimal

element of travel dc, occupying an infinitesimal

element of time dt. C = total travel during the experiment. T = total time occupied liy the experiment.

Then the total work delivered = f K d c ; and let 'Li — = 1{ ^ which may be termed the " travel-mean " of the reaction. And, in the same way, let —7^ — — E/, which may be termed the " time- mean " of the reaction.

Then, the true work delivered being C He, the indicated work was C P, and the error consisted in the difierence between E^. and P. Xow (and here lay the important pointj, the only condition which determined the relation between reaction and preponderance was this. T P = the total amount of angular momentum that would have been imparted to the lever by its preponderance, if this were unresisted ; T B.t = the amount that would have been imparted to it in an opposite direction by the reaction. And, since these two values must in the long run be equal, P = E;, and the error consisted in the difierence between Ej and He, namely, between the time-mean and the travel-mean of the reaction. The difierence between these two means depended, as he had indicated, on how the fluctuations in speed chimed with the fluctuations in reaction. Thus, if in a friction-brake of constant preponderance (as Figs. 3 or 4) the lever was oscillating, and the weight rising and falling, the consequent error would depend on how far the moments of excess of reaction, denoted by upward acceleration of the weight, coincided persistently ■with the moments of excess or defect of speed, and vice versa. In an ordinary friction-brake, since friction was practically independent of speed, there was no reason why such alternations of speed as were due purely to alternating action of the steam-governor, for example, should persistently coincide in any particular v^aj with alternations in reaction ; hence errors due to this cause might be expected in the long run to balance one another. On the other hand, alternations in the reaction would themselves tend to originate speed alternations, which, as modified by the

Proccediugs.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 45

action of the governor and other causes, might conceivably chime Mi'- Froude.

with the reaction alternations in such a way as to introduce

error.

He would now refer briefly to his second subject, namely, the turbine dynamometer invented by his father, the late Mr. W. Froude, M. Inst. C.E., to which reference had been made by Professor Barr and Dr. E. Plopkinson. It was devised specially to meet the case of dynamometric trials of large marine-engines, and his father's thoughts were turned in the direction of an instrument of that type, mainly by the difficulty of devising an instrument of the ordinary type sufficiently compact to be applied to a ship, and yet capable of absorbing such an amount of power with safety. He could not say that Dr. Hopkinson's description of the instrument conveyed an entirely correct impression of it. He could not then give an intelligible description of the instru- ment, but he would refer to the Paper on the subject read by his father before the Institution of Mechanical Engineers in 1877.^ The only feature on which he would comment was one which had a direct bearing on Dr. Hopkinson's remarks, namely, the method of regulating the reaction. That was accomplished by the use of sliding shutters, which covered the faces of the cells through which the water circulated, and so by more or less impeding the circulation of the water diminished the reaction of the instrument. That expedient answered simply to perfection. At will, the reaction of the instrument could be regulated from the maximum of which it was capable, to an amount of about one-fifteenth of that maximum, with the same revolutions of the engine. The instrument devised for marine-engine trials was made by Messrs. Easton and Anderson in 1878, and was tried in H.M.S. " Conquest," at Devouport, in 1880, after his father's death. A much smaller instrument was subsequently made by his brother, Mr. Hurrell Froude, for testing small rapidly-running engines, chiefly for driving electric-light dynamos, and it was sold by him to the Admiralty for tise at Portsmouth Dockyard. It had remained there in use ever since, and he had been informed that it had given the highest satisfaction. He therefore could not admit the correctness of Dr. Hopkinson's remarks that, " though exceed- ingly powerful, it was very difficult to regulate its resistance," and that " it was not imtil it had been reconstructed by Professor Osborne Keynolds that the brake had become a prac- tical piece of mechanism," also that "there were great difficulties

lustitutiou of Mechanical Engineers. Proceedings 1877, p. 237.

46 DISCUSSION ON FKICTION-BRAKE DYNAMOMETERS. [Minutes of

Mr. Froude. in keeping the couple or turning moment on the casing con- stant." He was curious to learn on what authority Dr. Hop- kinson had made those statements, which, to his mind, were entirely incorrect. He admitted that the instrument was " exceed- ingly powerful," in the sense that in an incredibly small compass it was capable of absorbing an enormous amount of power ; but he regarded that as a merit rather than a drawback. But the juxtaposition of the words, " exceedingly powerful " and " difficult to regulate," appeared to him to be intended to convej^ the im- pression that the instrument exercised its power in some capricious and irregular manner, which it certainly did not do. He ad- mitted, that before the trials he had some apprehensions that there might be difficulties at the moment of starting the engine, but those apprehensions were illusory. There was nothing to be done but to turn on the steam, and the engine steadied itself immediately and quietly to the speed at which the amount of re- action balanced the steam-pressure. The contrast between his apprehensions (which he thought were excusable in dealing with an instrument capable of absorbing 2,000 or 3,000 HP., har- nessed to a large marine engine Avithout an intervening fly- wheel), and the gentleness of the phenomena resulting, was almost grotesque.

Although this instrument might perhaps be legitimately termed a " friction-brake dynamometer," it was important to notice that the part played in its action by fluid friction diftered essentially from that played by mechanical friction in an ordinary brake. True, in both cases the friction was the iiltimate absorbent of the power; but, whereas in the ordinary brake the friction was the direct agent of the reaction, in the turbine this agent was the centrifugal force of the circulating water ; and the speed of this circulation, and consequently the reaction, was diminished, not increased, by increase of friction. Hence, in order to develop the maximum absorbing-power of this instrument, it was necessary to have the surfaces as smooth, and the flow of the water as little obstructed, as possible. In this brake, unlike the friction-brake proper, the reaction naturally varied with speed, and it would be impossible by any self-acting regulating arrangement to equalize the reaction for such rajiid alternations of speed as might arise, in default of sufficient fly-wheel, from want of balance in the engine, and the variation in the turning moment due to the steam-jjressure in different parts of the stroke. In such cases the turbine would manifest a pulsation of reaction chiming with, and slightly sub- sequent to, the primary- pulsation of speed; and hence, in the use

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 47

of this instrximent nnder such conditions, as compared with an Mr. Frmule. ordinary friction-brake, it was perhaps specially requisite for ac- curacy that careful attention should be paid to the principles which he had attempted to indicate in the first portion of his remarks.

Mr. P. W. WiLLANS desired to say a few words on one or two Mr. Willans. trials which he made with brakes twelve years ago. He made one of the A})pold brakes. He did not say that he had designed it in the best possible way, but his experience with it satisfied him that there was something wrong about it. He found that on being lubricated freely with soap and water the engine ran away, and it did not seem to him that, as the brake was there to measure the work, the engine ought to run away in consequence merely of a difference in lubrication. The engine was ungoverned, and this of course showed the defect at once. About four years ago he made another of the Appold brakes, and in his first or second trial the power which he measured on the brake was rather greater than the indicated power measured on the brake, and that did not seem to him to be right. The fact was that these brakes were only very rough measurers of power, and he thought that the best way of realizing the difficulty was to imagine that, instead of the wooden blocks, rollers perfectly frictionless were placed between the brake and the strap. It would then be seen that the weight could be lifted, by simply using the tightening-screw, the point X, Fig. 5, bearing the whole strain. The amount of the error was simply a question of the coefficient of friction. The Author had summed it up by saying (p. 25 ) that " under such conditions " (that was, when there was very little pull on the lever) " the lever does not affect results, and adjustment of the frictional grip and position at which the load is carried has to be made by the hand-screw S." In other words, the brake was a compensating one, which might be a good one when it did not compensate, but which it was not safe to use, and which he would venture to predict would never again be used by observers who desired accurate and reliable indications. He had made up his mind to use one of Mr. Froude's brakes, the one of which Dr. E. Hopkinson had spoken, and he still intended to use it for large powers ; but a few months ago Mr. Coope, of Grantham, had brought him drawings of a brake which seemed to be free from the errors to which he had referred (Fig. 23). In this brake it woiild be seen that there were two weights W and W^. The weight W was lifted and the weight Wj lowered if the brake blocks began to rotate with the wheel. In the brake actually made Wj was suspended by four cords c passing over a part of the

48

DISCUSSION OX FRICTIOX-BRAKE DYNAMOMETERS. [Minutes of

Mr. Willans. circumference of the brake-wheel, and the brake-strap proper was di\dded into two parts, one on each side of these cords. The weight W was suspended by a flexible band, which was wound or unwound if the strap showed any tendency to rotate with the wheel. The cords c carrjang the weight W^ were connected at a point p with the brake-straps h h. The brake compensated per- fectly, because if the weight W was raised, owing to an increase in

Fig. 23.

the friction of the brake-blocks, the arc of contact of the cords c was reduced and the upward motion of W was at once checked. Ko readings were necessary during the trial, and little or no adjustment of the hand-screw, as the load was solely dependent on the weights W and Wj and their respective distances from the centre, the only error being the one due to a variation in the position at the beginning and end of the run of the two weights, a very minute one, and one which could be measured if necessary. There was a slight inaccuracy in the balancing due to the length of rope unwound, but this could be obviated. The rope (Fig. 23) ought

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 49

to he carried roiincl to the other side, as shown by the dotted Mr. Willans. lines, because of the slight difference in the length of cord over- hanging; but with that exception, he thought the brake was quite right in every way. It was designed by Mr. Coope, and it was on the same principle as Mr. Imray's, Fig. 4; but it was an improvement on it, because in Mr. Imray's the frictional resistance was only that due to the two weights resting on the wheel. In Mr. Coope's brake any required strain could be put on the main strap, the variation in friction being compensated by varying the arc of contact of an auxiliary strap only. The only part of the design for which Mr. Willans was responsible was the spring 8 instead of the rigid screw often employed ; he had found this a great help in obtaining a smooth working brake, but very likely it might have been often used by others also.

Professor Alexander B. W. Kennedy said, that recently, in Professor connection with Dr. Hopkinson and Mr. Tower, he had condticted '^"'^^^y* certain motor-trials for the Society of Arts, and these trials included a number of brake-experiments. Their results had not yet been published, but he believed it would not be improper for him to say a few words about the brakes used on that occasion. After some of the beautiful apparatus mentioned in the Paper and in the discussion, he was afraid that the ropes which he had exhibited would appear very shabby. But after consideration they had thought it best to revert to that simplest form of brake, a rope or couple of ropes, making one turn round the wheel, with a spring- balance pulling against the weight, as in Fig. 15. He would give a few figures to show how the method worked. The largest power that they took up from one wheel with a pair of such ropes, the wheel running perfectly dry without lubrication, with an engine having a fly-wheel weighing about 1,650 lbs., and a diameter of 5 feet 5 J inches, the rim 9 inches wide, and running at 60 revo- lutions per minute, was just under 15 brake HP. Under these conditions the brake worked perfectly steadily, but the wheel became heated. In another case they had to take up 19 or 20 HP., and they thought it best to have a wheel made with a rim of trough-section. The wheel was 5 feet in diameter and 7 inches wide, weighing 910 lbs., running at 140 revolutions per minute, and it took up 19^ brake HP. without the slightest trouble. They continually let water drip into the trough and evaporate out, so that there was perhaps ^^ inch of water in the trough during the whole time. Working it out roughly, he found that the heat taken up by the water evaporated amounted to from 80 to 90 per cent, of the whole brake-power. There was, however, an uncertain

[the INST. C.E. VOL. XCV.] K

50 DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

Professor small amount of water lost by splashing, which would affect this Kemie y. pgypg^tage. In the case of a brake carried once round the wheel ' in that waj', of course the absolute value of the spring-balance reading was a matter of importance. It was found that the , average pull on the spring-balance was about yL of the large | weight, and never varied rapidly, but only slowly, and to an extent of only a small percentage of its own value. They had it noted every five minutes during a trial of many hours, and they imagined that they had a better average value of the spring- balance pull in that way than they had of the indicated HP. by taking the indicator cards every quarter of an hour, instead of registering the power continuously. In order to get the rope-brake to work thoroiaghly well, it was advisable that the pulley should be flat and not rounded, also that the little wooden clips, which it was safe to put on so as to keep the ropes from slipping off, should have the rope laced to them, and not fastened by a nail or screAv, or by any metallic substance which could heat and burn the rope. Mr. Kapp. Mr. GiSBERT Kapp some time ago had occasion to test electro- motors, and he wanted a very delicate brake for the purpose, which would not be liable to any disturbing error such as was introduced by the compensating-lever. He had found the brake, which he was about to describe (Fig. 24) very useful for small powers. In his tests the motor was placed on the floor, and was provided with a rope pulley which revolved in the direction shown by the arrow. On the table above was placed an ordinary pair of scales. Two holes were bored in the table through which the ropes were brought up. It was important that 4he height of attachment of the two ropes should be different. The attachment E was on the same level with the centre of rotation of the beam, and the attachment D was somewhat lower. When the pulley pulled down the cord on the right and allowed the cord on the left to rise, the leverage on the right was slightly decreased and on the left slightly increased. At the same time the rope got a little slacker. That, therefore, was a compensating action ; it allowed the weight to be lifted a little higher, and the rope to slacken until the weight on the scale would just balance the turning moment transmitted from the motor pulley through the ropes to the beam of the balance. The latter was then out of level, and, to bring it into the horizontal position, additional weight must be put upon the pan, and the thTimb-nut above the suspending spring S must be simultaneously tightened so as to put a little more strain on the rope. Two adjustments had therefore to be made before an observation could be taken, one the adjustment of the

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 51

weight on tlie scale, and tlie other that of the initial tension in the Mr. Kapp. rope; but he would here remark that the necessity of making two adjustments was common to all brakes. The brake he had described did not require more attention or labour in setting it than any other ; but it had the advantage that the compensating

Fig. 24.

action did not introdiice an error when the brake was in equi- librium. Of course, a reading was only taken when the beam was floating horizontally. In that case the leverage on both sides was equal, and therefore the reading was correct.

Professor W. E. Ayrton said his experience had been with small Professor transmission- and absorption-dynamometers ; but, in view of the ■'^y**"^* great use of electromotors in connection with the vast system of electric distribution promised next year, some remarks on dynamometers of this description might not be out of place. There was no doubt that of the two kinds of dynamometers the transmission-dynamometer was the less satisfactory; for, whereas it was quite easy to apply an absorption-dynamometer in place of a tool driven by a motor, it was not nearly as easy to apply a transmission-dynamometer, for example, in the case of a

K 2

52 DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [Minutes oi

Professor dynamo-machine which was driven by a belt. It was a reflection Ayrton. ^^ electrical engineers that they had no means of doing that. His colleague, Professor Perry, and he had tried various methods. He could not give any details of them, as they were far from being in a practical state yet, but he would endeavour to give an idea of what they were aiming at. Taking an ordinary belt driving a dynamo-machine, to estimate the difference in the tightness of the two parts, they had tried sending waves along them, setting the two sides of the belt in vibration, and seeing how many nodes were found in the one and in the other. Another plan they had thought of was having a wire fastened to the belt and running with it, and getting the electrical resistance of the two parts, thus ascertaining the difference of tension. At present, however, there was no mode available for at once applying something to a belt driving a dj-namo-machine, and estimating the power given to the machine. The nearest approach to it was the well-known belt dynamometer of Hefner-Alteneck, which measured the difference of tension in the two belts. It was not a very satisfactory appa- ratus that could be always left in place measuring the power given to the dynamo-machine. It was extremely noisy; there was no method of lubricating its many pulleys ; and it was not particu- larly accurate near the zero. For that reason they thought five or six years ago that it would be a good plan to make a sort of dynamometer-coupling, which should always remain in position in a shaft, and would, therefore, tell the power the shaft was trans- mitting ; not a laboratory or testing-apparatus, to be only employed when an investigation was to be made, but something w^hich shoiild always be indicating the power that the shaft was transmitting. He exhibited such a coupling, which had been running for the last five years in the main shaft of the Fins- bury Technical College. It was made in two parts, joined together by spiral springs. One-half was keyed to one end of the shaft ; and the face was keyed to the other. The torque transmitted was, of course, evidenced by the slight stretching of the springs. So far. Professor Perry and he had merely followed a great number of people who had preceded them, and who had measured a transmitted couple by the extension or bending of springs. The question was how to measure that slight stretching ; and it was in the method they employed for measuring the exten- sion of the rotating springs that the novelty existed. The plan which they had found best was to attach the arrangement by a simple link-motion to a light pointer, which carried at the end a bright bead. When the apjiaratus was rotating, the bead described

I

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 53

a bright circle, the diameter of which became less as the power Professor transmitted was greater. In the instrument, the motion was ^^ °"' exactly, as given by a curve of the calibration which he held in his hand, 3 HP. per inch. The diameter of the circle measured the power that the shaft was transmitting at a given speed of revolution. At a speed of 1 62 revolutions per minute, the diameter of the circle described by the bright bead diminishing by 1 inch meant 3 HP. The apparatus was very sharp in its action, arising from the springs being very strong and the pointer light, the sensibility being obtained by considerable magnification being effected by the way in which the pointer was fastened. If the load on the engine were changed, the bead moved instantly to a new place. For a given load, at a given speed, there was an oscillation of about ^ inch of the bead, so that the diameter of the circle described by the bead could be read accurately. The slight oscillation showed that the friction was not resisting motion. They used a gas-burner over the right shoiilder of the observer, which illumin- ated the rotating bright bead more conveniently. The idea of applying a dynamometer directly to the pulley of a dynamo- machine had been a long-cherished idea with them ; and he was delighted to find that Professor Smith had carried it out. He had replaced the ordinary pulley of the dynamo-machine by what looked like an ordinary pulley, biit which also measured the power transmitted by the belt to the machine.

The idea of a dynamometer-coupling had also some five years ago been employed by Professor Perry and himself, with the co-operation of Mr. Tomlinson, who was assisting them at that time, in order to get an automatic governing of the speed of an electromotor. The ordinary method was to have some sort of centrifugal governor which cut off the current when the speed was too fast. They, on the other hand, wanted to devise a dynamometer for governing, so that the motor should be regulated by the work it had to do, and not by the sj)eed it ought to go at, but was not going at. In their motor, instead of having stationary field-magnets, stationary brushes, and a rotating arma- ture and commutator, they had a stationary armature, rotating field-magnets, and rotating brushes. Under ordinary circumstances in one of their ungoverned motors, the brushes were keyed rigidly to the field-magnet; but, instead of that, in their " dynamo- metrically governed " motor the brushes were keyed to the tool driven by the motor, the briishholder being connected with the field-magnet by a spiral spring. If, further, matters were so arranged that when the spring was stretched, when there was

54 DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

Professor much torque transmitted by the motor to the tool, the brashes Ayrton. -^gj-g j^ their best position, whereas when the torque transmitted was small, and the springs were much less stretched, the brushes were in a less favourable position, there had been an extremely] simple dynamometrical governing of the motor; the motor went] at a constant speed, because the greater the pull it had to give, ot\ the greater the couple it had to exert on any tool, the more were the springs stretched, and the more was the brushholder left behind, and the better was the position of the brushes electrically. Dynamometrical governing had been subsequently devised by Professor Silvanus Thompson in England, and by Professor Elihu | Thomson in America.

In all compensating absorption dynamometers, there was a certain object to be carried out. The coefficient of friction varied. People had tried in a variety of ways to make up for this. Something must be varied if they did not want the weights to be thrown over. One plan was to let one of the forces vary by means of the spring- balance ; but he did not think it was a good plan, as it was likely to alter the load on the dynamometer. Another plan was to vary the arc of contact. It was curious that every person who had arrived at that idea seemed to regard it as his own. It had been invented over and over again, and he now heard, for the first time, that Mr. Coope had invented it. He believed it was due initially to Professor James Thomson. It was afterwards re-invented by Mr. Carpentier, and modified by Mr. Eaffard, who made compen- sating-dynamometers on that principle several years ago. The late Professor Fleeming Jenkin, M. Inst. C.E., while working with them in telpherage, hit on the same idea and thought it was his own. He and his colleague had tried several dynamometers utilizing the same princijDle, and, as far as he had seen, the auto- matic variation of the arc of contact was the best method of compensating for alteration in the coefficient of friction.

It had occurred to them, however, four or five years ago, whether there might not be another way of compensating for variations in the coefficient of friction when the difference of weights was kept constant. The idea was to use two cords, not two different cords, but a cord in two parts, with different coefficients of friction, joined together, and their arrangements had been referred to by the Author in his introductory remarks. The plan had become much simplified after several years of trial. It was found that a small knot tied on the cord answered all purposes. It was so simple that he hardly liked to show it ; but its sim- plicity was its great charm. There was a cord running in a

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 55

grooved pulley and a string was tied ronnd the cord to make a Professor softish knot or protuberance in the cord. It was not at all a ^y''*°"- bad compensating-dynamometer. The load might be altered considerably without getting the weights thrown over. Such an absorption-dynamometer with a knotted cord had been used on an electromotor for several hours at a time without either of the weights falling, and without any adjustment being made by hand. Every now and then, when the coefficient of friction diminished and the heavier weight fell, the knot came up against the pulley and prevented it going over ; then it went back again and oscillated slightly. He might mention that a form of knotted-cord absorption-dynamometer had been employed by them in experimentally calibrating the transmission-dyna- mometer with rotating bright bead which he had been describing. The absorption dynamometer in this case consisted of a cotton belt some 4 inches wide, hanging over an ordinary flat-rim pulley, carrying weight at each end, and the knot was formed of a leather lace that was passed two or three times backwards and forwards through the cotton belt. They had found that the automatic compensation was very good, and that, by having a stream of soapy water running over the belt, this form of dyna- mometer could be used to measure up to 27 HP. The transmission- dynamometer was calibrated by transmitting through it the power from the engine which was temporarily all wasted, and at the same time measured with this knotted-belt absorption-dynamometer.

Mr. John Imray said that, in the Paper, reference had been made Mr. Imray. to a brake which he had invented many years ago (Fig. 4), and it might be interesting to the members to know something of the circumstances under which it was brought out. It was many years since the late Mr. William Fronde and he investigated, at consider- able length, the conditions of the frictional hold of belts upon pulleys, and the result was communicated in the Paper by Mr. Froude to the Institution of Mechanical Engineers.^ The first thing they had to look at was this. At that time amongst engineers there was a fallacy prevalent that, the larger the pulley, the greater was the frictional hold of the strap upon it. They disposed of that by trying pulleys of all sizes from 5 inches to 5 feet, with straps on them loaded with weights, and there was not a shadow of difference between them. The diameter of the pulley had nothing to do with the frictional hold. They then investigated the question, and they thought that they were the first who had

' lustitutiuu of Mechanical Eugiucers. Proceedings. 1858, p. 02.

56 DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

Mr. luiray. come to a formula for frictional hold, which was very much like the one given by Professor Ayrton. It appeared from that formula that a change in the number of degrees of the arc of contact made a great change in the frictional hold. For instance, they found that if the one weight was 1 lb., and the other weight was 3 lbs., when half the circumference was embraced, then the latter would be 9 lbs. when the whole circumference was embraced, when three halves 27 lbs., and so on according to the formula. It there- fore appeared to him that the best way of making a brake auto- matically adjustable was to make it vary for itself the amount of the circumference which was embraced by the strap. For that reason he schemed the brake shown in Fig. 4. There were two arms, one on each side of the wheel. Those arms carried metal straps, by which the large weight was hung; and to the top of those arms at B the brake-strap was attached. Whenever the weight rose it took a less part of the circumference ; when it descended it took a greater part of the circumference, so that it always cured itself, and it kept very steady. He believed that Mr. Froude used it, and to a large extent had foimd it suc- cessful. About the same time Mr. Froude produced a brake for measuring the power transmitted to a machine. He had two pulleys, and he passed the strap round one of them, back round the driving-pulley and then round the other. Those two pulleys were mounted on a lever with a spring-balance, that measured the difference between the tight strap and the loose, and recorded that upon a card which gave a good indication of the power transmitted. Professor Professor E. H. Smith stated that he had been asked to explain the construction of his transmission-dynamometer, which he had brought for the purpose. One of its chief advantages was that it tested the machine to be indicated exactly under its ordinary working conditions. The ordinary driving-pulley was taken from its shaft, and the dynamometer-pulley was clamped in its place. The machine exhibited could be clamped upon any shaft, from 1 inch to 2 inches in diameter. It consisted of a central body, which was bored right through to 2 inches diameter, and in which were cut three taper slots, up which were driven three wedge keys, all advanced together hj a ring and a single nut, the nut being 2y\^ inches inside diameter and having its thread on the outside. Thus the instniment was clamped upon the shaft so that, if the shaft itself ran true, the dynamometer-pulley invariably ran true, if it was clamped Avith sufficient tightness. If it was allowed to get slack it began to vibrate. At the end of the inner body ran

Proceecliugs.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 57

loosely an outer collar or flange to which was bolted the pulley Professor upon which the driving-belt ran. The pulley was made separate ™' * from the flange for the sake of changing it, so that difierent sizes of pulley might be readily put on. Between the flange which ran loosely upon the inner body and another flange which was keyed to it at its other end, lay a couple of spiral springs. The belt, therefore, of the pulley could only drive the shaft through the intermediation of the sj^ring; and, having calibrated the strength of the spring, the twist of the spring measured the moment being transmitted. It only remained to render that twist of the sj)ring visible upon a scale, which was done by an indicating nut. Upon the end of the inner body was cut a sharp-pitched, double-threaded screw, a nut running upon it. Along with the outer j^art of the instrument, namely, the driving-pulley, ran a steel tube in which were cut two longitudinal or axial slots, opposite each other, and the indicating nut had two corresponding wings which could slide along the slots. Thus, since the rotation of the nut equalled the twist of the spring, and therefore indicated the driving moment that was being exerted, and, since the nut must move axially along the screw a distance proportional to its rotation, the axial movement of the nut indicated the moment being transmitted. That was magnified by means of a long light lever. Two little steel blocks mounted on the pins in the ends of the short double or forked lever rested continually against the flange of the indicating nut ; they were always kept pressing one side by a weight on a small lever. The end of the long lever indicated the driving moment upon the scale. Two sets of springs were supplied ; one a right-handed spiral spring, and the other a left-handed spiral, suitable for machines which rotated in opposite directions.

The transmission-dynamometer of Messrs. Ayrton and Perry had, no doubt, been greatly improved since its introduction. The first of these was made in his own works from his own drawings. Professor Perry supplied a sketch of what he wanted ; but in this sketch the arrangement of the springs was identical with that of General Morin's original dynamometer, and quite different from that actually adopted in this machine.

With reference to friction-absorption dynamometers, he thought that the moral to be drawn from the first part of Mr. Fronde's remarks, illustrated by the diagrams, was simply that the whole dynamometer should be made as light as possible ; and a corollary from that moral was that, instead of using weight forces, spring forces should be employed, because spring-balances were less massive than the weights for the same forces exerted. Professor

58 DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [IMinutes of

Professor Kennedy had referred to the good action of the rope-dynamometer.

Smith, jjg might mention that during the present year he had used a similar dynamometer made of an ordinary leather belt, running upon a 5-foot cast-iron fly-wheel, with a spring-balance at the slack end, and a weight at the other, and it ran very steadily. He now wished to make two classifications of friction-dynamometers. In the first place they might be practically divided into friction- brake dynamometers for small and for large powers. The dyna- mometers for small powers were easy to construct, but it was difficult to construct one that would work well to indicate large powers. The reason of that, he thought, was because the coefficient of friction must be kept small in order to get smooth, easy work ; and the necessary consequence was that very large surfaces must be employed, wide brake-pulleys, and large diameters. The whole thing became clumsy and awkward for large powers. He had sketched a diagTam to show why, in order to get smooth working, the coefficient of friction must be kept down. The reason, in mathematical language, was that the second differential coefficient of the ratio of the large tension to the slack tension at the two ends of the belt, taken with respect to the coefficient of friction, was very rapid. Taking, for instance, an arc of contact equal to a whole circle, and the coefficient of friction O'lO, 0*20, 0*30, 0-40, and 0"50, successively, the ratios of the tension at the tight to that at the slack end woiild be 1*87, 3 '51, 6*59, 12-30, and 23 "10. If that was indicated graphically by a curve, it would be noticed that the curve turned up very sharply. Thxis, at high coefficients of friction a small variation of friction- coefficient created a gTcat disturbance in the mechanical working of the machine, while with small coefficients there was a com- paratively small disturbance. The variation with regard to the arc of contact was of exactly the same kind. Thus, to avoid irregularity of working, it had been frequently found necessary to use unguent in order to keep down the coefficient of friction. At first sight, the use of unguent might seem most irrational in a machine of that kind, which was a machine for creating friction. The second classification which he would like to make was one with regard to the compensating attachments. Careful distinction should be made between two kinds of compensation, firstly, momentary compensation, compensation for variations of coefficient of friction that lasted only for a moment or two ; and, secondly, compensation for gradual and permanent variation of the coefficient of friction.

The first kind might depend on the inertia of the mass of

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS.

59

one part of the instrument. This was the principle of the designs Professor shown in the three diagi'ammatic sketches, Figs. 25, 26, and 27. The first two of these he had used successfully, the second. Fig. 26, being what was ordinarily used at Mason College. In this,

Fig. 25.

however, the compensation-joint came near the tight end of the strap. If it was inserted at the slack end, as in Fig. 27, it should act more delicately and promptly. In each case the strap was tightened at the compensation-joint by an auxiliary lever a c, which was bound to the main lever by a spring c d (usually an

60

DISCUSSION ON FKICTION-BEAKE DYNAMOMETERS. [Minutes of

Professor india-riibber band).

Smith.

The tension of this spring was conveniently, but not necessarily, made adjustable; but this tension did not need to be measured. On gripping of the blocks on the drum occurring, the main lever was carried round, but the inertia of the auxiliary lever caused its end c to lag behind the main lever in its motion, and thus to slacken the strap at the joint a h. There was one minor fault inherent in these brakes, namely, that the mo- mentary compensation took place in spite of the tightening of the spring c d. The auxiliary lever should be made massive at the end c, in proportion to the average tension put on the spring c d ; but, since the whole weight of the apparatus was accurately balanced over the centre of the brake-wheel by the adjustable weight G, no force measurement except that of the indicating spring came into the calculation of the brake-power. In each

Fig. 28.

ScoJecL

SvrLng

case the centre of mass g of the aiixiliary lever, and the joints a and h, should all be on one radial line. It would be an improve- ment to bell-crank the auxiliary lever, and place the spring binding it to the main lever radially as in Fig. 28. In this arrangement the weight of the auxiliary lever being left mo- mentarily behind to the left of the vertical through the joint on which it rested, exerted a txarning moment tending towards an extra slackening ; that was, a moment partially counteracting the tightening effect of the spring, which moment increased the further the brake was drawn round by the gripping of the blocks.

Deprez's brake (Fig. 3), was an example of compensation for gradual and permanent change of coefficient of friction. The rope or leather strap with weight and spring-balance at the slack end also belonged to this class. Fig. 29 showed another form which it would be interesting to try. In this the tension t at the slack end

'Proceedings.] DISCUSSION ON FEICTION-BRAKE DYNAMOMETERS.

61

was produced by a spring acting throngli a bell-crank, shaped and Professor supported so that, when the brake gripped and was carried round, ^""*^"' compensation took place in three different ways ; first, the spring shortened and exerted less pull ; second, its leverage round the fulcrum decreased while that of the end of the strap increased, so that the ratio of slack-end tension to the pull of spring diminished ; third, the strap was lifted off the brake-drum through a small increasing arc, so that the arc of contact was diminished. The objection that the ratio of t to the pull of the spring was variable, and entered into the calculation of the HP., might be met by inserting a second scaled spring between the end of the strap and the bell-crank, in which case the tension of the lower spring would

FiCx. 29.

not need to be measured. The apparatus then became, however, more complicated and expensive. It might be interesting to note the correct formula that took the place of the well-known logar- ithmic or exponential formula (which was only ajjplicable to a continuously-covered arc of contact), when the friction was pro- duced by blocks at finite angles apart. If there were n blocks from the slack pull t to the tight pull T spaced equally at the

2 TT

angle -:j^ apart ; and if </> was the " angle of repose " (tan 0 =

coefficient of friction) ; then the correct ratio between the pulls at

the two ends was

~ 1

T

T

sin|.(i-	-ky^\
sinj.Q-	~k)-	-*!

62

DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

Professor on the Supposition that the blocks were connected by links with

Smith, frictionless pin and eye-joints.

r. Halpin. Mr. Druitt Halpin could not agree with the statement in the Paper, that the dynamometer could never give scientifically accu- rate measurements of the rate of the absorption of the work done. He thought, if proper precautions were taken in using the instru- ment, and in registering its indications, one of the most accurate , results might be obtained with which engineers were acquainted. He would endeavour to show how such results had been secured in the trials carried out at Leeds by Professor Barr and himself ; and it mis:ht be of interest to describe the means taken to arrive at the truth in those trials. He had interpolated a column in itabcs in Table III (p. 27) in the Paper, and some remarks which he would explain.

Messrs. McLaren's Trial.

Halpin's Brake.

Brake made like R.A.S. " corrected."

Brake made

like R. A. S.

but calculated

by R.A.S.

Rule.

Indicated HP 23

Brake HP 20

Coal per brake HP. per hour . j 2

Feed-water per brake HP. per\ „„

hour /; ^-^

Mechanical efficiency ... 0

22-20

19-10

2-14

22-00

0-86

R.A.S. Trial

at Newcastle.

22-20

29-95

1-39

14-26

1-54

24-020

22-770

2-267

21-530

0-860

All tests to which he referred were made with the same com- pound-engine, indicating approximately 20 HP., which had been tested at Newcastle in 1887 by the Eoyal Agricultural Society of England. The first tests mentioned in the Paper were on a brake similar to that sho\\Ti by Fig. 15. That brake carried the load directly at P, and at the top, at the tail-end of the brake- strap, there was a spring-balance, which had been accurately tested with dead-weights. The rim of the brake was as shown in Pigs. 14 and 15, and water was run into the rim on one side, and -continuously removed at the other. Passing a constant quantity of water through the rim of a brake-wheel was a very different thing from getting water in, and letting it boil off at any varying tempera- ture ; as the power of maintaining the temperature constant made it possible to keep the coefficient of friction constant, and thus to ensure the uniformity of the brake resistance. A further precau- tion to obtain accuracy was that the spring-balance was connected

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 63

to the recording-paper of a Moscrop recorder, which he certainly Mr. Halpin.

thought was a most delicate instrument, both in connection with

the indicator and with the brakes. In that way, an absolutely

autographic record was secured of the variations in the pulls in

the tail-strap of the brake. Those variations occurred from one

reason : the brake had run only a day or a day and a half before

use, and he had found from experience that, if a brake was running

in proper order, having the temperature completely under control,

the blocks having been at work a reasonable time, it could be

treated absolutely in the same manner as a counter ; it must not

be toiiched fi'om the beginning to the end of the run. If a brake

Fk;s. 30.

381 lbs

was not running in that condition, to his mind it was not running correctly. That was the first test made of which the figures were given in the first column, the indicated HP., and the brake HP., and the result was that a mechanical efficiency of 85 per cent, was obtained. The second column referred to a test conducted with the same engine. It was made on a brake as nearly as possible like the brakes used at Newcastle by the Eoyal Agricultural Society, except that the precaution was taken of measuring the whole of the forces acting on the brake, it not being assumed that any negligible forces were passing through certain fixed points. The brake was shown (not to scale as far as the lever was con- cerned) in Figs. 30, and the data in connection with the test were given in the second column of the Table, abstracted from

64 DISCUSSION ON FRICTION-BEAKE DYNAMOMETERS. [Minutes of

Jlr. Halpin. the Paper, with the addition in inverted commas of a single word, " corrected." The results were given as foxmd ; but with the addition that the whole of the forces, and not merely some of them, as at Newcastle, were measured. The inner end of the compensating- lever was left free, and two powerful spring-balances, carefully calibrated, were attached to it ; the effect was that, when carrying a load of 381 lbs. vertically at the end of the lever, the pull was 313 lbs. This altered the result ; for had the brake-power been calculated, as was done by the Eoyal Agricultural Society, and the number of revolutions taken multiplied by the load lifted and multiplied by the radius, the brake HP. would have been 29 • 95, with an indicated HP. of 22 • 2, giving a mechanical efficiency of the brake of 1 • 54. Under the circumstances, he thought that error w^as worth noticing. A fourth column showed the results as obtained by the Eoyal Agricultural Society at Newcastle ; and the Author stated, quoting from a Table published by one of the con- sulting engineers of the Society : " It will be seen that the figures agree very closely." The Author stopped there, but Mr. Anderson went on to say : " Hence we are bound to assume that the experi- ments at Newcastle were substantially accurate, or that Messrs. Halpin and Barr have erred to the same extent we are supposed to have done." He thought the accidental closeness of the figures was arrived at by introducing a necessary correcting coefficient of 56 per cent. In making the test referred to in the second trial, a most interesting fact had been graphically recorded. The brake was similar to that shown by Fig. 5, with a compensating-lever, and water was run on to it, as at New- castle. The pull at the end of the lever was graphically registered by the Moscrop recorder. Everything was measured — the water, coal, and all else used by the engine in the experiment — and the boiler was fed from a large tank, containing a weighed quantity of water. In the middle of the run it became necessary to refill the tank ; and as only one water-service of pijjes was available, both for replenishing the tank and for cooling the rim of the brake-wheel, the tank was filled from that service ; therefore the Moscrop recorder showed immediately the variation of the pull at the inner ends of the compensating-lever ; so soon as the water began to enter the tank, and the available supj^ly of water on the brake-wheel was limited, the coefficient of friction changed enor- mously. The brake HP. went down from 29 • 95 to 23 or 24. The speed remained constant, and the indicated HP. decreased, owing to the action of the automatic expansion-gear, on the steam admis- sion in the high-pressure cylinder. When the cock was shut again.

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 65

after the tank was filled so as to feed tlie boiler from, it, and the Mr. Halpiu. water was again txirned on to the rim. of the brake-wheel, there was the same evenness in the pull on the inner end of the com- pensating-lever. To obtain accuracy, the curves were plotted and treated with a planimeter. Autographic registrations had been made, during both trials at Leeds, by means of the Moscrop recorder ^ (Fig. 31), where the top line in each case showed the variations in the boiler-pressure ; the centre line the percentage of variations in the revolutions of the engine, an increase or decrease of speed of 2 J per cent, being measured by the vertical distance between the lines ; and the bottom line the variations in pull on the tail of the brake-strap during the first trial, and on the inner end of the com- pensating-lever during the second trial. A great decrease in this pull was noticed about the middle of the run, owing to the shortness of the water-supply to the rim of the brake-wheel, as explained above. If those curves were treated by a planimeter, as indicator- diagrams were, very accurate results could be obtained. A year before the trials were made, he had suggested to the Eoyal Agri- cultural Society, at a meeting of the Institution of Mechanical Engineers, the desirability of not working with the usual form of brake, but working with a water-cooled brake, which would be under complete control ; and also, instead of introducing a further unknown error, by using the present form of brake separate from the engine, and having two universal couplings with an inter- mediate length of loose shaft, plus another shaft with two additional bearings carrying the brake-wheel, to obviate the existence of all such sources of error by putting the brake directly on the crank- shaft of the engine to be tested. What the value of that error might be he was unable to say ; but when the brake was running directly on the engine-shaft, the consumption of coal was 2 • 1 1 lbs. per indicated HP. per hour (Table III, line 3) ; and when it was running, having the power absorbed by means of the indirect transmission, and using the necessary correction, it rose to 2 "14 lbs. Trustworthy results, however, could not be expected from the indicator-diagrams at the Newcastle trials, as they were not taken at the same time as the brake-tests were made, and an inspection of some of the sample ones, published by the Consulting Engineers, bore an error on their face of 27 per cent. When a comparison of the indicated and the brake HP. produced a mechanical efficiency of 1-082 per cent., the Consulting Engineers gave a foot-note,^ to

' The Engineer, vol. Ixv., 1888, p. 23.

' Institution of Mechanical Engineers. Proceedings, 1886, p. 370; and Journal of the Royal Agricultural Society, 1887, p. 725.

[the INST. O.K. VOL. XCV.] ¥

66

DISCUSSION ON FEICTION-BKAKE DYNAMOMETERS. [Minutes of

Mr. Halpin. the effect that " there was an error in the observations, due pro- bably to defective indicator-pijjes," but when they had to record

Fig. 31.

1

f-

~^

'

J^

\

<

■m

J^

i

ro

f

—

ilA

Hr

■ ^

f

f-

-4

^ POWER STEAM

>

<f

- Start -

g^

^

f!

<-,.-.

V,

^^

1 z

V-

£

SPEED Co r- * I C -O m o -( t'' &» -0 r^ - 5--*=* ^ r- B" 1 •» -Clto' — * * -0 Qa PI o> a 00 = ' k tn ^? 2| ^- 33 g

k-

i

(

•1

\

—

■ —

—

-m

•VJ '

1

r Co PI ■1 S OS

Sf££0

3~

•^a

^=

k

pi—

"~ —

-^^

_,_

'^m

-

J

_

~S

f

9

— <

...

" '

; 1

^"

k—

1

-

—

^^g

^

■s.

:

J7

""^

w.

■m

I

1

»r-

<"

1

V

uZ^^

^

4

I.

^

^

w

1 ^

k»>-

^

—i

1

/

C

f-

*

—

I

^

1 ]

I

-

k

L_

—f-

1

T

'

—

^

J

K

'

-

*».•

^

'

f

_

-

^

tk

i

""

c

_

t

^

1

9

^"■

H

i

■

^

£

__

1

k

:

d

^

\

f

f

1

*

— ^

i

"7

J POWER STEAU

POWER 1

— <

H

k

<-

ft«r

/-

^

STEAM 1

it

ASl

ni

,,s^

1

PU

LL ■)N

.,

SPEED 'iNlSH -

►

S/'££D

LANCE

1

(d

^.

^:^(i

(

f

;fl/

3^

N

AFi

: or SI

EEt

/t.

"^

LA

FA

r.<

t

* '-^

r,H

ftfW

SIrtft

ftp

PRT

ir

4

f6^

fARI

ATI

ON

or

5/'£

ED

an efficiency of merely 0-943, they did so without comment, evidently crediting this result as a physical possibility. He asserted that, unless the wliole of the external forces acting on a

•ProceediugB.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS.

67

dynamometer were accurately measured, the results were worthless, Mr. Halpin. and did not give the proportional efficiencies of various motors tested on the same brake. He could confirm one statement in regard to the Leeds trials. Professor Barr stated that he was informed that the engine when in their hands ran much better than it did at Newcastle. Mr. Halpin was close to the engines during all the trials, and observed their work, and he could fully bear out that statement. When the engine was brought to Newcastle, it was new; it came into the hands of Professor Barr and Mr, Halpin after it had been running two or three weeks, and the difference in working was very marked.

Mr. John Goodman said that none of the speakers had referred to Mr. Goodmaiu

Fig. 32.

65

V

s*

\ s

\

s.

\ \

^

^'

£ATI

ea

V

N

—

"^.

VIOO

40

\

^

^

^

'

'^^

^_

•RICA

■to LI

ATHE

R

-*..

^._

I "^

%

X

i-<J..

-p,^

•15 10

'"^><..

'--.—

•^-.r

~-'*r

^\:

-em

£

"

LUBI

\ICAT

D rtd

>r--

'—•1

"SB..

"^

—

'

10 20 30 40 50 fiO 70 80 90 100 110 120 130 140 lEJ 160 170

a method, in use in America, for making dynamometer-brakes run more smoothly than with the ordinary wooden blocks, namely, by covering each block with a piece of leather. Some time ago when he was making a brake trial, the brake vibrated a great deal ; and instead of putting a dashpot on in the usual way, he covered each of the blocks with a piece of old leather belting, and immediately the brake ran as steadily as possible. During the past week he had made several experiments with leather, wood and rope on a friction-machine, the results of which were shown by Fig. 32.

F 2

68 DISCUSSION ON FRICTION-BEAKE DYNAMOMETERS. [Minutes of

lan. The lubricants were plumbago and tallow, which showed that by- using rope, or by lubricating the surfaces, the friction was much more regular than with dry leather and dry wood. The machine was driven by a gas-engine. At each ignition of the gas there was a distinct vibration due to a momentary change of speed and friction. With dry wood, at every ignition of the gas, there was a diiference in the coefficient of 0-030; with dry leather, 0-014; with dry rope, 0 • 002 ; with wood lubricated with tallow and plumbago, 0 • 007 ; with lubricated leather, 0 • 004 ; thus showing that with a variation in speed, due to each ignition of the gas, the vibration due to dry wood and dry leather was more than ten times as great as with rope and lubricated wood. That might throw some light upon the circumstance that the rope-brake used by Professor Kennedy ran so very smoothly, even under the most trying circumstances.

ont. Mr. W. WoRBY Beaumont, in reply, said, with reference to the Table exhibited by Mr. Halpin, the conditions under which the brake was used when those figures were obtained were certainly not the conditions that existed when the Eoyal Agricultural Society's brake was tried at Newcastle. At Newcastle the lubricant gave a much higher coefficient of friction than water ; and, as stated in the Paper, Messrs. McLaren found that with tallow the pull ui")on the small lever was as low as 4 lbs., and in that case the error was insignificant. Of course, one person could manipulate an instrument so as to make very different indications from those obtained by another person using it under different conditions. He therefore did not think it was right that the figiires quoted should be taken as final, because the conditions were diiferent. Accordingly, the 64 per cent, of correction need not be made. That, he thought, was proved by the fact that the figures, when corrected, came to the same within about ^i^, as ascertained by the Eoyal Agri- cultural Society's brake, showing that it was only the ditFerence of conditions that rendered the correction necessary. He had not mentioned the steam-engine indicator referred to by Professor Barr, and had no intention of siaggesting the possibility of com- paring either the relative values or importance of indicators and friction djmamometers. Concerning the rise and fall of the load P, and the work done by the engine but not recorded in its favour, the statement in the Paper was substantially correct, in view of the action which took place in ordinary cases. With brakes such as those shown by Figs. 5 and 12, the difference between the velocity at which the load was lifted or fell was not, he believed, sufficient to' allow the engine to obtain any note-

Proceedings.] DISCUSSION ON FRICTION-BRAKE DYNAMOMETERS. 69

worthy relief during the fall ; and, if this were the case, the raising Mr. Beaumom of the load was work done but not acknowledged. If the load were raised slowly, and dropped quickly, the engine might be the gainer by relief during the fall. On the other hand, it was sup- posed, as by Professor Barr, that when the load was lifted less work was done against friction. This would be true if the load could be caused to rise without change in the friction ; but inas- much as the lifting of the load was a result of increase of friction, the supposition was not allowable, except under conditions which did not seem to occur in practice. But, the falling of the load was assumed to be attended by the performance of more work against friction ; as the fall was the result of decrease of friction, this also was not a permissible assumption. Hence, the work of raising the load was uncompensated work, and not recorded in favour of the engine. The spring mentioned at p. 7, as pre- ferable to a weight jj, was intended to be understood as prefex'able in practice ; and with brakes that did not work " with a steady load at the tail end," as stipulated by Professor Barr. If the very simple and good brake, which would work with such a steady tail load, coiild be had, it would be easy to work with accuracy without any tail load. The rise and fall of the load P would, however, if considerable, without doubt be attended with less inaccuracy with a dead-weight tail load than with a spring. A brake, as Fig. 2, but with a spring at S similar in effect to the springs shown in Coope's brake, could probably be thus worked. Fig. 23. The reference by Professor Barr to the use of the dashpot was a most important one, as it showed how inaccurate the indications of a brake might be if a dashpot were used. The coefficients mentioned on p. 18 were simply numbers proportional to the surface passed over by the brake-blocks per HP. per minute in a number of cases ; and from these were deduced constants, which might form a guide for the determination of the minimum surface for such brakes. Or two brakes equal in every respect, except that one was fitted with compensating-levers, that brake which was not so fitted might be employed to deal with more power, because the necessary increase in the tension upon the brake-strap would not, after a certain maxi- mum was reached, be attended by the inaccuracy which the levers would introduce. There was no abuse of the mathematical method in the calculation of the maximum tension on the brake- straps, for the blocks, though not many in most cases, were sufficiently numerous to permit the method of calculation adopted with almost inappreciable error. With the small range of flexure demanded, the brake-strap might be considered flexible without inaccuracy.

70 DISCUSSION ON FRICTION-BKAKE DYNAMOMETERS. [Minutes of

[r. Beaumont. The low coefficients of friction, assumed for the calciilation, were purposely taken, so that a maximum pull upon the compensating- levers would be arrived at ; and hence, if a larger coefficient were more nearly that actually occurring, then there would be a still smaller pull, with the loads taken, on the upper end of the com- pensa ting-lever, and the error due to it would be less than that given on pp. 20 and 23. He therefore considered a pull of 11 lbs. at the upper end of the compensating-lever, and under the circum- stances mentioned at p. 24, less than half what had been found by Messrs. McLaren. If the engine referred to by Professor Barr did work better at Leeds than at Newcastle, it was not shown by the figures given on p. 27 ; but the boiler did appear to have done so. The friction-brake described by Mr. Willans seemed likely to give excellent results. The introduction of the springs at the adjusting screws was a good feature, and would probably materially reduce the necessity for other compensation. The information given by Mr. Goodman would be \erj useful to engineers dealing with friction-brakes and similar mechanism.

Correspondence.

'rofessor Professor T. Alexander and Mr. A. W. Thomson considered that

Hr. Thomson. *^® Appold brake gave quite accurate results when it was used

properly. There was a verj^ simple method of arriving at the

result (p. 26)—

a W = c P + F «'.

Suppose the compensating-lever to be as shown in Fig. 8 ; but for simplicity, with the points of attachment A and B in a straight line from X ; the cranking of the lever was of no consequence, so far as the calculations were concerned. Let the lever take some definite fixed position, say that of Fig. 6, when the engine was working smoothly ; in this position the lever might be supposed to be fixed to the gi'ound. Let the tensions of the lever on the brake- blocks, towards the left at B, and right at A, be represented by Tj and T3; then, since there was equilibrium, the sum of the moments round 0 the centre, of (a) friction of brake-blocks, (h) weight W, and (c) the tensions T2 and T3, was zero. Taking the lever now, not as fixed to the gi-ound, but as pivoted at X ; then E, the resultant of To and T3, must pass through X. T2 and T3 might now be replaced by R ; and the sum of the moments round 0 the centre, of (a) friction of brake-blocks, (b) weight W, and (c)

Proceedings.] CORRESPONDENCE ON FRICTION-BRAKE DYNAMOMETERS. 71

the force R, was zero. Eesolving E into vertical and horizontal Professor

Alexande Mr. Thomson.

components, V and P acting at the point X; then, since X was exan er an

vertically under 0, the line of action of V passed through 0, and its moment was zero ; and therefore the sum of the moments round 0 the centre, of (a) friction of brake-blocks, (b) weight W, and (c) the horizontal force P acting towards the left at X, was zero ; that

was —

aW = cP -f Fa'.

The amount of this horizontal force P could be easily measured by a sjiring-balance. With a low coefficient of friction, the tension on the brake-strap had to be increased ; and since the ratio existing between T2 and T3 was constant, depending on the proportions of the lever, it followed that P might be of considerable amount ; and any quantitative results calculated without taking it into account would be erroneous. With a high coefficient of friction the force P might be small, and the results might probably be not far wrong, even if P was left out of account. In every case, however, where accuracy was desired, the moment of P must be considered.

Mr. B. DoNKiN, Jun., submitted the results of some experiments Mr. Donkin. with an Appold brake made in 1876, to determine the pull on the compensating-levers. Por this purpose the fulcrum pins of the levers were removed, and the levers retained in jjosition by strings, of which one pair ran horizontally over jiulleys at d (Fig. 33) ; another pair was made fast at e, and a third was threaded round the shaft and carried vertically up to the handrail/. In working, the pairs e and /were put on as safeguards, those marked ad g being the only ones used. Before making an experiment the engine was run for about half an hour, till the wheel was well warm, the strings over the pulley at d being made fast at g, and the weights on the brake lifted by means of the adjusting screw k. The strings were then unfastened and weighted sufficiently to bring the load to the proper position, the centre a being vertically below the centre of the wheel. To effect this, it was necessary to adjust slightly the screw h and the weight at g ; and, by disregarding the position of the centre a, and screwing up or slacking out the adjusting screw /r, a greater or less weight at g would keej^ the brake central. The distance of the string a d below the centre of the shaft was 6| inches full in all the following experiments. With 10 HP. (200 lbs. at 2 feet 7.V inches radius) on the brake, the weight at g was from 10 to 12 lbs., 11 lbs. being apparently the best. With 7 HP. (140 lbs.) on the brake, the weight at g was from 6j to 7^ lbs., 6j lbs. being apparently correct. With .5 IIP. (100 lbs.) on the brake, the weight

72 COKKESPONDENCE ON FRICTION-BRAKE DYNAMOMETERS. [Minutes of

Mr. Donkiu. at g was from 4^ to 5^ lbs., 5 lbs. being correct. With the strings a' shown by the dotted line a' d' g', a being 1 foot 6| inches below the centre of the shaft, with 5 HP. (100 lbs.) on the brake, the weight required at g' was from 8.V to 9 lbs. Another experiment was made by fixing strings to the tops of the levers at a, 6^ inches below the centre of the shaft, and also another pair to the levers 4 inches above point 5, or 2 feet 2| inches below the centre of the shaft, the strings being led over pulleys as before, and either or both pairs could be used to keep the levers in position. At first the top strings were used and the bottom ones left slack. With 7 HP. (140 lbs.) on the brake, 3 to 4 lbs. on the string were sufficient

Fig. 33.

Experiments on Appold Brake with Ajios AcjrsTiNG Leters.

when the wheel was cool and the tallow stiff; after a time, however, 6 lbs. were required, and this was enough during the whole run of three hours. The top string was next slacked out and the bottom one brought into use, and the necessary weight hung upon it to keep the brake on its centre was 21 lbs. ; when the requisite load on each string was found, the strings were used alternately, the one not in use being looped up. The engine was afterwards run with the governor disconnected at 92 revolutions per minute, the top string being in use, and when the steam I)ressure was steady the lower string was brought into use instead of the top. The revolutions of the engine increased to 130 and dropped down to 120 per minute, and, to adjust the engine, 16 lbs.

Proceediugs.] COERESPONDENCE ON FRICTION-BRAKE DYNAMOMETERS. 73

were added to the load of 140 lbs. on the brake, while to bring the Mr. Doukin.

brake to the centre 1 lb. was added to the 21 lbs. on the lower

string, when the speed of 92 revolutions per minute was again

obtained. This was repeated several times ; also, when the weights

from the lower were taken on to the upper string, the speed was

much reduced. The point of attachment of the lower string was

next altered and brought 18^ inches below thg centre of the shaft,

when it was found that, although the speed of the engine varied

as the strings were reversed, the difference was not enough to be

measured by putting more weight on the brake, as the engine was

not running j^erfectly uniformly. Lubrication of the brake by

tallow made the weights rise a little, oil caused them to fall.

When an attempt was made to run the brake without the com-

pensating-levers, the addition of oil sent the weight down with a

force of 8 or 10 lbs., but tallow alone instead of oil sent them up

with a force of 2 or 3 lbs. An experiment by running the brake

without the levers was not successful, as the weight could not be

kept properly balanced for more than a minute or so.

Mr. Frank Garrett remarked that he did not attach the para- Mr. Garrett, mount importance to the subject of the construction of friction- brake dynamometers generally, which had been accorded to it by those who were possessed by the mania for engine-racing, and to whom the capacity of a dynamometer to register the last foot- pound appeared to be vital. He nevertheless esteemed the water- cooled brake (Figs. 12 to 14) very highly, as a most trustworthy, steady-going instrument, upon which he could rely for comparative results, whether with regard to the duty of steam-engines or to the steam-making capacity of coal. He had found it a most invahaable assistant when experimenting with the governor of a steam-engine, the compensating-lever acting admirably under the trying con- dition of removing the weights suddenly from the scale, sometimes one by one at intervals, and sometimes as quickly as possible, until the engine had been relieved of all its load. The experiment communicated by him to the Author had interested him greatly. He might perhaps claim to have established the fact that, with a water-cooled brake, a check, more or less approximate and reliable, upon the duty of the engine might be recorded by means of the heat-units transmitted to the water. In his experiments he had disregarded the heat-units absorbed by the increased temperatures of the brake-wheel, and these would probably go far to account for the difference existing between the foot-pounds recorded by the brake-wheel and the mechanical equivalent of the heat-ixnits trans- mitted to the cooling water as given in Table II (p. 14).

74 COKEESPOXDEXCE ON FRICTION-BRAKE DYNAMOMETERS. [Miuutes of

Professor Professor A. Jamiesox stated that about a year ago he had Jamieson. occasion to make a series of tests on a Griffin gas-engine. The brake proposed by the makers of this engine was the same as that used by the Eoyal AgricuUiiral Society described and com- mented upon in the Paper. Owing to the evident defects of the Agricultural Society's brake, one of the following form, Fig. 34, was adopted instead, which gave fairly good results with the gas- engine, developing 13 "6 brake HP. Thechief objections, however, to it were :— 1. That even for that small jDower, it was necessary to have two brakes, one brake upon each fly-wheel. 2. That the lubrication of the brakes required considerable attention. 3. That the back pull, indicated by the Salter's balances, varied considerably, and hunted up and do^vn within limits which necessitated some guessing and frequent observations. 4. That the oil or grease for lubricating the fly-wheels bespattered the floor, the wall opposite, and the observer's clothes, when reading the Salter's balances.

He had again, December 14th, 1888, had an opportunity of testing an identically similar gas-engine at Kilmarnock. This time he employed only one brake fixed on one of the fly-"R"heels. Fig. 35 illustrated this form of brake, which he understood was the same as that which had been used by the judges at the late gas- engine trials under the auspices of the Society of Arts. The following Table showed the more important results : —

Mean revolutions of brake fly-wheel per minute . . 205 Maximum deviation from mean speed, per cent. . . 5|

Dead load, W, in lbs 157

Mean back pull on balance, in lbs 4

Radius of dead load, W, from centre of brake wheel = r\ o . c:;9

1

1 2

in feet }

Size of each of the two small ropes, diameter in inches 0

Mean brake HP. during two hours' run 15

Gas-consumption per brake HP. in cubic feet per hour 21

indicated HP. „ „ 18

He considered this form of brake preferable to any one of the numerous forms that he had tried, and believed it could be adopted for large powers, and for long continuous runs for the reasons : —

1. It could be constructed on very short notice from materials always at hand in every factory or worksho]), and at very little expense. 2. It was so self-adjusting that no very accurate fitting was required. 3. It could be put on and taken oif in about one minute ; being very light and of small bulk it could be hung up or laid by in a cupboard. 4. It needed little if any attention for lubrication. 5. The back' pull registered by the spring-balance

Proceedings.] CORRESPONDENCE ON FRICTION-BRAKE DYNAMOMETERS. 75

was steady, and might be made a minimtim by properly adjusting Professor the load W before commencing the trial run. 6. The brake-wheel •^''"■"'i^^oii' soon attained such a maximum temperature that the radiating heat balanced the heat being generated by friction. 7. It might be

Fig. 34.

Pointer

^^\\\\\\\p\\\\\\^\^

used for small as well as for large powers, without any special attendant apparatus except a weight and a spring-balance. 8. For larger powers only more, or larger, or flatter ropes, or a larger brake-wheel, were required.

76 COEKESPONDENCE ON FKICTIOX-BrtAKE DYNAMOMETERS. [Minutes of

Mr. Schon- Mr. W. ScHONHEYDER wished to state that, having conducted a heyder. considerable number of brake-tests of engines, steam, gas and caloric, he preferred a brake similar to that illustrated by Fig. 16, with the strap clipping one-half, three-foiarths, or the whole cir- cumference of the wheel ; for smaller powers without any cooling arrangements, but for larger powers with the mode of cooling described, whereby water was continuously being run into the trough-shaped rim and continuously withdrawn therefrom. With such a brake the load P could be weighed with any desired degree of accuracy ; the horizontal distance of the load from the centre of the shaft could be measured to the greatest nicety, and the counter- pull of the spring-balance could be read as frequently as desired, or a diagram could be traced indicating all the variations of pull during the whole test. Not a single factor need remain un- determined, and therefore the amount of brake-power obtained must be undisputable. He could not agree with the Author that " In practice, generally speaking, the adjustment required by means of the screw S is as necessary with the compensating-lever as without it, and its value may therefore be questioned for this reason alone ;" for he had made several tests with brakes fitted with compensating-levers, and had run the engines for six hours and more withoiit once having to touch the adjusting screw ; but he did not consider this could have been accomplished without the continuous cooling arrangement, which permitted complete control over the temperature of the wheel, and therefore insured a uniformity of friction otherwise unattainable; the wood blocks M^ere slightly lubricated with oil. He believed the mode of cooling a brake-wheel thoroughly by water, contimially circulating through the wheel, was first practised by Mr. Halpin. As to using the Appold compensating-levers, more than twelve years ago he directed attention ^ to the grave errors which might be committed by not measuring the jiressure on the fixed point, and advised experiments to be undertaken to ascertain the probable amount of such errors. For a friction-brake, which he lately designed for a new engineering college, he placed the arms of the wheel at one side of the trough-shaped rim, boarded up the spaces between the arms, and overhung the wheel, so that all the water- siipply and take-off arrangements were at the one side only, and the splashing from the water would be kept away from the engine and shaft-bearing; the proportions were maximum brake HP. 150, revolutions 150 per minute, diameter of wheel 9 feet, width of

EngineaHng, vol. xxii. (1876), p. 75.

Proceedings] CORRESPONDENCE ON FRICTION-BRAKE DYNAMOMETERS. 77

wheel 10 inclies, maximum pressure per square inch of wood Mr. Schon- blocks 8 lbs, ; the straps were made of steel, very thin and flexible, "'^y^''^''- and they encircled the wheel for its whole circumference.

Mr. J. E. Sweet in practice, for the sake of convenience, reversed Mr. Sweet, the method adopted by nearly all other engineers. Instead of producing resistance by raising a weight, he caused it by pressure on the platform of a weighing-machine, at a point 5 feet 3 inches from the centre of rotation of the brake-wheel, or the radius of a 33-foot circle (Fig. 36). By this means the load could be changed hy simply sliding the weights on the beam. To ascertain the amount

Ficr. 36.

for any given HP., it was onlj'- necessary to make a single division. Thus, to set the brake for 15 HP. it was only necessary to divide 15,000 by the numl)er of revolutions made by the engine, and the result was the number of lbs. at which the weighing-machine must be set. The wheel was cooled by water ; the brake-blocks were white pine, though he believed hard wood was better ; cylinder-oil or salt pork was the lubricant. Notwithstanding there was no compensating-device, there was very little trouble. The service was simply making short tests to indicate and set valves, and to adjust governors at varying loads.

20 November, 1888.

Sir GEOEGE B. BRUCE, President, in the Chair.

The discussion on the Paper by Mr. William Worby Beaumont, on " Friction-Brake Dynamometers," occupied the whole evening.

78 WILLIAMS ON THE WITHAM [Minutes of

«

27 November, 1888.

Sir GEOEGE B. BRUCE, President, in the Chair.

(Paper No 2257.)

" The Witham New Outfall Channel and Improvement

Works."

By JoHX Evelyn Williams, M. Inst. C.E.

Probably few questions have received more attention from time to time at the hands of engineers than the improvement of the Eiver Witham.^ Following the observations of Kinderley in 1736, will be found, amongst others, the reports of Smeaton, Sir John Rennie, Sir William Cubitt, and Sir John Hawkshaw. In 1878, the Author, in reporting upon the subject to the Witham General Commis- sioners, derived much valuable aid from those earlier inquiries about the regime of the river.

The Eiver Witham rises near the village of Market Overton, on the northern boundary of Rutland, and thence flows in a winding and northerly direction, past Grantham, Long-Bennington, and Aubourn, to Lincoln. Here it passes over weirs at Bargate and at Stamp-End, and then proceeds in a south-easterly direction through the Fens to the Grand Sluice at Boston, and thence to the Wash (Plate 1, Fig. 1). The length of the Witham is about 89 miles.

At Lincoln, the Witham is held up for navigation in connection with the Foss-dyke, which joins the Trent at Torksey ; and its flow seaward is regulated by the weirs at Bargate and at Stamp-End. The water from Bargate weir, on the south side of Lincoln, passes through Sincil dyke into the south Delph, and enters the Witham immediately below Bardney lock. The water on the west side of Lincoln, joining the Foss-dyke navigation, passes from Brayford JVIeer, through Lincoln High-bridge, to Stamp-End weir and lock ;

> Previous communications on the Eiver Witham will be found in Minutes of Proceedings Inst. C.E., vol. xxviii. p. 59 ; and vol. Ixvii. p. 205.

Proceedings.] OUTFALL IMPROVEMENT WORKS. 79

and the water between this point and Bardney lock, a length of about 9 miles, is held tip for navigation. Between Bardney lock and the Grand Sluice at Boston, a distance of 23 miles, the Witham is also held up for navigation purposes, and a minimum depth of 5 feet is maintained.

The Grand Sluice is constructed across the Witham at Boston, and has four openings, the sea-doors or gates of which are self- acting, and close, so as to prevent the tide flowing further up the Witham (Plate 1, Figs. 1 and 8). On the tide receding, the sea- doors open with the pressure of the land water, which then passes off to the sea. On the land side of the sluice, land-doors are fitted, so as to regulate the fresh-water level and the scouring of the channel during dry seasons.

The area of land draining through the Witham Outfall is about 762,215 acres; but only about one-fourth of this area, or 194,649 acres, consisting of the six districts of the General Commissioners and the Black Sluice district, contribute towards the expenditure incurred in improving and maintaining the Outfall. To defray the costs of the new channel works these lands have been charged for a period of thirty-five years, from the 6th of April, 1881, with a uniform tax not exceeding 1«. per acre in any one year. The flood- water passes into the Outfall through the Grand Sluice, the Black Sluice, the Maud Foster Sluice, and the Hobhole Sluice (Plate 1, Figs. 1 and 2).

In 1878, the Outfall, or tidal j^ortion of the Witham, was most unfavourable, both for drainage and navigation. The re- lative capacity of the river, for the discharge of ordinary floods for a reach of 12 miles aT)ove the Grand Sluice, compared with the sluice itself and the Outfall Channel through Boston, was as follows: — Eiver Witham, 150; Grand Sluice, 100; and Outfall Channel, 75. The restricted section of the circuitous channel through Boston was further impeded by vessels moored in front of the several wharves. The head of water, necessary to overcome these obstructions in the Outfall Channel, increased the eleva- tion of the floods inland, and nullified, in a great measure, the benefit of all interior works of improvement. To remedy this, the Author proposed the cutting of a shorter and more direct channel for the flood-water, commencing in the river about 60 chains below the Grand Sluice, and terminating immediately above the Maud Foster Sluice, and that the circuitous loop of the old channel, thus severed, should be converted into a scouring reservoir or floating dock (see dotted channel, Plate 1, Fig. 1). This would have removed the shipping from the flood channel, and rendered

80 ^^^LLIAMS on the WITHAM [Minutes of

the 0\itfall to the estuary more direct. Owing, however, to a coutlict of opinion between the interests affected, this was not con- sidered advisable ; and ultimately the enlargement and improve- ment of the old channel through Boston was adopted. The most restricted section of this was immediately seaward of the Grand Sluice; and its enlargement was commenced in December, 1878 (Plate 1, Fio-. 4). The effect of this minor work of improvement in the Outfall was so apparent, that it furthered materially the o-eneral feeling in favour of the larger works that ultimately followed.

The channel between Maud Foster Sluice and Hobhole Sluice, a leno-th of 2f miles, was more favourable in direction and about 200 feet wide, but required regulating and dredging. The lower reach of the channel between Hobhole Sluice and the estuary was, however, not only shallow and circuitous, but untrained and broken through a mass of shifting sands (Plate 1, Figs. 1 and 2). The flood-waters rushing seaward, and impinging on the harder and unyielding strata of the Scalp, were deflected to the westward into the shifting sands, cutting frequent and successive steeps from 10 to 12 feet in height. These steeps kept tumbling into the channel and chokino- it up, until the channel ultimately became distorted and irretnilar, sometimes shifting a mile from east to west in a few weeks. The flood-tide from the estuary swept over these shifting sands, and rushed up the river to Boston, carrying a large quantity of sand in suspension. This was deposited, during slack-tide, in front of the sluices in the upper reach of the Outfall. In dry seasons, owing to the absence of the fresh-water scour, this deposit reached the level of 11 feet 6 inches above the sills of the Grand Sluice in Boston, and rendered the flow of neap-tides insensible at the sluice. In July, 1878, a tide which rose 21 feet in Clayhole, onlv rose 14 feet 11 inches at Hobhole Sluice, and 12 feet 6 inches at Boston. The duration of flow was five and a half hours in Clayhole, three hours at Hobhole, and two hours at Boston. The average rise of spring-tides in the estuary is 23 feet 4 inches, and of neap-tides 9 feet 2 inches. At low-water of spring-tides, when the river was swollen with flood-w^ater, the fall between the Grand Sluice and Clayhole, a distance of 8 miles, was 1 7 feet 6 inches, or an average fall of 26] inches per mile ; whereas the fall in the surface of the water, in a reach of 23 miles above the Grand Sluice, was only 4 inches per mile. This clearly indicated where works of improvement were most needed ; and the Author advised the cutting of a new and shorter outfall-channel to Clayhole, clear of the shifting sands, and also the dredging and

Proceedings.] OUTFALL IMPROVEMENT WORKS. 81

improvement generally of the Outfall between this new channel and the Grand Sluice. The cutting of a shorter channel to the estuary had been previously recommended, from time to time, by many engineers, including Sir John Eennie, Sir William Cubitt, and Sir John Ilawkshaw.

In August, 1879, the General Commissioners invited Committees from the trusts interested to meet and consider the matter. It was ultimately decided that the General Commissioners should promote a Bill in Parliament in the ensuing Session ; and though strongly opposed in Committee by the Eiver Welland Trustees, the Eiver Witham Outfall Improvement Act, 1880, was obtained. Under the provisions of this Act, the Witham Outfall Board was constituted, and, on the l-lth of December, 1880, the cutting of the new Outfall channel to Clayhole was commenced.

The new channel is curvilinear on plan, and its length from the commencement at Hobhole to the termination in Clayhole, in the estuary of the Wash, is 3 miles (Plate 1, Figs. 1, 2, and .5). The bottom was cut 100 feet wide at Hobhole, increasing gradually to 130 feet at its seaward termination in. Clayhole. The average bottom width of the channel is therefore 115 feet, compared with the Suez Canal of 72 feet, and the Amsterdam Ship Canal of 87 feet. The bottom of the new channel at Hobhole was cut to the level of 3 feet below the sills of Hobhole Sluice, or 1 1 • 20 feet below ordnance datum, with a gradual fall of 1 foot per mile to its termination in Clayhole. The depth of water in the channel at Clayhole is 27 feet 6 inches at spring-tides, and 20 feet 6 inches at neap-tides. The excavation through the enclosed land was about 23 feet in depth and 300 feet wide at ground level ; and the strata consisted chiefly of silt and of brown and blue clay, with intersecting patches of peat, under- lying which was a hard bed of boulder clay, m which the bottom of the channel was formed for about two-thirds of its length.

The excavations were carried on by means of three powerful steam-navvies, advancing in echelon or nearly abreast, together with numerous barrow-and- wagon roads, attended by eight locomo- tives and about seven hundred men. The material excavated was deposited in embankment on each side of the new channel, and a portion was utilized in forming the embankment to close the old channel. Outside the old sea-bank, the new embankments were extended simultaneously towards the low- water mark on the outer foreshore, and then united in horse-shoe fashion, so as to exclude the tides from as large an area of the work as practicable. By this means the middle and heaviest section of the excavation was executed in dry ground ; whilst the outer ends of the channel were

[the INST. C.E. VOL. XCV.] ij

82 WILLIAMS ON THE WITHAM []\[iiiutC8 of

completed by dredging. Between the edge of the cutting and the toe of each side embankment, there was left a flat or foreshore 60 feet in width.

The total quantity of excavation in the new channel was about 2,000,000 cubic yards, the contract price for w^hich was Is. per cubic yard, including dredging. The most suitable portion of the excavations was selected for forming the exposed slopes of the embankments, and fascine-work and cliff stone were used for the pro- tection of the slopes where necessary. The ends of the new channel were trained and protected by fascine-w^ork. In forming a training- wall, the fascines were laid in the water so as to overlap in heading- courses, and then weighted with a uniform layer of clsij so as to form a solid mattress, the width of which depends uj^on the depth and situation of the work. As each layer of fascines and clay sinks, another is formed, and so on until the level of half-tide is reached. The training-Avall thus constructed is finally covered with a layer of cliff stone, to protect it from the effects of abrasion and scoiir ^ (Plate 1, Fig. 7). The first sea-going vessel passed through the new channel on the 7th of April, 1884, and the permanent closing of the old channel was then actively proceeded with. To effect this, it was necessary to form an embankment i mile in length across the old channel and foreshore, below Hobhole Sluice, con- necting the parishes of Fishtoft and Wyberton. The maximum height of this embankment was 35 feet, and its top was formed to the level of 8 feet above ordinary spring-tides, with a minimum top width of 15 feet, and side slopes of 5 horizontal and 1 vertical, faced with clay puddle 3 feet in thickness. The embankment was pushed forward simultaneously from the enclosure banks on each side of the old channel ; and when the two ojiposite tip-heads had advanced within a distance of 10 chains of meeting, two parallel groynes, formed of fascine-work and cliff stone, were run out on each side of the old channel. These groynes afterwards formed the inner and outer toe of the new eml)ankment across the fair- way of the old channel ; and as they advanced towards meeting, the scouring energy of the tidal water increased so rapidly, that it became necessary to load the gi'oynes with old rails before effecting the final closure ; this was successfully accomplished on the 29th of August, 1884. The advantage of the new Outfall then became obvious, by the strengthening and concentration of the ebbing current, and the general depression in the low-water level. The incoming tides, instead of rushing up the river with a bore

' " Faseinn-Work at the Outfalls of the Fen Rivers, and Keclamatiou of the Foreshore." By W. H. Wheeler. Minutes of Proceedings Inst. C.E., vol. xlvi. p. Gl.

Proceedings.] OUTFALL IMPROVEMENT WORKS. 83

as formerly, heavily charged with matter in suspension, now flow in gently from the estuary comparatively clear ; and vessels above 2,000 tons navigate the new channel with ease, compared with vessels of 300 tons, the largest the old channel could accom- modate.

In addition, notwithstanding the exceptionally dry summer of 1887, and the consequent absence of scour down the Witham, the accumulation of dejiosit by the tides only reached 6 inches at the Grand Sluice ; whereas in 1874, with a larger rainfall, it reached 11 feet 6 inches, and neap-tides were insensible at the sluice.

The Outfall between the new channel and the Grand Sluice has been improved generally by training-works and dredging, and the easing of the abrupt bends through Boston. The channel is now 8 feet deeper than it was in 1878, and there is a navigable depth of 22 feet at spring-tides up to the Boston Dock, con- structed by the Corporation of Boston under their Act of 1881. The restricted section of the Outfall through Boston has been enlarged and improved by dredging and other works ; and a very contracted portion of the channel, where the houses abutted immediately upon the river, has been widened by the construction of a breastwork of timber piling (Plate 1, Figs. 4 and 6). These works have been carried out by the Witham Outfall Board from the designs and under the direction of the Author.

In contimiation inland of the improvement of the Outfall, the Author advised, in his report of 1878, the enlargement of the Grand Sluice, and also the widening and improvement generally of the Witham for a reach of 12 miles above the sluice. Owing to the seaward flow of the flood-waters being arrested during high-water in the Outfall, the river above the Grand Sluice may be termed, in addition to a conduit, an elongated reservoir for the storage of the flood-waters diiring the period of suspended dis- charge through the sluice. During this interval, at spring-tides, the flood-water sometimes rose as much as 5 feet on the land side of the Grand Sluice. It was therefore evident, that the enlarge- ment of the storage capacity of the channel, above the sluice, would tend to diminish the rise of the flood-water during this period. To effect these improvements, the General Commissioners obtained the necessary powers in 1881 ; and the works were undertaken forthwith.

The Graiid Sluice was designed by Langley Edwards, and com- pleted in 1766. The sluice had four openings, three 17 feet wide and one 15 feet wide. The latter was a lock, but it was also available for the passage of the flood-waters when necessary.

G 2

84 WILLIAMS ON THE ^^^THAM [Minutes of

The enlargement of the sluice comprisefl the demolition of this lock, and the construction of a wider and deeper lock on its site (Plate 1, Fig. 8). The sills of the Grand Sluice were 6-80 feet above low-water of ordinary spring-tides in the estuary, 8 miles distant ; and the area draining through the sluice was about 500,000 acres, only 34,726 acres of which, situated Lelow Lincoln, contributed to the cost of enlarging the sluice and the improve- ments of the river, under the -Act of 1881. The maximum flood recorded at the Grand Sluice was during the winter of 1876-7, when the flood-levels above the sill were 15 feet 11 inches at high- water, and 10 feet 11 inches at low-water. Ordinarj^ spring- tides rise 16 feet 6 inches above the sill, and neap-tides aljont 7 feet less.

The new lock is 30 feet wide, or double the width of the old one ; and the sills are 3 feet lower, or 3 • 80 feet above low-water of ordinary spring-tides in the estuary (Plate 1, Figs. 9 and 10). The excavations were carried down to the hard clay, upon which a bed of concrete was laid, 7 feet in thickness. The concrete was composed of 6 parts of ballast to 1 part of Portland cement. Upon this the floor was formed, consisting of 3 parts of ballast and 1 part of Portland cement. The pointing-sills were formed of sandstone ashlar, 3 feet in thickness, with radiating joints on plan so as to project into and bond with the concrete floor. The heel-stones, to receive the gate-pivots, were formed of granite, 5 feet by 4 feet, and 18 inches thick; and the quoins and copings were formed of sandstone ashlar. The hollow quoins were 4 feet by 4 feet, laid in alternate courses of header and stretcher so as to bond with the brickwork in the face of the side walls. The coping of the lock Avails is 12 inches thick, drafted and picked. The Portland cement was specified to be of such quality that a test-bar, moulded 1.', inch square, or 2j inches in sectional area, should, after seven days' setting in water, bear, without breaking, a tensile-strain of 500 lbs. During the progTess of the Avorks one hundred and five tests were made, and the average breaking-strain was 674 lbs.

The doors, or gates, six in number, were built of English oak (Plate 1, Fig. 11). The outer pair are self-acting, and close so as to prevent the tides floAving up the Witham. Their height is 24 feet 6 inches, and the span between the centres of the juA^ots is 32 feet 10 inches. The heel and nutre-posts were formed out of timber 18 inches square ; and the top and bottom ribs are 18 inches by 12 inches at the ends, and 21 inches by 12 inches in the middle. The intermediate ribs are 18^ inches bj^ 12 inches, tenoned into the mitre and heel posts, ajid secui'ed Avith doiible Avrought-iron

Procccdiugs.]

OUTFALL IMPROVEMENT AVORKS.

85

straps 4 inches hy j inch, and bolted through the ribs and posts with 1-inch bolts. The planking is of oak, 2^ inches thick, tongned and grooved ; and the gates turn freely ujion their j^ivots without rollers. The top of each heel-post is secured to the side wall by a wroiight-iron collar bolted through a cast-iron block forming the top hollow quoin. The inner gates are fitted with valves for scouring and locking purposes ; but during floods all the gates of the Grand Sluice are thrown open, and secured back in recesses formed for them in the side walls. The four openings of the sluice now form a waterway 81 feet in width for the free passage of the flood- .waters to the sea. The new lock was opened on the 8th of December, 1883 ; and the beneficial effect of the enlarged and lower waterway was immediately felt in the upper reach of the river.

The widening and improvement generally of the Withara between the Grand Sluice and Tattershall Bridge, a reach of 12 miles, consisted of the removal of the continuous projecting flat or foreland of the banks on each side of the channel (Plate 1, Fig. 3). These projecting forelands, below flood-level, were a source of expense and obstruction, owing to the large quantity of flag that grew upon them, thereby virtually raising the bed of the river, and obstructing the seaward flow of the flood-water. The greater portion of these forelands were removed for the entire length of 12 miles of channel; the waterway was thus in- creased in width about 50 per cent, at the ordinary navigation level ; and, in addition, the shoals were removed by dredging, and the bends in the channel made easier. The slopes next the channel were nowhere less than 3 to 1 ; and the excavated material, con- sisting of clay and silt, was deposited chiefly so as to enlarge and strengthen the flood-banks. Fascine-work was used for the pro- tection of the river slopes at points where the strata were very silty and yielding in character. The total quantity of the excava- tions in this section of the works was 602,059 cubic yards, and the average cost was 9 • 2d. per cubic yard.

The works have altogether efiected a material improvement in the drainage and navigation. Not only has the depth of the navigable channel of the port of Boston been increased 8 feet, and the accumulation of deposit by the tides become almost nil ; but the absolute gain or dejiression acquired in the low-water level at the several sea-sluices is as follows : —

	Ft.	In.		Ft.	In
Hobhole Sluice .	. 5	6	Black Sluice . .	. . 4	0
Maud Foster Shiice	. 4	3	Grand Sluice .	. . 4	0

This large depression in the Outfall practically means the raisinj

86 WILLIAMS ON THE WITHAM [Minutes of

of the Fen lands, and an immense benefit to the whole of the districts and interests concerned.

The works were completed in May, 1887, and the total expen- diture was approximately as under, viz. —

£.

The New Sea Channel and Outfall Improvements"! ,gg qqq Act. 18S0 J '

The Enlargement of the Grand Sluice and Eiver"! ^, „qq Witham Act, 1881 /

212.000

The plans and sections were presented to Parliament by the Author, and the works were afterwards designed in detail and carried out under his direction as Engineer-in-Chief.

The Contractor for the new channel was Mr. Thomas Monk ; and the dredging of the iipper reach of the Outfall was partly executed by him, and also by Mr. W. N. Smith. The Contractor for the Grand Sluice Works was Mr. William Eigby; and for the river piling, Mr. Samuel Sherwin. A portion of the dredging and fascine- work, together with the whole of the works in connection with the enlargement and improvement of the river above the Grand Sluice, were carried out without the intervention of a Contractor.

The Paper is accompanied by six sheets of tracings, from which Plate 1 has been prepared.

[Appendixes.

Proccodiugs.] OUTFALL IMPROVEMENT WORKS. 87

APPENDIXES.

I. Flood Keport. — 1883. To the General Commissioners for Drainage by the River Witham.

My Lords and Gentlemen,

I regret to report that since your last Meeting the district under your juris- diction lias been subject to a rainfall unparalleled for its severity in the history of the Witham Drainage.

The rainfall registered at the Grand Sluice, Boston, for the twenty-four hours ending 9 a.m. on the 30th ult. was equal to 310 tons per acre. This extra- ordinary rainfall is 143 tons per acre in excess of the maximum for the same period during the previous ten years, and 242 tons per acre in excess of the rainfall of the disastrous flood of 1877.

The water-level in the Witham at Bardney rose 11 feet in twenty-four hours, the maximum flood-level being 18 feet 8 inches on the 30th ult.

Considering the trojiical character of the rainfall, I feel convinced that, were it not for the extensive improvement works now well advanced in the Witham below Chapel Hill, the previous maximum flood-levels would have been ex- ceeded and disaster would have ensued.

In the East Fen the rainfall was even more excessive than that registered at the Grand Sluice. For the twenty-four hours ending 9 a.m. on the 30th ult. the rainfall at Lade Bank was equal to 357 tons per acre, or fourteen times tlie maximum quantity assumed to be pumped in the same time by the Lade Bank engines.

When this enormous downfall came upon the district the gi'ound was already saturated with a rainfall of 15G tons per acre during the previous week.

The upland waters came pouring down in immense volume, and the flood- level in the catchwatcr drains rose far above any previous recorded height. These drains rapidly filled, until the flood-waters overflowed their banks at various points.

Steeping Eiver, with a watershed of nearly 30,000 acres outside your juris- diction, became surcharged with the flood-waters from the uplands, which rapidly topped its banks and flowed over them into the Fen, in some places 12 inches deep.

Much of this flood-water found its way into the upper end of the East Fen catchwater, already surcharged and overflowing at intervals. Two small breaches occurred in the left bank on Sunday, the 30th ult., the more important of which I saw practically closed in a few hours.

In the interval a breach occurred in the right bank of Steeping Eiver above Thorpe Culvert, which I had successfully closed the following morning.

Several dangerous slips had already taken place in the banks of Steeping River, and it was only by cradging where practicable and other protective measures that more serious disasters were prevented.

Under the exceptional conditions the Lade Bank engines are performing their duty efficiently, but the flood-water in Hobhole drain has flowed back over the doors in much greater volume than heretofore.

88

WILLIAMS ON THE AVITHAM

[Minutes of

It is perhaps unnecessary for me to add that a rainfall of 357 tons per acre in twenty-four hours, or more than 60 per cent, in excess of the average rainfall of a whole month during the previous ten years, represents a sudden and ex- ceptional strain far in excess of the capacity of the works under your jurisdiction.

They were not designed on a basis to meet so unprecedented an emergency, and the utmost that can be done under such appalling conditions is to minimise disaster.

I cannot conclude this Keport without gratefully acknowledging the valuable and timely co-operation so cheerfully extended to me by the Chairman of the Fourth District Commissioners, and other gentlemen, during this disastrous flood. — I have the honour to be, my Lords and Gentlemen, your obedient servant,

JOHN EVELYN WILLIAMS.

Witham Office, Boston, 5th October, 1883.

MAXDirM D.MLT RaINFALL.

1873. July.	1874. July.	1875. Aug.	1876. June.	1877. Jan.	1878. Aug.	1879. Aug.	1880. Sept.	1881. July.	1882. Oct.	1883. Sept.
Inch.	Inch. 0-90	Inch. 1-67	Inch. 1-40	Inch. 113	Inch. 0-91	Inch. 1-10	Inch. 1-62	Inch. 1-06	Inch. 1-53	Inch. 310
117

Note. — The 3-10 inches for the twenty-four hours ending 9 a.m. on the 30th September was preceded by a rainfall of 1 • 56 inch the previous week.

The rainfall during September was 6 '78 inches, the greatest monthly fall registered.

II. Flood Eepoet.— 1885. To the General Commissiojiers for Drainage by the Uiver Witham. My Lords axd Gentlemex,

Since your last meeting the whole of the districts under your jurisdiction have been subject to a long and excessive rainfall.

The rainfall during October was 419 tons per acre, or 4G per cent, in excess of the average of the preceding ten years. The maximum daily fall was 155 tons per acre on the 23rd, and during the last week in the mouth there was a fall of 246 tons per acre.

This was followed by a rainfall of 214 tons per acre during the first six days

Proceedings.]

OUTFALL IMPROVEMENT WORKS.

89

in November. I am pleased, however, to report tliat, comjjared with simikir conditions of rainfall, and tidal oscillation, the depression in the level of the flood-waters was conspicuous and general.

During the low-water interval the depression compared with the previous maximum level was 3 feet 9 inches at Hobhole Sluice, 2 feet 4 inches at the Grand Sluice, and 2 feet 1| inch at Bardney Lock.

In the East Fen, at the Lade Bank engines, the water only rose 2 inches above the summer level.

This favourable condition of the drainage system is no doubt due to the Works carried out under "The Witham Outfall Improvement Act, 1880," and "The Witham Drainage Act, 1881."

With the improved and shorter Outfall Channel to Clayhole, together with the enlargement of the Grand Sluice and the Witham below Chapel Hill, a much greater volume of flood-water passed off between tides, and at no time was disaster imminent as heretofore. — I have the honour to be, my Lords and Gentlemen, your obedient servant,

JOHN EVELYN WILLIAMS, Engineer.

Witham Office, Boston, 14th November, 1885.

III. — Witham Outfall. Tidal Curves at Boston before and after

COMPLETION of NeW ChANNEL. — FiG. 1.

4.A.II £

C

TIME 8 " 9 10 tl 12

2 3 Ira

!

?*

hm:o

sr H3

34-

/

r^

\,

.-/

/^

- ^

7Q

.^y ■ ,

N

\,

^'^ SI

N

, \,

il $1

\ \

Q/ W

N "

HM' 0

iWcn^js

16 27

\

\

IS'

b/ V

N

\

1 f/

\\

\ ^

7 'V

s.

\

./ ^

k\

10

\ ^J

^

J

1

kr;

/

LW C

yeifi

; 707

\

/

X

s

/

^^^

<;

y

^|-~-,,_

Lo

•c«(. /,

M' at if

Wr.rk\

zoo

0

DATU

M UNEJ.X W(_

Sia:,h

"<<■-- or

lO f.^^

Mfjy^

Qrd^

inrv Ai*"^ 1 ,

0

90

WILLIAMS ON THE WITHAM

[Minutes of

O

C »H =

« r^

.

c;

O

-+I

^

CO

CO

CO

lO

^^

<M

t-

CO

t^

CO

o

i^

lO

-t<

CO

lO

-ti

c-

l>

C3

00

^

,__)

o

,_4

J.^

1— (

o

o

PH

i-H

I— 1

1-^

o

-tl

-^

.

t>

»

,^

Tj

00

"O

CO

CO

CO

C5

t-

l>

00

«

CO

CO

■^

CO

-tl

o

o

CI

CO

CO

CO

«>

CO

o

(M

o

(M

^

'"^

^

<M

p^

o

IM

CO

CO

CO

"

(M

.

tr>

CO

C5

CO

o

CO

Ci

-t<

t-

(33

o

CO

^

00

S;

OO

00

t-

w

CO

lO

CO

o

'T'

CO

05

CO

,__l

p_l

O

y—i

,.H

(M

o

,— t

CO

'tl

CO

©

CO

^

IM

.

M

,

CO

^^

CO

■M

CO

(M

oo

CO

CO

-+<

«

CO

o

a

a

CO

(M

lO

(M

CO

CO'

CO

CO

CO

'"'

o

o

'"'

'^

o

<N

o

(M

'"'

'"'

(M

CO

_

.

ro

o

C<I

l>

CO

o

-+i

CO

00

CO

00

,_!

CO

o

I<I

S5

-*l

lO

<jj

CO

OO

o

t^

in

00

CO

lO

^

1-1

<M

^^

,__f

iM

CO

C<l

r_^

CO

!M

IM

,_f

-M

'^

^

CO

00

oo

M

O

C-l

CO

C5

(M

05

1—1

^

IM

CO

lO

00

>o

o

CO

(M

00

C5

<N

CO

t-

CO

00

T-l

■—1

00

^

1-H

^

(M

^

cq

cq

(M

CO

■*

(M

CO

s

o

00

o

lO

^^

CM

CO

CO

-tl

t^

■*

CO

CO

00

(M

o

t^

O

-+I

IM

tO

IM

CO

CD

o

CO

,_,

>-(

o

M

CO

-h

IM

S<1

IM

(M

CO

'-'

(M

o CO

rr

-+<

CO

(M

»c

■*

"+1

CO

CO

CO

iM

CO

CO

CO

<N

i>

c»

o

o

o

C5

-fl

00

l>

CO

a>

'*<

oo

=

o

■"*

o

CO

'"'

'i^

lO

!M

Tj<

CO

1—1

1—1

-tl CO o

o

^

o

o

o

l^

(>J

-^

o

o

no

-tl

^

(M

o

CO

—I

<M

O

-tl

(M

CO

Tj

"-I

CO

o

I— 1

IM

o

(M

CO

CO

CO

CO

CO

O

^.^

o

o

IM

CO

.

r^

t^

CO

o

CO

00

CO

OO

CO

IM

CO

CO

00

Tt<

O

00

o

CO

C5

(N

CO

CO

^

00

CO

o

•H

»-H

'"'

'"'

o

■"^

CO

1— 1

'"'

lO

1—1

(M

CO

'"'

lO IM

1-^ 1^

3?

,^

^

CI

^

CO

l^

iM

o

t-

CO

CO

CO

^

2

o

^

CO

'-'

CM

<Xi

t^

Ci

l>

"O

»o

1-1

(M

(M

^

CO

'"'

■"^

(M

!M

r-(

(M

'"'

'"'

-tl (M 1—1

CO

(T

^

O

CO

00

-+I

IC

^+1

CO

CO

CO

CO

CO

2

■—I

O

CO

— i

l^

lO

»o

CO

t-

(M

t-

IC

CO IM

•S

<M

(M

<>]

CO

o

CO

^

'-'

tH

^

IM

CO

C2 IM

ITS

.

CO

-*i

<M

CD

CO

C5

CO

CO

^_^

CO

05

-tl

O

05

lO

-+I

■*!

o

Ci

Ci

lO

IM

CO

00

00

m

00

tH

'-'

o

o

o

I— 1

CO

IM

IM

CO

■*

o

(M

?i

a:

o

(?^

c;

^_,

o

e<j

-H

-tl

lO

CO

O

lO

C3

■*

«

cc

00

i^

O

t-

o

o

C5

CO

O

i-(

-

'"'

'"'

o

o

o

o

iM

^

(M

0-1

'"'

IM

OO

CO

CO

CO

^

00

-H

^_^

o

o

03

CO

00

>o

00

■*!

CO

CO

o

uo

O

00

C5

IM

IM

00

ri

1— t

,—1

,—1

o

(M

r— (

CO

IM

1—1

.^

r_i

o

C3

"■

'"'

>.

o

o

A

S 1-5

^

O

^

CO Ei

a o CO

"o o o

g o

rO g o o o o

< EH

Proceedings.]

OUTFALL IMPROVEMENT WORKS.

91

00

CO

CO

CO

t-

rji

(M

-H

CO

-tl

t>

t<

^^

^

CO

•-I

(>J

CO

CO

(M

>o

CO

00

eo

(M

1-1

'-

o

o

o

o

O

o

O

o

o

o

o

o

CO oo

o

CO

CO

<M

CO

t^

t^

o

IM

CC

o

o

/3

-tl

o

CO

C^

cc

CO

CO

(M

o

Cj

o

00 r-l

o

o

o

o

--^

o

o

o

O

o

o

-^

US 00

(M

-+I

o

oo

(M

t^

CC

(M

iO

1«

(N

o

-a

■*!

iO

CO

Tt<

CO

Oi

■*

CO

lO

00

eo

00

c

o

o

o

o

o

o

o

o

o

'-'

o

o

00

lO

o

CO

lO

oo

(M

^

r-

(M

(N

^

CC

ja

CO

Tt<

CO

Ol

CO

(N

CO

<M

O

(M

N

tl

rH

o

o

o

o

o

O

o

o

o

O

o

o

CO DO

r^

t-

o

t^

cq

>o

-+<

!M

o

o

^

CO

ja

CO

-+1

CO

eo

CO

■*!

CO

t^

lO

C5

"O

00

B

o

o

o

o

o

^

o

o

eo

o

o

o

00

o

-H

,_,

lO

CO

o

o

CO

o

CO

lO

o

J3

lO

CO

t-

lO

CO

o

tl

l^

00

»o

CO

CO

00 rH

C

o

o

o

o

o

o

o

o

o

^

o

o

TfH

m

CO

o

^

CO

CO

-H

o

f

iM

o

GO

ja

CO

in

CO

iM

CO

o

o

tl

»o

tl

o

OO 1-1

-^

o

o

o

o

o

o

'-'

^

o

o

o

o

00

00

o

lO

l>

-H

l>

(35

o

(M

t^

eo

CD

•s

o

CO

(>J

lO

CO

C5

>o

lO

CO

»o

t<

(M

00 T-l

B

o

o

o

o

o

o

l-(

o

rH

rH

o

o

^

t^

I>

o

Ci

t-

o

-n

o

CO

CO

t<

,

J

-^^

-ft

l^

CO

CO

CO

c^

IM

IM

1-t

o

o

o

o

o

^

o

o

O

o

O

o

00 i-

1— 1

CO

lO

o

o

CO

lO

1—1

-+I

oo

Oi

^

^

-tl

(M

CO

lO

CO

oo

•^

Cft

CO

tl

00

•o

00

S

o

o

o

o

o

o

o

o

o

o

o

o

1^

CO

tH

CO

^

CO

<M

CO

t^

l>

t-

o

,_,

•s

1— (

l>

OJ

CO

IM

CO

lO

lO

CO

CO

CO

tl

00 iH

B

'-'

o

o

o

O

o

o

o

o

o

o

o

CO 1:-

o

CO

CO

CO

C5

o

CO

CO

o

oo

tl

^

J

00

CO

"O

-ri

o

CO

o

CO

lO

"^

i-«

C

o

o

o

'"'

o

'~^

'"'

o

o

o

o

o

^

-tl

t>

-*<

,_,

rji

t-

b-

CO

t<

t^

tl

J

CO

CO

I— 1

(M

-+i

CO

CD

CD

eo

00

(N

00 rH

B

o

o

o

o

O

O

o

'-'

o

'-'

o

o

^

o

^

o

l>

o

_M

o

o

^

t--

CO

C3

"2

l>)

lO

<M

c^

<M

CO

Ol

eo

CO

co

CO

t<

t^

B

o

o

o

o

o

o

o

o

o

o

o

o

eo

CO

t-

t-

00

lO

o

l>

»o

(N

<M

05

00

ja

lO

CO

CO

•*

05

f

cq

o

00

a

o

o

o

o

o

O

rH

rH

o

o

o

o

S OS

ci >-

ci

<

1-5

3

o

o "S o

hi o a o

3 a § o Q

[Discussion.

92 DISCUSSION ox THE WITHAM [Miuutcs of

Discussion.

Mr. J. EvELYX Williams said that on the completion of the new channel to Clayhole, which not only dried the sills of the Hobhole Sluice, but effected a depression of 5 feet 6 inches in the low-water level at that point, it became evident that before the fourth district of the Witham drainage, nearly 74,000 acres, could obtain the full benefit of the new outfall channel, an enlarged and lower outlet at Hobhole Sluice was absolutely necessary. The General Commissioners accordingly applied to Parliament, and obtained the necessary powers under the " Witham Drainage Act, 1887." This Act enabled the Commissioners to apply the surplus funds of the fourth district, acquired under the " Witham Outfall Improvement Act, 1880," to the jnarposes of the works, so that no drainage taxes were levied to meet the expenditure incurred in constructing the additional and lower outlet at Hobhole. He had just carried out that work, and as the sill was 3 feet lower than the sills of the old sluice, the result was that a large area of low land had now for the first time natural drainage, and the pumping- engines at Lade Bank, which indicated 400 HP., erected under the Act of 1867, had practically become so much power in reserve.

Mr. J. Clarke Hawkshaw thought every engineer who had had anything to do with the Witham, must be gratified at seeing the new outfall carried out at last. The old saying, of putting the cart before the horse, could now no longer be applied to the works on the Witham. It was not the fault of engineers that this great improvement had not been carried out before. Mr. Lewin, in his report in 1860, said that the matter had then been under agitation for one hundred and twenty years. The reason why this new outfall had not been made before was a consequence of the divided jurisdiction which existed on the river. In a Paper, on river-control and management, which he had read before the British Association at Belfast, he quoted the Witham as an example of a river where such divided jurisdiction existed. In that Paper he described in some detail the jurisdiction of seventeen prin- cipal sets of commissioners who ruled on the Witham, between Grantham and Lincoln, and the Trent and Boston. He knew the district pretty well himself. When Sir John Hawkshaw was called upon to report on the drainage of the Witham in 1877, Mr. Clarke Hawkshaw had spent some weeks there with Mr. Vernon-Harcourt, and he traversed nearly the whole district on foot, map}>iug out the area that was flooded at the time. In

Procccdiiif^s.] OUTFALL IMPROVEMENT WOEKS. 93

Jannary of that year there were about 40,000 acres under Jir. Il.iwk- water, inchiding nearly 100 acres of the city of Lincoln, parts of '^^''^^'' which were flooded to the dei)th of 3 feet. He did not know of any work which would tend more to i)revent such a disaster occurring in future than the new outfall. As early as 1861, his father had recommended it, and had always referred to it in subsequent reports as one of the greatest improvements that could be carried out. He thought the Author was to be congratulated that it should have fallen to his lot to carry out such an important work, which engineers had been trying to get done for nearly one hundred and fifty years.

Sir Charles A. Hartley, K.C.M.G., observed that the results of SU- Chark-s the work described in the Paper had been three-fold. First, the ^■'^^'''.^'• navigable channel of the Witham outfall had been deepened 8 feet by dredging, so that there was now a minimum depth of 23 feet at high-water of spring-tides, and of 10 feet at high- water of neap-tides, over a minimum width of 100 feet through- out the 7 miles of improved channel between Boston and the sea. Secondly, the drainage of a large area of the fen lands had been very materially benefited by the widening of a reach of 12 miles, immediately above the Grand Sluice, by the enlargement and lowering of the Grand Sluice itself, and by the depression ac(pired in the low-water level of the three sea sluices between the Grand Sluice and the head of the new cut at Ho1)hole. Lastly, and this was a highly interesting fact, the accumulation of deposit by the tides below the Grand Sluice had become almost inaj^preci- able since the completion of the works. These gratifying results afforded a striking illustration of the great advantages to be gained, both as regarded improved drainage and navigation, by commencing the correction of a tidal river of the Witham type at the sea outfall, and by carrying the low- water level of spring-tides as far inland as practicable ; that was to say, as far up-stream as the cost of the work would seem to be justified by the attainment of the end in view. The Author had made no mention of the changes which had taken place in the bed of the Estuary at Clayhole since the completion of the new cut. It would be interesting to learn from him the importance of the changes. For his own part, Sir Charles Hartley was of opinion that, to perpetuate the improve- ments already effected, it would be necessary to prolong the banks of the new channel seawards, from time to time, to kee}) j)ace with the advance of newly formed sandbanks at the mouth of the river, and to ensure the free flow of the ebb current into deep water. He desired to express his admiration of the skill and energy

94 DISCUSSION ON THE WITHAM [Minutes of

displayed by the Author in the rai^d execution of a difficult and important work at a comparatively small cost.

Mr. L. F. Verson-Harcourt observed that he had had the oppor- tunity, as Mr. Clarke Hawkshaw had remarked, of making an intimate acquaintance with the Witham in 1877, and he also had the advantage of seeing the works during their construction in 1883, and of going into the new cut when the excavators were at work, and through which the river now flowed. It was not the first time that the question of the Witham had come before the Institution, for twenty years ago there was a discussion upon it, and in that discussion the two chief points discussed were the question of the removal of the Grand Sluice, and the outfall cut which had now been made. There were considerable diffi- culties in connection with the question of the removal of a sluice which had once been established. There was the difficulty of arranging about the levels of the water, the adequate strengthening and raising the banks along each side to secure the land against an inroad of the sea, and of introducing a large amount of tidal water into a part where the tidal water had been excluded ; and it would seem that it was the more natural thing to carrj^ out the enlargement of the Grand Sluice. The great difficulties under which the Witham had always laboured, had been the small amount of fall from Lincoln down to the Grand Sluice, and the impeded shallow shifting channel towards the sea. It was im- possible to get a greater fall along that part of the Witham than 4 inches in the mile right down from Lincoln. Above Lincoln there was a greater fall, therefore the water came down to Lincoln more rapidly ; and in Lincoln itself, at the present time, there was a great contraction of the channel and considerable obstruction to the escape of the flood-waters. It was above and a little below Lincoln that the chief flooding occurred in the winter of 1876-7, as de- scribed by him in a discussion on the " Conservancy of Eivers," ^ in 1881. At that time Sir John Hawkshaw proposed, amongst other means, to make a cut, away from the jiresent channel through Lincoln, along some waste land which could be got outside Lincoln, for carrying the flood-waters of the Brant and of the Witham down to the South Delph, which was an artificial cut, carrying the drainage waters of the Witham. Another defect in the Witham Avas that the fresh water, which would naturally go down the main stream, was, from the necessities of drainage, diverted to several sluices and drains, such as the Hobhole, the Maud Foster, and the

' Minutes of Proceedings Inst. C.E.. vol. Ixvii. p. 283.

Proeecdings.] OUTFALL IMPROVEMENT WORKS. 95

Black Sluice. Therefore, the waters that should naturally be ^Ir- ^'l^l■^on- scouring the upper portion of the tidal channel were diverted to '"^'^"' • other places. There was also in those days, the great imjiediment of the outfall, as well as the very narrow and inadequate channel through Boston itself. The large projecting quay walls and jetties as he saw them in 1877 were in a very neglected state. Then there was the very sharp bend through Boston, which the Author had proposed to change by introducing what would have been an excellent thing for the river itself — the cut across it — but that, of course, would have altered existing arrangements considerably, and it might not have been acceptable to the Black Sluice Com- missioners. Moreover there was the difficulty in dealing with the Grand Sluice. The Grand Sluice not only provided a very inadequate waterway in flood-time, but also was an absolute im- pediment to the proper flow of the tide, which in old daj^s was supposed to have flowed up as far as Lincoln. But even in 1877, Sir John Hawkshaw did not propose taking away the Grand Sluice, as had been suggested before, he merely suggested an enlarge- ment ; and at that time it was proposed to increase considerably the sectional area of the Witham above up to Lincoln, so as to enable it to carry oft" the flood- waters, because it was thought that it was practically almost impossible at that time to induce the Witham Commissioners to approve of the outfall scheme which had previously been unsuccessfully urged upon them. If the enlargement of the Grand Sluice proposed by Sir John Hawkshaw had been carried out, there would have been 110 feet width of opening in place of the present 81 feet, and the sills of the sluice were all to be lowered 7 feet ; in fact, the sluice, according to his proposal, was to be rebuilt a little higher up than the existing sluice, and its water-way more than doubled. There could be no doubt as to the great advantages accruing from the alteration of the outfall. It was an immense advantage to the river itself; l)ut, of course, a great disadvantage to the Eiver Welland, that river having previously had the benefit of the scour of the Witham, which it had now lost. In most of the schemes proposed for the improvement of the Witham, there were also training- works proposed for the Eiver Welland to join it at Clayhole. He believed it was want of funds that prevented the carrying out of the training-works ; these would be very desirable, in order that the Witham and Welland, as in former times, but in a better direction, might have the same outfall. The Author had compared the new channel with the Amsterdam canal, and the Suez canal in section ; bxit it should be borne in mind that

90 piSCUSSION ON THE AVITHAJH [Minutes of

they were for different purposes. The Amsterdam canal had to be made 16 miles long, just sufficient in section for the traffic carried through it; whereas the Witham outfall was for the tidal flow and the discharge of flood-water, as well as for navigation. The same thing might be said with regard to the Suez Canal, 100 miles long, which was solely for navigation ; it, however, was going to be widened, so that it would be considerably larger in section in future than the Witham cut. The Manchester Ship Canal, also, would have a bottom width of about 120 feet. But there could not be a doubt as to the advantage of the new cut, not merely in lowering the low-water level, but also in preventing the silting up that used to occur close to the Grand Sluice by the stoppage of silty water brought up on the flood tide, which was now said no longer to come up on account of the outfall channel being taken direct into the deep Avater of Clayhole, instead of meandering through sandy shoals. Credit was due to the Author for having at length carried out an oftentimes proposed scheme, not on account of any special difficulties in the work itself, but owing to the Witham Drainage Commissioners having been at last induced to approve, and raise money for a work which was beyond the limits of their ordinary jurisdiction, and Avhich, whilst of great Ijenefit to the drainage by improving the capacity for discharge of the outfall channel, and lowering the low-water level at the several sluices, as predicted in previous reports, might appear mainly a work of navigation improve- ment, such as was usually carried out by a Harbour Board. In spite of the greatly improved condition of the outfall, the enlargement of the Grand Sluice, and the increased section of the channel between the Grand Sluice and Tattershall Bridge, he considered that the land bordering the Witham higher up, near and above Lincoln, could not be regarded as secure from floods in very wet winters, and that further works Avould be necessary, in continuation of the present improvements, higher up the river, to protect the Witham basin adequately from serious inundations.

Mr. A. Giles, M.P., Vice President, said that looking at the vast improvement which had been effected in the navigation of the Witham by the new cut, and by the straightening and deepening of the channel, he could only express his surprise that the work was not undertaken long ago. It appeared that the low-water level of spring-tides was now within 4 feet of the dock sill at Boston ; and as the rise of a spring-tide was 23 feet, it followed that there was a depth of 19 feet over the sill at Boston, That gave a navigation fit for a vessel of 2,000 tons, whereas before the navigation was only fit fov vessels up to 300 tons. It was

Proceedings.] OUTFALL IMPROVEMENT WORKS. 97

evident when tliat improvement was going on that the Boston Mr. Giles. people awoke to the importance of having a dock in their own town, and they made it in consequence of the improvement of the Witham. Some years ago he was connected with the Bill brought into Parliament for making what was called the Boston Ocean Dock. The navigation of the river was then so bad, that it was thought impossible that it should ever Ite amended. The Ocean Dock Bill was lost, and the Boston Dock Bill was carried, and he congratulated the Boston joeople uj^on the fact. But the im- provement to which he had referred represented only half of the advantages of Mr. Williams' work, because, in consequence of the improvement in the channel, and the capability of lowering the drainage sluices all through the country, many thousands of acres of land had been saved from being flooded. Would any gentleman connected with agriculture say what was the vahie of those acres now, compared with their value when sul)ject to floods ? The improvement would pay the interest over and over again on the money expended. Although so small a portion had contributed to the cost of the work, he thought all the land that had been saved from being flooded, in consequence of the lowering of the sills of the sluices, ought also to contribute. Great credit was due for the way in which the work had been carried out, and for the smallness of the cost. He remembered the Witham when the cut was being made, and he recollected the state of the navigation before that time, when there were many feet of m\id against the sill of the Grand Sluice ; but during last summer, the river appeared as clear as could be expected, there being, 2)erhaps, 6 inches of mud on the sill, instead of several feet as before. The question was one which was creating a considerable degree of interest at the present time, not only in regard to the Witham, biit to other rivers in the country, especially the river at Preston, where, he thought, a similar course of procedure might be followed with advantage.

Mr. H. J. Marten stated that he had visited the works on one or Mr. Marten, two occasions whilst they were in progress. With the exception of the enlargement and alterations at the Grand Sluice, they consisted principally of estuary-excavation and embankment, and were exe- cuted, to a large extent, by the stalwart limbs of the old English navvy. When visiting the works, the onlj^ apprehension which he had as to their success, was lest, after the old channel of the Kiver Witham had been filled up by the accumulation of sand, owing to the cessation of the planing action of the tidal- and flood- waters consequent on the construction of the new embankment,

[tHK INST. O.E. VOL. XCV.] H

98 DISCUSSION ON THE WITHAM piinutes of

Mr. Marten, sand miglit then begin to drift round tlie end of the south embankment into the new channel. The area between the south embankment and the Eiver Welland, in which the bed of the old channel of the Eiver Witham formerly lay, formed at the present time a comjoaratively still-water reservoir or dejiositing-ground for large quantities of sand ; and since the construction of the new bank across the old channel of the Witham, this deposit had accumu- lated at so rapid a rate, that in some parts, he understood, it reached a height of 8 feet above the former level. So long as this great reservoir had still some depositing-capacity left in it, he did not apprehend the sandbank would spread beyond the ends of the new embankments ; but when this depositing-groand was full, the incoming sand, having no further room within the area indicated, must necessarily go elsewhere, and in that event it would probably creep round into and be deposited in the new channel, in larger quantities than up to the present time. The accumulations of sand in the old channel of the Eiver Witham were now commencing to drive the channel of the Welland somewhat further to the south, away from the mouth of the new channel, which point it had hitherto passed. The entrance to the new channel would therefore shortly lose the effect of the scour of the Welland in front of it, which might have an important bearing on it in the future. He therefore agreed that, as originally proposed by Sir John Hawkshaw, the ends of the embankments of the new channel would in time have to be carried a little further seaward. From the marked effect of the embankments, however, in lessening the quantity of sand- laden water entering the channel, from the much greater scour, and from the better conservation of the erosive force of the current passing along the bed of the new channel, the Author had probably done wisely in not — at least for the present — continuing the ends of the embankments further seaward. There could be no doubt of the success of the work up to the present time ; and if at some future period the banks should require to be extended, any such extension could be carried out at a comparatively trifling cost, and with the further advantage of experience. Any further expense required for that purpose could well be borne, considering the great benefit derived from the new channel, both above and below the Grand Sluice. Eeferring to this latter work, he under- stood that the outfall at the Grand Sluice had been practically lowered some 8 feet below its former level. This was of great service to the large area of land drained by that Sluice, not only from an agricultural point of view, but even still more so on sanitary grounds, as the lowering of the outfall 8 feet was practically

Proceedings.] OUTFALL IMPROVEMENT WORKS. 99

tantamount to raising the whole of the area so drained something Mr. Marten.

like that height above previous saturation-level. This meant a

greater freedom from diseases due to damp subsoil, namely, less

ague and fever, and fewer pulmonary complaints than heretofore,

with increased length of life and vigour over the whole of that

area.

Mr. W. Shelford said he had been familiar with the Witham Mr. Shelford for thirty years. It was one of the group of four interesting rivers which discharged the drainage water from a district of 6,000 square miles into the Wash, and which in fact formed the outlet for the drainage of the great Fen-land of England, covering an area of 1,306 square miles. The existence of that land was due entirely to engineers, and its rivers had exercised the minds of some of the greatest members of the Institution. The Fen-land was also considered thirty years ago a splendid school-ground for yoimg men. He could not help joining with other speakers in congratulating the Author upon having to carry out a scheme which had been desired by part of the inhabitants of the country for one hundred and twenty years, and which had been proposed more than once by men whose names were honoured by the Institution. He also congratulated him upon the successful accom- plishment of the work. But he wished to ask a few questions upon the subject. It was a remarkable fact that in the Paper on the Witham by Mr. W. H. Wheeler, M. Inst. C.E., of Boston, twenty years ago, also in one published subsequently, and in the present Paper, there was an omission of all tidal observations and tidal diagrams. He thought such diagrams would be exceedingly useful, and he hoped the Author would sTipi:)ly them. If they were taken at sufficiently frequent intervals of time and space, and were plotted, so as to show by isochronous lines the level of water during the flood and on the ebb, he believed it could be shown from the state of the diagram what was the cause of the de- terioration of the old channel, and how far that cause had now been removed. He also wished to ask why the Author had adopted a slope of 4 to 1 in the new channel. He believed that was the slope proposed by Sir John Hawkshaw; but it was worth while to ask the question, because, in the adjacent rivers, the outfalls of which were all artificial, and had been entirely made by engineers, the slopes were much steeper, varying from 1^ to 1 to 3 to 1 ; and it was found in the fen-rivers, as a matter of experience, that when the tide had a free flow, and the silt which formed the bed of the river was allowed to arrive at its own angle, 3 to 1 was the angle of repose. But there was a peculiarity about the Witham.

H 2

100 DISCUSSION ON THE WITHAM [Minutes of

Whereas most of the fen-rivers ran through silt and sand, the "Witham had a bed of boiilder clay. The boulder clay extended from riamborough Head all along the east coast, underlying the silt, and it was found in the bed of the Witham. The Scalp, shown upon the chart and described in the Paper, was a clayey shoal ; it had been cut through in order to make the new channel. It would be interesting to know whether the slope of 4 to 1 was adopted because the material was clay, in the expectation that the clay would stand without any protection in the way of fascine-work or stone ; and also to know whether it had been successful. He would further ask whether it was the fact, as stated, that the Grand Sluice at Boston had lieen designed by Mr. Langley EdAvards. Mr. Smeaton had always been credited with it, and his recollection was that it was proposed in a joint report by Mr. Smeaton, Mr. Langley Edwards, and Mr. Grundy, in 1761. The Grand Sluice was the characteristic feature of the Witham. It was put in, apparently, as the result of a contest between the landowners and the mer- chants, in which the landowners prevailed. In other words, it was a case of drainage versus navigation, and the advocates of drainage had the best of it. The sluice was 8 miles from the sea, and cut off 23 miles of the tidal wedge, or three-quarters of it ; and what was the result ? There were plenty of instances elsewhere which confirmed the experience on the Witham. In the first l^lace, the high-water line was raised above the level of high-water at sea. At Boston it was raised 8 inches. That was one effect always found in cutting off" the tidal wedge. Then slack water being necessarily produced for some distance below the sluice, any matter in suspension in the water would fall to the bottom and raise the bed. That again raised the low-water line, and so there was a deterioration of the river set up which went on in an aggTa- vated form and with accelerated speed until the bore was formed ; the action was then analogous to a breaker on the sea-shore coming up, and bringing with it sand and shingle in its advance, which in receding it was unable to take away. A river like the Witham, under those circumstances, would silt up and be blocked, were it not that the great fresh-water land floods kept it moderately open. At the present time the Grand Sluice remained, and the water, instead of being charged "vvith silt, came up comparatively clear. Instead of there being a deposit of silt in summer to the depth of 11 feet 6 inches, there was, during the last dry year, only a deposit of 6 inches. The question, however, was whether the channel would be self-maintained. It was fouled with silt before, and now

Proi'ee(liu<,'8.] OUTFALL IMPROVEMENT WORKS. 101

the only reason why the water came iii^ clear was that the new -^Ir. Shelford. channel was cut through clay, and there was no silt to bring. In the course of time, the silt brought down from the upper country would, in all probability, accumulate in the channel, and the question might then arise whether or not the channel would maintain itself in the absence of freshes by the simple action of the tide. It appeared to him to be a case of the new broom sweep- ing clean. That was, to his mind, by far the most interesting point in connection with the Paper. A channel cut through clay of that kind must, he thought, be in its best condition when first executed. He believed it had been intended that the low-water level should be lowered by the works 3 feet at the different sluices, but it had in fact been lowered 4 feet ; and any one acquainted with the fens would know how very important that result was. It was practically lifting uj) the whole country, which was below the level of the sea, bodily 4 feet.

Mr. G. J. Symons said that the Appendix to the Paper contained Mr. Symons. an interesting report of a flood in 1883, which must have taxed all the Author's skill and energy. It was a serious resj)onsibility to be in charge of so large a watershed, during a great rainfall, with the banks of the rivers liable to be broken down in all directions. But the figures which the Author had given merely showed the results of two rain-gauges, one at the Grand Sluice at Boston, and the other at the Lade Bank pumping-engine, both in the south-east corner of the watershed. However, the floods were obviously not purely local floods falling in that one corner of the gathering-ground. An organization now existed all over the country for determining the quantity of the rainfall, and he thought that when engineers had to deal with floods they should obtain all possible information on the subject. Speaking on behalf of rainfall-observers he wished to say that they were always willing to do everything they could to co-operate with their engineering friends in that matter. In regard to the storm in question, he thought that it would be interesting to give some actual details of what had occurred, and he had jire- pared a map for the purpose. The thick line showed approximately the watershed of the Witham. He had not carried the line down to the sea-coast because it was stated in the Paper that the flood in the Steeping Eiver watershed was so tremendous that some of the water got over into that of the Witham. At Boston and Lade Bank, the fall exceeded 3 inches, and at Skegness 4 inches. He had plotted all available records on the map (Fig. 2), and had indicated by dot and dash lines the areas over which the rainfall was

102

DISCUSSION ON THE WITHAM

[Minutes of

respectively above 1 inch, 2 inches, and 3 inches. This proved that the flood to be dealt with was not represented by the 3 inches in the extreme corner of the watershed, but it might probably be taken to be 1^ inch or 2 inches. He was anxious to see a little more attention paid to the relative flow of water from large areas of ground. It appeared to him that nearly all the water oif the gathering-ground, which did not evaporate, came down through

Fig. 2.

the Grand Sluice, and if that were the case, valuable informa- tion as to the actual flow from that large watershed could be obtained by establishing instruments at the Grand Sluice so as to get a continuous record of the volume of water passing through it. There would be the input and the output from that large area, and if it could be done, all woTild agree that the results would be very valuable. The annexed Table contained the particulars of the rainfall in Lincolnshire on the 29th of September, 1883.

Proceedings.]

OUTFALL IMPROVEMENT WORKS.

103

Rainfall.

Mr. Symons.

Stamford, Northfields . Bourne, Wytham-on-the-Hill Long Sutton .... Grantham, Saltersford . „ Heydour Vie

Boston, High Street

„ Grand Sluice Leake, Lade Bank engines Stubton (Newark) . Sleaford, Bloxholm .

Navenby

Skegness

Inches.

■14 •38 •49 •63 •10 •10 •10 •57 •49 •10 •38 •31

Horncastle, Miningsby . Lincoln, Branston . Spilsby, Partuey Horncastle, Bucknall . Lincoln Waterworks . . „ Skellingthorpe . Horncastle, Hemingby . Alford, The Sycamores .

„ Suttou-by-the-Sea Louth

„ Westgate .

Inches.

•43 •26 •82 •87 •19 •00 •69 •40 •25 •48 •37

Mr. A. C. HuRTZiG imagined that the new bed of the river shown Mr. Hurtzig. on the section was the bed of the river as originally improved. The Author had referred to the important fact that improvement in the tidal propagation had also improved the tidal development, and that, consequently, there would be a larger volume passing in and out of the river after the old outlet was closed. He should have thought that this action of the tide would be further to deepen the channel which was originally dredged to a certain depth. He also wished to ask why sandstone had been used in the hollow quoins of the lock ? Was it on the gTound of cheapness ? Kubbing- surfaces at important places, as in the case of hollow quoins of the lock-gates, should be as hard and durable as possible ; and he should have thought that the slight additional expense of granite would not have been prohibitive, bearing in mind the possibility of the gates getting leaky in a few years. He wished to ask, further, whether in on-shore gales there was any tendency to silt up the entrance? There might possibly be a tendency to sand-banks forming close to the entrance, where they had not previously existed. At an early date a scheme of works had been contemplated with more refer- ence to drainage than to navigation. Since then the ideas on the subject had developed, and there was now an important navigable channel for the admission of large vessels to Boston, which might in future become an important port. The development and improvement in these days of a work like the Witham outfall, from a navigation point of view, was of much greater interest to engineers, at least marine engineers, than the other questions concerned.

Mr. F. Wentworth-Sheilds asked what was the particular direc- Mr. Went- tion of the flood-tide ? Judging from the indications presented by ^^'orth-SheilJs. the outline of the coast, he presumed that the flood-tide was in the

104

DISCUSSION ON THE WITHAM

[Minutes of

Willi

direction of the immediate mouth of the channel ; in other words, that the new channel had been so designed as to catch the flood- tide, and so the deep-sea current entering the channel would consist of clear water, which would account for that absence of silt mentioned by the Author as a remarkable feature in the case.

Sir George B. Bruce, President, said a communication had been received from Mr. Duff-Bruce, asking a question to which, perhaps, the Author would reply, namely, whether he had any special appliances for landing on the banks the material dredged from the river between the Grand Sluice and Tattershall Bridge. The total quantity of excavation, he said, on that section was stated to be 602,059 cubic yards, and the average cost 9 '2d. per cubic yard. Of that, how much had been spent in landing the dredged material on the banks ?

Mr. J. Evelyn Williams, in reply ujwn the discussion, stated that no change injurious to either the drainage or the naviga- tion had taken place in the bed of the estuary at Clayhole since the completion of the works. The condition of the new channel itself had actually improved, a fact illustrative of the advantage attending the confining of a stream within a regular channel, so as to concentrate its scouring action. There was no apparent necessity, at all events at present, to prolong the banks of the new channel, and he did not think anything further in that direction would be required than the extension of the training walls, so as to keep pace with any possible recession of the low- water contour at Clayhole, and at the same time ensure the free flow of the ebb current into deep water. The scour of the Witham was not always advantageous to the Welland. During heavy floods the rivers impinged at right angles, and the Witham, being the more powerful, turned over the shifting sands into the fairway of the Welland, thereby choking and deflecting the stream in its jjassage to the sea. Frequently the two rivers flowed throiigh separate channels into the Wash, and he was informed that the drainage by the Welland had not been injuriously affected since the opening of the Witham new outfall. With a view, however, to secure the obvious advantage of the permanent and combined scour of the two rivers into Clayhole, a clause was inserted in the River Witham Outfall Improvement Act of 1880, to the effect that if, at any time within twenty years after the comjiletion of the Witham new channel, the Eiver Welland trustees should decide to form an improved channel, from the point of confluence of the two rivers at the Elbow Buo}' to the entrance of the new Witham channel in Clayhole, one-h^lf of the cost of such works should be

Proceediugs.] OUTFALL IMPROVEMENT WORKS. 105

paid by the Witliam Outfall Board. The area benefited by the Mr. Williams.

works was certainly not less than 200,000 acres, and over this area

the aggregate cost of the works, described in the Paper, represented

a capital sum of only 21 -28. per acre. The large area affected by

the works was not only materially enhanced in value, but the

owners and occupiers were relieved of the anxiety and loss which

attended frequent and disastrous floods. An ordinary spring-tide

in Clayhole rose 23 feet 4 inches, and flowed for five and a half

hours. Before the opening of the new channel, spring-tides

rushed up to Boston with a bore 3 feet in height, heavily charged

with silt, and flowed for two hours. On the completion of the

works, the bore disappeared and the tides flowed gently in and

comparatively clear from the estuary. Sj^ring-tides now flowed for

three hours and neaps from four to six hours, according to the force

and direction of the wind. With reference to the slopes of the

new channel, a slope of 4 to 1 was a successful and economical one

in so exposed a position. It saved in a great measure costly works

of protection ; the same slope had been adopted on the Norfolk

Estuary Works by Sir John Eennie and Mr. Eobert Stephenson,

many years ago. It was the first time he had heard Mr. Smeaton

mentioned as the engineer of the Grand Sluice, local history

credited Mr. Langley Edwards alone wdth the work. He could

bear testimony to the great labour and care bestowed by Mr.

Symons in compiling rainfall statistics, the perusal of which often

aflbrded him much valuable aid. Mr. Hurtzig, like himself,

preferred granite to sandstone ; but no granite whatever was used

in the construction of any of the sea sluices. Quoins of Bramley-

fall ashlar were sufficiently hard and durable for such works, and

their cost was less than one-half that of granite. The new channel

was designed to catch the flood-tide, and at the same time to ensure

the free flow of the ebb current into deep water. He might state

in reply to Mr. Dufi'-Bruce that no special apjiliances had been used

for landing the material dredged from the river between the Grand

Sluice and Tattershall Bridge. The cost of landing the material

dredged averaged od. per cubic yard.

Correspondence.

Mr. P. Caland considered the cutting of a new outfall at the Mr. Caland, mouth of the Witham to be an important work of river-improve- ment. It comprised at once the widening and deepening of the Grand Sluice above Boston, the draining of the lands formerly

106 CORRESPONDENCE ON THE WITHAM [Minutes of

Caland. exposed to flooding, the reinforcement of the action of the tide, and with it the improvement of navigation, the two last mentioned features being the purpose of every scheme of river-improve- ment in the estuarj". Not every scheme of river-improvement was so fortunate. It appeared from the Paper that while formerly the depth of water at the mouth of the Eiver Witham at high-water spring-tides was barely 15 feet, and the navigation difficult, if not imiwssible, for vessels of more than 300 tons, the result of the works had been to add 8 feet to the effective depth ; so that now vessels of 2,000 tons, and drawing 23 feet, could reach Boston. As the level of the low- water had been lowered not less than 5^ feet at Hobhole Sluice, and 4 feet at Grand Sluice, the tidal action had been greatly intensified, and the period between slack-water and high- and low-water had been correspondingly lengthened, thus greatly benefiting the navigation at flood-tide, and the drainage at the ebb. And not only the depth of the river, but the direction of the fairway, had been much improved by the construction of the new outfall between Clayhole and Boston. In future, both on the ebb and at the flood, the navigation of the channel would be available at a much earlier period of the tide, owing to the avoidance of the tortuous narrows, and channel encumbered by sandbanks, which formerly rendered the navigation difficult, and at times impossible.

Mr. M. F. FitzGerald observed that the Paper brought out strongly the importance, where low-lying lands about a flat reach of river were to be dealt with, of securing a free outfall at the lowest possible level. Where such was not provided, the increase of surface slope, which necessarily occurred during floods, might cause flooding at a moderate distance above the discharging sluices, even when this area was so great as to prevent any rise, during floods, at the sluices themselves. He would like to suggest, with reference to the great weight of floods referred to in the Appendix, that the flow of rivers had sometimes been observed notably to exceed that due to the rainfall, not only for a week or a day, but for several weeks together, especially in winter floods after long-continued wet weather. This point had been alluded to in a Paper by Mr. Robert Manning, M. Inst. C.E.,^ and observations of river flow giving such a result had been sometimes treated as anomalous and necessarily incorrect. Fig, 3 represented the average flow of the Shannon for a period of ten years, along with the flow which would have taken place if the rainfall had been

Minutes of Proceedings Inst. C.E., vol. xxv. p. 459.

Proceedings.]

OUTFALL IMPROVEMENT WORKS.

107

discharged as fast as it fell. In this river the flow considerably Mr.FitzGerald. exceeded the rainfall for the greater part of January and Fehruary. The ratio of river to rainfall-flow for the months of November, December, January, and February together was 1*2 to 1 ; for the remaining months of the year it was, however, only 0 • 39 to 1 ; and for the months of May, June, July, August and September, taken together, 0'33 to 1. Evidently the flow of a single month could not be compared with the rainfall of the same month, any more

Fig. 3.

cuaic ner

',000,000.

aoo.ooo-

600,000-

<oo,ooo-

200,000- 0.

									1
									• y
			■••-.,.	,
		^^	■■						/
				^^				-^
J A NY	fCBT	■AACH	APRIL	MAV	JBN£	juur	AUCT	SEpr	QtTf	NOV?	oecf

Fig. 4.

CUBIC FtBT P£f1 MINUTE	•	"	\
		/^ ^	\
		/	\,
		/	\
	.___ ^^
	JUNE	JULY	AUGUST

RIVER DISCHARGE RAINFALL „

than that of a day or an hour. The effect of continued wet weather in increasing the ratio of river- to rainfall-discharge, embracing the summer flood of 1861, was shown by Fig. 4. The data for both these Figs, were partly derived from Mr. Bateman's Report on the Shannon (1866), and partly from the gauge records kept at Killaloe and elsewhere. The importance of paying regard to the seasonal variation of the rainfall-discharge coefiicient was particularly great in rivers bordered by low land of a spongy nature, and it would be well if more attention were directed to this point as

108 CORRESPONDENCE ON THE WITH AM [Miuutes of

Hr.FitzGerald. the conditions were different from those regarding the quantity of water which could be impounded by a given catchment area. The effect of the new outfall channel on the Welland discharge- channel should be beneficial, by diminishing the tendency of the channel between Holbeach Marsh and the Scalp to shift ; but much depended on the effect of the altered direction of the flowing tide currents, about the Clayhole, reacting on those over the sands between Holbeach Marsh and the Scalp. There seemed in some instances, a tendency to produce an effect, which in this case would be represented by the flushing over of the Welland Channel in a northerly direction up to the Scaljo, followed by a closing of the channel along its south-eastern border, and a deej^ening of that leading towards the Gat Sand.

Mr. C. Bloys van Teeslong observed that three works of different character, though nearly connected, had been almost simultane- ously executed on the River Witham. By the excavation of a new outlet a serious impediment to navigation had been removed, and an unobstructed access of tidal waters obtained; by the deepening of the tidal compartment a complete filling up and emptying of the river had been assured ; whilst by the enlarge- ment of the upper part and the Grand Sluice, the area drained by the Witham had been greatly improved. He attributed these favourable results to the measures being well combined, and simultaneously executed. Measures of amelioration of the upper part of the district, taken in former years, consisted chiefly in shortening the length of the river, in the excavation of new drains, and in deepening the bottom ; but these works produced only a temporary alleviation. The shortening of a river, and the subse- quent acceleration of the cuiTent, generally gave rise to a more rapid erosion of the banks and bottom, and disturbed its eqiiilibrium ; and if the river was prevented from regaining its original length, the transport of materials, the result of erosion, tended to destroy the effects of the improvement. By deepening a river the regime of the stream was often injuriously affected, and the old conditions gradually re-ajjpeared. The formation of new canals for drainage, separated from a river, frequently proved injurious by abstracting water which, by its scouring jjower, maintained the river. No objection, however, could, he thought, be oflered to the widening of a river, as had been done in the case of the Witham ; and especially was this the case where the removal of material chiefly took place above high flood-level. The erection of banks in early times reduced the sectional area of river beds during floods, and it was by neutralizing the prejudicial influence of this diminution of

Proceed in j?s.] OUTFALL IMPROVEMENT WORKS. 109

area, that excavation beyond tlie navigable channel conld be useful. With respect to the Grand Sluice, he considered it was difficult to decide, at present, whether its erection had been a mistake, as its former surroundings were not now known, and might have altered considerably since early days. He considered the relatively small width of the Withani an important factor in maintaining the tidal portion of its bed free from sand, which would otherwise be brought in during flood-tide, especially in stormy weather ; the influence of the fresh- water flow in the last of the ebb would be especially beneficial. He approved of the cut through the Scalp. Where two rivers, such as the Witham and the Welland, joined at right-angles, and their common outlet was encumbered with shifting sandbanks, a radical remedy seemed perfectly justifiable. Moreover, by the formation of the new mouth, a prolongation of the training-works could be easily ex- ecuted iu the future if necessary. The cutting of the Hoek van Holland, at the mouth of the Eiver Maas, the waterway to Rotterdam, was the most interesting instance of a similar work on the Continent. In the case of the Maas, it had been originally contemplated to resort mainly to the scouring action of tidal currents ; but though there had been considerable erosion, it was found necessary, owing to the shifting and the shallowing of the navigable channel at the entrance, to undertake extensive excava- tion. The result was considered successful ; although, probabty, owing to the relatively great width and straightness of the channel, sufiicient depth could only be maintained by yearly dredging. In another part of the Rhine delta, two small rivers, the Donge and the Oude Maasje, as in the case of the Witham and Welland, formerly joined nearly at right-angles to one another at the point of discharge into the Estuary of the Amer. The first-named river had in the course of time gradually altered its course in the lower part, and now occupied a position much resembling that of the Witham. At a short distance beyond the entrance, the navigable channel shallowed, and moderate dredging was required annually to maintain the depth of water equal to what it was higher up. With regard to the question of keeping the outlets of the Witham and of the Welland separate, in Holland, as a general rule, rivers were kept apart as much as possible, if their drainage areas were of a different nature, as for instance the Rhine and the Maas ; in other cases, the uniting of branches was esteemed favourable to their general condition. On the point of cost he thought it would be advisable to bring the outlet of the Welland to Clayhole rather than to Lynn Deeps. He found, from ancient charts, that

110 CORRESPONDENCE ON THE WITHAM [MinuteB of

Ir. Bloys van formerly both the Witham and the Welland discharged alternately res ong. j^^^ Boston Deeps and Lynn Deeps, and that since the commence-

ment of the present century neither of the Deeps had materially altered in capacity, though the latter had become narrower and deeper. The width of the Witham, in comparison with other marine waterways to an important town, was relatively small. For a commercial town of the rank of Boston, a width of channel of about 180 feet would have been adopted in Holland; its nearly straight channel somewhat resembled a canal, but the navigation was rendered more difficult by tidal currents. By the shortening of the river, consequent on the new outlet, the reservoir of tidal water had been diminished, impairing the action of the tidal scour. He was of opinion that it depended, in some degree, on the more or less rapid obliteration of the Wash, whether the results obtained by the works on the Witham would prove durable. Hitherto that had not occurred ; but he thought it possible that this would occur, as soon as the reclamation of land had reached some limit, though probably not for many years. Boston Deeps was said to be maintained by tidal action. If true, this might be, to some extent, explained by the manner in which the tidal flow entered the W^ash. The tidal water, following the deepest channel, did not enter Boston Deeps directly, but took the direction of L;yTin Deeps, and passed across the Long Sands. As to the ebb currents, it might be deduced from the position of the banks and channels that they equally diverged at the entrance of Boston Deeps. Under these circum- stances, it was evident that the raising of the banks would greatly improve the general condition as regarded navigation. To provide an outlet for the Welland into deep water, he should recommend the construction of a bank from the outer end of the W^itham bank in a south-easterly direction towards Holbeach Marsh, leaving an opening at Clayhole for the discharge of the W^elland waters. The space, thus partly enclosed, would be protected against the action of waves and strong currents, but filled up and emptied at every tide, and rapid silting up, es;pecially at the angles, might be anticipated. The lands thus formed might, as soon as suitable for reclamation, be enclosed ; while the banks, constructed for this purpose, would at the same time, serve to train the river. The secondary branch, leading the Welland to Lynn Deeps, would in this way be dammed off; that leading to Clayhole would remain unaltered, and would probably not need deepening, as it originally carried seawards the combined waters of the Witham and the Welland. After some time,/ a situation would l)e obtained similar

Proceedings.] " OUTFALL IMPKOVEMENT WORKS. Ill

to that proposed by the Lincolnshire Estuary Company. At first, Mr. Bloys van

the Welland would be liable to shift its course in the large enclosed ^^^ °"°'

space ; but not so much as at present, secondary branches being

shut o&, and the position of the mouth being fixed. The outlet

of the Witham would, of course, be benefited by the scouring

action of the water passing the entrance to the partly enclosed

space.

Mr. J. Evelyn Williajms, in concluding the correspondence, Mr. Williams. desired to add the following communication received from Mr. Goodwyn Archer, the Clerk and Solicitor to the Ouse Outfall Board : — " The improvement seems to be for the Boston Watershed equivalent to what was done for the Lynn Watershed by the Eau Brink Cut. Taking into account the difference of the areas of land affected, the Eau Brink Scheme cost the contributors more than twice the amount which your contributors have to bear." This statement Mr. Williams felt sure could not be other than gratify- ing to the land-owners in the Witham Valley.

112

ELECTIOXS, ETC.

[IMinutcs of

4 December, 1888.

Sir GEOEGE B. BEUCE, President,

in the Chair.

The following Associate-Members have been transferred to the class of

Members.

Edward Appletox.

Haevet Bagxall, M.A., B.E.

Percival Fowler.

Philip Affleck Fraser.

Arthtr Egbert William Fulton.

James Edward Fflton.

Eai Bah-vdee Gaxga-Ea3i.

Arthur Staples Gerrard. "William Harper. William Hughes, B.A., M.E. Charles Erxest Norman. Willia^i Stuart Eendel. Henry Goulton Sketchlet. Edmund Casttell Bowyer Smijth.

The following Candidates have been admitted as

Students.

George Francis Adams. Charles Frederick Bamford. Sa^iuel Harry Hill Barratt. Thomas John Bayne. Cyril Holm Biss. Frank Walker Bottle. Arthur Walter Bradley. James Wood Bragg, B.A. Harry Bucknall. Willia3i Edward Burgess. James Hubert Cochrane, A.K.C. Wilfred Arthur Cope. Frederick Nutter Cox. Henry Haryey Dare, B.E. John Thomas Llewellyn DA^aES. Albert Edward Dawxey. John Smith Dawson. Fraxk Bridgewater Debexham. Archie Eussell Emdix. NoRMAX Fitz, B.E. John Fraxcis Foster. Colin Proud Fowler. Alexander Eraser.

Frax'k Horace Frere, A.K.C.

Herbert Ferdinand Friederichs.

Charles Humphrey' Gilbert Wh.Sc.

Henry Hallett.

WiLLiA^i Leigh HA^rrLTON.

Francis Joseph Harvey.

Harold Hawkins.

Charles William Hobley.

John Holliday.

Percy Winstanley Hull.

Alfred Jajies, B.A.

Charles Frewen Jexkin, B.A.

Herbert Harry Jones.

Harry Birch Killox.

Joshua Lambert.

William Hexry FitzEoy Landon.

Sydx"ey Aspland Lang.

Herbert Willia^i Lax'gley.

Alec George Yaugh.^n Lee.

Edward Sargint Lixtdsey.

Frederick Lambert Lordex.

Ernest Loyegroye.

Harold ]\Iacaxdrew

Proceedings.]

ELECTIONS, ETC.

113

Students, continued.

Lessel Stephen McKenzie.

James Maie, Wh.Sc.

Edward Charles Egbert Marks.

Ernest Meteor Martin.

Harry Powell Miles.

Charles Julius Alfred Mittel-

hausen. Charles Joseph Anthony Patrick

Moore. Alexander Douglas Moriarty. Charles Shelley Oakes. James Geofprie Musgrave O'Hara. Frederick Charles Osborne. Francis Davidson Outeam. Percy John Paterson. DiGBY Prescott Pedder. Charles Leon Emile Pitot. Walter Playfair. John Portsmouth. Alan Railton. Arthur William Ranken, A.K.C.

Robert Newby Hartley Reid.

Joseph Peter Robinson.

George William Roome.

Alfred Schwartz.

Henry Chawner Hine Shenton.

William Sillem.

Holman Fred Stephens.

Edward Duncan Stoney.

Hugh Sidney Streatfleld.

William Archer Porter Tait, B.Sc.

Francis Manley Shawe Taylor.

Charles Wilson Thompson.

John Thompson.

Nicholas King Turnbull, Wh.Sc.

James Vicars, B.E.

James Whitaker, Wh.Sc.

Henry Thomas White.

Francis Houlton Wrench.

Alan Wyatt-Smith.

Maurice Edward Yonge.

The following Candidates were balloted for and duly elected as

Members.

Alexander Anderson. John McKenzie Bell. Oliver Budge. William Lind Buyers. David Cowan. William Davidson. Theodore Newel Ely. Arthur Sumner Gibson. John Richardson Hewitt.

Andrew Oliver Lyons. Loudoun Francis MacLean. Henry William Martin. James Price, Jun., B.E. Alan Wood Rendell. Richard Oswald Robson. Charles Arthur Rowlandson. Frederick Cook Stephens. Berkeley Deane Wise.

Associate Members.

William Wallace Andrews. Llewelyn Birchall Atkinson, A.K.C,

Stud. Inst. C.E. Walter Attard. Philip Port Ayres. John Banks. Charles Arthur Albert Barnes

Stud. Inst. C.E. Onward Bates. John Beveridge. Edward Philip Binet.

[the INST. C.E. VOL. XCV.]

William Blackshaw.

John Vaughan Brenchley, Stud.

Inst. C.E. Samuel Edwin Burgess, Stud. Inst.

C.E. Ernest Sydney Burman. Henry Robert John Burstall,

Wh.Sc, Stud. Inst. C.E. Clerke Burton. William Robert Butler, B.E. Fked Smith Button.

I

114

ELECTIONS, ETC.

[Minutes of

Associate Members, continued.

George James Chapman.

Edward George Clark, Stud. Inst.

C.E. William John Clarke. Henry Lowthian Cleaver, Stud.

lost. C.E. CorRTENAY Thornton Clieton, Stud.

lust. C.E. William Horace Coomber, Stud. Inst.

C.E. Robert Cra^rtord. Frederick Sol'thwell Cripps. Charles Eichard Ernest Crook,

Stud. Inst. C.E. John D'Aeth. Morgan Williams Davies. WiLLiAsi ArorsTUS Davies. Thomas Ei'Sholm Dickinson.

WiLLIAJI DiESELHOEST.

Pai'l Doiier, A.K.C, Stud. Inst. C.E.

Edward Ellis.

George Wadhaisi Floyer, Stud. Inst.

C.E. Egbert Maynard Gloyne, Stud. Inst.

C.E. George Lacy Good. James Walter Grimshaw. George Higgins. George Hobbs, Stud. Inst. C.E. Alfred James Hodgkinson. Edward Holmes.

Arthur John Ikin, Stud. Inst. C.E. Nicholas Pavl Jasper, Stud. Inst. C.E. William John Jenkins. Henry Sydney Jones, Stud. Inst. C.E. James Keith. Eichard Johnson Lawton. George Hilder Libbis, Stud. Inst. C.E. Henry Littlejohn, Stud. Inst. C.E. Theophilus Septiml'S McCallum. Donald Grant Macdonald. James Waddell Boyd Maclaren,

Stud. Inst. C.E.

William Snell Tandy Magee.

Warine Ben Hay Martindale, Stud. Inst. C.E.

Stephen Martin-Leake.

Edward Alworth Mitchell Mere- wether, B.E.

George Percival Milnes, Stud. Inst. C.E.

Charles Edward Cage Montresor.

Percy Nevill, Stud. Inst. C.E.

Thomas Nisbet.

William Anthony Morgan Par- tridge.

AUGUSTPS TiCHBORNE PeNTLAND.

Eichard Douglas Perceval. Williaji Marshall Philip. Arthur Powell. Harry Ernest Prescott. Eeginald Seymour Prinsep, Stud.

Inst. C.E. Eichard Watkins Eichards. WiLLiAJi Henry Eobins. JA3IES Eochfort. Ealph Baron Eogers, M.A. Frederick Eose, Jun., Stud. Inst. C.E. Norman William Eoy, Stud. Inst. C.E. John Sampson. James Joseph Shaw. EoRERT Skelton, A.K.C, Stud. Inst.

C.E. Edmund Paley Stephenson. Percy Kendall Stothert, Stud. Inst.

C.E. Clement Moriscrip Sykes.

EOBERT COCKBURN SySON.

Arthur Drew Thomas, Stud. Inst. C.E.

Alexander Walker.

Charles Leslie Walker, Stud. Inst.

C.E. Thomas Henry Ward. Williaji Warner. William Wearing. Walter James Weightman.

Associates.

Edward Miller Gard Eddy. William Henry Jaques, Lieut. U.S.N.

William Alfred Perry. Benjamin Willcox.

Proceedings.] ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. 115 QPaper No. 2286.)

" On the Influence of Chemical Composition on the Strength

of Bessemer-Steel Tires. "

By John Oliver Arnold, F.C.S.

The importance of the subject dealt with in this Paper can hardly be over-estimated. It is a question affecting the safety of thou- sands of lives. In such a case it is impossible to have too great a margin between working and breaking-strain.

There is a growing tendency amongst railway engineers to specify for tires a steel possessing a high resistance to tension. This no doubt is conducive to economy in wear ; but it remains a debatable question whether such material is not more liable to the risk of sudden fracture than a more ductile if less durable steel. It is true that the framers of such specifications apparently provide against such risks, by insisting also on steel of a high resistive power to rupture under the impact of a falling weight, and of a fairly high capacity to elongate and to contract in area when broken by tension. But, in the Author's opinion, the chemical composition necessary to obtain the anomaly of high tensile-strain, together with high elongation, is such as to render steel liable to those sudden or gradual molecular changes (the nature of which is at present imperfectly understood), which doubtless produce from time to time disastrous results. The Author purposes to lay before the Institution the data upon which he bases this conclusion. The question is rendered intricate : —

(a) By the fact that steel is a complex body, and that the influences of its elements upon each other, with reference to physical effect, have hitherto defied all attempts to reduce them to formulas.

(6) The difficulty of insuring in pieces of steel, identical in com- position, a uniform and constant molecular structure. In connec- tion with the latter remark it may be stated, as a general principle, that the more abnormal the proportions of the foreign and harden- ing elements in a tire-steel, the greater the liability of the material to injurious molecular change. In addition to the iron, the chemical composition of normal tire-steel is approximately per cent. : —

Carbon.	Silicon.	Manganese.	Sulphur.	Phosphorus.
0-28	0-07	1-25	0-08	008

I 2

116 AKNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. [Minutes of

A test-piece, 2 inches parallel and 0-564: inch in diameter, 0-25 inch area, planed out of a tire possessing the above analysis, would give on an average the following mechanical results, the figures being calculated on the original dimensions : —

Te^^StZn.! Elongation. \ «f-f Fracture.
Tons per Square inch. 37	Per cent. 26	Per cent. i Grey granular, ■with silky 47 1 edges ; shape convex I and concave.

Such a tire, with an inside diameter of 2 feet 8 inches and a sectional area of 11 inches, would behave under the falling- weight- test in a manner indicated by the subjoined figures, the weight of the tup being 22 cwt.

Fall in feet . . .	2	4	6 1 8	10 i 12
Deflection in inches .	1 *	3 13 q 1	4H	6J	Unbroken.

The falling- weight-test might have been continued much further ; in fact until the tire was doubled up ; and often rupture can only be brought about after rej^eated blows at 30 feet. Such a tire is perfectly adajited to fulfil all requirements except that, like every- thing else, it wears out in time. It possesses one highly important property ; it is little liable to molecular change under sudden heavy and rej^eated shocks. This is proved by the fact that a test- piece, planed out of an untested tire, gives miich the same result on the testing-machine as a piece planed out of a tire which has been subjected to the falling-weight-test.

It may be accepted as an axiom that in Bessemer tire-steel the carbon should lie between the limits of 0-25 and 0*32 per cent. Taking the foregoing analysis as a fixed starting-j^oint, the efi'ect of an increase in the proportions of silicon, sulphur and phosphorus respectively will be touched upon. High silicon in Bessemer steel is due either to underblowing, or to a very impure manganiferous addition, and 0-20 per cent, is a highly dangerous amount in a tire. High phosphorus is due to the use of originally impure pig-iron, and 0-15 per cent, is dangerous. High sulphur may be due either to impure pig-irons or to inferior cupola-coke. The influence of this element is not so marked as that of silicon

Proceedings.] AENOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. 117

and phosphorus, but it should nevertheless be kept as low as possible. The effect of an abnormal proportion either of silicon or of phosphorus in a tire is to produce brittleness. All three elements must be regarded as impurities, under all circumstances to be kept low, and to take no part in any adjustment of the hardening elements of steel, to obtain any desired increase in the tensile- strength of tires.

Suppose that it is desired to increase the tensile-strain of a normal tire 5 tons, that is to say, to 42 tons per square inch, and at the same time to obtain without rupture the normal deflection, 2 inches per foot of inside diameter, under the falling weight. Such a result (always remembering that Bessemer tire-steel, from the nature of the process by which it is made, must contain, to insure soundness, about 1 • 25 per cent, of manganese) cannot be safely obtained by an increase of carbon, because the additional percentage of this element, necessary to give the required strain, would render the tire liable to break under the falling-weight-test, before the necessary deflection had been obtained. In other words, a tire containing 1 • 25 per cent, of manganese and 0 • 40 per cent, of carbon would give the required strain on the machine, but would fail under the drop-test. Therefore the only means of arriving at the desired result, without introducing a new element, is to increase the manganese.^

The effect of such increase is denoted by the following data : — Steel was made, having the following composition per cent. : —

Carbon.	Silicon.	Manganese.	Sulphur.	Phosphorus.
0-25	0-03	1-75	0-12	Oil

An ingot of this steel was hammered into a 4-inch bloom ; the bloom was then rolled down into a bar, 1^ inch square; and from this four test-pieces, 2 inches parallel, 0-5G4 inch in diameter, and 0 • 25 inch area, were turned.

* In the case of open-hearth steel, the comparatively small proportion of Fe O, present in the metal at the termination of the oxidation, enables the steel-maker to leave a much lower percentage of manganese in the finished steel, and consequently allows greater latitude in the range of carbon and hardening elements other than manganese.

118 ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. [MinuteB of

Hesults.

Tons. Mean maximum strain per square inch . . . .42-1 Highest „ „ .... 43-3

Lowest „ „ .... 41*6

Per cent.

Mean elongation 18-0

Highest „ 18-8

Lowest „ 17 •!

Mean reduction of area 26 ■ 3

Highest „ „ 31-0

Lowest „ „ 21 "5

The fractures were flat, and devoid of that uniform fine grey granular appearance and convex and concave shape characteristic of highly ductile steel. They exhibited about 35 per cent, of grey granules and 65 per cent, of fine crystals.

Although the actual falling-weight-test of the above steel cannot be given, the mean result of the tests on many tires practically identical in composition, 2 feet 8 inches inside diameter and having a sectional area of 1 1 inches, is as follows, the weight of the tup being 22 cwt. : —

Fall in feet	2	4	6	8	10	12	14
Deflection in inches	5	i	If	9 3 Q9	51	6|

So far the conclusions to be drawn are that an increase of manganese produces an increase of hardness, whilst at the same time the tire is sufficiently tough under the impact of the falling- weight. The loss of ductility is indicated by the rate of de- flection.

Assuming the lower percentage of carbon and silicon in the high manganese-steel, when compared with normal steel, to be compensated for by the somewhat higher proportion of phosphorus and sulphur, the two steels are practically identical in composi- tion, except in the percentage of manganese ; and the influence of this element on tensile-strain, elongation, reduction of area, and rate of deflection under the falling weight may be approximately summarized as follows: That the addition of 0-50 per cent, of manganese to the normal tire has : —

(1) Raised the strain supported from 37 to 42 tons per square inch.

(2) Reduced the elongation from 26 to 18 per cent.

Proceedings.] AENOLD ON THE 8TEENGTH OF BESSEMEB-STEEL TIRES. 119

(3) Decreased the rediiction of area from 48 to 26 per cent.

(4) Kequired an additional 15^ foot-tons to produce equal deflec- tion.

The life of a tire possessing a tensile-strain of 42 tons is no doubt longer than that of one breaking under tension at 37 tons per square inch ; but experience shows that the former steel is much more sensitive to the influences, rate of cooling, shocks and vibration, which cause that mysterious rearrangement of particles, converting steel from a tough substance, yielding under tension a grey granular fracture, to a brittle material presenting a fracture consisting of bright crystals.

But 42 tons by no means form the limit of the demands contained in engineers' specifications. A tensile-strain of 48 tons, an elonga- tion of 15 per cent., and a deflection of 2 inches to the foot have been specified.

From what has been advanced it will be obvious that, to obtain such a result as a breaking-strain of nearly 50 tons by means of manganese, about 2 • 50 per cent, of that element would be neces- sary. No steel-maker would risk the inevitable brittleness of such a metallurgical deformity. Therefore the aid of another element has to be called in, and that element is chromium. This substance has been praised as a most valuable addition to steel ; it has also been condemned as an altogether noxious ingredient. The facts of the case, as far as its application to mild steel is concerned, will now be stated. The conclusions drawn, from data to follow, as to its efi'ects are : —

(a) That chromium added in small quantities raises the tensile- strain of steel in a remarkable degree, without seriously diminishing the ductility,

(b) That when added in too high a proportion it induces brittleness.

The following is the percentage analysis of tire-steel, required to stand a strain of at least 48 tons per square inch : —

Carbon.	Chromium.	Manganese.	Silicon.	Sulphur.	Phosphorus.
0-28	0-42	1-54 0-08	010	0-09

A test-piece was prepared by hammering a 14-inch square ingot down to Ij inch square. This bar was then turned to 2 inches parallel, 0*564 inch in diameter and 0*25 inch area, when the fol- lowing results were obtained : —

120 ARNOLD ON THE STKENGTH OF BESSEMER-STEEL TIRES. [Miimtes of

Maximum Tensile-Strain.	Elongation.	Reduction of Area.	Fracture.
Tons per Square Inch. 49-8	Per cent. 150	Per cent. 26-0	(Flat and finely \ crystalline.

A tire, 2 feet 8 inclies in inside diameter and of 11 inches sec- tional area, behaved as follows under a falling weight of 22 cwt. : —

Fall in feet . . . ' 2	4 6 8 10	12	14	16	18	20	25
Deflection in inches . |	1 111 lis Q I	*i	51	7i	9i	11^5 Broke 1

A test-piece, 2 inches parallel, 0*564 inch in diameter, and 0-25 inch area, was planed out of the broken tire and gave the following figures : —

'l\^r" Elongation.	Reduction of Area.	Fracture.
S^uZl'lZ,. Percent. 47-7 3-0	Per cent. 6-4	Large crystals.

The molecular change set up by the shock and vibration of the falling weight is thus indicated most clearly.

Although it is possible to get a strain of 50 tons per square inch, together with great strength, under the drop-test, such tires are very uncertain. Their molecules may or may not assume a struc- ture capable of resisting sudden shocks. Their action is delicately poised, and a slight external influence, such as too rapid cooling after leaving the rolls, turns the scale.

The influence of slow cooling, on the arrangement of the molecules into a form capable of resistance to rupture under shocks, is well indicated by the subjoined tabulations of the deflections obtained on a tire of 2 feet 8 inches inside diameter and 11 inches sectional area. This tire was made from the same steel which, under ordinary treatment, showed the deflections last tabulated, but which, when buried in hot ashes and allowed to cool from a blood-red heat during forty-eight hours, gave, under the falling weight of 22 cwt., the following superior result : — ■

Fall in feet ....	2	4	6	8	10	12	14	16	18	20
Deflection in inches	1 3	9 16	1|	915	ih	Gl	81	10 12,'	15,',

Proceedings.] ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. 121

The mean of four closely-agreeing tests from pieces planed from various parts of the punished tire, the test-bars being 2 inches parallel, 0-5G4: inch in diameter, and 0*25 inch area, is embodied below : —

Maximum Strain.	Elongation.	Reduction of Area.	Fracture.
Tons per Square Inch. 46-2	Per cent. 20-7	Per cent. 45-7	Granular.

This remarkable result indicates that not only does annealing raise the ductility of steel, but also causes a molecular arrangement, capable of great resistance to alteration under vibration and shocks, achieving this result on a material previously very sensitive to physical influences and to local molecular disturbance referred to later on.

The liability of high-strain tires to assume a brittle crystalline structure, instead of that arrangement of interlaced (?) molecules so conducive to ductility, is indicated by the following data : —

A steel gave on analysis these results per cent. : —

Carbon.	Cliromium.	Manganese.	Silicon.	Sulphur.	Phosphorus.
0-32	0-30	1-46	Oil	0-05	0-07

A piece of this material, cut from an ingot and drawn down to II inch square, then turned to 2 inches parallel, O'oG-l inch in diameter, and 0*25 inch area, behaved on the machine thus : —

Maximum Strain.	Elongation.	Reduction of Area.	Fracture.
Tons per Square inch. 50	Per cent. 16	Per cent. 29	Coarse gi-anular.

A tire, of 2 feet 8 inches inside diameter and 1 1 inches sectional area, tested under a falling weight of 22 cwt., acted as follows : —

Fall in feet	10	12 14
Deflection in inches	n	31 Broke

122 ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. [Minutes of

A test-piece, 2 inches parallel, 0-564 inch in diameter, and 0*25 inch area, was planed from the broken tire with this result : —

Maximum Strain.	Elongation.	Reduction of Area.	Fracture.
Tons per Square Inch. 48-4	Per cent. 60	Per cent. 6 ■ 0 Large crystals.

The loss of ductility resulting from the presence of too much chromium is seen by comparing the subjoined data with the analysis and tensile-test on the hammered bar of the material last dealt with, the mechanical conditions being identical. A steel possessed the following comj^osition per cent. : —

Carbon.	Chromium.	Manganese.	Silicon.	Sulphur.	Phosphorus.
0-28	0-64	1-41	0-11	1 0-07	0-07

A piece was cut from an ingot and drawn down to a bar 1^ inch square, the size of the test-piece, being 2 inches parallel, 0 • 564 inch in diameter, 0*25 inch area, with this result: —

Maximum Strain.	Elongation.	Reduction of Area.	Fracture.
Tons per Square Inch. 50-4	Per cent. 10-0	Per cent. 13-8	Crystalline.

It is well known that the amount of work put upon steel has a marked relation to the molecular structure, and consequently to the ductility of the material. This point is illustrated by the following experiments : — A 14-inch ingot was reduced by hammer- ing to a shaft 8 inches in diameter. Two test-pieces 8 inches parallel, 0*8 inch in diameter, 0*5 inch area, were planed out of this shaft, and tested on the machine with the following results : —

—	Maximum Strain.	Elongation.	Reduction of Area.	Fracture.
No. 1 . . . No. 2 . . .	Tons per Square Inch. 42-0 41-4	Per cent. Per cent. 91 18-9 11-3 89-9	>Flat and crystalline.
Mean .	41-7 ! 10-2 29-4

More work was now put on the shaft ; it was hammered into a 3-inch bloom, and the bloom was rolled into a bar 1^ inch square, and

Proceedings.] ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. 123

from this bar two test-pieces 8 inches parallel, 0 • 8 inch in diameter, 0 • 5 inch area, were turned, the results arrived at being as follow : —

	Maximum Strain.	Elongation.	Keduction of Area.	iFracture.
No. 1 . . . No. 2 . . .	Tons per Square Inch. 42-3 42-4	Per cent. 15-2 13-7	Per cent. 39-9 32-3	(Grey granular, with silky < edges. Convex and con- 1 cave in shape.
Mean .	42-35	14-45	36-1

An interesting, but rather expensive, addition to the above tests would have been the results yielded by a pair of pieces planed out of the unworked ingot. The elongation would probably have been practically nil. It is curious that specifications are sometimes drawn up requiring a certain percentage of elongation, but failing to specify the length of the test-piece to which such percentage refers. The eifect of varying lengths and diameters of test-pieces, on the results obtained, is exemplified by the following series of trials : — An ingot 14 inches square was hammered down to a 4-inch bloom. The bloom was then rolled into a bar 1^^ inch square. From this bar twenty test-pieces were turned. Their sizes and the figures jdelded on the tensile-testing machine are tabulated in the subjoined columns : —

Dimensions of Test-piece.
			Maximum Strain.	Elongation.	Reduction of
			Area.
Parallel.	Diameter.	Area.
Inches.	Inch.	Square Inch.	Tons per Square Inch.	Per cent.	Per cent.
2	0	564	0	25	44	7	21	9	36	0
2	0	564	0	25	42	2	22	6	39	9
Fig. A 2	0	800	0	50	44	1	22	0	32	3
2	0	985	0	75	42	6	24	2	36	3
4	0	564	0	25	43	3	19	5	39	9
4	0	564	0	25	44	2	17	6	31	0
Fig. B 4	0	800	0	50	42	3	2]	0	38	6
4	0	985	0	75	41	3	22	5	27	8
6	0	564	0	25	43	5	16	1	39	9
6	0	564	0	25	46	3	14	1	31	0
Fig. C 6	0	800	0	50	43	9	17	7	32	3
6	0	985	0	75	41	4	17	4	39	1
8	0	564	0	25	42	0	15	8	44	0
8	0	564	0	25	41	4	16	4	45	6
Fig. D 8	0	800	0	50	44	4	14	8	32	3
8	0	985	0	75	41	3	18	0	27	8
10	0	564	0	25	44	4	13	1	36	0
10	0	564	0	25	46	0	12	8	26	3
Fig. E 10	0	800	0	50	43	8	13	1	32	3
10	0-985	0-75	42-1	13-8	30-9

124 ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. [Minutes of The above results may be thus summarized in tabular form : —

Length of Test- piecfcs.

Inches. 2

10

Strain.

Elongation.

Reduction of Area.

Per cent. 22-2

20-2

16-4

16-2

13-2

Per cent. •1

Square Inch. 0-25

0-50

0-75

Strain.

Tons. 43-8

43-7

41-7

Elongation.

Per cent. 170

17-7

19-2

The analysis of the steel bar from which the test-pieces were turned indicated the following percentage composition : —
Carbon.	Chromium.	Manganese.	Silicon.	Sulphur. Phosphonis.
0-27	0-26	1-69	0-04	012	Oil

It will be noticed that the tests vary considerably inter se, although made on pieces identical chemically,^ and in the me- chanical treatment they had undergone. The obvious inference is that so sensitive is high-strain steel to physical change that it is impossible, without annealing, to get even in one bar homogeneous molecular structure.

Taking into consideration mean results, it will be seen : —

(a) That the length of a test-piece does not affect the strain or reduction of area.

(b) That, if the elongation yielded by a 2-inch test-piece = 100 roughly; then that by a 6-inch = 80; and by a 10-inch = 60.

* Certain metallurgists and engineers have attempted to account for the varying results, yielded by several mechanical tests from the same bar of steel, by attributing their variations to the heterogeneous chemical composition of the material. But the Author's tests on this point, extending over nine years, have led him to the conclusion that the chemical composition of ingots of ordinary size, which are set in a few minutes after casting, is practically uniform, the slight differences found being such as to warrant their reference chiefly to errors of analysis. At any rate, they are far too small to account, say, for the differences existing between the tests of the two 6-inch pieces preceding Fig. C in the Table, namely, 3 tons in the strain, 2 per cent, in the elongation, and 9 per cent, in the reduction of area. In ingots of large size, which remain liquid for some hours, there is no doubt a decisive variation in the composition of drillings taken from different parts of the block.

Proceedings.] ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. 125

(c) That a test-piece of large area gives a lower strain and higher elongation than one of small area.

Figs. A, B, C, D, E, have reference to five test-pieces, each

marked into eight equal parts, in order to ascertain the exact distribution of the elongation. Each pair of Figs, represents the test-piece before and after testing. The value of each division is expressed in decimals of an inch, and on each tested piece is also

126 ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES, [Minutes of

Proceedings.] ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. 127

expressed the percentage-distribution when the total elongation is represented by 100. The sketches are one-half full size, and the distance of the point of rupture from the nearest division-point is to scale. It is worthy of remark that the elongation obtained on the 2-inch test-piece, 22 per cent., is by no means coincident with that obtained in the 2-inch adjacent to the lines of fracture in piece B, 32-5 per cent.

In dealing with the question of tires, the Author has endeavoured to free his mind from bias, but his standpoint is naturally that of the steel-maker. The other side of the question, namely, the view of the engineer, has to be considered. The latter holds that a certain resistance to ruj^ture under a falling weight, together with certain results obtained on the tensile-testing-machine, form a criterion of the capacity of tires to safely meet strains they may be subjected to when put to work. This view is, no doubt, true in the great majority of cases ; but the fact remains that the theory has never yet been exhaustively proved. Engineers are naturally reticent to the outer world with reference to the breaking of tires or axles of the rolling-stock under their charge, though their silence does not extend to the makers of the faulty articles. Nevertheless, both engineer and maker often remain in the dark as to the cause of failure. But when the fracture of an axle or a tire leads to some fatal disaster, the engineer is to some extent called before the bar of public opinion to account for the accident. In such cases his explanation that the breakage was due to an " original but invisible flaw " has become |)roverbial. However, the explanation is given in perfectly good faith. An analysis and a tensile-test of the broken article have been made, and the results obtained have thrown no light on the matter. But the important question with regard to the mechanical test arises : Was the test-piece planed from the immediate vicinity of the fracture? The Author has obtained data which prove, beyond doubt, that injurious molecular change may be very local. For instance, a tire which has been much punished under the falling weight will sometimes give widely diverse results on the machine, such variations depending upon the position from which the test-pieces were taken. Let Fig. F represent a tire deflected under the droj?. Let a be the point of impact, b the point opposite. Then a will obviously be the point at which the most marked molecular change may be expected. Next, in order of liability, will rank the point b. Whilst the points c c, being the centres from which a bending motion of the i)articles has taken place, may also be assumed to fall under the influence of molecular change. The points x xxx are least liable

128 ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. [Minutes of

to physical rearrangement. This theory will be found with certain types of steel to be borne out in practice. Test-pieces planed from X will yield good elongations and graniilar fractures, whilst pieces from a will give poor elongations and crystalline fractures.

Returning to the question as to how far the tests, to which tires are subjected before leaving the makers' works, indicate their fitness for the strains and vibrations they will subsequently meet, it may be suggested that what is required to definitely settle the point is a patient and extensive investigation upon the following lines : —

At the works, let tires be selected from groups made from the same blows, each group being marked with a distinctive stamp. Let representative tires from each group be subjected to exhaustive chemical and mechanical tests. Let the exact mechanical treatment from ingot to finished wheel be faithfully recorded. When the life of one of these tires is finished, either from breakage or in the natural

course, let the chemical and the mechanical tests be repeated. Such a series of tests would indicate whether the influences en- countered by a tire are such as to bring about molecular change. They would also show the chemical composition and the mechanical treatment most conducive to injurious molecular rearrangement, and those tires most likely to retain a permanently tough molecular structure.

The data so far set forth in this Paper have reference only to tires ; but if the theories deduced from the various results herein tabulated are admitted, their application will extend to all classes of steel. There is still much ignorance concerning the causes of the various phenomena connected with the physical properties of steel, although the effects produced by such caiTses are well known. Thus it is known that a certain percentage of carbon endows steel with the property of hardness, and that steel low in carbon will not harden when plunged 'at a red heat into water or oil. The

Proceedings.] ARNOLD ON THE STRENGTH OF BESSEMER-STEEL TIRES. 129

visible eiFect of hardening is to change the fracture of the steel from a coarse to an exceedingly fine texture. On the machine the ductile, ragged, and convex and concave appearance of the fracture of unhardened steel is replaced by a flat, sharp break. The results obtained are exemplified by the following experiments.

A spring steel gave on analysis the following results per cent. : —

Carbon.	^Manganese.	Silicon.	Sulphur.	Phosphorus.
0-50	110	0-07 0-09 i 008

Three test-pieces, 2 inches by 1 inch l)y ^^^ inch, gave on the machine the following figures : —

r Hf f^,..^ Tons per ' Elongation Reduction Condition of Steel. Square Inch. percent. °f^^«^, ! j per cent.
Uuhardeued .... Water-hardened . Oil-hardened ....	50-8 69-4 88-0	14-9 31-4 10-9 .300 3-1 4-9

What has caused such astonishing molecular changes ? Chemical analysis has proved that the carbon has undergone some alteration in form, because in hardened steel it no longer communicates a brown colour to nitric acid to the same extent as in the unhardened material. There is a slight change in the specific gravity and dimensions, the former being, no doubt, the result of the latter. Beyond these three facts nothing is definitely known. Many theories have been put forward, ranging from an idea that the hardening is caused by the molecules being trapped, whilst ex- panded by heat, into a rigid form by the sudden cooling action of the water, to the dream of the theorists who hold that the harden- ing is due to a conversion of the carbon into microscopic diamonds throughout the steel. It is not at all clear why the expansion theory should not apply to all metals, and the veriest t;^T0 in analysis could easily demolish the diamond hypothesis.

On the other hand, it is found that a steel casting when it leaves the mould has its molecules so arranged that it is quite brittle. But the process of annealing, whilst bringing about no chemical change beyond a slight oxidation of the exterior, causes the molecules of the steel to assume some new arrangement, which converts the casting from an article very liable to fracture under the effect of

[the INST. C.E. VOL. XCV.] K

130 ARNOLD ON THE STEENGTH OF BESSEMER-STEEL TIKES. [Minutes of

sudden blows to one possessed of great toughness. Tlie laws which govern the molecular changes of steel, and their relations to chemical and physical causes, require searching investigation. Even the meagre knowledge at present extant is scattered in confusion like the parts of a puzzle, steel-worker, chemist, and engineer each holding a portion ; and until they jjlace them together, and call in the aid of the microscopist, the puzzle will remain unsolved.

Six drawings accompany the Paper, from which the Figs, in the text have been reduced and engraved.

[Discussion.

Proceedings.] DISCUSSION ON STRENGTH OF BESSEMER-STEEL TIRES. 131

Discussion.

Mr. Edward Reynolds said he had not had any personal experi- ^^^- Reynolds, ence in making Bessemer steel, and therefore could not approach the subject from the same stand-point as the Author had done in the Paper, which, although he might be disposed to criticise a few of the Author's conclusions, he considered was on the right lines, and must have required a large amount of patient investi- gation. The main point in it appeared to be the non-desirability of the modern practice of adopting very hard steel for railway tires. It was no new point that hardness alone was riot, even from a mechanical point of view, the crucial quality required to get great power of resistance to wear. In a Paper read before the Institution in 1875 by Mr. J. T. Smith, M. Inst. C.E.,i there was a record of observation on the wear of rails of different tempers which had been used on a long incline on the Furness Railway ; when it was found, contrary to expectation, that the softer rails showed least wear. Many analogies might be found to that experience, as in the case of tool steel. About the beginning of last year he received a piece of a tire from Mr. W. Dean, M. Inst. C.E., of the Great Western Railway. It was a tire made before Mr. Reynolds' Company had any other testing-machine than the falling-weight- test, and before they made any regular chemical analysis. The tire was 2 1 inches thick when rolled. It had been put on in October 1868, and was taken off in December 1886. The diameter of the wheel was 5 feet 9 inches, the weight on the wheels was 1 1 tons, and the mileage 411,349. The piece was long enough to take a tensile-test, and it was found that the tensile-strength was only 36 tons. That, he thought, went to show partly that hardness alone would not suffice, and that it depended upon how the hardness was produced. A certain kind of homogeneity, which offered great resistance to dis- ruption of the particles, was wanted. He was aware that attention had been directed in the Paper to the sixbject of chemical composi- tion, which would lead up to the point he had mentioned ; but, unfortunately, in spite of the vast improvement in processes of production, neither chemists nor engineers, nor users, knew exactly how or why certain good irons gave the required quality. In other kinds of steel the workmen termed it " body." In tool steels the test for body was made thus : A thin cold chisel was made by a

Minutes of Proceedings Inst. C.E., vol. xlii. p. 74.

K 2

132 DISCUSSION ON THE STRENGTH [Minutes of

Royiiolds. man year after year, who was therefore enabled to get it approxi- mately the same, and after being hardened it was tried by being struck with a heavy hammer vertically upon a lump of iron. The steel which would stand most of that treatment, when really hard, without splintering, was held to have the most body. The class of iron which would give that body, would give a sort of homo- geneity, which in that, and in many other cases, led to long wear. He might mention an example in a twin-screw ship, in which the shafts on the two sides were by different makers, the softer one showing scarcely any wear and tear of the journals, while the harder one required frequent adjustment. That proved nothing, but it Avas an example of the fact to which he wished to direct attention. So far he agreed with the Author ; but there were certain points in the Paper which he thought misleading, in the manner in which he had arrived at his deductions. In regard to molecular changes, an example was given (p. 122) of a test-piece prepared by hammering a 14-inch square ingot down to 1^ incli square. The bar was then turned, and showed 15 per cent, elon- gation ; and then a piece planed from a broken tire, after being tested, only gave 3 per cent, elongation. That could not be accepted as any datum from which to form an opinion. Two test-pieces, prepared under wholly different circumstances, could not properly be compared. It would have been easy to take two tires out of the same batch, and to try one without previous shock testing, and the other as described in the Paper. But his reason for mentioning the matter was not so much to cavil at the mode of expression with regard to molecular change, as to express a doubt, whether such change, in the true sense, existed at all. Of course, breaking was a molecular change ; but he thought that was not what the Author meant hy the expres- sion. There was an old idea (which he had himself entertained forty years ago) that there was a change of crystals, and so on, in iron, simply from long use or from some reason of that kind ; but, he thought, it had never been demonstrated that such changes existed. The same bar of iron might be broken, alternately with fibrous and crystalline fractures from one end to the other, accord- ing to the way in which it was treated ; and although it was well known that materials did get to a certain extent altered by long use, he held that it was by what had properly been called " fatigiie." He had before expressed his opinion that materials which were jjcrfectly elastic, without friction amongst their particles, were absolutely unknown, and that when a shaft, like the shaft of a screw-steamer, Avas rotated under deflection, the friction amongst

Procccdiugs.] OF BESSEMEK- STEEL TIEES. 133

the particles gradually caused a separation, just as it was found ilr- Reynolds.

that by Lending a piece of lead backwards and forwards the

parts began to sever. The tire to which he had alluded went to

show that the mere vibration of use had not materially altered its

qualities, which were exactly what might be inferred from the

analyses that had since been made. A comparison of the analyses

with those mentioned in the Paper did not go for much, because,

as the Author had mentioned in a foot-note, superior classes of steel,

such as those produced by the Siemens process, could be made of

materials without so high a percentage of silicon, phosphorus and

other impurities, as was almost necessarily involved in the Bessemer

manufacture. The Bessemer manufacture was to a large extent a

question of cost, and it was in the highest degree creditable to those

who worked it out, that so good a result had been obtained with

what was, in his experience, an inferior material. But he fully

agreed with the Author that it was no nse to attempt to get a

very high tensile-strain with such material. If tool steel, for

example, were tried to be made of the material which had to be

used for the Bessemer process (cost being taken into account), no

" faking xip " by manganese would make it stand the body test

which he had described. His own feeling on the subject, which

was perhaps natural considering that he had been associated with

the higher class of steel-making, was that quality alone was to be

relied upon if high results were expected.

Mr. R. A. Hadfield had much pleasure in complimenting his Mr. Hadfield. fellow-townsman on his very practical Paper, in which he had brought forward important questions that would have to be dealt with in fviture by the metallurgist. In regard to the Author's statement as to hard steel being suitable to stand great wear and tear, he might mention some interesting American experiments in iron and steel practice by Dr. Dudley, specially with reference to the manufacture of steel rails. This gentleman thought that possibly some definite chemical statement might be drawn up for the maniifactiire of steel for rails. But after much investigation, and numeroiis tests, both chemical and mechanical, it was found that no invariable rule or standard could be determined ; and that chemical analysis alone could not be taken as the sole guide, in ascertaining the kind of steel that would stand the most severe wear and tear. In the Author's tests (p. 117) with steel containing 1 • 75 per cent, of manganese, he clearly found that a brittle per- centage had been approached. If he had gone as far as Mr. Had- field's experiments in regard to the effects of manganese upon iron, he would have noticed that manganese between 2 • 50 and 7 • 00 per

13-1 DISCUSSION ON THE STRENGTH [Minutes of

Wr. Hadfield. cent., whilst adding to the hardness, added also very much to the hrittlenessi It might also be interesting to refer to some tests carried on by the Terre Noire Company in 1878 with test-pieces of similar dimensions to those referred to by the Author, only on a 4-inch length, averaging about O'oO per cent, of carbon, the manganese increasing from 0 • 50 per cent, to 1 • 00 jier cent., 1 • 30 per cent., and 2 • 00 per cent. The results showed that 0 • 5 per cent, of manganese gave a tensile-strength of 34 tons per square inch with 24 per cent, elongation ; when oil-tempered, 43 tons and 14 per cent, elongation ; 1 • 00 })er cent, manganese gave a tensile-strength of 41 tons, the elongation being 23 per cent. Steel with 1 • 30 per cent, of manganese gave a tensile-strength of 57 tons, and 9 per cent, elongation. In the oil-hardened sjiecimens the tensile- strength went as high as 82 tons, but with scarcely any elonga- tion. The Author mentioned that he used up to 1"75 per cent.; this was getting dangerously near to the amount that might be used with safety in ordinary steel. In fact, it was surprising to hear that he had got so good a result. No doubt this w^as owing to the carbon being so low, only 0*25 per cent. The Author had referred to the effect of working the metal, which was a very important jioint ; but he thought he had somewhat overrated the effect of this on steel, because it had been clearly pointed out by General Maitland,^ that there was very little differ- ence in the tests obtained from blooms forged from 10-inch ingots down to 7, 5, and 2 inches square. The last bloom with the smallest area, when a test-piece was taken from it, gave no better results than the first. He thought, therefore, this was a point that should be carefully considered before pronouncing a decisive opinion upon it. The matter had also been referred to by Mr. J. Eiley, of the Steel Company of Scotland, in a Paper read before the Iron and Steel Institute in 1887, on the treatment of mild steel. '^ There was no more eminent authority in the steel trade, and Mr, Eiley had expressed his opinion in reference to mild steel (he did not know whether the same thing would apply to hard steel), that abundance of work put upon steel ingots rather tended to produce a strong steel than to give more ductility. Indeed, he added that if ductility was required, an excessive amount of work should not be put upon the steel. That was also confirmed by Mr. William Parker, of Lloyd's, who mentioned that from an ingot 24 inches by 15 inches, a plate 1 inch thick was rolled, Avhich had a tenacity

' Minutes of Proceedings Inst. C.E., vol. Ixxxix. p. 126. - The Jt)ninal of the Iron and Steel Institute, 1SS7, p. 121.

Proceedings.] OF BESSEMER-STEEL TIRES. 135

of 27" 7 tons, with an elongation of 23 per cent, on 8 indies; and Mr. Hadfield. a plate of similar size rolled from an ingot of half the sectional area gave 2 per cent, more elongation. Although, perhaps, not absolutely conclusive, this tended to show that the value of work on steel might be too highly estimated. He specially mentioned this matter, being personally interested in the manufacture of steel castings in which it was necessary to produce a tough but un- worked material. If a sound test-piece was obtained from a mild steel casting, very considerable elongation, though perhaps not with the same certainty, was obtainable, as with forged steel. The Author had mentioned (p. 128) his method of selecting samples for tests. Mr. Barker, of the Great Indian Peninsula Eailway, had a very good system for testing steel tires ; but he did not know whether all steel-makers agreed with it. He tested each one by dropping it from a certain height, so that if there was the slightest strain in it this was found out, and the tire rejected. He under- stood, from a recent conversation with Mr. Barker, that in conse- quence of that very careful method of testing tires, there had never been any breakage in service. The subject of the structural changes in steel formed a very important part of the Paper. It was a subject with which steel-makers were not as conversant as they should be. The methods of analysis were not perhaps perfect, but they had been so much improved during the last ten years, that he thought chemists could not well tell more than they had done in regard to the composition of steel. Moreover, it had been proved that chemists in steel-works laboratories could not detect analytically the diflerence between two pieces of steel, which might give very different results in the testing-machine. That applied to castings as well as to forgings. What, therefore, seemed necessary in the future was a closer examination of the structure of steel. As the Author and he had pointed out, there was much to be done in that direction. If the microscopist could help, by all means let his aid be called in. But there was another point on which a great deal was to be learned, and no more able expositor of the subject had been found than Mr. Howe, of Boston, U.S.A., in his work on the "Metallurgy of Steel," now being published. He had found that in the heating of steel there were several critical points of vital importance in determining its future quality and temper. These critical j)oints varied according to the particular temper of the steel being operated upon, and did not occur at the same temperatures in hard as in mild steel. Unfortunately the means of ascertaining these changes were not at present satis- factory. When speaking of the heat to which a piece of steel had

13G DISCUSSION ON THE STRENGTH [Jlinutes of

Hadfield. been subjected, at present vague and loose terras were employed. In fact, they were often quite indefinite, as diiferent persons often judged diflerently of the same temperatiire. Dark red, Idood red, cherry red, and the like, were used indiscriminately in describing- lower temperatures, and the same remark applied to higher ones. Yet from the investigations of Chernofi', Osmond, Pionchon, and others, slight differences in temperature might make considerable difterences in the future characteristics of the steel being treated. This applied especially to hard steels, which were more liable than the softer kinds to structural changes by the degree of heat employed in their treatment, and therefore it would be of great service to metallurgists to have more accurate means of knowing what degrees of heat were really used in their operations. A good p;yTometer would be of the greatest value, but although several l^jTometers were in the market, they did not seem thoroughly reliable. He had emjiloyed one made by Murray, of Glasgow, which had given fairlj^ good results, but still at times variations occurred with them, and these necessarily caused confusion in the experiments. Many leading metallurgists fully agreed on this point, and therefore it was to be hoped that some accurate, and yet sim2:)le, instrument would before long be perfected.

Professor W. Chandler Egberts- Austen observed that the Author had remarked that the process of annealing brought " no chemical change beyond a slight oxidation of the exterior." Surely it was well known that although ultimate analysis might show no dif ference in the amount of carbon present in hard steel and in soft, nevertheless the mode of existence of the carbon was totally different in the two materials, and the mechanical properties of the steel entirely depended on the mode of existence of the carbon, that was, whether it was combined or free. The Author had further stated that "Chemical analysis has proved that the carbon has undergone some alteration in form, because in hardened steel it no longer communicates a brown colour to nitric acid to the same extent as in the unhardened material. There is a slight change in the specific gravity and dimensions, the former being, no doubt, the result of the latter. Beyond these three facts nothing- is definitely known." Professor Eoberts-Axxsten thought that cer- tainly a little more than that was known. Le Chatelier and Pion- chon had shown, by independent methods, that, on cooling from a temperature of aboiit 700° Centigi-ade,iron passed from one allotropic state to another. Apparently the a, or soft, modification of iron was present in the greatest quantity in slowly-cooled iron, and when the iron was heated it really passed to the hard, or /?, modi-

Proceedings.] OF BESSEMER-STEEL TIRES. 137

fication. Osmond had further shown that, in carburized iron, a Prof. Roberts- change in the relation of carbon and iron took place during cooling -^^i^^*^"- from a high temperature. He had also shown that the final tem- })erature at which that change took place sank lower and lower with the proportion of carbon, until, in a steel containing 0 • 8 per cent, of carbon, the passage of iron from one allotrojiic state to another exactly corresponded with the temperature at which carl)on itself passed from one allotropic state to another. When carbon was in the proportion of 0 • 80 per cent, it simply hindered the iron, which was being rapidly cooled, from passing from the hard modi- fication to the soft. A pure iron would not harden at all, however rapidly it might be cooled, because there was no foreign element present to prevent it passing from the hard modification to the soft. There was thus direct evidence of the possibility of iron being prevented from passing from one allotropic state to another under the influence of an added element ; and if, as was known to be the case, added elements produced allotropic changes in metals, the influ- ence of such added impurities ought to be governed by the periodic law of Newlands and Mendeleef. Professor Eoberts-Austen had tried to show that in the case of gold the influence of the added impuritj^ was strictly governed by that law; and he had stated that if certain elements were added to gold it might become as brittle as sugar, and the larger the atomic volume of the added impurity the greater would be the disturbance produced. It was difticult in the case of iron to trace the influence of the added impurity, because so few experiments had been made on the in- fluence of a single element added in small quantities to pure iron ; but he ventured to predict that investigation would ultimately show that the larger the atomic vol^^me of the added element, the less work would it be possible to do on the metal, to which it was added, without rupturing it, either by longitudinal or by transverse stress.

Mr. W. Mattieu Williams observed that there was one statement Mr Williams, in the Fa]}er which he had heard with some surprise, namely, the statement as to the growing tendency amongst railway engineers to specify, for tires, steel possessing a higher resistance to tension than formerly. He had discussed the subject twenty years ago when writing on phosphorus in steel, and since that time he had given further attention to the matter, and was satisfied that, in increasing the tenacity of steel, as measured by a gradually applied strain, the brittleness was proportionally increased until it reached the brittleness of glass. Colonel English had made many experi- ments at Sir John Brown's works in Sheffield, in connection with

138 DISCUSSION ON THE STRENGTH [Minutes of

Mr. Williams, the subject, especially with regard to the bolts used for bolting on armour-j^lates, which were, of course, subject to violent shocks. In practical work the qiiestion was not how much pull might be put upon a thing, even the suspenders of a bridge, but how, when the pull was ujion it, it would stand a vibratory shock. He might mention an experiment which he had tried when he was chemist at the Atlas works. He took a test-piece of plate-metal 3 inches by ^ inch, and put it under strain, and w^hen it began to yield a little he gave it a smart tap with a small rod of wire, and it suddenly broke. On one occasion a steel rail when pitched on a lorrie, in the yard of the works, struck the iron edge of the lorrie, while the other end struck the iron pavement, and it broke in half merely by the action of the double shock. It seemed hardly credible, but on consideration the rationale of it might be understood. He had himself occasionally tried a simj^le experiment bearing on the subject. He had taken a rod of steel, of any convenient size, and clamped it firmly to a table ; he then placed a marble at one end, just touching it, and gave the other end a tap with a hammer. Al- though the bar was not moved at all, the marble was shot off from the other end. A wave of compression travelled along, terminating at the end, and thrusting the marble off. In like manner, he imagined that the breakage of the rail, like many other breakages, was caused by a wave of compression, or perhaps by two waves going in opposite directions, and meeting like two waves of water. A great amount of compression and also of extension was obtained in that way. The difficulty was to overcome it, to enable the metal to resist a vibratory shock when under strain. He had little doubt that the Tay Bridge gave way in consequence of its inability to do that. The wind was producing a strain upon every part of the bridge, and when the train arrived a vibratory shock was produced, which, acting with the strain of the "wind, proved fatal. He had been surprised to hear what had been stated in the Paper, and elsewhere, with regard to the effects of manganese, since he had arrived at a totally different conclusion on the subject, namely, that manganese was simply a mischievous impiirity in steel. He did not refer to Mr. Hadfield's alloy of iron, about which he knew nothing ; he simply referred to manganese in the proportion of 1 • 00 or 0 • 50 per cent. He once thought that manganese was a great improver of steel and iron, and found an easy method of introducing the required quantity, esj^ecially into puddled steel or iron, and he had taken out a patent for the purpose. His method was to use as the fettling of the puddle furnace, either alone or mixed with hematite, a paste consisting of

J

Proceedings.] OF BESSEMER-STEEL TIRES. 139

black oxide of manganese. The metal was then poured in. Its Mr. Williams. action produced a large supply of oxygen, which assisted one part of the puddling process; and by the reduction of the oxide of manganese, by the carbon contained in the iron, a certain amount of manganese was produced. The results were tested by himself and others, and the conclusion arrived at was, that the more manganese there was in the iron, the worse the resulting metal proved to be. Taking a given sample, its quality was improved in proportion as the manganese was eliminated from it. The same thing occurred with tool steel. His conclusion was that the man- ganese first of all combined with the oxygen existing in minute particles of the black oxide of iron, which seriously damaged the iron, and then the oxidized manganese combined with the residual silicon, forming a fusible cinder readily squeezed out in working. In regard to the question of manganese improving the quality of steel, and giving it a high tensile-strength, he might mention that at the time to which he referred they were using spiegeleisen, and the ordinary carbon test was 0 • 40 and 0 • 50 per cent, for rails, and 0*10 per cent, more for tires. In regard to the requirements of rail- way engineers to which reference had been made, he would direct attention to the specification of the Admiralty ten years ago for ship-steel, which demanded that it should be raised to a cherry-red heat and then plunged in water, and that the tensile-strain should not be less than 26 tons nor exceed 30 tons jier square inch. If it exceeded 30 tons the steel would be rejected, as in his oiunion it ought to be ; 20 per cent, of elongation was required. The matter was a very important one, it being a question of overdoing the hardness and tenacity, at the expense of the brittleness of steel, whether used for tires or rails or anything else liable to vibratory shock.

Mr. E. A. CowPER said he entirely agreed with Mr. Eeynolds in Mr. Cowper. thinking that a few blows on a tire would not alter its molecular arrangement. If a piece of steel were bent, and a straight piece ciit out of it, it would have undue strain on it, compression on the one side and tension on the other. That would not be a fair comparison with a straight bar ; and therefore a piece cut out of a tire did not bear the same strain as it would when it was straight, and before it was put into the tire. That was indepen- dently of any 2)roof of molecular disturbance being set up in the tire from the blows. If a tire was put on a wheel or circular block and then hammered, the strain necessary to pull it in half would not be found much less than what was required when it was a straight bar. Some experiments were made many years ago on

140 DISCUSSION ON THE STKENGTH [Minutes of

the London and North Western Railway; as axles were said to be crystallizing. These axles were made somewhat longer, and after two years, when the ends were broken off, no difference conld be found in the quality of the iron. But if an axle were turned down to a square shoulder, and was much larger behind the wheel than through the wheel, it formed a nick and would break there. He had many years ago introduced the plan of having a very small rounded shoulder behind the wheel. Steel axle castings were said in the Paper to be quite brittle before being annealed. That he must deny. Many steel castings in use were by no means brittle, but they were not quite so tough as after they had been worked. He also took exception to a 14-inch ingot drawn to Ij inch square as a fair way of trying the diictility against a tire 5j inches by 2 inches. They were two different samples, and one of them had had. a great deal more work put on it than the other, so that a great difference would be expected. The Author had stated that " the amount of work put ujion steel has a marked relation to the mole- cular structure, and consequently to the ductility of the material." That was well known, and he agreed with the statement. The reduction of area and the extension dejiended upon whether the bar gave way uniformly or at one point. But when it gave way at one point, the reduction of area at that jDoint was nearly the same whether the bar was long or short. It was very seldom that the elongation was equal throughout the whole bar, hardly once in one thousand times. The addition of a small quantity of chromium, if it gave the reqiiired hardness, and at the same time did not injure the ductility, would be a great point gained. He denied that tires gave way frequently with the present degree of hardness required. Accidents with broken tires were very few in number, far fewer than formerly with wrought-iron tires, and than when steel tires were first introduced. That might possibly arise from the improved methods of manufacture, not entirely from the tires being harder. Broken tires were now fewer than ever, and their hardness was greater than ever. There had been recent examj)les of large contracts for tires with a high tensile-strength and a very con- siderable amount of ductility ; but he was sorry to say, that those contracts had gone abroad, the conditions being found difficult to fulfil in this country at the time the contract was given out.

Mr. J. A. F. AspiNALL stated, with reference to the question of crystallization of steel after it had been at work for a long period of years, that he had sent to the Institution some years ago a short Paper, together with a series of photographs showing the results obtained bv breaki'n"; a number of steel crank-axles whicli

Proceedings.] OF BESSEMER-STEEL TIRES. 141

had been in use for a long time. It so happened that in many ^Ir- Aspinall. cases, the webs which had been originally cut out of these crank- axles had been kept, so that tlie fracture of the material which had not undergone any work could be compared with the fracture of the crank-axle wliich had been in use. At the same time a number of crank-axles of Yorkshire iron were broken up, but no com- parison could be made with pieces of their webs, as they had not been retained. The steel axles sliowed that they had under- gone no change after many years' work, and the tensile-tests were exceedingly good. Every one of the iron axles showed very large crystals. With reference to the subject of annealing mild steel, he had recently put some plates into the testing-machine, and after they had been stretched 10 per cent, they were annealed ; and this operation of stretching and annealing was repeated seven or eight times with several specimens, which in each case gave an elongation of from 85 • 5 per cent, to 89 per cent., the maximum load gradually rising until it was 20 per cent, higher at the final test than witli several specimens which had previously been tested witliout any annealing, and which gave an elongation of from 2-t to 28 per cent. With regard to tires, it was difficult to trace the work of a tire from the day it left the manufacturer's shop until the day it was taken off the engine. A good tire would last from five to ten years, and it was not easy to watch experiments for so long a period. The amount of wear to be obtained out of good 6 feet G inches tires, on coupled engines carrying 15 tons on a pair of wheels, was about -I inch for 40,000 miles.^ He could not agree with the statement that engineers were demanding too high a tensile-test for tires. He had had large numbers of tires made of steel by the open-hearth process, giving a tensile-strength exceeding 46 tons per square inch, and he had never known one of them to fail. The steel of some of the tires made by this process which gave the best results, contained 0-65 per cent, of carbon, 0*28 of silicon, 0*09 of sulphur, 0*06 of phosphoriis, and 0*86 of manganese; and in the case of another maker 0 • 60 jier cent, of carbon, 0 • 23 of silicon, 0 • 07 of sialphur, 0 • 02 of phosphorus, and 1-11 per cent, of manganese. The composition advocated by the Author for tires would not, he considered, be a desirable one, if a long period of wear was held in view, for the wear and tear of the tire was often the measure of the period during which tlie engine could be kept out of the shops, which was a most important consideration. If a tire went on to an engine 3 inches thick, and had to be taken oif again when it was about 1^ inch thick, the maximum amount of wear and tear

' Minutes of Proceedings lust. C.E., vol. Ixxxi. p. 121.

142 DISCUSSION ON THE STRENGTH [Minutes of

shoiild be taken out of the first 1^ inch. There were tires in the market possessing every desirable qualit}', which were perfectly reliable so far as breakage was concerned, and which gave a reasonable life.

Mr. W. B. Lewis was sure the members would agree that steel- workers, chemists, and engineers should collate all the informa- tion they individually possessed for the common benefit. He thought engineers were using a steel of a very much higher character than that dealt with in the Paper, and he rather objected to the insinuation that they were doing so at a considerable risk. The Victorian Government placed a contract with Messrs. Cockerill and Co. for three thousand tires, the si^ecification being that the steel was to stand a tensile-strain of 45 tons, with an elongation of 20 per cent, in a length of 5 inches. Some trouble was at first experienced, biit after sundry experiments and failures, the firm succeeded in manufacturing uniform steel to answer those require- ments. The analysis of the steel, as made by the chemists at the works, was : carbon, 0 • 330 per cent. ; silicon, 0 • 220 ; sulphur, 0 • 041 ; phosphorus, 0 • 057 ; manganese, 0 • 750 ; iron, by diff"erence, 98 • 602. That, as a qualitative analysis, he believed was perfectly correct : but he had reason to doubt it as a quantitative analysis. The Inspector, in forwarding it, wrote : " I believe the standard used for the carbon test gives a much lower percentage of carbon than that generally used in the English works." That was confirmed by an English analysis. General Maitland, who felt interested in the subject, asked him to send some of the steel for examination at Woolwich. This was done, and the General found that the tensile-tests agreed with those taken at the works and those also made by Mr. Kirkaldy and Professor Kennedy. He wrote, saying : " The steel is very good, and tempers well ; our analyst makes the carbon 0'516, the manganese !• 038, and the silicon 0-122." Thus the carbon was much more than had been stated at the works. He wished to direct attention to the circumstance that the steel contained the same ingredients as those mentioned in the Paper as belonging to the standard, or what the Author seemed to take as his normal type. There was no chromium and no special material to make it hard. As to the safety of the tires, they were all subjected, not only to the test mentioned, a bar being cut out of one in fifty tires ; but a complete tire was tested by an impact-test of a 1-ton weight falling 5 feet, 10 feet, 15 feet, and then ten blows of 20 feet. The extreme deflection under the 15-foot blow was 3^ inches; from the 20-foot blow it varied from 2i to 5^\ inches. He thought tires that would stand such an

Proceedings.] OF BESSEMER-STEEL TIRES. 143

amount of knocking about, might be used with confidence. The Mr. Lewis. English makers had refused to tender for the tires at that time ; but he was happy to say that since then they had made a great number to similar si^ecifications, excepting that the elongation was less. Three leading English manufacturers had subsequently been engaged in supplying these tires. Two of the firms had made steel possessing 45 tons tensile-strength, with 15 per cent, elongation on a 5-inch test-bar ; but the area of the test-bar was ^ inch, whereas in Messrs. Cockerill's case it was 1 inch. In the case of the third, the tensile-strength was 45 tons, with 10 per cent, elongation. But these tires had been subjected to greater tests than any others. The ordinary test had been that of 1 ton falling 5, 10, 15, 20, and 25 feet, and 30 feet twice. The highest deflection obtained under the 15-foot blow was 3 inches, and under the 20-foot blow 4^ inches. It was therefore much harder steel in every way than that which the Author had mentioned. Out of a contract for four hundred tires, six tests were made. One tire broke at the 30-foot fall, having stood the 5, 10, 15, 20, and 25-foot falls. The five others stood the two 30-foot blows, and remained unbroken. Under another con- tract for nearly two thousand tires, in which thirty-four tests were taken, four tires were broken ; one at the first 30-foot blow, and three at the second, all standing the 25-foot blow. Another manu- facturer made one thousand, with results equally satisfactory. Again, another made over two thousand tires. Forty tests were taken, and three tires failed; one at the 10-foot blow, qne at the 15-foot, and one at the 20-foot. The tests in this instance were not quite the same. They were 1 ton at 5, 10, 15, and 20 feet, repeated five times. He thought that steel which seemed to contain the same ingredients as those mentioned by the Author, but was very much harder, might be safely used. The tires, however, had been made only a few years, and he could not yet speak as to their wear in the same way that Mr. Reynolds had spoken -of the Great Western Railway tire. They had not existed long enough ; but there need be no fear to use a tire that could stand so miich knocking about. Mr. Reynolds had referred to the experiments tried on the Furness Railway, which, according to the contention of Mr. Smith, of Barrow, showed that a soft iron rail wore better than a hard one. In the first j)lace, he was not sure that the analogy held good between a rail and a tire ; and in the second place, he thought that the experiments were of a very narrow range. He believed that the steel sustained a strain of from 28 to 34 tons, which was very different from steel sustaining 45 tons. The universal practice of manufacturers, who had the

144 DISCUSSION ON THE STRENGTH [Miuutes of

highest reputation for making good tires, had been to strive after harder steel. He hojjed, therefore, that the Aiithor was not war- ranted in insinuating that engineers, in Tising hard steel of that kind, were nsing that which involved any risk. The steel had been made by the open-hearth process, and there had been no necessity to nse chromium, nor any other unusual material, to get the requisite hardness.

Dr. H. Clifton Sorby said he would confine his observations to that part of the subject which had attracted his attention for many years, namely, the microscopic structure of steel. He thought that what might be seen by examining suitably prepared objects, with high magnifying powers, would throw a good deal of light on some facts that had been ascertained in an independent manner. The microscopical structure of Bessemer steel was very interesting and curious. It somewhat resembled, though to some extent it differed from, the structure of certain portions of charcoal Swedish iron containing a certain amount of combined carbon. The structure was very complex. There was no difficulty in recognizing by the microscope two totally distinct substances. One of them he took to be free iron, or iron containing next to no carbon, if any, and the other a chemical compound of a certain amount of carbon with iron. The appearance of the two under the microscope, in prepared sections, was so different, that one could not possibly be mistaken for the other ; they were as distinct as any two minerals could be. The conclusion at which he had arrived was that, in cooling from a state of fusion, what he regarded as a comi^ound of a certain amount of carljon with iron crystallized oiit first ; and that in crystallizing it threw off the superfluous amount of iron and entangled some of it, but threw out the greater portion towards the outside ; so that, in looking at a section of the steel in the micro- scope, a i^eculiar kind of network was seen, that network being the outside of the crystals, also extending in a most intricate and curious manner along what he took to be some important planes in the crystals of the other compound. On studying the matter more fully, he thought it might safely be said that there was good evidence to show that the original crystals which crystallized out in the first case, on cooling down to a certain temperature, under- went a complete molecular change, so that they might be regarded as not exactly pseudomorphs, but (to use a mineralogical term) paramorphs ; that, in fact, after cooling to a certain point, they underwent molecular changes, breaking up into a number of crystals not necessarily related to the crystalline structure of each original larger crystal. That appeared to him to be a satis-

Proceedings.] OF BESSEMER-STEEL TIRES. 145

factory way of explaining the peculiarities to be found in properly Dr. Sorby. prej)ared sections of an ingot. In annealing the specimen, how- ever, a complete change took place. Instead of having the curious complex mixture to which he had referred, the two constituents were separated from one another more completely, and on a smaller scale ; the crystals were smaller and each one more simple. The difference between the annealed ingot and the ingot in its original condition, as seen under the microscope, was so great that no one could hesitate to say that most important changes had occurred. That, he thought, agreed well with some of the conclusions of the Author. When the specimen was rolled or hammered, he believed that his own specimen was rolled, a complete change took place; the crystals were much smaller, and the two distinct substances were broken up and made more uniform. It might well be believed that the difference in structure between rolled and unrolled steel was very great indeed. There were the same two constituents, only on a smaller scale and more uniform, as in fact might well be supposed would be the case. Light was also, to some extent, thrown on the question of hardening. His experiments on that subject were not with rolled steel but with the ingot itself. He thought that, having a coarser structure to deal with, the differ- ence in the structure after hardening would be more apparent, and so it was. Taking a piece of an ingot in its natural state, raising it to a red heat, and plunging it suddenly into cold water, and then making a sectiou and examining it with a micro- scope, a most remarkable and complete change could be seen ; the whole was uniform as far as could be judged with the micro- scope. The ultimate constituents were so small, that even with a power of 400 linear, which could be used with perfect efficiency, they could not be distinguished. The question of hardening would probably prove one of the most difficult to deal with from the microscopist's point of view, because the ultimate constituents after hardening were so minute, that it would be difficult to distinguish them even with very high powers. It appeared to him that there were two important changes. In the first place, when a portion of the ingot was raised to a red heat, it seemed that the two original constituents mutually dissolved, so that at a high temperature there was, to a certain extent, a reproduction of a state of fusion, though not strictly fusion — the constituents were more uniformly distributed — and when hardened, there was no time for them to separate. He believed that, independently of there not being a separation which occurred in annealing, and in a slowly cooled ingot, there was some most important change very difficult to

[tHK INST. C.K. VOL. XCV.] L

146 DISCUSSION ON THE STRENGTH [IMimites of

investigate. Those facts, lie thought, threw much light on some of the qiiestions under disciission. As far as he had been able to judge by applying the microscope to the structure cf iron and steel, there was so much still to be learned on the subject, that speculation should not be carried too far. In order thoroughly to investigate it, some of the young men of the day should devote twenty or thirty years' study to it, as he had himself done to the structure of rocks. After such an investigation they might possibly know something about the constitution of that most remarkable element, iron. His own work had shown him how much yet remained to be learned, and if he had succeeded in attracting attention to the subject, as one well worthy of in- vestigation, he should feel amply repaid.

Mr. W. G. KiRKALDY desired to make a few remarks on the subject from a mechanical point of view. He had had occasion to make many examinations of tires under pulling and other tests. No doubt the subject was interesting from a chemist's point of view; but he thought engineers in their specifications could deal with it very well by means of mechanical tests, without tjdng down the manufacturer to a certain chemical composition. It was no doubt needful to have very strict specifications as to the mechanical requirements which would be suitable for tires to stand wear and tear. It was not, he thought, necessary to make tires of very hard steel, but rather to aim at great toughness. He had examined tires from many different makers, and most of those received from Krupp had shown remarkable properties for tough- ness, and had given a very high strain, yet with such large contraction of area that he did not think there could be any risk in using them. The results showed that an increase in hardness did not necessarily mean an increase of tensile-strength. In the case of one specimen the ultimate tensile-strain was 50-9 tons and the contraction 41 • 5 per cent., a high contraction for such a strain. In another case the ultimate tensile-strain was 22 • 3 tons and the contraction was only 2 ' 2 per cent. For his own part he preferred specimens which bore witness that they were actually out of tires, the heads showing the section. Unless the engineer was very strict, to see how the tests were obtained, he might be easily misled by results obtained from the small specimens customary with manufacturers and others. Evidence of one effort to obtain unduly high results was demonstrated by the small specimens which Le submitted for examination. He had called the engineer's attention to the fact that, although the specimens sent prepared for testing might have been out of tires,

Proceedings.]

OF BESSEMER-STEEL TIRES.

147

o

o

^

C b

S '='

o

o

o

CO

05

O

O

o

C 00

IC

o

o

o

O

o

9< oJ

"<»

(O

05

05

CO

o

P 05

C5

o

O

o

52

a

_Ph

'"'

c

'"'

^£

>a

go

O

o

o

r^

^

o

o

o

i^iM

UO

o

o

o

o

o

11

■*

o

o

US

-+I

o

o

»

Fh

'"'

'"'

>

^

'"'

1

1?=

•<*<

<M

CO

00

(M

tH

CD

iM

o

CD

00

(M

00

UO

CO

Q

IM

CO

fN

CO

O

lO

CO

CO

to

iO

iM

iO

lO

IM

Ol

;^

feiN

<N

IN

<M

IM

CM

(M

(M

ja

;3

Il4

° S £ ,.;

S<=>

IM

CD

-t<

^

■^

g H

00

CO

o

en

n

:ll^ -<

'^.^

CO

00

00

o

00

*

(^

'"'

c

l^H

S^^

'tl

00

o

-tl

-tl

00

-tl

-tl

CO

g<x>

lO

CO

-tl

lO

00

IM

-tl

CD

I-

c

.Bg^

£ 1— 1

,_,

Oi

IM

CO

(M

■*

Tt<

-tl

o

•,"1

P^

dj

o

^

o'=-S-S

O

o

o

CO

IM

CO

-H

CD

-tl

IM

-t<

Cf)

IM

o O

o

o

o

<M

o

lO

CD

CO

-tl

1^

-tl

O

(M

■^ -S-c

mO

o

o

o

^

^

;.^

^

<M

^

;^

^

•*

-t<

<1-

^

^

^•a

CO

CO

t^

CO

rn

CO

1^

P^

t^

CD

-+<

t^

r^

O

IM

-t<

o

^

o

IM

r^

CO

lO

IM

-tl

IM

Ift

r-

CO

>o

-H

CO

p.

3 ^ S »S

,/t^

(N

t-

CO

iO

IM

00

CO

lO

03

r^

o

IM

CM

/

~ 00

lO

t^

-tl

o

r^

CT)

05

r^

lO

t^

-tl

-tl

J~{

r— t

CO

" 00

Ci

00

00

CO

CO

-+l

CO

00

o

t^

o

CD

O

CO

OQ

'^

'"'

'"'

^

'"'

^

'"'

^

'"'

'"'

g

_ -i^ «5

^

«sl

"o

Irt

CO

o CO

00 <M

IM

CO in

o

o

CO 00

IM IM

(M -tl

CO CO

IM IM

C<l Ol

O

-^1 ? -3 2 .

rt<

•*!

-*l

IM

-tl

CO

10

■«

^

i-S|i

s®

l-(

(N

1-H

o

■*

>ei

lO

T-H

o

>c

t^

t--

oo

-+I

IM

r-o

^

05

rsi

CO

t^

^

o

o

^

o

o

O

o

<7i

CM

^S3-

feiO

iCi

<o

o

-*<

-*l

<o

lO

CO

CD

»o

CD

CO

CD

t-

CD

pL,

s

ajM

o

o

CO

cc

CO

IM

o

CD

CO

IM

CD

CO

in

CO

CO

O I— (

o

CO

CO

-tl

>o

CO

CO

QO

CD

Ci

IM

CO

CO

<M

in

n 1

H«5

lO

■*

-tl

-tl

-tl

-tl

-tl

(M

CO

CO

CO

CO

CO

(M

CM

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

t S.2

lO

o

o

r^

o

o

o

o

o

o

lO

lO

ITS

o

o

lO

1 ^ 5 ^

lO

(M

-tl

tM

00

CO

Oi

00

lO

r^

OJ

r^

CO

Ol

OO

^•^

IM

CO

-tl

(M

CO

-tl

Ci

IM

00

o

i-H

o

CO

;^-ti

-^

-tl

-tl

o

IM

CO

CD

CO

,_!

r^

CO

■o

UO

o

t-

o

o

o

O

03

Ol

CO

00

00

t>

t>

t-

lO

lO

«o

CO

UO

lO

t^

O

(M

c^

IM

,^

00

00

■<*<

-tl

00

CI

£ -g

^f-

r^

r^

t^

o

._!

IM

,_!

t-

IM

05

C5

o

o

r^

in

E-ic^

IM

iM

IM

IM

(M

IM

IM

IM

(M

(M

c

C u^

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

II

2P.S

©

o

O

o

o

O

O

o

o

O

o

o

o

o

o

o

o

o

o

o

O

O

o

o

O

o

o

o

o

o

o

S W

00

l^

l^

CO

o

00

t^

lO

CO

-tl

-tl

t^

l^

00

t-

fc

ij CO

5^'

^H

,_,

,_,

CO

00

05

00

in

05

-+<

-tl

"O

IC

Ol

lO

CD

o

CO

CO

-tl

~n

-tl

-t<

CO

^

-tl

-tl

-t<

-t<

CO

CO

o-;o

o

o

o

o

o

o

o

o

o

c

o

o

o

o

o

CO

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

MO

o

lO

lO

o

o

o

o

o

o

o

o

o

o

o

o

a

-<)

^^

(M

^

^

IM

IM

<M

(M

-'

^

IM

0<l

(M

(M

(M

CM

t^

r-

(M

(M

r^

t^

t^

r^

00

00

r^

r^

t^

r-

t^

l>

3

■f=!^

05

cti

or)

C5

02

03

C3

IM

(M

Ol

05

en

Ol

Ol

Ol

lO

CO

CO

lO

lO

lO

lO

lO

lO

lO

in

•o

lO

5

o|

"-1 1—1

1— 1

1—1

r-H

1—1

^

^

1— 1

-^

T-t

1-1

1—1

i-H

^

1— 1

6

(M

CO

CO

r^

(N

CO

l-ft

<ri

o

^

^

(M

CD

r^

Cf)

Ol

^

00

00

o

o

O

lO

iM

CO

^H

lO

lO

lO

'O

00

00

L^

l^

CO

CO

CD

CD

l^

c^

o

t>

Ci

C5

IM

CM

H

CO

CO

CO

CO

pH

rH

(M

iM

CM

M

C-l

;-j

«

■S

>

>

c

O

Mr. KirkaUly.

L 2

148 DISCUSSION ON THE STRENGTH [Minutes of

Mr. Kirka^dy. as represented by the maker, yet as they had been forged down, instead of being turned from tires in normal condition, no reliance should be placed upon them. The matter Avas one that should be very carefully dealt with for the experiments to be thoroughly reliable. There should also be a standard size in order that the experiments might be compared. At present an engineer often found that he could not compare the specimens tested at one maker's works with tests sent to him from another maker ; but if the same size were adhered to, which had been a great point . witli his father for many years, that difficulty would be removed, and the tests could be compared so far with each other. Engineers should be protected from sham experiments. He had known many cases in which they had been in danger of being misled by results put forward by interested persons, but such results had proved ultimately unfavourable to the English maniifacturer. It was naturally important to makers to obtain good results; but if the material on going out did not come up to what was repre- sented, a reaction would set in. He had known several cases in which, in consequence of failures in that respect, orders had been given to a foreign firm, which had managed to keep its footing. Keference had been made to some experiments said to prove that work on steel did not necessarily improve it very much. Accord- ing to his own experience, the fact was the very opposite. He had seen the most conclusive proofs that work on steel had a great deal to do with its character, and referred to an extensive series of experiments upon Fagersta steel, by his father, which had been piiblished in full detail. Any gentlemen, desiring to see the niimerous specimens collected at his works having a bearing on the subject under discussion, would be heartily welcomed, and he felt confident they would gain information that would prove of practical value. The tests in the Table (p. 147) had been selected for the purpose of showing the difference in the qualities and behaviour of various grades of tire-steel. They were a j^ortion of the forty-two specimens of tires which he. exhibited. Jlr. Berkley. ]yjj._ CrgoRGE BERKLEY, Vice President, Said it would be readily admitted how very important chemical knowledge in regard to steel, and other materials used by the engineer, had been, still was, and was likely to be. Without this knowledge, Bessemer steel and Siemens steel, two aids to the engineer probably as useful as any that could be mentioned, would still have been unknown. He was not satisfied with the reasons assigned by the Author for attributing importance to the particular subject he had mentioned. The Author had referred to the question as one affecting the safety

Proceedings.] OF BESSEMER-STEEL TIRES. 149

of thousands of lives. That was scarcely suitable for a scientific Mr. Berkley Paper. He had also stated that it was impossible to have too great a margin between a working- and a breaking-strain. Surely some- thing more definite than this was required. The point was to ascertain the admissible difference between the working- and the breaking-strain. The Author had further stated that there was a growing tendency amongst railway engineers to specify for tires possessing a high resistance to tension. Looking at the specifica- tions of other engineers, and also having regard to his own, he did not think that the statement was a correct one. Eight or ten years ago there was, unquestionably, an advance, which was justified by the improvement in the manufacture of the material. The Author probably did not mean it, but he appeared to attribute to engi- neers a consideration for economy, even at the sacrifice of safety. Nothing could be less accurate than such an inference. It had been his duty to send tires abroad for thirty-seven years. He had not been sufficiently industrious to examine all the analyses and experiments during that time, but he had carefully scrutinized them for the last ten years. During that period he had sent abroad about thirty-five thousand tires, of which about thirty thousand were for wagons and carriages. He had examined hundreds of tests in regard to those thirty thousand tires, and they showed that the average tensile-strain had rather exceeded 44 tons per square inch ; the maximum had been 47 tons, the stretch on the average had been over 20 per cent, in a 2-inch test-piece, and the deflection under a drop-test had averaged 6 inches in a foot of the diameter of the tire, instead of what the Author had called the normal deflection of 2 inches in a foot. He assumed the Author meant, by normal deflection, that commonly specified by many engineers, who were satisfied that they had secured the safety of the travelling public, if imder a drop-test the deflection amounted to 2 inches in every foot diameter of the wheel. That had not been his own practice. As the result of many hundreds of tests he had obtained an average deflection of 6 inches. He had examined fifty-six tests of engine tires made during the same time, and the tensile-strength of those had been fully 46 tons per square inch ; the stretch had been 22 per cent., and the deflection under the drop-test had averaged from 3i to 8 inches in a foot. Fifty-six tires had broken after receiving from five to twenty-one blows, and on the average ten blows of 1 ton weight falling through 30 feet. The tire was placed on a block of cast-iron weighing 5 tons. A fact, not altogether without interest, was that on a colonial railway, which liad a very great length and very uiany

150 DISCUSSION ON THE STRENGTH [Minutes of

curves of 300 feet radius, it was foimd necessary, in order to ensure safety, to have much harder tires. With the softer tires the flanges ground away so rapidly that they became like blunt knives after running a comparatively short distance. Application was accord- ingly made to a w^ell-known firm pre-eminent in the manufacture of hard tires. What had been the result of sending out the objection- able tires referred to by the Author, as those who had read the Paper might, he feared, regard them? He had only known one tire break about seven years ago, during the ten years he had mentioned. That might perhaps be considered only negative evidence. He had, however, the advantage of speaking during the last few days to a locomotive superintendent who had received in India about thirty thousand of the thirty-five thousand tires, who stated that during his experience of thirteen or fourteen years he had not known of a single broken tire. Other engineers had without doubt been eqiially successful, so that he thought the Author's apprehension might to some extent be soothed, and that he need not anticipate, or lead others to anticipate, serious mishaps, and the want of safety. The Author observed that it was accepted as an axiom that from 0"25 to 0"32 per cent, of carbon was the proper proportion for Bessemer steel, and that to get a strain of 42 tons was rather hazardous, and that if that were obtained, and if 0 • 40 per cent, of carbon were put into the steel for tires, they would fail under the tup-test. All the tires of which he had sjioken had been under the tup-test, and with very rare exceptions had not failed. The Author had further stated that the strength of tires should not be increased by adding carbon, and suggested that good might be done by the addition of chromium. He had not a word to say against the addition of chromium, and he knew that those who used it considered that it added to the ductility of steel. But he wished to point out that the Author's limit of 0*40 per cent, of carbon was not the limit of safety. The Author, at p. 122, gave a steel with 0 • 28 per cent, of carbon, 0 • 42 chromium, and 1 "54 manganese, while (p. 129) he gave a steel with 0*50 per cent, of carbon and no chromium. From his own statement something very different in the ductility of the two steels might be expected. But what were the facts ? Both of them bore as near as possible a strain of 60 tons per square inch, and they stretched, one of them 15*0, and the other 14-9 percent.; the stretch represented duc- tility equally with the bending from the tiip-test. He might be permitted to refer to some experiments made for him by Professor Kennedy and Dr. Riley in 1882. Of six specimens two had carbon under 0 • 40 per cent., one With 0 • 425 of carbon had a tensile-strength

Proceedings.] OF BESSEMER-STEEL TIRES. 151

of 43 • 39 tons per square inch, and stretch of 23 per cent., breaking Mr. Berkley, under a tup- fall of 18 feet; the deflection being 4 inches per foot. In the next case the carbon was 0*433, the tensile-strain 45*5 tons, the stretch 22 per cent., breaking at 27 feet fall, the deflection being 5 inches per foot. In the next case the carbon was 0*441, the tensile-strain 44 tons, the stretch 18 per cent., breaking at a fall of 30 feet, with the deflection of 4 inches per foot. The next, carbon 0*577, tensile-strain 45*74 tons, stretch 15 per cent., break- ing under 20 feet fall, and deflection 3^ inches per foot. The main object of the Paper appeared in the statement of the Author's views, as to the test necessary to indicate the fitness of tires for strains that they would subsequently meet. He said, " At the works, let tires be selected from groups made from the same blows, each group being marked with a distinctive stamp." That was the practice of almost all engineers. " Let representative tires from each group be sulijected to exhaustive chemical and mechanical tests. Let the exact mechanical treatment from ingot to finished wheel be faithfully recorded. When the life of one of these tires is

finished let the chemical and the mechanical tests be repeated."

It was a charming idea to go through an experience of years, from the commencement of the manufacture to the end of the life of an article. It was delightful to contemplate such a state of things, but it was not practicable. The Author had naturally suggested that chemical, as well as mechanical, tests should be made of the material. No doubt occasionally an analysis of steel was desirable to give the engineer information as to the chemical composition of a material, which he had perhaps tested many times. But he knew that, to a great extent, chemical tests could not be carried out in practice. His principal objections to them were that they were very costly, and that often they could not be carried out in time. It would be a great inconvenience if manufacturers had to add chemical tests to their other tests ; and even assuming that they were trustworthy, which, however, he could not admit, he did not think that an indirect method of ascertaining the mechanical qualities of a material would be so good as a direct method, by subjecting the material to the kind of work which it would sub- sequently have to undergo. It was an assumption that he should not like to vouch for, that, by the analysis of a piece of steel, it could be foretold what the results would be when the specimen was tried by a mechanical test. He had examined analyses made for himself and others, and he did not hesitate to say, that if a chemical test were specified, no engineer would be able to reject a batch of tires which did not exactly conform to the chemical

152 DISCUSSION ON THE STRENGTH [Minutes of

analysis. As a practical engineer, therefore, he regarded a chemical test as of little value, except as giving some information of the general character of the steel. He would impress upon all present the undesirability of having two tests. They might be incon- sistent with each other, and then they would give much trouble. It was far better to have one test and to stick to it. There was one point on which he did agree with the Author, namely, his recommendation to cool the tires slowly. It was of importance, when a tire was made, to put it in some place, or protect it, in such a way as to ensure its being cooled slowly, so that local strains might not be set up by unequal cooling. If engineers would press upon manufacturers to do that more systematically, they would find a great advantage from it. He might, perhaps, be permitted to enumerate shortly the mechanical tests which he had found to be necessary. In the first place, he limited steel for tires of wagons and carriages to a strain of 42 tons per square inch, and for tires of engines to 44 tons, with a stretch of 20 per cent, in 2 inches. The next test he believed to be a very important one, though it was not generally practised. Every tire after it was cold should be dropped on to a block of cast-iron, weighing at least 2 tons, first in one direction, and then turned through an angle of 90°, and dropped in the other direction. If the steel was too hard so as to be at all dangerous, or if there were any flaws in it, that test would discover the defect. Tires of different diameters shoiild be dropped different heights. With diameters of 3 feet 6 inches they should be dropped o feet ; with diameters from 3 feet 6 inches to 4 feet 6 inches, 4 feet; with diameters from 4 feet 6 inches to 5 feet 6 inches, 3 feet 6 inches. More tires had been broken in that way than had been broken by a tup of 1 ton falling 30 feet. Every blow of the apparatus, or every charge of the furnace, should, in his opinion, be tested. In his own practice, the tires were put under a tup of 1 ton on a block of 5 tons, and the drop was one of 30 feet. He was in the habit of specifying that the tire should receive two such blows, which almost invariably gave a greater deflection than 2 inches per foot of diameter of tire. He also specified that the tires should be broken with the tup. It took on an average about ten blows to break the tire, and the deflection was at least equal to 6 inches in 1 foot of diameter. r. Tickers. Mr. T. E. YiCKERS should have attached more value to the Paper

if it had Iteen prepared with a little more experience. In the first place, more attention should have been given to the size of the sections of the tires testetl, and the tup for testing should have

Proceediugs.] OF BESSEMER-STEEL TIRES. 153

been of the usual weight of 1 ton. He did not think that the Mr. Vickers, Author was justified in finding fault with engineers for specifying a hard material which would wear well. They wanted a material to wear as well as possible, and to have safety combined with it. From many years' experience he held that the falling-weight-test upon a tire was the best that could be applied to it. By adopting uniform sizes, an idea could be obtained both as to the toughness and hardness of the tire ; for tires of a given diameter and section and equal hardness would always be uniform in their deflection. He thought the results at which the Author had arrived were not due to the hardness of the steel, but to its composition. His experience had apparently been derived from steel containing 0*80 per cent, of sul})hur, 0 • 80 per cent, of phosphorus, and 1 • 25 per cent, of man- ganese. Mr. Vickers had no experience of tires of such a chemical composition. His idea of a good tire was one containing from 0 • 50 to 0 • 70 per cent, of manganese, not over 0 • 035 of phosphorus, about half the amount of sulphur mentioned by the Author, and as much as 0 • 60 per cent, of carbon. Such a steel* would stand the dropping- weight-test thoroughly well, even if the carbon reached 0"70 per cent., and if a tire would stand this test it was perfectly safe. It used to be said that this test was absurd, because a tire was never subjected to such treatment in practice, and that the tensile-test was the more suitable one. But there must be some test, and it was not the tire itself, but the material of which it was made that was tested, and for that purpose the falling- weight-test, he thought, was the best. With uniform sections, it would show both the ductility and the hardness ; but as the sections and the diameters varied very much, it was necessary to have the tensile-test in addition, to ascertain the hardness, in order to ensure suflScient wear. Those were mechanical tests which he had adopted for nearly thirty years. The Author had spoken of the molecular changes to which steel of high temper was liable. He did not know whether by that was meant the old idea of crystallizing by vibration, which he had thought was now exploded. Neither iron nor steel, which had small crystals to begin with, could by any process of vibration be transformed into a structure containing larger crystals. The Author, in one instance, had contradicted himself. He had stated (p. 116) : " This is proved by the fact that a test-piece, planed out of an untested tire, gives miich the same result on the testing-machine as a piece planed out of a tire which has been subjected to the falling- weight-test " ; while later on he stated that : " A tire which has been much punished under the falling weight will sometimes give widely diverse results on the machine, such variations de-

154 DISCUSSION ON THE STRENGTH [Miuutce of

pending upon the position from wliicli the test-pieces were taken." If there was no difference between the tire tested and a tire untested, there oiight to be no difference between pieces taken from different parts of a tire which had been subjected to that punishment ; and in practice he had found, from many trials, that tensile-test- pieces, cut from different parts of a tire so tested, did not materially vary. He had not much experience of chromium, except experi- mentally, because he had always been able to get steel of the necessary quality without the use of it. He simjjly relied on not having too much impurity, and adding carbon, which was one of the best constituents of steel to get hardness. Alhision had been made to annealing, and Mr. Berkley had referred to the slow cooling of the tires before testing. Although he agreed with almost everything he had said, he certainly could not agree with him in that. It was never known w^hat a tire would be subjected to after it had gone abroad; it might be re-heated and allowed to cool qiaickly, or it might when being j^ut on the wheel be heated too much and dipped into water, the results of which would be to undo all the effects of an- nealing. An experimental tire containing chromium, and cooled in ashes during four or five days, was reduced in tensile-strength from 60 to 45 tons per square inch. The same tire re-heated, and allowed to cool in the ordinary way in the open air, regained its former tensile- strength, showing that it assumed the original form Avhich it had when it left the rolls, and thus lost the effect of the annealing. In order to ascertain the probable safety of a tire, it ought to be subjected to a test in its most disadvantageous state, namely, its state when it left the rolls. He did not olgect to tires being re-heated, if they had been allowed to become too cold in the rolling- mill, and afterwards cooled in the open air ; but he thought it was a mistake to allow tires to be cooled in ashes, when removed from the rolls, in order to facilitate the testing. He never ditl anything of the kind unless expressly told to do so by the engineer, and even then he always did it with great unwillingness.

Mr. W. Stroudley observed that the Author had advocated the use of soft tires. Mr. Stroudley had used Bessemer tires for wagons, and had worked ten thousand or eleven thousand pairs. He had also used them on heavy goods engines running at low speeds. He had found that the wear of Bessemer-steel tires in the case of the engines, was more rapid than that of the superior quality of steel made by the crucible process, or by the open-hearth system. He could only find a record of one Bessemer-steel tire breaking on a wagon out of the number he had mentioned. He had also examined the records of engines fitted with steel tires, and, in

rroceedings.] OF BESSEMER-STEEL TIRES. 155

100,000,000 miles distance travelled, he had found only one tender Mr. Stroudley.

and one engine tire broken. He had tried the experiment of

locomotives with tires made of soft steel possessing about 36 or

37 tons tensile-strain. The wear of those tires was very rapid ;

they spread out quickly, so that they had to be frequently

turned. He had since adopted tires with a tensile-strain of about

47 tons per square inch. He had removed them after they had

been worn out, and cut sections from them, and had obtained

a tensile-strain of from 47 to 48 tons on a 3-inch specimen. He

therefore thought there could be no advantage in making a tire

soft ; but there was a great advantage in making it of steel that

would sustain a tensile-strain of 50 tons in the case of higher

class steel, and very near that in the case of Bessemer steel. No

accident had resulted from the fractures he had mentioned, and

the tire had not left the wheel in any case. The process used

in jiutting on the tires was to bore out the tire carefully, with

the usual allowance for shrinkage, to place it on the wheel and

let it cool without water, or any other means. As a railway man,

he saw no necessity for meddling with the chemical constituents

of the steel. He attached great importance to the thickness of

the rim of the whesl. He had observed that tires of wheels having

thin rims broke, whereas those of wheels having thicker rims did

not break. He thought that part of the wheel was very often

neglected.

Mr. Alexander McDonnell said he had not intended to speak Mr. McDonnell, but for the remarks made by Mr. Stroudley and Mr. Vickers. He was glad to hear the remark of Mr. Vickers on annealing, which he regarded as exceedingly valuable when juoperly done ; but he thought that it had sometimes done mischief. Where forgings were left to soak, and to lie in hot sand for a long time after being forged, instead of being allowed to cool in a slow uniform way, they had been damaged, and their tensile-strength reduced. Almost all the tires on wheels that he had seen broken of late years had been on very thin rims, or where the rim had been broken. He had observed a considerable number of wheels running with a broken rim, and he had seen more tires broken upon defective wheels of that kind than on any others. The Author woiild have been j^erhaps frightened if he had lived in the time of welding tires. Mr. McDonnell had stopped many hundreds of wagons for broken tires, and great numbers of engine tires were broken when tires used to be welded. That was almost unknown now.

Mr. J. Oliver Arnold, in reply, considered the Paper had been Mr. Arnold, amply justified by the various and often mutually destructive

156 DISCUSSION ON THE STRENGTH [IMinutes of

Ir. Arnold. opinions expressed by the different authorities -who had taken part in the discussion. For example, the Paper advocated a strain of nearly 40 tons per square inch for Bessemer-steel tires ; but the majority of the railway engineers who had spoken advocated a strain of nearly 50 tons, while Mr. Mattieu Williams nailed his colours to the mast at something under 30 tons. Mr. Eeynolds had spoken of a property of steel which he termed " body," and which he defined as a certain kind of homogeneity, which offered great resistance to the disruption of the particles. That property depended upon two factors : the first, suitable chemical composition, and the second, appropriate physical treatment. Mr. Arnold altogether denied the existence of " body," in the sense that it was a mysterious essence to be distilled from certain brands of iron. Messrs. Eeynolds and Cow^ier had both taken strong exception to the hammered bar, mentioned on page 120. He fully admitted that the difference in the treatment of the two steels somewhat vitiated the accuracy of the comparison ; biit he could not admit that it accounted for the reduction in the extension from 15 to 3 per cent. The former was inclined to deny the existence of molecular change, and to substitute for it what he called fatigue ; and he defined fatigue as a separation of particles arising from the imperfect elasticity of the steel. Ko doubt there was a great deal in favour of that view, but he failed to see that the term " molecular change " was not a better expres- sion to apply to it than the meaningless term " fatigue." The tire that bore a strain of 36 tons, referred to by Mr. Eeynolds, fully bore out the statement in the Paper that tires of moderate strain were not subject to alteration from vibration. Again, a doubt had been expressed whether hard tires had a greater wearing capacity than soft tires. He thought that point had been pretty conclusively settled by other engineers who had taken part in the discussion. Mr. Hadfield supj^orted the view that steel, identical in chemical composition, might vary very largely on the machine. That he regarded as an important feature of the discussion, and it was to be hoped that the theory that such varia- tions proceeded from heterogeneous chemical composition was now exploded, and that future research would be carried out on the only lines likely to lead to good results, namely, that of heterogeneous molecular or physical structure. Professor Eoberts-Austen had denied the chemical identity of an annealed and an unannealed steel casting, and had asserted that it was well known that annealing totally changed the state in which the carbon existed. So far, however, from its being well known, it was totally unknown

Proceedings.] OF BESSEMER-STEEL TIRES, 157

to many practical steel-makers, and his own experiments led to an Mr. Arnold. opi)osite conclusion. Professor Roberts-Austen was j^erhaps con- fusing two distinct sets of phenomena. It was well known that, in rolling hard steel, a separation of the particles of graphite sometimes took place, producing what was techically known as black steel. That often occurred in the case of file steel. When the steel was hardened, the whole of the carbon reassumed the combined condition ; and Mr. B. W. Winder had obtained results on annealing which showed that graphite reappeared in exactly the same position as before hardening. But that was very different from the case of a mild steel casting containing 0 • 50 or 0 • 60 per cent, of carbon ; because if that steel, after leaving the mould, was drilled, having been carefully cleared from scale, and a careful colour-comparison was made and repeated after annealing, very little diiference could be observed in the two colour-tests. He was afraid that Professor Roberts-Austen's dream of a metal- lurgical millennium, when the influences of the hardening elements of steel were to be ruled by the periodic law of Newlands and Mendeleef, would never be realized. Mr. Cowper had expressed an opinion that unannealed steel castings were not brittle, and stated that many steel castings were in use. Whether unannealed steel castings were or were not brittle would be best proved by the following experiments. Some two years ago he had to superintend the metallurgical portion of the manufacture of several hundred steel buffer cases. They were cast from crucible steel of the highest quality, and contained perhaps, 0 • 50 or 0 • 60 per cent, of carbon. The cases were tested by placing them base down- wards on an anvil, and striking a heavy blow with a 4-ton steam-hammer. The unannealed cases went to splinters at the first blow, but the annealed cases stood from four to six blows ; and the manner of rupture was totally different. In the latter, a longitudinal crack was developed from the base, and in some cases the bearing was actually crushed into the sjjring chaniber. Messrs. Yickers and Reynolds had both expressed an opinion that the tires, of which analyses had been given, were too impure. The former particularly objected to phosphorus ; but Mr. Aspinall had stated that the Siemens tires which gave the best results had the following analysis: carbon, 0-65 per cent.; silicon, 0 • 28 ; sulphur, 0 • 09 ; jihosphorus, 0 • 06, and manganese, 0-86. That was a direct contradiction to the views as to quality expressed by Messrs. Tickers and Reynolds. Mr. Berkley had altogether mistaken him on one point. He was rather severe on him for stating that the limit of carbon should be 0*32 per cent.

1 58 DISCUSSION ON THE STEENGTH [Minutes of

ilr. Arnold. He was really referring to Bessemer steel, and Mr. Berkley must have missed the foot-note on page 117 referring to that subject, in which it was distinctly stated, that in steels low in manganese there was a much greater latitude for the carhon. Mr. Berkley thought it very curious that the spring steel (p. 129) with 0*50 per cent, of carbon, and 1-10 per cent, of manganese shoiald elongate as much as the tire-steel (p. 122), containing 0*28 per cent, of carbon, 1 • 54 of manganese, and 0 • 42 per cent, of chromium, and argTied that these results proved the inconsistency of the reasoning in the Paper. Here was an example of the use of that chemical analysis which Mr. Berkley valued so little. By its indications, the chemist found in this particular case that 0*40 per cent, of manganese, and 0*40 per cent, of chromium had much the same hardening effect as 0*20 per cent, of carbon; and he thought it inadmissible to build up an elaborate argument on the carbon contents alone, ignoring altogether the influences of the variations in the proportion of manganese and chromium. The same speaker had also said that the assertion that the strength of steel tires was a question affecting the safety of many lives, was not suitable for a scientific Paper. But it was admitted in the Paper that the precautions taken by engineers to ensure the safety of tires were, in the great majority of cases, effectual. The information given by Messrs. Cowper, Aspinall, Berkley and Stroudley, as to the rarity of tire breakages was satisfactory as far as it went ; but it was evident that tires did occasionally break, and should such fractures lead to accidents, it would be very small consolation to the injured to be assured that their case was exceptional, and that so many tires ran so many thousands of miles with a minute percentage of breakages. The question was. Why should there be any breakages ? there must be a cause for them, and it might be removed. The statistics given by Mr. Berkley were very valuable and interesting, biit a similar record of the behaviour of 60-ton tires on English railways would be still more instructive, because on these, conditions in a great measure absent on colonial railways existed, namely, high speed and continuous brakes, and consequently much greater vibration. It would appear from statistics recently })ul)lished, that English were more fortunate than German engineers in the matter of tire fractures, the German record for 1887 having been three thousand five hundred and fifty-two broken tires, and that for the fracture of six hundred and thirty, or 33 i^ per cent, of these, no cause could be assigned. Mr. Berkley's idea that so intricate a problem, as the influence of molecular structure on the strength of steel, should be solved \>y the indications of a single method of

Proceedings.] OF BESSEMER-STEEL TIRES. 159

testing, would not commend itself to those accustomed to research Mr. Arnold, on o])Scure i)henoinena ; because experience has shown that such questions were only satisfactorily answered by experiments made from every point of view, namely, in this case, chemical, physical and microscopical. Mr. Aspinall had referred to the superiority of steel over wrought-iron crank-axles ; and it was highly de- sirable that the use of iron for the manufacture of railway tires and axles should be abolished, with a substitution of steel of moderate tensile-strength, the latter being uniform in chemical comjiosition ; whereas the lines of continiiity in wrought-iron were broken Tip by the layers of admixed slag. Mr. Aspinall as well as Mr. Berkley spoke favourably of the effects of annealing, and it was to be presumed that tires made from steel of the extraor- dinary composition quoted by the former were carefully annealed, otherwise a tire containing 0*60 per cent, of carbon, nearly 0"25 per cent, of silicon, and 1-10 per cent, of manganese, if allowed to cool in the ordinary way after leaving the rolls, would give a strain approaching 60 tons on the machine, and snap under the impact-test, long before the necessary deflection had been obtained. Tlie contriluition of Mr. Lewis to the discussion would be rather unpalatable reading for English steel-makers, indicating as it did the superior skill of continental steel-makers. The analysis of the Belgian steel showed it to contain 0 • 06 per cent, of phosphorus, or nearly twice as much as the ideal tire of Mr. Vickers, and yet it was asserted that the quality' of this steel was so high, that it yielded tests which no English maker dared guarantee. The divergences between the Belgian and English analytical results were, however, so great, that it was difficult to believe that the steels on which the chemical analyses were made were identical. Engineers would do well to give careful attention to Mr. Kirkaldy's significant remarks, and more par- ticularly where he urged that "engineers should be protected from sham experiments." Mr. Vickers depreciated the value of the Paper because of the inexperience exhibited in its preparation, founding this opinion, firstly, on the fact that the tup used in the impact-tests weighed 22 cwt. instead of 1 ton ; and secondly, on the want of uniformity in the sections and diameters of the tires experimented upon. The aj^plication of a little arithmetic would remove any inconvenience arising from the weight of the tup ; and the second objection was wholly imaginary. The diameters and sectional areas of the tires used were uniform throughout. The sentences taken from their contexts (pp. 116 and 127), which Mr. Vickers held to be contradictory, had reference tc steels dif-

160 DISCUSSION OK THE STRENGTH [Minutes of

fering widely in their chemical constitution, and consequently, juechanical properties. It would be just as reasonable to charge with inconsistency any one who stated that a steel containing 1 • 00 per cent, of carbon would harden, whilst a steel containing 0*10 per cent, would not harden. In conclusion he trusted that the anoma- lous and discrepant experiences, brought to light during the discussion, might prove valuable as finger-posts pointing out paths for further research.

Correspondence.

Ir. Aiijileby. Mr. C. J. AppLEBY regretted that the record of investigations had been limited to metals containing chromium, and what seemed to him exceptional proportions of manganese, and that but brief reference (p. 128 J had been made to tempering in oil, and none at all to the usual mode of annealing by lengthened exposure to heat in a closed chamber. He had found that the results obtained by the last named method were irregular and unreliable. As re- gaided oil-tempering, he had made many experiments with open- hearth, Bessemer and crucible steel, and, whether high or low in carbon, he had found that immersion in oil, at temperatures suitable to the character of the metal, invariably imjjroved its quality, and more especially its tenacity. He thought that the following list of tests of oil-tempered Bessemer-steel axles, made in Sheffield, had a direct bearing on the question under con- sideration, because, after all, a quality of steel was sought which would be reliable under the severe conditions which tires and axles must constantly fulfil. The axles referred to were forged from ingots of the usual size, and the tensile-strength of the steel when forged was 26*1 tons per square inch, with an elongation of 28' 6 per cent, in a length of 8 inches, and reduction of area of 60-2 per cent. The tests of the steel, after oil- tempering, gave a tensile-strength of 27-5 tons per square inch, an elongation of 26*87 per cent, in a length of 8 inches, and reduction of area of 58 • 44. The dimensions of the axles were : — length, 7 feet 4 inches ; diameter at wheel-seat, 5 j inches ; diameter at centre, 4^ inches. The first test consisted of twenty blows from a 1-ton tup, with 20 feet free drop on the centre of the axle, the bearings being 3 feet apart. The axle was turned half round after each blow, and the deflections were : — 1st blow, deflection 2| inches ; 2nd, straight; 3rd, 2^ inches; 4th, straight; 5th, 2j inches; 6th, ^ past straight; 7th, 2j^ inches; 8th, j inch past straight; 9th,

Procccdinf^s.] OF BESSEMER-STEEL TIRES. 161

2| inches; lOth, -^ inch past straight; 11th, 2 J inches; 12th, Mr. Appleby. ^ inch past straight; 13th, 2 inches; 14th, ^ inch past straight; 15th, 21 inches; 16th, ^ inch past straight; 17th, 21 inches; 18th, f inch past straight; 19th, 2} inches; 20th, ^ inch past straight. The tests were then continued by giving the tup a fall of 26 feet with the following results : — 21st blow, deflection 2 J inches ; 22nd, I {r inch past straight ; 23rd, 2 j inches ; 24th, 1 inch past straight; 25th, 2-^ inches; 26th, 11 inch past straight; 27th, 2V inches; 28th, 1} inch past straight ; 29th, 2 J inches; 30th, 1 g^ inch past straight, thus showing gradual fatigue ; but there was no fracture, and nothing to indicate that these severe tests had caused permanent injury to the steel. The bending-tests were made on pieces 1^ inch square by 8 inches long, machined out of the axles before and after oil-tempering, and were bent double when cold. The untempered steel showed signs of tearing in the bend, whilst the tempered pieces gave no indication of this kind. There was a wide difference in the fractures, the un- tempered steel being open and crystalline, whilst the tempered specimens were much closer in the grain, with an entire absence of crystals, and had a silky appearance, the colour being darker than in the untempered steel. As might have been expected, steel, higher in carbon, which had a tensile-strength of 35 tons per square inch before treatment, increased in strength in a higher ratio than the low-carbon steel above referred to. The increase in tensile- strength after oil-tempering was 3 to 4 tons per square inch, but there was no material alteration in the elongation nor in reduction of area. The well-known fact, mentioned by the Author, that steel low in carbon did not harden, might have gone far towards creating the widely-held impression that Bessemer steel would not harden. This was a fallacy; probably but few engineers were aware how large a quantity of Bessemer steel was daily worked up into springs, and even cutting tools of all sorts. Bessemer steel had not quite outlived its character for uncertainty in quality ; but his experience showed that steel, singiilarly even in quality, was produced in those works where analyses of the pig-iron and spiegeleisen were systematically made before they were charged into the cupola, and of the ingot when cast, due care being exercised in manufacture by using pure coke, &c. It would be remembered that the effect of im:uersing heated steel in heated oil was referred to at considerable length in the discussion on General Maitland's Paper on "The Treatment of Gun-Steel."*

' Minutes of Proceedings Inst. C.E., vol. Ixxxix. p. 114. [the INST. C.E. VOL. XCV.] M

162 CORRESPONDENCE OX THE STRENGTH [Minutes of

Mr. AjiiiloLy. In nearly all cases it was spoken of as " oil-hardening," perhaps " tempering " would be a more accurate expression ; and it seemed desirable that the difference between " hardening," which increased the tensile-strength and largely diminished the ductility, and " tempering," which added greatly to the tenacity of the metal without materially reducing its dixctility, should be clearly recog- nized. It seemed to him extremely doubtful whether exceptionally high tensile-strength gave to tires and such like things the qualities attributed to it. But even if high tensile-strain should be con- sidered essential, he believed a judicious use of oil-tempering would give results more reliable than any Avhich could be obtained by the use of chromium, and the high percentage of manganese men - tioned by the Author ; because the metal, being purer, would be less liable to deterioration during manufacture, and to failures due to change in molecular structure, than the less pure steel not oil- tempered.

Mr. Brustlein. Mr. H. A. Brustlein observed that the metal experimented on by the Author was really a manganese-steel with a small per- centage of chromium ; whilst in his communication to the Iron and Steel Institute,^ he referred exclusively to a carbon and chrome- steel compared with carbon-steel alone, and his later practice had confirmed the statement he had made on this subject. In his opinion he could hardly admit that the diSerence between the results shown in the first and third tables (p. 123) could be attri- buted to the effect of the blows of the falling weight. He rather attributed it to other circumstances which it would take too long to explain. At the Unieux Works he had never recorded a like observation under a blow, although since tool steel was especially made there, steels containing so much sulphur, phosphorus and manganese, and so little carbon, were never worked. As the Author too observed, chrome-steels were verj'' sensible to annealing and tempering ; but the readiness with which they could be put in one state or the other, might be turned to advantage. Moreover chromium had a great superiority over manganese, in that il behaved well when associated in large proi:)ortions with carbon ; whilst manganese alloyed with carbon, except in the very high percentage adopted by Mr. Hadfield, rendered the steel brittle and untrustworthy. The influence of the length of the test-pieces on the results of tensile-tests had been carefully studied by Mr. J. Barba of Creusot, who had communicated a Paper on the subject to the Society of Civil Engineers of Paris. ^

> The Journal of tlik Iron and Steel Institute, 1886, p. 770. ^ Me'iuoircB de la Socie'tc' dcs Inge'nieurs Civils, ISSO, p. G93.

Proceedings.] OF BESSEMER-STEEL TIRES. 163

Mr. F. W. Harbokd stated that his experience in the manufacture Mr. Ilarbord. of mild steel had led him to the same conclusion as the Author, namely, that high tensile-strains in steel could only be obtained by increasing the proportion of metalloids, and this was always at the expense of ductility and general reliability of the material, especially when it was subjected to continued vibration. He had also found that chromium was a most useful element where high tensile-strength was required ; and that within moderate limits it could be employed without diminishing the ductility, or otherwise deteriorating the material.

Mr. J. W. King observed that since the experiments quoted by Mr. King, the Author had been made, now several years ago, there had been a great advance by steel-makers in general in controlling the state of crystallization of finished pieces of steel, such as axles, tires, guns, &c., so much so that tire-steels could be made with great regularity to register, after falling-weight-tests, strains as high as from 45 to 50 tons per square inch, at the same time giving an elongation and reduction of area equal to steel 20 tons lower in tensile-strength.

Sir Alexander Eendel, K.C.I.E., could give no accurate infor- Sir A. Readel. mation on the results of the present testing of steel tires, as he got none from India. But there were few failures here in testing, and he had no complaints, so far, of failures in Indii, except in two cases where tires had been found broken on wheels just arrived in India, the result evidently of their having been put on too tight. He had raised his tests to their present standard, because well- known makers of tires which had a high reputation, and which were made to very high tests, wanted him to specify tlieir make. This he declined to do, but specified their tests. Since then he had experienced no difficulty in getting tires capable of sustaining these tests from other manufactures, quite as well as fiom the makers to whom he referred. He only specified tires i.y liiese high tests for locomotives and tenders. They were made by the Siemens-Martin process. He had been obliged to give up steel for carriage and wagon axles altogether. The following was au analysis of a steel tire made at Leeds by the Siemens-Martin process. The analysis had been recently carried out by Mr. Arnold Philip, one of the chemists employed at the Eoyal Indian Engineering College, Cooper's Hill. There was no chromium in the sample. The tensile-strain obtained with this tire was 45-3 tons per square inch. The contraction of tested area at tlie point of fracture was 21 per cent. The extension on a length of 6 inches was 15-4 per cent. A weight of 20 cwt. was allowed to

M 2

164 CORRESPONDENCE ON THE STRENGTH [Minntes of

Sir A. Rendel. fall seven times from a height of 12 feet, until the tire deflected ^ of its diameter (5 feet 6 inches external^, but it showed no sign of fracture : —

Per cent.

Carbon (by combustion) 0

Silicon 0

Phosphorus 0

Manganese 0

Sulphur 0

Copper 0

Tungsten 0

Titanium 0

Chromium 0

Arsenic 0

Iron (by difference) 98

100

Carbon (by colour-test) 0 • 575

527 148 052 492 101 050 000 000 000 018 612

000

Mr. W. SowERBY observed that the experiments recorded in the Paper -were hardly on the exact lines as tests of the ability of tires to endure the treatment to which they were subjected when in use ; for, except when there was collision, they seldom or never had to undergo such violent concussions as were indicated in the experiments; nor were they ever subject to any tensile-strain. But they had to undergo and endure constant and continuous impact on the rails over which they travelled, and the amount of such impact depended greatly upon the velocity of the trains, and the gradients of the railways over which they passed. A train, especially when at high speed, had a tendency to follow the parabolic line of a projectile ; and, as it descended the gradient, the foreparts of the wheels were very slightly raised above the rails, and there was a continuous fall or thump. In ascending a gradient this thumping was behind the wheel, and so distinct was this thumping on steepish gradients, that anyone with a quick ear could easily detect whether the train was going down or up an incline ; and it was this continuous thumping to which railway wheel-tires were subjected. If the material of which the wheels were made was good metal, then this continued impact would have a tendency to improve, toughen, and strengthen the tire ; just as old horseshoe nails were said to be toughened by being long worn, and thus subjected to similar impact. It was this kind of continued impact or hammering that a Persian or Indian sword- blade underwent to make it perfect. If the steel from which the tires were made was not good, then the constant impact would

Proceediugs.] OF BESSEMER-STEEL TIRES. 165

have a tendency to crystallize and destroy the strength of the Mr. Sowerby. metal, just as the constant blows of a hammer would destroy and crystallize the best bar of iron if suspended vertically or hori- zontally. The gradients of a railway having such an eiFect, the natural conclusion was that they should be made to suit the cir- cumstances. Italian engineers, he believed, had made some lines with parabolic gradients ; and more than forty years ago he graded the sections of some lines he was then engaged upon, one being the London District of the Great Northern Eailway. He found that such a system of grading would have had several advantages, amongst others that of saving 25 per cent, in the earth- work ; but it would be troublesome to lay down such gradients, and he was somewhat surprised to find the idea carried out literally on the Switchback Eailway at the American Exhibition. Whether the Americans had adopted parabolic gradients on their railways he was unable to say, but if they had, they had done wisely. Much, doubtless, had been learnt from the primitive workers in iron in India and in Spain, especially in the use of manganese and chrome ores for making steel, and also in the mode of tempering steel by cooling it in oil ; and there was still much that might be learned from them in the way of preparing and mixing ores, fuel and fluxes, and subsequent manipulation. It was by constant and continued hammering or impact that the finest Persian, Damascus, and Toledo blades were made, and they were tempered and marked by waving or swinging them in the air whilst cooling. Even the mode of making great steel guns by coils was not new to the Indian " Lohars " ; for many years ago he saw just such a gun lying at the hill fort of Teree in the Himalayas ; it was about 12 feet long, 5 to 6 inches bore, and it was built ujjon a spiral form exactly on the same principle as the Armstrong guns, and that gun must have been one hundred and fifty, or two hundred years old.

Mr. Bartlett W, Winder observed that his work had been more Mr. Winder, especially confined to the higher grades of tool steel, and here similar phenomena were met with, particularly and strikingly in high-carbon steels. A portion of carbon, which in the ingot or hammered bar had existed as combined carbon, assuming during the rolling the free or graphite state ; the fracture, which should have been of a close silky white colour, appeared as a coarse black fracture, owing to the presence of the free carbon. The free carbon, on hardening, disappeared or took another form, but did not again return to its previous combined state, as on annealing it again appeared in exactly the same position. What state the carbon assumed on hardening could not, up to the present, be explained

166 CORRESPONDENCE ON BESSEMER-STEEL TIRES. [Minutes of

Mr. Wiuiler. even after the most careful analysis. How far this molecular alteration would affect the after-working of the steel, he was not prepared to say. In many other kinds of steel, equally marked molecular change occurred, and very generally in irons. Fibrous iron rapidly assumed the granular state. He fully endorsed the Author's ojiinion, that until maker, consumer, chemist, and micro- scopist united, this most vital question in the steel trade — the change appearing so unexpectedly, so powerful in its action, and about which so little was known — would remain unsolved.

11 December, 1888.

Sir GEOEGE B. BRUCE, President, in the Chair.

The discussion on the Paper by Mr. John Oliver Arnold, "On the Influence of Chemical Composition on the Strength of Bessemer- Steel Tires," occupied the entire evening.

Pi'0ceediug3.] ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 1 G7

18 December, 1888.

Sir GEORGE B. BRUCE, President, in the Chair.

(Paper No. 2138.)

"The Fric:ion of Locomotive Slide-Valves." By John x\udley Frederick Aspinall, M. Inst. C.E.

The attention of the Author has been drawn to the fact that few, if any, trustworthy data exist of the friction of slide-valves under steam. Hence he was led to make experiments on locomotives, with a view of ascertaining whether the amount of slide-valve friction was as great as it is commonly assumed to be.

The only previous investigations of the friction of slide-valves with which the Author is acquainted are as follow: — In 1866, a Paper by Mr. Thomas Adams, " On the Friction of the Slide- Valve and its Appendages," read at the Society of Engineers.^ In 1871, a Paper by Mr. W. G. Beattie, " Description of a Balanced Slide- Valve for Locomotive-Engines," was read at a meeting of the Institution of Mechanical Engineers.^

Both these Papers were intended to show the sujieriority of balanced valves over the ordinary valve. But the Author is not aware that either form of balanced valve is now in use, as the mechanical difficulty of keeping them in order outweighs their other advantages. Mr. Beattie states that a valve 10^ inches long and 17 inches wide, with a steam-chest pressure of 125 lbs. per square inch, requires 6,160 lbs. to start it into motion. Mr. Adams's Paper contains very extraordinary statements as to the friction of valves. For example, that a force of 9,752 lbs. is required to maintain the motion of a valve 18;^ inches by 9 J inches, with a pressure on the back of 160 lbs. per square inch.

It seemed to the Author desirable that, instead of attempting to deduce the friction of the valve from the difference of the steam- chest- and cylinder-pressures, the valve should be made, by suitable mechanism, to describe a diagram giving the exact force required to

' Transactions of the Society of Engineers for 1866, pp. 6-21.

^ Institution of Mechanical Engineers, rroccediugs, 1871, pp. 35-40.

168 ASPINALL ON FKICTION OF LOCOMOTIVE SLIDE-VALVES. [Miuutes of

move it at each poiut of the stroke. An apparatus was constructed, as shown in Plate 2, Fig. 1, consisting of a cylinder and piston, packed with cup leathers, which could be used as a pulling link, replacing the ordinary pulling link which gives motion to the valve. A steam-engine indicator was screwed on to a nipple at one end of this cylinder, for recording the pressures in the cylinder during the stroke of the valve. On the other end of the cylinder, an air-valve was screwed on to a corresponding nipple. The pistons, both of the indicator cylinder and of the air-valve cylinder, were fitted with cup leathers, to prevent the escape of the fluid when the apparatus was working. The hydraulic cylinder with attached indicator and air-valve is shown on a larger scale in Plate 2, Fig. 4. A second indicator was fixed on the front cover of the steam-chest, its barrel receiving motion from the valve-spindle. Thus, two diagrams are taken simultaneously, one from the hydraulic-pressure cylinder which is driving the slide-valve, and the other from the steam-chest. These give the force required to move the valve, and the corresponding pressure on the back of the valve. The steam-chest pressure-line has been plotted on the hydraulic-cylinder diagram in each case.

The hydraulic cylinder was filled at both ends with oil, from which air was expelled. In some of the earlier experiments, air in the cylinder was found to disturb the diagrams so as to render them unintelligible. The indicator cylinder below its piston, and the air- valve below its piston, were also both carefully filled with oil before attachment to the hydraulic cylinder. The cord from the indicator barrel was attached to the back end of the cylinder, so that, as the indicator moved with the valve, the length of diagram was equal to the stroke of the valve. The air-valve on one end of the hydraulic cylinder ensures that the pressure at that end is always simply atmospheric pressure; consequently the pressure in the other end of the hydraulic cylinder, which is recorded by the indicator, is that which balances the force required to move the valve. Without an air- valve it would be necessary to have an indicator on both ends of the hydraulic cylinder, and to take the difi"erence of the pressures recorded.

In the diagTams, the return-line comes a little below the atmo- 6j)heric line, but the pressures have been measured to the atmospheric line only. Theoretically, during the return-stroke there should be a negative pressure in the indicator end of the hydraulic cylinder. For, when the piston is moving away from the indicator, the air- valve closes, and the oil in that end of the cylinder is slightly compressed. Consequently a vacuum must be produced in the

Procccdiugs.] A8PINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 169

indicator end of the cylinder, if its piston did not allow the passage of air downwards. An accurate Bourdon gauge was attached to the indicator end of the hydraulic cylinder, and the air-valve replaced by a solid plug screwed on. In this condition, with a calculated pull on the valve-spindle of 1,482 lbs., the gauge showed a vacuum of 5^ inches, corresponding to a pressure on the hydraulic piston of 40 lbs. If the tube of the gauge could have been completely filled with oil, the vacuum would no doubt have been greater.

It will be seen that, when the diagrams are being taken, there is a continuous very small movement, backwards and forwards, of the hydraulic-cylinder piston. As the indicator piston rises in obedience to the pressure, the hydraulic-cylinder piston must move also, the volumes described by both being equal. The diameter of the indicator piston was 0-75 inch; that of the hydraulic piston, 6 inches. Hence for a movement of j inch of the indicator piston, that of the hydraulic-cylinder piston must be j X 0 • 75'-/ 6'-^ = 0 • 0039 inch.

Determination of the Friction of the Apparatus. — The recorded pressures are affected by the friction of the hydraulic-cylinder piston, that of the sjjindle of the hydraulic-cylinder piston, of the friction of the indicator piston and of the air-valve piston. Pre- liminary experiments were made, with the arrangement shown in Fig. 5, to determine the amount of the friction of the apparatus. This was filled with oil and arranged in working order. One end of the spindle was fixed, the other attached to a bell-crank lever, weighted on the long arm. The difference between the pressure in the hydraulic cylinder, calculated from the weights and that registered by the indicator, was taken to be equal to the total friction of the apparatus.

Tai	!LE I. — Experiments on the Fkiction of tlio Apparatus.
Weight on Lever in lbs.	Pressure in Cylinder calculated from Weight.	Pull on Valve-Spindle calculated from Weight.	Pressure shown by Indicator.	Pull on Valve- Spindle shown by Indicator.	Difference of Actual and Indicated Pull on Valve-Spindle.
76 132 188 244 300 356 ■ 412	Lbs. per Square Inch. 11-9 20-8 29-5 38-3 47-1 55-9 64-7	Lbs. 316-5 549-6 782-8 1,016-0 1,249-2 1,482-4 1,715-6	Lbs. per Square Inch. 5-0 12-0 18-0 26-0 33-0 400 48-5	Lbs. 132-5 318-1 477-2 689-3 874-8 1,060-4 1,285-7	Lbs. 184-0 231-5 305-6 326-7 374-4 422-0 429-9

170 ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Assuming that the pressures shown in cohimn 4 are a linear function of those shown in column 2, the Author obtains the following expression. Let L be the pressure shown by the indicator, and L^ the real pressure in the hydraulic cylinder. Then

Lj = a + feL,

w^here a and h are constants to be determined from the experiments. The following values prove to be suitable, and are used in reducing the indicator-diagrams of the hydraulic cylinder : —

Li = 5-61 + 1-26 L.

It may be stated that the friction of the valve-s2:)indle in the valve-chest stuffing-box is inappreciably small. A diagram taken with no steam-pressure on the valve shows only a slight thickening of the atmospheric line, and the valve-spindle, when disconnected from the pulling link, could be easily moved by hand. Many preliminary trials were made with the apparatus to get it into perfect working order, and to eliminate the small causes of error always met with in new apjjaratus.

Method of dealing with the Diagrams obtained. — Diagrams have been taken from two classes of engines, with valves of three different forms, in both good and bad condition. A summary of the results is given in Table II. In all the diagrams from the pulling-end of the hydraulic cylinder the jiressures are lower than in those from the pushing-end. The reason is that, as the valve-spindle works in a dummy gland on the front of the steam-chest, the steam-pressure on the end of the sjiindle is always acting in one direction. This has been allowed for in calculating the diagTams. Sets of diagrams were taken from one engine successively on the same day. These agree closely both as regards form and height. On the other hand different sets of diagrams, taken on different days and under dissimilar conditions, often varied considerably. Either the cylinders differed, or the valves, or the method of lubrication, or the speed at which the engine was running. Besides this, there must have been other causes of variation less easy to give an account of, as, for instance, whether the steam was dry or wet, or whether the boiler was priming.

The average pressure shown by the diagrams has been found by the usual method employed for indicator diagrams. Although theoretically this is not quite accurate for diagrams of the form of the hydraulic-cylinder diagrams, it is sufficiently so for the purpose.

i

Proceedings.] A8PINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 171

	O -7-	O i !-<	con- nder ired ; tion. soft. and very n.
	'^ o	t£ a o
	-^1
	'^^	O . _ — O	3^ ^22 =o.9'-S o.2oo~;=: ^:So
a!	a > a	•« - ., "^ a S o'^ - - 0.2 o
	o P o	> a > -w o
	m _ _	>^ >_^___>_^^>___ «_^_ 1
	Fl	fl a	'T^ ;i > .9 TJK?
9 "3	-3 "^.S		P2 • or^O^OgO s| -T ^.9 «.9-T.9
	■^3 p-iM	S -!;»
	6 2	PI CO
	— - — . —	' —~ V — '	^ V ' -' V '
		-t-Jll . a>-i- CQ.CQ.-*->'l	^ ^^TT^, — ~ :;rT"~; — ■
	a§| • ^^ • ^ -^ -ass =^^-2 • ao • , -'^ '^-^^	a o a ■ a o a ■ 2 2 3- g 2 § • S i^ § s «
	-§ fl t a • o .fc. 9 ci ■
	>^ ^ +: -::; sr	-Eaa 'z.St
a	o tc 3 =° =^	73 •'- . S .t^ . ^ . O . to — '	a) — . . » — ■ .
a 1	^ '^ -►- " o	^1.MM. 1.^1.1	^.9 . '^.9 .
3	« g^ a ^	IIS • ^3 • ZsBs^s'z	_o -2.-- 3 -^-^ *^ ^ a • ^ S a •
	-w ^ O aQ o	o .2 S a ® (D a ,^ a o .;:i 3	0.2 § 2-2 c3
	a 2-2 S-5	g a 2 -5^ EP'" ^ .fl-aiaag-::	. a^^-s. a aj-^
	^ ^_ ^	fi^ Q S cc «	« «
M 1	ooO	-.--.s--,----'^
	.a,t< 5 S S
|£3	>-:;•-(	'"'
lest res- ire.	.00 05 t^ 00	00 OOO lO O lO O O O OO O CO CO o	OOOSt^O rH00CO(N
.So o Oi lo	^ O rt C3 CO rt O O lO l> >n CO (M IM oo	CO ■» a ~t< cgcosooo
|"f^-	►H (M (M eg -M	(MC^ICO^CSCilMCOO-MOS^I COCO-f	fococgco Mcoog-^
	O CO lO 05	1> CO (M C; O O lO C5 C<l O i-H S5 05 30 M	osc-cqi^ loortjo
m 4< "O f- fil	OlO-tl<005C<ICO-t<<M0200 t^lOt^	lO^COOO (Mt^-MCO
rt'S «	.o c3i QC OS OS	ooor^coi^s^r-ico-t<oo30D lOM—i	CO t^ i^ o oi CO Oi r5
^^l-	i.:]l> GO c» cs	iOCOC^IOCO'tHC^COiOOrtiO CO-*il^	t^ CO CO O i-H O 01 03
1-1 rt S<I C<l rt r^ CO C^ --< eg rt i-T i-H -H ^	,_r ^ rH i-T i-<^ M
olu- ank ■per ute.	CO - - -	...,go;:S CO ...o .-
S.25-^1	lO - - -	- - - - S^i^ ^ - - . 00 - "
w'-s-^s		. • .
bo bb	bca ^a fl.3 a. 3	fcC ^ be g^ b£ ^C tD %^ .9-s .9 •? .9 -5 .9 , . . . ..•§	bD ^ bD ^C .^ be .„ ^_3.3_a .s _a .3 _a
	•^^■^^
— o Ja
, — > a^ , — > Ul
S 1	fl 3 a a	s :s :i s :2 s :3 a	a a a a a a a
PhPhPhPh	Ph P4 Ph fL, Ph Ph Ph Ph	p^pLipH PuiPhP^p^
!■">	.a -*l ^nm -	rtw ^H* -Hl-J* -. ..H-l*— txH*— CO H* ^—m	H*..-- -.„_
£ 0-3	OCO -<N -	CO 'N -eo -<N -coiMiOcg eo 'og	CO--- -.»„
M >	a
	r^ . . .	rcJ'*-'^''''^"'^*'^'*	^3
	f-i	»-l »-> ^ »H ^	»H
	^ 1	d rrt eS cS c3 03	03
	&
	;h c^	^ OS !-l >.! ^4 ^H	»H
c2 =g =
^ =2	^ t2 'B "a'-^'a^a^ a
	a ^	a a <v aoaofl o	a
	P^ H*	li, H* fe O fMOf^Qf=H O	P^
2 «	Tt* 00 !M CO	•^ GO eg CO o -tH t> -H lo 05 J5 C; O -+I 00 CDCOl>t>O000COO5O3O52S <^5g<^'	oocgcoo (McoO'*
1.^	1— 1 I— 1	Tt<»Oir5CO COCOTj<r*<
O ~ 2 2
		""•*" "'^"-^
^ s	rt "rt c; M	-H IC O; ^5 t^ -^ O O) -N CO O -H t^ — 1 lO	lOOiCOt^ Ol^^t^-H
	cococoi^i^ooaoaoojoioo rnc^icg	•r)t -rr m fi CgCOCO"*!

172 ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Miuutes of

Samples of the diagrams and calculations selected at random are given in Plates 3 and 4. For instance, taking Diagram 17, Plate 3, Fig. 6, the average pressure per square inch from the hydraulic- cylinder diagram is 27*25 lbs. By the formula put forward, this, corrected for friction, gives as the true pressure, 5'614- (1-26 X 27 • 25) = 39-94 lbs. per square inch. The area of the back end of the hydraulic piston was 26*51 square inches. Hence the total pull was 26*51 x 39*94= 1,058*8 lbs. To this must be added 254*6 lbs., the steam pressure on the end of the valve- spindle, making a total force to move the valve of 1,330*1 lbs. For diagrams taken on the other end of the hydraulic cylinder, the valve-spindle pressure must be subtracted.

In Table II the sixth column gives the average force acting on the valve-spindle, from four diagrams in each case taken in quick succession, calculated precisely in the way described. The Table thus contains the results of one hundred and seven separate experiments. In almost every case, the force calculated from the pushing-diagTams exceeds that of the pulling-diagrams by an amount varying from 4 to 10 per cent. The valve-spindle, pulling link, and eccentric rods, form a long jointed strut, which, when the valve is being j^ushed, will deflect to an extent depending on the slackness of the joints, and this will cause a side pressure on the back steam-chest gland, which will be absent when the valve is pulled. It is difficult, however, to see that the friction due to this side pressure could produce even as much as 4 per cent, additional force to move the valve.

Discussion of the Eesults.

Diagrams, 1 to 16 and 61 to 107, were taken from a goods engine with cylinders 18 inches in diameter and 24 inches stroke, with six coupled wheels, 5 feet in diameter, and 4-inch steam-pipe. Diagrams, 17 to 60, were taken from an express engine, with cylinders 17 inches in diameter and 22 inches stroke, with four coupled wheels, 6 feet 6 inches in diameter, and 3 2-inch steam-pipe. In Table II are given the steam-chest pressures, the kind of lubrication of the valves, the descrijition of the valves, and notes on the condition of the valves and valve- chest faces.

In some cases, indicator-diagrams were taken from the cylinder simultaneously with those from the pulling-link and the steam- chest. Three of the cylinder diagrams are given on Plate 3,

Proceedinf^s.] ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 173

Figs. 6, 8, and 10, plotted above the corresponding pnlling-link diagrams. Of these, diagram 1 7 is a pulling diagram in full gear ; diagram 21 is a pushing-diagram in full gear; diagram 25 is a pushing-diagram with the valve notched up to within 25 per cent, of the centre.

It will be observed in the diagrams that the valve-resistance is not uniform during the stroke, the variation being greatest with the Allen valve. Table II shows that the valve-resistance is greater with a short stroke than with a long stroke, and that to an amount greater than is accounted for by the increase of pressure when the valve has a short stroke.

Among the causes of variation of the valve-resistance, during the stroke, may be mentioned : —

1. The variation of pressure on the back of the valve due to variation of the steam-chest pressure during each stroke. This in some of the full-gear diagrams amounts to 10 or 12 lbs. per square inch.

2. The variation of pressure on the face of the valve, over the area corresponding to the steam-ports. This can be determined from the indicator-diagrams, when these were taken, as in the case of No. 17.

3. The variation of pressure on the exhaust-area of the valve. If this is assumed to be the same as that in the cylinder, it can also be determined from the indicator-diagrams.

4. The form of the valve.

5. The inertia of the mass between the hydraulic piston and the valve.

Variation of pressure on the back of the valve necessarily involves variation of the friction. This is well seen in diagram 17, Fig. 6, in the depression at each end of the steam-chest pressure- line X Y, amounting to about 10 lbs. per square inch. This fall of pressure is due to the opening of the steam-ports, and is absent in the short-stroke diagrams where less steam is drawn off.

To examine the variation of pressure on the face of the valve corresponding to the steam-ports, consider diagram 17, which is a pulling-diagram ; that is, the valve is at the end of its travel inwards when the diagram begins at the right-hand side. Then, at the beginning of the stroke, the back steam-port is full open and the front port is open to exhaust. At the completion of the stroke, the front port is open to steam and the back port to exhaust. The changes of pressure in the cylinder during the stroke are shown on the indicator-diagram, the line A B C D E F corresponding to the back end of the cylinder, and 5 4 3 2 1 to the front end.

174 ASPrSTALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [IVrmutes of

These steam-pressure lines are transferred to the valve-diagram, where they necessarily ajjpear much distorted, because the phase of motion of the valve differs from that of the piston by the angle of advance. By marking on the valve-diagram the positions cor- responding to the points ABODE, 54321, in the travel of the piston, the pressures for those points can be transferred. Two lines on the valve-diagrams are thus obtained, representing the simultaneous steam-pressures in the two ends of the cylinder for every position of the valve. Above these a short dot-line has been drawn, showing the sum of the pressures at each point. This line, which may be called the line of relieving-pressure on the valve- face, gives the total pressure measured over the area of one port opening (in this case 13^ inches by 13 inch) tending to counteract the steam-pressure on the back of the valve. This line shows a drop of about 40 lbs. per square inch, which, reckoned on the area of a steam-port, gives a relief of pressure of 742 lbs. Assuming a coefficient of friction, deduced later on, of 0*068, the variation of effort on the valve-spindle will be 50-5 lbs., equivalent to a rise of about 1 • 9 lb. in the hydraulic-cylinder diagram. The small peak at p in diagram 17, which also occurs more or less markedly in all the full-stroke diagrams at a point corresponding to the opening of the steam-port, appears to be due to a vacuum, formed under the valve-face by the sudden rush of steam into the cylinder. The diminution of jDressure under the valve causes an increase of friction, \yhen the valve is working near the centre of the link, the compression is much greater, and the rush of steam into the cylinder when the port opens is much less rapid. Hence, in the short-stroke diagrams, an increase of valve-resistance at this point is not shown.

There is no inside lap on the Allen valve, and the distance between the inside edges of the steam-ports in the valve is equal to the distance between the outside edges of the cylinder- ports, the width of the bar in the valve being equal to the width of the steam-ports. When the valve begins to open on one side to exhaust, the valve-passage is already filled with steam at the exhaust-pressure, which is suddenly released into the other end ■of the cylinder where compression is just beginning. The effect is to raise a little the compression-line of the indicator-diagram {at 2-3 in diagram 17), and at the same time to diminish the relieving-pressure on the valve-face. Calculation for diagram 17 shows that there is a diminution of relieving-pressure of about S3 ll)s. per square inch on 10*1 square inches of valve-face, or altogether 939 lbs. This, \vith a coefficient of friction of 0"068,

Proceedings.] ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 175

gives an increase of valve-resistance corresponding to a rise of about 2-4 lbs. in the hydra \ilic-cylinder diagram.

The third cause of variation of valve-resistance is change of the pressure acting on the exhaust area of the valve. It is doubtful if the exhaust pressure, shown on the indicator-diagram, is quite the same as that under the valve, at least when the valve is just opening or nearly closed. But taking the indicator pressure as approximately the same as the pressure under the valve, then, when exhaust opens, an area of 13 V X 6^ = 93 square inches is suddenly exposed to the pressure shown at the point D, or say 90 lbs. per square inch. This gives a total relief of pressure of 8,370 lbs. With a coefficient of friction of 0"068, this would cause a diminu- tion of the valve-resistance amounting to 570 lbs. This diminution would continue, though decreasing in amount, while the valve travelled from D to E on the lower Fig. in diagram 17. The diagram shows, however, not a diminution, but an increase of valve- resistance, and this is due to the fourth cause of variation of the valve-resistance. Diagram 17 was taken with an Allen valve of the form shown in Plate 2, Fig. 3. All full-stroke diagrams, taken with this f(jrm of valve, show a more or less sudden increase of valve-resistance at this part of the stroke. With ordinary valves there is a less marked increase. The explanation may be that the steam, rushing through a narrow opening, strikes the further side of the exhaust-sjKice, making the pressure much greater against that side than against the near side. This not only neutralizes the eftect of the increase of relieving pressure, but even increases the valve-resistance.

The last cause of variation of the valve-resistance is the inertia of the moving parts driven by the pulling link. That inertia will increase the effort necessary to move the valve during the first half of the stroke, and diminish it during the second half.

Suppose a mass of weight W, moved by a uniformly rotating crank of radius r, the crank-pin having the velocity V. Then the resistance due to inertia at the beginning of the stroke is —

Wt;2

ijr

Now in this case the valve has a travel of 3]^ inches, so that its motion is practically the same as if it were driven by a crank of radius 1-626 inch, or 0-1354 foot. Taking the sjieed of the engine as 20 miles per hour, or 29-3 feet per second, and the

176 ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

driving-wheels as 78 inches diameter, the velocity of the crank-pin

driving the valve is

31 29-3 X — = 1-22 foot per second. 78 ^

Consequently the resistance due to inertia at the beginning of the stroke is, for a weight of 155 lbs.

155 X 1-22'^ 32-2 X 0-1354

62-8 lbs.

This would correspond to about 2 lbs. on the hydraulic-cylinder diagram, a quantity to be added to the frictional resistance of the valve at the beginning of the stroke, and deducted from it at the end. At higher speeds, the eifect of the inertia of the valve will be much more marked. Thus, in diagrams 77 to 91, at speeds of 168 to 224 revolutions per minute, the resistance due to inertia at the beo'inning of the stroke of the valve will be from 203 to 361 lbs., corresponding to from 7*7 to 13*7 lbs. on the hydraulic-cylinder diagrams.

Valve-Friction. — To determine the valve-friction it is probably more accurate to take, not the mean resistance during the travel of the valve, but the resistance at mid-stroke where the inertia resist- ance vanishes. The valve-chest pressure is taken at the same point. Table III has been calculated in this way.

Table III. — Pressure on Yalve and Yalve-Eesistance at Mid-stroke.

	Starting	Friction	Dimensions of	Pressure in
E.xperimenter.	Friction.	of Motion.	Valve.	Steam-chest.	Remarks.
	Lbs.	Lbs.	Inches.	Ll)8. per Square Inch.
Adams		9,752	18J X 9^	160
Bcattie	6,160	4,620	17 X lOJ	125
Halpin	2,629		lOf X 9^	80
Aspiuall .		1,321	161 X 10	134	(■Brass Allen valve ; ( full stroke.
)>		1,096	16J X 10	,na (Brass plain valve ; |\ full stroke.
„ . •		982	16J X 10	127	(Cast-iron valve; \ full stroke.

The average of the eight diagrams, 17 to 24, gives for the valve- resistance at mid-stroke 1,321 lbs. Assuming the whole area of the valve to be subjected to the steam-chest pressure, the load on the valve is 16^ x 10 X 134 ="22,110 lbs. The relieving-pressure on

Proceedings.] ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 177

the area of one steam-port is 13.V X If X 97 = 1,800 lbs. The relief due to steam in the valve-passage isl3^Xfx97 = 980 lbs. The relief pressure over the exhaust area of the valve is at this point of the travel zero. Thus the resultant load on the valve is 19,330 lbs. The coefficient of friction is therefore 1,321 -^ 19,330 = 0*068. Similar calculations, for diagrams 29 to 36 and 37 to 44, give valve-resistances of 1,096 and 982 lbs., and coefficients of friction 0-054 and 0*051. The former set of diagrams was for a brass valve, and the latter for a cast-iron valve, working together on the same pair of cylinders, the diagrams being taken on the same day and under the same conditions. Both valve- and cylinder- faces were in good condition. The lowness of the coefficient of friction is surprising. It is ordinarily taken at 0*08 to 0 * 09 for well-lubricated surfaces. It is the more remarkable as the valve- surfaces were at a temperature of about 350^ Fahrenheit, so that any oil reaching the surfaces must either have been vaporized or very thin. The pressure per square inch of bearing surface is about 380 lbs.

Table IV. — Percentage of Power Lost in Friction of Valves and Eccentrics.

atrake of Valve in Inches.	Revolutions per Minute.	Indicated HP.	HP. to drive Valves.	Per cent, of Power lost.
2J	56 56	343 212	4-6 4-8	1-34 2-26

The difference due to different modes of lubrication is shown in Table II. Diagrams 77 to 91 were taken during a run of 50 miles without any oil ; diagrams 92 to 99 with a sight-feed lubricator feeding seven drops per minute ; diagrams 100 to 107 with the same lubricator feeding sixty drops per minute. The valve-resistances, for the full-stroke pulling diagrams, when reduced to the same steam- chest pressure of 120 lbs. per square inch, were 1,868 lbs., 1,764 lbs. and 1,359 lbs. For the short-stroke diagrams the corresponding valve- resistances were 2,432 lbs., 1,898 lbs. and 1,475 lbs. These figures show a decrease of valve-resistance with increase of lubrication.

In the case of the goods engine, the full stroke of the valve is 3| inches, and the mean valve-resistance is 1,592 lbs. Hence the power required to keep two valves moving is per revolution

1,592 x-| X 4 = 1,791 foot-lbs. The work lost at the eccentrics

[the INST. C.E. VOL. XCV.] N

178 ASPINALL ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [MinuteB of

may be estimated thus : CircumfereBce of sheaves, 44 inches ; pressure normal to sheave, 1,592 lbs. ; coefificient of friction (as ordinarily- assumed), 0 • 08. Then, work lost per revolution at two eccentrics

44 = — X 1,592 X 0-08 X 2 = 933 foot-lbs. Work lost in friction of

two valves and eccentrics 2,723 foot-lbs. per revolution, or at 10 miles an hour, or 56 revolutions per minute, 4-6 HP. A similar calcixlation for the short-stroke diagTams gives 4 • 8 HP.

The work lost in the friction of the valves and eccentrics is about the same as that required to drag two and a half 10-ton wagons.

The Author has to acknowledge the valuable assistance rendered in the conduct of the exj^eriments and in the preparation of the Paper by his assistant, Mr. E. Coey.

The Paper is accompanied by three sheets of illustrations, and seven blue tracings of indicator diagrams, from which a selection has been made for reproduction to form three Plates, 2, 3, and 4.

I

[Discussion.

I

Proceedings.] DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 179

Discussion.

Mr. J. A. F. AsPiNALL wished to mention that most of the Mr. Aspinall. experiments referred to in the Paper were made in 1885, and some in 1886. The Paper had been considerably abridged, so that some of the explanations with regard to the progress of the experi- ments might not be quite as clear as they otherwise would have been ; but he should be happy to give any additional explanation if required. The valves used were, of course, valves between the cylinders as illustrated in Plate 2, Fig. 1. He had no doubt that if experiments were tried with valves working on the top of the cylinders, or with valves under the cylinders, somewhat different results would be obtained. He regarded the whole series of experiments as of a somewhat tentative character. He had not been able to obtain much information upon the subject, and he hoped the matter would be fully discussed. He might mention that the chief difference between the diagrams taken with the Allen valve and the ordinary valve would be seen by comparing the Diagrams on Plate 3 with those on Plate 4. There was not the sudden rise with the latter which took place at the period of exhaust with the former. The serious drop, as between the boiler- pressure and the steam-chest, was due to there being too small a pipe. In the case he had mentioned the engine had a S^-inch steam-pipe, and it was not large enough to take the steam of 17 by 22-inch cylinders when the engine was going at so low a speed as 4 miles an hour.

Mr. W. Cross said the subject of the Paper was one, as tlie Mr. Cross. Author had stated, about which very little was known, and very few experiments had been published. The coefficient of friction 0-068 was extraordinarily low, especially as compared with ordinary marine practice, which was usually taken at 0 • 250. From the very careful manner in which the experiments had been con- ducted, there was no doiibt that they were ai^proximately true; but if so, how was it possible to account for the valve-spindles breaking, as they undoubtedly did in locomotives, and especially, as far as his experience went, in inside-cylinder engines ? He had known two or three cases of valve-spindles of l|-inch diameter breaking. The valves were of brass, and the valve-spindles of the best Yorkshire iron. On changing the valves to cast-iron no further trouble was occasioned. He would not enter into the reason of it, but it showed that under certain circumstances the coefficient of friction must be enormously increased. It would

N 2

180 DISCUSSION ON FKICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Mr. Cross, therefore he imjDossible, as would appear at first sight, to reduce to any great extent the area of the parts driving the valve. A few years ago his firm was engaged in running trial trips with a high- speed man-of-war. He believed that the vessel was one of the first fitted with high-speed piston valves. The dimensions were : cylin- ders, 43 by 82 by 36 inches; 116 revolutions per minute, indicating 6,500 HP. ; both cylinders Ip. and hp. were fitted with double piston valves, two to each cylinder. At certain speeds, with a certain number of revolutions per minute, difficulty was experi- enced owing to the vibration of the valve-gear. At about 90 revolu- tions per minute, the engine would run quietly, but at 100 it would be vibrating severely ; at a gradually increasing speed the vibrations would die away altogether, and then at 112 or 115 revolutions, the vibration would be unsui^portable. The matter

Fig. 1.

Indicating Cylinder for MEASURmG Actual Strains in Valve.

Scale 1.

was, of coiarse, carefully investigated, and Mr. Marshall and he finally constructed the apparatus shown by Fig. 1. It was almost a reproduction of the Author's, except that there was no friction, and that they trusted only to mechanical fitting. The strains upon the two ends of the cylinders were exactly similar, as the spindle went through, and consequently the acting area was only the area of the brass piston. The stroke was exceedingly small, only 0 • 05 inch at each end. This dynamometer cylinder was fixed into the centre of the valve-spindle in the same manner as shown by the Author. There was an indicator at each end, the cord was fastened to some convenient portion of the valve-gear. The stroke of the valve connecting-rod being 7^ inches, it was too much for the drum of the indicator ; but by connecting it with some moving portion the stroke could be reduced to the amount required.

Proceedings.] DISCUSSION ON FEICTION OF LOCOMOTIVE SLIDE-VALVES. 181

Various liquids were tried, but thick glycerine was found to be Mr. Cross. the most suitable material. The machine ran for many hours without giving the slightest trouble. The friction of the piston- valve, properly speaking, was practically nothing. The weight of large valves was very serious. The cylinders were originally fitted with piston-valves weighing 1,500 lbs. each, and it would be seen (Figs. 2 and 3) that the shock at the two ends of the stroke was exceedingly severe. They then stopped the experiments, and instead of having heavy cast-iron valves they made phosphor- bronze valves as light as possible, bringing the weight down to 600 lbs. Not only was the size of the diagram much reduced

Fig. 2.

Fig. 3.

110 REV$

(Figs. 4, 5 and 6), but taking also into account the weight of the valves the shock upon the valve-gear was very much reduced. The result had been that, although the engines had been running some years, there had been no further trouble from this cause. The indicator-cylinder was fitted with a compensating arrange- ment, a kind of small by-pass valve fitted on the outside, whereby if the glycerine leaked from one end to the other, by stopping the engine it could be replaced in the centre. In some respects he thought the apparatus was an improvement upon the Author's. In the first place, he thought that the friction of the cup-leathers in the latter was extremely doubtful. It varied enormously with the age of the cup-leathers, the variation amounting sometimes to 60 or 70 per cent. As far as he could follow the Aiithor's diagram, the dummy end of the spindle was not balanced, and consequently the areas

182 DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Mr. Cross, were diiferent ; but lie presumed that a correction would be allowed for this. That the friction of the cup-leather had been excessive was shown in Table I, in the case in which the machine was tested by means of dead-weight. He therefore thought that the friction- less apparatus, trusting entirely to mechanical fitting, was decidedly siiperior for giving accurate results. He might be permitted to refer to a Paper read by his partner, Mr. F. C. Marshall, and his assistant, Mr. E. L. Weighton, on " High Speed Engines," ^ in which the results of the diagrams exhibited w^ere fiilly discussed. He hoped to hear something on the friction of diiferent metals at

Fig. 4.

100 REV?

Mr. Halj^in.

high temperatures ; this was exceedingly important, especially for marine engineering in these days of very high pressures.

Mr. Druitt Halpix observed that indicating an engine under the most favourable circumstances was not a pleasant thing, and indi- cating an engine under the difficulties described, running 40 miles an hour, was still less pleasant, and was often dangerous. The Author had referred to previous contributions on the subject exist- ing when his Paper was written, but since then a communication on the same subject by Mr. C. M. Giddiugs, had been published with illustrations, diagrams and results, in the Transactions of the American Society of Mechanical Engineers,^ 1886. The experiments

* Transactions of the North-East Coast Institution of Engineers and Ship- builders, vol. ii. p. 287.

* Transactions, vol. vii. p. 631.

Proceedings.] DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 183

were made with the apparatus shown by Plate 2. Looking at the Mr. Halpin. matter broadly, he certainly thought that the Author's results were almost too good to be true. Fig. 7 showed the general arrange-

FiG. 7.

ment of the instrument in side elevation. Fig. 8 was a diagramatic section on the centre line, illustrating the relative positions of the various parts of the instrument when a pushing force was applied

Fig. 8.

PUSHlNQi

Fig. 9.

to it. Fig. 9 was a similar section showing the instrument under compression, and Figs. 10 and 11 were diagrams taken by the instrument recording the actual friction due to the valve and the

184 DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Mr. Halpin. stuffing-box. He should be afraid to design a valve-gear with the expectation of such very low frictions. Mr. Cross had referred to the errors that might arise by the use of the hydraulic cylinder with his packing, and Table II contained the results of the actual tests made by the Author, with the apparatus shown in Plate 2, Fig. 5, giving the calculated and the measured pressures. Taking the first, the calculated pressure was 12 lbs., and the recorded pressure 5 lbs., being a difference of 138 per cent. Taking the last, the calculated was 65 lbs. and the recorded pressure 49 lbs., or a difference of 33 per cent. These differences occurred when the apparatus was at rest, and might become magnified when it was in motion. He had plotted out the whole Table graphically,

Fig. 10.

Fig. 11.

and the result was a straight line, showing that the friction was pretty uniform. Still it was a large amount, and, of course, it was a varying amount, taking the condition of the packing leather into consideration. The apparatus used by Mr. Giddings was very ingenious and simple. Unfortunately the results obtained could hardly be checked. One of the engines was a small engine with a 6j-inch by 10-inch cylinder, going at various speeds ranging from 125 to 200 revolutions, and the friction varied from 2 per cent, down to 1 • 4 per cent, of the whole dynamometrical power given by the engine. That was with a balanced valve. With an ordinary valve such as the Author had been using, not a balanced valve, a 9-inch by 12-inch engine was used, and there the friction varied from 4*5 per cent, to 7 -3 per cent. That was very different from the frictions given in Table IV, which in one case was 1*34 per cent., and in another 2-20 j)er cent.

Proceedings.] DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 185

That Table, he thought, needed a little further explanation. Jlr. Halpin. In the first experiment the valve moved 3^ inches with 56 revo- lutions, taking 4 • 6 HP. to drive it ; whereas, in another experi- ment with the same number of revolutions, at only two-thirds of the speed, it took rather more HP. to drive it. The Author had stated as the first cause of variation of the resistance during the stroke, " the variation of pressure on the back of the valve due to variation of the steam-chest pressure during each stroke. This in some of the fixll-gear diagrams amounts to 10 or 12 lbs. per square inch." The Author had given a verbal explanation of that, stating that the steam-pipe was not good enough. Mr. Halpin was of the same opinion ; but he thought that the Author might have gone further. On referring to Table II, instead of finding the variations of pressiire 10 or 12 lbs., he would find the total difference of pressure between the boiler and the steam-chest to reach as much as 50 lbs. Taking the Table as it stood, in the case of the goods engine, which was the first, making 56 revolu- tions a minute, assuming a constant piston speed throughout all parts of the stroke, the velocity was 5,800 feet per minute in tlie steam-pipe ; and taking the extreme case with 224 revolutions, there was a velocity of 23,200 feet per minute, which was, of course, far beyond anything it could be hoped to keep up, without great loss of pressure, by means of steam-pipes of the proportions now usually adopted in locomotive practice. He had never been able to obtain a velocity of more than from 3,600 to 4,000 feet per minute without a very serious loss of steam-pressure.

Mr. John Goodman said, that in 1885 he wrote a letter to Tlie Mr. Goodman. Engineer,'^ giving a diagram, of which he now exhibited a copy, of a similar apparatus to that used by the Author for ascer- taining the friction on slide-valves. He was going to carry out some experiments upon the same subject; but could not get an engine of suitable dimensions. He thought, however, that the apparatus might be improved by working upon a method of which a great deal had been heard lately, namely, the Emery system of testing machines. Here, instead of a piston, a thin diaphragm was introduced in the cylinder, and an indicator with a Bourdon tube was used instead of an ordinary piston indicator. In that way the difference in the length of the valve-rod might be reduced to a minimum, and the apparatus made to work frictionless. In the apparatus which he had designed, he had arranged the paper to work on a continuous roll. He thought it must be a

' Vol. lix. p. 186.

186 DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Mr. Goodman, difficult operation to put a card on the indicator-drum when the engine was going at a high speed, and that a continuous roll of paper would be a great improvement. The Author had referred to the friction of glands which, he said, was quite inappreciable. It might have been inappreciable in that instance ; but persons acquainted with the working of steam-engines would know that friction on the glands might be enormous. Indeed, it was an easy- matter to pull up an 8 or 10 HP. engine simply by tightening the piston-glands. He admitted that it was a very injudicious practice, but still it could be done ; and hence in any experiments upon the friction of slide-valves, it might come in as a serious error, and it might, to some extent, account for difference in the results obtained by different experimenters. In the experiments, for example, mentioned in the first paragraph of the Paper, the coefficient of friction came out as 0"2o, while Mr. Halpin's results, and those obtained in America, were much lower ; the difference being probably due to the extra friction on the slide- valve spindle-glands. Eeference had been made to the enormous friction that sometimes occurred on slide-valves from the simple reason that the buckles and slide-valve rods were not infrequently broken. He thought that was easily accounted for. It was well known that the friction of rest on all materials was far greater than the friction of motion ; and not only with dry- surfaces, but also with lubricated surfaces; in this case even more so, because the valve when at rest, would be metal to metal with the steam-chest ; but as soon as it began to move, a film of water and steam would get under it, and would act as a lubricant, and gTeatly reduce the friction. Eighteen months ago he saw a balanced slide-valve, and the makers were showing how easy it was to move the valve backwards and forwards with one hand. A valve, about 8 inches square, was moved very easily in that way. He tried several experiments with the valve, allowing it to remain at rest for some minutes, tapping the steam-chest cover, and then pulling with the hand as hard as he could. He could not move it, although it was supposed to be a balanced valve ; but after it was once started into motion it could be moved backwards and forwards without trouble. The experiments of Mr. Adams and Mr. Beattie were made by hanging a scale-jian with a pulley on the end of the slide-valve spindle. Weights were placed in the scale-pan until the valve began to move. That, of course, would be the friction of rest, and the greatest possible amount of friction that could come on the slide-valve. Hence, in designing a valve-gear the maximum load likely to come upon

Proceedings.] DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 187

it should be taken, and he thought it would be advisable to Jlr. adhere to the old-fashioned coefficient 0"25. The question of lubricants was one of very great importance. Some of the so-called valve oils were of no value. He knew of a case in which an enormous quantity of a well-known lubricant was sent out to one of the colonies, and after it had been at work three or four months complaints were received to the effect that all the j)orts and all the parts of the piston were choked with a hard mass of carbon. The oil had been largely advertised as a valve oil, but it was far from being suitable for the work. A few years ago he had carried out some experiments for Mr. Stroudley, with reference to suitable lubricants for slide-valve and piston work ; and he be- lieved that the method adopted was the only one for ascertaining whether an oil was siiitable or not. The oil proposed to be used was heated on a plate to about the temperature at which it was to be used in the cylinder. If after a reasonable time it turned thick and gummy, and ultimately solid, it was rejected ; but if it stood the heat for several hours, and ultimately almost entirely evaporated leaving practically no residue, it was deemed to be suitable oil for use in the cylinders. Recent experiments, especially those of Mr. Tower, and some which he had himself carried out, distinctly showed that with the ordinary system of lubrication, such as that used with siphons, there was a great waste of oil. He thought there was a wide field open for inventors to devise some means of lubricating slide-valves by pad lubrication, instead of the present method of allowing the lubricant to go drop by drop into the cylinder. There was a method that he thought might be worth trying, something similar to that adopted by Messrs. Aveling and Porter at Eochester for lubricating their traction-engine axles. It was a piece of wood ^-inch in diameter, floating in a bath of oil touching the bottom of the axle. As the axle revolved, the wood revolved and brought oil on to the axle. That method had proved economical and efficient, and he would suggest whether some such method could not be used for slide-valves. A small roller might be inserted in some part of the bars of the steam-chest (of course it would not be of wood but of metal), and small quantities of oil could be admitted to the centre of the rollers. That, he thought, would be more economical and efficient in reducing friction than the method of putting in oil drop by drop. The pressure required to move the valves was considerably less when the valves were work- ing full stroke than when they were notched up to one-quarter stroke. It would be found that the work done in moving the

188 DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Mr. Goodman, valve throngli the stroke was approximately constant. If some apparatus of that kind were adopted for the use of piston-rods, so as to indicate the work done by the piston, some valuable and interesting results might be obtained, that would throw light upon the friction of pistons and glands as well as of slide-valves. Mr. Tower. Mr. Beauchamp Tower asked the Author how he had obtained the results given in Table I. He presumed that he had added the weights one by one from 76 lbs. to 412 lbs. on the lever ; that in the first place he found that he had 132 lbs. indicated when the actual pull was 316 lbs., and so on until he had 1,765 lbs. when the indication was 1,285-7 lbs. He wished to ask whether the weights were then taken off and the indication noted. In the case where he had 1,715 lbs., and only 1,285*7 indicated, the friction amounted to 429 • 9 lbs. It would therefore appear that he could go on taking off the weights until the pull was 429 • 9 lbs. less than 1,285 • 7 lbs. ; that was 855 • 8 lbs., without overcoming the friction of the appa- ratus, and between 1,715*6 lbs. and 855*8 lbs. his indicator would not have moved. The Author's formula, given on p. 170, would clearly be useless in that case. With a frictional apparatus where there were changes of force going on inside, of which there was no external indication, such a formula would not apply. Finding that the frictions were so great, he thought it was a pity that the Author had not contrived some less frictional method, which would have made his results much more valuable. He might have adopted some such plan as Mr. Cross had mentioned, which, from Mr. Tower's own experience, answered perfectly well. With oil or glycerine there was practically no leakage and no friction. The straight horizontal lines, which were characteristic of the diagrams, clearly indicated the excessive friction of the apparatus. Within certain variations of load the Author's indicator showed no variation of force at all ; hence there was a straight line where there ought to be a wavy one.

Mr. Stroudley. Mr. W. Stroudley thought the Author had done well to bring the matter before the Institution, as it was one of great im- portance, and deserved to be carefully studied. About the year 1855 he had charge of some locomotive express engines very trou?jlesome to keep in order, the slide-valves having a very small bearing-surface, and the pressure being high. He then devised a plan of drilling two rows of holes down each face of the valve, and filled them with pure block tin, which reduced the friction so as nearly to double the life of the slide-valve. He afterwards saw some extreme cases of friction, some engrines havino; slide- valves 20^ inches by ll.j inches, working at 160 lbs. pressure, and

I

Proceedings.] DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 189

having 4\ inclies travel. They also were express engines, and Mr. Stroudley. the slide-valves wore down a good deal within a month or six weeks. The metalling of the valves just turned the scale. Before being metalled he had seen the slide-valve turned over on the edge, so that it would measure j inch more across, and the exhaust-port would be reduced by f inch. That was the result of extreme friction. Guided by previous experience, he had en- deavoured to reduce the size of the valve as far as possible, so that the load should not exceed a reasonable one, and that removed much of the friction, as was proved by the wear that took jDlace. The B and C class of engines on the Brighton Eailway, with slide- valves placed below, ran for ^V inch wear on the face (equivalent to 1 lb. of metal in a phosphor-bronze valve) 14,460 miles, the wear including the re-facing of the valve. With a brass valve the distance run was in some cases only 7,000 miles. There was, however, extreme difference in the wear of valves, dependent, no donbt, on the nature of the metal of the cylinders, as well as the difference in the care exercised by the driver, the extreme varying from 7,000 to as high as 77,000 miles per lb. of metal worn and turned off the face. The phosphor-bronze valve was nearly twice as durable as the ordinary brass valve. With the D class of valves placed vertically, the wear was ■^, inch for 7,000 miles with a brass valve, and 13,000 miles with a bronze valve. Nearly as good results were obtained with a vertical phosphor-bronze valve as with a horizontal brass valve. The friction had, in his opinion, always been assigned a higher value than it ought to have. Many years ago he tried experiments with the balanced valves (Adams's), changing the proportions and reducing the friction. With a 17-inch cylinder engine he could hold the reversing gear with one hand standing at right-angles when running at full speed. In that case the friction must have been very small, yet the consumption of coal was not appreciably altered. The quantity of oil required to keep the balanced valves from squeaking, and in working order, was much greater than for the ordinary valve. The result of his experiments was that the balanced valves were taken out and thrown away. That went to prove, in his opinion, that the friction must have been much smaller than it was genei'ally supposed to be. With reference to the Author's estimate of the friction and the power used by eccentrics and eccentric straps, he had now a great number of eccentric straps of cast-iron that had been running since 1872-3-4 with all classes of engines. If the friction were as great as had been stated by previous writers on the subject, those straps would have been worn to a much greater extent than

190 DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Mr. Strouilley. they were, they not having been refitted in any way since their construction. He thought that experiments made under any other conditions than actual working ones were not of much value. The Author's arrangement was well designed to illustrate what actually took place in work.

Mr. E. A. CowPER said the Author had not done fall justice to himself in Table III, where the various frictions of motion were given, namely, Adams, 9,752 lbs. ; Beattie, 4,620 lbs. ; Aspinall, 1,321 lbs. with a brass Allen valve full stroke, 1,096 lbs. with a brass plain valve full stroke, and 982 lbs. with a cast-iron valve full stroke. He considered the Author had not laid sufficient stress on the last being cast-iron, because it appeared that with the cast-iron valve far less driving power was required than with the other valves. He believed many people thought that it was best to have a brass valve, so that the faces of the ports should not suffer. With regard to the Allen valve, the Author had mentioned cutting out the inside equal to the length of the ports. That was a principle which Mr. Cowper had long adopted ; indeed he had cut out the inside considerably more than the distance between the ports, when working a single common high-pressure engine expansively, so as not to have too much compression ; more thorough expansion with a single valve was then possible. He should have liked to have heard more about balanced valves. For instance, there was Eobert Wilson's valve, which had a plate on the back of it, carried, not by the slide-valve itself, but by two sides or supports, one on each side of the slide-valve, and these touched the slide face. This plate and its supports was stationary, and the slide-valve worked under it, the plate taking the pressure of the steam on its own back. Then, if water occurred in the cvlinder, and the valve lifted off its face, the valve and plate lifted together, and no harm occurred to the cylinder bottoms. The mode of fitting up the plate to its place was to put a pressure on the back, to spring it as much as the steam would, and then get it up true, the back of the slide-valve being also got up true (without pressure) ; then the two were put together, and the faces to bear on the ports were got up true together, and thus no stuffing-box or relieving-piston was required for the back of the slide-valve. This valve worked very freely, and, he was informed, assisted in making the steam-hammer quite easy to handle. He had further been informed, by a thoroughly practical locomotive engineer, that if the plate was allowed to travel a little, it wore away like the slide-valve, and thus avoided the necessity of having a little taken off the face, to keep it to fit to the back of the slide-valve. This

Proceedings.] DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 191

would seem to be an important point. With regard to a large Mr. Cowper. valve with a short stroke, requiring more power to move it than if it had a long stroke, it would appear that any lubricant would be worked out more with a short and reciprocating motion, than with a longer motion, and as a proof of this, he might instance the fact that the axles of a common carriage placed on a railway truck would, in a long journey, set fast in their boxes, if the wheels were not tlioroughly well fastened to prevent motion. Also the old rock shaft, formerly employed in Bury's engine, and others with the "Gab-Motion," frequently cut, and even seized with the slight motion they got when running. With regard to slide-rods break- ing in marine engines, or engines with very wide ports and slides, he had no doubt but that in some cases the edges of the slides had caught against the ports, from their springing down or dropping into the ports, from the pressure of the steam on their backs. This might be expected if the valves were made too weak, as they often were, and if the edges of the ports were not rounded off a little, as they ought to be.

Mr. E. Woods, Past President, thought that the Author's re- Mr. Woods, searches and experiments oi:ght to receive the attention of engineers, especially as showing that the friction of slide-valves was much less than had been commonly supposed. The first application of the balance valve to locomotives was by Mr. John Gray, about the year 1838, who applied it on the Liverpool and Manchester Eailway to one of the engines used to assist in propel- ling trains up the Sutton and Whiston inclined planes. To one of the two, the " Sampson," or the " Goliath," he placed a short cylinder at the back of the valve-chest with a piston in it, and a connection between the piston-rod and the back of the valve, so as to balance the pressure of the two. But, as Mr. Stroudley had pointed out, the complexity of the method told against it, and it was finally abandoned, it being found that on the whole the friction was not actually reduced.

Mr. AsPiNALL, in reply, said he did not find that the breaking Mr. Aspinall. of valve-spindles, as mentioned by Mr. Cross, often happened. Speaking from the experience of a very large number of engines, he found that it was not a frequent occurrence with valve-spindles that had been properly proportioned, unless the glands were screwed up too tight, or the valve was allowed to wear too thin, and caught in the steam-ports. With regard to the use of thick glycerine, he had no doubt that it would answer extremely well. He did not think he had mentioned that oil was tried with the apparatus shown in Plate 2, Fig. 5. No doubt, as Mr. Halpin

192 DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VAIVES. [Minutes of

Mr. Aspinall, had said, the age of cup-leathers might affect the experiments, and the greatest possible care was taken in that respect. Only once diiring the experiments were the leathers changed, and then a fresh set of experiments "was tried with the apparatus, and it was found that the friction was the same as before. As to the question of greater power being used in a short stroke, that had been answered by Mr. Goodman. That it was more difficult to lubricate valves when notched up than when going full stroke, was a matter of practical experience. In reference to Mr. Halpin's remark as to the reduction of jDressure, he meant that the very great reduction, amounting to more than 30 lbs. in some cases, as between 160 lbs. boiler-pressure and the steam-chest pressure, was due to the small steam-pipe. In speaking of 10 lbs. or 12 lbs. he was referring to those depressions in the steam-chest pressure due to the gulp of steam taken by the cylinders. Mr. Goodman had stated that he would prefer a diaphragm to leathers. No doubt such an apparatus would be better in some ways as the movement of the piston was so slight, and the result of his experiments had been to satisfy him that his apparatus could be considerably improved. If he were to start again he had no doubt that he could get better tackle all round. But such matters required a long time, and there was not always the opportunity of trying such experiments. The question of gland-friction of course applied to the present case. The greatest possible care was taken to ensure that there should be no gland-friction, and that the packing should be properly done, and when the valve-spindle was disconnected from the rest of the gear, it could be put in by hand with ease. With regard to the increase of friction, which might be caused by screwing or packing the gland improperly ; it was only necessary to look at valve- spindles when they came into the shop deeply cut by wear, after being in the hands of a careless driver, to see how a driver might increase the friction of the glands. In the case under consideration the aim was to eliminate the gland-friction, and to get the valve- friction. Mr. Goodman was right in saying that there was the greatest difficulty in obtaining good oil for lubricating. Many of the so-called cylinder oils produced a choking-up of the ports with a black substance, which had to be chipped out when the engine came into the shop. He had experienced something of that kind quite recently. The objection to the balanced valves, apart from any question of their deficiency as valves, was the fact that there were so many loose parts to deal with. If Mr. Goodman's suggestion as to little rollers were carried out, it would be intro- ducing loose pieces into a part of the engine where it was most

Proceedings.] DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. 193

desirable to keep them out. Mr. Tower had spoken of the results Mr. Aspinall, in Table I obtained by trying the ap2)aratus shown by Plate 2, Fig. 5, and he appeared to think that if the weights were put on and taken off one by one, the apparatus woiild not record any movement. That was what he himself had anticipated, and the apparatus was most carefully tested on that account. The weights were placed on the end of the lever one by one, and taken off one by one, and for every additional weight taken off or put on, the indicator responded at once. Having tested them in that way, they were tried by screwing up the apparatus at the end as a whole, to see whether, starting with the whole weight (1,200 lbs. in one case, or 1,700 lbs. in the other), the same results followed ; and he was satisfied that, so far as the results were concerned, they were as accurate as they could be under the circumstances. Mr. Tower also spoke of the straight horizontal lines on diagram 17 (Plate 3) as indicating friction in the apparatus. These straight lines were, however, not due to that cause but to the use of the Allen valve. If the diagrams on Plate 4, taken with an ordinary valve, were compared with those on Plate 3, taken with the Allen valve, the former would be found to show curved wavy lines. When the Allen valve was removed from the engine, and the ordinary valve substituted, these straight lines disappeared ; when it was put back again they reappeared. The explanation might be that given on p. 175. He coiild quite bear out the remarks of Mr. Stroudley as to the very slight wear with cast-iron eccentric straps, which after running for a long time came into the shops without the tool-marks being worn out. Mr. Cowper had spoken of cast-iron valves, and had pointed out the fact that the friction was lower within than with brass valves. That was shown clearly by Table II, where it would be seen that the force reqiiired to move the valve amoiinted to 992 lbs. in one case, and to 998 lbs. in the other. As long as the kind of valve was right, cast-iron seemed to do very well, and for engines starting and stopping repeatedly extremely good results were obtained. Mr. J. C. Park, had told him that, during the last two or three years, all the brass valves had been taken out of the engines of the North London Eailway Company, and cast-iron valves substituted. The result was a good deal less wear and tear, and consequently less lubrication ; and there had not been the slightest difficulty in reversing the engines with the steam full on at a pressure of 160 lbs. He did not think there could be better evidence as to the value of cast- iron valves, under certain circumstances, where they could be well lubricated by the steam. It was true that the North Lot'dou

[the INST. C.E. VOL. XCV.] 0

194 DISCUSSION ON FRICTION OF LOCOMOTIVE SLIDE-VALVES. [Minutes of

Proceedings.

Mr. Aspinall. engines did not run any great distance at one time. With an express engine running long distances, he doubted whether cast- iron valves would give the same results.

Correspondence.

llr. Park. Mr. J. C. Park remarked that slide-valve friction had been under consideration for many years on the North London Railway, and that experimental trials, with valves more or less balanced, had been made from 1856 up to the present time, but without success. On the North London Railway, where more than 7,000 stoppages were made daily, the engine-drivers much preferred cast-iron valves, as they were able, without effort, to work the reversing lever quite easily with steam at 1 60 lbs. per square inch full on ; whereas with brass valves they were compelled to partially close the regulator. The wear was as 1 to 3 in favour of iron valves. For the last six years cast-iron valves had been used on the North London line, and he hoped in a short time all the engines would be fitted with them. The valves were made of 1 part of No. 1 Carron and No. 3 Cleator hematite iron, and 2 parts of clean scrap. The average mileage of these valves, if properly lubricated, was 82,491, with a wear of barely ^V inch, whereas brass valves were completely worn out under the same mileage. He would, later on, present fuller details to the Institution on this subject.

Sclocted GEIBBLE ON SURVEYING IN NEW COUNTRIES. 1^^

Papers.]

SECT. II.— OTHER SELECTED PAPERS.

{Paper No. 2301.}

" Preliminary Survey in New Countries, as Exemplified in the Survey of Windward Hawaii."

By Theodore Graham Gribble, Assoc. M. Inst. C.E.

In January, 1887, the Author was requested to make a survey of the windward side of the Island of Hawaii, Sandwich Islands, and report upon the feasibility of a narrow gauge railway, 70 miles long, to carry the sugar and other produce to the port of Hilo. The country is entirely volcanic, sloping down from three extinct, and one active crater to the sea. The engineering difficulties were mainly those presented by the gorges, or " gulches," so graphically described by Miss Bird in her " Six Months in the Sandwich Islands."

The exports of the districts of Hilo and Hamakua have now reached 45,000 tons of sugar alone. The fruit trade is increasing ; and coifee of very fine quality has been produced above the sugar belt, in small quantities only; but the new Government are portioning out homesteads for cofiee-cultivation, which are being freely taken up by the Portuguese settlers with their larg-e families. At its present price, 005*06 will pay better for cultiva- tion than sugar. The present outlets for the commerce consist of timber landings with cranes, built out on promontories underneath precipices. Here the sugar is received from the rope-way in a tram, or cage, and conveyed to schooners or coasting steamers upon boats or punts, when the weather permits. Sometimes the planta- tion has to suffer a month's detention, at others the loss of a boat- load of sugar.

The advantages of a railway in such a situation were undoubted ; but it was a question whether a line could be constructed at a cost which would yield a profit to the stockholders. The gorges above referred to are about a hundred in niimber, running down from the volcanoes to the sea, intercepting in every case the line of route. The methods adopted on the survey were mainly those in vogue in America; and the object of this Paper is to describe them for the assistance of young English engineers who may use

o 2

196 GRIBBLE ON SURVEYING IN NEW COUNTRIES. [Selected

them in the Colonies. The survey occupied five months of field work, and one month of oflSce work. The cost averaged £20 per mile.

The original intention was to run in and out of the gorges by means of forty back shunts. These were all dispensed with by means of curves of not less than 143 feet radius. The curve limits were combined with gradients as follows : a radius of 1 50 feet with a gradient not exceeding 1 in 50 ; a radius of 300 feet with a gradient not exceeding 1 in 40 ; and a radius of 450 feet with a gradient not exceeding 1 in 33. Within these limits, one engine will haul all the load required by the traffic. The estimate was for a 3-foot gauge line, with a maximum engine "weight of 25 tons. Two types of engines, four wheels coupled and six wheels coupled, were used for ascertaining the maxima moments of flexure on bridges ; and from these an equivalent uniformly distributed load was determined for each separate span. Plate girders, on the deck system, were adopted up to 50 feet span, and beyond that truss bridges. The estimated cost was somewhat under the limit prescribed by the promoters for assurance of financial success. The demonstration of the practicability of the line was largely due, in th.e Author's opinion, to the methods of telemetry adopted. A list of the instruments used, and particulars about them will be found in Appendix I.

A route survey, or reconnaissance, was rendered unnecessary by the existence of excellent maps of the district, prepared by the Surveyor General, Professor Alexander, and his assistant, Mr. Curtis Lyons ; these maps show the coast-line, Government road, the outline of the gulches, and most of the earlier mill buildings. The geodetic survey, on the same lines as those of Great Britain and the United States, is plotted to 500 feet per inch, and referred to rectangular co-ordinates of latitude and longitude, which proved of great service in checking the azimuths. The route was determined by the situation of the mills, all of them close to the sea.

It was brought out very clearly upon this survey that neither triangulation of the gorges, nor chainwork in the more even country, could afford the same degree of accuracy as the stadia work ; and in respect of desjjatch, they were not to be compared with it. The problems presented by the gulches were of great variety, from huge gaps, a quarter of a mile wide and 400 feet deep, necessitating a two-mile detour, down to openings which could be cheaply spanned. The gulchwork was all done by optical measure- ment ; chaining was out of the question. The sides of the gorges

Papers.] GRIBBLE ON SURVEYING IN NEW COUNTRIES. 197

sloped 35° on an average, but often reached 60°, with bh;ffs here and there of crumbling lava rock, giving a bare foothold to the men, and no place for an instrument. In addition to this, the vegetation was extremely dense, having grown undisturbed for centuries on the richest soil in the world. The Hau, or yellow- hibiscus, is the greatest foe to the climber; its roots run several hundred feet above and below ground, and its branches mat into a web, pliable yet hard ; it resists alike axe, hatchet, or cane-knife, and to cut a trail through it would take days. The system of optical measurement only requires the clearing of a small spot here and there to catch a sight of the staff, yet this often took from an hour to two hours and a half. The location between the larger gulches, including a considerable number of small ravines, was done with the chain, transit, and level, by Mr. Lincoln Cabot of Boston and a field-party. The whole of the traverse was plotted to astronomical azimuths, corrected every now and then by obser- vations. The compass was quite unreliable, as the magnetic attrac- tion of the hills frequently caused a deviation of several degrees within a few hundred feet. The azimuths were all determined by solar observations.

The following four methods are recommended, of which only the two latter were used —

(1) Observation of Polaris at its greatest elongation.^

(2) Observation of a pair of circumj)olar stars when in the same vertical.'-^

(3) Equal altitudes of the sun.

(4) A single observation of the sun out of the meridian.

The two first require an illuminated axis in the transit. The observer must, in either case, sit up to catch his stars, and chance their being clouded. Methods 3 and 4, being diurnal observa- tions, are more convenient ; but the calculation is more lengthy. A close approximation to No. 3 is obtained, without calculation, by taking the equal altitudes about two hours from the meridian, and taking the mean bearing to be the meridian. This method has the disadvantage of bringing back the observer to the same point of observation to take his second altitude. The last is by far the most useful method, both as being diurnal, and, as con- sisting of only one observation. The calculation needs an ordinary

' This method is most common. See Trautwine's pocket-book, ' Hints to Travellers,' ' Raper's Navigation,' &c.

^ This method is described by Professor Stockwell, of Cleveland, U.S., in the Journal of the Association of Engineering Societies, and may be used without » transit.

1£8 GRIBBLE ON SURVEYING IX NEW COUNTRIES. [Selected

table of logarithmic sines and cosines, and a Whitaker's almanac. The importance of the obserTation for azimuth can scarcely be overstated ; any error, either in the field or in the plot, is thereby at once detected. Even a good surveyor will rarely run a line for several miles in roiigh country without making some error; but a frequent check of azimuth will enable him so to distri- bute minute instrumental errors as to render them unscaleable quantities.

The modus operandi of the optical work at the gulches varied someAvhat according to the nature of the obstacles ; but the follow- ing course was adopted where the diiBculties were greatest. The transit was fixed on one side of a gulch, and foothold dug out ; the observer gave his entries to a calculator, who booked them, and worked up the position of each sight there and then. On the other side of the gulch, the leader of the staff-gang held the levelling- staff, the head chainman carried only a sight-stake, the tail chain- man carried the compass clinometer, only the tape being used ; and short distances of from thirty to fifty, and occasionally a hundred feet, were run. The tail chainman booked distances, bearings, and vertical angles. One of the axemen took the slopes of the hill at right angles to the line. Thus the line was run through the bush until a spot was reached suitable for a clearing. All hands then went to work with axes, hatchets, and cane-knives, and cleared a sight of the levelling-staff for the observer at the transit-station. This observation determined : (1) the actual distance of the staff from the transit, by observing the stadia hairs ; (2) the true azimuth of the same by observing the horizontal limb ; (3) the elevation above sea-level of the same, by observing the vertical lines. These points were termed primary points, and were laid off first upon the plot, from the transit-station. The compass-traverse, made by the staff- gang, was then plotted on tracing-paper, and superimposed upon the plot of the primary points ; the intermediate points were then pricked through, the traverse completed, and the contours filled in. When one side of the gulch was finished, the two parties changed sides, and repeated the same operation. By this means, a net-work of triangulation was obtained, serving as an independent check to the stadia work. The use of the compass was thus limited to short stretches between points accurately determined in elevation and azimuth. The survey was plotted on a scale of 100 feet to the inch on account of the steepness of the side hill, a difference of 10 feet in height showing xerj little, even on this scale, when the slope is 60"^. The contours were drawn at every 5 feet of height in the open country, and every 10 feet in the gulches. The alignment

Papers.] GRIBBLE ON SURVEYING IN NEW COUNTRIES. 199

was laid on from the contours ; after which the profile was plotted on ruled profile paper, to a scale of -tOO feet per inch horizontal, and 40 feet per inch vertical.

Telemeter station (T. S.), and staff station (S. S.), mean respec- tively the positions occupied, at the time of observation, by the instrument and the staff. When the instrument is shifted, say from A to B, a backsight is first taken from B to A. The elevation of the staff station B is now re-booked as the elevation of the telemeter station B, and vice versa with A (Appendix II). The plotting of the survey is simply the reproduction, on paper, of points obtained by radial distances from the telemeter station, the angles being the actual astronomical azimuths of the radial lines. The detail of buildings was filled in upon the same principle, but with the plane-table ; this instrument, of the simplest construc- tion, was placed over one of the points previously fixed by the theodolite, and carried a field tracing of the spot showing the skeleton traverse. It was generally arranged to fix the plane-table within 100 feet distance of the corners of the buildings, so that the subsidiary radial lines could be taped at once. By this means, the labour was divided and time saved.

The term elevation is the elevation of the point above mean sea- level. The term optical axis means the intersection of the vertical axis of the pivot with the axis of the telescope, and the abbrevia- tion for its elevation is 0. A. The vertical component (V. C.) is the product of the direct distance measured along the line of sight by the sine of the vertical angle. The horizontal component (H. C.) is the product of the said distance by the cosine of the said angle. The backsight or foresight is the reading on the staff of the axial hair. It is a backsight (B. S.) when used to determine the elevation of the optical axis from the known elevation of a staff station, which is only done to commence work and in shifting the instrument. It is a foresight, or intermediate (F. S.), when the reverse oj)eration is performed. The working out of the vertical component is some- times done by the slide-rule ; but it is better to use a table of sines to four places. The horizontal component is obtained from the graduations on the inner side of the vertical limb of the instru- ment, "ratio of hypotenuse to base." The multiplications of the sines by the direct distance are done very rapidly by two office hands, the one with the sine-table, the other with Dr. Crelle's calculating tables.

In the example of field-book chosen (Appendix II), the direct distances given are all just equal to the difference of the upper and lower stadia readings, multiplied by 100. The divergences actually

200 GEIBBLE ON SUEVEYING IN NEW COUNTRIES. [Selected

existing at each 100 feet are usually booked on the fly-leaf of the field-book, so that, without any further tables, the direct distance can be corrected and entered at once from the sight. The lower hair was, whenever possible, directed on the first joint of the staff, so that a simultaneous reading of upper and lower hairs could be obtained. The central or axial hair was always a mean reading between the other two. It was not read except at a change of telemeter station as a check, but it was left to the recorder to work out.

Eule 1. — To obtain the elevation of the optical axis (0. A.) from a backsight on a bench mark or other datum point, the known elevation of which is booked as elevation of staff station (S. S.) :

(a) When the vertical angle is p/ws, 0. A. = S. S. -|- B. S. - V. C, where B. S. is the backsight and V. C. the vertical component.

(6) When the vertical angle is minus, 0. A. = S. S.-f B. S.-f-V. C.

Eule 2. From the elevation of the optical axis, ascertained as above, to obtain the elevation of any staff station (S. S.) :

(a) When the vertical angle is plus, S. S. = 0. A.-f-V. C.-F. S., where F. S. is the foresight or intermediate.

(b) When the vertical angle is minus, S.S. = 0. A. - (V. C. -f F. S.). When the instrument is shifted, the new elevation of the optical

axis is obtained, as at first, by a backsight upon the known station just left ; but an independent check is obtained by actual measure- ment of the height of the optical axis above the new telemeter station. This may be done by an ordinary tape ; but in the Author's instrument, the plummet-chain terminates in a hook exactly 2 feet below the optical axis ; a steel tape is hooked on, and measures the height more -quickly and correctly. Even with the iitmost care, there will generally be a slight error discovered here, sometimes arising from fault of adjustment, sometimes from incorrect reading of the stadia. If the discrepancy is divided equally between the two stations, the error will be removed if it arises from the first-mentioned cause ; if from the second cause, it is impossible to locate it, and if it amounts to anything serious, the sights should be repeated. The Author has levelled 100 feet in one shot with only a divergence of y^ foot between backsight and foresight; the vertical angle was about 15°, so the direct distance was nearly 400 feet. Generally the discrepancy in such cases averaged from 3 to 6 inches.

In conclusion, it is suggested that for rapid and effectual pre- liminary survey —

(1) A telescope should be used which will discern Jupiter's satel- lites, mounted either on the ¥, or transit plane of the theodolite.

Papers.] GRIBBLE ON 8UEVBYINQ IN NEW COUNTRIES. 201

(2) The micrometer and stadia systems should be combined in the one instrument.

(3) The staff should be provided with sights and biibbles, to dis- pense with double calculations by insuring a position of the staff at right-angles to the line of collimation.

(4) An independent check of the levels should be made with each change of telemeter station, by means of the measurement of the height of instrument with the steel tape.

(5) The bearings should be all taken from the north point, ascertained and frequently checked by astronomical observation.

(6) The plane-table should be used, both as a sketching-board and also on its tripod ; but only as an auxiliary, and without any of the expensive attempts at making it a universal instrument.

[Appendixes.

202 GKIBBLE ON SURVEYING IN NEW COUNTRIES, [Selected

APPENDIXES.

APPENDIX L

List asd Description of Instruments.

1. One 6-inch transit theodolite, with a telescope of 14" focal length eccentri- cally fixed and counterbalanced in place of the ordinary one. Stadia hairs were added, reading 1 foot per 100, correct at 500 feet, the intermediate values being determined by a measured base in the usual way. The short telescopes usually fitted with stadia hairs would not have sighted across the ravines.

2. One micrometer telescope, 2 feet focus, by Elliott, as supplied to the army. It measures distances approximately from the height of infantry or cavalry. It was modified to suit with observations upon a pair of disks 10 feet apart. A tripod was added, fitted with tangent-screw ; the same tripod carried a plane- table of cheap construction, as the two instruments were not used simultaneously.

3. One 14-inch dumjiy-level.

4. One box-sextant, by Troughton and Simms.

5. One 5i-inch aneroid, in sling-case, by Steward, reading with vernier to single feet. This instrument is very distinct in its gi-aduation. The difterential scale of elevations, and uniform scale of pressures, are transposed by means of a " snail," so that the pressure-scale becomes ditferential, and the scale of elevations uniform, thus permitting the introduction of the vernier. The scale of elevations was specially worked out from Guyot's formulas, and found to be somewhat at variance with those usually adopted by instrument makers. The instrument was verified at Kew Observatory, and the errors at dift'erent pressures registered. The climate of the Sandwich Islands is very equable, causing less fluctuations in the barometer than in most countries ; but a lighter, cheaper, and more compact aneroid is recommended, as the gain through precision of gi-aduation will be often annulled by atmosjiheric changes. A 3-inch Sopwith aneroid would be much handier.

6. Two Abney levels.

7. One compass-clinometer. Colonel O'Grady Haley's patent, by Elliott. The prismatic compass is all that could be desired ; but the clinometer, being of the plummet type, is too slow in its action. Messrs. Elliott now make a combined prismatic compass and Abney level, of the Author's design, which will be found quite as accurate and more rapid than the O'Grady-Haley instrument. The same makers also construct the Author's telemeter-theodolite. It is a 7-inch Y theodolite, surmounted by an 18-inch telescojje of twenty power, which may be also carried in a sling-case for reconnaissance. Another telescope of forty power is provided in the same box for astronomical purposes, and long-range stadia measurements. The telemeter-theodolite is a combination of the micrometer-telescope and the stadia-theodolite (Fig. 1). Though heavier than the ordinary theodolite, it can be carried by one man, and is much less bulky than two instruments, each with its own tripod. "\\'hen used upon its tripod for the actual survey, the ordinary Sopwith staff is used, being read by the horizontal stadia-hairs. Two small levels are attached to the staff, and a sight-vane for placing it always at right- angles to the line of sight ; also a pair of sliding disks are fixed to the staff when, by a signal from the leader," the assistant knows that an extra or check-

Papers.] GKIBBLE ON SURVEYING IN NEW COUNTRIES.

203

sight is required, by means of the movable micrometer hairs. The stadia readings are used up to the limits of such observation ; and the micrometer readings are taken occasionally as a check. Beyond these limits the micrometer only can be used. The stadia hairs are fixed at 1 per cent., as is customary,' subtending 1 foot on the staff for every 100 feet of actual distance. The instru- mental constant is obtained, and the values of readings registered, by actual observation at every 100 feet, and the intermediate by interpolation.

The Author's experience does not go to prove the extraordinary accuracy claimed by the various inventors for the instruments of this type. If the registration just referred to be perfectly made, it is clear that the limit of accuracy with the stadia measurements will be in the same ratio as the power of the instrument. An ordinary 5-iuch theodolite will only read the hundredths ou

Fig. 1.

Micrometer Head to Theodolite.

the staff at about 300 feet. The Author's two telescopes will read them at 600 and 1000 feet respectively. Therefore the limit of accuracy is 1 foot up to these distances, and beyond them depends upon the experience with which the trained eye can estimate the portion intercepted upon the tenths by the stadia hairs. Beyond that limit, the accuracy is reduced to estimation between the feet-marks on the staff. If the telescope were of sufficient power to observe with the micro- meter hairs the precise extremities of the disks, the accuracy would be as gi-eat at a mile as at 100 yards, but this is impracticable ; and the one method is about as accurate as the other. The stadia hairs are needed for the shorter distances, as being very much more conveniently reduced in the field-book ; the micrometer cannot be dispensed with for the longer distances. The combination of the two in one instrument has been a long-felt want of surveyors.

Minutes of Proceedings luat. C.E.. vol. xci. p. 285.

204

GRIBBLE ON SURVEYING IN NEW COUNTRIES. [Selected

APPENDIX II.—

Tele- meter Station.

Staff Station.

Horizontal Limb.

Vertical Limb.

Stadia Hairs.

Direct Distance.

Horizon- I Vertical tal Com- I Compo- ponent. | nent.

BM A. A,. A3 A, A, B

A

B.

B,

B3

B,

B.

Ba

B,

Bb

B.

B.o

Bn

B,3

C

273° 05' 205° 18' 149° 28' 122° 30' 111° 40' 98° 48'

278° 48'

295° 35'

300° 05'

303° 03'

307° 38'

314° 07'

351° 29'

359° 42'

26° 20'

31° 40'

54° 35'

104° 27'

104° 27'

56° 05'

157° 18'

+ 2° 05'

— 3° 14 + 3° 35'

— 1° 25'

+ 0° 06'

— 1° 20'

500\ 780 f 5001 870/ 500 \ 535/ 500 \

1,115/ 500 \

1,260/ 500\

1,390/ 200\

1,430/

300\

1,530/ 500 \

1,075/ 500 \

1,085/ 500 \

1,050/ 500 \ 980/ 500 \ 900/ 500 \ 790/ 5001 665/ 500\ 745/ 500\ 850/ 500 \

1,130/ 500 \ 553/ 500 \ 976/ 500 \

1,148/ 500\

1,250/

280 370 35 615 760 890 1,230

1,230 575 585 550 480 400 290 165 245 350 630 53 476 648 750

280

1,228

1016

69

76

26

87

Papei-8.] GREBBLE ON SURVEYING IN NEW COUNTRIES.

205

— FlELD-BoOK.

Elevation of Tele- meter Station.

Height of Instru- ment.

Backsight.

Foresight or Inter- mediate.

Elevation of Staff Station.

Elevation

of Optical

Axis.

Remarks.

494

91

422

25

400-93

4-75

90

511

6-40

15

15

503-42

422 494

408 420 400

499

66

427

15

N. corner of hedge.

Hedge.

Junction of hedge.

»> >>

Hedge.

Hedge.

at corner.

Centre of road.

206

GRIBBLE ON SURVEYING IN NEW COUNTRIES. [Selected Fig. 2.

	^ ^^j£^ Bg It	-t:	/■■
	' f:::^- '«. '" .>-^«	^
\	"^Si^, ^.
		rT7" \	f
	£"" DJS
hs \ \ / \ ^	/
Y*i-t. ■•% }\^ _^__i:^	/ t
|d,3 _.W;>^^ y j	/
	Hh
	\
	n
"io / ''2*.?. \ /	4
	1 ---f«\/	i
	*=«	'" /xr—AG
	F.8		i
: --^ \>^y

IlorUontaZ Scale 83S f'-l huJv Vertu>i.l Scale 75 f* — 1 1ruJo

Section of Px^bi.ic Road at Occvpation Road.

I

Papers.] GEIBBLE ON SURVEYING IN NEW COUNTRIES. 207

APPENDIX III.

General Principles of Telemetry and Telemeters.

The term telemeter, which was introduced by surveyors, has been appropriated to so great an extent by electricians, that it is likely to be abandoned by the former for the term tacheometcr. The simplest form of telemeter is the plane- table ;' it is a graphic triangulation to attain the same end as the optical tele- meters. The optical telemeter triangulates for distance upon the same principle, but with the additional precision of the telescope.

1. One class of telemeters is that in which the measured base forms part of the instrument itself. Such are Adie's 18-inch and 3-foot telemeters, described in Heather's " Instruments," in Weale's Series, Piazzi Smyth's 5-foot telemeter, Colonel Gierke's 6-foot telemeter, and Otto Struve's Ti-inch telemeter. The disadvantage of this class is the error produced by expansion and contraction. They are not much used now.

2. The second class is that where the measured base is at the point observed, generally consisting either of a graduated staflF, or a pair of disks connected by a rod. In this class there arc two subdivisions.

(a) Those which have a fixed base and a varying angle, as the Rochon micro- meter telescope, furnished with a reflector by which the images of the disks are made to coincide, as in a sextant, and the angle is then measured in terms of distance. Elliott's army telescope, in which the fixed base is the height of infantry or cavalry, where two wires fixed in the diaphragm are caused to approach to or recede from one another by a micrometer screw ; the wires are actuated so as to exactly include the object observed, and the distance is read ofif either on the infantry or cavalry side of the screw, as the case may be. Binoculars are made upon the same principle. The advantage of this method is that it may be used approximately, by observing a man on foot or on horseback when impossible to send out a man with a target ; but it may be also used with precision in cases where the rod-man can hold up a target with a pair of disks whose extremities are equal to 5 feet 11 inches, or 8 feet 10 inches.

Eckhold's omnimeter is a combination of the first class and subdivision (a) of the second class just alluded to. The instrument is a transit to begin with ; it is furnished with a graduated distance-plate on the horizontal limb, worked by a micrometer screw, and read by a very powerful microscope at right-angles to the telescope. At one operation, by observing top and bottom of a 10-foot staff, the distance and elevation are obtained by means of calculation. The instrument has given much satisfaction both in India and the Colonies ; it is, however, more complicated than the stadia principle, takes longer to adjust and to work, and is not more accurate.

(b) Telemeters which have a variable base and a fixed angle. This is the stadia principle, described by Mr. B. H. Brough in his Paper on " Tacheometry, or Rapid Surveying." ^

3. The third class of telemeters is that in which the base is measured on the ground, at the point of observation. The Hadley sextant, though not, strictly

' Minutes of Proceedings lust. C.E., vol. xcii. p. 187. - Ibkl, vol. xci. p. 282.

208 GRIBBLE ON SUEVEYIXa IN NEW COUNTEIES. [Selected

speaking, a telemeter, is used as such ■when a distant vessel is observed simul- taneously from the deck and maintop. The Dudge-Steward omnitelemeter,' the " Bates " range-finder, and many others of this class, are more suitable for military than civil engineering. The measurement of a base upon the ground at the point of observation, unavoidable wiih military engineers, has not yet been successfully accomplished with that combination of celerity and accuracy which is reached by the stadia and micrometric measurements.

Engineering, August 20th, 1886.

Papers.] TOPHAM ON RAPID SURVEYING. 209

(Paper No. 2330.)

"Rapid Surveying."

By Francis David Topham, Assoc. M. Inst. C.E.

Engineers are sometimes called upon to execute works, such as roads or railways, in countries of which no maps at all exist. In such cases the engineer has to explore and consider the advantages of different routes before a detailed survey is begun. It will then be of great subsequent convenience if he can make some rapid sketch survey of the lines of route and surrounding country. The Author has practised a method of making such surveys while marching, and without retarding the speed of the march. A rapid route survey will serve as an illustration, which he once made from a place in the interior of Asia Minor to a seaport town, under the directions of Mr. C. E. Austin, M. Inst. C.E., who has largely employed this method. The route lay through a country of which no maps exist with the least approach to accuracy, even such features as mountains and rivers being entirely wrongly placed on the best maps. In a survey of this sort, it is impossible to actually measure the distance traversed ; but in a fairly flat and even country, it is best determined by stepping, and recording the number of steps by a pedometer, a trained man, by preference a soldier, being entrusted with the duty. But the country through which the above survey was taken, is most rough, rugged and moun- tainous, where wheel traffic does not exist, and where the travelling is scrambling and climbing, rather than marching. The only roads are narrow mule-paths winding over mountain passes and pre- cipices. Over such a road, the only way of arriving at the distance is by estimating the pace at which the leading muleteer is marching. This, after practice and checking by actual measure- ment, can be done with a much closer approximation to accuracy than one would think possible. All angles are taken within the magnetic meridian by the prismatic compass, which gives a result surprising in its accuracy ; for all angles being taken with the same meridian, there is no accumiilation of errors, as there would be if the angle on to each foresight were taken from the last back- sight. The procedure is as follows : — The observer from horse- back takes an angle on to the next forward bend of the road, and notes the time of leaving the point at which he took the anarle. and

[the INST. C.E. VOL. XCV.] I'

210 TOPHAM ON RAPID SURVEYING. [Selected

the time of arriving at the point on to which he took the angle, and notes also the estimated pace at which the distance from one point to the other was travelled. Arrived at this point, he similarly takes an angle on to the next bend beyond. After a little practice these angles may be taken without halting. Wherever practicable, a check angle should be read back on to prominent points of the road behind, and on to all the prominent moimtain tops. The levels of all mountain passes, river-crossings, and other important points are taken by an aneroid barometer. The survey had to be done at the rate of about 25 miles a day, much of it in very wet weather ; and it can hardly be taken as an instance of what could be done by the same method, if not more than 10 miles a day had to be traversed.

A scale was prepared so as to plot off the distances, as expressed in minutes, without reducing them to actual feet as follows : — Divide any line into equal parts by the points A, B, C, D, and E. At E, erect a perpendicular, and divide it into parts corre- sponding to the distance travelled in a minute at the rate of 4 miles an hoiir. At B, C, and D, draw lines parallel to the line at E. From A draw lines to these points in the line E, intersecting the lines B, C, and D. Then if the spaces of line E represent a minute of time at 4 miles an hour, the spaces cut oif by the same pencil of rajs on lines B, C, and D, represent the distance travelled in a minute at 1, 2, or 3 miles an hour respectively. In the same way the scale for any other pace can be added. In plotting the work, it will be necessary to choose a scale in which the shortest interval of time between the angles, at the slowest pace travelled, is not too small to measure conveniently with the dividers.

Having made a complete circle of some 250 miles, the Author found that the two lines traversed plotted in meeting at the starting point very closely indeed. This shows the accuracy with which work may be done with the prismatic compass ; and although failing of itself to prove the accuracy of the scale, it indicates that the Author had estimated the pace uniformly throughout.

The Author claims for this manner of surveying great rapidity, and even tolerable accuracy, especially in an open country where long straight sights can be taken, and that such a survey would be of great use when making a subsequent detailed and carefully measured survey.

TajxTS.] VICKERY ON SURVEYING IN AUSTRALASIA. 211

(Pajyer No. 2348.)

" The Practice of Surveying in the Australasian Colonies."

By Samuel Kingston Vickery, Assoc. M. Inst. C.E.

The conditions under which surveying operations are conducted in countries in which settlement is in progress diifer so widely from those prevailing in older countries, that the method obtaining in the Australasian Colonies may be of interest to members of the Institution, and of service to young engineers who may contemplate practising their profession there. In those Colonies, there is no branch of the profession of a civil engineer — if those which belong to the architect and mechanist may be excepted — which is not practised directly under the control of the State ; and this rule applies more especially to the branches in which surveying occupies a prominent part. Railway surveying, as well as construction, is in the hands of the Railway Departments ; building surveyor's work, so far as public works are concerned, is undertaken by the Public Works Departments ; irrigation surveys, including storage and reticiilation, by the Water Supply Departments ; underground surveying by the Mining Department ; whilst subsidiary road surveying, though now chiefly carried out by the local governing bodies, is under departmental supervision, and the officers employed are required to qualify themselves by examination- Surveys which are immediately under the control of the departments of Lands and Survey are performed almost exclusively by salaried officers, or by those who are " authorized," on either Crown lands or sold lands intended to be Ijrought under the operation of the Transfer of Lands Statute. Plans of surveys made by unauthorised persons are not officially recognized. All authorized surveyors are obliged to comply with the regulations issued in the form of departmental instructions, or bearing the force of an " Order in Coiincil." Particulars about the examinations for selecting certificated sur- veyors are given in Appendix I.

Each colony is usually divided into survey districts, and^eack district is placed under the control of a salaried inspecting officer. These districts are sub-divided into divisions, each being assigned to an authorized surveyor, to whom orders for departmental work are transmitted. This officer is not salaried, being remunerated by fees fixed by regulation, and is removable at the pleasure of the

i> 2

212 VICKERY ON SURVEYING IN AUSTRALASIA. [Selected

Surveyor-General. "When a vacancy occurs in one of these divisions, selection is made from the list of those holding a certificate of competency obtained by examination. The inspecting oftlcer has full control in his district. He makes periodical examinations of authorized surveyors' work in the field, inspects their instruments, field-books, &c., and certifies to all accounts. Inspecting surveyors are under the immediate control of the Surveyor-General. The higher classes of surveying, including minor triangulation, are performed exclusively by staif officers ; and ordinary sectional or block surveying, embracing the laying out of towns and roads, is chiefly executed by authorized survej'ors. The latter work, though theoretically simple, demands the exercise of considerable care, judgment, and observation, on the part of the sTirveyor, who is expected to study the interests of the State in all his operations. The system of examination and departmental supervision is now so complete in the various colonies, that a very high standard of efficiency has been attained by those employed.

The Author has had occasion to test the accuracy of the siu'vey of single blocks of 320 acres in the heart of steep, scrubby, and heavily timbered ranges, in which the error of arc was found to be under 1', and the error in linear measurement under three links in either latitude or departure. Appendix II gives an illustration of a block of 320 acres, through which a check line was projected by the Author showing the errors of " closing " to have been limited to 0^ 0' 30" in angular, and 0*1 link in linear measure- ment. The survey had been made with a Cook's five-inch theodolite, and a steel band of loQ feet. The form of plan shown is that adopted by the Victorian Survey Department ; but those required by other colonies are substantially similar. The dimen- sions on the plan furnished must " close " geometrically, and the maximum diflerence admissible between it and the field-notes is not to exceed 2' in arc, or the tangential equivalent on the length of any line. Chainage by the " stepj^ing " process is permitted ; but when the slojie of the ground exceeds 5° from the level, it is preferred that the angle should be measured on the vertical arc of the instrument, and used for reducing the measurements to the horizontal plane. Although in the illustration given, the " closing " errors are very trifling, it does not follow that the work is as accurate as the test would at first sight make it appear. Every operation in the field, no matter how simple or carefully done, is attended with a certain practical error ; and the summation of such errors may be either cumulative or counteractive, and possibly in the instance given they may have acted in the latter way Had

II

I!

Paj^crs.] VICKERY ON SURVEYING IN AUSTRALASIA. 213

the survey represented by the plan been effected on a hot day in summer, and the check line measurerl on a cold winter day, there would have been a possible error in "closing" of 3*4 links in latitude, and 4*2 links in departure. For a line 1 mile long, measured in the winter time with a steel tape 66 feet in length, may vary nearly 5 • 5 links from a line measured with the same chain in the summer time, from the effects of the variation of temperature alone in the Colony of Victoria. The mean monthly range of temperature to which a surveying chain would be subject there, varies from 43-8^ in winter to 54* 7^ in summer; and the variation in length in 1 mile would be : —

Winter . Summer

Length i Tenipera-

of ture of

Band. Atmusphere.

Increase in	Number
Lenf,'tli of 1 foot	of
of Steel lianii for	Chains
1"^ Fahrenheit.	In I mile.

Feet. Links.

66' X 43-8° X 0-000006886 x 80 = 1-592484 = 2-41 66' X 54-7° X 0-000006886 x 80 = 1-988787 = 3-01

or, between the winter and summer measurements (2*41 -f- 3*01) 5 • 42 links ; whilst, as there are some days in summer when the heat is phenomenally great, it is possible for the expansion and con- traction to exceed this estimate. Another cause which operates against acciaracy in chaining is the error due to the unequal tensile strain employed ; the extension in 1 mile with a 66-foot steel band, having an effective sectional area of h" X -^s", being 0*667 link where a strain of 30 lbs. is used, the results varying according to the strain and sectional area. A still more potent cause of error is when the measurements are made on inclined ground, necessitating the suspension of the tape used, " a steel tape of 6 lbs., used with a tensile strain of 30 lbs., giving an error of 13-44 links per mile," ^ due to the sagging or curvature of the tape.

Even with the aids of a thermometer to regulate the allowance for differences in temperature, and a spring attachment to regiilate the tensile strain, practical errors in linear measurement in the general survey can only be eliminated where they are based upon, or connected with, standard lines laid down in connection with geodetic or trigonometrical work.

It is patent that in a new colony where the population is still sparse, and the territory unalienated, a geodetic or trigonometrical

' Traiisactious of Vict. lust, of Surveyors, vol. ii. Paper by Ca]ttain Kielly.

214 YICKERY ON SURVEYING IN AUSTRALASIA. [Selected

survey should be iindertaken in the first place ; then the principal streams and mountain ranges delineated, and the quality of the land noted and classified ; and finally, the most suitable courses for main roads determined before the sub-division of the colony into blocks for the purjDose of settlement is commenced. This course is, however, unfortunately seldom practicable, the preliminary vrork being liable to be interfered with through influence brought to bear on Parliament by those anxious for immediate settlement, and the too fi'equent readiness of the Government of the day to abandon or postpone scientific but unremunerative operations, in favour of those which tend directly to feed the revenue. In some of the colonies, however, minor triang-nlation, generally depending on either major triangulation or meridional circuit, has preceded or is being carried on concurrently with settlement and sectional surveys, by which means future trouble is avoided by permitting the technical descri2)tion of boundaries of each holding to be expressed correctly with relation to the true meridian. But while neither the meridional circuit nor major triangulation surveys are sxafiiciently refined for purely scientific purposes, the data which they supply will be of great practical value in the future, the referring points being permanently fixed and available for obtaining azimiiths.

In the colonies, or parts of them to which geodetic operations have not been extended, settlement or sectional surveys miist be connected with or based on the nearest boundary, the accuracy of the bearing of which has been recognized and adopted. Where no reliable datum for bearings can be obtained within a reasonable distance, the surveyor is reqiiired to project a true meridional line on the ground by one of the approved methods, and base his work ujion it.

Eepeated checks made by the Author have convinced him that, with the use of a light steel band of length varying from one chain to ten chains, according to the character of the country in which the operations may lie, and a good o-inch transit theodolite, the maximum degree of errors stated as being allowable by the Govern- ment survey departments is not too small.

I

[AlTKNDiXtJi.

rnpcrs] VICKEEY ON SURVEYING IN AUSTRALASIA. 215

APPENDIXES.

APPENDIX I.

Examination op Surveyors.

Each flepartment has its own board of examiners; and the subjects for examination, in order to obtain a license, depend upon the description of work required. For examj^le, municipal surveyors, besides acquaintance with the ordinary subjects, are obliged to jjossess all the knowledge necessary for road and bridge construction, and the setting out of works ; mining surveyors, in addition to underground surveying, must understand practical mining, and the machinery and appliances used in connection therewith ; water-supply surveyors must prove themselves acquainted with hydraulics ; and those employed by the Lauds Departments, with whom this Paper deals more par- ticularly, before receiving certificates of competency, are examined in (fi) the construction, adjustment, and use of the theodolite, level, and other modern in- struments ; (b) the principles and practice of subdivisional, topogi-aphical, and road surveying ; (c) practical trigonometry ; (d) computation ; (e) jilotting by ordinates and otherwise; (/) charting and drawing. Additional special subjects are (a) practical geodesy, including the determination of latitude and longi- tude by prime zenith distance, and by prime vertical observation of the true meridian by astronomical observations, and of the relative latitude and longitude by triangulation ; (b) spherical trigonometry and trigonometrical analysis ; (c) setting out of curves, and computations therewith ; (d) levelling and mensu- ration of earthwork. These examinations generally extend over a period of from five to seven days of eight hours each; and about 30 per cent, of the candidates presenting themselves succeed in passing. The subjects vary in detail in the dilfereut colonies ; but the examinations are practically of equivalent value. In some instances certificates are granted without examination, the fact of their being so obtained being stated. In Victoria, these are granted only under the following conditions : — Having passed some examination — eqiiivaleut in the opinion of the board to that prescribed in Victoria — in Great Britain, the United States of America, India, or some British colony, and having been in practice during not less than one year ; having been engaged in that colony under some authorized surveyor for not less than six months, and having a favourable report on his qualifications from the Surveyor-General, or the Insj^ecting Surveyor in charge of the district in which the applicant has been employed.

APPENDIX II.

Survey of Block, with Check Line. Parish of Bunnugal. County of Ripon.

Geological Formation. — Newer volcanic, covered by recent tertiary.

Physical Conformation. — Fern-banks and swamps.

Soil. — Poor, sandy. Vegetation. — She oak, gum, pei^permint, and ferns.

216

VICKERY ON SURVEYING IN AUSTRALASIA. [Selected

Fig. 1.

G WALKER

cJa.„^S

C H.ARMYTAQE

smSS' 2g?£ -'

%^ ^ ^^

r. CAMERON

{SelBction/J

(Seledian/)

CTuajtB JO

Scaler 30 Ovaiits = 7 AM/.

^^%mni$

Closure and Area of Block.

Bearings.	Dist.	N.	s.	E.	W.	W. L.	Areas. N. S.
S. 89° 59' E. South . . East. . . S. 0° 3' E. . S.71°53'W. S. 33° 47' W. S. 89' 58' W. N 0^ 3'AV.	Links. 6,706 1,079 1,001 1,100 3,471 3,379 2,525 6,070	Links. 6,070	Links. 2-0 1,0790 l,l66-0 1,079-0 2,808-5 1-5	Links. 6,706-0 i,o6i-o 0-9	Links. 3,298-0 1,878-9 2,525-0 5-2	Links. 6,706-0 13,412-0 14,413-0 15,414-9 12,117-8 6,910-9 2,537-0 6-8	0-13412 .. 144-71548 .. 169-56390 .. 130-75106 194-93518 0-03805 0-41276
	6,070 6,0707,707-97,707-1	0-41276 640-13779 0-41276
		Area, 319a. 3r. 18p.		2)639-72503 A. 319-86251 E. 3-45004 P. 18 00160

Pa]x;rs.]

VICKERY ON SURVEYING IN AUSTRALASIA,

217

Closure of Check Lines, West Portion of Block.

Bearings.	Dist.	N.	s.	E.	w.
S. 89° 59' E. . . S. 31° 3' E. . . S. 71° 53' W. . . S. 33° 47' W. . . S. 89° 58' W. . . N. 0° 3' W. . .	Links. 2,996-5 3,629-0 483-0 3,379-0 2,525-0 6,070-0	Links. 6,070-0	Links. 0-9 3,109-0 150-2 2,808-5 1-5	Links. 2,996-5 1,871-8	Links. 459-1 1,878-9 2,525-0 5-2
	6,070-0	6,070-1	4,868-3	4,868-2

218 HUNTER OX THE MANUTACTUKE OF OIL-GAS. [Selected

(Paper No. 2233.)

" The Manufacture of Oil-Gas on the Pintsch System, and its application to the Lighting of Railway-Carriages." ^

By Gilbert Macintyre Hunter, Assoc. M. Inst. C.E.

I. The Manufacture and Compression of the Gas.

(a) Manufacture.

The Author's object, in submitting this Paper to the Institution, is an endeavour to make the process of oil-gas manufacture better understood and appreciated, rather than to bring forward facts previously unknown.

Generally, the gas is made from once-retined paraffin oil in cast- iron retorts, the tar is removed, the gas condensed, washed, and purified, and then passed to the gasometer, from which it is drawn by compression pumps, and forced into cylindrical holders.

At the Cook Street (Bridge Street Station, Glasgow) Gasworks,^ which supplies the Caledonian Eailway,^ the arrangement is as follows (Plate 5, Figs. 1 and 2) : — In the retort-house there are two sets of benches that can be worked alternately, also an upright boiler for suppl;^dng steam to a duplicate set of compressing pumps. In the purifying-house are the condensers, washer, a double set of purifiers, and the meter ; while the storage cylinders are in a separate house at a short distance.

The retorts, Plate 5, Figs. 3, 4 and 5, are Q -shaped cast-iron tubes 5 feet 10 inches long, and 10 inches wide, placed one above the other, the lower one resting its whole length on a firebrick sole. The flame from the furnace passes round both sides of the lower tube, over the top, and there engages with the upper tube, finally reaching the flue or combustion chamber. Thus the flame plays directly upon all parts of the tubes, not unduly heating some parts and others partially. The flues are built of ordinary firebricks.

> A Paper on " Compressed Oil-Gas and its applications," by Arthur Ayres, M. Inst. C.E., was read and discussed at the Institution in the session 1887-88. Minutes of Proceedings Inst. C.E., vol. xciii. p. 298.

^ Minutes of Proceedings Inst. C.E., vol. Ixxxvii. p. 392.

^ The company has another gasworks at Perth for supplying the northern section of t)ie railway.

PajXii's.] HUNTER ON THE MANUFACTURE OF OIL-GAS.

2:9

Both ends of the tubes are fitted with mouthpieces or covers, admitting of thorough cleaning.

The oil is pumped from barrels into a 15-gallon cistern on the top of the retort-bench, which thus maintains a constant head and a steady flow. It enters the upper retort, Plate 5, Fig. 4, in a thin continuous stream from a micrometer cock, through a siphon pipe, and falls upon an iron tray on the bottom of the retort. This prevents the cold stream of oil coming in direct contact with the hot retort, and thus reducing its temperature. Here it is partially vaporized, and, passing through the tube, descends to the lower tube, where it encounters a higher temperature, rendering it permanent or "fixed." The object of this is to insure that every portion of the vaporized oil shall be acted upon by the highest temperatiire, without undue exposure to the heated surface of the retort. As the gas leaves the retort, it is largely mixed with tarry and other

Figs. 1.

Section of Washer.

Plan of Underside of Tray.

vapours, which give it the appearance of smoke of a more or less dense nature, in proportion to the quantity of such matters present. It now passes through a tar pit or hydraulic main, filled with water, where the first separation of the tarry matter is effected. The condensers consist of two vertical cylinders, through which the gas passes, entering at the bottom of one and passing up through it and then down through and out at the bottom of the other, the effect of which is to condense the tar. The gas now passes to the washer (Figs. 1).

The inlet-pipe is vertical, with a tray over it having curved arms or " spreaders " on the under or water side. The gas on entering is turned downward by the tray, and driven through the water in a circling fashion by the spreaders ; it finally passes out of the water, through a series of perforations round the edge of the tray, into an upper or gas chamber, and thence to the purifiers. This washing process assists in removing the brownish or smoky appearance of the gas, and arrests and deposits all the tar which

220 HUNTER ON THE MANUFACTURE OF OIL-GAS. [Selected

may pass with it. The quantity of tar in the gas varies with the temj^erature maintained within the retorts, and also with the quantity of oil run into the retort. There is also a great variation in the quality of tar given off by different oils. Some classes jield very light " thin " tar, which comes out at the overflow in the washer, being lighter than the water ; while other classes give off heavy tar, which sinks to the bottom, and is drawn off to the tar tank. Both the condensers and the washer are i:)rovided with siphons to remove the tar. Although the gas has been undergoing purification, so to speak, in the successive stages through which it has passed, it is now submitted to a final and thoroi^gh purification for the removal of sulphur compounds, carbonic acid, and sulphur- etted hydrogen. The purifiers are arranged in a double set, and by means of a four-way valve the gas can be made to pass through either one or other, or both. The purifying material consists of slaked lime and sawdust in the proportions of 2 to 1, which is renewed weekly. It is spread evenly about 2 inches deep over a perforated tray, in the same manner as a coal-gas purifier. The gas now passes to the meter, and thence to the gasholder, which has a capacity of 1,700 cubic feet.

At night the main valve is shut, the dampers are closed, and the covers over the tar pit are removed, to allow the gas in the pipe to escape ; the manhole covers are also removed, and the tar and pitchy matter drawn off. The mouthpieces are next taken off, the retorts cleaned out, and the covers replaced ; the fires are drawn, and afterwards banked for next morning. The fires are kindled with coal and dross ; thereafter they are wholly maintained with coke, which gives much better results than coal. After they have been kindled for about an hour and a half, the oil is turned on. The projDer heat has not been obtained yet ; but, owing to the demands on the work, longer time cannot be spared, so that the yield of gas per gallon of oil is comparatively small for the first hour or so. When the retorts are irregularly heated there is a large increase in the quantity of tar ; while, if they are overheated, the distillation is attended with a separation of carbon in the form of soot, and if insufficiently there is deposited in the retort a carbonaceous substance resembling coke, in addition to an abundant production of tar. The best temperature is a bright cherry-red heat, verging on white; then there is only a small quantity of carbon or soot.

A high illuminating gas ^ can be produced when the yield is

' Transactions of the lustitutioir of Engineers and Shipbuilders in Scotland, vol. XXX. p. 236.

PaiK'l-s] HUNTER ON THE MANUFACTURE OF OIL-GAS. 221

small ; and conversely, with a high quantitative yield, the light- giving quality of the gas will be diminished. A fairly good yield, which can be obtained in every-day practice, varies from about 70 to 85 cubic feet, or 18,480 to 22,440 cubic feet per ton of oil, of 50 to 60 candle-i^ower gas. Mr. J. B. Macarthur has shown ^ that a yield of 120 cubic feet per gallon was only equal to 40 candles, and that when the yield reached 158 cubic feet (which was at a bright orange heat) the candle power fell to 20 to 25.

While the oil is undergoing the process of distillation, there are several means of ascertaining if this is proceeding continuously, and under the most advantageous conditions. Supposing a given (pxautity of oil per hour is flowing into the retort, and a suitalde temperature is maintained, a glance at the meter will show what quantity of gas is passing, and if regular in volume. Should the latter be irregular, it indicates a stoppage somewhere, which can be ascertained by reference to the pressure gauges, whether in the tar pits, condensers, washer, or purifiers. At the same time the gas would " blow " out from the water safety-valve, alongside the main valve, and also at the siphon oil-pipe ; while, if any- thing goes wrong with the gasometer, the gauge falls, and a blow out takes place at the relief. The down pipe from the retort to the tar pit is provided with a cock, which allows a small jet of gas to escaj^e. If the escaping gas be of a light brown colour, the distillation is perfect ; if, however, it be white, the supply of oil has to be reduced, and if dark brown, or if it forms flakes of soot, the supply has to be increased, and sometimes the heat of the retort lowered. The tar must not be too thick when it overflows at the siphon on the tar catch-pit, but jet black. If it should contain oil, it is thin and flows freely, and a drop placed on a piece of white paper will produce a transparent greasy border round the tar.

No. 1 bench was worked continuously for fourteen weeks in the winter of 1887-8, during which time 8,294 gallons of oil were distilled, yielding 669,850 cubic feet of gas, or an average of 80* 76 cubic feet per gallon. This is the longest time they have been worked without requiring a renewal of the retorts or the flues. They usually last for about ten or eleven weeks.

During the year 1886 the average yield per gallon of oil was 72 cubic feet of gas, or 19,008 cubic feet per ton of oil; and, for the year 1887, 81-22 cubic feet, or 21,442 cubic feet per ton of oil. These are very satisfactory results from a work where the resoiirces

' The Journal of the Society of Chemical Industry, vol. vi. 1887, p. 811.

222 HUNTER ON THE MANUFACTURE OF OIL-GAS. [Selected

are taxed to the utmost. A certain qTiantity of gas must be manu- factured each day, and at times it is difficult to supply the demand , therefore no time can be spared in allowing the oil to be " drawn " to its utmost, as it might be, were there time. Dr. Macadam, F.C.S., while conducting some experiments,^ found that the apparatus yielded from the gallon of oil, on the first occasion 90 "70 cubic feet of gas of 62-50 candle power ; on the second occasion 103*36 cubic feet of 59 • 15 candle power, or an average of 97 • 03 cubic feet of 60 • 82 candle-power gas. In both cases the firing of the retorts was moderate, though in the second trial greater care was taken to secure uniformity of heat, and the oil was run in more slowly, so that there was more thorough splitting up of the oil into permanent gas.

Approximately it takes about 5 tons of coke, 4^ tons of dross, and 10 cwt. of coal to make 100,000 cubic feet of gas, or about 4-46 lbs. of fuel per cubic foot of gas. This includes kindling and getting up the heat in the retort.

The cost of making the gas, including oil, fuel, wages, repairs, &c. is 6s. 7'24f?. per 1,000 cubic feet.

(h) Comjyressian.

The gas is drawn from the gasometer l)y compression-pumps, and forced into cylindrical store-holders at a pressure of 150 lbs. per Square inch, or 10 atmospheres. The compression is effected by a double-acting pump,'^ which has a ram-diameter of 6 1 inches, com- pressing the gas to 60 lbs. per square inch, and a diameter of 4 inches, compressing to 150 to 180 lbs. per square inch. Although the holders are tested to a pressure of 180 lbs., or 12 atmospheres, they are seldom charged to more than 150 lbs. The compressing cylinders are kept cool by a jet of water running round them. These pumps are worked week about.

The freezing cylinder retains in cold weather, in the shape of ice, any water particles that may be absorbed by the gas in the gasometer. These water particles are much impregnated with carburetted hydrogen, and, when thawed, may be drawn oif by a cock. As a rule, there is never more than a trace of this. The cylinder cools the gas, and serves the purjwse principally of a reservoir, from which the pumps can draw without directly affect- ing the gasometer. In fact, it serves the same purjjose as a gas

The Gas Institute. Transactions for 1887, p. 41. Kevue Geuerule des Chemius do For, 1882, p. 127.

i

Papers.] HUNTER ON THE MANUFACTURE OF OIL-GAS. 223

bag to a gas-engine. In tlie engine-room there is a suction-gauge, for sliowing if the pumps are drawing the gas regularly. There are relief cocks on the pumps, which shut oif the high-pressure gas from the cylinder valves after the store-holders have been charged, and the column of gas in the pipe is allowed to pass back into the gasometer.

The store-holders are mounted on trestles, each pair being con- nected underneath the front end by a small pipe which leads to a recipient encased in a box of sawdust for the hydrocarbon, from whence it is drawn off into a drum. The pressure of the gas in the holders drives oiit the hydrocarbon into the recipient, and thence to the drum ; but, as a certain quantity of gas always passes with the hydrocarbon into the drum, provision is made for this gas to pass from the drum into the gasometer. The hydro- carbon heing the principal constitiient of the gas, any diminution affects the illuminating power of the gas. On an average it loses fully 5 candle power by compression. The compression causes the deposit of the hydrocarbon at the rate of about 1 gallon per 1,000 cul)ic feet of gas. It gives off an inflammable vapour at a temi)era- ture of less than 45' Fahrenheit. It is, accordingly, rather a dangerous substance to work with, and is generally drawn off from the drum into smaller ones for transmission to manufacturing chemists, on a dull, moist day.

II. Application to the Lighting of Eailway-Carriages.

From the works to Bridge Street Station (}r mile), and also to Central Station (.V mile), jiipes are laid so that trains can be charged with gas while standing in either of these stations. A pipe is also laid to Giishetfauld's carriage-shed ^ (^ mile), and carriages can be charged there before being taken out and mar- shalled. Filling cocks are placed every 34 feet, so that a carriage may come near each cock ; a 30-foot length of india-rubber hose is attached to the filling cock on the carriage, and the pressure-gauge records the volume of gas supjilied. It takes three minutes to couple the hose and charge a standard carriage carrying six lamps, if they be empty, but this is seldom the case.

Plate 5, Fig. 6 shows the general arrangement of the pipes and fittings on an ordinary carriage. Each carriage is provided with one or two recipients or cylinders, of a size proportionate to the number of lamps it carries. The recijiients are made of y\-inch steel

' Miuutcs of Proccediugs lust. C.E., vol. Lxxxvii. p. 397.

224 HUOT^ER ON THE MANUrACTURE OF OIL -GAS. [Selected

plate, carefully bent. The joints are both, riveted and soldered, and the flanged ends are then fitted in, fastened by screws, and soldered over. The cylinder is afterwards tinned both inside and outside, so that it is perfectly gas-tight. It was at one time imagined that dangerous explosions might result from the presence of so much gas under each carriage, in the event of a railway accident. In order to settle this point, an experiment was made with a recipient 6 feet in length and 16 inches interior diameter, made of ^-inch plate. This recipient was filled with gas at a pressure of 6 atmo- spheres (90 lbs.), and placed under a scaffolding, supported at its ends with a fire on each side. A weight of 400 lbs. was dropped on it from a height of 15 feet, the weight falling exactly on the riveted seam. At the first blow it was merely bent in, and became slightly loosened about the rivets. A pointed piece of iron was then placed on the recipient, and the weight allowed to drop as before. This made a hole 5 inches in diameter, and, in opposition to the effects which would have been exhibited by the explosion of a steam- boiler, the gas simply escaped from the hole, and, being ignited by the fire, burned with a high flame. There was nothing of the nature of an explosion, and scarcely any noticeable concussion of the air at a few paces distant. The test sufiicit n ly proved the com- paratively harmless nature of the small volume of gas enclosed in the recipient, and that the gas can escape only through external jxjwer or an accident. This has been confirmed in several cases where a broken tire made a hole in the recipient, the gas escaping, harmlessly into the air before the train had come to a standstill. Each recipient is tested under hydraulic pressure before being attached to a carriage.

The size and position of the recipients vary with the different modes of under-framing, and also the number of wheels for each type of carriage. For a standard carriage they are fixed longi- tudinally to the under-framing, one on each side, and connected by a pipe. The filling cocks are fixed outside the framing on each side, as also the pressure-gauge. A regulating or governing valve ^ reduces the pressure of the gas after it has left the recipient on its way to the lamps. The valve consists of a cast-iron pan 12 inches in diameter and 6 inches iil depth. The upper part is closed by a membrane of specially prepared leather, to the centre of which is fastened a rod connected at its lower end with a lever controlling the regulator-valve. This lever is also controlled by a double-leaf spring acting in opposition to the membrane ; and by this means

' The Etigtneer. vol. xlix. p. 293.

Papers.] HTTNTER ON THE MANUFACTURE OF OIL-GAS. 225

the regiilator-valve is rendered independent of the movement of the carriage, so that accidental extinction of the light never takes place. The gas passes the valve, which is adjusted for a pressure of only that due to a -^-inch column of water ; and when the pres- sure under the membrane is reduced helow this, the valve is closed, and the jirocess of opening and closing is repeated until a balance between the admission and the consumption is obtained. These re- gulators are adjusted according to the number of lamps which they will supply ; thus a foiir- or five-lamp regulator is adjusted to 30 millimetres, or 1^\ inch.^ From the regulator a low-pressure pipe is carried up the end of the carriage on to the roof, and connects the several lamps. On the end of the carriage there is a by-pass valve, worked by a draw-rod, by which in the event of a train standing emj^ty at a station, there being no occa- sion for the gas burning, the lights can be shut down p^

without being extinguished. There is likewise a con- cussion-box, which always contains a supply of gas, so that, in the event of a sudden pull of the draw-rod, the lights are not extinguished, as they otherwise might be. The pipe is simply cut ; both ends are closed, and the pipe is perforated with a few holes through which the gas escapes, and fills the box, making it a g^s ^^^^^^^^^f,^^^ chamber, from which it is again drawn oft'. Above this cussios-Box. is the main cock, which shuts off" the gas from, or admits it to, the whole carriage.

The recipients are charged to 6 atmospheres (90 lbs.) for thirty- six hoiirs' burning ; and no carriage is allowed to leave Central or Buchanan Street stations with a lower indicated pressure than 2 atmospheres, or for twelve hours' burning.

In Fig. 3, which is a section of a standard lamp, the arrows show the direction in which the foul air escapes, and the fresh air enters to support combustion. It will be observed that sudden gnsts of wind cannot directly affect the light ; that, the globe being- perfect inside, the opening of the carriage-door has no eff"ect uj^on the light ; and that sparks from the engine or other matters cannot get into the inside of the lamp. The rich nature of the gas renders the adoption of a special burner necessary ; because, the greater the illuminating jjower of a gas, the smaller will be the consumption to maintain a light of given brilliancy. The burner, a small fish- tail, made of "speckstein" or steatite, gives a well-shaped, steady

All the pipes between the recipients and the filling cocks and the regulator are under high pressure, and arc made of best special i)ipc of y^^^-inch bore and 5-iuch outside diameter.

[tHK INST. C.E. VOL. XCV.] y

I

226

HUNTER ON THE MANUFACTUBE OF OIL-GAS.

[Selected

flame. It is easily cleaned. The funnel and all the casing is^H made of best tinned copper, and the body of cast-iron. Each jet ^^ has a thumb-cock, so that any light can be turned off, while the others remain burning in the same carriage. There is also a regulating spindle, which can be worked with a special key from the inside of the carriage, for raising or lowering the flame. The connecting pipe between each lamp is screwed into the socket, a. For saloon and such types of carriages the globe of the lamps is hinged on the inside, to admit of the passengers lighting them at pleasure. They also have hoods for shading off the light.

The luminary being a permanent gas, it can remain for any length of time in the recipients without deteriorating, either in

Fig. 3.

Section of Lamp.

Scale i.

illuminating power or in pressure. Dr. Macadam says : ^ — " I have also examined samples of the gas taken by me from various cylinders, where the gas had been stored for several months under a pressure of 10 atmospheres ; and in all cases the gas was found to be practically equal to the above quality,^ and hence was of a permanent cha- racter." The permanency of the gas may be further illustrated by the fact that the carriages, which are charged at Gushetfauld's shed or at Central Station, maintain the same pressure on the gauges as

' The Gas Institute. Transactions for 1887, p. 45.

^ The quality of the gas obtained was very high, owing to its containing a large percentage of heavy hydrocarbons, of which there were respectively 39 -25 and 37-15 per cent., or an average of 38-20 per cent.

Papers.] HUNTEK ON THE MANUPACTUKE OF OIL-GAS. 227

those at the works, and that after having travelled J mile through pipes to each place.

The disadvantages of the old system of lighting carriages with oil-lamps are well known. The removal of dirty lamps to the lamp-room, cleaning them, filling the cisterns and replacing them in the carriages, rendered necessary the employment of a staff of men at terminal stations. Again, the breakage of globes, and damage to the framing, formed a heavy item of cost per annum. This is now obviated, as, the lamps being fixtures, all that is required when cleaning the lamps is to lift the cover, remove the reflector, and lift the gas-bracket. The lamp being now open, the inside of the globe, the enamel on the reflector, and other parts can be cleaned and all replaced.

Taking the present price of oil as Is. lid. per gallon, and the cost of gas as Gs. 7-24(Z. per 1,000 cubic feet less the residual pro- ducts, the oil will burn for one hundred and eighty-four hours, while the gas will burn thirteen hundred and thirty-three hours, and the cost would be 0-125(Z, and 0-05944fZ. respectively per hour. This is a saving of 0-06556d. per hour, which, for 1,000 cubic feet, represents a saving of 87-391, or 110-29 per cent. The value of the residual products, hydrocarbon and tar, for 1887, amounted to Is. l-24:d. per 1,000 cubic feet.

The cost of fitting a standard carriage with six lamps, including recipient, regulator, gauges, valves, pipes, &c., is £33.

The tests to determine the illuminating power of the gas, as given in the Appendix, were undertaken with the view of showing the value of the gas as used in the carriages. The fittings on the photometer were supplied by Pintseh's Company.

In conclusion, the Author begs to acknowledge his indebtedness for information to Mr. Stewart Kershaw, Manager for Scotland of Pintseh's Patent Lighting Company, Limited ; and to Mr. Samuel Stewart, F.I.C, Chemist, Stores Department, Caledonian Eailway, who supplied the photometric tests.

The Paper is accompanied by three tracings and several dia- grams, from which Plate 5 and the Figs, in the text have been prepared.

[AppENorx.

228 HUNTER ON THE MANUEACTUKE OF OEL-GAS. [Selected

APPENDIX.

Experiments to test the Caxdle power of Gas.

The gas was taken from a recipient charged to over 10 atmospheres.

Specific gravity of the oil (Clippeu's " Straiton ") at 60^ Fahr. . . 848-2

Weight of 1 gallon of oil 8-482 lbs.

Number of gallons of oil per ton 264

Flashing-point of the oil, Fahr. (close-test) 268''

Firing-point of the oil, Fahr- 294°

Specific gi-avity of the hydrocarbon at 60° Fahr 845-2

Weight of 1 gallon of hydrocarbon 8-452 lbs.

Firing-point of the hydrocarbon, Fahr. (below) 45°

Gas.

Time of test

Temperature, Fahr

Pressure of the gas in tenths of an inch

Eeduction in pressure, from 7 - 35 to 5 - 75 atmospheres .

Gas consumed per hour, cubic foot

Illuminating value of the gas as taken by i^hotometer, 1

and reckoned in standard sperm caudles, consuming } 8 - 25 8-25 8 • 0 8-0

120 grains per hour )

5 cubic feet are equal in intensity to 55*7 candles . .

Illuminating value of 1 cubic foot of the gas in gi-ains\

e ( 1,336

of sperm } '

Illuminating value of the gas from 1 gallon of oil in i

lbs. of sperm / ^^'^^

Total illuminating value of the oil yielded by 1 ton of 1

oil, and given in lbs. of sperm / ' "^^

Volume of oil running into each retort per half-^

minute, oz J *

A''olume of gas yielded by each retort per minute, as~i

shown by the meter, cubic feet j

Total volume of gas yielded per hour b}' the two'i

retorts, cubic feet /

Average yield per gallon, cubic feet 80 - 48

Which, calculated to the ton of oil, gave, cubic feet . 21 ,246-72

min. 0	mms. 15	mins. 30	mins. 45
63°	63°	64°	64°
7-5	7-75	7-25	7-5
	1-	60
	0-	74

960

Papers.] SMITH ON HUKST's TRIANGULAR PRISMATIC FORMULA. 229

(Paper No. 2,336.)

" Hurst's Triangular Prismatic Formula for Earthwork compared with the Prismoidal Formula."

By James William Smith, Assoc. M. Inst. C.E.

In this communication references to Hurst's " Handbook of For- mulas, Tables, and Memoranda," ai)23ly to the 12th edition, 1879 ; and references to Molesworth's " Pocket-Book of useful Formulae, and Memoranda," apply to the 19th edition, 1879.

In the first of these Handbooks Hurst's Triangular Prismatic Formula is given at p. 257 as follows : —

" To measure the solidity of earthwork over large areas of irregular depth : — Divide the surface into triangles, and multiply the horizontal area of each by one-third of the sum of the vertical depths taken at the angles, and the result will equal the solidity."

This is quoted by Molesworth, p. 52, but without the following im^portant note by Hurst :— " The surfaces of the triangles must be true planes, or they must be taken so small as to approximate to planes."

The well-known Prismoidal Formula is given by Molesworth at p. 46 as follows : —

[Sum of areas of both ends -|- (area of middle X 4)] x length

6 *

The convenience and simplicity of Hurst's triangular prismatic formula is incontestable ; but it has been looked upon with dis- favour, owing to the diiferent results obtained according to the manner in which a quadrangular figure may be divided by a diagonal into two triangles, and to the difference between either of those results and the result obtained by the prismoidal formula. The object of this Paper is to point out the reason of these divergences and to provide a remedy.

Hurst's formula is undoubtedly mathematically correct, provided, as he takes care to explain, the surface of each triangle is a true plane, and any three points of a surface can be connected by a true plane. But in practice, levels being taken for convenience at stated intervals, on parallel lines at stated distances from one

230 SMITH ON hurst's TRIANGULAR PRISMATIC FORMULA. [Selected

anotlier, the surface to be dealt with is primarily divided into quadrangular figures ; and the subsequent division into triangles is purely arbitrary. Of course, if the engineer knew that a line between any two angles would touch the surface all along, that line would represent the proper diagonal to be drawn on the plan, and there would be no reason to further investigate the ap- plication of Hurst's method ; but, as a rule, he cannot know this.

As, however, except under the rare condition of the whole surface of the quadrangle being a true plane, the result obtained by Hurst's method, if the quadrangle be divided by a diagonal between any two angles, will diifer from that obtained by dividing the quadrangle by a diagonal between the other two angles, a mean result, easily arrived at, is a desideratum. Though such a mean may not be based upon a strictly mathematical principle, yet if only a mean, it should be welcome ; and since it can be shown that the mean corresponds with the effect of the prismoidal formula, which has received universal acceptance, it should be deemed entitled to general adoption.

The mean is most easily obtained if the area be divided into parallelograms, and it is not often in practice an area would be otherwise divided. In these cases the earthwork (whether filling or excavation) would consist of prisms whose ends are parallelo- grams. The formula for the solidity, giving the mean of the two ways in which Hurst's formula may be applied, is simply : —

.„. Sum of the four depths

(Z) ^ X area.

4

In the examjiles which follow, the depths are enclosed within circles, and are taken more divergent from one another than they would generally be in practice, in order to more severely test the applicability of the above formula.

ExamjjJe 1. — Plan of rectangle, area = 72.

Eeqiiired the solidity of Fig. 1, depth at each angle being given.

{a) By Hurst's formula, if divided as in Fig. 2 : —

Solidity = { -\-

Total solidity = 732

I

Papers.] SMITH ON HURST's TRIANGULAR PRISMATIC FORMULA. 231 (h) By Hurst's formula if divided as in Fig. 3 : — 9 + 6 + 18 _ 72 Solidity = { +

X y = 396

18 + 11+9 72 ^^^

— ■ — X -TT = 456

3 2

Total solidity = 852 (c) Mean of a and & = 792.

Fig. 1. Fig. 2.

(d) By new formula (quadrangular) Z : —

18 + 11 + 9 + 6

Solidity =

X 72 = 792, i.e., the mean of a and 6.

(e) By prismoidal formula, observing the following directions in Hurst's Handbook, p. 257 : —

" The sections are to be taken parallel to each other, and the area of the middle section to be calculated from the mean dimen- sions, and not by averaging the areas of the ends."

The figure then becomes as follows, the depths (12) and (10) being calculated respectively as the means of 18 + 6 and 11 + 9 : —

Fig. 4.

e -■■-

18 + 11

2

6 + 9

2 12 + 10

X 6 = 87 = sectional area at one end, X 6 = 45 = „ „ „ other end,

X 6 = 66 = „ „ in middle.

232 SMITH ON HUKSt's triangular prismatic formula. [Selected

87

+ 45

+ (66x4=) 264

396 X 12 the length.

6;4752

Solidity = 792 i.e., the same as by the quadrangular for- mula (Z).

The prismoidal formula might be applied by taking the sections in the other direction (i.e. longitudinally), but the result would be ■precisely the same.

Example 2. — Plan of rhombus, area = 60.

Eequired the solidity of Fig. 5, depth at each angle being given.

Fig. 5.

(a) By Hurst's formula, if divided as in Fig. 6 : — Solidity = 210 + 190 = 400. Fig. 6.

(h) By Hurst's formula if divided as in Fig. 7 : — Solidity = 220 + 250 = 470. Fig. 7.

(cj Mean of a and h = 435.

(d) By new quadrangular formula (Z) : — ■

Solidity = i^±l±l±i,X 60

435, i.e., the mean of a and h.

Piipers.] SMITH ON HURSt's TRIANGULAR PRISMATIC FORMULA. 233

(e) By prismoidal formula, the depths (8^) and (6) heing cal- culated, the former as the mean of 10 -(- 7 and the later as the mean of 4 + 8 : —

Fig. 8.

>> - -- TO ■ — - >.

Sectional area at one end = 70 ; „ „ at other end = 75 ;

„ „ in middle = 72^.

Solidity = 435, i.e., the same as by the quadrangular formula (Z).

The same result would have been obtained if the sectional lines had been taken as 6 long instead of 10, and the length of the figure as 10 instead of 6.

When the area of the surface is a trapezoid, not often occurring in practice, a somewhat different method has to be adopted.

Taking the two parallel sides of the trapezoid, as the respective bases of the two triangles into which the trapezoid may have been divided, the formula, to obtain a mean result, will be as follows : —

Mean of the two depths opposite base |

/^rx <.,.■,. + sum of the two depths at the base I x < f ^®^ i (Y) Solidity = -!- — -i ^ 1 triangle

o J

The solidity in each triangle must be calculated separately. Exam^ple 3. — Plan of trapezoid. Fig. 9, area of smaller triangle = 24, area of larger triangle = 60.

Required the solidity, the depth at each angle being given.

234 SMITH ON hurst's triangular prismatic formula. [Selected

(a) By Hurst's formula, if divided as shown by the dotted diagonal : —

9 + 11 + 3 ^ ^ X 24 = 184

3 + 7 + 9

X 60 = 380

Total solidity = 664

(h) By Hurst's formula, if divided thus : — Fig. 10.

11 + 9 + 7

11 + 3 + 7

X 24 = 216

X 60 = 420

Total solidity =636

(c) Mean of a and h = 600. (d) By new formula (Y) : —

7 + 3

+ 11+9

X 24 = 200

11 + 9

+ 3 + 7

X 60 = 400

Solidity = 600, i.e., the mean of a and 6.

(e) By prismoidal formula, the depths 7 and 8 being calculated,

Fig. 11.

the former as the mean of 1 1 + 3, and the latter as the mean of 9 + 7 :—

Papers.] SMITH ON HURSt's TRIANGULAR PRISMATIC FORMULA. 235

Sectional area at one end = 40 ; „ „ at other end = 50 ;

„ „ in middle = 52^.

Solidity = 600, i.e, the same as by the formula (Y).

In the case of a trapezium, no simpler method has been found than to take the mean of the two results, arrived at from the application of Hurst's formula to the two triangles obtained by drawing the diagonal in one direction, and to the two other triangles obtained by drawing the diagonal in the other direction. The trapezium is, however, a figure to be avoided in areas of earthwork ; the surface should be divided into parallelograms and triangles, the triangles being formed only when necessitated by the irregular direction of the boundaries. Formula (Z) should be applied to the parallelograms, and Hurst's formula to the triangles.

A trapezium may be divided into a trapezoid and a triangle, an additional level being taken on the ground, at the point where the line drawn parallel to one side of the trapezium from one of the angles cuts the side opposite to the angle. Formula (Y) may be applied to the trapezoid, and Hurst's formula to the triangle.

When a surface is divided into parallelograms of equal area, the application of formula (Z) is much simplified by placing the depths in columns ; the depths which belong to only one parallelo- gram in one column, those which are common to two parallelograms in another, and those which are common to four parallelograms in a third column. The sums of the second and third columns are multiplied by 2 and 4 respectively, and the products added to the sum of the first column, and then multiplied by one-fourth of the area of one parallelogram, give the total solidity.

In calculations for earthwork, it will be found preferable to use the actual reduced levels, or those levels minus an easily deducted whole number, rather than depths, each one of which is a probable source of error, as it involves a more or less complicated subtraction ; besides, when the finished surface is to be sloped, a separate calculation is necessary to ascertain the reduced level of the finished surface at each point.

For instance, suppose that the lowest reduced levels of a surface to be excavated are between 80 and 90 on the datum, then if 80 be deducted from each reduced level, and the remainder used as a depth, the result will be the solidity if the surface be excavated down to 80 on datum. The difierence between 80 and the reduced level to which the surface is to be finished, if used as a depth in

236 SMITH ON hurst's triangular prismatic formula. [Selected

connection with the whole area, will, in one single calculation, give the contents to be either added to or deducted from the previously ascertained solidity, as the case may require, in order to arrive at the total or net solidity. Even if the finished surface is to be sloped, the mean reduced level will generally be seen almost at a glance, and one calculation for the addition or deduction will be sufficient.

It is obvious that quantities of filling may be arrived at in a similar manner.

Besides the benefit of avoiding a multiplicity of minor calcula- tions afforded by this method, there is also the advantage of all the actual reduced levels used in the computations being preserved and recorded in the measurement book. In the use of depths only, there is no record of the reduced levels.

The Paper is illustrated by sketches in the manuscript from which the eleven Figs, have been prepared.

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 237

(Paper No. 2339.)

" Alpine Engineering." By Leveson Francis Vernon-Harcourt, M.A., M. Inst. C.E.

The passage of the Alps has always been a fascinating object to mankind — not indeed from their forming the most insurmountable harrier in the world, for the Himalayas, the Andes, and other moimtain-ranges exceed them in height and inaccessibility — but owing to their being the chief obstacle separating some of the most fertile regions of the earliest civilized quarter of the globe. The crossing of the Alps by armies under Hannibal and Napoleon, has invested the passes which they traversed with an historical interest ; whilst the skill of engineers has crowned them with marvels of road-making and mountain railways, and pierced their iimermost recesses with the principal tunnels of the world, (riate 6.)

Alpine Passes and Eoads.

Plate 6, Fig. 2.

Some of the Alpine passes were known and used as means of communication in very early times ; but the first recorded instance of the conveyance of an army across the Alps was when Hannibal, marching from Spain for the invasion of Italy in the second Punic war, siirmounted the pass of the Little St. Bernard. Several of the passes were known in later times to the Romans, who desig- nated the highest point of the ascent as " Mons," from which the prefix "Mont," given to several of these passes, has been derived, referring to the summit of the pass, and not to neighbouring peaks of greater altitude. Caesar ai)pears to have crossed Mont Genevre with an army, to check the incursion of the Helvetii into Gaul ; and vestiges of Roman works maybe traced on the pass. ^ Augiistus made a carriage-road across the Little St. Bernard, which Avas allowed to fall into decay ; but nevertheless, being one of the easiest of the passes, it could be traversed by artillery and light carriages ; and it possesses the advantages of its approaches and

' "Passes of the Alps." W. Brockcdoii, vol. i.

238 VERNON-HAECOUKT ON ALPINE ENGINEERING. [Selected

summit, 7,192 feet above the sea, not being exposed to avalanches. The Brenner pass, on the route between Innsbruck and Yerona (Plate 6, Figs. 2 and 4), being the lowest of the passes across the principal Alpine chain, only 4,588 feet above sea-level, was the first of these passes regularly used by carriages; the first record of this pass dates back to the year 13 b.c. It was by the Brenner that Attila led his invading army into Italy; and this was the route of later barbarian incursions. The Austrians formed a road over it in early times, to provide proper means of communication with their jDossessions in Lombardy ; and it was made practicable for carriages in 1772. The Col di Tenda, with its summit 6,158 feet above sea-level, the most southern of the Alpine passes, lying on the road from Turin to Nice, was made j^racticable for carriages in 1789, and was improved by Xapoleon. A new road has been formed, passing through a tunnel, 2^ miles long, lighted by elec- tricity, in place of the numerous zigzags by which the old road surmounts the pass. These two passes, the Brenner and the Tenda, lying at almost the extreme limits of the Alpine range, possessed the only two carriage-roads previous to the commence- ment of a carriage-way across the Simplon in 1801. The pass of the Great St. Bernard, situated near the highest peaks of the Alps, with Mont Blanc to the west, and the Matterhorn and Monte Eosa to the east, was traversed by the Romans in 100 B.C. ; it was more used after the foundation of Aosta in 26 B.C. ; and the road was improved by Constantine in 339. It is now best known from its hospice at the summit of the pass, 8,120 feet above the sea, and the bravery of the monks and their dogs in rescuing travellers from the snow ; and it has been rendered famous by the passage of Kapoleon across it, in May 1800, with his army and artillery on his way to the campaign of Marengo. Owing to the difficulties experienced in this march, when the ordinary obstacles of the narrow and rugged tra-ck were enhanced by the quantities of snow •still encumbering the pass, and the danger of avalanches, Napoleon projected several roads across the Alps, to which he was also prompted by the desire to provide easier communication befrvi'een France and the annexed kingdom of Italy. The Simplon pass was selected for the route of a carriage-road across the Pennine Alps, constructed in 1801-6, which, though one of the most difficult of the Alpine passes for the formation of a road, is much lower than the Great St. Bernard (6,590 feet as compared with 8,120 feet), and more accessible ; and it provided a much-needed passage about midway in the Alpine range (Plate 6, Figs. 2, 5 and 9). It is, however, exjwsed to avalanches towards the summit ; and several

Papers.] VERNON-HARCOUKT ON ALPINE ENGINEERING. 239

galleries and refuges liave been provided for shelter. The pass of the Mont Cenis, with an altitude of 6,772 feet, lying on the direct road between Lyons and Turin, and situated about half-way between the Simplon and Tenda passes, has been the most fre- quented road between France and Italy (Plate 6, Figs. 1, 2 and 6) ; portions of the road were very difficult to traverse, even on foot, in early times ; but the second Duke of Savoy made the worst portion (known as les echelles) practicable for carriages in 1670; and Napoleon, by large works, begun in 1803 and completed in 1810, made the road passable at all seasons. A road was also constructed by order of Napoleon, between 1802 and 180-4, across the Mont Genevre, on the route between Grenoble and Turin, situated a little to the south of the Mont Cenis, and rising 6,102 feet above the sea ; but the road contemplated by Napoleon across the Little St. Bernard, to connect Grenoble and Aosta, was not carried out, thoiigh a carriage-road has recently been constructed. The St. Gothard is the next pass of importance to the east of the Simplon (Plate 6, Figs. 1, 2, 5 and 7) ; it lies on the direct line between the Lakes of Lucerne and Maggiore, and rises only 6,936 feet above sea-level ; but though used by barbarian hordes in invading Italy, and the scene of several conflicts between the French and Austrians in the campaign of 1799, it was not made available for carriages till 1832. Accordingly, in the early part of the century, the traffic was diverted to the new and more acces- sible routes, till the carriage-road, commenced in 1820 by the Cantons of Uri and Ticino, restored its importance. The road is in places much exposed to avalanches ; snow-storms and ava- lanches are most prevalent on the southern side, and the great snow-drifts do not always entirely disappear in the summer. Several passes intervene between the St. Gothard and the Stelvio, further to the east. The Bernardino pass, known to the Eomans, and having an elevation at the summit of 6,769 feet, was made accessible for carriages by a road constructed in 1818-24 by the inhabitants of the Grisons Canton. The Spliigen pass, a short distance to the east of the Bernardino, rising to an altitude of 6,945 feet, was one of the most frequented roads in the fifteenth century (Plate 6, Figs. 2 and 5) ; and both it and the Bernardino were made use of in the campaigns of Napoleon, the passage of the Spliigen by the army of reserve under Macdonald, in the winter of 1800, being one of the most difficult marches ever accomplished. A carriage-road was formed over the Spliigen by the Austrian Govern- ment in 1819-24. These two roads provided more direct com- munication between western Germany and Lombardy ; and roads

240 VERNON-HARCOUET ON ALPINE ENGINEERING. [Selectea

over the Julier (siimmit-level 7,503 feet), the Maloya (5,942 feet), and the Bernina (7,658 feet), have since been added (Plate 6, Fig. 2). The road over the Julier, completed in 1827, is little exposed to avalanches ; Augustus constructed a military road over the Julier and Maloya passes, and the road over the Bernina is the second highest in Europe. The carriage-road over the Stelvio pass is the highest in Europe, reaching at its summit an altitude of 9,213 feet above the sea — about 800 feet above the normal limit of perpetiial snow in those parts, and more than 1,000 feet higher than the summit of the Great St. Bernard. The construction of this road, in 1820-25, was dictated by political exigencies, for connecting the German and Italian dominions of Austria without passing through foreign territory, as its height, the heavy works required to protect the approaches to the summit from avalanches, and the cost of maintenance, would have precluded its selection for purely commercial purposes.

The road over the Arlberg pass, though not crossing the main Alpine chain, is of considerable importance in aftbrding direct communication between Switzerland and the Tyrol (Plate 6, Figs. 1, 2, 5 and 8). The summit of the pass, 5,902 feet above sea-level, lies on the ridge separating the basins of the Ehine and the Danube. Several other roads and bridle-jiaths have been formed across the Alps ; amongst the latter, the most remarkable is the path over the Gemmi pass, formed in 1736-41, which descends an almost perpendicular cliff, 1,660 feet high, to Leukerbad, a distance of over 2 miles, by zigzags cut into the rock.

Great skill has been displayed in the tracing of the Alpine roads, through rugged valleys, exposed in places to avalanches and land- slips, and crossing torrents which are subject to raj^id floods. The torrents, however, are confined, for the most part, in their rocky channels ; and the bridges, by which the roads cross from one side of the valley to the other, to avoid formidable obstacles, are usually considerably above the reach of the waters. The grandeur, indeed, of the adjacent moimtains, and the depth of the valleys, enhance the appearance of these works ; for the celebrated Devil's Bridge, on the St. Gothard road, crossing the Eeuss torrent at a consider- able height has a clear span of only 26 feet. Avalanches are avoided by carrying the road through covered galleries, or under overhanging rocks ; projecting points are pierced by tunnels ; steep slopes are skirted by terraces ; ravines are crossed by high viaducts ; and the gTadients are regulated, where necessary, by means of zigzags. The rugged, bleak, and desolate regions through which these roads pass, the steepness of the sides of the

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 241

deep valleys, and the short period during which works can be carried on in the higher portions of the passes, give an importance to these works, which roads formed over hard rock would not otherwise possess ; and most of these works were executed at periods when rock excavations and tunnels were exceptional and difficult engineering works, in the absence of modern appliances.

The great development of travelling, and the demand for rapid transit, produced by the introduction of railways, soon led to schemes for extending the facilities of railway communication across the barriers of the Alps. Locomotives, however, of the ordinary type could not ascend the steep gradients, nor turn the sharp corners by which the Alpine roads surmount the passes ; and accordingly, the crossing of the Alps by railways necessitated works, the execution of which was delayed by their unprecedented magnitude and cost.

Semmering Eailway. Plate 6, Figs. 1 and 3 ; and Plate 7, Fig. 1.

The first railway which surmounted the Alps does not cross the main Alpine chain, but traverses the outlying Styrian Alps at the Semmering pass, which is only 3,248 feet above sea-level. The object of the Semmering railway was to connect Vienna with its seaport, Trieste ; and the most practicable route lay along the Semmering pass, whereby any necessity of constructing a long tunnel was avoided, though the gradients and works were neces- sarily heavy and costly. The scheme was first considered in 1842; a definite line was jiroposed in 1844, which was finally approved in 1848, when the works were commenced; and the line was completed and opened for traffic in 1854,^ at a cost of about £98,000 per mile for a double line throughout.

The railway ascends the north-eastern slope of the pass, from Gloggnitz, by contouring the valleys in a winding course (Plate 6, Fig. 3), going through fourteen tunnels, and passing over sixteen viaducts,^ before reaching the Semmering tunnel at the summit,

' "Atlas Pittorcsque du Chemin de For du Semmering," Carlo di Gliega, p. 10, aud plates 1 and 2.

* The tunnels are from 14^ to G64 yards long, with a total length of 3,113 yards. The viaducts are from 33 to 249 yards long, with a total length of 1,620 yards, and from 36J to 150 feet maxima heights; and four of them have been constructed with two tiers of arches.

[the INST. C.E. VOL. XCV.] 11

242 YEENON-HAECOURT ON ALPINE EKGmEERING. [Selected

1,562 yards long (Plate 7, Fig. 1). This tunnel was constructed by the aid of nine shafts, of which five were left open, after completion, for ventilation. The summit-level of the line, 2,892 feet above sea-level, is inside the Semmering tunnel ; and the rise from near Bayerbach, where the steep gradients commence, is 1,305 feet in 13.V miles, making the average rise 1 in 53j. The average, however, is reduced by the introduction of level portions, or gentle gTadients, at the stations and towards the summit ; for the ruling gradient of 1 in 40 extends over 5V- miles, whilst the next most frequent gradient of 1 in 45 occujiies 3j miles. The south-western slope is much less rugged, so that without any special contouring, no tunnels or viaducts were required to bring the line down to Murzzuschlag, nearly 7.^ miles from the summit. As the difference of level in this distance is 710 feet, the average inclination on this slope is 1 in 55.V, which is less than on the oj)posite side, though the descent is more rapid near the summit. The worst gradient on this side is 1 in 41i for 700 yards; three gradients intermediate between this and 1 in 45 have a total length of over 1 mile ; and the most frequent and longest gradients are 1 in 45 and 1 in 47, extending together over 3j miles.

The length of the line between Gloggnitz and Murzzuschlag is 25^ miles ; and the line is curved for veiy nearly half this distance. Very few of the curves exceed 1 9 chains in radius ; and there are ninety-six curves of this, or smaller radius extending over 11 J miles, of which the sharjiest curves, thirty in number, have a radius of only 9^ chains for a total length of 4^ miles.

The gTadients of this line, though not quite unprecedented, were unusual on main lines in those days ; and as they extended over several miles almost continuously, and were combined with very sharp curves, special locomotives were required for traversing this portion of the line. From the results of trials of various locomotives submitted for conducting the traffic, Baron Engerth was enabled to design a locomotive suited to the requirements of the line.^ The passenger locomotives were able to draw trains of 100 tons, exclusive of the engine, up the inclines at a speed of 11^ miles an hour, and of 115 tons at 9.^ miles an hour, the respective speeds arranged for passenger- and goods-trains ; whilst the goods loconiotives, with six wheels coupled, could draw up trains of 115 and 130 tons at similar speeds res^^ectively ; and a

* " Die lokomotive dcr Staats Eiscnbahn iiber den Semmering," W. Eugcrtli ; Minutes of Proceedings lust. C.E., vol. xv. p. 353 ; and Zeitsclnift des Oester- rL-ichischeu lugeuieur- und Arcbilekteu-Vereins. 1854.

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 243

modified locomotive subsequently introduced, with, eight wheels coupled, could draw up a train of 175 tons at the lower speed.^ These loads had to be reduced by one-fourth in bad weather, fogs, or great cold ; but they could be somewhat increased in ascending from Murzzuschlag, owing to the more favoured situation of the south-western slope, and the shorter and slightly better gradient on that side — a fortunate circumstance, as the greater traffic is in that direction towards Vienna. The goods-trains, weighing generally 350 tons without the engine, had originally to be taken up in three divisions ; but with the modified locomotives, they could be taken in two divisions. For some time it was considered that the curves of only 9^ chains on the Semmering, sometimes reversing without the intervention of a straight portion, precluded the employment of an engine behind the train, to assist the engine in front, and thus draw and push up a goods-train of 350 tons in one operation. Some experiments, however, showed that with the helj:) of a counter-pressure steam, or a vacuum-brake to control the motion, an engine could be placed in the rear of a train wdthout any disadvantage. Accordingly, since 1869 this system has been adopted on the Semmering line, obviating the inconvenience of dividing the goods-trains, and avoiding accidents from the breakage of couplings.

The cost of traction is necessarily much heavier on the Semmer- ing section than on the rest of the line ; but it has been gradually reduced by increasing the grate-area of the engines, and thereby diminishing the consumption of fuel, and also by improvements in the engines, the better utilization of their power, and the reduc- tion in the price of materials, advantages which have been shared by the rest of the line. Thus, in 1860, the cost of traction, including maintenance of rolling-stock, which was 2s. 5d. per train-mile on the remainder of the southern lines of Austria, amounted to 3s. 7§cZ. for passenger-trains, and 10s. llfL for goods- trains, drawn up in three divisions, on the Semmering section. These prices were reduced by 1865 to Is. 3d. per train-mile on the whole of the southern lines, and to 2s. 4d. and 4s. Sd. for the passenger- and goods-trains respectively, on the Semmering, the goods-trains being separated into two parts. By 1868, the con- tinuous reduction in price appears to have reached its limit ; for though the cost per train-mile was Is. 7^d. in 1877 on the Semmering, as compared with Is. 9^d. in 1868, it had risen

' Memoires de la Socie'te des Ingeuieurs Civils, 1862, p. 117; 1861, p. 208; 1865, p. 241 ; 1866, p. 140; 1867, p. 359; and 1868, p. 477.

u 2

244 VERNON-HAECOUET ON ALPINE ENGINEERING. [Selected

higlier than in 1868 in some of the intervening years ; and on the other soiathern lines, it was Is. 2},d. in 1877, or higher than the price of Is. Id. per train-mile in 1868. As, however, the mean load of the trains rose from 125 tons in 1868, to 132 tons in 1877 on the Semmering, and from 193 tons to 227 tons on the other lines, the ton-mile affords a fairer standard of the varia- tion in the cost of traction. ^ The cost per ton-mile decreased from 0-173^. in 1868, to 0-152(7. in 1877 on the Semmering, and from 0 069(?. to 0-064d. on the remainder of the southern lines; but was higher than in 1868 in several of the intervening years. The above fignires indicate that, as regards cost of traction, the inclines of the Semmering section are equivalent to the addition of nearly 50 per cent, to the length of this section to the total length of the southern lines, or of 26 miles X Ij = 39 miles, in a total length of l,035j miles, comprising the whole of the State southern lines of Austria with their branches, and including the Semmering section. This virtual addition to the length illustrates the well-known importance of steep gradients in the relative working expenses of railways, and is a consideration of special importance in judging of the respective advantages of different transalpine routes in shortening the distance between certain points. In this particiilar instance, the addition, which is \erj large as regards the Semmering section alone, becomes some- what merged in the length of the whole system, and is, moreover, somewhat reduced by the existence of other steep gradients across the Julian Alps near Trieste. It appears to have been imagined that another line connecting Vienna with Trieste, by a detour of 110 miles, opened in 1868, would have diverted the heavy traffic from the Semmering. This might have occurred if the cost of trac- tion and method of working on the Semmering inclines had remained the same as in the earlier years ; but the improvements in these respects had become so great by 1868, that the Semmering route was able to compete successfully with the new line ; and instead of a reduction in traffic, the distance run by passenger and mixed trains, on the Semmering section, rose from 61,010 miles in 1868, to 80,034 miles in 1877, and by goods and military trains, from 171,800 miles in 1868, to 271,567 miles in 1877, in conjunction wdth a steady gradual increase in the gross average load of the trains.

Though the railway does not ascend as high as 3,000 feet above the sea, the trains on the Semmering are sometimes impeded by

' Memoircs de la Societe' des Ingc'nieurs Civils, 1878, p. HO, and Table 10.

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 245

snow ; and the average annual cost of clearing the line from snow lias been reckoned at £13 8s. per mile.

The Semmering railway serves also to connect Vienna with Italy by means of a line which, crossing the eastern Alps, and reaching the summit-level at St. Lambrecht, only 2,917 feet above the sea, descends by Tarvis to Udine in Italy, and thence proceeds on to Verona and Bologna (Plate 6, Fig. 1).

Brenner Eailway. Plate 6, Figs. 1 and 4 ; and Plate 7, Fig. 2.

The success achieved in crossing the Semmering led the Austrian government, in 186-i, to embark upon the more formid- able enterjirise of crossing the main chain of the Alps, to provide railway communication between the Tyrol and Venetia, and the rest of the empire (Plate 6, Fig. 1). The route selected was naturally the Brenner pass, which, besides being suitably situated between Innsbruck and Verona, is the lowest of the main Alpine passes, and, lying at a distance from the highest peaks, is freer from snow than the others. The slopes on each side of the ridge are such that, by contouring the hills with curves having a minimum radius of 14 J chains, it was possible to attain a summit- level 4,497 feet above the sea, 91 feet lower than the pass of the road, with gradients not exceeding 1 in 40 on the northern slope, and 1 in 44 on the southern slope, and without any tunnel at the summit (Plate 6, Fig. 4 ; and Plate 7, Fig. 2). The works were commenced in March 1864, and the line was opened for traffic in August 1867. The line is double between Innsbruck and the Brenner station at the summit, with the exception of a portion of single line, 2^- miles long, between Matrei and Steinach, where the gradients are less steep ; but it is single on the opposite side of the summit, between the Brenner Station and Botzen, except from Franzensfeste to Brixen, where there is a double line for 6.\- miles. ^ The total length of the line from Innsbruck to Botzen is 78?, miles, of which only 26? miles are laid with a double line.

The railway rises 2,586 feet between Innsbruck and the summit, in a distance of 23 miles, with gradients of 1 in 40 along 17^

' Nearly 50 miles of the heaviest portion of the line appear to have cost about £70,000 per mile, considerablj^ less than the Semmering line, which, however, is a double line throucrhout.

246 VERXON-HAKCOUKT ON ALPINE ENGINEERING. [Selected

miles of this length, so that even with the levels, or flat gradients, at the stations, and the easier gradients between Matrei and Steinach, the average gradient between the Innsbruck and Brenner stations amounts to 1 in 46 • 88. This portion of the line passes through fourteen tunnels ; the length of the longest is 948 yards ; and their total length is 3,839 yards, some of them being very short. The descent from the sxammit to Botzen amounts to 3,624 feet, in a distance of 55^ miles, giving a mean inclination of 1 in 80-86; but the steepest gradients occur between the summit and Sterzing, where the fall is 1,385 feet in 13i^ miles, or an average gradient of 1 in 52 • 54, the maximiim gradient, on the southern slope, of 1 in 44, extending over 10| miles of this length. In the next 18 miles, between Sterzing and Brixen, the fall is 1,240 feet, so that the average gradient along this section is reduced to 1 in 76*74; and the 1 in 44 gradient only occurs along a length of 5i miles, with about § mile of 1 in 46, and 1 mile of 1 in 50. There is a fall of only 999 feet in the 23| miles from Brixen to Botzen, making the average gradient 1 in 125 '21 ; and the steepest gTadient on this portion is 1 in 67, for a length of 5i miles. There are three tunnels between Brenner and Sterzing, with a total length of 1,038 yards, of which one is 832 yards long ; and there are five tunnels between Brixen and Botzen, having a total length of 823 yards, one being 426 yards long. No tunnels occur between Sterzing and Brixen ; so that as regards both gradients and works, the southern slope is more favoiirably circumstanced than the northern slope. The line is curved along nearly half its length; the sharpest curves of 14,V chains have a total length of 91 miles, or nearly one-eighth of the whole distance ; and 4^ miles of this length occur between Innsbruck and the summit.^ The line is laid with Bessemer steel rails, weighing 65|- lbs. per yard, on wooden sleepers ; and admits of locomotives with 13J tons on a pair of wheels running over it. The locomotives employed are similar to the improved Semmering types. The express passenger-trains are drawn up the inclines by an engine with six wheels coupled, at a speed of lo^ miles per hour, and descend at a speed of 233- miles an hour. The mixed trains and goods-trains, drawn by engines with eight wheels

• Full particulars of the gradients and curves, with the exact length of each, and of the tunnels, between Innsbruck and Botzen, together with a longitudinal section of the railway, are in the library of the Institution, having been furnished by IMr. Karl Jenny, in response to iinjuiries made by the Author through Mr. Forrest, Sec. Inst. C.E.

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 247

coupled, ascend and descend at a rate of from 9.\ to Hi miles an hour, the weight of the train being 200 tons. Owing to the easier curves on the Brenner than on the Semmering, heavy goods-trains were always taken up the Brenner inclines in one piece, by engines at each end of the train ; so that, in 1867, a train of 3G9 tons, with two engines and tenders weighing 131 tons, making a gross load of 500 tons, surmounted the incline. The sharp curves, however, on the Semmering aided the brake-power in the descent ; so that the working of the trains in descending the Brenner was more difficult, till improved brake-power on the Le Chatelier system was introduced.

The traffic on the Brenner line increased steadily during the first ten years of its existence, the number of miles run by pas- senger and mixed trains having risen from 125,154 miles in 1868, to 229,398 miles in 1877, and by goods and military trains, from 154,639 miles to 310,535 miles; whilst the gross mean load of the trains increased from 100 tons in 1868, to 138 tons in 1877. In comparing the Brenner railway with the Semmering, it must be remembered that the working of the Brenner was commenced with the great advantage of fifteen years' experience on the Semmering. The cost of traction on the Brenner, which was Is. lO^d. per train-mile in 1868, or Id. higher than on the Semmering in the same year, was reduced to Is. Gd. in 1877, or l^d. less than on the Semmering at the same period. The cost per ton-mile, however which forms the proper basis for comparison, shows a more steady and greater reduction, owing to the considerable rise in the train- loads, having fallen from 0-178tZ. in 1868, to O-OQld. in 1877, slightly greater than the similar cost on the Semmering in the first year, but less than two-thirds of the cost on the Semmering in the latter year.^ This latter difference may be attributed mainly to the good loading of the trains in both directions on the Brenner, whilst the main traffic on the Semmering is from Trieste to Vienna ; and also to the Brenner railway being three times the length of the Semmering railway, which reduces the proportionate expenses, and to the curves being easier, and the total average gradient lighter on the Brenner than on the Semmering, namely, 14i-chain curves, instead of 9.V chains, and 1 in 66*7 average gradient as compared with 1 in 60 • 7.

The reduction in cost of traction, between 1868 and 1877, was still greater on the rest of the Tyrolese lines than on the Brenner, having fallen from O-l-ild. per ton-mile in 1868, to O-OG'Id. per

' Mcmoiios dc la Socicte des lugeuicurs Civils, 1878, p. 110, aud TaMc ^<'.

248 VERNON-HAECOUET ON ALPINE ENGINEERING. [Selected

ton-mile in 1877, partly owing to the increase in the aA'erage train-loads from 121 tons in 18G8, to 209 tons in 1877. In the latter year, the cost of working on the Brenner was about half as mnch again per ton-mile as on the rest of the Tyrolese lines ; so that the Brenner gradients were equivalent, in this respect, to an addition of 39 miles to the whole system. This length is identical with the virtual addition of length due to the Semmer- ing gradients and curves at the same period, in spite of the much greater actual length of the Brenner railway, owing to the great reduction in cost of traction on the Brenner ; so that under the conditions existing in 1877, the 78^^ miles of steep gradients on the Brenner did not add more to the cost of traction per ton- mile than the 25 \- miles on the Semmering. The above addition, however, only relates to the cost of working, and not to the time of transit, which is aifected by the length of the ascending gradients, and has an important influence on the route to be preferred for the mails and express passenger-trains. The Brenner railway has a much more considerable bearing on the Tyrol system, of which it forms part, than the Semmering on the southern State railways, owing to the comparatively short length of that system, which comprised only 190 miles of railway altogether in 1877, the Brenner line accordingly being about two-fifths of the whole.

The Brenner railway, besides being the first to cross the main barrier of the Alps, has its summit-level higher than any of the lines which have hitherto been constructed across the Alj)s, with the exception of the temporary Fell railway; for the summit- level in the Mont Cenis tunnel is 10-i feet lower, in the Arlberg tunnel 198 feet lower, and in the St. Gothard tunnel 711 feet lower, than at the Brenner (Plate 7, Fig. 7). It also retains the peculiarity of being the only Alpine railway without a tunnel at the summit, which the moderate elevation of the j^ass, and its favourable position, rendered it possible to dispense with.

The Brenner railway not only connected Austria with its TjTolese possessions, and with Yenetia, which it lost before the completion of the line, but also placed Germany in direct railway communication with Northern Italy, and with the port of Brindisi, and thus established a very valuable link between northern and southern Europe.

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 249

Mont Cenis Fell Eailway. Plate 6, Fig. 2.

When the passage of the Alps by railways was contemplated, attention was naturally directed to the principal passes, which afford the easiest and most direct access by road between various places. No other pass, however, across the main Alpine chain, except the Brenner, is suitably situated at a low enough level to enable an ordinary railway to be carried over the summit ; for the line of the celebrated Corniche road, avoiding the Alps in entering Italy by skirting the sea-coast, is too circuitous to be of more than local importance in providing railway communication between Marseilles and Genoa, and to the intervening towns. Accordingly, the Alpine railways have gradually been carried up the valleys on each side of the principal passes, till at length the exigencies of through traffic have led to schemes for traversing the gaps between the two ends of the several lines, by a railway rising with steep gradients, and piercing the insurmountable intervening ridge by a summit tunnel of considerable length.

The Mont Cenis pass, forming the main line of communication between France and Italy, was proposed for the route of an Alpine railway in 1 852 ; but the uncertainties attending the con- struction of a tunnel over 7^ miles long, through the hardest strata, without shafts, delayed the commencement of the works till 1857. The piercing, however, of the tunnel by hand labour, at the beginning, progressed so slowly that Mr. Fell, in 1863, revived the idea, first suggested in 1830, of using horizontal wheels on a locomotive, gripping a central rail, for obtaining additional adhesion, and thus enabling trains to travel safely over very steep inclines and sharp curves. Mr. Fell jiroposed to utilize this system for traversing the break in the line between St. IMichel and Susa, a distance of 48 miles, by constructing a railway of this type along the Mont Cenis road. After two experimental trials,^ the first on the High Peak Eailway in Derbyshire, in 18G3 and 1864, and the second on a portion of the Mont Cenis road,'^ 1^^ mile long, in 1865, permission was obtained to construct the line for conveying the traffic till the completion of the tunnel. When

' Minutes of rroceedings lust. C.E., vol. xxvi. p. 313 ; Report of the Meetiug of the British Association in 1866, p. 143; and Ecport on the Mout Cenis Eail- way to the Board of Trade by Capt. Tyler, R.E., 1865.

' Auuales des Pouts et Chaussees, 4th series, vol. xi. 1866, p. 95, plate 116; and "Etudes sur la loeomoliou au moyeu du Kail Central," M. Desbrierc, 1866.

250 VERNOX-HARCOURT ON ALPINE ENGINEERING. [Selected

the works were commenced, in March 186G, it was supiDosed that the line would be open for traffic in about a year ; whilst it was estimated that the tunnel would require over eleven years more for its completion.^ The estimated cost was £320,000 ; and it was calcTilated that, with only seven years of working, the railway would, beside paying 7 per cent, per annum interest on the capital, both repay the capital expended, and also leave a hand- some margin of profit for division on winding uj) the company at the opening of the tunnel.'^ None of these golden anticipations, however, were realized. The completion of the railway was delayed by an unusually high flood of the Arc in the autumn of 1866, which not only washed away portions of the railway, but also did much damage to the road for some miles above St. Michel. The railway followed the road, except in a few places where diversions were made to obtain a more uniform g-radient, and to avoid villages ; it occupied 13 feet in width of the outer portion of the road, and was laid to a gauge of 3 feet 7f inches, with curves having a minimum radius of 2 chains, the central rail, laid on its side, being raised 7.j inches above the ordinary rail-level. The maximum gradient of the line was 1 in 12 ; but the greater portion of the slope on the French side is so much gentler than on the Italian side, that the mean gradient from St. Michel to Lanslebourg, a distance of 24 miles, was 1 in 60, with a very- short length of 1 in 12; whereas, for the remainder of the ascent from Lanslebourg, at the turn of the road some distance beyond Modane, to the summit, 6,772 feet above sea-level, and in the descent to Susa, the average gTadient was 1 in 17, and the j^revailing gradient 1 in 12, as only at two places, one near the summit and the other close to Susa, did the gradient exceed 1 in 25. The rise from St. Michel to the summit is about -±,600 feet ; and the fall from the summit to Susa about 5,300 feet. The central rail was laid along all gTadients exceeding 1 in 25. The railway passed through 9 miles of covered way, to protect it where most liable to snow-drifts, made of masonry where exj^osed to avalanches. The first train was taken over the line in Aiig-ust 1867 ; but owing to defects in the working of the locomotives first em- ployed, and other causes, the railway was not actually opened for traffic till June 15th, 1868, when an engine, weighing 22 tons when loaded, drew a train of 17 tons from St. Michel to Susa, the time occupied in transit, exclusive of stoppages, being four

* Engineering, vol. i. p. 7. - The Eiifjiiuei; Jau. 18(y(;, p

Papers.] VERNON-HARCOUKT ON ALPINE ENGINEERING. 251

hours and fifty minutes, or a rate of 10 miles an liour.^ The railway was worked regularly from this time till the opening- of the tunnel line in September 1871, with only a few interruptions, once by torrents of rain two months after the opening, and two or three times by snow; but for a time the want of sufficient loco- motives hampered the traffic. The trains ran with remarkable freedom from accident, by aid of the central rail and brakes, in spite of the steep inclines and very sharp curves, for the only casualty during the working of the line was a goods-train leaving the line, owing to the negligence of the engine-driver in reversing his engine, for returning down the incline on a stormy night, without first applying the brakes. The greatest load taken over the railway was a train weighing 36 tons, and the heaviest loco- motive employed weighed 26 tons.

The financial resiilts of the enterprise were naturally not satis- factory, for the limited life of the railway was shortened at the beginning by delays in opening, and at its close by the increased rate of progress of the tunnel, so that the railway was only in operation for about three and a half years, instead of the antici- pated seven to ten years. Moreover, during that period consider- able modifications in the locomotives, and additions, proved necessary ; and the working expenses were very heavy, owing to the frequent repairs of the engines, which experienced great wear- and-tear in going round the very sharp curves, and the cost of keeping the line clear of snow in the winter. The cost of con- struction also had risen to about £450,000, an increase of two- fifths over the original estimate.'^ The railway, too, never quite sujDcrseded the diligences, for in the last complete year of its working it carried twenty-eight thousand passengers, out of a total estimate of forty-two thousand persons annually crossing the Mont Cenis, or two-thirds only of the whole number.^ In con- sequence of these various causes the railway never paid a dividend, and a large portion of the original capital was sunk.

From a purely engineering point of view, the railway over the Mont Cenis was a decided success ; it proved that a railway could be worked in perfect safety on the steep inclines, and over the lofty summit of an Alpine road ; it efi'ected a saving of about six hours in the transit as compared with the diligences, and it enabled an accelerated mail-service to India, via Brindisi, to be established

Enrjineering, vol. v. pp. 598 aud 620.

Ibid., vol. vii. p. 10.

The Engineer, Oct. 1871, p. 231.

252 VERNON-HAECOUET ON ALPINE ENGINEERING. [Selected

more tlian a year earlier than would otherwise have been prac- ticable. If the scheme had been started in 1852, when the Mont Cenis route was first proposed, instead of when the tunnel works had made considerable progress, it is possible that the adoption of a long tunnel might have been deferred ; and in any case the railway would have had ample time to pass completely beyond the experimental stage, and to acquire engines fully adapted to its requirements, and would have had a fair prospect of financial success.

Though the suggestions, made at various times, for crossing the Simj)lon, St. Gothard, and Lukmanier passes by a Fell railway, in preference to a long tunnel, so as to reduce the capital cost, have not met with approval, the experience gained on the Mont Cenis has not been fruitless. The central-rail system has subsequently been adopted for steej) inclines on the Wellington and Feather- stone Eailway, in New Zealand,' and on an extension of the Cantagallo Eailway in Brazil.- The New Zealand line, with a 3 feet 6 inches gauge, ascends a continuous gradient of 1 in 15 for about 2^ miles, for crossing the Eochfort saddle, with curves having a minimum radius of 5 chains, and a train weighing 53 tons is drawn tij) by an engine of 36 tons ; whilst the saving in capital cost by the adoption of the steep gradient, instead of an ordinary line, was estimated at £100,000. The Brazilian line crosses the Serra about 3,000 feet above the Cantagallo Eailway, rising to this altitude in about 10 miles, with gradients of from 1 in 20 to 1 in 12, and with curves round the projecting spurs of from 2 to 5 chains radius, and descending with easier gradients to Nova Friburgo. The gauge of this line is 3 feet 7-^ inches, like the Mont Cenis ; and the locomotives used on it resemble the latest type of the Mont Cenis engines, with the machinery of the ordinary and of the horizontal wheels disconnected, which made the working much easier, and with further improvements.^ It was anticijjated that these new locomotives would effect a large reduction in the cost of working, which amounted to 5s. per train- mile on the Mont Cenis. The system has also been recently proposed for carrying branch railways over the Tenda and Genevre passes, for which the cost of long tunnels could not be entertained (Plate 6, Figs. 1 and 2). The central-rail system has the advantage of effecting a very considerable saving in capital cost in mountainous

' IMiuutcs of Proceedings Inst. C.E. vol. Ixiii. p. 50.

■ lleport of the Meeting of the British Association in 1870, p. 210.

^ 3Icmoires de lu tfociete des IngJuieurs Civils, 1S72, p. 118, and 187(5, ]). 155

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 253

districts, with, in some cases, as on the Xew Zealand line, a rednctiou in length as well. On the Mont Cenis railway the length was 6}} miles greater than by the tunnel-line, and the time of transit ahout four hours more, taking into account the time spent in trans-shipment. But the saving in cost, even for a more solid and more efficiently pro- tected permanent surface-line (estimated by Mr. Fell to cost £20,000 a mile) would have been very large as compared with the tunnel- line. Against this saving must be set off the additional cost of working with special engines, and the lower speed ; the cost of clearing the line from snow, at the greater heights in Alpine regions, and the probable interruptions of traffic by storms and snow-drifts ; the delays and inconvenience of a double trans-ship- ment, unless the line was made to the ordinary gauge; and the smaller loads which could be conveyed, necessitating a division of the trains, and involving a considerably less capacity for traffic.

Considerations affecting Main Lines through the Alps.

A variety of circumstances have to be considered in determining the proper routes for railway communication across the Alps. Besides selecting the lines of greatest probable traffic, it is essential that the shortest and easiest route should be adopted, as far as physical conditions will permit, otherwise a competing line may be constructed, and divert the traffic which has been secured at a large cost. Moreover, the gradients of the approach railways, on each side of the Alps, as well as of the connecting line, must be taken into account, as a more circuitous route with lighter gradients may be able to convey the traffic at a lower cost ; and a large outlay on the connecting line may prohibit an adequate lowering of the rates to attract the traffic which should naturally pass over the system. Eapidity and regrxlarity of transit attract passengers, and low rates determine the route adopted for goods ; and these are incompatible with heavy gradients, exposed situations, and a large capital cost. In deciding between two or three adjacent passes, the height to which the line must ascend in the open air, the nature of the works with the practicable gradients and curves of the lines of access, and the length of the summit tunnel, and its maximum depth below the surface must be con- sidered. If the railway is carried high up the valleys, on each side, before entering the summit tunnel, the gradients must be made steeper and the curves sharper, or the works heavier, as the slope of the valley increases on ascending, and the railway is much

254 YErtNOX-nARCOURT ON ALPINE ENGINEERING. [Selected

more liaLle to l)e blocked with snow ; -whilst if the tunnel is entered at a comparatively low level, the tunnel is longer, and the internal heat in the tunnel, towards the centre, during con- struction is gTeater, owing to the increased depth below the surface. The limit of height for an open line, in the centre of the Alps, appears to have been reached by the Mont Cenis railway, 4,270 feet above sea-level ; for this line, and also the St. Gothard, with a maximum height outside the tunnel of only 3,756 feet above the sea, have been occasioiaally blocked by snow. The limit of depth below the surface appears to have been nearly attained in the St. Gothard tunnel, where the heat during con- struction became trying to the workmen, with a maximum depth below the surface of 5,733 feet. A divergence of opinion, however, on these points is exhibited in some of the proposed schemes for traversing the Alps ; for the open portion of the Great St. Bernard route would greatly exceed the first limit, with the object of avoiding the second ; whilst the Mont Blanc route, though much below the height of the Mont Cenis, would have the depth of its tunnel, below the highest peak, nearly twice the maximum depth of the St. Gothard tunnel (Plate 7, Fig. 7). Assuming other conditions to be similar, the preference should undoubtedly be given to a low-level route, provided the length of tunnel is not much greater, and the internal heat not likely to be excessive, as moderate gradients, a smaller ascent, and immunity from interrup- tion of traffic, are important factors in the competition for traffic. The distance also between the ends of the completed railways, on each side, has a bearing on the choice of route, as affecting the cost of the reqixisite connecting link.

It is impossible that many lines across the Alps could all afford an adequate return for the large cost which each would necessitate, and therefore it is the more important that the lines selected should be the best attainable. Not only, however, have the interests of the two countries, which are to be connected by the proposed railway, to be consulted, but if the line is to pass through Switzer- land, the approval of that country is essential. Moreover, the success of the line depends, not merely upon the traffic between the two or three countries combining to construct it, but especially upon the throiigh traffic from other countries it may attract. The Brenner line, though lying wholly in Austrian territory, profits largely by the through traffic between Germany, Italy, and the sea-coast. The Mont Cenis line also, connecting France and Italy, obtained a considerable through traffic between north-western Europe and Italy, and with the East via Brindisi, though some of this

Papers.] VEUNON-HARCOURT ON ALPINE ENGINEERING. 255

traffic was diverted from Marseilles. These two first Alpine routes appeared marked out Ly nature for serving central and western Europe without any antagonism ; but when a third intermediate route was proposed, various alternative schemes were suggested. The St. Gothard line, which was given the preference by the in- fluence of Switzerland and Germany, has gathered its traffic from regions formerly served, more or less, by one or other of the older lines; it is acknowledged that the St. Gothard has drawn away a considerable amount of traffic from the Mont Cenis line ; and it has ])robably affected the traffic on the Brenner railway. The injury, indeed, to French trade by the diversion of traffic from France into Belgium and Germany, by the opening of the St. Gothard railway, has been so marked, that a fourth Alpine line is proposed, which, whilst improving the means of communication of certain districts, is mainly designed to compete with the St. Gothard, and to bring back into France a portion of the traffic which originally passed ])y the Mont Cenis. Accordingly, the novel element of regulating the balance of national trade has been added to the considera- tions affecting Alpine routes.

Mont Cenis Eailway and Tunnel. Plate 6, Fig. G, and Flate 7, Fig. 3.

As early as 1840, Fourneaux to Bardonneche was recognized as the proper direction for a sub- Alpine tunnel to connect France and Italy; but the first definite proposal for the construction of the tunnel was made in 1852, and the works were only commenced in 1857.

Mont Cenis Railway. — The railway across the Mont Cenis had to connect the French terminus at St. Michel with the Victor Emmanuel Eailway at Bussoleno, the junction for the branch line to Susa about 5 miles distant. The section of the Mont Cenis railway fairly resembles, in its general outline, the section of the Brenner line (compare Plate 7, Fig. 3 with Fig. 2), though the necessity for a long summit tunnel completely modified the conditions of its construction. The railway had already commenced the ascent before reaching St. Michel ; for the portion between St. Jean de Maurienne and St. Michel j^asses through three tunnels, and has some gradients of 1 in 5U to 1 in 37^ on it, with an average gradient of 1 in 70, rising 577 feet in about 7?- miles (Plate 7, Fig. 3). The distance between St. Michel and Bussoleno is 4G^ miles; the rise from St. Michel to the summit, inside the

256 VEHNON-HARCOURT ON ALPINE ENaiNEERING. [Selected

tunnel, is 2,061 feet in nearly 17 miles, and the fall from the summit to Bussoleno is 2,806 feet in 29j miles, giving an average gradient of 1 in 43 • 3 on the French side (reduced to 1 in 49 • 1 if reckoned from St. Jean de Maurienne), and 1 in 55 on the Italian side. The steepest portions of the line are near the Fourneaux entrance, and in the tunnel on the French side, where gradients of 1 in 43_j to 1 in 38 ^ are continuous for 5 miles, and between Salbertrand and Bussoleno, where the gradients average 1 in 38 for 13 miles. There are 4i miles of gradients of 1 in 40 ; and steeper gradients than 1 in 40 extend over 6^- miles on the French side, and lOj miles on the Italian side, the steepest gradient on the line being 1 in 33^ for a total length of 7| miles, of which 5f*Q miles are on the French side, comprised mainly between St. Michel and Lapraz. There are fourteen tunnels between St. Jean de Maiarienne and the Mont Cenis tunnel, with a total length of 4,789 yards, one having a length of 1,143 yards, and the rest ranging between 743 yards and 61 yards. The works on the Italian slope are considerably heavier, for the line passes over eight viaducts, and through twenty-six tunnels in the 25 miles between the Bardonneche end of the tunnel and Bussoleno. These tunnels have a total length of 8,835 yards, the longest being 1,933 lineal yards, two others 1,203 and 1,196 yards long, and the remainder from 588 to 49 yards in length. The line is not so tortuous as either the Semmering or the Brenner railway, there being only one decided loop close to the Fourneaux entrance to the tunnel (Plate 6, Fig. 6) ; but it is curved for three-sevenths of its length. The curves are sharpest on the French side, for on that side, the most frequent curves of 24 • 8 chains extend over a length of about 03- miles, three curves of 20 and 17^ chains have a total length of over ^ mile, and there is one curve of 17 chains radius in the loop betAveen Modane and Fourneaux f mile long ; whereas, on the Italian side, the curves of 24*8 chains occupy only 626 lineal yards, and there are only three sharper curves, with a radius of 23 • 8 chains, having a total length of h mile.^

Comparing the Mont Cenis railwaj^, between St. Jean de Mau- rienne and Bussoleno, with the Brenner railway between Innsbruck and Brixen, it appears that the average gradient throughout is slightly steeper on the Mont Cenis than on the Brenner, 1 in 52

' Full particulars of the gi-adients and curves, with their respective tenths, together with some indications of the tunnels, viaducts, and bridges, are given in a longitudinal section of the railway between St. Jean de Maurienne and Bussoleno, in the library of the Institution, presented by Mr. Jules Michel, in response to a request made by the"^ Author through Mr. Forrest.

Papers.] VERNON-HARCOUET ON ALPINE ENGINEERING. 257

as compared with 1 in 55 • 5, but for a somewhat shorter distance, 53| miles instead of 5-lf miles. The ruling gradient also of the Mont Cenis is 1 in S'A^, whilst on the Brenner it is 1 in 40, and the ap- proach works are considerably heavier on the Mont Cenis ; but on the other hand, the curves are much easier on the Mont Cenis, for no curves on the Mont Cenis are as sharp as the curves of 14^ chains so common on the Brenner, and extending over a much gTeater distance than the most frequent curves of 24-8 chains on the Mont Cenis (See Appendix).

Mont Cenis Tunnel. — The first long tunnel through the Alps has generally been designated as the IVIont Cenis tunnel, on account of its aiibrding the same communication by railway which was formerly effected by road over the Mont Cenis pass. But though the railway skirts the road from St. Jean de Maurienne to Modane, it shortly after diverges, and entering the tunnel at Fourneaux, follows a south-south-easterly course, passing under the Col de Frejus at the French boundary, several miles south-south-west of the Mont Cenis pass, and emerging at Bardonneche in Italian territory (Plate 6, Fig. 6).

The Mont Cenis tunnel was driven in a perfectly straight line from each end, thereby reducing as far as possible the chances of error in direction, but with gradients of 1 in 43^ on the French side, and 1 in 2,000 to 1 in 1,000 on the Italian side, rising from the Fourneaux end to attain a higher level on the Italian side, and from the Bardonneche end solely to ensure a fall for drainage. The straight tunnel was subsequently joined, at short distances from its extremities, by two curved tunnels to the approach lines on each side, increasing the actual length of the tunnel, as used for traffic, from 7"6 miles to 7 "97 miles; the curved tunnel at the French end was made 464 yards long, to a radius of 24 • 8 chains and a gradient of 1 in 43^, and at the Italian end, 891 yards long, with a portion to the same radius, and a gradient of 1 in 33^. The exact line of the straight tunnel was determined by very careful triangula- tion, and indicated by marks along the surface ; and it was maintained by lines given from observatories established in the direct line on the further side of the valleys opposite each end of the tunnel.

The work of driving the heading at the face of the tunnel, which Avas commenced at Ijoth ends towards the close of 1857, progressed very slowly till the Sommeiller boring machines were introduced, in 1861 at the south end, and in 1863 at the north end ; the greatest advance made in one year at both ends by hand boring being 502 lineal yards in 1858, as compared witli 1,788

[tIIK INST. C.E. VOL. XCV.] S

258 VEENOX-HAKCOURT ON ALPINE ENGINEERING. [Selected

yards in 1870, the year of greatest progress with the machines, the strata penetrated at both jieriods being schist. The upper diagram (Fig. 1) shows the yearly rate of progress of the advanced lieadings at each end, from the commencement at the close of 1857 till the headings were joined on the 25th of December 1870, and the strata through which they were driven, together with the temperatures observed in the tunnel on the Italian side during construction. The rate naturally varied according to the strata ; and after the introduction of the machines, the greatest and the least monthly advance occurred on the French side, amounting to 12-9 feet per day in May, 1865, when traversing carbonaceous schist, and to only 1'17 foot per day in April, 1866, through quartz; whilst the total advance in 1866 was only 232 lineal yards at the northern forehead, being Avholly in quartz, and 901 yards at the southern forehead through calcareous schist. Two years later, in 1868, the progress was greatest at the northern forehead (745 yards as compared with 698 yards) when both the headings were traversing calcareous schist ; but, as a rule, a better rate was maintained in the southern portion, which was solely in calcareous schist, than in the northern portion where the strata varied. The perforators were worked by compressed air, the air being compressed by water-power at each end of the tunnel. The machinery emjdoyed, the method of driving the headings, the stages of enlargement, and the cross section and lining of the tunnel have been fulty described, and particulars of its cost have been given, in two Papers by Mr. T. Sopwith,^ M. Inst. C.E.

The tunnel proved to be 15 yards longer than calculated, and the heading on the French side 1 foot too high in level, jirobably a result of the miscalculation of length ; but the direction was quite correct. The driving occujiied thirteen years and one month ; and the average daily progress was 2*57 lineal yards. The total cost of the tunnel, amounting to £3,000,000, was equivalent to about £224 per lineal yard. According to Mr. Sopwith's estimates in 1863, the introduction of machinery increased the cost of driving the advanced heading's nearly in as large a proportion as it improved the rate of progress ; but the machinery was novel and on a large scale ; and it is certain that with the improvements effected in the machinery, and the increased advance, the cost of the machine

' IMiinites of Proccediugs lust. C.E., yoI. xxiii. p. 258, .iiid vol. xxxvi. p. 1. Further details will also be found in " Traforo delle Aljii tra Bardouuecbe o Modanc,Kclazionc," Torino, 18G3. G. Sommeiller ; Annales dos Fonts et Cliaussccs. 4th seri(!8, vol. v. p. 1 and plates 37 to 40 ; and Engmecrimj, vols. xi. and xii. " The Mont C'enis Tunnel," bv Francis Kossuth.

Papers.] VERNON-HAECOUBT ON ALPINE ENGINEERING. 259

. (879

\,

>^

*■ - XiX- ' - . '880 1

LANGEN

^,

° MICA

-ECM

■2 «l

:"!-.

.3

PT.O.i'

9CM

ST

j/

oy/M

f^ ST ANTON

260 VERNON-HARCOURT ON ALPINE ENCxINEERING, [Selected

work must have been considerably reduced before tbe close of the works. The large cost at first was due to the novelty and experimental nature of the work ; and the above cost per lineal yard must by no means be regarded as the actual cost at a later stage of the work, for it is supposed that the contractors made a profit of over 100 per cent, on the 4,798 yards of tunnel remaining to be constructed in 1867, let to them at £167 12s. per yard. The experience, however, of the application of boring-machinery worked by compressed air at the Mont Cenis tunnel, though purchased by the French and Italian Governments at a high cost, has been most vahiable for subsequent similar works. Whereas, also, the slow progress of the tunnel works, during the earlier years, diverted attention for a time to the more novel and more visiljle works of the Fell railway, yet as soon as the success of the tunnel was established, a great impetus was given to other similar schemes. The railway was opened for trafiic towards the end of 1871, nearly fourteen years after its commencement, the longer duration of the works of the Mont Cenis railway than of the Brenner railway, by about eleven and a half years, being wholly due to the construction of the summit tunnel.

Owing to a gradual settlement of the recent glacial deposits at the Fourneaux end of the tunnel, and consequent dislocation of the tunnel near its entrance, a new curved entrance, 1,718 yards long, was driven through the carbonaceous schist, clear of the glacial deposit, in 1879-81, which shortened the total length by 252 yards. ^

To promote the ventilation of the tunnel, the air-compressing machinery was left at Bardonncche to supply air under pressure in an 8-inch pipe, laid all along the tunnel, from which it coiild be drawn through cocks at intervals when required ; and the exhausters, which had been employed at the Fourneaux end for drawing out the foul air from the tunnel during construction, were maintained for drawing air along a passage, laid at the bottom of the tunnel, through apertures opened at pleasure in diflerent parts of the tunnel.^ The ventilation thereby afforded is very inefficient, except near the ends of the tunnel where it is least wanted ; but fortunately natural ventilation, j^artly due to differences in the atmospheric conditions at the two ends of the tunnel, partly to the greater heat towards the centre, and

' Minutes of Proceedings Inst. C.E., vol. Ixvi. p. 41.3.

^ Ibid., vol. liii. p. 1G4; iind Rapport des Experts siu- le rerecnient dn Simi>lon. p. 24.

Papers] VERNON-HARCOURT ON ALPINE ENGINEERING. 261

occasionally to the wind, keeps the tunnel generally fairly clear of fonl air.

Influences of the Mont Cenis Bailway. — The opening of the railway passing under the Col de Frejus, not only joined the Paris, Lyons, and Mediterranean Eailway to the Italian system of railways, and placed Paris, Lyons, and Geneva in direct communi- cation with Turin, Genoa, Florence, and Kome, but also rendered Brindisi the most suitable port to the East for Great P>ritaiu and a large portion of France, in place of Marseilles. The distance between London and Alexandria was reduced l)y adopting the Brindisi route, being 2,431 miles, instead of 2,534 miles via Marseilles. A great saving of time, moreover, was effected by substituting a route by land for a considerable portion of the sea route, since a train travels faster than a steamer. Thus the route by the Mont Cenis and Brindisi, by reducing the sea route from 1,701 miles to 954 miles, and increasing the land route from 833 miles to 1,477 miles, was reckoned by Captain Tyler,^ Assoc. Inst. C.E., to effect a saving in time in the journey from London to Alexandria of forty-two hours. The diversion from Marseilles was not favourable to the interests of the Paris, Lyons, and Mediterranean Eailway Company, which therefore naturally evinced no eagerness for the completion of the line. Nevertheless, the change must have come sooner or later by some route; and that company had the advantage of being the first in the field for accommodating Western Europe.

St. Gotharu Eailway and Tunnel.

Plate G, Figs. 5 and 7 ; and Plate 7, Fig. 4.

Proposals for constructing a third railway across the Alps were not delayed till the success of the Mont Cenis tunnel was assured, for Mr. Flachat suggested the Simplon route in 1859, and the St, Gothard scheme was brought into notice in 1866. If it had been decided at the outset that one line only besides the Brenner should be constructed across the main Alpine chain, the Simplon line would have possessed the best claim to be selected, as the most central between the Brenner railway and the Corniche road, and as the nature of its api^roaches was easier, the length of line remaining to be constructed was shorter, and its summit-level

' Report ou the Mout Ceiiis Eailway to the Uuard of Trade by Capt. Tyler, R.E., 18(;:).

262 VERNON-HAKCOIJET OX ALPINE ENGINEERING. [Selected

would be lower than that of any other route. The only objections that could be raised against the Simplon were the length of its tunnel, and the heat that might possibly be encountered in tunnel- ling at a considerable depth below the surface. As soon, however, as the Mont Cenis route was undertaken, the conditions were changed, as the St. Gothard route then became the most central between the Mont Cenis and Brenner railways ; and the Simplon line would have trenched upon the Mont Cenis zone of traffic, without fully accommodating the district lying between it and the Brenner (Plate 6, Fig. 1). Two other routes besides the Simplon were put forward as rivals to the St. Gothard route, namely, the Lukmanier route, which would have terminated at the same point, Biasca, as the St. Gothard railway, and the Spliigen route which would hai^e joined the main line at Milan (Plate 6, Figs. 1 and 5). The Spliigen route was much nearer the Brenner than the St. Gothard, and traversed only a narrow portion of Switzerland ; and it was rejected owing to its minor commercial importance. The Lukmanier route, lying between the St. Gothard and the Spliigen routes, would have started from the same point, Chur, as the Spliigen route, and therefore equally near to the Brenner; and though its approach railways would have been shorter, its gradients easier, and its summit-level lower, than those of the St. Gothard, its summit tunnel would have been longer. Eventually the Swiss Government gave the preference to the St. Gothard route over the Lukmanier and the Simplon, as being the most central of the three ; and it was naturally more acceptable to Germany than the Simplon, being far better situated for German traffic and trade.

St. Gothard Railway. — Till the St. Gothard line was opened, Switzerland possessed no direct railway communication from north to south, as both the Brenner and Mont Cenis raihvays passed outside its boundaries ; whereas the St. Gothard railway traverses it nearly centrally, and at its widest part.

The St. Gothard railway, authorized in 1869, is situated at an altogether lower level than the Mont Cenis and the Brenner, for not only is its summit-level considerably lower, but it also descends to lower levels at its extremities, especially on the southern side (Plate 7, Fig. 7). The ascent commences at Erstfeld, 1,501 feet above the sea, but becomes steeper at Amstag 3 miles further on ; and from thence to Goeschenen, at the entrance to the summit tunnel, the line rises 1,842 feet in 142- miles, giving an average inclination of 1 in 42^. This is the steepest portion of the line; and the gradients, excejit foi* short lengths at the stations, are from

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 263

1 in 45 V to 1 in 38^, the steeper gradients predominating. Owing, however, to the smaller rise between Erstfeld and Amstag, and the flatter gradient of 1 in 172 in the northern part of the tunnel, the average rise between Erstfeld and the summit, a distance of 23-j miles, is 1 in 56*3. The descent from the summit, 3,786 feet above the sea, to Biasca, is 2,815 feet in 32^- miles, giving an average gradient of 1 in 60 • 6 on the south side. The steep gradients are not so continuous on the southern slope as between Amstag and Goeschenen; but there are several gradients of 1 in 38^ scattered about, one being about 2,^ miles long; and there is a gradient of 1 in 37 on each side of Giornico station. The gradients average 1 in 45.^ between Fiesso and Bodio, a distance of 17 J miles; and the steepest section is l^etween Lavorgo and Giornico, the site of the helicoidal tunnels, where the average gradient is 1 in 42.^ for 4"; miles.' The curved portions of the line occupy nearly half of the length between Erstfeld and Biasca. Curves of 14 chains to under 15 chains, which are the sharpest on the line, have a total length of 2i miles; curves of 15 chains and under 20 chains are the most frequent, extending over 15 j miles; whilst curves of 20 chains and under 30 chains occupy 5^ miles of the line.

A special peculiarity of the St. Gothard railway is the manner in which the steep portions of the slojjes have been surmounted by a spiral and two adjacent loops at Wasen, and by two spirals between Fiesso and Faido, and two spirals close together near Giornico, by means of which the line doubles back upon itself at a different level, and enables the gradients and works to be kept within reasonable limits (Plate 6, Fig. 7). The railway passes at these places througli eight helicoidal tunnels, four on the northern and four on the southern slopes, two of the former and all of tlie latter being nearly 1 mile each in length, and the other two f mile long. Including these helicoidal tunnels, the railway passes through twenty-one tunnels on the northern slope, with a total length of 8,061 yards, and through twelve tunnels on the southern slope, having a total length of 8,685 yards, giving a total length of 9^ miles of tunnels on both approaches. These tunnels, moreover, are comprised within 13 miles between Amstag and Gojschenen, and within 7^ miles between. Fiesso and Faido, and Lavorgo and Giornico. With the sunnnit tunnel, the total length of tunnelling in the 56 miles between Erstfeld and Biasca amounts to 18 J miles,

' A section of the line giving the heights, gi-adicuts, distances between the stations, and the lengths of the tunnels, is appended to tlic Rapport Triiuestriel, No. 22, snr la liguc du St. (Jothard, I Jan. an :!l Mars, 1S7S.

264 VERNON-HARCOURT ON ALPINE ENGINEERING. [Selected

or actually one-third of the Avhole distance. The timnels on the northern slope were driven by hand-boring, except the Pfaffensprung helicoidal tunnel, where, after 170 yards in length of the heading at the lower end had been driven by hand, at an average rate of 1 • 0 foot per day, 204 yards were driven by the Frolich percussion drill ^ worked by compressed air, advancing on the average 3 j feet per day ; and the remaining 709 yards of heading, driven from the lower end, were accomplished by the Brandt rotatory drill,- worked by water-pressure, at an average daily rate of 5' 94 feet. The progTess of the heading at the upper end of this tnnnel, driven by hand, averaged 1 • 93 foot per day ; whilst the average daily advance of all the headings driven by hand on the northern slope was 1 • 67 foot. The Pfaffenspmng tunnel and four of the other tunnels passed entirely through gneissic gTanite ; the strata traversed by the tunnels included also granite, gneiss, mica schist, moraine, and debris. The Frolich percussion drill was employed in driving the headings of foiar of the tunnels on the southern sloj)e, after a commencement had in each case been made by hand. The total length of headings thus driven in the Freggio, Prato, Piano-Tondo and Travi helicoidal tunnels was 4,190 yards, at an average daily rate of 4.V feet, through gneiss and gneissic mica schist ; whilst the headings driven by hand in these and the other tunnels on the southern slope, amounting altogether to 4,409 lineal yards, advanced at an average rate of 1 • 8 foot per day. Machine-boring was partially adopted at the lower face only of the Prato tunnel; but both headings in the other three tunnels were mainly driven by the percussion drill. It had originally been proposed to construct the tunnels through hard rock for a single line only, leaving the doubling of the section till the traffic should require it. The variable nature, however, of the rock rendered considerably more length of lining necessary than had been anticipated, so that eventually the arch was built almost throughout for a doixble line, leaving the enlargement at one or on both sides till a future time where practicable.

The original scheme for the approach railways had been drawn wp from rough small-scale plans ; so that when the working sections were prepared, it was discovered that the railways could not possibly follow the lines originally traced out, and that the works would considerably exceed the original estimate.^ Accordingly,

' A description, with drawings, of this drill is given in Eapport Trimestriel, No. 32, de la ligne du St. Gothard, p. 85. - Ihid. p. 8'J. ^ Auuales des Pouts ct Chuuss'Jes, 5th series, vol. xiii. 1S77. p. -16.

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 265

the construction of the approach lines, as definitely laid out in 1876, had to be delayed till arrangements could he made for raising the additional capital ; and though the summit tunnel works were continued without interruption, the approach works were not pro- ceeded with till near the close of 1878. The bridges and viaducts were mostly constructed with iron lattice-girders resting on masonry piers, iron being adopted to exjiedite the works. ^ The approach works, however, owing to the delay, were not completed till the 1st of June, 1882, five months after the long tunnel section had been opened for traffic.

The St. Gothard railway between Erstfeld and Biasca, 50 miles long, slightly longer than the movmtain sections of the Brenner and the Mont Cenis, has an average gradient of 1 in 58*72, distinctly flatter than the 1 in 52 on the Mont Cenis, and slightly flatter than the 1 in 55 • 5 on the Brenner. The ruling gradients of the St. Gothard of 1 in 38^ on the northern slope, and 1 in 37 on the southern slope, are intermediate between the 1 in 33*3 of the Mont Cenis, and the 1 in 40 and 1 in 4-1 of the Brenner. Whereas, however, the St. Gothard has the advantage in gradients over the Mont Cenis, it has sharper curves than the Mont Cenis, and a much greater length of curves under 20 chains radius ; but it is superior to the Brenner as regards curves (see Appendix). The tunnels on the slopes of the St. Gothard are much longer than on the Brenner or the Mont Cenis ; their spiral form, to gain length for the rise, is peculiar to the St. Gothard ; and the approach works generally were specially heavy. The St. Gothard railway is formed for a single line, like the Mont Cenis and portions of the Brenner, except in some of the tunnels and at the stations.

St. Gothard Tunnel. — The summit tunnel under the St. Gothard, like the Mont Cenis tunnel, was driven in a straight line from end to end, and constructed for a double line throughout. As there is a difference of level of only 118 feet between the summits of the approach lines at each end of the tunnel, it could be formed with the easy rising gradients towards the summit of 1 in 172 on the north side, and 1 in 500 to 1 in 2,000 on the south side ; the gradient on the south side being simply to ensure drainage, and on the north side to allow for this rise and the higher level at the southern end. The straight tunnel, as driven, had a length of 9-26 miles; but with some modifications at the northern entrance at Goescheuen,

' Small elevations of the bridges and viaducts, and sections of the tunnels, defence walls, &c., are given in the plans belonging to the Final Ecport on the St. Gothard, in Kapport Trimestriel, No. 40.

266 VERNON-HAKCOURT ON ALPINE ENGINEERING. [Selected

and a curved portion to connect it with, the apjiroach line at Airolo, the actual length of the tunnel traversed by trains is 9 "SI miles. The driving of the timnel was commenced in September 1872, with the advantage of the thirteen years' experience at the Mont Cenis tunnel ; and dynamite was introduced for blasting, whereas gunpowder only was employed at Mont Cenis. The boring of the advanced headings at each end was commenced by hand; but machines were introduced at Ga3schenen in April 1873, and at Airolo in July 1873, though the regular compressors did not begin working till about a year after the works had been begun. The machine drills first employed at the two ends were Dubois and Franc^'ois', but they were superseded by the Ferrous drill ; and eventually, after various trials, the Ferroux drills were exclusively used at the northern heading, and McKean drills at the sox;thern heading, from November 1875 till the work was completed.' Water- power at each end of the tunnel worked turbines acting on machinery for compressing the air required for the drills. The progress of the advanced headings by hand-labour at the commence- ment was necessarily slow, with a maximum monthly advance of 29 yards on the north side, 43 yards on the south side, and 53 yards on the two sides together. The progress, however, increased rapidly with the introduction and improvement of the machinery ; so that the advance, which was only 8G0 yards at the two headings in the first year, 1872-73, during which hand-labour had pre- dominated, increased to 1,816 yards in 1873-74, and reached 2,653 yards in 1874-75, a rate which was only exceeded in 1878-79, when the advance in the year amounted to 2,793 yards at the two headings combined (see Diagram Fig. l,p. 259). The maximum progress in a month was 159 yards on the north side, which occurred in October 1878, and 187 yards on the south side in August 1878, during which month the maximiim advance on the two sides of 304 yards was achieved. The maximum advance in a year was 1,500 yards on the north side in 1878-79, and 1,316 yards on the south, side in 1874-75. The advance was greatest at the northern side in every year except the first, and the portion of 1879-80 pre- vious to the junction of the headings on the 29th of February, ] 880, so that the length pierced from the northern side was 631 yards more than the other. The driving of the headings occupied seven years and five months, only a little more than half the period

' Particulars about the drills and compressors will be found in the Eapports Trimestriels de la lignc du St. Gotliard, and also in Minutes of Proceedings lust. C'.E., vol. xlii. p. 22S, aud vol. Ivii. p. 2'S'J.

Papers.] VEKNON-HAECOURT ON ALPINE ENGINEERING. 267

required at tlie Mont Cenis ; lout as the tunnel is nearly 2 miles longer, the average daily advance was more than twice as great, namely, 6*01 yards as compared with 2-51 yards. When the junction of the headings was elFected, a difference of onl}^ 13 inches was found in the direction of the two lines, and only 2 inches in the level. The length of the tunnel proved to be 25 feet shorter than had been estimated.

The strata traversed were very variable, consisting chiefly of granite, schist, and gneiss with mica in the northern portion, and cpiartzous schist, mica schist, and gneiss in the southern portion ; and veins of serpentine, cipolin, and hornblende were also met with.^ The state also of the rocks varied considerably, lieing much disintegrated and fissured in places. The dip of the strata is considerable throughout, and in some parts is almost vertical, favouring the influx of water, which came in to a eonsideral)lc extent at various places where fissures were jiierced. The tem- perature of the rock in the tunnel rose gradually as greater depths were reached ; ^ and towards the centre of the tunnel in January and February 1880 attained on the average 86^-7, the highest observed being 87'^ "4. The actual temperature in the tunnel corresponded very closely with the predicted temperature ; but the undergroimd temperature showed a greater divergence from the temperature at the surface under the high plateaus than under the highest peaks. The temperature of the rock near the centre of the tunnel was found to have fallen to about 7-i:h^ in May 1882, or about 12° in two years and four months; and it was expected that it would eventually descend to 68°.

The workmen suffered considerably from the effects of the increased temjierature, combined with the humidity of the air, the inadequate ventilation, and the defectiveness of the sanitary arrangements ; and there was a great mortality amongst the horses. No provisions, however, had been made to meet these contin- gencies ; biit the ventilation was gradually improved by introducing more fresh air at the faces, and Mekarski's compressed-air engines were used for removing the excavated material ; ^ and as soon as the headings were joined, natural ventilation took place whenever

' A description and a geological section of the strata traversed by the timucl are appended to Rapport Trimestriol, No. 31, de la ligne du St. Gothard.

^ A diagi'am of the temperatures observed is appended to Rapport Trimcstriel, No. 30.

^ A description and drawings of these engines are given in Rapport Trimestriel, No. 14.

268 VERNON-HARCOUET ON ALPINE ENGINEEEING. [Selected

there was a difference in the atmospheric pressure at the two ends. A supply of fresh air can be procured at intervals along the tunnel from a conduit filled with comjiressed air, to supply the workmen who may have to he in the tunnel when an equilibrium in the atmospheric conditions at the two ends puts a stop for a time to the natural ventilation.

The tunnel was completed and opened for traffic on the 1st of Janiiary, 1882 ; and the whole line was opened on the 1st of June, 1882. The considerably longer period which elapsed between the junction of the headings and the completion of the tunnel at the St. Gothard than at the Mont Cenis, namely, twenty-two months in place of nine months, was due to the enlargement works not being able to follow so closely the headings, driven along the top at the St. Gothard, as the bottom headings of the Mont Cenis.

The tunnel cost about £142 per lineal yard, making a total sum of about £2,327,000 ; so that, in spite of its greater length, its cost was less than four-fifths that of the Mont Cenis tunnel, and only two-thirds of the cost of the Mont Cenis per lineal yard. Accord- ingly, both in rate of construction and cost, the St. Gothard tunnel exhibited a very considerable improvement on the Mont Cenis tunnel.

Aelberg Eailway axd Tunnel. Plate 6, Figs. 1, 5 and 8; and Plate 7, Fig. 5.

Soon after the completion of the St. Gothard railway, Austria, for the third time, undertook an Alpine railway, in order to obtain direct railway communication with France, without traversing any foreign countrj^ besides Switzerland. The Arlberg railway, like the Brenner railway, lies Avholly in Austrian territory ; but whereas, in crossing the main Alpine chain at the Brenner, it was possible to dispense with a summit tunnel, the Arlberg railway, though only crossing an outlj'ing spur of the Alps, rises almost as high as the Brenner and Mont Cenis railways, and necessitated works little inferior in magnitude to those of the Mont Cenis (Plate 7, Figs. 5 and 7). The line commences at Innsbruck, the starting-point of the Brenner railway, and proceeds almost due west, so that it runs nearty at right-angles to the Brenner, and after passing under the Arl mountain in a tunnel about 6| miles long, it terminates at Bludenz near the Swiss frontier, where it joins another railway connecting it with the Swiss lines (Plate 1, Fig. 1).

Arlberg Itaihcaij. — The actual ascent on the eastern side only

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING, 269

really commences at Landeck, 44:} miles from Innsbruck, as the average gradient up to this point is only 1 in 369. Fx'om Landeck to the summit, inside the tunnel, the railway rises 1,750 feet in 20 miles, giving an average gradient of 1 in 60 • 3, the maxima gradients of 1 in 39| to 1 in 45 occupying 10^ miles (Plate 7, Fig. 5). The railway descends 2,465 feet on the west side between the summit and Bludenz, a distance of 19| miles, making the average gradient 1 in 42 • 7 ; but the average gradient amounts to 1 in 36 for over 14 miles from the end of the tunnel to 1^ mile from Bludenz, and the maxima gradients of 1 in 32^ to 1 in 34 extend over 10;-- miles. ^

The line is curved for 16f miles out of a total length of nearly 40 miles ; there is a short curve of 10 chains close to Landeck, and curves of 12 V chains are the most frequent, extending over 8? miles, whilst curves of under 30 chains occupy altogether 15j miles.

There are three tunnels on the eastern slope, having a total length of 341 yards, and six tunnels on the western slope, having a total length of 837 yards ; the longest of them is only 1 furlong long. The principal bridges are the bridge over the Inn near Landeck, 197 feet sjian, and 60 feet above the river; the Trisana viaduct, with a central sjian of 377^ feet, raised 282 feet above the Eiver Trisana, and two side spans of 131{ feet; and the Schmeid- tobel bridge, with a span of 229 1 feet.

The Arlberg railway, between Landeck and Bludenz, has a general average gradient of 1 in 50, which is steeper than that of the Brenner, the Mont Cenis, or the St. Gothard ; but it is more than 10 miles shorter than any of these lines. Its sharpest curve of 10 chains radius is only exceeded by those of the Semmering of 9\ chains, and it has a greater length of curves under 14 chains than any of the other lines. It has the steepest ruling gradient of 1 in 32 J on the western slope of any Alpine line, but the ruling and average gradients on the eastern slope are less than those of the Mont Cenis and St. Gothard (see Appendix).

Arlberg Tunnel. — The tunnel under the Arl mountain, 6}^ miles long, was constructed, like the Mont Cenis and St. Gothard tunnels, in a perfectly straight line, and with the advantage of the experience gained at both of these works ; and like those tunnels, it was formed with gradients rising towards the summit, on the east side 1 in 520, to provide for drainage, and 1 in 72 on

' A plan and a section of the Arlberg railway, with the gradients and curves, and elevations of tlie principal bridges, are given in Zeitschrift des Oester- reichischen Ingeuieur- und Architekten-Vereins, 18S2

270 YEEXON-HAKCOURT ON ALPINE ENGINEEKING, [Selected

the west side to reach the higher level of the summit of the eastern slojie (Plate 6, Fig. 8, and Plate 7, Fig. 5). The driving of the headings was commenced in July 1880, and they were joined three years and foiir months later, namely, on the 13th of November, 1883, so that the average rate of progTess was about 2 miles a year. The Ferroux percussion drill, worked hy com- pressed air, was employed at the eastern heading, and the Brandt grinding rotary drill, worked by water-pressure, at the western heading, both which drills had been previoiisly used with satis- factory results in piercing the tunnels on the St. Gothard. The water-power for working the air-compressors, and providing the water-pressure, was obtained from reservoirs formed at each end. In the first four months the driving was performed by hand till the machines were ready, and the progress was necessarily slow ; but the average daily progi-ess with the machine-drills in 1881, 1882, and 1883, was 4*52, 5*73, and 5-95 lineal j-ards on the eastern side, and 3 "17, 5 -03, and o'9-i yards on the western side (see Diagram Fig. 1, p 259). The rock traversed on the eastern side consisted of hard compact schist, with a large i^roportion of quartz, approximating to gneiss, with a very little water ; whereas on the western side there was a large proportion of mica in the schist, which was fissured with veins of clay and yielded considerable quantities- of water.^ The variable nature of the rock at the western heading, and the tendency to slips, delayed the progress of the Brandt drills, so that they pierced 825 yards less than the Ferroux drills ; whereas when they worked in similar strata, towards the close of the work, their rate of progress was equal. Moreover, the Brandt drill expended less force and less ex2:)losives, and could be worked with fewer men. The heading was driven along the bottom ; and shafts, opened upwards from the heading to the roof at intervals, enabled an upper g-allery to be driven from them in both directions, and thus the enlargement and lining followed close up to the advanced heading.

The Arlberg tunnel cost £1,209,400, which, for a length of 11,027^ yards, is equivalent to £107 18s. 2d. per lineal yard, being a reduction of £34 per lineal yard on the cost of the St. Gothard tunnel, and less than half the cost per lineal yard of the Mont Ceuis tunnel. Moreover, the progress of the Arlberg tunnel headings was half as rapid again as the St. Gothard headings ; and the completed tunnel followed much closer on the heading at the Arlberg than at the St. Gothard, owing to the different method of

Minutes of Proceedings Inst. C.E.. vol. Ixxx. \). 382.

PaiKirs.] VERNON-HARCOURT ON ALPINE ENGINEERING. 271

construction adopted. The Arlberg tunnel, accordingly, exhibits as notable an advance on the St. Gothard tunnel as the latter does on the Mont Cenis, both in rate and cost of construction. The length, however, and the internal heat of the Arlberg tunnel w^ere less than at the St. Gothard.

The railway was opened for traffic in September 1884, more than a year earlier than was anticipated at the outset.

PROrosED Alpine Eailways and Tunnels. Plate G, Figs. 1, 5, and 9 ; and Plate 7, Figs. 6 and 7.

The St. Gothard railway, by drawing a large amount of traffic eastwards, has so affected the traffic and trade of France that schemes, which were dropped for a time when the construction of the St. Gothard railway was decided upon, have again been revived, with the object of trying to bring back a portion of the traffic to its former channels. The three routes proposed for this pur})ose traverse the Alps between the Mont Cenis and the St. Gothard, namely, the Mont Blanc route, the Great St. Bernard route, and the Simplon route, an old rival of the St. Gothard (Plate 6, Fig. 1). They would all be easily accessible from French railways ; the Mont Blanc and Great St. Bernard routes converge vipon Turin, and the Simplon line would join the St. Gothard line at Milan. Merely regarded as routes on a map, the Great St. Bernard route divides the district between the Mont Cenis and St. Gothard the most centrally, as the Mont Blanc route ap- proaches the Mont Cenis, and the Simplon converges on the St. Gothard. The distances between Calais and Brindisi by the several routes are as follows: 1,360 miles by the Mont Cenis, 1,386 miles by the St. Gothard, 1,32-1: miles by the Mont Blanc, 1,326 miles by the Great St. Bernard, and 1,322 miles by the Simplon.

Mont Blanc Railway and Tunnel (proposed?). — The line proposed for the Mont Blanc route would start from an existing railway near Bonneville, a little to the south of Geneva, and following the valley of the Arve would ascend by gradients of from 1 in 259 to 1 in 93, and after passing Sallenches and St. Gervais, would enter a summit tunnel not far from Chamonix, 11^ miles long, and emerge at Pro St. Didier in the valley of the Dora Baltea.^ It

* "Projet d'un Clicmin de Fer International a faibles pentes a travors Ics Appenuincs ct Ics Alpcs jiar la chainc du Mont-Blanc," Joseph Bonelli. Turin,

1880.

272 VERNON-HARCOURT ON ALPINE ENGETEERING. [Selected

would then descend along this latter valley by Morgex, Arvier, and Yilleneuve, with g-radients of from 1 in 06^ to 1 in 82^, to Aosta, where it would join the line leading to Turin and the other Italian railways (Plate 6, Fig. 1). The line of the tunnel passes under Mont Maudit of the Mont Blanc range, at a depth of about 11,390 feet below the surface ; but it would be possible to sink a venti- lating shaft, 1,550 feet deep, into the tunnel, from the Veni valley, 4 miles from the Italian entrance. As the entrance to the tunnel on the French side would be not more than -40 feet lower than the Italian entrance, both the gradients rising to the summit in the centre oi the tunnel would be gentle. The summit proposed is only about 3,280 feet above sea-level, much lower than the summit of any Alpine railway hitherto constructed, except the Semmering ; for it would be 500 feet lower than the St. Gothard summit, and more than 1,000 feet lower than the summits of the Arlberg, Mont Cenis, and Brenner railways (Plate 7, Fig. 7). The strata traversed by the tunnel would consist of calcareous and aluminous schists, crystalline gneiss, and granite. The cost of the tunnel is estimated by Mr. Bonelli at £1-45 -is. per lineal yard, or about £2,938,000; and the total cost of the line, assuming that the authorized exten- sions towards Chamonix can be utilized, is estimated at about £3,480,000 ; but this can only be regarded as a rough approxi- mation, as the scheme does not ajDpear to have been worked out in any detail. Other schemes for the same route have been proposed besides the one sketched out, diifering somewhat in the line of the tunnel and its altitude, and consequently in the length of the tunnel, its maximum depth below the surface, and the gradients of the approach lines. ^ They, however, all indicate a line with gentle gradients and a long summit tunnel.

The route possesses the merit of having a low summit and easy gradients, as compared with previous Alpine railways ; but it has the disadvantage of a tunnel at least as long as the St. Gothard, and in most schemes about 2 miles longer, and reaching an un- precedented depth of at least 9,800 feet below the surface.

Great St. Bernard Bailway and Tunnel (proposed). — The Great St. Bernard railway, as sketched out in some detail by Mr. Vautheleret in 1884, would start from the Martigny station, at 1,545 feet above the sea, and by contouring the valleys with three long loops (Plate 6, Fig. 1), would rise by gradients of from 1 in 62 to 1 in 47 '4, to a height of 5,314 feet above sea-level, in a distance of 36| miles, passing through several tunnels in its

I

' Minutes of Proceedings Inst.'C.E., vol. Ixiii. p. 380; and vol. Ixiv. p. 391.

rapcrs.] VEKNON-HAECOURT ON ALPINE ENGINEERING. 273

course (Plate 7, Fig. 7).^ It would then enter the summit tunnel, 5f\f miles long, under the Col de Ferret, a little to the west of the Great St. Bernard pass, and traverse the houndary between Switzerland and Italy, on the summit of the ridge, at a depth of 3,478 feet below the surface. The gradients in the tunnel, rising towards the centre, would he gentle, as the entrance on the Italian side would be only 7 feet higher than on the Swiss side ; and it is proposed to sink three shafts in the tunnel, 722, 754, and 410 feet deej) respectively. The railway woixld then descend to Aosta, 43^ miles distant, and 1,903 feet above sea-level, with gradients of from 1 in 53 to 1 in 80, passing through several tunnels, and forming long loops near Courmayeur and Morgex. The line is laid out to be in curve for half its length. The sharpest curves on each slope are designed to be 17g chains, and the most frequent curves are 20 chains radius, the remainder being mainly 25-chain curves. The summit tunnel would pierce sandstone and slate, intersected by numerous layers of quartz. There would be fifty- six tunnels on the approaches, having a total length of 16,- miles, and seven viaducts from 240 to 650 feet long and from 72 to 95 feet high, together with numerous bridges. The works, accordingly, would be exceptionally heavy on this line, which, even with its comparatively short summit tunnel, would be in tunnel for a little over a quarter of its total length of 86 miles, irresi:)ective of the 9 J miles of covered galleries proposed on each side of the summit tunnel, where the line rises higher than 4,600 feet above the sea-level, as a protection against snow. The approach works would, indeed, compare with those of the St. Gothard in magnitude, whilst extending over 30 miles greater distance. The estimated cost of the line from Martigny to Aosta is about £3,403,000 or £39,570 per mile.

The Great St. Bernard railway presents a remarkable contrast to the Mont Blanc railway, for whereas the Mont Blanc railway affords easy gradients, short approaches, and a low summit-level, the Great St. Bernard railway takes a circuitous course, with very long approaches, sharper curves, steeper gradients, heavier approach works, and an exceptionally high summit-level, in order to obtain a much shorter summit tunnel. In fact, the gradients are mainly considered in the Mont Blanc scheme, and the summit tunnel in the Great St. Bernard route. The estimated cost is approximately

' A plan and section and other particulars of this scheme are given in Mdmoires de la Socie'te' des Ingenieurs Civils, 1884, vol. i. p. 454, and plates 74, 75, and 70.

[the INST. C.E. VOL. XCV.] T

274 TEEXOX-HAECOUKT ON ALPINE ENGINEEKING. [Selected

the same for both lines ; but the Great St. Bernard line has been examined in greater detail, and is not made dependent, like the Mont Blanc line, upon the completion of somewhat -uncertain extensions.

Simplon Eailway and Tunnel. — The Simplon route has long been the object of special study, and since it was first proposed by Mr. riachat, in 1859, up to the present time, few j^ears have passed without the appearance of some publication dealing with the project. For some time it was a serious rival to the St. Gothard scheme, till the influence of Germany and Switzerland procured the construction of the St. Gothard railway. Though the carrying out, however, of the St. Gothard raihvay diminished the advantages of the Simjjlon route, the scheme was never allowed to drop ; and now that the prejudicial influence of the route by the St. Gothard to the traffic and trade of France, which was feared at the outset, has become an accomplished fact, increased interest has been manifested by Frenchmen in the Simplon scheme, as a means of restoring the equilibrium of their trade, and bringing back the through traffic to the East into French territory. In 1886 the various schemes for forming a railway communication between France and Italy through Switzerland, across the Simplon, were laid before a Commission of experts, who reported their decisions in the same year.^ The scheme, as approved in this report, consists of an ordinary single-line railway, starting from the line to Brieg, near Yisp, in the Ehone Valley (Plate 6, Figs. 5 and 9), 2,133 feet above sea-level, and rising by a uniform gradient of 1 in 50, for 5j-L- miles, to 2,690 feet altitude at Glyss- Brieg station; after which it enters the summit tunnel, nearly 10 miles long, having easy gradients on each side, rising to a summit-level of only 2,773 feet above the sea, as the level of the southern entrance is only 33 feet higher than the northern entrance (Plate 7, Fig. 6). From thence the line descends, near the Swiss boundary", into Italy, to form a junction with the railway at Domo d'Ossola, 14^ miles distant, with gradients of 1 in 40 between the stations, down to 896 feet above sea-level. The average gradient from Visp to the summit is 1 in 90 • 6, much easier than any of the existing Alpine lines ; and from the summit to Domo d'Ossola it is 1 in 54-2, steeper than the southern sides of the existing lines, but easier than the average gradients on the northern sides of the Brenner and the Mont Cenis, and than the west side of the Aidberg railway. The total average gradient of

' Eapport lies Experts sur le Percemeut du Simplon. Lausanne, 188G.

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 275

the Simplon line is 1 in G3 • 2, for a total length of about 30 miles, which is easier than the average gradient of any existing Alpine line ; and it extends over a much shorter distance than any of the other lines. The line, as laid out, is curved for less than one- third of its length, but the sharpest curves, of 15 chains radius, extend over 3i miles ; and, as regards curves up to 20 chains radius, it is less favourably laid out than the Mont Cenis (Appen- dix). The line passes through one tunnel on the northern slope, and through seven tunnels on the southern slope, with a total length of 1 mile 717 yards, the longest tunnel, as designed, being 492 yards long. Except close to Glyss-Brieg station, the heavy works will be confined to the southern slope (Plate 7, Fig. 6).

The summit tunnel will traverse gneiss, granite, mica-schist, limestone, and cipolin. It is proposed to make it in two straight lines from each end, meeting at an angle near the centre, so as to avoid the highest ridges of the mountain, and reduce the thickness of superincumbent rock, with the object of keeping clear of exces- sive underground temperatures. Nevertheless, the surface of the ground, at its highest j)oint, will be 6,895 feet above the tunnel, 1,102 feet more than over the St. Gothard, and with a tunnel f mile longer. After careful consideration, however, the experts came to the conclusion, with the experience of the Mont Cenis and St. Gothard tunnels before them, that though the temperature of the rock in the Simplon tunnel might exceed the maximum in the St. Gothard tunnel of 87° -4 for a distance, in the centre, of 2^ miles, and even reach 100° -4 to 104° for 1^ mile, it would be possible to execute the work by special methods of ventilation, and by cleansing and cooling the air at the faces. ^ It is pointed out in the report that the progress achieved in electricity would render the transmission of power much easier, and would afford a more complete illumination for the work in the tunnel.

The cost of a tunnel for a single line is reckoned at £1,877,000, or, for a length of 17,573 yards, £106 16s. per lineal yard, to which is added a sum of about £80,000, to provide for the cost of reducing the temperatiire in the central part of the tunnel, in the event of its being high. The total cost of the line, from near Visp to Domo d'Ossola, is estimated, for a single line throughout, at about £2,118,000, very considerably less than the estimated cost of the other two schemes.

The Simplon railway would afford the shortest route between Paris and Milan; but the St. Gothard line would be shorter for

' Eappoi'ts clos Experts sur le Pcrccmcnt du Simplou, p. 28.

T 2

276 VERNON-HAKCOURT ON ALPINE ENGINEERING. [Selected

going between Belfort and Milan, Bale and Milan, Belfort and Genoa, and Bale and Genoa. The actual distance between Boulogne and Piacenza would be shorter by the Simplon ; and the virtual distances, allowing for the gradients, would be 856 miles by the Mont Cenis, 829 miles by the St. Gothard, and 812 miles by the Simplon.

Comparison hetween the three Alpine Schemes. — The Simplon route possesses the lowest summit-level of any existing or proposed Alpine railway (Plate 7, Fig. 7, and Appendix), and the Great St. Bernard considerably the highest; whilst the Mont Blanc tunnel would be the longest, and at much the greatest depth below the surface. The Great St. Bernard route has much the longest approaches, with the heaviest works, but a tunnel of about half the length of the others ; the Mont Blanc has the easiest gradients ; and the Simplon, with the steepest gradients of the three, has a length little over a third of the Great St. Bernard line.

The Mont Blanc line, with its easy gradients, might compete successfully for traffic with the Mont Cenis ; but, from its position, it could have little influence in diverting the traffic from the St. Gothard. It would be of little use to Italy, which is adequately served in that quarter by the Mont Cenis ; and its only advantage to France would be in expediting somewhat its traffic, and as this could only be done by diverting the traffic from the Mont Cenis, it does not appear worth the expenses it would entail. The only place of importance it would decidedly benefit is Geneva, by affording it more direct access to Italy. Moreover, the great length of tunnel, the uncertain character of portions of the strata to be traversed, and especially the excessive heat that might be experienced at the gTcat depth below the surface which the tunnel would have to traverse, might well cause hesitation in attempting even the most desirable scheme in other respects. The little advantage that the Mont Blanc route ofiers to France, and the excessive heat liable to be encountered in constructing its summit tunnel, appear to the Author adequate reasons to preclude its adoption, in spite of its easy gradients.

The Great St. Bernard line, with its comparatively short tunnel, might have offered inducements when the experience in long tunnels was small ; but its very long ascents render it unsuitable for a competing line ; whilst its high elevation would render it very liable to be blocked by snow-drifts, and therefore not advan- tageous for a through route, whose prosperity depends upon its regularity. Its position, liowever, is more favourable than that of

Papers.] VERNON-HARCOURT ON ALPINE ENGINEERING. 277

tlie Mont Blanc roiite for opening ont fresh communications, and it would be of more value to Switzerland.

The Simplon line, though less intermediate between the Mont Cenis and the St. Gothard than the Great St. Bernard route, is well situated for Switzerland, and for opening up fresh lines of traffic by Milan ; and, from its very proximity to the St. Gothard, is specially well suited for competing for its traffic, and drawing it back over French railways. Its low summit-level is particularly favourable for quick through traffic ; and, as pointed out above, it would shorten the distance from Paris and Boulogne to Brindisi, and thus probably secure the carriage of the Eastern mails. More- over, its construction is estimated to cost much less than either of the two other schemes. The only objection that can be raised against the route is the possible high temperatures which may be encountered in constructing the summit tunnel ; but with special provisions for meeting this contingency, it does not appear sufficiently formidable to bar its execution, as might be the case under Mont Blanc. The prospects the Simplon route offers to France appear adequate to induce that State to afford financial aid, without looking for a direct return from the capital invested ; for the State benefits by the prosperity of the nation, and a restoration of lost trade and traffic to France would fully compensate for a considerable outlay.

The Paper is illustrated by three sheets of tracings to a small scale, from which Plates 6 and 7, and the Fig. in the text, have been engraved.

[Appendix,

278

VERNON-HARCOUET ON ALPINE ENGINEERING. [Selected

I— I P

Ph

<1

	m '='	t; i^	CO	-H	r^	0		0
	■53	0 i^	Cfc	IC'	-+1	T
o	6^	S 0	<M	C5	0	<M		•"^
eft	E'g t.			■*	»o	»o		(M
	•5 « ■s =■	s Si			i>	05
	^ S 3	S 1-1	t-	1— 1	I-H	'"'		CO
witl ins. and der		t^		(M	CO		C5
^ lO	CO		CO	>o		I-
3	0- =	S '^	t^		(M	(M
	CO			s	"* T-H
	p •	g >o			3t3 I. '-'
	S"'	g ^			■3 0	3
•jnaipBjg	H C-	a lo	a --1	a c^	ao		a <M
93BjaAy i^jox	_^ 0			— 00 a'^	^ 0		-g
			^ ^	a=^	a®	a^		a=^
	■a	83BJ9AV		■:;s		"" 0 rH CO	"^ 0 I-H 0 • a^ .a rH CO	d	rH lO
■}n3tpBJ{) ummixBj\[		a9 I— 1 Tt(	a9	a9 T-l CO	a? '"* 0 rH Tfl
	ja						02^---' ^ 0	^	t-
	0	VTi	CD	tn
			■S ^H	(N			m m
	&	•asiH	V 0	CO	00		A "^	a
						H	0
						^		02	y—^
	.^ <»	0	!M	(M	CO
m		•qjSaa'T	S -^	ao	(N	CO	0		<M
c			1 t-				0
		CO	C^	CO	!N
		-•^-'^^	/— '^— ^	^_^^*__^	^_A_^	'— '^— N		^— -^'— ^
		•jnaipBjf)	a 30	a c»	ai-H	a <M	a t-		a CO
O		a2BJ3Ay		-^			c5.a'P		. 0
•jnaipBjf) ranraixBjt	■" 0	a9 "^ 0	a?^ "^ CO	a"-?	a? "^ 0 rH 0
			CO	CO	t-	»o	"S '^	0
	i-,		■s <=>	00	CO	IM	s <»		-M
	■A	•asia	g 10	>n	CO	(M	IT-	0 '?',	CO
				(M	(>>	0
	1— (	t-	TjH	CO
		•q^Snai	S 05	10	«o	t-	OS		CO
			i ^	(M	■*!	CO	a>
			C^	(M	(M CO			10
	0		CO
'laatrnx Jsao aoBjang	■S S		t^	2?			Ci
JO iqSpH Isa^TOJO	fe		1«	10
	•pjBA			■*	<M	00
=3 ^		«ti		cq	TtH	0		0
11	jsd }soo			(M	T-H	rH		T^
	«; <»		t^	,_(	CO		<Xi
3 3	•q^uai	g CO		9	eo	CO		05
		1 ®		t-	ci	i		05
	!M	t^	CO	CD	05		CO
■[3A3'3-Bag 3A0qB	■s s	05	S	00	Ci		t-
^iraoins JO !!q3[aH	0 00	■*!	eo	t-	(>q		l>
	(M	tJ<	■*	eo	Tt<		(M
	0 tjD	0 ■	c 0 ■	0	0	'~	0 ■
	-« o3		1 '* "^				a
	4^1 a a a	ill ^ a ^	i « S a ^ a i	1	1 s-s	a
	OQ	PQ	S	02	<	02
<« c
	CO -t<	-H t^	r^ rH	0<l (M	0 •*!
« g"-g	-+< 0 in	CD 0 CO	LI 0 t-	1^ 0 00	00 0 CO
^""l	CO ■♦^ X!	OO -1- CO	yj ->-' CO	00 -*i CO	CO -tJ 00
	-, ^	^ r-.	I— ( f-H

Papers.] MACALISTER ON THE RIVER CLYDE. 279

(Paper No. 2341.)

"The Eiver Clyde." By Daniel Macalister, Assoc. M. Inst. C.E.

(Abstract.)

In this Paper the Author describes some of the changes which have taken place in the navigable and in the subsidiary channels, and in the shoals of the Clyde from Dumbarton to Greenock, between the years 1860 and 1880. He also makes a few remarks ^^23on tidal scour.

The first part of the Paper is based upon surveys made in the above-mentioned years by order of the Admiralty, of which the charts that have been issued afford a ready means of comparison. The history of earlier changes and improvements will be found in the Paper on " The Eiver Clyde," by Mr. Deas.^

The main features of these charts show a navigable channel near the centre of the river at Dumbarton, approaching the south bank as it proceeds seawards to Port Glasgow, from whence it continues along the bank until it reaches the estuary at Greenock. On the north side, a shoal extends from Dumbarton to opposite Greenock, and is divided by two subsidiary channels into three parts, named respectively, the Pillar, the Cockle and the Greenock banks.

The navigable channel is defined as that in which the depth is greater than 12 feet at low-water; and a good idea of the state of this channel will be obtained by examining the distance apart of the 2-fathom lines upon either side of the channel.

In 1860, the navigable channel was very irregular and tortuous, and presented considerable difficulties to safe navigation. The width at Garvel Point, Greenock, was only about 190 feet ; at Port Glasgow, the 2-fathom lines touched one another; and at Garmoyle, ^ mile below Dumbarton, the width was 180 feet. The depths at these points at low-water were respectively 14, 12, and 13 feet. Above Port Glasgow, a second channel, known as the " Old Channel," had a depth of 11 feet, or only 1 foot less than the main channel.

With regard to the subsidiary channels and shoals on the north

Miuutcs of Proceedings lust. C.E., voL xxxvi. p. 124.

280 MACALISTER ON THE EIVER CLYDE, [Selected

side in 1860, the channel between the Pillar and Cockle banks had a minimum dejith of 7 feet at low-water for a distance of 1^ mile, and a maximum depth of 17 feet ; and it seems probable that at no very remote time it was the main channel of the Clyde, or at all events, of equal imjDortance to that which occupied the site of the present navigable channel. The channel between the Cockle and Greenock banks had a minimum depth of 4 feet at low-water. The Pillar bank extended from the mouth of the Leven at Dumbarton, along the north bank of the river to Ardmore Point ; and the Cockle bank from opposite Port Glasgow for nearly 1^ mile along the northern side of the navigable channel. This bank had an area of 1,160,000 square yards above low- water, and its greatest elevation was 3 feet ; while the Greenock bank had an area of 1,261,500 square yards above low-water, and the greatest elevation was 2 feet.

In 1880, the irregularities in the navigable channel had been in a great measure removed ; and from Greenock to Dumbarton there was a fairly uniform channel with easy curves, presenting few difficulties to navigation. The following is a comparison of the state of the channel at certain points in 1860 and 1880 : —

Width between ; Depth below Width between Depth below 2-fatbom lines, i Low-water. 2-fathom lines.' Low-water.

Feet. Garvel Point . . ! 190

Port Glasgow . . . ' Nil ISO

Garmoyle, h mile below 1 Dumbarton . . . /

In 1860, a vessel drawing more than 12 feet could not get further up the Clyde at low-water than Port Glasgow; whereas, in 1880, a vessel drawing 16 feet could get up to Garmoyle.

The " Old Channel " at Port Glasgow had only 8 feet at low- water, having silted up 3 feet in twenty years.

The channel between the Pillar and Cockle banks had silted up to a remarkable extent ; but the " Pool," which had been 17 feet deep in 1860, was in 1880 only 1 foot less, which the Author attributes to a whirlpool action caused by the ebb of the river meeting the first of the flood-tide at this point. Among other changes, a slight scour had taken place in the Pillar bank, liut the area of the Cockle bank above low-water had been reduced to about one-half

Pfipcrs.] MACALISTER ON THE RIVER CLYDE. 281

what it was in 1860, and the whole of the Greenock bank above low-water had disappeared.

These changes are so great, and their effects on the navigable channel below Garvel Point so important, that the Author jiroceeds to investigate their probable cause, by considering the tidal capacity and the scour of the river. He estimates that in 1860, the capacity of the Clyde above Garvel Point was 2,750,000,000 cubic feet ; and that the sectional area between Garvel Point and Ardmore Point at low-water was 56,000 square feet, and at high- water 158,000 square feet, the mean being 107,000 square feet. The mean discharge of the upland water, or river-flow, was estimated by Mr. Ure at 48,000 cubic feet per minute. The duration of the flood-tide at Port Glasgow, 2 miles above Garvel Point, was six and a quarter hours, and of the ebb six hours ; so that the total flow of upland water during the flood was 18,000,000 cubic feet, and during the ebb 17,280,000. The volume of water l)assing Garvel Point on the flood-tide was therefore 2,732,000,000 ci;bic feet ; the mean velocity being 1 •1348 foot per second ; and on the ebb the numbers were respectively 2,767,280,000 cubic feet, and 1 • 1973 foot per second.

The Author further estimates that the tidal capacity of the Clyde had increased in 1880 to 2,804,000,000 cubic feet; that the vohime of water passing Garvel Point on the flood-tide was 2,786,000,000 cubic feet, and on the ebb 2,821,280,000 cubic feet ; that the sectional area of the river at this point at low-water had increased to 66,500 square feet, and at high-water to 159,500 square feet, the mean being 113,000 square feet; and that the mean velocity of flood and of ebb had diminished respectively to 1'0947 foot and 1*1559 foot per second. In other words, at Garvel Point in 1880, the tidal capacity had increased 2 per cent., and the mean sectional area of the river 5 • 6 per cent. ; but the mean velocity of the flood had decreased 3 • 6 per cent., and of the ebb 3 • 5 per cent.

From the foregoing it is evident that although the imjirovement of the navigable channel has increased the tidal capacity of the river in the upper reaches, this gain is more than counterbalanced by the increased sectional area of the Clyde at Garvel Point, due to the scouring away of the Greenock bank ; and the same remark applies in a lesser degree to the Cockle bank. Simultaneously the velocity of the current in the channel between the Pillar and Cockle banks was reduced ; the deeper portions formed reservoirs and silted up, and the shallower portions were deepened.

These great changes are illustrated by sections, which are

282 MACALISTER ON THE RIVEK CLYDE. [Selected

preserved for reference. They show generally that the up-stream sides, and the surfaces of the banks had been scoured away to a considerable extent, and that the materials had been deposited on the down-stream sides of the bank, opposite Greenock. The same action had been going on further down at Eoseneath Patch, a shoal in the centre of the estuary. With the increase of the sectional area from Greenock to Ardmore Point, due to the improvement of the navigable channel and the scouring of the bank, the velocity of the ebb woiild be reduced ; consequently silting had taken place in the north or convex side of the navigable channel opposite the Albert Harbour, Greenock ; and scouring had been caused by the flood-tide on the northern side of the Pillar bank.

The Author anticipates that unless the subsidiary channels are dammed up, and the faces of the banks forming the northern side of the navigable channel are protected by stone to above low-water mark, from above Port Glasgow to Greenock, the scouring away of the Greenock and Cockle banks will continue, until the sectional area has so much increased that this action ceases owing to the diminished velocity of the tide ; but as the velocity in the navigable channel will be correspondingly reduced, silting will take place in it.

By multiplying the square of the mean velocity of the river by the volume, the Author estimates the scouring power of the el)b- tide, at Garvel Point in 1880, to have been 13 per cent, greater than that of the flood-tide.

Papers.] FAILURE OF THE KALI NADI AQUEDUCT. 233

(Paper No. 2358.)

" The Failure of the Kali Nadi Aqueduct on the Lower Ganges Canal."

(" Selectious from the Records of tlie Grovernment of India. Public Works Department. No. ccxl. 1888.")

Abstracted by Walter Hampden Thelwall, M. Inst. C.E.

The following is an abstract of a series of reports and other official documents relating to the aqueduct for carrying the Lower Ganges Canal over the Kali Nadi stream. This aqueduct was designed in 1870-73, built soon after (the date of construction is not given), partially destroyed by a flood on October 2nd, 1884, and completely swept away by another flood on January 17th, 1885.

Original Design.

The drainage-area of the Kali Nadi above the aqueduct was estimated at 3,025 sqxiare miles. Very few data were available for estimating the maximum flood-discharge, but it was assumed that the greatest rainfall in twenty-four hours would be 6 inches, and if one-fourth of the fall were to flow off in the same time, the dis- charge would be 38 cubic feet per second per square mile of drainage-area, or a total of 114,950 cubic feet per second ; whereas the greatest recorded flood (calculated apparently from sections and high-water marks) was 20,382 cubic feet, or 8-7 cubic feet per second per square mile. That is to say, if the maximum rainfall was G inches in twenty-four hours, only ^ inch, or one-eighteenth of the fall, would be discharged in the same time. This difference between the estimated and the actual discharge was considered at the time to be due to two causes : first, that the greatest rainfall in any locality is not continuous over large areas, and secondly, that owing to the flatness of the district, the rainfall from the distant parts of the area must take a considerable time in reaching the point of discharge. It was considered that a sufficient margin would be allowed if the discharge were taken at 12 cubic feet per second (instead of 8-7), giving a volume of 30,300 cubic feet. To dis-

284 FAILURE OF THE KALI NAM AQUEDUCT. [Selected

charge this at a velocity of 10 feet per second, seven spans of 35 feet by 15 feet were provided in the first design, and it was calculated that during high floods the water up-stream would be headed up 2 feet behind the bridge.

The volume of water in the canal to l:ie carried across the aqueduct was 5,374 cubic feet per second ; the depth was fixed at 9 • 4 feet, and the width was determined by the consideration that the velocity should not exceed that against which boats could be towed by ordinary means. This velocity was taken to be 4^ feet per second, or 3 miles per hour, and the width of channel was accordingly fixed at 127 feet. A roadwaj', 12 feet wide, was also provided on one side of the canal.

Revised Design.

When the design was submitted to the Sujierintending Engineer, this officer rej^orted that he considered the waterwaj' proposed larger than was necessary. He calculated the highest known flood at 9,500 cubic feet per second, or 3*66 cubic feet per second per square mile, a revised survey of the drainage-area showing this to be only 2,593 square miles. This discharge was admitted to be very small, but was accounted for by the fact that the soil was very- sandy, and that the drainage was intercepted at various places by public works. These figures were moreover confirmed by the fact that about ^ mile down-stream there was a native bridge more than a century old, having seven openings of 10 feet 6 inches span and 14 feet 6 inches high from floor to flood-level, and two side openings 8 feet by 5 feet, giving a total water area of 1,146 square feet. It was found by watermarks that the stream was headed up 1^ foot, the velocity of approach was taken as 1 • 48 foot, and the discharge calculated (taking c = 0-6) at 8,436 cubic feet per second. A further check was obtained from measurements taken at a point 39 miles above the bridge, during a flood said to be the highest known. Adding to this measured volume a proportion for the additional drainage-area at the site of the aqueduct, the volume here was estimated at 7,033 cubic feet per second.

In spite of these calculations, however, the Superintending Engineer considered that a discharge of 3*66 cubic feet per second per square mile of catchment-basin was too small to be relied on, and determined to provide for 7 cubic feet, a qiiantity which he stated was often used in dealing with long and large catchment- basins, and, on such a sandy soil, might be adojited without risk.

Papers.] FAILURE OF THE KALI NADI AQUEDUCT. 285

The aqueduct and adjoining river-lied, if properly protected, might also easily stand a current equal to that which was supjwsed to have passed under the old road-bridge, viz., 7^ feet per second. From these data the waterway was fixed at five openings of 35 feet by 14 feet, or more than double that of the old bridge 1 mile lower down the stream, and the discharge estimated at about 18.000 cubic feet per second.

The width of the canal carried by the aqueduct was increased to 192 feet, so as to diminish the velocity to 3 feet per second, as it was considered that 4^ feet per second would be dangerous to the earthen embankment at the end of the aqueduct, and would involve a heavy expenditure in pitching the slopes and bed of the canal. The aqueduct was accordingly built to this design, namely, five segmental arches, each 35 feet span ; height, from springing of invert to sjmnging of arch, 14 feet 8 inches ; versed sine of invert, 1 foot 4 inches, and of arch, 4 feet 8 inches. The reduced levels of the principal points were as follow : —

Level of zero of flood-gaugro.

The abutments and piers were 212 feet long, and were founded on wells sunk in sand to a depth of 19 feet below the bed of the Nadi, or 514*5, with curtain wells sunk 10 feet.

A flood-gauge was fixed on the down-stream side of the aqueduct, the zero being fixed at the level of the springing of the invert.

Previous to October 1884 the highest flood recorded rose to 16*3 feet on the gaage, or E.L. 551*14.

Partial Destruction of the Aqueduct.

On October 2, 1884, the aqueduct was partially destroyed by a flood which rose to 18*7 feet on the gauge, and was headed up 3-5 feet higher (or 22*2 feet) on the up-stream face. The discharge while the aqueduct was still standing was calculated at 37,000 c^^bic feet per second, and the mean velocity (including a velocity of approach of 5 feet per second) at 11-7 feet, c being

Crown of invert* .	. . 533-50
Springing of „	. . 534-84
,, arch .	. . 549-50
Crown of arch .	. . 554-19
Canal-bed .	. . 558-07

' This was not a brick invert, but block kunkur pitching 2 feet tliick, grouted.

286 FAILURE OF THE KALI NADI AQUEDUCT. [Selected

taken at 0 • 74. The discharge, after the Ijreach in the work was made, was roughly estimated at 4-i,000 cubic feet per second. The destruction began by the stream, as it rushed through the openings, washing away a part of the kunkiir floor, and then undermining the wells, when about one-fourth of the structure gave way and fell into the river. Nearly all the curtain wells were washed away. Out of the seventeen foundation wells in each pier, one pier lost twelve, another five, and another one, while one pier and the two abutments remained intact. In the worst ydace, however, there remained a width of about 50 feet of the canal bed, and advantage was taken of this to form a temporary aqueduct, by bxiilding a revetment-wall for the canal, and in three months and a half after the accident occurred a reduced volume of 1,500 cubic feet per second was passed down the canal.

A flood of 44,000 cubic feet per second represented for the drainage-area of 3,025 square miles, a discharge of 14*7 cubic feet per second per square mile. The old road-bridge which had been relied on as justifying the waterway for the aqueduct was not injured, but the approaches at both ends were carried away, and the water poured over the road for a length of ^ mile and cut through the embankment in several places.

When the extent of the damage had been fully investigated, it was decided that it was useless to attempt to repair it, but that the remains must be removed and a new aqueduct built consisting of nine spans of 35 feet. No reasons are given for the adoption of this waterway, nor does it appear that any further investigations were made in regard to possible floods in the future.

Complete Destruction of the Aqueduct.

On iTuly 17, 1885, a far heavier flood than that of October 2, 1884, came down the stream and completed the destruction of the aqueduct.

Mr. Good, the Executive Engineer, witnessed this catastrophe, and described it in his report on the subject. Between noon and 6.30 P.M. on July 16th the down-stream water-level rose from 546-4 to 549-0. By 4.30 a.m. on the 17th it had risen to 553-20 down-stream, and to 559 - 0 up-stream, but no damage had appa- rently occurred to the work. Immediately after, however, the water-level up-stream rose 4 feet in one wave, and the work began to fall in. At about 5 a.m. it became evident that the canal channel was doomed, as cracks began to appear in the new right revetment-wall, and the arch rings of Nos. 3 and 4 arches Ijegan to

Papers.] FAILURE OF THE KALI NADI AQUEDUCT. 287

separate from the backing. The left revetment-wall and roachvay over arches Nos. 3 and 4 and part of the arches themselves subsided at 6.15 a.m. By 8 a.m. the flood at up-stream side had risen to w^ithin 6 inches of the top of the new right revetment-wall (567-50), and surged over it in waves 10 or 15 feet high, and then the whole of this revetment, from abutment to abutment, and arches Nos. 3, 4, and 5 collapsed. Arches Nos. 2 and 1 were blown up shortly after, but leaving about 50 feet of the old revetment-wall supported by the piers, which obstructed the passage of the flood, and was at 9.30 a.m. the only vestige of the structure visible. The surface velocity of the current, taken by timing floating bodies through the 219 feet length of the abutments, was 18*25 feet per second.

The action of the stream below the aqueduct was terrific. The water in the main stream was piled up in waves 20 feet high at intervals of about 100 feet apart, and laterally, especially towards the right, spread out into a formidable swirl which was fast cutting away the outer slope of the bank.

Ultimately the water cut its way through the canal-banks on both sides of the bridge for a length of 50 feet behind the right abutment and about 300 feet behind the left, and the only parts of the aqueduct left standing were about two-thirds of one abutment and two wing walls.

The maximum up-stream flood-level was about 569-00, the down- stream level at the time being 556-00, the water being headed uji 13 feet. The gap behind the left abutment was 100 feet, that behind the right abutment, 50 feet. Before the arches gave way the up-stream level was 566-00 and the down-stream 556-00. The length between the up- and down-stream faces of the aqueduct was 212 feet.

The rainfall in the district on July 16th was 17-6 inches, and 3 inches more fell on the 17th, the whole 20 inches having fallen in a little more than twenty-four hours.

Design for new Aqueduct.

Although engineers were on the spot diiring the catastrophe and noted the facts as they occurred, they were unable to obtain sufficient data from which to calculate accurately the volume of the flood. For instance, though the lengths of the gaps behind the abutments were measured, it was obviously impossible to obtain sections of them, and hence the sectional area of the flood

288 FAILURE OF THE KALI NADI AQUEDUCT. [Selected

could only be approximately estimated. The dimensions taken for the calculations were : —

Width between abutments 5 X 35' + i X 7' = 20.3 feet.

Breacli behind left abutment 100 ,,

„ right ,, 50 ,,

Up-stream flood-level 569-00

Down-stream „ 556 '00

Bottom of stream 532 • 00

The breach behind the left abutment must have been obstructed by the fallen masonry of the heavy land wings, and by the kunkur pitching on the left bank of the Xadi, and that behind the right abutment by the up-stream land- and river-wings, which were still standing, so that the width of the breach was taken at

( 200 -| ^ ) = 275 feet. The volume was then calculated,

taking the width as 275 feet, depth at the down-stream face 24 feet, heading up on up-stream face 13 feet, velocity of approach 3 feet per second. Taking c at 0 • 5, the volume was found to be 132,475 ciibic feet per second. To provide for a similar flood in future, a design was prepared for a new aqueduct, having eleven spans of 60 feet. The soil in the bed of the Nadi under the aqueduct, and for 50 feet above and below it, was to be dredged out to 15 feet below the bed, or E.L. 519, and a sunken floor laid of heavy concrete blocks 5 feet in thickness, the surface-level of the floor being E.L. 524. The waterway available down to this level, with a velocity of 8 feet per second, would have been sufficient for a flood of 137,456 cubic feet per second, equivalent to 58 cubic feet per second per square mile of drainage-area, estimated at 2,377 square miles.

The length of the aqueduct being increased from 200 to over 700 feet, it became a matter of great importance to reduce the width, and it was necessary to consider what depth and velocity could be given to the water in the canal consistently with the safety of the earthen channel along the embankment. This channel was 192 feet wide at the bottom, with slopes 1^^ to 1, and the gradient was 0*5 foot per mile or 0*095 per 1,000. The supply required was 4,100 cubic feet per second. After making a series of calcxxlations as to the relative volumes which could be passed down the earthen and masonry channels, it was decided to allow a depth of 8 feet, a width in the latter of 130 feet, and a gradient of 0'15 per 1,000, giving a supply of 4,111 cubic feet per second at a velocity of 3 • 95 feet. In the event of the depth being increased to 9 feet, which was the greatest that the earthen

Papers.] FAILURE OF THE KALI NADI AQUEDUCT. 289

channel could carry, tlie mean velocity in the aqiiednct would be 4-21 feet and the discharge 4,926 cubic feet per second. A reference to the Solani Aqueduct on the Upper Ganges Canal showed that the mean velocity with the full supply was 4*24 feet per second, and proved that the proposed velocity of 4' 21 feet was not excessive. To protect the earthen channel it was decided to pitch the slopes for a distance of 310 feet above and below the aqueduct, and the bed for a distance of 310 feet below and 100 feet above.

When the design was completed it was submitted to Colonel Brownlow, E.E., Insj:)ector General of Irrigation, who, on examining the reports of the flood and the calculations founded on them, considered that the discharge had been greatly underestimated. With the data previously given he was of opinion that the co- efficient c should have been taken as = 0 • 75 at least, instead of 0 • 5, which would add 50 per cent, to the estimated volume.. However, he considered that it would he sufficient to give fifteen spans of 60 feet. These spans he proposed to divide into three sets of five each by two abutment piers, and this design was adopted. In addition to the width of 130 feet for the water channel, a road 11 feet wide was provided on one side of the canal and a bridle-path 6 feet wide on the other, the whole width from face to face of the arches being 148-7 feet.

In determining the number of foundation-wells and the depths to which they should be sunk, it was decided that the pressure on the foundations should not exceed 2^ tons per square foot, and that 140 lbs. per square foot might be taken as the frictional resistance of the wells in undisturbed soil. Where the ground had been previously dredged out no reliance was to be placed on frictional resistance. Each abutment and pier-abutment was to be founded on a double row of wells of 12 feet diameter, and each pier on a single row of 20 feet wells. Ultimately the calculations as to stability appear to have been used only to ascertain the minimum depth allowable, and it was determined to sink all the wells to a depth of 50 feet. The total number of wells will be 268, and the aggregate depth of sinking 15,000 feet. The backs of the abutments and river wings will have a row of sheet piles 15 feet long and 4g inches thick driven behind them to prevent any settlement of earth.

The available height from the bed of the Nadi (E.L. 534) to the bed of the canal (558-27) is 24-27 feet, and has been disposed of as follows : — Concrete over arch 0-35 foot, arch ring 4-15 feet, versed sine of 60^ arc 8 feet, Nadi bed to springing 11-77 feet.

[the INST. C.E. VOL. xcv.] u

290 FAILITKE OF THE KALI NADI AQUEDUCT. [Sc4ected

The sunken floor will be laid with its upper surface at E.L. 524, or 10 feet below Xadi bed, to a distance of 25 feet above and below the faces of the aqueduct, and will end with aprons 25 feet long, slo2:)ing downwards at an angle of 1 in 6 to E.L. 519, so as to protect the main portion of the floor. The excavation for this floor is to be done partly by pumping and ordinary excavation, and partly by dredging from barges by hand and by steam-power. The thickness of the arches varies from 4*15 feet at the crown to 4*58 feet at the springings. Spandrel arches of 4 feet span, 0"83 feet thick, with piers 1 • 7 foot thick, Avill be built. Over them will be a course of flat bricks in Portland cement. The floor and sides of the channel will be rendered with Portland cement to make them watertight. In building the arches each set of five will be turned separately. There will be five sets of centerings, each about 32 feet long, or eqiaal to one-fifth of the length of the arch. Each arch will therefore be built in five sections, the difierent sections l)eing joined together by masonry wedges built afterwards.

The whole site of the works will be enclosed by embankments raised well above maximum flood-level, so that there can be no interruption to the work except during the periods of actual rainfall.

The estimate for the aqueduct itself is Es. 30,52,000, but other subsidiary works, including canal and Xadi diversions, railway and sidings for l;>ringing materials, buildings, tools, plant, establish- ment, and miscellaneous expenses, bring up the total estimated cost to Es. 49,98,810.

The collection of papers concludes with a report showing the amount of work executed up to the end of October, 1887.

Vapers.] LOPES ON THE REPARATION OF BETCHWORTH TUNNEL. 291

(Paper No. 2338.)

" The Reparation of Betchworth TunDel, Dorking, on the London, Brighton and South Coast Eailway."

By George Lopes, B.A., Camb., Assoc. M. Inst. C.E.

The Betchwortli Tunnel, on the direct Portsmouth line of the London, Brighton and South Coast Railway, is about j mile south of that company's Dorking station, in the Lower Greensand formation, 136 feet at its maximum depth under Deepdene Park.

The tunnel is straight, 385 yards in length, and has a rising gradient of 1 in 80 southwards towards Holmwood. It was originally designed with slightly curved side walls 1 foot 10^ inches thick, and with a segmental arch of similar thickness, and, except for a length of 12 feet near the north face, an invert was dispensed with.

In the northern portion of the tunnel the greensand is quite hard, in some places approaching rock, presenting a straight and solid face, and needing but the merest skin of brickwork lining to secure its retention in place ; but gradually this changes to sand, mixed with boulder-stone, which again gives place to sand piire and simple, of the finest quality, wholly dry, and devoid of the least cohesion. Indeed, with the exception of a portion about 25 yards long, 40 yards from the north entrance, no water was found.

The original tunnel was opened for public traffic in May, 1867, and formed the most important work on the Horsham and Dorking section of the railway.

During the twenty years that had elapsed since the construction of the tunnel until the collapse, on the 27th of July, 1887, when, without warning, a portion of it fell in, only ordinary superficial repairs of decayed bricks were necessary or were made. In fact, repairs of this character were in progress in a portion of the tunnel when the collapse occurred ; but no kind of doubt had been enter- tained of the perfect security of the work, as it exhibited through- out every sign of stability and solidity. The sudden failure is all

u 2

292 LOPES ON THE KEPAKATION OF BETCHWORTH TUNNEL. [Selected

the more surprising as tlie summer had heen exceptionally dry, and in that part of the tunnel no repairs were, or had been, in hand. A failure was first suspected at about 5 p.m. on that day by a foreman of plate-layers, who noticed that the brickwork at about 70 yards from the south end was badly cracked, and appeared to be giving way, a fine stream of dry sand coming through the fissure ; the fracture extended longitudinally, at a height of about 12 feet above rail-level.

At the time this was noticed a train from Horsham was due, and having waited to see it through the tunnel, the man, observing that the crack was increasing, at once ran out towards Dorking station and stopped a " do'mi " train which was about to enter ; on his return the tunnel was wholly blocked.

Early on the 28th of July the tunnel was visited by the Chief Engineer, accompanied by the General Manager of the railway and the Author. From the north end up to the stoppage the brickwork appeared sound and intact ; but here the tunnel, from rails to soffit, M^as completely filled with fine dry sand, which had been probably driven down by the weight of the superincumbent mass from the surface of the park above, and had extended far beyond the place where the collapse had occurred. The merest cracks could be perceived in the soffit at the crown, close to the sand. On examining the ground above the tunnel, a depression, extending for 45 yards, and varjdng in width from 10 yards to 27 yards, and at its deepest part 21 feet, was found (Figs. 1). The south end was then visited, and it was apparent that the dry sand had run in as through a funnel, choking the line for a length of 58 yards, and filling every crevice until a slojie of about 2 to 1 at both ends was formed. Eather clearer signs of settlement in the soffit were visible, but practically the sand hid everything, and the full extent of the failure could only be surmised.

The result of this inspection was a prompt decision to tunnel through the slip in a right line, and make good the work from below, the execution being entrusted to Mr. J. T. Firbank, as contractor, and the Author being appointed Eesident Engineer.

Both north and south of the slip the arch showed signs of weakness, and the fii'st thing undertaken was shoring it against falling in. Six skeleton ribs of elm were fixed under the arch at both places, tightly wedged and supported on sills and struts, carried down into the solid ground and to the level of the founda- tions of the side walls.

To do this, a trench about 2 feet wide was sunk through the slij), from the soffit of the~ old arch to the foundations, the end

Papers.] LOPES ON THE REPARATION OF BETCHWORTH TUNNEL. 293

nearest the slip and the side being close-poled with 3-inch deals, and stretched to the walls ; the polings were liberally packed with hay to stojD any of the sand running. Then 12-inch by G-inch

Figs. 1.

CROSS SECTION N9 I.

CROSS SECTION N° 2.

CROSS SEaiON N9 3.

Scale 1 inch = 50 feet.

timbers were laid, and 9-inch by 9-inch props placed njion them, on which a top sill, 12 inches by G inches, was put, and these received the skeleton ribs. Between the ribs and the soffit,

294 LOPES ON THE REPARATION OF BETCH WORTH TUNNEL. [Selected

laggings, about 2 feet apart, were run in and tightly fixed against the arch by page-wedges driven between them and the ribs.

The arches having thus been practically secured, a start was made to close-timber the broken end by poling from the crown. Poling-boards, 2 inches thick by 3 feet long, were inserted verti- cally, and a stretcher, aboiit 6 feet long by 10 inches square, was got in about 3 feet down from the soffit, the ends being carried well behind the old work. The sand was thrown back on the slope, to assist in forming a scaffold, and also as a counterbalance to the more central body of the slip. The poling was then con- tinued, the 10-inch stretcher acting as the waling, and another stretcher was duly inserted 3 feet lower as before.

In the same manner the face timbering was carried down to the heel of the rib, when it was stopped and a top heading commenced. The bottom waling was previously angle-strutted from the side walls by 12-inch by 12-inch timbers, let into the walls and tightly wedged therein, the other end being birds-mouthed. The top heading, 4 feet wide by 6 feet high, was then cut through the face polings at the level of the soffit, and driven 20 feet into the slip, at a rising inclination of 24 inches in its length, and, this point gained, the heading was returned towards the face, rising 24 inches more, and again returned with another 24 inches. This, allowing 2 feet 3 inches for the thickness of the arch, gave 2 feet for the crown bars and 1 foot 9 inches drop. The heading, close-timbered throughout, was poled at the sides with 1^-inch boards, and on the top with 2-inch boards. The head and side trees were 9 inches die-square, placed 3 feet 6 inches from centre to centre on 12-inch by 6-inch foot-blocks each 18 inches long. The feet of the side trees were stretched across with the same sized timbers, all well dogged with iron " brobs " and laced with long boards 1 inch thick.

When the 20-foot length of heading was finished, a crown bar of pitch-pine, 2 feet square and 18 feet long, was put in position, one end being run back 3 feet over the crown of the existing arch, and the leading end being supported by a 10-inch square back prop, placed on a foot-block. The heading was widened piecemeal, by removing setting after setting and close-poling, until room for another bar alongside the first was obtained, great quantities of hay being used all the time, lantil five crown bars had been fixed and propped.

A nipper sill 16 inches by 8 inches and 12 feet long was then laid in, and the five bars were prop})ed again from it. The top heading, in its width, thiis-disa2ipeared, and mining for a 12-foot

Papers.] LOPES ON THE REPARATION OF BETCHWORTH TUNNEL. 295

length proceeded, the sides always being close-hoarded, and three bars on each side inserted until the sill bed-level was reached. The bars, with few exceptions, were of larch. Back props, 10 inches by 10 inches, were then got in, resting on foot-blocks, the sill of pitch-pine, 16 inches square, was drawn into place, and a wooden saddle, 8 feet long by 16 inches by 8 inches, was fastened to the upper side.

Each bar was supported over the sill by 1 0-inch props, radiating with the curve of the arch. Stretchers, 16 inches square, were fixed between the old brick face and each end of the sill, and angle-struts or judkins raking from the brickwork to about 4 feet on each side of its centre. The two rakers, 46 feet long, running 7 feet into the solid ground, with foot-blocks, were fixed and wedged, birds-mouthed at the sill, and with two wrought-iron glands to each near the upper ends. The rakers were themselves stretched by a 9-inch by 9-inch stretcher near the top end, and by a 12-inch by 12-inch stretcher about halfway down.

The length was then deemed practically secure, and a gullet, about 6 feet wide, was driven through the loose sand from the face at the middle sill-level, the sides being still close-poled and stretched. The face near the upper sill being gained, three back props, 10 inches by 12 inches, were got in with foot-blocks to each, and the gullet was widened until the upper sill was supported by eleven back props, all being close-j^oled, and the sides of the tunnel were reached. Three more bars on each side were then fixed, and the middle sill followed similar to the upper one. Twelve upright props were fixed between the upper and middle sills, and sill- stretchers from the brickwork to the sill ends were put in, and rakers with foot-blocks. A bottom gullet, 6 feet wide at rail-level, was now driven as before, and eleven back props were got under the middle sill, and close-poled until the sides were again reached, when two more bars on each side were fixed. The bottom sill, 14 inches square in one length, was then placed in position, and eleven vertical props, 10 inches square, were inserted between it and the middle sill, and 12-inch sill-stretchers put in; 10-inch square rakers, with foot-blocks, followed, and the excavation was complete, being fully timbered to the level of the rails.

The mining for the invert was then proceeded with, the face only being timbered. The curve for the invert, being shallow, was not poled.

Between the top bars, twelve in number, short stretchers were fixed about 3 feet apart, and single stretchers between the rest, a 9-iuch cross-stretcher, or cock-roost, running between the fourth set

296 LOPES ON THE REPARATION OF BETCHWORTH TUNNEL. [Selected

of bars. All props were dogged top and bottom, bi;t no bolts were used, driving and page-wedges, however, being plentiful. The back props were fastened to the sides by 9-inch by 3-inch angle-

FiG. 2.

TRANSVERSE SECTION

Scale 1 inch = 8 feet.

ties, and the bottom uprights and the lower sills were laced by 1-inch boards.

Figs. 2 and 3 represent the full timbering for a 12-foot length of inverted section.

Papers.] LOPES ON THE REPARATION OF BETCHWOETH TUNNEL. 297

The brickwork was then put in hand, invert moulds and leading frames being first placed in position. The side stretchers and bars were removed as the brickwork was got up until the side walls were ready for the skeleton centres. These, three to each 12-foot

afcf- i.i' e'

LONGITU Ol N Al. SECTION.

Scale 1 inch = 8 feet.

length, were made entirely of elm in three thicknesses. The middle centre leaves 1\ inches thick, and the leading centres 3 inches and grooved, with laggings 7 inches by 3 inches, were supported on *J-inch props on brick piers 14 inches square, built out from the

298 LOPES ON THE REPARATION OF BETCHWORTH TUNNEL. [Selected

invert with 12-incli by 6-incli sills, and placed on beech slack blocks 18 inches by 7 inches by 3^ inches.

As the arch progressed the side bars were taken out until six crown bars remained, which were built in, and the spaces between them filled in tight with brickwork to the poling-boards. The keying-in for 1 foot 10 inches wide was effected by block laggings fixed transversely on grooved laggings speciallj- prejDared. These })lock laggings were 22 inches long by 7 inches by 3 inches, notched at each end to fit the grooves, and the key of the arch was made good from one end outwards. The whole of the poling- boards were left in.

The length being thus finished, the mining for another pro- ceeded as the first, and the whole length was executed in precisely the same manner. Another length followed until three lengths from each end were completed, when a top heading was driven through to connect the two faces. Both faces were kept in hand until an interval of 36 feet remained, when the north end was stopped and the south only jiroceeded with. When this length was finished, the north was completed and the junction only was left. Here the bars were supported on the toothings of the arches on either side, and as the brickwork was finished eight bars were built in with it, thus 37 yards were completed.

The skeleton ribs were left standing about ten days, by which time the cement was well set. Notwithstanding, and in spite of the bars being bricked in solid, the excessive weight of the super- incumbent sand in the slip broke the toothings at the junction of every length, causing a slight crack, the gTeatest being ^ inch wide ; these cracks were immediately stopped, and there was no further settlement.

Where the old arches at both faces showed signs of cracking and weakness, which was for a length of 10 yards at the south end, and 8 yards at the north end, the inverted section was continued, and here it was built in lengths of 6 feet. The skeleton ribs already fixed were made the base for the work, 12-inch timbers being thrown across from sill to sill with planks over them. An upper scaffold was then made of 3^-inch by ^-inch double irons slung from the sides of the ribs with cross-bolts, upon which die-square timbers were placed and plank -sheeted to enable the miners to start breaking into the key of the old arch.

As the brickwork was cut away, short polings were got in until the 6-foot length was removed, when a crown bar 12 inches in diameter, one end resting on the new brickwork and the other on the old work, was inserted. The mining was continued, bars being

Papers.] LOPES ON THE REPARATION OF BETCHWORTH TUNNEL. 299

put in about 1 8 inches apart for the extent of the arch, below which bars of 9 inches, die-square and 2 feet apart, sufficed, until all the old, brickwork in the length had been removed, the face of the sand being close-poled all the while and the end face secured by waling 12 inches square strutted by 9-inch timbers from the new brick

Fig. 4.

Segmental Centre Supporting Relieving Arch.

Scale 1 inch = 8 feet.

invert. Each 6-foot length was carried out in the same way, except at the ends where the inverted section was to be abandoned. Here the arches were gradually projected over longitudinally beyond the walls until at the crown they led for about 9 inches. This was to act as a shield to a lower arch.

The very unsatisfactoiy character of the old brickAvork, thus

300 LOPES ON THE REPARATION OF BETCHWOKTH TUNNEL. [Selected

exposed, suggested a complete examination of the rest of the tunnel, and with hardly an exception the work was found to be defective. Accordingly Sir John Fowler, K.C.M.G., Past President Inst. C.E., the Consulting Engineer to the Company, was invited to confer with Mr. Banister and to determine on the course to he pursued as re- garded the remaining 330 yards of the tunnel.

It was then decided to remove the whole of the side walls, to rebuild them in Portland cement, to turn a relieving arch, also in cement, at a lower level than the old one, and to dispense with an invert. The new side walls, being 18 inches thick instead of 1 foot lOj inches like the old walls, were backed with cement concrete to the sand, and carried up with the new relieving arch of sufficient thickness to underpin the old arch until the relieving arch was clear of it. The new arch was 18 inches thick, and the space between the crown and the old soffit was filled in with old dry bricks tightly wedged and packed (Fig. 4).

The work was built in cement throughout and ten break-ups were put in hand at once, which, as the work proceeded, was found to be the extreme limit for safety. Four skeleton ribs were fixed at each place, the middle ones 9 feet apart and sheeted with 7-inch by 3-inch laggings, tightly wedged. The level of the under- pinning was first ascertained and the brickwork cut away for 9 inches thick, when die-square timber was put in and strutted off the ribs, and another 14 inches got out under that and strutted in like manner. The back being thus reached, the brickwork was gTadually removed to the footings, and the sides were close-poled, packed with hay, and stnitted by larch bars 12 feet long, 9 inches thick and 2 feet apart, their ends being let into the brickwork on each side.

As soon as the old brickwork had been removed the new was inserted, great care being taken in the underpinning, which was most tedious and difficult, the old brickwork being practically loose with every tendency to fall. As length after length was executed, the old side walls were in many places splintered and cracked, and more skeleton ribs had to be erected, and the mining lengths reduced from 9 feet to 7 feet 6 inches. By this arrange- ment, however, the work was completed without accident. AYhere the relieving-arch section joined the new inverted arch (Fig. o) a wall 14 inches thick was built as face packing, and at the south entrance the whole of the spandrels and parapet were taken down and rebuilt.

The ribs for the relieving arch were made in two sections of 6-inch timber, the struts raking from crown to heel with cross-ties

Papers.] LOPES ON THE REPARATION OF BETGHWORTH TUNNEL. 301

and queenpost, and could be taken in pieces and reunited The tie-beams were kept sufficiently bigb above the rails to leave room for the contractor's six-wheeled locomotive which worked throuo-h the tunneL ^

The time occupied in the 55 yards of new inverted section in the slip was from the 8th of August to the 30th of December, 1887 and m the remaining 330 yards from the 7th of November, 1887, to the

Fig. 5.

	e
j
i^ INVEnj-£0 SECTION J
<...60 I^.. .J^ IZ'q' Lengths.. .^ _ ^ £ nti «
._ .i ies.o' - i -A
^^^^m	v//////////////////////^^^^^
1	j	M
//////////////y
	RAIL V LEVEL \

Vertical scale, 1 inch = 30 feet. Horizontal ,, „ = go

7th of February, 1888, continuously day and night, with the exception of Christmas-day.

In the reconstruction, one of the two lines of railway throu-h the tunnel was taken up and the other slewed to the centre for the convenience of the contractor.

Although there was little water in the tunnel, the long cuttin- at the south end on a gradient falling towards the tunnel rendered

302 LOPES ON THE REPARATION OF BETCHWOBTH TUNNEL. [Selected

necessary tlie preservation of the original drainage arrangement, which was by a 12-inch pipe.

During the busiest part of the reconstruction three hundred men were employed night and day with six horses and one locomotive. The bricks, of which two and one-third millions were used, were the best Horsham stocks. Cement was supplied by the Sussex Portland Cement Company, Newhaven, to the extent of 780 tons, and this was mixed with 2,000 cubic yards of sand, part from pits at New Cross, and part from beds at Oxted. Water was at first provided in tanks from Horsham and Holmwood, but afterwards from the main of the Dorking Water Company, by a 2-inch pipe laid through the tunnel with junctions and bib-cocks where necessary.

The following materials were supplied, viz. : — 62 elm skeleton ribs; 114 segmental centres; 16,500 cubic feet of timber; 3,030 cubic feet of larch ; 37,400 lineal feet of 3-inch by 7-inch battens ; 40,600 lineal feet of 3-inch by 9-inch deals ; 52;^ fathoms of poling- boards ; 48,700 page-, driving-, and raking-wedges ; 760 slack blocks; 10 tons of hay; 10,525 gallons of naphtha, and 3,474 lbs. of candles.

The slip comprised about 33,000 cubic yards, and of this the bars in a 12-foot length of the inverted section bore 10,500 tons during the mining.

The cost of the work per lineal yard of the inverted section was £145, and of the relieving-arch section £30. The exceptionally heavy expense of the former was caused by the enormous quantity of timber necessary to support the slip, and to the necessity of building in so much of it.

The tunnel was opened for traffic on the 1st of March, 1888.

The Paper is accompanied by several diagrams, from which the Figs, in the text have been prepared.

Papers.] WOKTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. 303

(Paper No. 236 1.)

" The Permaiient-Way of some Eailways in Germany and in Austria-Hungary." ^

(Translated and abstracted by William Barton Worthington, B.Sc., M. Inst. C.E.)

This Paper is a digest of a series of illustrated articles describing in considerable detail tlie permanent-way of certain railways in the German and in the Aiistro-Hungarian States. The principal dimensions of rails and metal sleepers are given in Tables I and II, in which will be found some further particulars relating to each railway.

Table I. — Rails.

[Length.

Alsace-Lorraine Railways.

I. Steel rails on wooden cross- slcejiers

29-50

II. Steel rails on iron cross-sleepers j III „ „ [ „

with Haarmaun hook-plates | IV. Rteel rails (Hartwich) for local I nq . en

lines /

Austrian North-Western Bailway.

I. Steel rails on steel- sleepers, Vgq. en

Holienegger system . . . / ' II. Steel rails on wooden cross- 1 nq . -^

sleepers j

III. Steel rails on specially large 't I

bearing-plates ^ . . . • /, "

Deptb.

Width I Thick- Width ^^'eight

of ness of Head. ; Web.

of Foot.

per

I/neal

Yard.

Inches. Inches. Inch. Inches. Lbs.

514

4-90 4-90

2-31

7-07 1-96

2-23 2-23

0-54

0-39

0-43 0-40

I 3-96 74-30

4-71

3-54

4-09

73-76

58-86 66-52

' The original articles appeared in the Organ fiir die Fortschritte des Eisen- bahnweseus, 1888, p. 1 et seq.

" Bearing-plates 13-7 inches x 5-1 inches.

304 •R-OKTHINGTON ON THE PERMANENT -WAY OF EATLWAYS. [Selected Table I. — Rails — continued.

	Length.	Depth.	"Width Thick- of ness of Head. Web.	Width of Foot.	Weight per Lineal Yard.
	Feet.	Inches.	Inches.	Inch.	Inches.	Lbs.
Bavarian State Eaihcays.
I and II. Main and secondary lines III. (a) Local lines (h)'Do (c)Do. (Hartwich). . . .	29-50 29-43 29-50	5-10 4-00 4-32 5-90	2-28 1-81	0-43 0-31 o'-'35	4-10 2-67 3-54 4-71	63-00 38-20 44-25 58-46
State Bailways of Saxomj.
I. Main and branch lines. II. Local lines, normal gauge in. „ „ narrow ,,	24-60 24-60 29-50	5-11 4-32 3-43	2-27 2-00 1-60	0-43 0-39 0-31	4-12 3-37 3-10	69-24 49-00 31-40
Bhenish Bavarian (Palatinate).
I. Main lines, iron cross-sleepers II. „ „ wood „ III. Secondary lines, wood ,,	26-24 >>	5-26 4'-'l2	2-27 2'-'03	0-42 o'-'39	4-12 2'-"54	68-50 50'-'34
Hessian Ludicigs Bailivay.
I. Steel rails on iron cross-sleepers ,, „ wood „	24-60	5-11	2-27	0-56 J)	3-93 >>	71-8
Baden State Bailways.
I. Main lines, steel rails on iron'i cross-sleepers / II. Local lines, steel rails on wood | cross-sleepers . . . . /	29-50 24-60	5-08 4-08	2-36 2-04	0-54 0-43	4-12 3-54	73-00 53-40
Austrian Southern Baihcay.	32-80	5-00	2-25	0-50	4-00	68-54
Austrian State Bailicays.
I. (a) Steel rails, iron cross-^ sleepers (Heiudl) . . / (b) Steel rails, wood cross-1 sleepers / II. (a) Steel rails, iron cross- "1 sleepers (Heindl) . . j (h) Steel rails, wood cross-1 sleepers j	24-60 24-60	4-91 4-71	2-27 >> 2-23 »>	0-46 >> 0-46 >>	4-32 )> 4-32	71-10 » 63-93
Huncjarian State Bailways. I. Main lines, 1st rank . . . | II. „ „ 2nd „ • • •)	26 23 19 26 23 19	251 00 68 1 251 0 68)	5-00 4-32	2-28 1-85	0-46 0-43	4; 08 3-54	67-00 50-22
III. Secondary lines |	19 16	68, 40/	314	1-49	0-39	2-75	30-24

.it

Papers.] WORTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. 305 Table II. — Metal Sleepers.

Width Width ' „ , below. I above. , P '	Thick- Length. ! ness of top.	Weight.
Alsace-Lorraine Hallways.	Inches.	Inches.	Inches.	Feet.	Inch.	Lbs.
II. Steel rails ou iron cross-sleepers	8-05	4-32	2-90	7-87	0-35	110-00'
III. „ „ „ „ \ with Haarmaun hook-platesj	10-34	511	3-54	8-85	0-31	150-42 2
Austrian North-Westerii Railway.
I. Flat-bottomed steel rails oui mild steel longitudinal- ( sleepers (Hohenegger sys- tem) j	11-81	6-6	2-94	29-25	0-35	577-60
Bavarian State Railways.
I. Steel rails ou iron cross-sleepers,"! closed ends (Heindl system) /	9-44	5-12	3-54	8-20	0-35	138-88

II. Secondary lines. Commencing in 1888, these lines will be laid with per- manent-way similar to No. 1, but lighter.

III. (a) Steel rails on iron longi-"l q.nc tudinal-sleepers . . ./ (h) Steel rails ou iron cross- M n.nq sleepers /

Rhenisli Bavarian (Palatinate) Railways.

I. Steel rails on iron cross-sleeiDcrs

Hessian Ludivigs Railway.

I. Steel rails on iron cross-sleepers closed at the end .

5-91 3-90

2-32 29-43

2-96

j9-25j (6-301 (4-93|

to > < to > < to 14-30, 3-20 |2-76

Austrian State Railways la and Ila. Steel rails ou iron cross-

spersj g.gQ

Ilrt. Steel rails ou iron cross-"! i a . oc sleepers (Heindl system) . /

Baden State Raihvays.

iel rails on iron cross-sleejiersl q r,p. (Hilf pattern) . . . .]\

4-70

5-90

4-60

2-75

3-93

2-40

-20

8-50

-20

7-87

7-95

0-31 0-35

0-43

0-40

0-40

0-35

346-10 87-00

114-603

114-20

157-50

94-30

' Old pattern. ^ New pattern.

^ Minutes of Proceedings Inst. O.E., vol. xci. p. 492.

[the INST. C.E. VOL. XCV.]

306 WORTHINGTON ON THE PERMAXENT-WAY OF RAILWAYS. [Selected

Alsace-Lorraine State Eailways.

There are foiir kinds of permanent-way in use on these railways.

I. Steel rails on wooden cross-sleepers ^ II. Steel rails on iron cross-sleepers I

1 'ill 1^9-lD.

Ill Steel rails on iron cross-sleepers with ,.

Haarmann hook-plates IV. Steel rails (Hartwich system) for local lines.

I. The inner cover of the fish-joint is 18-7 inches X 3-29 inches x 0-78 inch, and weighs 10-86 lbs. The outer cover is an angle of the same length and weighs 20 lbs. The fish-bolts are 0-93 inch diameter, and weigh 1-6 lb., and are spaced 4-4 inches apart centre to centre.

The rail-bearing plates are 6 • 87 inches x 6 • 87 inches X 0 • 47 inch, with raised edges 1 • 4 inch wide and 0 • 4 inch high, and weigh 5 • 9 lbs. each. They are used only at the sleepers on each side of the rail joint, and are let into the sleeper at the inner edge so as to give the 1 in 20 cant to the rail.

The rail is attached to the sleeper by three screws passing through holes in the bearing plate, two on the inside and one on the out- side of the rail, the head of the screws bearing on the edge of the foot of the rail. These screws are 4 • 1 6 inches long, including the head, and 0-74 inch diameter below the head; they weigh ^ lb. each. The sleepers are of creosoted oak, 8-2 feet x 10*4 inches X 4-1 inches.

There are ten sleepers to the 29 feet 6 inches rail-length, spaced from centre to centre, 23*9 inches at the rail-joint, 29*5 inches next, 37-4 inches next, and 39*4 inches (1 metre) the remainder.

The weight of iron and steel in one rail-length, 29 feet 6 inches, of this permanent way is : —

2 rails 1,464

4 rail-joint plates 61

4 fish-bolts 6

5 rail-bearing plates .... 23 60 screws 44

31 lbs. 75 „ 50 „ 40 „ 91 „

Total . . . 1.600-93

or 162 '7 lbs. per lineal yard.

II and III. The rails and rail-joints are the same as in I. The older iron sleepers differ from the Yautherin section only in having a thickening of triangular section, the apex of the triangle being

rapprs] WORTHINGTON ON THE PERMANENT -^V AY OF RAILWAYS. 307

downwards, along the bottom edge of the sides ; the newer sleepers are of larger dimensions. The ends of the sleepers are closed.

II. In the older form the sleeper is bent at the middle so as to give the 1 in 20 cant to each rail. The rail rests directly npon the sleeper, and is held in place by two clamp-plates, throTigh each of which passes the bolt which fastens it to the sleeper. The bolts are 0 • 78 inch diameter, and weigh 0 • 8 lb. each. The head of the bolt is oblong, as is the hole in the sleeper through which it passes, half a turn being given to the bolt to fix it. Between the clamp- plate and the sleeper is a packing with a projection on its lower side which fills up the open portion of the oblong bolt-hole in the sleeper and prevents the bolt from turning.

III. In the newer pattern the 1 in 20 cant is given to the rail by a wedge-shaped " hook-plate " (Haarmann's pattern). At the outer end this plate is formed into two hooks, one of which passes down, through and forward under the top plate of the sleeper, while the other turns wp and back over the foot of the rail. The inside of the rail is secured to the sleeper by a clip and olilong-headed fang bolt, the clip having on its lower face a projection which fills the superfluous portion of the bolt-hole in the sleeper. The hook-f)late weighs 4 lbs. and the clip about 1 lb.

In the 29,;^ feet rail-length there are ten sleepers spaced from centre to centre: 23*2 inches at the point, 29-9 inches next, 37*4 inches next, and 39*4 inches the remainder.

The weight of one rail-length of this newer pattern is : —

2 rails

4 rail-joint plates

4 fish-bolts 10 sleepers 20 hook-plates 20 dips .... 20 bolts . . .

Total . . or 326 '3 lbs. per lineal yard.

1	,4G4	3i lbs
	61	•76 „
	6	50 „
1	564	40 „
	80	25 „
	21	15 „
	13	22 „
3	211	59 „

IV. The rails are of steel and laid on the Ilartwich system.

The angle joint-plates are 24 inches long, 5 • 9 inches high, and Aveigh 34 • 6 lbs. each. The rail-joint is made with eight bolts 0'78 inch diameter in two rows 2*36 inches apart, centres. The distance of the centre of the bolts from the end of the rail is, in the upper row, 2-75 and 8*26 inches; in the lower row 5*11 and 10-62 inches.

The gauge is kept T)y cross-ties 1 inch in diameter, 9 feet 10. inches

X 2

308 WORTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. [Selected

apart (centres), and weighing, together with the two nuts at each end, which form the attachment to the rail and by means of which the gauge is regulated, 16* 75 lbs.

The 1 in 20 cant of the rails is obtained by wedge-shaped washers between the web of the rail and the nuts of the ties.

The weight of one rail-length, 29 feet 6 inches, of this permanent- way is : —

2 rails 1,443-65 lbs.

4 angle-joint plates 138 '43 „

8 fish-bolts 8-10 „

3 cross-ties and nuts . . . . 50 ■ 35 „ 12 washers 10 85 „

Total .... 1,651-38 „ or 167-78 lbs. per lineal yard.

AusTPJAX North- Western Eailway.

The permanent-way for main lines is of three kinds.

I. Flat-bottomed steel rails on mild steel longitudinal sleepers (Hohenegger system), II. Flat-bottomed steel rails on wooden cross-sleepers with ordinary bearing plates. III. Flat-bottomed steel rails on wooden cross-sleepers with specially large bearing plates.

I. The fish-plates for rail-joints are 19 '68 inches long. The inner plate is a flat plate of 11 lbs. weight, the outer is an angle-plate with a projecting member at the top reaching half-way up the head of the rail, and weighs 25 • 62 lbs.

The joint of the longitudinal sleepers is made by a wrought-iron box with open ends, the bottom of which is a flat plate 16 j inches wide, and 15f inches long, on to which is bolted a plate bent to fit the inside of the sleeper. To this bent plate the sleepers are fastened by six bolts. The bottom plate of the joint-box is hori- zontal, and the top plate is so shaped as to give the inward cant ot 1 in 16 to the sleeper and rail. The joint-box weighs 35 j lbs.

The bolts for the rail-joints and sleeper-joints are 0*86 inch in diameter; the former weigh !•! lb., the latter 1*3 lb.

The rail is fastened to the sleeper by bolts of 0 • 74 inch diameter, passing through a square clip, one edge of which fits exactly to the sloping outer edge of the rail foot, and the other fits against a sloping rib formed along the edge of the top surface of the sleeper.

Papers.] WOETHINGTON ON THE PERMANENT -WAY OF RAILWAYS. 309

The weiglit of one rail-length of the road is : —

2 steel rails 29 feet 6 inches long . . . 1 , 158 ■ 74 lbs.

2 mild steel sleejjcrs 29 feet 3 inches long 1,155 -21 „

4 angle-irou cross-ties 238 '09 „

2 rail-joint angle-plates 51'14 „

2 „ flat-plates 22-04 „

2 sleeper-joints 70 • 53 „

8 rail-joint bolts 10-58 „

20 sleeper bolts 22 ■ 04 „

16 cross-tie bolts 14-11 „

36 rail bolts 27-78 „

130 washers, &c 95-15 „

Total .... 2,865-41 „ or 290-7 lbs. per lineal yard.

II. The fish-plates at the rail-joints are angle-plates ; the outer one has a projection at the top, is 27*55 inches long, and weighs 26 "44 lbs. The inner one is 25 '19 inches long, and weighs 19" 17 lbs. The bolt-holes nearest to the joint are 6*7 inches from centre to centre, and the outer holes are 4*6 inches from these. The fish-bolts are 0*86 inch diameter, and weigh 1*3 lb. each.

Between the rail and the sleeper is a bearing-plate, 7*9 inches long, 5 • 5 inches wide, and 0 • 4 inch thick, with a rim ^ inch high along the outer edge, against which the outer angle fish-plate rests. The rails and plates are fastened to the sleeper by three octagonal spikes, 0 • 7 inch diameter, and 6 • 3 inches long, two on the outside, and one on the inside of the rail.

The wooden sleepers are 8*2 feet long, 9*8 inches broad below, at least 5 • 9 inches above, and 5 • 9 inches deep. At the joints they are spaced 23*6 inches from centre to centre ; the remaining spaces increase from 32*27 inches to 33*85 inches at the middle of the rail-length.

The weight of iron in one rail-length of this road is : —

2 rails 29 feet 6 inches long . . . 1 ,309-55 lbs.

4 angle-joint plates 91-25 „

8 fish-bolts 10-58 „

8 washers 1 - 38 „

22 rail-bed plates 10-14 „

66 spikes 55-34 „

Total .... 1,478-24 „

or 150-2 lbs. per lineal yard.

III. This permanent- way combines the rail fastenings of I with the wooden sleepers of II. The rail rests upon a bearing-plate

310 WOKTHINGTON ON THE PERMANENT- WAY OF RAILWAYS. [Selected

13* 7 inches x 5"1 inches, to whicli it is attached hy two holts passing throngh clips, the oblique outer edges of which rest against the sloping faces of corresponding ribs formed upon the top of the bearing-plate, while their inner edges lap orer the edges of the rail foot. Thus by slackening the bolt of one clip and tightening the other, the gange of the line can be slightly altered.

The bearing-plates are fastened to the sleepers by three spikes, tAvo on the outside, and one on the inside. The large size of the bearing-plate gives a greater resistance to the tilting over of the rail than is the case with the ordinary plate in II.

The weight of the ironwork of one rail-length of this road is : —

2 rails 29 feet 6 inches long . . . 1 ,309-54 lbs.

4 angle-joint plates 91-25 „

8 fish-bolts 10-68 „

8 washers 1-38 „

22 large rail bearing-plates . . . 247 • 34 „

44 bolts for „ ... 48-49 „

44 washers 1 - 99 „

66 spikes 55-34 „

44 holdfasts 81-57 „

Total 1,847-58 „

or ] 87 - 7 lbs. per lineal yard.

Bavarian State Eailways. The lines may be divided into three classes : —

I. Main lines. II. Secondary lines. III. Local lines (of normal and narrow gauge).

I. Main lines : Greatest wheel-pressure 7 tons, and highest speed 46 miles per hour.

(a) With steel rails and iron cross-sleepers (Heindl system) ;

(6) With steel, or steel-headed rails, and wood cross-sleepers.

The rails (flat-bottomed) are the same for both a and h.

There are eleven sleepers in the rail-length, spaced as follows : — At the joint, 19-7 inches, centre to centre; next, 29*6 inches; next two, 31* 5 inches ; four middle spaces, 35 "4 inches.

The angle fish-plates have the same cross-section for a and 6, the length being 19" 7 inches for the outside plate in both cases, and 26-2 inches and 29-5 inches for the inside plate in a and h respectively. The fish-liolt holes are 2 inches and 7 inches from the joint.

Papers.] WORTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. 311

The fish-bolts are 0*87 inch diameter, and weigh 1*3 lb.; the foot-bolts which attach the rails to the iron-sleepers are 0 • 8 inch diameter, and weigh 1 lb. ; the spikes for the wooden sleepers are 0 • 6 inch thick, 6 inches long, and weigh 0 • 6 lb.

The rail bed-plates on the iron sleepers are wedge-shaped, with an inclination of 1 in 20 ; those on the wooden sleepers are 0 • 47 inch thick under the rails, and 0 • 7 inch at the edges, the 1 in 20 inclination being cut in the sleeper. The plates are 7 • 08 inches X 5' 90 inches at the joint, and 5*9 inches x 5*59 inches at the intermediate sleepers.

The wood sleepers are 8*2 feet long, 10 "23 inches wide below, and at least 6 • 29 inches wide above, and 6 • 88 inches thick.

The weight of one rail-length of this road (29*5 feet) is : —

2 steel rails 1,238-10 lbs.

11 cross sleepers 1,527'80

2 outside fish-plates 39 • 24

2 iuside „ 52" 46

8 fish-bolts 10-14

22 bed-plates 47-61

44 foot-bolts aud fittiup-s . . . . 100-08

Total .... 3,015-48 „

or 306 lbs. per lineal yard.

II. Secondary lines : Greatest wheel-pressure, 6 tons ; greatest speed, 15 miles per hour.

Commencing in 1888, these lines will be laid with material similar in design to that used for main lines, but lighter.

III. Local lines : Greatest wheel-pressure, 5 tons ; greatest speed, 15 miles per hour.

(a) Flat-bottomed steel rails on iron longitudinal sleepers.

The total weight of this permanent- way is 164*4 lbs. per yard run.

(6) Flat-bottomed steel rails on iron cross-sleepers.

The total weight of this permanent-way is 195*86 lbs. per yard run.

(c) Eails on the Hartwich system, without any sleepers.

The rails are kept to gauge, and to their cant of 1 in 20, by cross- ties (four in the 29 feet 6 inches length of the rail), 3|^ inches by I inch, bolted to the web of the rail by two bolts.

The total weight of this permanent- way is 139*5 lbs. per yard run.

In addition to the local lines laid on these three systems, there

312 WORTHIXGTON ON THE PERMANENT- WAY OF RAILWAYS. [Selected

are some narrow-gauge (1 metre) lines, laid witli a flat-bottomed rail on light iron cross-sleejiers, the total weight of the permanent- way being 123-58 lbs. per yard run.

State Eailways of Saxony, These may be divided into three classes : —

I. Main and branch lines. II. Local lines of normal gauge. III. Local narrow-gauge lines.

I. The permauent-Avay of these, as of the local lines, consists of flat-bottomed steel rails, laid ujDon wooden cross-sleepers. The cant of the rail is at present 1 in 16, but it is intended to make it 1 in 20.

The sleepers are of fir, impregnated with zinc chloride ; 6 feet 4 J inches to 8 feet 2^ inches long; 6*7 inches wide above, and 7*87 inches below, and 6-29 inches thick. At the rail-joint they are 21^ inches apart (centres), the next pair 30*7 inches, and the remainder 35 • 4 inches.

The angle fish-plates are 25' 19 inches long outside, 17*54 inches inside; the former weighing 26*8 lbs., the latter 13*59 lbs. The outside plate is prolonged upwards under and round the top flange of the rail, so that the top of it is only about f inch below the top of the rail. The fish-bolt holes nearest to the joint are 3f inches from centre to centre, and the outer holes are 5*18 inches from these. The bolts are 0*97 inch in diameter, and with the nut weigh 1*93 lb.

The rail bearing-plates, which are fixed on every sleeper, are 7-56 inches X 5*96 inches, 0*47 inch thick, and weigh 6*72 lbs. On each side of the rail-foot they have a rim 0 • 23 inch high, and on the under side have three pointed ribs, 0*12 inch high, to pre- vent the plate from slipping on the sleeper.

The rail is fastened to the sleeper by three | inch square spikes, 5f inches long, two on the inside and one on the outside.

II. Local lines of normal gauge, with maximum wheel-pressure of 5 tons.

The sleepers are the same as for main lines, and spaced in the same way.

The outside angle fish-plate is 20 • 3 inches long ; the inside one 18 "10 inches. Their ends butt against the spikes which attach the rails to the sleepers, and which are the same size, and arranged

Papers.] WORTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. 313

in the same way as those for the main lines. The fish-bolts are the same as on main lines, and spaced in the same way.

The rail-hearing plates on each sleeper are 7 • 07 inches hy 5 • 89 inches, 0 • 43 inch thick, and weigh 4 • 84 lbs. They have no ribs on the under side.

The cant of the rails is 1 in 1 6.

III. Local narrow (2 feet 5^ inches) gauge lines.

These are laid with flat-bottomed steel rails on wooden cross- sleepers.

The sleepers are of fir, impregnated with zinc chloride, and are about 5 feet long, 7 • 9 inches wide at the bottom, 5 • 5 inches at the toj?, and 4 • 3 inches thick. There are twelve sleepers to the rail- length, spaced 19" 7 inches apart (centres) at the joint, 30*4 inches for the rest.

The varioxis fastenings are similar to those of the main lines described above, but lighter.

Rhenish Bavarian (Palatinate) Eailways. The permanent-way may be divided into : —

I. For main lines with iron cross-sleepers. II. For main lines with wood cross-sleepers. III. For secondary lines with wood cross-sleepers.

I. The angle fish-plates are 19-7 inches long, and weigh 17*6 lbs. each. The fish-bolts on each side of the joint are 3 • 9 inches apart from centre to centre, and the other two are 5 • 1 inches from these. The bolts are 0-86 inch diameter, and weigh 1*2 lb. each. The rail foot-bolts are 0 • 74 inch diameter, and weigh 0 • 66 lb. They hold down the clips which press upon the foot of the rail, and which have on the under side a projection which fits into the oblong bolt-hole in the sleeper. These projections are of different sizes, and by changing the clips the gauge of the line can be regulated.

The sleepers are of varying section, and are of the pattern fully described in the Minutes of Proceedings Inst. C.E., vol. xci. p. 492. The varying thickness is obtained by rolling in eccentric rolls, and the bending is done by hydraulic pressure while the metal is still warm. The sleeper is deep and narroY^r in the middle, and widest and thickest under the rail, where it has an inclination of 1 in 20. It weighs 114-6 lbs.

314 WOBTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. [Selected

The nine sleepers are sp^aced in the 24 feet 3 inch rail-length from centre to centre as follows : —

22-04 inches at the joint, 33-06 „ the next, 37-76 ,, the remainder.

The weight of one rail-length, 24 feet 3 inches, of this permanent

way IS : —

2 rails 1,199

4 anc(le fish-plates 70

8 fish-bolts 9

9 iron sleepers 1,031

36 rail foot-bolts 23

36 „ foot-clips 26

36 washers 1

Total . . . .2,362

31 lbs.

54 „

34 „

74 „

82 „

19 „

14 „

or 270 lbs. per lineal yard.

II. The permanent-way for main lines with wood sleepers is, as regards rails and rail-joints, exactly similar to I. The sleepers of oak in the natural state, or of fir treated with corrosive sub- limate, creosote oil, or zinc chloride, are 8 feet 2^ inches long, Q\ inches thick, 10 "6 inches wide at the bottom, and at least 8*6 at the top. They are adzed to a slope of 1 in 20 to receive the rail-bearing plates, which are 7*7 inches by 6"3 inches, and 0*4 inch thick, with a rim 1 • 8 inch x 0 • 3 inch along the outer edge to hold the rail. The weight of the bearing-plate is 6 • 6 lbs. ; it has one hole on the outer and two on the inner side of the rail for the spikes, 6*5 inches long, and weighing 0*7 lb., which are driven into the sleeper to hold the rail-foot.

The spacing of the sleepers is the same as that of the iron sleepers in I.

Bearing-plates are used on every sleeper, whether in curves or straight lines.

- III. For local lines a small flat-bottomed rail on wood cross- sleepers is used.

The angle fish-plates of Bessemer steel are 15j inches long, and weigh 9*1 and 9'5 lbs. inside and outside respectively^ They project below the bottom of the rail to the extent of the thickness of the rail-bearing plates, against the ends of which their ends rest.

The fish-bolts are 0 • 7 inch diameter, and weigh, together with nut and washer, 0-7 lb. each. The rail-bearing plates are 6"3 inches X 5 • 9 inches X 0 • 3 inch, without any raised outer rim, and weigh 2 • 8 lbs. each. The spikes (one on the outside and two on the inside of the rail at each sleeper) are 5-9 inches long, and weigh about ^ lb. each.

Papers.] WORTHINGTON ON THE PERMANENT- WAY OF RAILWAYS. 315

The sleepers are of fir, 7 feet 6^ inciies long, 5 • 9 inches thick, 7 • 9 inches wide at the bottom, and at least 5 • 9 at the top. The seats for the hearing-plates are adzed to a slope of 1 in 20, and hearing-plates are used on every sleeper both in straights as well as curves.

The sleepers are spaced from centre to centre : —

20 "86 inches at the joint,

34-82 „ next,

37 '39 „ the remainder.

Hessian Ludwigs-Eisenbahn.

On this line two forms of permanent- way are now used — T. Flat-hottomed steel rails with iron cross-sleepers on the lines

of great traffic. II. riat-hottomed steel rails with wooden cross-sleepers on the

lines of small trafiBc.

The further use of the iron longitudinal sleeper permanent-way (Hilf's system), which was laid on the lines from Frankfort-on-the- Main to Limburg, from Wiesbaden to Niedernhausen, and part of the line from Mainz to Worms, has been given up, the cost of wages for maintenance of the 78 miles of line laid with longitudinal iron sleepers having been found to be, for the years 1881-1886, 36 per cent, higher than in the case of road laid with cross-sleepers.

I. The angle fish-plates are 24 inches long ; the fish-bolt holes on each side of the joint are 4 • 7 inches from centre to centre, and the other two holes are 5 • 5 inches from these. The foot of the angle- iron is cut away at each end so as to rest against the clip which holds the foot of the rail down. These angles weigh 17*6 lbs, each. The fish-bolts are 0-78 inch diameter, and weigh 1*1 lb. each.

The rail foot-bolts, which are provided with an eccentric where they pass through the sleepers, by means of which the gauge (4 feet 8i inches) can be regulated, are 0-78 inch diameter, and weigh 0-95 lb. each. The clips through which they pass, and which hold the foot of the rail, weigh 0*55 lb. each. The iron cross-sleepers are closed at the ends. The sleeper is bent to give the 1 in 20 cant to the rails for a length of about 6 inches outside and 10 inches inside the rail. The sleepers weigh 114*2 lbs. each, and are spaced from centre to centre : —

24 '8 inches at the joint,

30-0 „ next,

35 '0 „ the remainder.

316 WORTHINGTON ON THE PERMANENT- WAY OF RAILWAYS. [Selected

The weight of one rail-length (24 feet 7 inches) of this permanent- way is : —

2 rails 1,177

9 sleepers 1 , 027

4 angle fish-plates 70

8 fish-bolts 8

36 rail foot-bolts 34

36 clips 19

Total .... 2,338

25 lbs. 79 „ 54 „ 82 „ 13 „ 85 „

38

or 284 • 09 lbs. i^er lineal yard.

II. The rails and rail-joints are the same as for I.

The rail-bearing plates are 7 • 1 inches long, 6 • 6 inches wide, with a rim 0 • 2 inch high and 1 • 6 inch wide on the outer edge, and weigh 7 • 7 lbs. each. There are four holes punched in them, through which the ^ inch square and 6^ inches long spikes are driven in.

The wooden sleepers are of oak and fir, 8 feet 2J inches long, 8-83 inches wide at the bottom, and at least 4*72 inches at the top, and 5 • 9 inches thick. The oak is used in its natural state, the fir is treated with con'osive sublimate. The sleepers are spaced in the same way as the iron sleepers (I). With the oak sleeper in straight lengths, bearing-plates are only used for the sleeper on each side of the rail-joint, while in curves of from 15 to 30 chains radius four additional bearing-plates are used on the intermediate sleepers ; in curves of from 30 to 60 chains radius only two additional plates are used. Where, as is sometimes the case, oak and fir sleejiers are used together, in straight lengths, five oak sleepers and four fir are used in each rail-length, and in curves oak only. Main lines are never laid with all the sleepers of fir.

As yet no special permanent-way is used for local lines.

Austrian State Eailways.

The present normal permanent-way is divided into

I. For first-class lines —

(a) Steel rails (system X) on iron cross-sleepers on Heindl's

system. (h) Steel rails (system X) on wooden cross-sleepers.

II. For second class lines —

(a) Steel rails (system XI) on iron cross-sleepers on Heindl's

system. (h) Steel rails (system XI) on wooden cross-sleepers.

Papers.] WORTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. 317

The rails and joints are similar in la and lb ; the sleepers and fastenings in 16 are the same as those in II&, while the rails in Ila are the same as in 116. It is therefore only necessary to describe la and 116.

la. The outside angle fish-plate is 23 • 6 inches long, and weighs 17 '2 lbs., the inner 26 "25 inches long and weighs 23-5 lbs. The fish-bolts nearest the joint are 4*08 inches from centre to centre, and 4*91 inches from the outer bolts.

The wedge-shaped bearing-plates are 5*22 inches long, 2*63 inches wide, and varying from 0 • 23 to 0 • 52 inch thick, thus giving a cant of 1 in 1 6 to the rail. They weigh 3 lbs. each. Along the outer edge is a raised rim against which the outer edge of the rail- foot rests.

The fish-bolts are 0 • 85 inch in diameter, and weigh 1 • 25 lb. Grover's washers are used to prevent the nuts from working loose. The rail is attached to the sleeper by two bolts 0*85 inch diameter, with tee-heads, and weighing 1 • 1 lb. each. The bolt passes through a clip about 2 inches square, and weighing about ^ lb., which rests on the foot of the rail, and is separated from the sleeper by a packing, on the lower face of which is a projection which fills the hole in the sleeper through which the tee-head of the bolt has passed. These packings are of four different patterns, so that the gauge can be regulated by their means. Their average weight is 0-8 lb. Grover's washers are used with these bolts also.

The distance apart of the sleepers from centre to centre is : — 19-7 inches at the joint, 31*5 inches next, and 35*4 inches the remainder.

The weight of one rail-length (24 feet 7 inches) of this per- manent-way is : —

With Rail System X. With Rail System XI.

Lbs.

2 rails 1,167-35

9 sleepers 1,418-67

4 angle fish-plates . . . . 81-36

8 fish-bolts 10-00

18 beariug-platea 53-55

36 rail foot-bolts 40-00

18 outside clips 11*90

18 inside „ 8-84

36 packings 28-15

Something omitted in original^ 07. 7f; ? 44 Grover washers . . . /

Total .... 2,847-58

Lbs

1,049

1,418

81

10

53

40

11

8

28

27

2,729-23

or weight per yard run 347 - 20

332-30

318 -WOETHINCtTON on the permanent-way of railways. [Selected

II&. The angle fisli-plates on the inside of the rail are 21 • 6 inches long, and weigh 16*7 lbs. ; on the outside they are 23 • 6 inches long, and weigh 18 lbs. The foot of the angle-iron is cut out to take the heads of the spikes by which the rail is attached to the sleeper. The fish-bolts are 0 • 86 inch in diameter, and weigh 1 • 3 lb. each, and they are spaced as in la.

The rail-bearing plates are 7^ inches long, 5"1 inches wide, 0 • 35 inch thick, with a ^ inch rim at each side to hold the rail- foot, and weigh 4i lbs. each. They are let into the sleejier to give the rail cant of 1 in 1 6. To take the spikes they have two holes on the inner, and one on the outer, side of the rail.

The spikes are 0 • 6 inch square, 6 inches long, and weigh 0*7 lb.

The sleepers are 6 feet 8j inches long, 6 inches thick, 8'8 inches wide at the bottom, and at least 6 inches at the top. They are either of oak, larch, or soft wood impregnated with chloride of zinc. They are spaced in the same way as the iron sleepers in la.

In straight lengths and curves of more than ^ mile radius with oak and larch sleepers, bearing-jDlates are used only on the sleepers on each side of the rail-joint and the sleeper at the middle of the rail. "With soft wood sleepers a bearing-plate is used on every alternate sleeper. On curves of 15 chains radius and under, bearing-plates are used on every sleeper for all kinds of wood.

Baden State Eailways. These railways may be divided into two classes : —

I. Main lines with a permanent-way of steel rails upon iron

cross-sleepers. II. Local lines with steel rails upon wood cross-sleepers.

The outer cover of the rail-joint is in the form of a channel iron with the two legs bent outwards, the inner is a flat plate ; both are 21 J inches long, the outer weighing 22*6, the inner 11^ lbs. The bolt-holes on each side of the joint are 4^ inches apart, centre to centre, and the outer holes are 6 • 1 inches from these. In order to prevent the creeping of the rail, the end of the channel iron cover butts against the clip which holds the rail down to the adjoining sleeper.

The fish-bolts are 0*8 inch diameter, and weigh 1 -14 lb. each.

The iron cross-sleepers are of the Hilf pattern, without the middle rib. From a point about 10 inches inside, to a point about 5j inches outside the centre of the rail, the sleeper is bent to a slope of 1 in 20 to give the rail its cant, and from this outside

Papers.] WORTHINGTON ON THE PERMANENT- WAY OF RAILWAYS. 319

point it is bent down again with a slope of about 1 in 17, the end being closed (bent over). In the 29 feet 6 inches rail-length there are eleven sleepers ; those at the joint are 22 • 4 inches apart from centre to centre, the remainder being 33 • 2 inches apart. The rails

are fastened to the sleepers by bolts passing throiigh | 1 shaped

clips. The bolts are f inch diameter, and weigh 0 • 9 lb., and the clips weigh 1 lb. each. In the hollow under the clip is fitted a plate If inch sqiiare, and ^ inch thick, with a ^ inch hole for the bolt, so placed that the circumference of the hole is 0*27, 0*4, 0"54, and 0*66 inch from the four sides of the plate respectively, so that by turning this plate through an angle of 90° the gauge can be altered.

Various kinds of washers are used to prevent the bolts from working loose.

The weight of one rail-length (29 feet 6 inches) of this permanent- way is : —

2 rails l,43G-53 lbs.

11 cross-sleei^ers 1,037 •94 „

2 channel fish-jjlates . . . . 45'19 ,,

2 flat fish-plates 22-93 „

8 fish-bolts 9-17 „

44 clips 4G-56 „

44 gauge-plates 15'52 „

44 rail-bolts 40-73 ,,

Total .... 2,654-57 „

or 269-72 lbs. per lineal yard.

II. The fish-plates are similar in design to those in I, but smaller, being 16.^ inches long, and weighing, the outer 13^, the inner 5^ lbs. The bolt-holes on each side of the joint are 3*8 inches from centre to centre, the other two being 4 • 7 inches from these. The fish-bolts are f inch diameter, and weigh 0-9 lb. The rail-bearing plates are 6*13 inches long, 4*72 inches wide, and 0 • 3 inch thick, with an outer raised rim 0 • 3 inch high, and weigh 2f lbs. Each has two rectangular holes through which pass the spikes which hold the foot of the rail. These are 5 • 9 inches long, 0 • 6 inch square, and are slightly widened oiit at the head to enable them to be drawn out.

There are nine sleepers to each rail-length (24 feet 7j inches). Of these the two at the joints and the middle one of each length -are of oak, the remainder of red wood, both impregnated with corrosive sublimate.

The sleepers are spaced from centre to centre; 21^ inches at the joint ; 33 inches next space ; 34j inches the next two sjiaces ; and

320 WORTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. [Selected

35^ inches the middle two spaces. In ordinary cases rail-bearing plates are used only on the three oak sleepers, the rail being attached to the others by two spikes, one on each side. In sharp curves bearing-plates are used on every sleeper.

The weight of the ironwork for one rail-length (2-i feet 6j inches) of this permanent- way is : —

2 rails 876

2 channel fish-plates 2G

2 plain „ H

8 fish-bolts 7

6 bearing-plates IG

36 spikes 21

33 lbs. 46 „ 13 „ 06 „ 41 „ 43 „

Total . . . . 758-72 „

or 116 -87 lbs. per lineal yard.

Austrian Southern Eailway Company.

The permanent-way of this line is laid with Siemens-Martin steel rails upon wood cross-sleepers.

The outer cover of the rail-joint is a channel iron with the upper limb bent outwards, 23^ inches long, and weighing 17-8 lbs.; the inner cover is flat, 21-65 inches long, and weighing 10 -35 lbs. The l)olt-holes on each side of the joint are, from centre to centre, 4-16 inches, the other two being 5^ inches from these.

The lower limb of the outer cover is slotted near each end to take the spike which attaches the rail-foot to the sleeper. The fish-bolts are 0-86 inch diameter, and weigh 1^ lb. each. The nuts are provided with washers.

The rail-bearing plates are 7-35 inches long, 5-17 inches wide, 0 • 4 inch thick, and have a raised rim 0 • 4 inch high along each end to hold the foot of the rail. They weigh 5 lbs. each, and have two holes in the inside, and one on the outside of the rail for the spikes. These are 6^- inches long, 0-58 inch square, with T heads and weigh 0*7 lb. each.

The wood cross-sleepers are 7 feet 10^ inches long, 6-4 inches thick 10 inches wide at the bottom, and at least 5" 2 inches at the top. Those which are of oak and larch are not treated with any antiseptic, while those which are of beech are impregnated with the fumes from gas tar oil (Paradis system). In the 32 feet 9 J inches rail-length there are twelve sleepers spaced from centre to centre ; 20 inches at the joint; 30-9 inches the next space; 34-6 the remainder. In curves up to 15 chains radius rail-bearers are used

Papers.] WOETHINGTON ON THE PERMANENT-WAY OF RAILWAYS. 321

on every sleeper; from 15 to 25 chains radius they are used on eight out of the twelve including those at the joint, two spikes being used on the inside, and one on the outside of the rail at the sleepers without plates ; from 25 to 40 chains radius plates are used on four only out of the twelve, the spikes on the other sleepers being used as above. In curves of 40 chains radius and upwards, there are plates on four out of the twelve, but the remaining sleepers have one spike on each side of the rail.

The weight of the ironwork of one rail length (32 feet 9 j inches) of the permanent-way is : —

2 rails 1,499-14 lbs.

2 outer joint covers 35'71 „

2 inner 20-72 „

8 fish-bolts 9-88 „

8 washers 0-35 „

8 rail-bearing plates 39-69 ,,

56 spikes 38-89 „

Total .... 1,644-38 „

or 150-4 lbs. per lineal yard.

Hungarian State Railways.

The permanent-way of modern Hungarian State Railways may be divided into —

I. For main lines of the first rank. II. For main lines of the second rank. III. For lines of the third rank.

all of which are laid with flat-bottomed rails on wooden cross sleepers with rail-bearing plates.

I. The outer fish-plate, which is an unsymmetrical channel iron, is 24 '8 inches long, and weighs 21*6 lbs. The inner fish-plate is plain, 22 inches long, and weighs 13 "9 lbs. The fish-bolts on each side of the joints are 4 • 3 inches apart from centre to centre, and 5 • 9 inches from the next bolts. The fish-bolts of j-inch diameter weigh 0-9 lb. each.

On the wood sleepers rest rail-bearing plates with an inclination of 1 in 16. They are 7-08 inches long, 6-28 inches wide, 0*35 inch thick, and weigh 4 lbs. They are plain plates without any outer raised rim, and are punched for one spike on the outside, and two on the inside of the rail. With a view to their durability these plates are laid with the direction of the fibre at right angles to the length of the rail.

[the INST. C.E. VOL. XCV.] Y ♦

322 WORTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. [Selected

In straight lengths and curves of 35 chains radius and over three bearing plates are used in each rail length ; in curves of from 20 to 35 chains radius, four, and in curves of less than twenty chains, six under the outer and four under the inner rail.

The rail spikes have T heads, are 6 j inches long, 0 • 62 inch by 0 • 70 inch in section (the long sides being at right angles to the rail) and weigh ^ lb. each. There are three spikes to each bearing plate. Where there are no bearing plates in straight lengths and in curves of 50 chains radiiis or more, there are four spikes to each sleeper, one on each side of each rail ; in curves of 35 to 50 chains on two sleepers, there are six spikes, two on the outside, and one on the inside of each rail, and on the remaining sleepers one on each side of each rail. In curves of 20 to 35 chains radius, each sleeper has two spikes ou the outside, and one on the inside of the outer rail, and one on each side of the inner rail. In curves of less than 20 chains radius the arrangement is the same as for 20-35 chains with the addition of an extra spike on the oiitside of the inner on two sleepers. The classification above refers to rails of 8 metres (26 feet 3 inches) lengths, the same proportions being adhered to for shorter rails.

The cross sleepers are of oak and beech, the former occasionally, the latter always impregnated with zinc chloride. They are 8 feet 2;^ inches long, 5-9 inches thick, 9-8 inches wide at the bottom, and at least 6 • 7 inches at the top. They are spaced from centre to centre 22 inches at the joint, and from 31 • 9 to 38 • 2 inches the remainder.

The weight of one rail length (26 feet 3 inches) of the iron and steel in the permanent-way is : —

2 rails 1,172-85 lbs.

2 inside fish-plates 27'82 ,,

2 outside „ 42-22 „

8 fish-bolts 7-19 „

6 rail-bearing plates . . . . 23-81 „

42 spikes 31-47 „

Total .... 1,306-37 „

or 149 '32 lbs. per lineal yard.

II. Main lines of the second rank.

The outer fish-plates (angles) a:re 20 • 1 inches long, and weigh 13-2 lbs., the inner, which are plain, are 22-4 inches long, and weigh 7 • 1 lbs. The bolt holes nearest to the joint are 4 • 04 inches apart, and the other two bolts are 5*7 inches from them, centre to centre. The fish-l)olts are 0 • 62 inch diameter, and weigh 0 • 6 lb.

Papers.] WORTHINGTON ON THE PERMANENT-WAT OF RAILWAYS. 323

Some of the sleepers liave rail-bearing plates 4- 3 inches long, 5 • 9 inches wide, 0 • 3 inch thick, weighing 3 • 2 lbs. They have no raised rim along the outer edge, and have two spike holes on the inside, and one on the outside of the rail. The rail has a cant of 1 in 1 6. The rail spikes have T heads, are 5 ■ 9 inches long, and 0-6 inch x 0*55 inch in section and weigh ^-Ib. each. The number of spikes used with each sleeper varies according to the radius of the curve as in I. The wood sleepers are of oak, 7 feet 3 inches long, 5^ inches thick, 7*9 inches wide below, and at least 4 J inches above. They are spaced from centre to centre, 20*1 inches at the joint, and from 35 '1 to 37 inches the remainder.

The weight of one rail length (26 feet 3 inches) of the iron and steel in this permanent-way is : —

2 rails 879

2 inside fisli-plates (plain) ... 14

2 outside ,, (angles) ... 26

8 fish-bolts 4

6 rail-bearing plates 19

42 spikes 21

Total . . . . 965-30

02 lbs.

20

22

76

18

93

or 110 '.34 lbs. per lineal yard.

III. The lines classed of the third rank are 1 metre gauge (3-28 feet).

The fish-plates (plain) are 12 •35 inches long and weigh about 2\ lbs. each, the inner being a trifle thicker than the outer. Of the four bolt holes the middle ones are 4'1 inches apart, centre to centre, and the outer ones are 3 inches from these. The fish-bolts are ^-inch diameter, and weigh 0 • 28.

The rail-bearing surfaces of the wood sleepers have an inclination of 1 in 1 6, and are in part laid with bearing plates, 5 inches square and 0"2 inch thick, weighing l£ lbs. These have on each side of the rail a raised rim 1 inch wide, and 0 • 2 inch high, and have four spike holes. In curves of 10 chains radius and flatter, two bearing plates are used in each rail length ; in sharper curves four are used.

The rail spikes, which have a single-hooked head to hold the rail foot, and two side lugs for drawing out by, are 5^- inches long, 0 • 46 inch square, and weigh 0 • 35 lb. each. There are four spikes to each bearing plate. For sleepers without bearing plates, if the radius of the curve is 10 chains or more, there is one spike on each side of each rail ; if under 10 chains, there is one spike on each side of the inner rail, one inside, and two outside the outer rail. The wood cross sleepers are of oak, 5 feet 7 inches long,

Y 2

824 WOKTHINGTON ON THE PERMANENT-WAY OF RAILWAYS. [Selected

5^ inches thick, 7' 9 inches wide at the bottom, and at least 4*7 inches at the tojD. The rail joints are supported, not suspended. There are eight sleepers to the rail length (19 feet 8^ inches wide) spaced from 24*6 to 30 • 3 inches from centre to centre.

The weight of iron and steel in one rail length (19 feet 8^ inches) of this permanent-way is : —

2 rails 376-83 lbs.

4 fish-plates 10-05 „

8 fish-bolts 2-31 „

2 rail-bearing plates 3 - 47 „

36 spikes 12-38 „

Total . . . . 425-04 „

or 04-72 ll'S. per lineal yard.

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 325

(Students' Paper No. 243.)

" The Speed-Trials of the latest additions to the Admiral Class of British War- Vessels." ^

By David Sing Capper, M.A., Stud. Inst. C.E.

The speed trials of H.M.SS. " Camperdown " and " Anson," in April, 1887, brought to a conchision the trials of the Admiral class of British war-vessels.

The results show, that both in speed and economy, they siirpass any of their predecessors except the "Howe"; and the perform- ances of the " Anson," perhaps, are unequalled by those of any heavily armoured vessel afloat. The " Camperdown " excelled the "Howe" in speed, but was at light draught; while the "Howe" was tried at load draught. This, being allowed for, places the " Camperdown " second.

The " Camperdown " and the " Anson " are twin shijis, each 330 feet long between perpendiculars, 68 feet 6 inches in extreme breadth, and with a depth of hold of 26 feet 2 inches. Their displacement is 10,000 tons at load draught. The " Camperdown" was built at Portsmouth dockyard, the engines being manu- factured by Messrs. Maudslay, Sons, & Field, of Westminster ; while the "Anson" was built at Pembroke dockyard, the engines being manufactured by Messrs. Humphrys, Tennant & Co., of Deptford.

These ships thus afford examples of two of the three types of propelling machinery, recently described by Mr. S. H. Wells, Stud. Inst. C.E.'^ It will therefore be needless to enter minutely into their differences. A short summary of the dimensions of their leading parts will, however, be necessary. Each shij) has twin screws driven by two sets of engines of the three-cylinder compound inverted type. Each set has one high-pressure cylinder, 52 inches in diameter, and two low-pressure cylinders 74 inches in diameter. The stroke

* This communication was read and discussed at a meeting of the Students on the 20th of AiH'il, 1888, and was awarded a Miller Prize in the Session 1887-88.

" Minutes of Proceedings lust. C.E., vol. xci. p. oG6.

326 CAPPER ON SPEED-TKIALS OF BRITISH WAR-VESSELS. [Selected

is 3 feet 9 inches in each case. The engines of the " Camperdown " are in all leading particiilars identical with those of the "Benbow," supplied by the same firm. The cylinders are supported on four cast-steel frames, to which the double guides for the cross-heads are bolted. The condensers, one to each set of engines, are of brass, vertical, containing eleven thousand five hundred and fifty tubes apiece, giving a total cooling-surface of upwards of 17,000 square feet. The air-pumps, 30 inches in diameter, also of brass, are worked by wrought-iron beams from the low-pressure cross- heads. There is one air-pump to each low-jiressure cylinder. The circulating water is set in motion by centrifugal pumps, driven by separate engines with cylinders 12 inches in diameter. The diameter of the impeller is 4 feet. The condensers can be used as jet condensers should the centrifugal pumps break down. All working parts are of steel. The crank-shafts, made of Whit- worth compressed steel, are of annular section. The cranks are placed at angles of 120^ to each other. The boilers, twelve in number, in four stokeholds, separated by water-tight bulkheads, are oval with flattened sides, 12 feet 4 inches wide by 14 feet 1 inch high and 9 feet 10 inches long. There are thirty -six furnaces, each 3 feet 2 inches in diameter, with grates 7 feet 3 inches long, giving a total fire-grate area of 82 6i square feet. The iron tubes are three thousand four hundred and thirty-two in number, 2j inches in external diameter, and 7 feet long, having a heating-surface of 17,000 square feet. The total heating-surface is 20,400 square feet. The working-pressure is 90 lbs. per square inch. There are two funnels, oval, 10 feet by 6 feet, rising to a height of 75 feet above the lower fire-bars. The air-pressure under forced draught is maintained by eight fans, 5 feet in diameter, two in each boiler room. These are driven at a high velocity by small horizontal engines ; four fans 4 feet 6 inches in diameter are fitted for ventilating purposes. There is the usual comi^lement of auxiliary engines for donkey, feed, fire, bilge, and other services, which, together with electric-lighting and other small engines, exhaust into an auxiliary condenser.

The propellers are 16 feet in diameter, and 2 feet 8^ inches long, with a pitch variable between 18 feet and 21 feet, and struck at a mean of 19 feet 6 inches. The area of each blade is 19*38 square feet, giving a total to each propeller of 77 •52 square feet. They are made after a pattern supplied by the Admiralty, deduced from the experiments of Mr. Froude at Torquay, known as the modified Grifiiths type. In this type of propeller the boss is enlarged to 1 or j, or sometimes even a larger proportion of

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 327

the diameter of the propeller, and the i;seless portion at the tips is reduced. In the present case the propellers are four-bladed.

The engines of the " Anson " are similar in design to those of the " Rodney," " Collingwood," and " Howe," which were engined by the same firm. Several modifications, however, have been introduced.

The condensers, instead of being of brass, and placed separate from the engines, are of cast-iron, and form the frames upon which the low-pressure cylinders rest on the midship side. By this arrangement much space is saved without increase of weight. It is this firm's practice to fit air-pumps worked direct from the main pistons, without the intervention of rocking levers. Hitherto two pumps have been fitted to each condenser, to guard against the destruction of the vacuum resulting from any injury to a single pump. This by long experience has proved a totally unnecessary precaution, and in the " Anson " only one brass air- pump has been fitted to each condenser. It is placed behind the low-pressure piston-rod guide (Plate 8). The aggregate cooling surface in the two condensers is 17,000 square feet. The diameter of the air-pumps is 15 inches. The crank-shafts are made of Whitworth compressed steel in three interchangeable lengths. The cranks are jilaced at somewhat uniisual angles, the low- pressure cranks being at right-angles to each other, and the high- pressure crank, bisecting the supplementary angle between them, being at 135'^ to each low-pressure crank. The high-pressure crank leads, and is followed by the forward and aft low-pressure cranks in that order. The reason for this arrangement will appear from the twisting-moment diagram (Fig. 5).

The valve-gear is Stephenson's solid bar-link with indirect action. This gear certainly has disadvantages compared with the double-bar direct-acting links fitted by Messrs. Maudslay. It requires a larger eccentric, and consequently is more liable to cause heating in the eccentric straps at high speeds. It is, moreover, found convenient to hang it from the near eccentric-rod end. This gives a short drag-link, and consequent large influence of drag-link ujion the valve-motion. Yet with the very short valve-travel in use with double-ported valves, these objections are more theoretical than real ; and the advantages of greatly reduced first cost and solidity of construction more than counterbalance the slight disturbing efi'ect of the short drag-link, while a balance-cylinder relieves the weight upon the eccentric straps.

The boilers, eight in number, constructed throughout of Siemens- Martin steel, are 16 feet in diameter, and 14 feet 2 inches long.

328 CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. [Selected

They are of the high-cylindrical type. The thirty-two furnaces (Fox's corrugated) have a mean width of 3 feet 7^ inches ; sixteen of them are 7 feet long, and sixteen of them 6 feet 9 inches long. The total fire-grate area is therefore 797*5 square feet. They contain in all three thousand four hundred iron tubes, 2j inches in diameter, and 7 feet long between the tube-plates, giving a heating surface through the tubes of 17,150 square feet, and a total heating surface of 20,300 square feet. The two funnels are oval, 9 feet by 5 feet 6 inches. Their height above the lower fire-bars is 75 feet. The fans for forced draiight are eight in number, driven at a velocity of 400 to 600 revolutions per minute by Brotherhood's three-cylinder single-acting high- speed engines. The other auxiliary engines and fittings are similar to those of the " Camperdown."

The propellers are of gun-metal, 15 feet 6 inches in diameter, with a pitch variable between 18 feet and 21 feet, struck at a mean pitch of 19 feet 6 inches. The area of each blade is 18*14 square feet, giving a total of 72 • 56 square feet.

The manner in which Admiralty official trials are conducted has been already described by Mr. Wells ; but additional particulars are required, for the proper understanding of the trials which follow. The official trial consists of a four hours' run at fiill speed, both under "natural draught" and "forced draught." These are the terms in general use, although the trials would now be more accurately described as "open" and "closed" stokehold trials; for the stokeholds being so far below the deck, it is found advisable even on the natural-draught trial to keep the fans running at a moderate speed, althoiigh as the air-tight doors are open, the pressure cannot rise much above that of the atmosphere. During these trials several runs are made on the measured mile in both directions. The influence of the tide is thus eliminated by taking the mean of two runs, one with, the other against, the tide. This constitutes the first mean speed. The means of a number of such double runs, taken in pairs, then give second means. If sufficient runs have been made, a third and a fourth mean are taken, until finally the true mean speed is arrived at. The result is thus rendered practically free from the varying tidal influence.

Every half-hour during the trial indicator diagTams are taken, and the steam-pressures, vacuum, &c., recorded. The revolutions are taken from mechanical counters, the mean revolutions per minute for the four hours being used as the constant for working out the indicated HP. This gives a fair mean result. During the closed-stokehold trial, the pressure of air in the stokeholds on

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 329

the water-gauges is registered. The maximum limit allowed by the Admiralty regulations is a pressure of 2 inches of water. In torjDedo-boat practice this is largely exceeded ; a pressure of 4 and 5 inches being quite common in such craft, and some builders permit as much as 7 inches of water. It is doubtful whether any advantage is gained by so strong an air-blast. To supply too great an amount of oxygen is apt to cool the burning fuel unnecessarily ; while portions of unburnt and half-burnt coal are forced up the funnel by the draught and wasted ; and the temperature of the waste gases is raised, and heat by that means thrown away, which, with a more moderate draught, would have been utilized. Experiment proves that, even with a pressure of 2 inches of water, this source of loss is serious. Moreover, the deterioration of boilers and the burning of fire-bars are serious considerations. Some attempts have been made to obviate the latter difficulty by evaporating water in the ashpits, and these have proved very successful ; yet with a pressure of 5 to 7 inches it is doubtful if this arrangement would suffice. No doubt the Admiralty limit of 2 inches is therefore a wise one, and with such a moderate draught and careful stoking no serious injury is done to the boilers.

It is to be regretted that complete data are not obtainable for calculating the efficiencies of the engines and the boilers at these trials. Some of the data, moreover, are almost worthless from the roughness of the methods adopted for obtaining them. Such, for instance, is the method by which the coal consumed is measured. A certain number of buckets at the commencement of the trial are weighed, and a mean is taken of the weight per bucketful. This is multiplied by the number of bucketfuls used during the trial. But no particular care is taken to fill the bucket equally each time. No account is taken of the amount of coal in the furnaces when the trial commences, this being assumed to be equal to that which remains at the end of the trial, though no special care is taken that it shall be so. A more accurate criterion, of the economies of engines of different types, is obtained by comparing the amount of feed-water used per HP. per hour. By combining this with the amount of coal used, the true evaporative efficiency of the boilers would be obtained. This would be by no means difficult to accomplish, though it would undoubtedly entail extra expense and trouble during the trial. All the boilers are supplied with feed- water from a common tank, to which is returned the condensed water from the hot-well. The amount wasted, in evaporation and in condensation, is made up from a fresh water tank; or, as is now contemplated, from double distillers. So that,

330 CAPPER OS SPEED-TRIALS OF BRITISH WAR-VESSELS, [Selected

by some form of water-meter, all the feed-water diiring the trial could be automatically registered. The surplus water from the overflow could be similarly measured, and its amount deducted from the total feed-water. Care must then be taken that the water in the boilers stood at about the same level at the end of the trial as at the beginning ; the difference being estimated and the amount of water it represented being calculated from the known capacity of the boilers, or from an observation easily made when the boilers were being filled. The auxiliary engines take steam from the main boilers, and exhaust into an auxiliary condenser. The amount of steam used by them could be measured ; and the HP. developed by them estimated, and added to that of the main engines.^ The}" generall}^ take steam throughout the stroke.

Experiments as to the internal resistance of the engines would be exceedingly valuable ; but there is the difficulty of obtaining a means of measuring the effective HP. imparted to the propellers. The method, which most readily presents itself, is to measure this quantity by the amount of torsion produced in the propeller shaft. This would entail previous experiment to ascertain the coefficient for each shaft, to which there is the objection of trouble and cost. The only method of determining the internal resistance, in the absence of special exj^eriment, is by the approximation obtained from the tangent to the curve of indicated thrust in the manner pointed out by the late Mr. AV. Froude, M. Inst. C.E.

It is certain that such experiments would amply repay their cost; for by them an accurate comparison of the relative effi- ciencies of various tyj)es of engine could be made, which could not but result in increased care in design, and reduction of loss by internal friction. At present, labour in this direction is only carried on by surmise, and without definite data to go upon.

The trial of the " Camperdown " with open stokehold took place at Portsmouth on the 14th of March, 1887 (Aj)pendix, Table I). The ship was in light trim, drawing 22 feet 3 inches forward, and 24 feet 5 inches aft, the mean draiight therefore being 23 feet 4 inches, or 3 feet 5 inches under the designed load-draught. This represents a displacement of 8,292 tons. The trial was prolonged for six hours, and the results were very satisfactory. Full particulars are given in Appendix, Table II, where the table of half-hourly observations shows that the lowest collective HP. indicated was in the second half-hour, namely, 7,931 • 6 indicated HP., while the

' The reason for this addition is to supply correct data for comparison with engines, where the pumps are worked direct from the main engines by rockiug- levers or otherwise.

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR- VESSELS. 331

highest was 9,151*6, or 1,651 indicated HP. above the contract. The mean for the six hours was 8,622 indicated HP., or 1,122 above the contract. The fans were running, as above explained, for ventilating purposes, at a mean speed of 403 revolutions per minute, representing an air-pressure of about 0*3 inch of water; but the pressure was not recorded. The mean boiler- l^ressure for the twelve boilers was 83-3 lbs. per square inch. The pressure in the main steam-pipe in the engine-room was 80-3 lbs., showing the very moderate loss of 3 lbs. pressure due to friction and condensation. The high-pressure cylinder was in these engines placed forward of the two low-pressure cylinders, to shorten the steam-j^ipe ; and this, no doubt, consider- ably helped towards this result. The revolutions reached a mean of close upon 95 per minute for the two engines. The power was very nearly equally divided between the high-pressure and two low-pressure cylinders. The mean indicated HP., developed by the high-pressure cylinder, was: starboard, 2,151*5, and port, 2,173; while for the two low-pressure cylinders together, the collective power was : starboard, 2,222*8, and port, 2,074*9. The coal consumed was, by the above method of measurement, 974 cwt. 86 lbs. during the four hours, or over 2 lbs. per indicated HP. per hour. This would be reduced were the HP. of the auxiliary engines taken into account. The coal used was Harris's Deep Navigation, and the stokers employed were supplied by the dock- yard at Portsmouth, The mean speed on the measured mile was 16*3 knots.

The closed-stokehold trial of the " Camperdown " took place on Wednesday, the 16th of March. The draught of the ship had been increased to 22 feet 4^ inches forward, and to 24 feet 4.^- inches aft, or a mean of 23 feet 4^ inches. At this draught the displacement is 8,313 tons. The sea was rougher than on the former trial, tending to diminish the speed.

The mean indicated collective HP. was 11,742*16. The dis- tribution of power was not as it had been in the former trial. The three cylinders here developed approximately equal powers. The high-pressure cylinder developed less than on the former trial, the figures for the starboard being 1,885*5 indicated HP., and for the port, 1,828*2; while the low-pressure cylinders combined developed a mean of starboard, 3,948*9, and port, 4,079*6 indicated HP. This was accomplished by opening communication through a pass-valve direct from the main steam-pipe to the low-pressure receiver. By this means the pressure in the receiver was raised from 11-2 lbs. to 29 • 15 lbs. per square inch, and the mean pressure

332 CAPPER ox SPEED-TEIALS OF BRITISH •\VAK-VESSELS. [Selected

in the Ligh-pressure cylinder correspondingly lowered. The power developed by the high-pressure cylinder during this trial was 13 per cent, less than during the open-stokehold trial, although the toiler-pressure was about 5 lbs. per square inch higher, and the number of revolutions was greater. The extra steam evaporated was therefore utilized in the low-pressure cylinders, which greatly increased the power, though the advantages of a high rate of expansion were thereby impaired.

The weight of steam per HP. used must therefore have been increased, and this is perhajis the cause of the much greater con- sumption of coal per indicated HP. For the four hours during which the trial lasted, the total consumption of coal was 1,368 cwt. 42 lbs., being a mean of more than 3^^ lbs. per indicated HP. per hour, or an increase of 53 per cent, over the consumption during the open-stokehold trial. The advantages of increased power by so very simple a method, however, counterbalance the disadvantages, where the increase is required only under special circumstances.

It is to meet this division of power, when the strains are a maximum, that the cranks were placed at equal angles with one another. To place them at the most desirable angle for their maximum power, therefore, they are at a disadvantage when working at lower powers.

During this trial the air-pressure somewhat exceeded the regu- lation maximum, reaching the mean of 2*47 inches of water, the maximum being 2 "87 inches, and the minimum 1*87 inch. The fans were run at a mean velocity of 558 revolutions per minute. The speed on the measured mile was 17 '14:4 knots per hour, showing an increase of 0*85 knot for the additional 3,000 indicated HP. over the sj)eed at the open-stokehold trial. The mean revo- lutions were : port, 101-4; starboard, 101 -86 per minute. Details of the circle-turning, and reversing-engine trials are suj)i3lied in the Appendix, Table III.

The open-stokehold trial of the " Anson" took place on Monday, the 4th of Ajjril, 1887. The weather was very favoiirable, the mercurial barometer standing at 30 inches, and a light breeze blowing with a force of 2. The draught on this trial was 22 feet 6 inches forward, and 24 feet 2 inches aft, the mean being 23 feet 4 inches, and the displacement 8,318 tons, or rather more than that of the " Camperdown " had been. No preliminary trial had been made, yet the machinery worked smoothly, with no hot bearings, and the results obtained were highly satisfactory. The fans were kept running, for ventilating purposes, at a mean speed of 397 revolutions per minute. This, with the stokehold

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 333

doors open, produced a pressure of 0*2 inch of water in the stokeholds. The average steam-jiressure in the boilers through- out the trial vras 94*4 lbs. per square inch, vs^hile that in the engine-room, close to the high-pressure slide-jacket, was 89 • 65 lbs., the loss being a little under 5 lbs. from friction and radiation. This loss was somewhat greater than in the " Camperdown," the diiference being accounted for by the greater length of steam- pipe entailed by placing the high-pressure cylinder in the middle between the two low-pressure cylinders. The loss is, however, slight in any case, and shows the absence of ground for the objec- tions raised to the arrangement on this score.

The power developed was about halved between the high- pressure cylinder and the two low-pressure cylinders combined. The high-pressure developed 1,884-95 indicated HP., and 2,201-57 indicated HP. for starboard and port respectively, while the corre- sponding collective powers for the two low-pressure cylinders were 1,995-73 and 2,179-63. It is to maintain a good balance with this unequal division of power that the unusual angles of cranks above referred to were adopted. The effect upon the indicator-diagrams is noticeable ; for the high-pressure cylinder, exhausting just as the aft low-pressure cylinder is full open to steam, produces a slight bump aboiit mid-stroke in the aft low-pressure diagram, the receiver pressure being raised at this moment. This has the effect of reducing the variation in twisting moment for this cylinder, and at the same time reduces the evils of fluctuating pressure in the receivers, these being of small capacity. The back-pressure line on the high-pressure diagram, therefore, shows very slight deviation from a straight line.

The coal-consumption for the four hours was 699 cwt. 102 lbs., being a little over 2^- lbs. per indicated HP. per hour. The power steadily increased during the trial from 7,624 indicated HP. in the first half-hour to 9,700 collective HP. in the last. This was due to the small and broken coal at the bottom of the bunkers, which necessitated bucketful after bucketful having to be thrown away at the beginning of the trial. It was not, therefore, till some time after the trial commenced that the boilers evaporated their full steam. After the large coal was reached, steam escaped freely from the safety-valves, and no further difficulty was en- countered. This probably also increased the apparent consumption of coal.

The speed, from the mean of six runs on the measured mile, was 16-5 knots per hoiir. Details of the circle turning are contained in the Appendix, Table IV.

834 CAPPER ON SPEED-TEIALS OF BRITISH WAR-VESSELS. [Selected

The official full-power trial with " closed stokeholds " was made on Wednesday, the 6th of April. The results exceeded anything hitherto accomplished by engines of a similar weight of the com- pound type. There was a considerable breeze, so that the sea was rough, making it necessary to keep to the inside of the Isle of Wight. This, with the consequent frequent turning and extra strains on the engines, somewhat reduced the speed by log.

The safety-valves were blowing off freely throughout the trial, a proof of the ease with which the eight boilers generated the reqiiired volume of steam. During the last half-hour it was deemed advisable to ease down the fires somewhat. The average air-pressure during the trial was 1-9 inch, being below the maximum allowed. This pressure was maintained by the eight fans with an average speed of 44,1^ revolutions per minute.

The distribution of power between the cylinders remained the same as in the former trial. The means were : High-pressure, star- board, 3,028 indicated HP. ; port, 3,070 indicated HP. ; while the collective HP. for the two low-pressure cylinders was : starboard, 3,369; port, 3,122 indicated HP.

The economy was, if anything, increased by the use of forced draught, as far as can be judged by the coal consumption. The total for the four hours was 1,004 cwt. 85 lbs., the mean being a little under 2^ lbs. per indicated HP. per hour, or rather under the mean for the open-stokehold trial. The total mean indicated HP. developed reached 12,584, or 3,084 indicated HP. over the contract power. The speed, reckoned from the mean of five runs on the measured mile, was 17,435 knots per hour, while the patent log registered a mean of 17*6 knots. The ship had been somewhat lightened since the open-stokehold trial, the mean draught having been diminished to 23 feet 3i inches, giving a displacement of 8,277 tons. The reversing-engine trials are detailed in the Appendix, Table IV.

The sjiecial conditions, which influence the design of the ma- chinery of war-vessels, have been enumerated by Mr. Wells. ^ Of these conditions, those upon which the most stress must be laid are the keeping of the engines below the water-line, and the reduction of weight to its minimum safe limit.

It is to reduce the weight that the piston-speed of war-vessels has been gradually increased, until its mean value for vessels of the class under consideration is now from 750 to 850 feet per minute. The necessity for keeping the engines as low down as

' Minutes of Proceedings Inst. C.E., vol. xci. p. 366.

Tapers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 335

possible, and the consequent reduction of stroke, lias led to a gradual increase in the number of revolutions to attain this piston-speed, until they have reached 110 per minute. In some triple-expansion engines, 130 to 1 50 have been attained for engines of about the same power. The chief objections to so short a stroke, accompanied by a small ratio of connecting-rod, are, the larger effect upon steam dis- tribution produced by the angularity of the connecting-rod, the much larger value of the normal thrust upon the guides, and the consequent increase of the absolute value of the friction upon them. The connecting-rods are rarely more than four times the length of the crank, and sometimes less than that. The effect ujion the twisting moments is shown by Figs. 4 and 5.

It will be instructive to examine briefly how far the two sets of engines fulfil the conditions above referred to. The engines of both ships are entirely below the water-line. The diameters of the cylinders are the same ; they have the same length of stroke, 3 feet 9 inches, and the same length of connecting-rod, which is in each case four times the length of the crank.

The total weight of the machinery of the " Camj)erdown " exceeds 1,276^ tons; that of the "Anson" is 1,150.^ tons, while the contract weight was 1,225 tons. Thus the engines of the " Camperdown " exceed the contract weight by upwards of 50 tons, while those of the " Anson " are over 74 tons below it. There is thus a saving in weight of more than 125 tons in favour of the "Anson." In comparing the weights per indicated HP. developed, this advan- tage is further increased. With open stokeholds the weight per indicated HP. per hour of the machinery of the " Camperdown " is 364 lbs. ; of the " Anson," 309 • 6 lbs. On the closed-stokehold trial the corresponding figures were, for the " Camperdown," 267*3 lbs. per indicated HP., and for the "Anson," 204*7 lbs. The weight of the engines of the " Inflexible," built about eleven years ago by Messrs. John Elder and Co., of Glasgow, was nearly the same as of those of the " Camperdown"; while the maximum power developed was only 8,483 indicated HP., or nearly 25 per cent. less. This remarkable increase in power for a given weight is largely due to a more lavish use of cast-steel and brass than was possible at that time ; but still more is due to forced draught, and the consequent reduction in weight of the boilers, and to the higher piston speeds.

The adoption of triple-expansion engines, and a more extended use of cast-steel, as the foundries become better able to cope with the difficulties of casting, will probably give results surpassing even those above quoted.

336 CAPPER ON SPEED-TFJALS OF BRITISH WAR-VESSELS. [Selected

It may be pointed out that triple-expansion engines are neces- sarily heavier, power for power, than simple or compound engines with a given boiler-pressure. The advantage in point of weight lies in the possibility of using much higher steam-pressures with economy. By this means the size and weight of the cylinders is greatly reduced ; while the boilers may be fewer in number.

The weight of the boilers, when full of water, always bears a large ratio to the total weight of the engines. By computing the indicated HP. developed per ton-weight of boilers, a fair idea of the relative efficiencies of various sets of engines, and the advantage of forced draught may be obtained. The method usually adopted is to compare the ratio of indicated HP. per square foot of grate. Biit, as pointed out by Mr. E. Sennett, M. Inst. C.E., in a Paper read before the Institution of Naval Architects,^ this is by no means a fair method of comparing boilers dissimilar in design and construction. For reference, however, the indicated HP. per square foot of grate has also been given in the accompanj^ing Table IX in the Appendix. The total weight of the boilers of the " Cam- perdown," with uptakes, funnel, water, and all fittings, is 662 tons. With open stokeholds, therefore, the engines developed 12-4 indi- cated HP. per ton of boiler; while with closed stokeholds they developed 1 6 • 8 indicated HP. per ton, or above 35 per cent, more with closed than with open stokeholds.

The total corresponding weight of the boilers of the "Anson" is 590 tons. With open stokeholds, therefore, the " Anson " developed 14-1 indicated HP. per ton, and with closed stokeholds 21*3 indicated HP., showing an increase of 50 per cent, by the use of forced draught. The difference between the two vessels, by grate areas, is not so marked; but as the boilers are of different types, this, as above pointed out, is not so fair a way of comparing them. The "Camperdown" has a total fire-grate area of 826 '5 square feet, and therefore developed 10*43 and 14-2 indicated HP. per square foot for closed and for open-stokehold trials respectively. The "Anson," with a grate area of 797 "5 square feet, developed 10 "43 and 15*78 indicated HP. for closed and open-stokehold trials respectively. The "Anson" no doubt derived some advantage by the reduction in the number of boilers to eight, partly due to the increase of 10 lbs. per square inch in boiler-pressure, this being 100 lbs. per square inch to 90 lbs. in the " Camperdown." Yet this was again counterbalanced by their extra weight and size, so that

' Transactions, vol. xsvii. p. 17C. 1886.

II

Papers.] CAPPER ON SPEED-TRIALS OP BRITISH WAR-VESSELS. 337

the boilers of the " Anson " were Init 20 tons lighter than those of the " Howe," which had the same niiniher of boilers and of the same type as the " Camperdown."

The slight difference in weight of the engines is chiefly account- able to the difference in type. In the " Camperdown," double steel frames and massive cross-head and double guides replace the single frames, light columns, and single guides of the "Anson." The engine-room is thereby rendered much less open and free of encumbrance. The position of the condenser and pumps of the " Camperdown," behind the engines on the wing side, also takes away from the available platform space. The pumps, as already stated, are worked by rocking-levers from the low-pressure cross- heads. This, with the low horizontal condensers which obtain in the merchant service, is no doubt a convenient method of working them. But when the condenser is vertical, and separated from the framing, as in the " Camperdown," these pumps are clumsy and block up the passage-way to an unnecessary extent. Long- continued experience has proved that the direct-acting pumps of the " Anson " type give resiilts as good, if not better, than those from pumps worked at a lower speed. A series of experi- ments made with the pumps of H.M.S. " Collingwood " gave at from 88 to 90 revolutions, and a piston-speed, and therefore plunger-sj)eed, of 630 feet per minute, and a vacuum of 28 inches, a maximum pressure in the barrel of only 5 lbs. above the atmo- sphere. The speed of piston on the "Anson" reached at the closed- stokehold trial a maximum of 820 feet per minute. This is considerably greater. But experiments show that where there is a short delivery, and a small volume of water to be set in motion at each stroke, the maximum pressure in the pump-barrel does not increase with the speed of plunger, but tends rather to diminish with a well-designed pump ; while at higher speeds, where the inlet and outlet are of good proportions and unobstructed, the variation in pressure is much reduced and a much steadier diagram obtained.

There is some advantage in the rocking-lever over the direct- acting pump as regards the balancing of the weights of the recipro- cating parts ; yet with a three-cylinder engine, with cranks placed at the most suitable angles, and carefully designed parts, no additional aid of this kind is needed. In horizontal engines, where it is always more difficult to obtain a good vacuum, direct-acting pumps are universally used.

The diagrams and tables which accompany this Paper need a word of explanation. The midship section and displacement co-

[tHE INST. C.E. VOL. XCV.] Z

338 CAPPER ON SPEED-TRIALS OF BRITISH TVAR-VESSELS. [Selected

efficient curves are shown in Figs. 1 and 2 respectively. The former is of little value except in a case like the present, where ships similar in design are compared.

Fig. 1.

MIDSHIP SECTION COEFFICIENT.

	,^	•
		,^	-n	^--	^^	•v
	^					^	s.
							-■s	N
								W
									\
									■"nA
									.■■■k 1
									\	\\
									\

B 9 10 II 12 13 /♦ IS le II Id SPBEO IN KNOTS

Hcwc

ColLuigyvoocb

Alison

CojnpcTLwwn/

Fu

280 2S0 2W 220 200

tea

160

		DISPLACEMENT	DOEFFIC	lENl
1. H P.
			— -	k
J	^					^	•«t^
								V^
										V
									\	[\
										\ ^

IZ 13 J* IS 16 n IS

SPBEO IN KFIOTS

The ratios of indicated HP. to heating-surface and grate area have also been calculated, and the results recorded in Table IX. Fig. 3, which represents the curve for indicated thrust of the

propellers, is calculated by the formula - — ^ — ' — , where

P = Pitch of propeller,

N = Number of revolutions per minute.

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 339

As no j^rogressive trials were made of the " Camperdown " and " Anson," the curves for the " Collingwood " and " Howe " have

THRUST III TOMS

too

Fig. 3.

INDtCATED THRUST.

Sf€ED IN KNOTS

Howe

0)lLLn(jwoo(L

Anson.'

						j
A«
							i
						1
						l\-
						'
					.'	1
fi";					/
					/
				/:
				//
			//
		^	/y
	—	/"
in

CcLnipcrdx)wn,

been added for comparison. The above foruuila gives the thrust exerted by the propeller in tons. The curves are obtained by

z 2

340 CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. [Selected

plotting this with vertical ordinates equal to the thrust and abscissas equal to the ship's speed in knots. The late Mr. W. Froude, who proposed this system of plotting the efficiencies of the propeller, j^ointed out that such curves showed a persistent refusal to cross the vertical asymptote at the zero of the scale, and that the point where the curve did cross the asymptote showed the amount of internal resistance in the engines. To obtain this, the tangent must be drawn to the curve at the lowest calculated point on the curve, which should correspond to a sjieed of not more than 3 to 5 knots per hour. The abscissa under this tangent being divided in the ratios of 0-87 to 1 (the resistance being found to vary for these speeds as the 1*87 power of the speed), and a

Fig. 4.

MEAN !<90 / V^ ,'...*> ...

TwiBting moment on crank-shaft H.M.S. " Camperdown " starboard engines. Full power.

Closed stokeholds.

vertical drawn through this point, a horizontal line through the point where this vertical ordinate cuts the curve will cut the vertical asymptote at a point which corresponds to the power absorbed by internal friction in the engines. The lowest speed for which data are obtainable, for the " Collingwood" and the " Howe," is 8 knots per hour. The curve for the " Anson " lies considerably below the curves for the above-named ships ; while for the " Camper- down " it nearly coincides with the upper part of that for the " Collingwood."

The slip-ratios for the propellers are given in Table IX. In the " Camperdown," on the open-stokehold trial, the ratio for the star- board propeller was 11 per cent, and for the port 12 per cent. On the closed-stokehold trial it was 12 per cent, for l)oth starboard and port.

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 341

For the " Anson," on the open-stokehold trial, the ratio for star- board was 9 per cent., and port, 12 per cent. ; on the closed-stoke- hold trial, for starboard, 16 per cent., and for port, 17 per cent.

The much greater value of the slip-ratio, for the propellers of the " Anson " on the closed-stokehold trials, is partly accounted for by the unexpectedly high number of revolutions at which they were run, and for which they were not designed. The slip is calculated from the mean speed of ship and mean number of revolutions. The propellers of the " Camperdown " show little more slip for the higher than for the lower speeds.

Fig. 5. mjit.ikeo _

MEA N 2132.

/ -

Twisting moment on crank-shaft H.M.S. "Anson." Tort engines. Full power. Closed

stokeholds. In these diagrams allowance has been made for weight and inertia of reciprocating parts. Scale for vertical ordinates 1 Inch = 1,100 inch-tons.

In the curves of twisting moment on the crank-shaft in the " Anson," Fig. 5, allowance has been made for weight and inertia of the reciprocating parts. The moments have been calculated from the ordinates taken from indicator diagrams, obtained during the last half hour of the closed-stokehold trial. The method adopted was as follows : —

The indicator diagrams were divided by vertical ordinates at ■^V part of the stroke apart. A true diagram of pressures was constructed by subtracting from these ordinates the values of the

342 CAPPEE ON SPEED-TRIALS OF BRITISH WAR-VESSELS. [Selected

back-pressures, shown on the indicator card for the opposite end of the cylinder. The curve, given by the inertia of the reciprocating parts, was then constructed by the ixsual method, and algebraically added to this diagram. The low-jiressure diagram was read off on such a scale as would harmonize its ordinates with the scale of the high-pressure diagram, so as to represent the moments from the two cylinders on the same scale. A correction was made for the dead weight of the reciprocating parts. The ordinates thus obtained were plotted on a horizontal base, representing the circumference of the crank circle at points corresponding (allowing for angularity of connecting-rod) with the ten positions of the ordinates on the indicator diagrams. The sum of the ordinates of the three diagrams thus obtained, at any point, gives the true twisting moment for that point.

The effect of the inertia of the reciprocating parts is to shift the maximum moment for any crank, from the first and third to the second and fourth quadrants of the crank circle ; and, in combina- tion with a finite connecting-rod, it considerably affects the maxi- mum and the minimum moments of the combined diagram. For the diagram of twisting moment on the crank-shaft for the " Cam- perdown," Fig. 4, the Author is indebted to Mr. S. H. Wells, Stud. Inst. C.E.

The Paper is accompanied by several illustrations, from which Plate 8 and the Figs, in the text have been prepared.

[ArrENDix.

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 343

APPENDIX.

Table I. — Abstract of Mean Results obtained	\t Trials of H.M.SS.
" Anson "	and " Campekdown," in 1887.
	H.M.S. " Camperdown."	H.M.S. "Anson."
	Open	Closed	Open	Closed
	Stokehold.	Stokehold.	Stokehold.	Stokehold.
Date of trial	14 March. 16 March.	4 April.	6 April.
Where tried	Portsmouth.	Inside Isle of Wight.
Draught of water {^J™^	22 ft. 3 ins. |22 ft. 4^ ins. 24 ft. 5 ins. :24 ft. 4| ins.	22 ft. 6 ins. 22 ft. 4 ins. 24 ft. 2 ins. 24 ft. 3 ins.
Area of midship section	1,342 sq.ft.	1,345 sq.ft.	1,339 sq. ft.	1,330 sq.ft.
Displacement ....	8,292 tons.	8,313 tons.	8,298 tons.	8,277 tons.
Load on safety valves .	( 90 lbs. per| \ sq. inch /	|90 lbs. perj \ sq. inch j	flOOlbs.perj \ sq. inch /	1100 lbs. per \ sq. inch.
Air pressure (inches of water)		2-47	0-2	1-91
Steam in boilers (mean pres- 1 sure /	83-3 lbs.	87-5 lbs.	94-4 lbs.	101-0 lbs.
		Ford. Aft.	Ford. Aft.
Vacuum (Starboard engines \Port engines . .	28-3 ins.	26-9 ins.	29-3 27-928-2 26-8
28-3 ins.	27-4 ins.	28-2 27-627-8 26-8
Steam pressure in receiver .	11-2 lbs.	29 -15 lbs.	3 -53 lbs.	11-96 lbs.
1 Starboard .	94-4	101-4	95-2	108-6
Revolutions < Port .	95-4	101-9	98-7	108-9
Mean . . .	94-9	101-7	96-9	108-8
	Starb. Port.	Starb. Port.	Starb. Port.	Starb. Port.
Mean pressure in Jhp. .	47-7 46-738-5 37-240-9 47-457-8 58-4
11-2 11-919-9 20-510-7 11-315-8 147
	StarboardJM-^^Hip^' ( Total .	2,172-9	1,885-5	1,884-9 3,028-2
	2,074-9	3,948-9	1,995-7 3,363-8
	4,247-8	5,834-4	3,880-6 1 6,392-0
LHP.		' '
Port, .j^'^-'^'^fli?' 1 Total .	2,151-5	1,828-1	2,261-6	3,070-8
	2,222-8	4,079-7	2,179-6	3,122-1
	4,374-3	5,907-8	4,441-2 1 6,192-8
Collective HP	8,622-1	11,742-2	8,321-9 12,584-8
Fan engines, mean revolutions	403	558	397-5	441-3
(Joal consumption, lbs. perl LHP. per hour . . . /	2-11	3-26	2-35 coal verv small.	2-23
Description of coal used	Harris's Deep Navigation.
Time at full speed without) stopping /	6 hours.	4 hours.	4 hours.	4 hours.
Stopping and starting —	Starboard.	Port.	Starboard.	Port.
Stopped from full speed	6 seconds.	5 seconds.	18 seconds.	14 seconds.
Being stopped started astern	10 „	8	12	13 „
From astern full speed tol ahead full speed . . . /			14 „	15
Speed, If/ l-^g ■ ■ ■ ■ (Measurcil mile .	16-3	17-06 17-14	16'-52	17-6 17-44

344 CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. [Selected

Table I — continued.

ClRCLE-TuENING.

j H.M.S. " Camperdown.'

Starboard

Circle.

Port Circle.

H.M.S. " Anson."

Starboard Circle.

Port Circle.

Budder |

Helm to starboard or port . ' aport

Angle of rudder . . . . ' 34°

Time taken to put helm to this) k- ,

1 -ii i • • Wo seconds

angle with steering engine f

Turns of wheel from amid-i

ships position of helm

Number of men at wheel

Half circle made in . Full „ „ Diameter of circle, yards Revolutions of engines after "I helm was put up, jjer circle/

astarboard 34°

11 seconds

2min. 10 sec. 2miu. 20 sec. 4min. 538ec. 4min. 42 sec.

630

650

Area = 190 square feet, aport I astarboard 341° I 34 J° 18 seconds 18 seconds

4 4

2 2

2min. 46 sec. 2min. 28sec.

5min. 47 sec. 5min. ISsec.

652 I 632 Starb. Port. Starb. Port. 492 455 496 516

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 345

= W

iSa-

<1

C3COC5(N(N-*lrtCDr^COOlO

CO <X) CO CO

1-1 -t< Oi

_CO_l> oo_oc__oo_

lO oq lo !-i ■*!

r-H iO

CO CO C5

■* Ttl CO >o ■* OJ >0 05 IM 05 O i-H

C0CO-*0DCOO5t^(M COO^ICOCOO-^OCT)

■^a-^rt*-^-^-^-^-^

PH O

Oi <M C5

O GO -t<

-fl iO CO

O O (M

c^ c^ cq^

(M CO l>- (H

t> CO Ci

t^ 05 05 O O

(M (M <N <N (M (M (M

0OiOCOO5Tt(-*(-*(-*rH0OQ0O5

O 05 O

(M 1-1 i-H

CO •* I-H lO i-H IM

IMi-H(M(M<?q(MlM!N(NIMC<)!N

»o-*iTt<i-icoco-*io(Neocot>

CO t^ -t<

(M CO

CO -JH t^

Oi 05 (M

CO CO ■*

Ci CO CO

c:2 CO c»

CO (M lO

(M 05 00 (M

1-1 Ttt

*^ ^5^ TfH Tfi

eoi«>«<Na3iocoi>i>i-(Oso5

l>iCQ0IM-*t~l>O5 »O0000C0C005OI> 005COt>OpHt-(CO

1-H CO CO CO O pH

(NrHi-Hi-l!M<M(M(M(M(M(MIM

cqOSOJCnoOCOQOOOCOIMCOt:^

CO I-H »o Oi (M lO 05 l> I-H 05 O I-H

eoc<iiococoi>oo

O lO O 05 -t< i-H 03

(M (M 1-H (M

Cq I-H (M IM (N ~Q0 CO CO ^C^ C<l~

(M (M IM

IM iM IM !M^

rH 05 C<I CO O

IM (M t-H CO

iM050i-liO(M05COCO»0-*t>

(M lO t- 05

00 CO 1-H lO l> 1-1 .Q

i-J 1-H O O 05 1-H

00 IM 00 1-H CO (M CO

I-H I-H !M 1-H 1-H

.eOOOIMCOlMi-H-*HIMCOCOOr-l

eoi>GOoooocooocnoocoo

■*! 1*1 -^ i*< lO

CO

CO

00

o

(M

CO

IM

(M

t-

CO

CO

05

■*

CO 05

CO 05

(M 05

-*l 05

CO 05

05

05

05

CO 05

05

CO 05

<J>

05

CO

CO

O

05

IM

05

CO

t-

05

00

r

l>

-+<

. rHiw n|N H« i-tM -^IN «HlM Hoi

£00 OOOOOOOOOOOOOOOOOOOOOO £<M <M(MIMIM(M(M(MIMIM(MIM

, H|« H« H« Hn HlN f-i|oi i-<I« w|« Hn F-lfj

2COODCOCO00COCOt^Q0CO0CO0 £ IM IMIMIMlMIMCMiMIMIMIMIM

;3t-

niTji ^i« ^b) H« p-lN H* mwi e:wi COOlCOCOCO-tlCSi+l l>t>QOCOOOO0t^O0

O CO CO

3t- t-

COCOOOCOOOOOCOOOOOCO

3		.^	^	rJ3	^	^	-d	ri	rd	^
iM				CO	t>	00	05	O	^H

5 g

OQ B

CO S

CO ci

346 CAPPEK ON SPEED -TRIALS OF BRITISH WAR- VESSELS. [Selected

GO

!M

t^

t^

(M

_M

CO

^

o

C5

o

CO

t-

C5

■*

CO

= .•

«

o

o

C:

o^

C5

Ci

o

!M

S-

CO

t-

-+I

00

t-~

00

CO

-H

:^

o

'^^

O

00

O

00

t-

t-

^

n;

;i;

;::3

m;

^;

<M

^3

I-H

^J

o

,— «

o

o

^

(M

CO

CO

CO

o

C5

»o

1—1

t-

CQ

t>

t~

•s

lO

■*

t>

00

CO

00

C5

'^iH

t^

o

i>

^

I-H

a

Ttl

CO

o

H

C5

t>

o

00

O

00

CI

lO

lO

ui

CO

»o

CO

m

lO

(M

•f

CO

CO

lO

CO

-tH

(M

•rH

<M

CO

i-i

a

CO

:-•

d

o

-f

o

CO

CO

CO

Ci

o

Ci

&

t^

c^

•fl

CO

CO

CO

CO

00

t-

t>

00

<-<

<— 1

o

CO

o

o

M

CO

■*

Tt<

-<*< •o

_'*_

-*

t-

(M

c^

t>

00

•*!

l>

c;

o.

C5

C5

(^

o

CO

cq

o

CO

■x

o

CO

-t*

o

IM

>-i

■—1

00

o

CO

l>

00

CO

CO

<x>

(N

?-l

rH

1—1

I-H

I— 1

^

I-H

,_,

^

<M

(M

r^

t-

CO

,_^

GO

.H

I— c

CO

r-(

l>

CO

(M

(N

CO

CO

5

eS

CO

,_(

CO

©

o

^^

o

o

-tl

*— '

o

-+I

Ci

t-

lO

!M

I>-

CO

oo

o

»

o

00

a

OD

CO

>o

m

>o

m

CO

tn

m

lO

o

■

o

tn

-*l

■*

,_l

•s

o

»o

CO

o

(M

lO

lO

CO

C5

H

o

d.

,—1

r-l

t-

(M

CO

o

CO

00

CO

C5

t-

o

o

■*

d

ao

!M

co

o

00

C5

o

C5

ai c

;«

,o

M

•*!

CO

CO

tH

CO

CO

1*<

CO

Tt<

cq

o

(M

CO

C5

C5

CO

CO

I— 1

CO

l>

(N

■*

CO

CO

o

'tl

■

c.

O

CJ

o

00

CO

o

CO

t^

10

3

CO

CO

(M

o

C5

-*<

^

o

00

^ .

l>

C5

o

o

OS

00

00

3 "^

,—1

r-l

,_,

1—1

N

,_!

I-H

CO

^

o '

1

lO

lO

lO

•-^ c

aiO

t-

CO

o

o

t--

o

m

^

o

o

.

— "

o

,—1

I-H

o

1-H

o

o

o

=0

^

Si =

(S

"—I

<M

CO

.O

t^

lO

PH

S.-S

C-

5m

CO

CO

t-

4i

O

i)

cb

t>

. —

HH

I-;-*!

CO

CO

co

CO

CO

CO

CO

CO

1

m

w

lO

x-O

o

o

CO

05

05

C5

<N

o

a

c

CJ n

'^ 5^

t1

^JCi

r_4

o

00

C5

00

o

o

C5

.2

r— 1

— ^-

t-l

~io~

~<N~

(M

uo ^

CJ

.9 '■

l>

S

M

c.

,a CO

CO

LO

^

^

O

05

CO

GO

o

r-l _

h^CO

CO

CO

rjH

-*l

^

CO

CO "^00

CO

1 s

t>

1— 1

TfH

l>

O

(M

'^ - oi

f

g'la

s

1— (

o

C5

1—1

-ifi

CO

CO

J—i

,— t

^

(S

o

o

C5

o

o

o

o

o

o

c^;^^

r-H

I-H

I— (

!M

l-H

I-H_

CO

-H

a

#■

.•^

CO o

<M

CO

t^

d X s

o

o

O

o

CO

(N

(N

CJ

S.3 ~

o

o

O

o

o

O

o

o

o

^

r— 1

r— 1

I-H

1-H

T-H

I-H

1-H

Ttt

o

.^

„.•-*!

H?!

Hm

1-^

&= ^

^

c t^

t^

t^

t^

l>

t-

l^

t-

t^

§^£S

;i4

(M

(M

OJ

(M

(M

(M

iM

a

^

^^1

^ c3

c t^

t-

t-

CO

l>

t^

t^

t~

CO

^^

►=1M

(M

(M

(M

(M

(N

(M

(M

<»

»

ss^-

•li

m

H-i<

WW

-*)

-*l

1^

20

Oi

O

<M

-ri

CO

(M

o

<M

"CO

O

00

00

CO

00

00

00

00

c r u

•

CO

^

UO

o

a-«i

r4«

H*

HN

Kt*

KM"

M*

< ""^

'^ »

O

■*!

r--

O

o

t-

lO

t-

"00

00

CO

00

C5

o

00

00

»

- «

^

-r

_

^

-

_

-

-

s ?

-^

Cl

- C-:

-f

i^

—

t-

cc

1^

—

,3 a

OD CO (M (N ^

-^^ a t I

Papers.] CAPPEB ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 347

= W

S 9

c of

* to ^

0.3

sal

c3 O 3 Of See

1-H rH !M

i-H i-H (M

O T— ( T— I

I— I 1— I I— ( (N

05 i-H r-(

05 T-l 1— I

Ct-

Hm «!■* nw KW Kl-J" «Hi

SB

ho o

o 2

05 Oi di 05

CO Sh ^

		o
	lO	03
	CO	^
f^	oq
'i
o		r^
rJ3	o	o

•3 a: °

848 CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS, [Selected

W

w

"^ V. ^

So

c p.

c3 n ^ S s 5

. Ot/J

© l-H l-H i-H M

eoooMiweoOTMef:

CO T-H

t> 1-1

O 1-1

CO CO CO CO

c g Ec o

c ?

rH CO -H rH

CO <N t^ O

Tt< l> Tfl 1-1

O IM	r^	t>	«o	t-	CO	05	CO
C5 O CO »o	O	o	CO	CO	(M		CO CO
CO CO	CO	CO	CO	CO	CO	CO	CO

jj ^ ^ lO T}H ^ Tt^ "^ ^

;2C5

• 1—1 O 1—1 1— I

2 O CO CC lO CO

5"^

. Kb*

1^

[^o o o o o o

'^ ^ A ^

s	'a
,ja	A
'S	en

r3 2 ci t-i CO

c2 "Z '"' c^ '-'

rt oj - -

* -2 * a a

=; -: o

-^1 a -? s s

O

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 349

CV.

>y\

c-^

°^

-tl

00

00

o

o

GO

o

00

o^

CO

Ifi

lO

CD

CO

o

CO

»o

^

CD

00

lO

CO

o

C5

CO

05

^

00

00

C5

00

05

OD

05

CO

<

■H

CO

.

<S

ci^ t^

00

(>)

o

^^

o

(M

,_^

C5

e5

C5

00

C5

Oi

o

O

C5

CO

o

fe

o

00

Ph

^

o'-'S

CO

o

CD

lO

t-

t^

GO

CO

^

°00

00

05

GO

oo

oo

00

oo

00

TS

;-(

•T3 U

lO

Pm

°S

CO

o

CO

<M

<M

CO

o

^

05

Ol

as

Oi

a>

05

05

crs

o

2

t^

"3

05

M

^

n'^

-H

-¥

00

00

CO

t-

CO

o

03

^

^

°00

CO

00

00

00

00

00

00

oo

■a

IC

Oj if

IM

o

(M

^

^^

iM

^

C^l

cq

oo

o

o

o

o

o

o

o

o

■s

1

rH

OS

o'-fi

CO

CO

o

CO

CO

CO

C2

t-

^

°co

00

00

C5

00

00

00

00

00

■a

;-,

03

1

■s

"oo

o

CO

^

00

o

o

^

1—1 CO

e5 t-< o

o

05

O

05

o

o

o

Oi

1—1

1—1

s

rH

J

o

-H

CO

^H

CO

C<1

CO

_l

fM

-

5

°t-

t^

c~

t>

t>

t-

o

t^

t-

lO

•

;

i

„^

iM

o

o

1—1

o

CO

o

o

g

:

"CO

t>

t-

t-

t-

t-

c-

l^

o

dH

'

^

^

3

05

tn

3

o'«

CO

CO

00

CD

CO

(M

CO

-H

a

£

t>

t>

c-

t^

o

t>

t>

t>

l>

o

3

2

t

4

fl

~co

•5s

^

J

^-n

t^

t^

o

o

o

00

lO

t^

0

-

<

°L-

t^

t^

00

00

00

t-

l>

c-

C5

1

-

3 3

IM

00

C<1

■*

o

cq

(M

CO

g

•

t^

t-

o

t^

t^

t^

t-

c-

CO

'

^

lO

3

oq

3 >

o"^

CO

CD

00

o

05

o

cq

GO

°C-

t^

t^

t^

00

t-

00

00

t^

3

{.

<„ m

TS

J

^

^

^

^

a

"^ 3

^

c5

c =

'M

CO

•tl

o

s

t^

v>

350 CAPPEK ON SPEED-TKIALS OF BRITISH WAE-VESSELS. [Selected

w

o

.

^

^5

C-1

^

C-l

CO

(M

(M

oq

c

1

'+1

'JH

^

-tl

Tf

■*!

-+I

-*i

~eo

^

o

CO

C5

C5

■M

O

C5

00

<!

o

C5

C5

c;

o

o

C5

C5

ct2

<j

■g

'^

3

c«

CO

(M

M

-H

CO

(M

O

r>

■2

^02

CO

C5

o

o

o

o

o

C5

fa

'"'

'"'

■J

o

y^

Ch

>»

^5''

i-O

M

cc

**t

-+I

lo

>o

o

"=c;

o

c;

o

cs

C5

C5

C5

«

T!

P<

e3

i

■g

OS

fe

3

,^

-^

o

o

o

o

C5

CO

o

•

&

CO

o

o

o

o

O

O

o

2

o

f— t

f— t

I-H

rH

rH

r— *

I-H

I-H

I-H

"3

fa

J

~co

2

^

^§

C-1

o

CO

-H

o

o

CD

lO

CO

<

CI

o

Ci

C5

o

crs

C5

C5

d

<

t

-+I

=5

c^

t^

r;^

o

C5

,_!

CO

CO

>-

^

o

o

,— (

o

o

o

•H

p

"—I

■—I

r- 1

=s

.= ^

"?*

d

«

^C5

o

C5

o

(M

o

o

<='C5

o

C5

o

o

o

o

o

•a

S

"S

o

[zi

o

O

00

-f

o

o

CO

o

C5

o

Co

o

o

o

o

o

o

C5

o

I-H

1— 1

1— 1

fa

rH

^

r'-^

l>

CO

CO

o

C5

C5

Ci

CO

C

^

'^l-

l^

t^

t~

l>

t-

I-

l^

t-

■*!

J

_H

C<l

^^

CO

CO

_H

CO

C<l

(M

^

^

■^t^

t-

t-

l^

t-

t^

c-

t>

t-

■^

;. F

s

3

O

•

3

°£2

co

CO

C5

C5

o

©

C5

CO

t^

l-

l-

l^

t^

00

l>

t-

c

o

D

p

t

2

■gj

J

f-

-o

-o

I>

'O

c^

'■O

cc

co

it

^i^

l^

t>

l^

l^

I>

\r~

t^

t^

«

>1

,— ,

1

J

n-f^

o

-^

'O

lO

o

-t<

o

o

•^t-

t^

t^

t^

t>

1>

t>

l-

l^

;-•

If

a

^

«

o

t^

:$

r.C5

o

o

<M

I-H

(M

o

<M

o

^

CO

00

CO

oo

00

00

00

00

3

u

-(

..•

^

■A

^ a

d

3

'fl

o

_o

C^l

CO

-t^

o

o

l^

(JJ

f-H

s

Tapers.] CAPPEE ON SPEED-TKIALS OF BRITISH WAR- VESSELS. 351

Table VIII. — Particulars of Machinery : (I).

H.II.S. " Camperdown."

H.M.S. " Anson.''

Number of cylinders

Diameter of „ < , ^

Length of stroke Number of boilers Diameter of „ Length of ,, Number of furnaces Width of

Length of fire-grate Total area of „ Heating surface in tubes

Total Number of funnels . Dimensions of ,, Height above fire-bara Propellers, Type of .

,, Number of blades

„ Diameter

Pitch.

,, Area of each blade

Cooling surface condensers

Diameter of circulating pumps, four (centrifugal) .

Water discharged per hour by two pumps, each working separately

Do. by all four together

G (2 hp. and 4 Ip.) 6 (2 hp. and 4 Ip.)

52 inches. 74 „

3 feet 9 inches.

12

( 14 ft. 1 in. X 12 ft. \ 4 ins. (oval)

9 feet 10 inches.

36 3 feet 2 inches. 7 „ 3 „ 826-5 square feet. 17,000 20,400

2

52 inches. 74 „

3 feet 9 inches.

16 feet. 10 „ 32

3 ft. 7J in. (mean width)

("16 furnaces 7 feet, and \16 furnaces 6 ft. 9 ins.

797 • 5 square feet. 17,150 20,300

2

10 feet X 6 feet (oval) 9 ft. X 5 ft. 6 ins. (oval)

75 feet

Modified Griffiths'.

4

16 feet.

19 feet 6 inches.

19 '38 square feet.

17,000

4 feet.

75 feet. Modified Griffiths'. 4 15 feet 6 inches. 19 „ 6 „ 18 • 14 square feet. 17,000

3 feet 10 inches.

2,205 tons. 3,465 „

352 CAPPEE ON SPEED-TRIALS OF BRITISH WAR-VESSELS. [Selected

Table IX. — Particitlars of Machinery : (II). Eatios and Coefficients.

	H.M.S. "Camperdown."	H.M.S. '	'Anson."
	Open Stokehold.	Closed Stokehold.	Open Stokehold.
	Closed Stokehold.
Heating surface (tubes)	0	83	0	84
Heating surface (total)
Heating surface (tubes) Fire-grate area	20-5 1	21	5
Heating surface (total Fire-gi-ate area	24	7	1 25-3 1
Propellers: ^r: .... Diameter	1	22	1	26
LHP. per sq. foot fire-grate . .	10-43	14-2	10-43	15-78
„ „ heating surface .	0-42	0-57	0-40	0-62
„ ton displacement ' .	1-04	1-4	1-0	1-52
„ „ machinery . . .] (total) . . ./	6-75	9-2	7-2	10-9
„ „ boilers ....	12-4	16-8	14-1	21-3
Cooling sm-face per I.HP. sq. feet	19	1-44	2-03	1-3
Coal consumed ,, ,, lbs.	211	3-26	2-35	2-23
,, „ ,, sq. foot fire-grate	220	46-3	24-5	35-2
CT . c Starboard . Slip, per cent.; IPort	11 12	12 12	9 12	16 17
Displacement coeflEicient , „p .	207-7	176-0	222-0	172-9
A V^ Midship section „ frr-p-	680-3	576-8	725-4	631-3
( Starboard	708 -Oft.	751 -5 ft.	714-0 ft.	814-5 ft.
Piston speed per minute < (Port . .	715-5 „	764-3 „	740-3 „	814-8 „

' In trim as at trial.

Papers.] CAPPER ON SPEED-TRIALS OF BRITISH WAR-VESSELS. 353 Table X. — Compakative Table of Weights of Engines and Boilers.

Weight of engines

,, of water in condensers .

„ Screw propellers and shafting

„ Boilera

„ Water in boilers ....

i Engine-rooms . Boiler-rooms . Screw-tunnels i Engine-rooms . Boiler-rooms . Screw-tunnels

Total weight .

H.M.S.

" Camperdown."

Tons. cwt.

403 12

24 6

100 13

488 1

174 0

14 9

21 19

5 19

9 19

11 9

21 19

qrs. lbs. 0 8

1,276 12 1 15 1,150 10 2

H.M.S.

"Anson."

Tons. cwt. qrs. lbs.

377 1

22 17

98 446 144

40 1 0 24

21 14 2 13

Total contract weight 1,225 tons.

	H.M.S. "Camperdown,"	H.M.S. "Anson."
	Open-Stoke- Closed-Stoke- hold Trial, i hold Trial.	Open-Stoke- hold Trial.
	Closed-Stoke- hold Trial.
Weight of enginee (complete)! per indicated HP. developed/ Weight of Ixiilers (water, &c.)\ per indicated HP. developed/ Total weight of machinery per \ indicated HP. developed . f	Ll)s. 150-9 213 1 364-0	Lbs. 110-8 156-5 267-3	Lbs. 150-7 158-9 309-6	Lbs. 99-6 105-1 204-7
Indicated HP.developed per ton j weight of boilers, including > uptakes, water, and all fittings )	12-4	16-8	14-1	21-3

Table XI.— Data for Twisting-Moment Diagrams. Closed-Stokehold Trials.

Port or starboard engines .... Half hour during which cards were taken

Boiler-pressure

Revolutions

Vacuum

Mean pressure per sq. inch . hp. cyl.

No. 1 Ip.

No. 2 Ip.

Indicated HP hp.

No. 1 Ip.

No. 2 Ip.

Total collective indicated HP. .

Maximum moment

" Camperdown."

Minimum

Starboard 7th half hour 84 lbs. 103 27 inches 39-0 lbs. 18-2 „ 22-6 „ 1,940-1 „ 1,832-4 „ 2,275-4 „ 6,047-8 „ Forward stroke 1,760 Backward „ 2,010 Forward „ 1,320 Backward „ 1,190

'Anson.

Port

8th half hour

lOOi lbs.

110 26i inches

58 14 15 3,064 1,472 1,694 6,230

•6 lbs.

•7 „

■0 „

•0 „

•4 „

•2 „

•0 ,.

2,980 1,090

[the INST. C.E. VOL. XCV.]

2 A

354 SANDBERG ON THE USE OF HEAVIER RAILS [Selected

(Paper No. 2366.)

On tlie of Use Heavier Eails for Safety and Economy in Eailway Traffic."

By Christkk Peter Sandberg, Assoc. M. Inst. C.E.

(Abstract.)

The following is an abridgment of a communication which should be considered as a sequel to the Pajier written three years ago on " Eail-Joints and Steel Eails." ^

In that Paper the Author gave a design for a 100-lb. flange rail, which has been called the Goliath rail, and recommended its adoption for railways with heavy traffic and high speed. The Author has since devoted considerable time to persuading railway authorities, both in America and on the Continent of Europe, to make a trial of this heavy rail. The Paper now presented gives the result of his labours, and further suggests the adoption of steel base-plates in order to secure a longer duration for the wooden sleeper, and to permit of the use of a harder steel in the rails, with safety against fractures.

The Belgian State Railway was the first to adopt the idea, and after a trial of a lOo-lb. rail for an order in 1886 of about 300 tons, a second trial order of 1,000 tons was given in 1887, and in 1888 it was adopted for the whole of the replacements, about 10,000 tons having been ordered for that purpose. These figures show the satisfaction given by this heavy rail, as regards both safety and economy in the maintenance of the road and rolling- stock, as well as comfort in running at high speeds.

Other countries are working in the same direction ; on some of the French railways the weight of rails has been increased from 60 lbs. to 86 lbs., and on the German railways to 80 lbs. per yard. But considering the high sjDeed run on some of the continental main lines, there is still room for a further increase in weight of rails. In America, the proposal has met with much favour, and the 60-lb. rail which had been generally used there has been supplanted by rails of 70 lbs. and 80 lbs., and even 90 lbs. per yard on some of

' Miimtcs of Pi'0('e(?ding8 lust. C.E., vol. Ix xiv. ]). 3C5.

Papers.] FOR SAFETY AND ECONOMY IN RAILWAY TRAFFIC. 355

the heavier worked lines. But in America, however, the heavy- rails of modern section have not given as good wearing results as the former light rails. The Author would explain this by the fact that the heavy sections are designed with a wide thin rail-flange, which necessitates the employment of softer steel than was the case in the former sections with narrow thick flanges. Eail-makers must guard against fractures of rails, and the harder the rail and the wider the flange, the greater is the risk ; consequently, not- withstanding the desire of the engineer to obtain a harder steel, in order to resist the wear as well as the crushing from the increasing weight of the rolling-stock, manufacturers have been obliged to make the heavy sections softer than the light rails used in former times ; and engineers, so far as experience goes, have not had the economically good results from the heavy rails that they expected. Makers have done all they could to obtain safety with these wide and thin rail-flanges, by curving the rails when hot, so as to avoid cold straightening, which much deteriorates the strength of the rail. It would be well for rail-makers in Europe to adopt the same process when meeting the demand for harder rail steel, in order to secure safety against fractures. In Europe there is a general demand for harder rails. That this hardness must be carefully dealt with, the Author proved by experinients made during the summer and winter of last year upon an extensive scale, at the Domnarfvet Works, in Sweden, where it was desired to increase the proportion of carbon from 0'3 to 0*4 per cent. The rails were cut in halves ; and the one half was tested in the winter at a temperature of — 30^ Centigrade = — 22° Fahrenheit ; the corre- sponding half being tested in the summer at a temperature of 4- 30^ Centigrade = -f- 90° Fahrenheit. The result was, that out of the twenty-one rails tested, nearly all broke in the winter, and all stood in the summer with a specified drop-test of 1 ton falling 15 feet for a 63-lb. rail. It was consequently thought unsafe to increase the proportion of carbon to 0'4 per cent. ; but that it was better to go back to 0 • 3, which had given satisfaction as regards safety ever since the introduction of steel rails, notwithstanding the excessively cold climate of that country.

The rail-section has a good deal to do with the determination of hardness. For instance, a bull-headed rail or a flange rail with a narrow, thick flange, can be made much harder than a rail with a wide thin flange, such as that which has been adoj^ted in America. The danger of copying specifications as to the constituents for different sections of rail is therefore evident. However, as where flange-rail sections have been adopted, harder rails are neces-

2 A 2

356 SANDBERG ON THE USE OF HEAVIER RAILS [Selected

sary, and it is at the same time desirable to have rails with wide flanges to lessen the damage to the sleepers, a problem is presented which does not exist with the rail of bull-headed section fixed to a cast-iron chair, or when metal sleepers are used, namely, how to obtain good wearing-results combined with safety in a rail with a wide flange. The engineer demands a hard and at the same time safe rail, while the manufacturer prefers a rail with thick flange, to enable him to make it of hard steel without risk of fracture. Thus, the wide flange must either be made thick, or the narrow flange must be provided with a steel base-plate fixed on to wooden sleepers. Several railways in Eurojie have already been provided with steel base-jjlates, but these have not yet reached the dimensions of the cast-iron chairs used on the English roads. The Author suggests the adoption of a wider base-plate, the conversion of the light flange rail into a heavier one, by giving greater thickness, both to the head and to the flange, and by using a base-plate at least 10 inches long, as shown by Plate 9. The Swedish 63-lb. section is proposed to be increased to 80 lbs., and 10 lbs. additional weight given to the base-plate ; this is instead of making the flange wider and thinner, as has been the case with the American roads. Thus a hard and safe rail could be obtained that should satisfy both producer and consumer, besides providing a much broader base, which would prolong the duration of the wooden sleej^er to an extent impossible with a flange rail.

The English roads are the best in existence, and a proof of that is given by last year's experience in the run from London to Edin- burgh in eight hours (which means over 50 miles per hour) without accident ; but this should not tempt the managers of continental roads of weaker construction to imitate such sjieed until these have been strengthened. High speed requires heavier rails for safety and economy, both of road and rolling-stock. High speed is dangerous on light rails, producing fractures, widening the gauge, and thereby causing accidents. The rails used on the continental lines are generally too light for the speed actually run upon them. The price of steel rails is now much less than formerly, and as at least 50 per cent, is recouped by the value of the old rails, there is no excuse for delaying the change to heavier rails for all main lines. The cost and qiiality of sleepers should decide what system should be adopted, either steel sleepers, or steel base-plates, or rails with wide flanges laid upon sleepers of hard wood; but thin- flanged rails should be avoided, as incompatible with hardness com- bined with safety.

Papers.] FOR SAFETY AND ECONOMY IN RAILWAY TRAFFIC. 357

In Plate 9 is shown the English permanent way with 80-lb. or 90-lb. bull-headed rails fixed to cast-iron chairs attached to wooden sleepers ; also steel sleepers of different designs ; and Irish, Scotch, and Welsh rails applied to wooden sleepers. The second row of Figs, shows the 100-lb. flange-rail section with angle fishplates on wooden sleepers, according to the Author's design as published in the former Paj^er ; also the 105-lb. Belgian Goliath, as it is called, with but slight alterations from the Author's designs. Dutch, Danish, Swedish, French, German, Eussian, and American sections of rails, of the heaviest weights used in those countries, are also shown ; and, finally, the Author's design for converting a light weak road into a strong road, by increasing the rail head and the thickness of the flange, and by the application of steel base-plates.

Here may be given the approximate cost of the different types of road. Metal sleepers being dearer, but lasting longer, could not be compared with the other systems ; and as fishplates, bolts, spikes, wood screws, or fang-bolts, are used for local reasons, or fcr reasons arising from climatic conditions, their cost may be left out.

English Permanent Way, 1 mile, Single Line.

£. s. £. i-. For SO-lb. bull-headed rails, 1251 tons at £4 per ton . . 50.3 0 3,625 chairs, 40 lbs. each, at £2 10s. per ton . . . 162 0

Total cost jjer miJe . . £665 0

Goliath Flange Eails, 100 lbs. per yard. 156 tons at £4 per ton. Total cost per mile . . . £624 0

Flange Rails, 80 lbs. per yard.

1251 tons at £4 per ton 503 0

2,921 base-plates (two for every sleeper), at 10 lbs. [x-r) r-a ^^ plate, at £6 per ton /

Total £581 12

These calculations show that there is ample room for increasing the weight of the flange-rail system before it comes up to the cost of the biill-headed rail with cast-iron chairs, and that even a 100-lb. rail would be cheajier than an 80-lb. bull-headed rail with chair.

To balance economy with safety and all other conditions, an international discussion, not only between English and continental engineers, but also between them and their American brethren would be highly valuable. The Author having shown his designs,

358

SANDBERG ON THE USE OF HEAVIER RAILS

[Selected

in comparison with others, has no other object in view than an improvement in safety and economy in railway traffic, and hopes that some practical good may result from this communication.

The Paper is accompanied l)y a tracing and sketch, from which Plate 9 and the cut have been engraved.

SUPPLEMENT.

On the 29th of October, 1888, the Author handed the above com- munication to the Secretary of the Institution, and on the same day, almost the same hour, the accident took place to the Czar's train near Borki station, on the Kursk-CharkofiT-Asov Eailway in the south of Eussia. The coincidence is singular, as the main cause of this most disastrous accident lay in the rails being too weak for the heavy Imperial train, run at a comparatively high speed, pre- cisely the danger which the Author warns against in this Paper.

The following Pig. shows the effect of the accident in displacing the carriasces : —

a O

ACCIDENT TO THE CzAR's TRAIN AT BORKI.

Here Nos. 1 to 4 represent two heavy engines and tenders ; No. 5, the luggage van ; No. 6, brake-van ; No. 7, the minister's carriage (the first to run off) ; No. 8, the cook's carriage ; No. 9, the kitchen; No. 10, the servants' carriage (where most were killed); No. 11, the dining-saloon (in which the Czar and the Court were

Papers.] FOR SAFETY AND ECONOMY IN RAILWAY TRAFFIC. 359

travelling at the time of the accident); No. 12, the Imperial children's carriage; No. 13, the Czar's sleeping carriage; No. 14, the Czarevitch's carriage, followed by others.

The lesson, technically speaking, to be drawn from this ought to be invaluable, not only to Eussia, but also to other countries on the Continent of Europe, namely, not to attemjjt to exceed the speed which the condition of the road will permit with perfect safety. Taking into consideration how slowly a change can be brought about in strengthening the road, a commencement should be made without delay on all main lines where it is desired to increase the speed, and to use heavier rolling-stock than was originally contemplated.

300 CHARLES SHAW-LEFEVEE. [(Obituary.

OBITUARY

The Right Hon. CHARLES SHAW-LEFEVRE, Viscount EVERSLEY, was Lorn in London, on the 22nd of February, 1794. He was descended from one Pierre Lefevre, a French Huguenot who settled in England after the Revocation of the Edict of Nantes, and whose descendants became the possessors of estates and dignities sufficient to secure their enrolment among the untitled aristocracy of England.

The larger part of Lord Eversley's celebrity having been achieved in connection with a political career, abundantly recorded elsewhere, it is not necessary in these pages to do more than give an outline of a life remarkable alike for its success and for its long continuance. He was the eldest son of Mr. Charles Shaw, who subsequently took the additional surname of Lefevre, himself a considerable figure in the politics of a century ago, and whose imposing and somewhat pompous manner provoked the remark of Canning : " There are only two great men in the world — Shah Abbas and Shaw-Lefevre." The younger Shaw-Lefevre, on leaving Cambridge, commenced life as a barrister of Lincoln's Inn; but while pursuing his studies with all diligence and success, he proved himself so keen a sportsman that his serious mother, slow to discern the greater qualities of her firstborn son, sorrowfully observed: " As for Charles, he is only fit to be a gamekeeper." But whatever her estimate of his intellect, she had reason to be proud of his person, for he grew to be one of the tallest and handsomest men of his generation. On the death of his father in 1823, he settled in Hampshire on the family estate, and soon acquired an important position in the county, as an able magistrate, an influential member of the Court of Quarter Sessions, and a zealous officer of Yeomanry. In 1830 he entered the unreformed Parliament as Member for Downton, in Wiltshire, which, though one of the oldest boroughs in the United Kingdom, was, two years later, disfranchised by the passing of the Great Reform Bill. He was subsequently elected one of the representatives of North Hants, and retained his seat until he was made a Viscount in 1857. In less than ten years from his first entry into the House of Commons he was elected Speaker, and, when after eighteen^years of service he retired with the usual

Obituary.J CHARLES SHAW-LEFEVRE. 361

title and pension, he left behind him the reputation of having been the best Speaker that ever ruled debate. On resigning his dignified office, he was raised to the peerage as Viscount Eversley of Heck- field, in the county of Southampton, and there, on the 28th of December, 1888, he died at the patriarchal age of ninety-four. The thirty-two years of his life after retiring from the Speakership were mostly devoted to the avocations of a country gentleman ; for though he was pretty constant m his attendance at the House of Lords, he took no great part in its affairs, the Upper Chamber being too cold and listless for one who had spent his best days in the exciting atmosphere of the popular assembly. In his own county Lord Eversley was an active and zealous public man, serving at different times as Chairman of Quarter Sessions, High Steward of Winchester, Hon. Lieut-Col. of the Hants Yeomanry, and Aide-de- Camp to the Queen ; Governer and Captain-General of the Isle of Wight, and Church-Estate Commissioner, which post he resigned on his acceptance of that of Ecclesiastical Commissioner. He was also a Trustee of the British Museum, and in 1885 was made a G.C.B.

As a relief to his public duties he devoted himself to practical farming and gardening, and it was in connection with the former pursuit that he was, on the 20th of June 18-12, elected an Honorary Member of this Institution; his qualification setting forth the " encouragement he has given to the application of mechanics to Agriculture, and his general attachment to Scientific pursuits."

Lord Eversley, alike by the gifts of fortune and b}' their limita- tions, was the type of what a modern English gentleman should be ; and, in the opinion of those who knew him best, he fulfilled that ideal in every relation of life.

JOHN BROWN died on the 24th of August, 1888, in his sixty-fifth year, having been born at Stafford in 1823. He began his technical education as a clerk in the office of Mr. John T. Woodhouse of Ashby-de-la-Zouche, and became an articled pupil in this office in 1848, subsequently acting as Mr. Woodhouse's principal assistant. After serving in this capacity for a few years, he commenced business on his own account in Barnsley about 1854. He had been engaged along with Mr. Howel in a difficult investigation regarding Lord Granville's collieries in Staffordshire, and his extensive information and accurate and cautious reports soon gained for him the con- fidence of a large clientele. For about ten years he resided at

362 JOHN BKOWX. [Obituary.

Harbro' House, Barnsley, and while there was entrusted with many important arbitration cases. He was appointed Engineer to the Lundh.ill Colliery Company after the occurrence of a disastrous explosion in that pit in 1857, when the bodies of eighty victims of the explosion were recovered under his superintendence. He here introduced the " Dumb Drift," a device whereby the gases coming with the ventilation-current from the workings were delivered into the upcast shaft at a higher level than, and without coming in contact with, the furnace. Xo"n' that furnace-ventilation has almost ceased to be used, this improvement may have lost interest ; but it was a novelty at the time in this district, and in spite of its evident advantages, Mr. Brown incurred much hostile criticism in introducing it. From the management of this pit he retired, assured of the respect and good wishes of the officials and workmen of the place, which were testified to by the presentation of a handsome silver-gilt cup. Mr. Brown and Mr. T. W. Jetfcock now became partners, the firm having offices both in Barnsley and Sheffield. While still at Barnsley an exciting episode of Mr. Brown's life was the explosion at Oaks Colliery in 1866. He had previously been the Viewer of this colliery, and this knowledge of it enabled him to recover the bodies of many of those killed on this occasion.

Mr. Brown's consulting practice was not confined to this country. While at Barnsley he was called upon to report on mining pro- jects in Denmark, and on others in Eussia, the latter in 1868, and again in 1874 on other jirojects in Portugal. In 1869 he became Engineer to the Cannock Chase Colliery Company, having been at an earlier date the late Marquis of Anglesea's adviser in the enterprise that has led to the development of this important coal-field. As early as 1852 he had made the plans from which No. 2 pit of this field was constructed. It was called the " Fly " on account of the high speed of winding adopted, and it is still in work. In the Hednesford pit Mr. Brown designed the apparatus for closing the mouth of the upcast shaft when winding ; and throughout the colliery he made gTeat use of compressed-air engines for underground haulage, and winding from a lower seam to the foot of the main winding-shaft. His stay here, however, was also short, and the same sort of difficulties as those at Lundhill led to his retiring from his post in 1873. He now removed to Hednesford, and was appointed Engineer to the South Staffordshire Mines Drainage Commission. He had also a large private business, and from 1874 to 1879 acted as Consulting Engineer to the Mid- Cannock Colliery Company.- In 1880 he shifted his head-quarters

Obituary.] JOHN BROWN. 3G3

to Birmingham, where he continued and extended his consulting and arbitration business. He worked with Sir Eobert Clifton in developing the Clifton Colliery near Nottingham, and acted as Consulting Engineer also for the Duke of Portland, Lord Wharn- cliffe and others. One of his important arbitrations was the North Staffordshire Coal and Ironstone Trade Arbitration in 1875. Mr. Brown was elected a Member of The Institution of Civil Engineers in 1858. He was also a Fellow of the Geological Society, and a Member of the Iron and Steel Institute ; the North of England Institute of Mining Engineers ; the Cleveland Institute of Mining Engineers ; the Chesterfield and Derbyshire Institute of Mining Engineers ; the Midland Institute of Mining Engineers ; and the South Staffordshire and North Staffordshire Mining Institute, of which last he was for a period the President. On the institu- tion of the Professorship of Mining in the Mason Science College, Birmingham, he was appointed to fill the chair, which he held for one year.

He had a very extensive knowledge of the mining characteristics and conditions of the Midland district. This, combined with a retentive memory, a strictly methodical system of working, a markedly judicial habit of mind in criticising the cases laid before him, and a frank and kindly manner, led to his being widely trusted in arbitration cases, and to his advice in private enterprises being highly valued. Unfortunately a certain quick- ness of temper led to his frequent premature abandonment of posts where his more permanent service would have been desirable. His pupils describe him as a kind and considerate master, and his generosity towards his personal friends, not only in money matters, but also in imi^arting professional information, was lavish to a high decree.

WILLIAM ARMITAGE BROWN was born at Plymouth on the 8th of August, 1830. His father was a surgeon, and young Brown studied for a few years with the intention of following the same ])rofession, passing some examination for the pxirjiose. He, however, abandoned the idea, and became pupil and apprentice to Mr. J. E. Hodgkin, of Suffolk Works, Birmingham, in January 1855. In 1858, Mr. Hodgkin retired from business, and Messrs. Charles and Walter May, and George Mountain became his successors, with whom Mr. Brown went through the ordinary routine of the shops and drawing-ofiice. After this he spent some time in the London

364 WILLIAM AEMITAGE BROWN. [Obituary

office as drauglitsman. Thence, in 1860, he went to Pernambuco, Brazil, for Mr. (now Sir) Charles Hutton Gregory, and was engaged on bridgework for the Eecife and Sao Francisco Eailway Company. Mr. Brown spent nearly two years in Brazil, and on his return to England, he in 1862 entered the service of the London, Chatham and Dover Eailway under Mr. William Mills, the Engineer to the Company. Mr. Brown served first as draughtsman, and subse- quently as Chief Office Assistant. He continued in that position until the early part of 1879, when he was appointed Inspecting Engineer for the permanent-way materials for the same line, which post he held until his death. In 1863 Mr. Brown took out a patent for a train-speed indicator, and in the same year he, conjointly with the late Mr. Eichard Lionel Jones, patented a single-handed signal-lamp. He was engaged on a patent for steel sleepers at the time of his death, which occurred on the 28th of August, 1888. Mr. Brown was elected an Associate of the Institution on the 3rd of February, 187-i, and was transferred to Member on the loth of December, 1886.

HENEY CAEE was born in Derby, on the 24th of November, 1817. On the death of his father, when he was only six weeks old, his mother with her two children, the subject of this memoir and his sister, went to Holbrook, near Derby, to reside with her father- in-law, the Eev. John Carr, and here Henry Carr remained sixteen years, until the death of his grandfather, when the family removed to Duffield. The intention was to educate him for Holy Orders. He was sent to Oakham Grammar School, as well as several other schools. He, however, evinced no inclination or taste for classical studies, but took the greatest interest and delight in carpentry, and indeed in any kind of mechanical work. Before he was nine years old he had saved up his money to buy a lathe, and all his spare time was spent at his lathe or carpenter's bench. It was then that the wise decision was come to of allowing him to choose a profession in accordance with his own inclinations and taste, and in January 1837 he was articled as a civil engineer to Mr. (after- wards Sir William) Cubitt, and started work in Mr. Cubitt's London office.

Mr. Carr's first expedition was with Mr. Cubitt to the North of Ireland, a trip he always looked back upon with the greatest pleasure as having afforded him the advantage of close contact with Mr. Cubitt during the time the latter was examining various

Obituary.] HENRY CARR. 365

engineering works and projects ; a note-lDook, still preserved, evinces the interest Mr. Carr took in all these engineering works. He then went to Tunbridge for three months, at the time the South Eastern Railway was under construction, and had his first railway experi - ence ; but so little work was doing there that Mr. Culiitt placed him under Mr. John Wright, then Resident Engineer on the line from Folkestone to Dover, and he had almost the entire supervision during the construction of the Shakespear Tunnel. While at Folkestone he devised an instrument for setting out the slopes of cuttings. He was about eighteen months at Dover, when a very severe attack of typhoid fever necessitated his return to Derby- shire, and left him an invalid during the remainder of his ])upilage. In the spring of 1841, three months after his articles had expired, his health was so far recovered that Mr. Cubitt recommended him to Mr. George Turnbull, then Engineer in charge of the important works at Middlesbrough Docks. The following extract from a letter to Mr. Turnbull shows the high opinion Mr. Cubitt had formed of him : — " He is highly talented, and if his health gets re-established will make a very good engineer, his inventive talent being great and his judgment good." He remained at the Middlesl)rough Docks about two years, and while there made a survey of the town and docks for the late Mr. Pease, for which he received his first pay.

His next work was building a weir across the River Trent, near Trent Junction, of which he had sole charge under the late Mr. Peter Barlow, during the spring and summer of 1843. He was at Folkestone for a short time the year following on harbour work, and afterwards on the London and Croydon Railway, then under construction as an atmospheric railway. He was next apjiointed Resident Engineer on the extension to Epsom, and as all road-crossings had to be by over- or under-bridges, he had great scope for his love of bridge-designing in the thirty-two bridges be- tween Croydon and Epsom. Before this extension was completed, the difficulties experienced in working the atmospheric system from Croydon towards London decided the Directors to abandon it, and the Croydon and Epsom line was laid for the ordinary locomotive. While at Epsom Mr. Carr preioared plans, working-drawings, &c., for the bridges on the proposed extension to Dorking.

On the completion of the work at Epsom, Mr. Carr removed to London to prepare drawings for the Great Northern Railway, and to Doncaster in September 1847, on being appointed Resident Engineer on the section of that railway between Askern and Retford. While at Doncaster he designed and superintended

366 HENKY CAER. [Obituary.

the Imilding of t-svo bridges at Sprotbro', over the Eiver Don and canal, for the late Sir Joseph Copley, Bart. On their completion he furnished a Paper on the work to The Institution of Civil Engineers, for which he received a Council premium. In September 1850 Mr. Carr went to Eetford as Resident Engineer on the 18 miles from Eetford to Xewark, which included two bridges over the Eiver Trent and the Newark Dyke. He also superintended the building of the extensive machine- and locomotive-shops at Don- caster. In August 1853 Mr. Carr moved to Peterborough, when for some time he had charge, as Resident Engineer, of the whole of the Great Northern main line from London to Doncaster. About this time he designed and built a bridge across the Eiver Trent at Kelham for Mr. Manners-Sutton. Previous to going to Peter- borough Mr. Carr had designed and patented a crossing to resist the crushing of the rail-flange in the ordinary crossing. This brought him in contact with Mr. Wild, of points and crossing fame ; and in 1855 he resigned his position on the Great Northern Eailway, and took offices in Victoria Street, Westminster, com- mencing practice on his own account. During the first two years he was engaged on Mutford Sluice, near Lowestoft, and on several small engineering works, and from this time he also took great interest in architecture, preparing plans and specifications for several country residences, two being parsonages at Hartington and near Carlisle, as well as making additions and alterations to a number of large houses.

He had also business relations with Mr. Barrow, of the Staveley Ironworks, and in 1861 he commenced preparing the elaborate plans for the large engine-erecting and repair ing-shops at Long- hedge, Battersea, for the London, Chatham and Dover Eailwa}'', which occupied nearly three years in execution. At this time Mr. Carr, in conjunction with the late Mr. Joseph Cubitt, designed the new bridge over the Thames at Blackfriars, and although his name did not appear with Mr. Cubitt's as Joint Engineer, it is well known that the drawings were prepared in ]\Ir. Carr's office under his immediate superintendence, and that all the calculations and a great deal of the drawing were done by Mr. Carr himself. The temporary timber Ijridge, which carried the traffic during the building of the new bridge, and which was very favourably noticed at the time, was entirely the work of Mr. Carr. Blackfriars Bridge was in hand for the seven years 1862-69, and during this time Mr. Carr, under Mr. Cubitt's direc- tions, prepared the drawings for the bridge which carries the London, Chatham and Dover Eailwa}' over the Eiver Thames,

Obituary.] HENRY CARR. 367

niakinp; all the calculations of strains and scantlins^s himself. In 1869 Mr. Carr reported to the Bridge House Estates Com- mittee of the City of London on the Metropolitan connnunications between the north and south sides of the Eiver Thames. He prepared schemes for improving the gradients of Southwark Bridge (it was mainly by his eiforts the bridge was eventiially thrown open to the pul)lic), and also for widening the footpaths of London Bridge. In 1872 Mr. Carr read a Paper at the Eoyal Institute of British Architects on " The Bridges of London."

Not having good health, Mr. Carr was obliged in later years to decline undertaking work which would involve nu;ch exposure. Having no family, and being possessed of some private means, he was able to a certain extent to choose his work. His great forte was designing, and the rapidity with which he worked was astonishing to those with whom he came in contact. It was 'his habit to get all his drawings out in pencil, passing them over to others to " ink in," colour and finish, and it was wonderful to note the accuracy and care with which they were drawn ; there was seldom a line to alter, and they were neatness itself. It has been said of some engineers that they were made by their staff; of Mr. Carr it may be said that he made his staff.

Soon after the completion of Blackfriars Bridge Mr. Carr retired from the active duties of his profession, but he was never idle ; his carpenter's shop and lathe kept him much occupied, as many of his friends can testify by the number of useful as well as orna- mental things he made and presented to them. Occasional tables for the drawing-room, in walnut and oak, he was fond of making, and the work and finish was so good that no cabinet-maker need have been ashamed of them. He also interested himself in several other matters. Amongst the first was the subject of canvassing for admission to charitable institutions ; some specially distressing cases that came under his notice first drew his attention to the subject. As a result of his efforts the Charity Voting Reform Association was formed under the presidency of the Duke of Northumberland, with many influential noblemen, clergymen and others as vice-presidents, the Eev. E. J. Simpson of St. Clement's Danes, and Mr. Carr being honorary secretaries.

During the last ten years of his life Mr. Carr took up the sul)ject of the use of arsenic in domestic fabrics, more especially in wall- papers. He wrote a pamphlet, " Our Domestic Poisons," which had a very wide circulation ; also a Paper read before the Society of Arts, and he delivered a lecture on the subject at the Health Exhibition in 188-1. He was supported by man}" of the leading

368 HENRY CARR. i[Obituary.

medical men in London, some amongst them having themselves and their families suffered from the effect of arsenic in wall-papers, &c. He tested hundreds of papers for manufacturers and private indivi- duals, using both the Eeinsch and the Marsh processes, and largely in consequence of his efforts the use of arsenic has been to a great extent abandoned. Mr. Carr, as an engineer, had great courage in the pursuit of his profession, amounting at times to what would almost appear rashness. He was rapid in judgment, and accepted responsibility without hesitation, not only for his own work, but for the actions of his trusted assistants, to whom he was very loyal when satisfied they were doing their best, never forgetting any one who had worked faithfully for him, and it always gave him great pleasure to be of any help or assistance to his old hands by advice, encouragement, or even in a pecuniary way. In private life Mr. Carr was known for his scrupulously high tone of honour, com- bined with extreme kindness and sympathy with trouble and mis- fortune ; no one with any claim upon him ever sought his assistance in vain, and many without any claim have benefited largely by his generosity and goodness, and now mourn the loss of a true friend and liberal benefactor.

Mr. Carr's name was borne on the roll of this Institution for more than half a century. He was elected a Graduate on the 6th of February, 1838, transferred to Associate on the loth of June, 1847, and to full Member on the 2nd of March, 1852. He died on the 21st of August, 1888. He was assiduous in his attendance at the meetings, and took great interest in the proceedings, as will be seen by the following list of Papers, &c., contributed by him to the Minutes of Proceedings.

*' Description of an Instrument for setting out the width of Cuttings and

Embankments of Railways, Canals, or Eoads, as particularly applicable to

falling or side-lying ground." Vol. i., 1839, p. 52. *' Description of a Dynamometer, or an Instrument for Measuring the Friction

on Eoads, Railways, Canals, etc." Vol. i., 1840, p. 52. *' Description of an Instrument for describing the Profile of Eoads." Vol. i.,

1810, p. 56. •" Description of two Bridges over the Eiver Don and the Canal, with the Lodge

and Approaches, on the Estate of Sir Josejjh Copley, Bart., at Sprotbro',

near Doncaster." Vol. x., pp. 302-306.

A model of Carr's Eailway-Crossing was exhibited after the meeting of the 2ud of May, 1854.

Mr. Carr also freqiiently took part in discussions, especially on questions referring to railways, in which branch of the profession lie was an expert.

Obituary] HENRY CARR. 869

The following works by him are in the Library of the Institution : —

Cubitt, J., and Carr, H. — " Blackfriars Bridge. Contract for removal of existing bridge and for building the projxjsed new bridge, and other work connected therewith." Folio. Atlas of drawings. ISQi. (Reprinted in 1869.)

" Metropolitan Street TrafBc. Suggested improvements." Tract 8vo. Vol. 234. Plate. London, 1871.

" The Bridges of Loudon." Tract 4to. Vol.68. Plates. London, 1871. Ex- cerpt Ti-ansactions of the Royal Institute of British Architects.

" Our Domestic Poisons ; or, the Poisonous Effects of certain Dyes and Colours used in Domestic Fabrics." 2nd and 3rd editions. Paper read before the Society of Arts. 8vo. London, 1879-80-83.

ROBERT DENNY, the fifth son of the late Rev. Henry Denny, Rector of Churchill, Diocese of Ardport, of Churchill, in the County of Kerry, was born on the 5th of June, 1843. He was educated at Beaumont College, co. Cork, and in 1862 was articled for three years to Mr. W. R. Le Fanu, by whom he was employed on part of the Nenagh extension of the Great Southern and Western Railway, as assistant in charge of the works, and after- wards on the same line, under Messrs. Cotton and Flemyng, when Mr. Le Fanu had retired from the engineering profession. From these gentlemen he received the most flattering testimonials. He subsequently worked with Mr. J. C. Biirke, of Westminster, and in the year 1870 started engineering business independently in Tralee. Thenceforward, until a short time previous to his death, he was associated in all the principal works carried out in Kerry, his native county. These works were of a varied nature, and comprised : The Castleisland Railway (a branch of the Great Southern and Western) ; the Tralee Waterworks ; the Limerick and Kerry Railway (in which work he was associated with Mr. W. Barrington ; the Fenit Railway, and the Tralee and Fenit Pier and Harbour works, of which Mr. Benjamin Baker was Engineer-in- Chief. The last work which Mr. Denny carried out was a scheme for providing a water-supply to the town of Killarney. All these works were successfully completed, and their success was in a great measure due to the local supervision exercised by Mr. Denny.

His genial and sociable manner caused him to be a universal favourite, while his attention to business and his integrity gained him the respect and esteem of his professional brethren. He died on the 25th of July, 1888, regretted by all who knew him.

Mr. Denny was elected a Member of the Institution on the 4th of December, 1883.

[the INST. C.K. VOL. XCV.] 2 B

370 JAMES EASTOX. [OLituary.

JAMES EASTOX, bom in Stamford Street, Southwark, on the 1-ith of September, 1830, was the second son of James Easton, the founder of the firm, of which the subject of this notice became subsequently the head. Mr. Easton's education commenced at a private school in Eamsgate, and was completed at King's College School, London. In his father's Avorks he obtained a desultory engineering training, until called uj^on, in 1851, to assume the management of the Wandsworth Paper Mills, now in the occupa- tion of Mr. McMurray. The business was not found to answer; he therefore joined his maternal grandfather, Mr. Shaw, in a large stationer's business, which comprised the manufacture of cheques and bank-notes; but resigned the occupation in 1856, in order to enter his father's ofBce in the room of his elder brother, who had become incapacitated from work by a severe accident.

The firm of Easton and Amos, which Mr. Easton joined in 1858, was at that time largely engaged in the manufacture of centrifugal pumps, on the system introduced by Mr, Appold, and Mr. Easton devoted himself very much to the business, especially in the larger sizes used for drainage purposes, his strong common-sense and large views rendering him a very eificient adviser to the various Drainage Boards in Somersetshire, and in the Fen counties. The drainage of Whittlesea Mere by a centrifugal pump in 1852 — a bold and thoroughly successful application of a comparatively new system — was followed by many equally important enterprises both at home and abroad, among which were the reclamation works at Wexford, the great pumping establishments erected, under Sir John Hawkshaw, for the Witham drainage at Lade Bank, near Boston, and on the Amsterdam Canal.

The experience which he had acquired as a paper-maker proved very useful in extending that branch of the business. His firm, in 1861, erected the Ettric Forest Paper-Mills on Dartford Creek, since become the property of the Daily Telegraph, the Stowmarket Paper-Mills, in 1865, and large mills in Egjq^t, Portugal, and Japan, besides imj^roving and adding to many existing establishments.

The Franco-German war in 1870-1 prevented Messrs. Call, of Paris, from completing some large contracts for cane-sugar factories which they had entered into with Ismail Pasha, the then Viceroy of Egypt ; that astute riiler, who had made Mr. Easton's acquain- tance during the construction of the Boulac Paper-Mill and the large pumping establishment of Gizeh, and who had already entrusted one large sugar-mill to him, judged rightly that he would be best qualified to complete, at short notice, the work

Obituary.] JAMES EASTON. 371

which the French firm was unable to carry out, and so prevent a ruinous loss of cane. Two more mills, each, like the first, capable of making 120 tons of white sugar per day, and having the largest cane-crushing rolls in the world, were ordered, and two out of the three were finished in the sti2:)ulated time, thus saving the crop which had actually been planted before the machinery was ordered. The third factory, though delivered in good time, was never com- pletely erected.

In the waterworks branch of the business, Mr. Easton's talents were chiefly devoted to the financial arrangements and negotiations necessary to carry out great undertakings abroad. He conceived the idea, in the case of such towns as Antwerp and Seville, of completing the works from his own resources and from those of many friends, whose confidence he enjoyed, and of disposing of the finished undertakings in complete working order to properly con- stituted companies. The case of Antwerp was surrounded with extraordinary difiiculties ; the concession had been revoked and renewed more than once, and in addition there were several pre- tenders to a share in the prospective benefits to arise from the enterprise. The difficulty of disentangling the complicated busi- ness was very great, and this was intensified by the objectionable nature of the only available source of supply, which, although it would have yielded a sufficiently good water for municipal pur- poses, could never have commanded a ready sale for domestic use ; he had, therefore, to take the bold step of introducing the iron process of purification on a far larger scale than had ever been attempted before, and did so with complete success. The case of Seville was less difficult ; but much skill and patience were needed to complete the negotiations with the authorities and with the Government.

Mr. Easton's health, never very robust, began to fail sensibly in the early part of 1888 ; his strength gave way gradually, and he died on the 28th of August in that year at Westgate, in Kent, and was buried in the old churchyard of St. Peter's, where many of his wife's relations lie at rest.

He married in 1855 Miss Annie Devonshire Sackette, daughter of Mr. S. Sackette, Deputy-Mayor of Eamsgate. The four daughters, the fruit of this marriage, all survive him.

JOHN FOWLEE was born in the year 1824 at Rubislaw, near Aberdeen. He was educated at the Aberdeen Grammar School, which he left at the age of fifteen- He does not at this time

2 B 2

372 JOHN FOWLER. [Obituary.

appear to have had anj' clear idea what vocation he should follow, and has left no information as to his tastes or inclinations beyond the fact of his great love of Botany. His family at this day hold a rather valuable collection of specimens gathered on the Aberdeen- shire hills ; and in his later life, his great interest in this science was evinced by the tenderness with which he would pluck flowers, fruit or ferns, and the care with which he examined the same. Probably influenced by his love of this subject, after leaving school, he was appointed as a gardener to Sir Thomas Gladstone, of Forgue, in Forfarshire ; but, after a month or two, he went to London for the purpose of getting employed on some of the great engineering works of the Metropolis. He did not stay there long, as at the age of eighteen he visited his half brother, Mr. James Johnson, then Superintendent to the Tees Navigation Company, Stockton, who persuaded him to stay in that town as his assistant. On the death of Mr. Johnson, Mr. Fowler succeeded to the position of Superin- dent; and in the year 1854, on the transference of the interests of the Navigation Company — a private company — to the Corporation of the Tees Conservancy Commission, he was appointed Chief Engineer. In this year a start was made with the series of works, which have converted the Eiver Tees from a shallow unimportant stream into a river easy of navigation, and capable of accommo- dating vessels of over 4,000 tons burthen ; and which reflects credit on Mr, Fowler as furnishing one of the most striking instances of river improvement, both with respect to economy and the per- manent benefits secured. The varioiis improvements designed and carried out by him have been described in Papers entitled " Dredgers, and Dredging in the Tees," and " Eiver Tees Improve- ments," contributed to the Institution.^ It may not, however, be out of place to say that, thirty-five years ago, the channel from Middlesbro' to the bar — a distance of 7 miles — was beset with shift- ing sand-banks and shoals, and sj^lit up, sometimes into four, and always into two. In the 3 miles of river from the bar upwards, forty-two buoys, besides beacons, were needed to mark out the channel, and required changing, more or less, every spring-tide. The available depth at low-water was 1 foot 10 inches in two places; and the channel crooked. Ships that would now be considered very small ran many times into the sand-banks, before getting through. During the thirty-three years that have elapsed since the commence- ment of the works, the channel has been gradually improved, so that, at the present time, there is an available depth of 15 feet at

• Minutes of Proceedings Inst. C.E., vol. Ixxv. p. 239 ; ibid., vol. xc. p. 344.

Obituary.] JOHN FOWLER. 373

low-water up to Middlesbro', a width of channel at that place of 170 yards, increasing to 350 yards at the fifth buoy-light. Ships of 3,000 tons are built and engined at Stockton, and vessels carry- ing 1,700 tons trade there. At Middlesbro' vessels carrying over 4,000 tons, and drawing 24 feet of water, are not unusual. The estuary has also been made into a safe harbour of refuge, and a gi'aving-dock, capable of accommodating the largest vessels, has been constructed by the Commissioners. To accomplish this, more than 20 miles of low- water training- walls, and 15 miles of high- water walls have been constructed; 21,000,000 tons of material have been dredged, and 120,000 cubic yards of rock removed. The breakwater on the south side was completed about two years ago. The northern one is in course of construction. On the 25th of October, 1888, the First Lord of the Treasury, Mr. W. H. Smith, M.P., formally opened the south breakwater in the presence of a large and influential company, and dedicated it as a " national work for ever." But Mr. FoAvler, the completion of whose hard and honest labour of thirty-five years was thus celebrated, was not spared to see the consummation of his task ; for a fortnight earlier, on the 11th of October, he was called to his rest.

In his position as Engineer to the Commission, Mr. Fowler had not opportunity of getting information and experience in any but his own branch of his profession, nor did he attempt or wish to do so. His whole professional life was devoted to the action of waves, tides, currents, and other matters contingent on his duties as a harbour-engineer. With application, he made himself so conver- sant with this subject that the value of his opinion was in a short time recognized. In few instances of improvements to our inland navigations was he not consulted ; his knowledge in these matters being so complete that he was in great demand as a parliamentary witness, being retained on every scheme of importance connected with waterways brought before the Parliamentary Committees. His evidence was always tendered in a straightforward, ready way, giving the impression of truth and conviction, and without equivocation in cross-examination.

He was a leading witness in all the parliamentary fights in cod- nection with the Manchester Ship Canal, and among the many rivers respecting which he was consulted were the Humber, Ouse, Trent, Mersey, Eibble, Eden, Clyde and Carron, and in Ireland, the Forth, Liffey, Boyne and Shannon.

In 1868, he was offered the appointment of Engineer to the Clyde Trustees, and it speaks greatly to his credit that his employers were unanimoiis in doing all in their power to induce

374 JOHN FOWLER. [ObituaiT.

him to remain with them, which, after some consideration, he decided to do, although at some pecuniary sacrifice. In 1885, he was consulted in reference to the improvements for the Port of Havre, but his engagements prevented him from entertaining the flattering proposal then made to him. He was also consulted some years ago on the harbour of Pobena, in Spain, At the time of his death, he was Engineer to the Trustees of the Ouse Navigation, who are carrjT.ng out improvements between York and Goole. The last work completed by him was the Naburn New Lock, 6 miles below York, opened by Prince Albert Victor on the 27th of July, 1888. He was also Engineer to the Pier and Harbour Commissioners of Whitby.

There are men who make a favourable impression at first, but are found, on more intimate acquaintance, to fall short of the estimate which has been formed concerning them. It was not so with Mr. Fowler. A natural reser\'e made it somewhat difficult to know him intimately ; but those who were privileged to get within the fence of this reserve, discovered the sterling excel- lences of his character, the deep true worth of the man. To knoAv him was to respect him ; to know him well was to love him much. He was an upright, honest and unselfish man, leal-hearted and true, in whose death all who knew him have to mourn the loss of a faithful friend ; those associated with him in work of any kind, that of a trusted colleague and ever ready helper ; while those more closely related to him by ties of blood, sorrow most of all for the removal of a kind and loving husband and father.

Mr. Eowler was elected a Member of the Institiition on the 4th of February, 1873.

WILLIAM FRANCIS was the eldest son of William Francis of Whitehall, Kenwin, Cornwall, and was born on the 2oth of September, 1831. He was educated at Plymouth Grammar School, and was afterwards articled to Mr. Nicholas Wliitley, Engineer and Surveyor of Truro.

In 1850-6, Mr. Francis was engaged on the Cornwall Eailway, first under Mr. Glenny in the preparation of the working plans, &c., subsequently acting as Assistant Engineer when the works were commenced, and latterly as Resident in charge of about 10 miles of that line, between Grampound Road and Liskeard. In 1856-7, he was employed as general Assistant in the Engineers' Office of the Great Eastern Railway, and subsequently became Resident Engineer on tlie Ipswich and Woudliridge Raihvay, under Mr.

Obituary.] WILLIAM FRANCIS. 375

Peter Bruff, the Chief Engineer. In 1800, Mr. Francis re- moved to Hamjishire, and under Messrs. Collister and Galbraith, superintended (1860-5) as Eesident Engineer, the construction of the Andover and Eedbridge Railway, now the property of the London and South Western Eailway Company. He was subse- quently engaged in the preparation of the Parliamentary surveys of the Northampton and Banbury Railways, and after the Act was obtained, took charge of a section of the works until the same were suspended during the financial crisis in 1866. Mr. Francis afterwards entered into business on his own account, and was extensively employed in arbitration work, the preparation of parliamentary plans, &c. During the latter portion of his life, Mr. Francis turned his attention to contracting, and super- intended the construction of the Ascot and Aldershot branch of the London and South Western Railway, for Mr. James Taylor. He also superintended for Messrs. Curry, Reeve and Co., the execution of the railway from Corfe to Swanage, now forming part of the London and South Western, and the Gravesend Extension of the London, Chatham and Dover Railway, for Mr. (now Sir George) Barclay Bruce. Mr. Francis was likewise interested in the con- tract for the first section of the North Cornwall Railway, from Halwill to Launceston. He was thoroughly conversant with tlie details of railway construction, which he had mastered both tlieo- retically and practically ; and from his genial disposition and good business habits, he was respected and esteemed by all who knew him.

For the last twelve or fifteen months Mr. Francis had been in failing health, and he died at Ealing on the 8th of November, at the age of fifty- seven years. He was elected an Associate of the Institution on the 5th of February 1867, and was transferred to Member on the 18th of February 1873.

FRANK ALEXANDER BROWN GENESTE, of Huguenot descent, was a son of the late Rev. Maximilian Geneste, Incumbent of Holy Trinity Church, West Cowes, Isle of Wight, and was born on the 28th of July, 1842.

Frank Geneste was educated at Winchester College, and was placed as a pupil with Mr. Habershon, architect, with whom he served a two years' pupilage, commencing in 1860. He was sub- sequently with Mr. H. C. Wilson, Civil Engineer, as jjupil, on the works of the Newport-Pagnell Railway. After this, he was em-

376 FRANK ALEXANDER BROWN GENESTE. [Obituary.

ployed on the following works : For nearly five years in India, on the construction of the Delhi Railway, for Messrs. Brassey, Wythes and Henfrey, the contractors for that line. He then went out as District Engineer for Messrs. Waring, Brothers, contractors, on the East Hungarian Railway, and was next engaged on railway surveys in the United States of Colombia for the Public Works Construction Company, first as Chief Assistant, and then as Chief Engineer.

The Government of Venezuela, requiring a survey to be made for a railway from La Guaira to Caracas, Mr. Geneste was appointed as the Chief Engineer. This survey was conducted by him, and the line has been constrvicted. He again was called upon to go to the United States of Colombia, this time for the Government of that country, as Chief Assistant Engineer. On returning to England, he entered into practice as a Civil Engineer in Westminster, on Parliamentary work, and on investigation of claims for compensation, &c. In September, 1877, he was nominated by Sir Charles Hutton Gregory, K.C.M.G., and appointed by the Crown Agents for the Colonies, as an Assistant Engineer on the Cape Government Railways, where he was chiefly engaged on the construction of the Beaufort West Extension of those railways ; first as an Assistant, then as Acting District, and finally as District Engineer. This included the erection of some large wroxaght-iron bridges, besides other works of importance, the work being carried out departmentally, without the intervention of a contractor.

On his return from the Cape of Good Hope, at the termination of his engagement, he, in Kovember 1880, undertook the con- struction of the Malta Railway, which he carried out with com- plete success. This work included a tunnel f mile long, under the lines of fortification ; also various military defence works. He was then appointed Engineer and General Manager by the Malta Railway Company, Limited, maintaining and working the line in a masterly fashion, as will be seen by any one who can read the local accounts of the different Festas, involving sudden and large traffic ; his greatest care having been to avoid peril to life or limb, and his pride being a clean bill. In November, 1887, he resigned his appointment at Malta, in order to take up the position of Resident Engineer and General Manager of the Santa Matia Railway, United States of Colombia, and left England in December of that year; but less than four months later, on the 1st of April, 1 888, he died of typhus fever.

His career makes it manifest that he was a capable engineer, of large experience. An accomplished artist, he rarely lost an oppor-

Obituary.] FRANK ALEXANDER BROWN GENESTE. 377

tunity of using this power ; and while always placing work before pleasure, his love of nature gave him that change of labour which to such minds is recreation. Comprehensive in his tastes ; able with all the aids at his command, especially pen and pencil, he was ever ready to amuse or instruct even a child ; and duty being always paramount, his genial character served him in making the way of life pleasant as well as consistent.

He was elected an Associate on the 7 th of April, 1871, and was transferred to the class of Member on the 13th of December, 1887.

CHARLES MAEKHAMi was the son of Mr. Charles Markham, Solicitor and Clerk of the Peace for the County of NortJiampton, and was born on the 1st of March, 1823. He was educated at Oundle School, and subsequently at Edinburgh University. It had been intended that he should follow an agricultural career, but his bent being towards mechanics, he went to France, and became Manager of the Marquise Rolling Mills, between Calais and Boulogne. The Revolution of 1848 destroyed this industry, and Mr. Markham returned to England and studied chemistry for twelve months under Professor Scoffern. He then served a pupilage under Mr. John Hunter, on the Eastern Counties Railway, and in 1852 entered the service of the Midland Railway Company, as an Assistant in the Locomotive Department at Derby. Here he rapidly rose, becoming Assistant Locomotive Superintendent under Mr. Matthew Kirtley. While occupying this position, Mr. Markham conducted a series of exjjeriments on the use of coal in locomotive-engines, that led to the total disuse of coke in the engines of the Midland Company, whereby that undertaking was saved nearly £50,000 a-year. Other workers in the same field shared with him the honour of this great advance ; but Mr. Markham's part in it was acknowledged to be a leading one, and his Paper, read before the Institution of Mechanical Engineers,^ marked a new era in locomotive practice.

In 1863 the Staveley Coal and Iron Works, which had been extensively developed by Mr. Richard Barrow, were transferred to a Limited Company. Mr. Markham accepted the post of Managing Director, and left Belper, where he had lived for some years, to take up his residence at Brimington Hall. No better choice for such a post could have been made. Mr. Markham was gifted

' The substance of this notice is taken from the Derbyshire local press of August and September, 1888. ^ lustilutiuu uf Mechanical Engineers. Procccdiugs, 18GU, p. 117.

378 CHARLES MABKHAM. [Oljituaiy.

"svitli a tlioroug'h business capacity; he had had a wide and far-reaching experience, and he worked with an energj^ and zeal which coukl hardly have been surpassed had the concern been his own. At the same time, under his superintendence, and the direc- tion of competent officials who had the immediate oversight of the various departments, nothing but work of the best quality — both as to material and execution — was turned out. The result was that a connection was built up which sustained the Company through the trying and prolonged period of commercial depression, which proved disastrous to so many undertakings of a similar character, and enabled it to pay its shareholders, in the worst of times, a fair interest for the money they had invested. Shortly after the formation of the Company, an ojiportunity was afforded to those of the shareholders who desired to avail themselves of it, of inspecting their newly-acquired property. An interesting gathering assembled, and after the visitors had been conducted over the works, and had examined the surface managements at several of the prin- cipal collieries, they were entertained to dinner in the Workmen's Hall at Barrow Hill. Mr. Barrow himself presided, and Mr. Markham was one of the speakers. Mr. Barrow referred in terms of pardonable pride to the growth of the works under his manage- ment, pointing out that the output had increased from 50,000 tons of coal and 15,000 tons of castings in 1844, to 700,000 tons of coal and 20,000 tons of castings in 1864. During Mr. Markham's management the development continued, though the ratio of increase, in consequence of the exigencies of keen and constantly- extending competition, had hardly been the same, the output at the time of his death being over 850,000 tons of coal, and 40,000 tons of castings. Mr. Markham's devotion to business prevented him taking a very active part in public affairs in the district, but when once he undertook a duty, he always performed it with assiduity and thoroughness. He was made a County Magistrate in 1869, and while health permitted he attended the Quarter Sessions for the county, and the local petty sessions, with exemplaiy regularity, and always manifested the utmost anxiety to administer even- handed justice.

Mr. Markham took a deep interest in the welfare of the Chester- field and North Derbyshire Hospital, of which he was one of the Vice Presidents. It was chiefly through his instrumentality that the area covered by tlie operations of this institution was divided into districts, each returning a working man to act ujion the Board of Management. He was one of the leading members of the Iron and Steel Institute ; he however disa2iproved of the forma-

Obituary.] CHARLES MAKKHAM. 379

tion of local institutes, believing that the necessities of the case were met by institutes of a national character, and hence he took no part in connection with the Chesterfield and Derbyshire Insti- tute of Mining Engineers. The erection of the Stephenson Memorial Hall was also in opposition to his views, and he declined to assist it by a contribution ; but on the day when the building was opened, he forwarded a cheque for £100, for the purpose of providing a supply of books for the Free Library of the borough, which was located in a portion of the building.

The poor always found a warm friend in Mr. Markham, and for several years he forwarded donations to provide Christmas treats for the inmates of the Chesterfield Union Workhouse, and the children in the Industrial Schools. In 1880 he gave a donation of £1,000 to his native town of Northampton, to found a "Markham Memorial" in memory of his parents. By Mr. Markham's wish, the money was devoted to providing prizes for Board and other Schools in Northampton, and to adding to the Free Library. Mr. Markham became some years ago the purchaser of Tapton House — where the last days of George Stephenson were spent — and he then removed from Brimingion Hall to Tapton House, where he afterwards resided.

Towards the end of 1887, Mr. Markham's health began to fail; and the death, shortly afterwards, of a son to whom he was much attached was a blew from which he did not seem to rally. He died on the 30th of August, 1888, in his sixty-sixth year.

Mr. Markham was elected a Member of the Institution on the 6thof April, 1864.

JULIUS PAZZANI was born on the 16th of June, 1841, at Brunn, Austria, where his father, who was a Greek by birth, had settled on becoming a naturalized Austrian subject. The younger Pazzani was educated at the Imperial Polytechnickum of Vienna, and in due time passed the principal examinations. In 1861, he entered the service of the Imperial Continental Gas Association as junior assistant at their Erdberg works, Vienna. In 1868, he became Engineer of the Belvedere works of the same company, also in Vienna, which he entirely reconstructed, and greatly extended, insomuch that of what he found there, nothing remained in a few years. In 1880, Mr. Pazzani was transferred to the Rotterdam station of the Association as Chief Engineer. Here he rebuilt the Delfshaven works, and generally reorganized the business of the company. Four years later, in 1884, he was

380 JULIUS PAZZANI. [Obituary.

removed to Amsterclam, also as Chief Engineer, with the special mission to construct two new gasworks, and to lay new mains and services all over the town. In this capacity Mr, Pazzani successfully carried out all the difficult and delicate negotiations involved in repijDing a city of 400,000 inhabitants, being exposed, moreover, to the hostility of the old gas interest, which it was his business to supersede. Mr. Pazzani just lived long enough to complete this undertaking, the first station being put into operation in September 1885, and the second or Amstel works in the course of 1887.

Mr. Pazzani's death was very sudden. On the evening of the 20th of October, 1888, he visited a friend, and on returning home complained of feeling unwell. He rapidly became unconscious, and died at seven o'clock the next morning — the immediate cause being a paralytic stroke.

Mr. Pazzani was elected a Member of the Institution on the 11th of January, 1887, and was also connected with the Gas Institute and the cognate bodies in Germany and Holland. In addition to being a notable engineer, Mr. Pazzani was a most amiable and accomplished man. He was a specially-skilled ling-uist, speaking Avith facility German, English, French, Italian, and Dutch.

WILLIAM EOGEES was the eldest son of the late Ebenezer Eogers, of Abercarne Vach, Abercarne, Monmouthshire, and was born on the 5th of February, 1843. At the age of sixteen, he entered the Ebbw Vale Ironworks, and went through the various shops, being afterwards engaged as an assistant in the engineering department.

Leaving these works in 1863, he was engaged for eighteen months as Assistant Manager under the late Mr. S. H. Blackwell, at the Eussell Hall and Corbyn Ironworks in Staffordshire, and ujjon the branch railways connected therewith. Having by this time acquired a considerable knowledge of the manufacture of iron, and an amount of useful experience in the construction of iron- work generally, which proved invaluable to him in the varied works which he had subsequently to carry out in foreign countries, he, in 1865, went to Australia. For a period of six years from this time he found employment uj)on the railways, waterworks, and other public works in the Government of South Australia, first as an Assistant Engineer, then as Eesident Engineer in charge of the surveys and construction of the Port Wakefield and

Obituary.] WILLIAM ROGERS. 381

Hayles riain Eailway, and of tlie Dry Creek Loop Line, under Mr. William Hanson, Eugineer-in-chief, and Mr. J. England, and subsequently as Chief Engineer to the contractors of the Northern Extension Eailway. In these various positions throughout this period, Mr. Rogers proved a thoroughly trustworthy surveyor, and showed that he possessed great natural capacity as an engineer.

Returning to this coimtry in 1871, he spent but a short time at home, for the same year he obtained an appointment as assistant on the staff of the Imperial Government Railways of Japan. During the next five years he filled in succession the posts of Resident Engineer during the constriiction of the Central Section of the Osaka-Kioto line, and Resident Engineer in charge of the open Yokohama- Yedo line, under Mr. R. Vicars-Boyle, C.S.I., the Engineer-in-chiefti' Owing to the abandonment at this time of further extensions b}' the Japanese Government, and the con- sequent reductions in their English staff, Mr. Rogers, in 1870, found himself out of employment, and came home.

After a few months in England, he accepted a post as Assistant Engineer on the East London and Queenstown Railway, under Mr. James Fforde, Chief Engineer for the Cape Colony, and was engaged on that railway until its completion, and on other minor works until his return in 1881.

The next field in which Mr. Rogers worked was Brazil ; as Chief District Engineer, he was engaged in the construction of the Alagoas Railway, and afterwards as Chief Engineer of the surveys, and preparing the plans and estimates of the Sergipe Railway, 115 miles in length, for Messrs. Hugh Wilson and Sons. Respect- ing his work on this latter line, the Consulting Engineers, Messrs. Hawkshaw, Son, and Hayter, write : " This line has been laid out with skill and jiidgment, and complete information obtained, and we have never known a work of the kind more rapidly, and at the same time more effectually done." The commencement of the const]*uction of the Sergipe Railway appearing to be postponed for some time, Mr. Rogers returned to England in 1885, and being unwilling to wait an indefinite time for the resiimption of active work in this district of Brazil, he accepted the post of Engineer- in-chief of the West Australian Railway then projected from Albany to Perth, and started for that colony in August, 1886. For a period of eighteen months, he carried out the multifarious duties of locating and inaugurating the construction of the line in a sparsely populated and almost unexplored country, throughout its length of upwards of 240 miles, when, on the 20th of February, whilst at his rooms at Albany, he was suddenly struck

382 WILLIAM KOGERS. [Obituary.

down with an attack of apoplexy, and expired in a few minutes. Thus, at the early age of forty-five years, ended a hright career, marked by hard and energetic work, and an amount of varied ex- perience in foreign lands that falls to the lot of few engineers.

Mr. Eogers was elected an Associate of the Institution on the 6th of February, 1877, and was transferred to the class of Member on the 1st of April, 1879.

GEOEGE HEXXET EOSS, the eldest son of the late Commander C. H. Eoss, E.X., was born on the 19th of March, 1843. At the age of- seventeen he was placed with Mr. E. Bagot as a pupil for three years, and was then employed for two years iu?. that gentleman's drawing-office, and on surveys for railways, waterworks, and town sewerao-e. In 1865 he entered the service of Mr. (now Sir Alexander) Eendel, and was employed for four years as a draughts- man in his London office, and in making surveys at Llanelly, the Victoria Docks, and elsewhere. Then Mr. Eendel appointed him Eesident Engineer on the Workington Harbour Improvements, where he had charge for three years of considerable marine surveys, the cutting of a new channel for the Eiver Derwent, the erection of a timber jetty, and the building of a breakwater in Portland- cement concrete, besides other minor works. In March, 1872, Mr. Eoss proceeded to Trinidad as Engineer and Surveyor for the Colonial Company, and was occupied in surveying large sugar- estates, and in laying out railways. "While in that colony he passed the Government examination as a Sworn Surveyor. He returned to England in December, 1873, and for six months of the following year was again in Mr. Eendel's drawing-office. In November, 187-4, he was appointed, on the nomination of Sir Charles Hutton Gregory, K.C.M.G., an Assistant Engineer on the Cape Government Eailway Surveys. Two years later he was placed in charge, under the District Engineer, Mr. G. D. Ather- stone, of No. 1 Section of District 2 of the Grahamstown Branch of the Midland and North-Eastern (Cape) Eailways. The survey and construction of this section, which included some heavy earth- works, such as cuttings in hard rock, an embankment nearly 70 feet high, and several retaining walls, a short tunnel through very treacherous strata of rock on clay beds, a temix)rary terminal station, with the necessary sidings and accommodation for goods, all such works as culverts, platelaj-ing, ballasting, and the telegraj^h, "were completed by Mr. Eoss, in the most careful and accurate

Obituary.] GEORGE HENNET ROSS. 383

manner. From September, 1879, to August, 1880, he took charge, for Mr. Faviell, the Contractor, of the construction of 10^ miles of the Midland System of the Cape Government Eailways, which included a tunnel, and some heavy cuttings and embankments about 30 miles south of Cradock. In August, 1881, Mr. Eoss was appointed District Engineer on the Survey Staff, under Mr. William Willcox, and placed in charge of survey camps on the Cradock and Colesberg Extension, and on proposed junction lines between the Midland and Eastern Railway Systems. This appoint- ment Mr. Eoss held until September, 1883, when he returned to England. In August, 1884, he was engaged by Mr. John Brunton and Mr. T. Claxton Fidler, Joint Engineers for the East India Tramways Company, as Eesident Engineer on the construction of the Karachi Steam Tramways, which work he carried out to their entire satisfaction, performing the duties of his position there with zeal, energy and ability. When the works were completed, and handed over by the contractors to the Company in September, 1885, Mr. Eoss returned to England. In the following November he obtained, again on the nomination of Sir Charles Hutton Gregory, the appointment of Colonial Engineer and Surveyor at Lagos. This post he held until his death, which occurred on the 16th of August, 1888, from fever.

Mr. Eoss was much regretted. A member of the Institution, under whom he served on the Cape Government Eailways writing after his death, said : — " He was a cheery bright fellow, was, I know, most hospitable and popular, and I am sure was thoroughly straightforward in all respects. His untimely death will be greatly regretted by all the members of the Cape Eailways who knew him, and by many others in the Colony, as he had a wide circle of friends, and, to the best of my belief, no enemies."

He was elected an Associate of the Institution on the 9th of April, 1872, was jilaced in the group of Associate Members on its creation in December, 1878, and transferred to the class of Members on the 14th of April, 1885.

EANSON COLECOME BATTEEBEE, eldest son of the late Joseph Batterbee, of Eeigate, was born on the 10th of March, 1838. In 1859 he was articled for three years to Mr. Eobert Sinclair, then Engineer-in-Chief of the Eastern Counties (now Great Eastern) Eailway. On the completion of his pupilage, Mr. Batterbee re- mained with Mr. Sinclair, and with his successor, Mr. H. W.

38-1 RANSON COLECOME BATTERBEE. i [Obituary.

Davis, for ten years. He left the Great Eastern Company in 1872 to become manager of the Brazilian Street Eailway Com- pany's line of tramway from Eecife to Caxanga. This under- taking, which was of a peculiar nature, had fallen into disrepute, and it required considerable tact and firmness to reinstate it. Mr. Batterbee gave great satisfaction in a very difficult jDosition, his management being distingiiished by sound good sense, attention to the interests both of the Company and the public, and strict integrity. About 1876 Mr. Batterbee returned to England, and did not resume professional employment until 1883, when he went to Parahyba de Xorte, also in Brazil, as Eesident Engineer of the Conde d'Eu Eailway. Here, again, a difficult task was imjiosed upon him, owing to disputes between the Company and the Con- tractors for the line. The worry and responsibility predisposed him to sickness, and an attack of heat apoplexy intervening, he was obliged to come home eighteen months later completely prostrated. Eecovering his health, he returned to Brazil in October 1884, and remained in charge of the line till April 1885, when, his health again giving way, he was compelled to resign his appointment. From that time till his death, on the 15th of July, 1888, Mr. Batterbee, though unable to engage in active work, took part in the settlement of the contract accounts, and performed other duties for the Company in London up to the end of 1887.

Mr. Batterbee was intelligent, courteous, and amiable, zealous in the discharge of his duties, and honourable in all his proceedings, and was much and deservedly esteemed by those with whom he came in contact. He was elected an Associate Member on the oth of December, 1882.

JAMES JOHN ALEXANDEE FLOWEE was born at the Baptist Parsonage, Chobham, Surrey, on the 21st of April, 1847. After being educated by private tutors, he served a pupilage to Messrs. A. and J. Inglis, Marine Engineers, Glasgow. He then proceeded to the Cape, and after Ijeing for a short time in private practice he was appointed Acting Engineer of Cape Town. During his tenure of that post he carried out an extension of the water-supply from Table Moiintain through Platte Klip, and constructed a small reservoir and filtering-beds. About a year after the return of the City Engineer, the latter Avas pensioned, and Mr. Flower was appointed his successor. During Mr. Flower's absence in England the duties of the office were fulfilled hy his father, and in the

Obituary.] JAMES JOHN ALEXANDER FLOWER. 385

result lie never personally took np the appointment, preferring to remain at home as the London rejiresentative of the firm in which he was a partner, namely, Messrs. James Flower and Sons, Mechan- ical Engineers, largely occupied in the construction and importation of machinery for gold-mining. He was thus occupied until the spring of 1888, when, owing to some misunderstanding with the Cape banks, the firm were obliged temporarily to suspend payment, being at the time in the full swing of business. This untoward event caused a shock from which Mr. Flower never recovered, and he died broken-hearted on the 12th of April, 1888.

Mr. Flower was elected an Associate of the Institution on the 2nd of December, 1873, and was transferred to the class of Associate Members on the separation of the former grade into two divisions.

SAMUEL HAEPUR was born at Derby on the 1st of March, 1820. At the age of seventeen he was articled to Mr. Samuel Henry Oakes, a civil engineer, whose daughter he subsequently married.

In 1841, shortly after completing his pupilage, he was appointed Borough Surveyor of Derby, where he carried out some important engineering works, amongst them the construction of the Flood- sewers and Main-drainage of the town, also the erection of a bridge across the Derwent. In 1846, while at Derby, he invented and first used the transverse invert-bricks, now so generally employed in the construction of sewers.

In 1851 he resigned his appointment at Derloy, in order to become Managing Partner in the firm of Thomlinson, Harpur and Harpur, Contractors. Among the works carried out by him in that capacity were : The main drainage and waterworks at Ashby-de-la-Zouch ; the Thornton Reservoir of the Leicester waterworks ; the main drainage and sewage-disposal works of the City of Coventry ; sewerage works at Burslem, and sewerage and sewage-disposal works at Cheltenham. These contracts occupied him until the year 1859, when he removed to Merthyr Tydfil, his firm having taken the contract for the construction of the filter-beds, depositing- tanks, pure-water reservoirs, engine-house, and erection of pumj^ing- machinery, as a portion of the scheme for sujiplying the district of Merthyr with water. With that contract began Mr. Harpur's connection with Merthyr Tydfil, and it became the means of the firm acquiring other work in the neighbourhood. On the failure of the original contractors for the construction of the storage-reservoir at Pentwyn, also in connection with the Merthyr waterworks, the

[the IMST. C.E. VOL. XCV.] 2 C

386 SAMUEL HARPUR. [Obituary.

contract was given to Mr. Harpnr, who carried it to completion in a satisfactory manner. The whole of the foregoing works were executed under the personal supervision and instructions of Mr. Harpur, his two partners seldom visiting the works.

In consequence of serious difficulties encountered in the con- struction of the Pentwyn Eeservoir, Mr. Harpur sustained heavy pecuniary losses, and as a result, on the completion of his contracts for these works in 1862, he severed his connection with the firm of Thomlinson, Harpur and Harpur, and resumed the profession of a civil engineer. About this time the appointment of Engineer and Surveyor to the Merthjn: Tydfil Local Board became vacant, and Mr. Harpur applied for the post, to which he was unanimously elected. The district of the Board was at that time in a lamentably unsanitary condition, the registered death-rate being upwards of 30 per 1,000. Mr. Harpur immediately set to work with a deter- mination to improve the health of the locality. Under his direc- tions the water-supply was at once extended to the outlying districts, and as the water came to be used by the inhabitants generally, the necessity for a proper system of drainage became more and more apparent. Mr. Harjjur thereupon prepared a scheme of sewerage for the district, and for disposing of the sewage by irrigation. The Board, however, after lengthy discussions, decided to carry out the sewerage scheme only, leaving the matter of the ultimate disposal of the sewage for further consideration, and in the meantime to discharge the sewage into the Taff". The contract for the complete sewerage of the district was let, and the Avork carried out in 1868-9. Scarcely, however, had the sewers been completed, when a Chancery injunction was obtained against the Board, restraining them from discharging sewage into the river. The Board were now forced to face the question of properly deodorizing and disposing of the sewage, and ultimately decided to adopt Mr. Harpur' s amended scheme of irrigation. Some of the lands selected for irrigation purposes were situated at a distance of 10 or 11 miles from Merthyr, and the line of the outfall sewer lay in places through very precipitous ground, crossed in numerous instances by deep ravines. In surmounting these, Mr. Harpur introduced a new feature in sewer-construction. Instead of follow- ing the contour of the ground, or siphoning under the depressions, he carried the sewer over by means of timber tubes supported on stone piers, the tubes being formed in segmental sections, banded together by iron straps. This mode of construction has proved highly satisfactory, and the traveller on the Taif Vale Eailway cannot but be struck by the novel features presented in numerous

Obituary.] SAMUEL HARPUR. 387

places by these tubes. Subsequently to the completion of the Merthyr Outfall Sewer, and when the lands of the Merthyr Local Board were being prepared for the reception of the sewage, the Merthyr Board arranged with the neighbouring Local Boards of Aberdare and Mountain Ash to treat the sewage of those districts on their land, and Mr. Harpur was engaged by those Boards as Consulting Engineer to carry out the work. The lands for irrigating the sewage of these three districts (some 400 acres in extent) were ultimately laid out by Mr. Harpur, and are now accomplishing all the objects which were sought to be attained, and are yielding a profit on working expenses. Upon the delivery of sewage from Aberdare and Mountain Ash on to the farm, the three Local Boards became joint-owners, and the farms were invested in a Committee representative of the three Boards, entitled the " Merthy and Aberdare Joint Sewage-Farms Committee." To this body Mr. Harpur continued to act as Engi- neer up to the time of his death.

In 1885, failing health compelled Mr. Harpui to resign his appointment under the Merthyr Local Board, but latterly his health had to outward appearance very much improved, and his death came at last very suddenly, on the 10th of November, 1888. after an illness of only five days' duration.

He was elected an Associate of the Institution on the 6th of May, 1873, and was transferred to the class of Associate Members on the formation of the latter grade.

Mr. Harpur was a man of great reserve ; he was earnest, conscientious, and upright in his dealings with all men, and scrupulously straightforward and impartial in the conduct of his business ; he was, moreover, very unpretentious in his manner, would never push himself forward in any way, and a marked trait in his character was a want of ambition. From his long experience in sanitary engineering, his advice and assistance was frequently sought, and he was often engaged as a witness in Parliament and elsewhere upon sanitary matters, though seldom seeking such engagements.

EENEST FEEDERIC MOEANT, son of Mr. John Edward Morant-Gale, was born on the 29 th of July, 1849, at Upham, in Hampshire. He was educated for the Indian Civil Service, but on reaching full age he resolved to try the more adventurous life of a sheep-farmer in Uruguay. Like many others, he was unfortunate.

388 ERNEST FEEDERIC MORANT. [Obituary.

and in 1873 found Mmself without capital or employment. Under the secirciimstances he joined the North Western of Uruguay Eailway as an assistant under Mr. W. G. Ferrar, by whom he was afforded every opportunity of acquiring engineering knowledge. Possessed of a keen intellect, he was, within two years, fully competent to undertake the ordinary duties of an assistant engineer, the practical knowledge gained in his uphill experience making his services of far more than ordinary value. From the North Western of Uruguay Company, Mr. Morant passed to the Buenos Ayres and Campana Eailway in the Argentine Eepublic. Owing to the completion of this line, he was for some time without regular emplo}Tnent, but proceeding to Brazil he was engaged by Mr. John Dixon, on the Eio de Janeiro Waterworks, and subsequently became one of Messrs. AVaring's agents on the Minas and Eio Eailway. While in Brazil he married a daughter of Mr. John Morritt, a prominent merchant of Eio, and joined him in several business undertakings, reorganizing a cotton-mill, and also aiding in the working of steam-navigation on the great Brazilian rivers. Having made a moderate fortune he returned to London, and entered into business as a contractor. In conjunction with another gentleman, Mr. Morant, in 1886, tendered for the construction of the great sewage-collecting tank at Barking. He was also much interested in the adoption of shale and other mineral oils for fuel purposes, under the form of gas. Mr. Morant was a director of the Amazon Steam Navigation Company, where his intimate knowledge of the river was of considerable service to his colleagues. Mr. Morant was an excellent business man, seeing at once the points of anything presented to him, and was able to go to the root of a matter.

Mr. Morant was elected an Associate Member of the Institution on the 2nd of February, 1886. His death, on the 26th of August, 1888, was the result of being thrown from a dog-cart.

EOBEET PINCHIN was born in the year 1821. In 1842 he was articled to Mr. W. J. Lord, formerly of the Eoyal Engineers, and on the expiration of his pupilage he was employed by Mr. (now Sir Joseph) Bazalgette, who was surveying a proposed railway from Birmingham to Stratford-on-Avon. On the completion of this work Mr. Pinchin, in 1846, had entire charge of the survey of the Parish of Moncton Combes, in Somersetshire, In the same year he sailed for India.- In taking his passage, however, provision

Obituary.] EGBERT PINCHIN. 389

was made to give Cape Town a call, and if professional prospects were at all bright, he proposed making the Cape Colony his future home. This was all the more desirable, as India was then univer- sally denounced as an unhealthy country for residence.

On arriving in Cape Town Mr. Pinchin called upon the several Government officers in charge of public works. The result was disappointing. Civil engineering at that date was confined to the construction of a few main lines of road and simple bridges, the erection of prisons, and other Government buildings, works which held out very little encouragement for gaining either pro- fessional experience or a suitable livelihood. When on the eve of continuing his journey to India, he was advised to call upon the late Mr. Charles Bell, the then Surveyor-General for the Colony, and at this visit Mr. Pinchin's professional future was decided. Mr. Bell pointed out that the only lucrative or permanent profes- sional employment in the Colony for some years would be Govern- ment land surveying, and advised Mr. Pinchin to devote his attention to this branch of his profession, and remain in Cape Colony. Through Mr. Bell's kindly advice and assistance he passed the necessary Government examination, and was licensed to practice as a land surveyor in the Cape Colony. No favourable opening offered itself in the Cape districts, which were at that time overstocked with surveyors, so Mr. Pinchin proceeded to the eastern province.

Port Elizabeth, in 1846, was insignificant compared -with its present influential standing, but was even then rapidly rising in mercantile importance. Mr. Pinchin's advent was, therefore, professionally propitious, and subsequently most advantageous from a financial point of view. The laying out of this " Liverpool of South Africa," Port Elizabeth and suburbs, may be said to have been Mr. Pinchin's lifework. For, though enjoying a wide- spread professional connection, he made his headquarters at Port Elizabeth, where he was an authority upon local land questions to the time of his death. About 1863 he was for a brief time in partnership with Mr. G. W. Smith, the Government Engineer and Surveyor now his successor, but depression in trade led to its termination, and Mr. Pinchin enjoyed for many years almost a monopoly of practice. The enlargement of the town and municipal improvements gave Mr. Pinchin opportunity for the exercise of his attainments as a civil engineer. Prize designs were invited by the Corporation for supplying Port Elizabeth with water, and Mr. Pinchin's scheme was many years afterwards adopted, and construction carried out, by the late Mr. J. H. Wicksteed, at an

390 ROBERT PINCHIN. [Obituary.

outlay of about £150,000. The works for a water-supply to Uitenhage were entrusted to Mr. Pinchin, but his extensive and lucrative land survey practice left him scant time to pursue other branches of the profession.

In 1872, when the Colonial Government entered upon the con- struction of railways, Mr. Pinchin's services were in constant demand for valuing, defining, and dividing town-lands for the various works and expropriations naturally attending the intro- duction of railways. About this time he became acquainted with Mr. H. L. Spindler, assistant to the Chief Eesident Eailway Engineer. Mr. Pinchin, on the part of the public, and Mr. Sjiindler on the part of Government, carried out the settlement of the many and varied land questions involved in the compulsory expro- priation of property. The professional ability and integrity dis- played by Mr. Pinchin averted any recourse to law, and his services were appreciated both by the Government and his clients. When Mr. Spindler retired from the railway service in 1879, and entered upon private practice, Mr. Pinchin became his professional associate, and remained so to the day of his death.

Mr. Pinchin enjoyed a well-deserved reputation as a geologist, and was for many years a member of the geological societies of London and Vienna. On several occasions his knowledge of the eastern province of the Colony saved prospectors and others fruitless search for coal, gold, &c. ; and his services were freely, and in most cases gratuitously, oifered to surrounding farmers to encourage search for precious stones, minerals, and in well-sinking. One of the last of many benefits conferred by him ujion Port Elizabeth was the careful arrangement of geological specimens acquired by the Eastern Province Naturalist Society.

From his scientific tastes and education he was not a man to attract many friends in a town of commercial men. He took no part in politics, nor did he seek municipal distinction, being out of touch with what may be termed strife for office. Upright, straightforward, and accurate in his business affairs, he made many friends among the Dutch farmers. Few enjoyed the privilege of his confidence, and though to most he would appear unusually reserved in manner, yet the young surveyor would find him animated and enthusiastic upon professional subjects. On his death-bed he sat up to write out, in Mr. Spindler's memorandum- book, a trigonometrical formula which he greatly admired for its simplicity. He had jiassed the crisis of a dangerous disorder, and had just returned to his office in a weakened condition, when a chilly south-east wind brought on an attack of inflammation of the

Obituary.] ROBERT PINCHIN. 391

lungs, to which he succumbed on the 9th of May, 1888. Mr. Pinchin was elected an Associate of the Institution on the 3rd of February, 1874, and was subsequently included among the

Associate Members.

CHAELES THOMAS SPENCER was born on the 12th of May, 1856. At the age of sixteen he was articled to Sir James Ramsden, of Barrow-in-Furness, for four years, and on the expiration of his pupilage, in 1876, he continued in the service of the Furness Railway until 1882. During that period he was engaged, as Assistant Engineer, in surveying and levelling, making plans, sections and drawings, superintending various works in connection with the railway and the Barrow Docks. In 1883-4, he was employed in the office of Mr. Stileman, as Assistant Engineer, in jDreparing contract and Parliamentary plans and sections. In January, 1884, he entered the service of the East Indian Rail- way Company, and was appointed as Assistant to Mr, Graham Peddie, District Engineer of the Jubbulpore Extension, and placed in charge of 140 miles of line. In 1885 he was, during the absence of the incumbent, placed in temporary charge of the next division of 83 miles. While thus engaged he carried out the Katni-Umaria (State) Railway Junction works with the East Indian Railway, and subsequently the junction of the Indian Midland Railway at Manikpur, with the East Indian system. While thus engaged, he died very suddenly of cholera, at Moghal Sarai, on the 7th of July, 1888.

Mr. Spencer was an exceptionally able man, thoroughly well up in his work, and very energetic. His death deprived the East Indian Railway Company of one of the most promising of its younger officers. In private life he was a universal favourite, being an eminently social and accomplished all-round man. He could dance, sing, play cricket and lawn-tennis, fight, preach, and play the organ in church, and was recognised as a type of true English manhood, kind, considerate, and fair. He was elected an Associate Member of the Institution on the 24th of May, 1887.

392 JOHN TRICKETT. [Obituary.

JOHN TRICKETT commenced his apprenticeship to engineering at the age of sixteen years, at the Butterley Ironworks, Derbyshire ; four years later he entered the Deptford works of the General Steam Navigation Company, then one of the largest enterprises of the kind in the country, as Assistant to the Managing Engineer, During the ten years he was with that Company he designed and constructed twelve pairs of marine engines and boilers, and super- intended the working of most of them on their first voyage. He also designed several land engines, constructed at the Company's works, as well as all the boilers for replacements in the ships, and supervised the new work in the shops, and repairs on board the ships.

Subsequently Mr. Trickett entered the works of Messrs. John Penn and Sons, at Greenwich, where he held a good appointment. In 1846 he was recommended by Messrs. Penn to the Admiralty, in answer to an invitation to submit the name of any one in their establishment suitable for employment in the public service. In July of that year he was appointed to the Steam Factory at Woolwich, and while there, as Assistant to the Chief Engineer, he carried out a series of elaborate screw experiments in the "Minx," " Teazer," " Eifleman," and " Sharpshooter."

In April, 1854, Mr. Trickett was appointed Chief Engineer and Inspector of Steam Machinery at Devonport and Keyham. Among the labour-saving arrangements made by him at those yards, were the multi-spindled machines for drilling six holes at one time in angle-iron, and for cutting four holes at once in tube-plates for boilers, which were then the first tools of the kind made for boiler-work. In 1855 he designed and fitted a rope-driving arrange- ment to all the travelling cranes in the workshops, driving them from the running shafting which is now generally adopted in private shops. He also designed and fitted the post-cranes in the foundry with steam-power for working them. In 1855-6 Mr. Trickett made and erected the first iron sheer-legs put up in the dockyards, and fitted them with steam-power. He also fitted steam- power to the 40-ton crane erected by Fairbairn at the north basin, and subsequently to the other cranes at Keyham, and to some at Devonport. In July, 1862, Mr. Trickett was transferred, at his own request, to Woolwich Yard, on the retirement of the late Chief Engineer Mr. Charles Atherton, where he remained until the closing of the yard at the end of September, 1869, when he returned to Keyham.

Obituary.] JOHN TIIICKETT. 393

Mr. Trickett gave much attention to the manufacture of yarn and rope, and made many improvements in the machinery, in all the details of manufacture, which resulted in much better and stronger cordage being produced, and the cost of manufacture lessened. He vras a member of the Boiler Committee formed in June, 1874, and was present in taking nearly all the evidence of persons examined by the Committee.

He retired from the post of Chief Engineer of H.M. Dockyard, Devonport, in December, 1879, and his long and able services were recognized by the Admiralty awarding him a special pension, which he enjoyed till his death, on the 9th of November, 1888.

Mr. Trickett was elected an Associate of the Institution on the 1st of February, 1853, and when the division into professional and non-professional Associates was made, he was graded Associate Member.

JOHN WAKEFOED was born on the 17th of December, 1833 He served a pupilage under the late Mr. E. Allen Stickney, Surveyor to the Public Health Commissioners of Brighton, who was employed at that time on trigonometrical surveys and in drainage and sea-defence works. Mr. Wakeford was then engaged by the Brighton, Hove and Preston Waterworks Company to prepare plans of the mains, hydrants, &c., belonging to the Company. In 1854 he became assistant to Mr. J. G. Poole, Surveyor to the Local Board of Health, Southampton, when he revised the Board's copy of the Ordnance Survey of the town. From 1854 to 1857 Mr. Wakeford was Assistant Borough Surveyor of Leicester, having the superin- tendence of extensive drainage operations, street- works, and surveys. During 1857-8 he was engaged temporarily with Mr. J. Gr. Poole, and also Mr. C. E. Bernard, of Cardiff. In 1858-9 he was Clerk of the Works and assistant to Mr. Edward Gibbs, Surveyor to the Local Board of Stratford-upon-Avon, in charge of the main-drainage works. In 1859-60 he was engaged by Messrs. Flint and Shenton, surveying extensions of the Borough of Leicester and in making a copy of the plans of that borough. In 1860 he was appointed Assistant Borough Surveyor of Wolverhampton, engaged in pre- paring a scheme of main drainage for that town, which was approved by the Local Government Board. He held this appoint- ment until 1865, when he became assistant to Mr. E. A. Marillier, Engineer to the Hull Docks. From 1866 to 1867 he was

394 JOHN WAKEFORD. [Obituary.

Surveyor to the Beverley and Barmstone Drainage Commis- sioners. The experience gained in these very varied engagements prepared him for the post in which he was destined to pass the remainder of his life. In 1868 he was engaged by the late Mr. J. J. Montgomery, Borough Surveyor of Belfast, as an assistant, and on that gentleman's death in 1884, Mr. Wakeford became Chief Assistant Borough Surveyor under his successor, Mr. J. C. Bretland. Mr. Wakeford held this oifice until his death on the 6th of June, 1888. Mr. Wakeford was elected an Associate Member on the 6th of May, 1886. He was a very careful and painstaking engineer of the second rank, and although not des- tined to attain an independent position, he left a considerable reputation for good and thorough work.

EEEBEET FEAXCIS WAEING, second son of Mr. Thomas Waring, of Cardiff, was born at that place on the 2nd of October, 1859. He was educated chiefly at Cheltenham College, where he took a good position in Physics and Mathematics. He entered his father's oflSce in 1875, and making rapid progress in the profes- sion, was elected an Associate Member of the Institution on the 4th of March, 1884. He soon took charge of the engineering portion of the business, and carried out a scheme of waterworks for the Borough of Aberavon ; a filtration system of sewerage for the City of Llandaff ; sewerage and sanitary works for various estates, and, up to the day before his death, was engaged in completing a scheme of drainage for the Cardiff Union Sanitary Authority.

Mr. Herbert Waring became a partner in the firm of T. Waring and Sons, Civil Engineers and Land Agents, in 1885. Being of a quiet and retiring disposition, he took no active part in the pro- ceedings of the Institution. His death, on the 12th of August, 1888, from a fit of apoplexy after a few hours' illness, was a great shock to his family and to a numerous circle of personal friends.

JOHN ASHWOETH was born on the 3rd of August, 1826. From his earliest youth he evinced an aptitude for mechanics, and was constantly making models of small pieces of machinery. When quite a boy he constructed several miniature water-wheels, which he fixed in small streams near his father's house. As he

obituary.] JOHN ASHWORTH. 395

advanced in years lie gave much attention to improvements in the steam-engine, and was one of the first, if not the first, in Bolton to see the advantages of iising steam at very high pressure. He spent some time in experimenting with a new railway-chair, and some specimens of the pattern he invented were laid on the Lanca- shire and Yorkshire Railway near Bolton. He afterwards brought this invention under the notice of Sir Edward Watkin, hut being at the time greatly occupied with other matters, he failed to follow it up. In due course Mr. Ashworth joined his father, who was in business as a cotton-spinner at Bolton, and turned his attention to the im- provement of spinning machinery, taking out several patents for that purpose. One related to what is called a lap-machine, used for making laps for combing-machines. The selvedges made from the machines then in use were not satisfactory, and after a con- siderable time in experimenting, he introduced a revolving plate in lieu of the stationary or fixed one. This was a great success and was adopted by many machine-makers, who paid Mr. Ashworth a considerable amount for royalties ; he also invented an improved cam for lifting conical valves, which was applied by several other engineers in Bolton.

About the year 1862 Mr. Ashworth left his father's mill, and built a large one for himself at Astley Bridge, near Bolton. This was a model structure in every respect, both as regarded the building and the engines, and machinery. Here he had a pair of engines of 500 or 600 indicated HP., working with an initial pressure of 100 lbs. to the square inch, and burning less fuel in proportion than any engines in the neighbourhood. In the necks of the steel crank-shafts of these engines, he cut elliptical grooves which caused the lubricating oil to pass along the necks in such a way as to be suifused equally over the whole surface, and so prevent heating.

Several other inventions also proceeded from his prolific brain, amongst them an improved cheese-toaster, which is a great favourite, all of the inventions being of a practical nature, and such as commended themselves to his associates in the spinning trade. Mr. Ashworth's motto was " thorough."

Mr. Ashworth was elected an Associate of the Institution on the 12th of May, 1874. He died on the 18th of September, 1888.

396 MAJOR AUGUSTUS SAMUEL WILLIAM CONNOR. [Obituary

MAJOR AUGUSTUS SAMUEL WILLIAM CONNOR, Bombay Staff Corps, son of the late William Connor, Honorary Magistrate of Alignrli, in the North- West Provinces of India, was born on the 18th of June, 184-i. He was ediicated at Mnssoori under the Rev. Dr. Lnoni, and afterwards at the Thomason Civil En- gineering College, Roorkee. From thence he was appointed to the Indian Trignometrical Survey, and did good work in it for some little time. In 1864, he entered the Public Works Department as an assistant engineer, and was employed on the Agra and Bombay road, and afterwards at Gwalior under Lieutenant (now Lieut. - Colonel) J. B. Sparks, Executive Engineer. Here he had the supervision of the construction of barracks, as also the defences of the fortress of Gwalior. His services were so highly esteemed, that although he had meanwhile been appointed to an ensigncy in the Ceylon rifle regiment, he was retained by the Govern- ment of India for eighteen months. He joined his regiment in Ceylon, but soon after exchanged into the 7th Royal Fusileers, and returned to India. The want of occupation during the peaceful period that now intervened in India proved irksome to his busy mind ; he therefore applied for re-employment in the Public Works Department, and was appointed a Second Grade Assistant Engineer on the railway branch, on the 1st of February, 1871, and was pro- moted to First Grade Assistant Engineer on the 1st of March, 1872. He Avas employed on the Indus Valley Railway for some time, but was sent to siiperintend famine operations on the Northern Bengal Railway, when he officiated as Executive Engineer from Janiiary 1874. These duties undermined his health, and he was sent to England on sick leave from March 1876 to September 1877. On returning he was jjosted to the Western Rajj)utana State Railway ; but on the Mysore famine breaking out shortly after, his previous similar services pointed to him as a valuable man, and he was sent there as an officiating Executive Engineer. In May, 1878, he was permanently promoted to executive rank. He returned to the Western Rajputana Railway in December, 1878 ; but again suffer- ing from exposure during the famine works, he was obliged to proceed to England on eighteen months sick leave in March, 1879. He was promoted to Third Grade Executive Engineer during his absence on leave. He received the thanks of the Government for his services during both these famines. His health was perma- nently affected by exposure during this trying time, and he never perfectly recovered. While in England on leave. Captain

Obituary.] MAJOR AUGUSTUS SAMUEL WILLIAM CONNOR. 397

Connor attended the Eoyal Engineer extra courses of study at Chatham, in electricity, telegraphy, army signalling, &c. ; also musketry at Hythe, and on his return to India in January, 1881, he was posted to Burma, where he remained two years, and while there he organized and commanded a company of Eailway Volun- teers. In August, 1883, he was transferred to the Madras Eail- way Surveys. In January, 1885, he was promoted Second Grade Executive Engineer, and transferred to the Sind Pishin State Eailway, where he was employed for over two years. For his services on this frontier line, he was specially commended by the Secretary of State and the Government of India. On the comple- tion of work here, his services were temporarily lent to the Central India Administration in February, 1888, and during an epidemic of smallpox at Mhow, in May, 1888, he caught the infection, and there died, on the 26 th of May.

Major Connor was appointed Ensign in the Ceylon Eifle Eegiment on the 13th February, 1867, he exchanged into the 7th Fusileers in 1869, and was promoted Lieutenant on August 31st, 1870; he then was transferred to the Bombay Staff Corps on April 30th, 1872; promoted to Captain February 13th, 1879, and Major, February 13th, 1887. He was elected an Associate of the Institu- tion on the 7th of December, 1875.

Major Connor was a most painstaking and hardworking officer ; and the conscientious performance of his duties under unfavourable climatic influences, was the indirect cause of his regretted death when yet in his prime.

GEOEGE HAWKINS was born at Daventry, in Northampton- shire, in 1815, and was left an orphan at three years of age.

He commenced his railway career on the Liverpool and Man- chester line under Mr. Braithwaite Poole, and in 1843, joined the London and Brighton Eailway as Goods Manager, at that Company's southern terminus. He was appointed Traffic Manager of the whole system in 1849, which position he held for twenty years. Amongst the able men who, in those days, thronged the Com- mittee Eooms of the Houses of Parliament fighting for and against the schemes of railway extension, Mr. Hawkins was for years a well-known figure ; and wherever crowds of passengers had to be accommodated at races, reviews, and similar gatherings in his

o98 GEOPlGE HAWKINS. [Obituary.

company's district in the South of England, he was alwaj's to be found superintending the arrangements in person. He was elected an Associate of the Institution on the 7th of December, 1858, and to the last he retained a strong interest in the success and develop- ment of the Institution. He had been member of various Lodges of Freemasons for over forty years, holding office in all. At the time of his decease, he was the oldest Past-Master of the Britannia. He died on the 1st of November, 1888, and was buried in the Brighton Cemetery on the 5th of that month. A large number of his old colleagues and employees attended the funeral, notwithstanding that nearly twenty years had elapsed since his retirement from active life.

GEAHAM HEWETT HILLS, Staff Commander, Eoyal Navy, was born on the 5th of July, 1826, at St. Lawrence's, near Eamsgate, his father, the late Captain John Hills, E.N., being then a lieutenant attached to H.M.S. " Eamilies," 74, stationed at Pegwell Bay. He received an education in mathematics and classics under a graduate of Cambridge, at Lancing, in Sussex, till April, 1842, when he embarked as a midshipman in one of Messrs. Soames's vessels then employed as a transport in the service of the Admiralty. This ship, during the first war with China, was sent to Hong Kong, where Mr. Hills contracted the illness which well nigh decimated the troops engaged in the expedition. But through the almost parental care of his captain, he was brought home alive. On his restoration, he embarked again in another transport of the same owners, engaged as a troopship for the service in the West Indies. Eeturning to England in December, 1844, he was soon after entered by the Admiralty as Master's Assistant, and joined H.M.S. " Victory," at Portsmouth, on February 20th, 1845 ; being thus settled in the navigating branch of the royal navy. On the 18th of March, 1845, he was transferred to H.M.S. "Vindictive," 50, carrying the flag of Vice-Admiral Sir F. W. Austen, and com- manded by Captain Sir Michael Seymour, with an old shipmate of his father. Lieutenant Franklyn, as First Lieutenant. In this ship he served till the 16th of June, 1848. By his log-book he can be traced as employed occasionally on surveying service, and drawing and copying charts. Particularly, he was several times engaged in buoying the outer shoals at Halifax. On the 19th of

Obituary] GRAHAM HEWETT HILLS. 399

June, 1848, Mr. Hills was again appointed to tlie old "Victory," tlie flagship at Portsmouth, and on the 28th of June, he received his commission for the rank of Second Master. On the 19th of Septem- ber following, he was removed into H.M. steam vessel "Dwarf," Lieutenant Sherard Osborne commanding. In company with another steamship, the " Dwarf " started the next day for Ireland, being destined to intercept the communications of the disaifected followers of Mr. Smith O'Brien, in the Eiver Suir, and to encourage and support the loyal. In this service the vessel was frequently patrolling the river from Waterford to Carrick-on-Suir, often with much doubt where her old-fashioned and nearly worn-out machinery and frame would land her. Mr. Hills utilized his opportunities to make a survey of the river, and drew up sailing- instructions for it which were published by the Admiralty. The collapse of the armed movement of Mr. Smith O'Brien enabled the services of the " Dwarf" to be dispensed with in the Suir, and on the 19th of January, 1849, the "Dwarf" proceeded to Haulbowline, in Cork harbour, for a complete refit. In June, the vessel returned to the Suir, and made one trip up the river, and was at Queens- town again in August, assisting in the honours of the visit of the Queen and Prince Consort. On the 15th of November, 1849, Lieutenant Sherard Osborne retired on half-pay, and for the remainder of the commission, Mr. Hills was commander.

This insight into coast service in Ireland, led to Mr. Hills being commissioned on the 13th of April, 1850, on the Irish coast survey. Accordingly, on the 15th of that month, he joined H.M.S. " Sparrow," at Waterford, on surveying service under Captain Eraser, E.N. ; under whom he took command of the ship whilst the survey was completed from Carnsore Point to Cork. Mr. Hills went to Ports- mouth in February, 1851, and on the 11th, passed at the Naval College for the rank of Master, with full numbers on the college sheet. The work of Captain Fraser being brought to an end, the " Sparrow " was paid off at Cork, on the 24th of January, 1852, upon which occasion the ship's company presented Mr. Hills with a regulation sword. His next commission, January 28th, 1852, was to H.M.S. "Neptune," 80, guardship at Portsmouth, which he joined on the 5th of February. The monotony of harbour duty brought his regularly kept log-book to an end at the end of April. On the 11th of September, Mr. Hills was glad to be "lent" from the "Neptune" for second-master's duty in H.M.S. " Tyne," carrying troops to and from Ascension, and visiting Rio de Janeiro. A fortnight after returning to Woolwich, on the 3rd of February,

400 GRAHAM HEWETT HILLS. [Obituary.

1853, he rejoined the "Neptune" at Portsmouth ; and on the 25th of April, he was again lent to take charge of H.M.S. " Illustrious," appointed as a hospital ship on account of an outbreak of sickness in the crew of the "Agamemnon." On the 19th of July, 1853, Mr. Hills was sent in the " Ehadamanthus," under the orders of the master attendant of Portsmouth dockyard, to assist in fitting out and bringing round the "Caesar," expected to he found launched at Pembroke dockyard. But the first attempts to launch that vessel failed, and it was not till the 9th of September that the " C^sar " was navigated into Portsmouth harbour, and Mr. Hills rejoined the " Neptune."

As the year 1854 opened, the probabilities of war with Russia were strong. In preparation for the despatch of a fleet to the Baltic, the Admiralty assigned the " Hecla " (s.v.) for special service to embark several scientific officers and specialists to ascertain par- ticulars of the navigation and of the fitness of certain ports for the reception of the fleet. Mr. Hills was one of the officers ordered to join the " Hecla " on this duty, and he sailed from Sheerness, on the llth of February. But on the 16th he was commissioned to H.M.S. "St. Yincent," and he was so borne whilst absent in the "Hecla." That ship proceeded to Hull, and embarked a number of Baltic and North Sea pilots for the common instruc- tion of all concerned. The first attempt to use photography for warlike purposes was made by Captain E. A. Scott, E.N., in the examination now made of the coasts of Norway and Sweden ; but the draughtsmen (chief of them Mr. Hills), were held to be more successful in carrying away likenesses of the coast, than the photographic apparatus, owing to the minute size of its repre- sentations. The process used was instantaneous. They recon- noitered the coast of Pomerania, crossed over to Copenhagen, embarked various Swedish and Danish pilots for information, and returned to the Downs on the 12th of March, where Captain "Washington, E.N., of the Admiralty hydrographic department boarded the ship and gniided them to the fleet under Sir Charles Napier. The "Hecla" proceeded to Portsmouth, where Mr. Hills was ordered to join the " James Watt," 91, fitting out for the Baltic at Devonport, which he did on the 25th of March. It became necessary to secure a handy anchorage for the fleet away from the enemy's to-^vns or fortresses; this was found in Baro- sund, on the north side of the Gulf of Finland, nearly opposite Eevel, and was surveyed, Mr. Hills being one of the officers employed. Shortly afterwards, Mr. Hills was employed on survey-

Obituary.] GRAHAM HEWETT HILLS. 401

ing for the attack on Bomarsiind, biat having been promoted Master for his services at Barosund, he was ordered home at once. Mr. Hills was next commissioned as Master to H.M.S. " Geyser," which lie joined on the 15th of December, under Captain Towers. In April, 1855, this vessel was sent as an advance ship to the Baltic, and on the 2nd of May, found Faro Sound frozen across ; never- theless, the fleet arriving, picked up several prizes, and on the 17th the "Geyser" was ordered to convoy them to England and reached Sheerness on the 30th. By the 20th of June, the " Geyser " had got back to Copenhagen, and soon after was with the fleet off Cronstadt, and took part in sundry contests with bodies of Russian troops employed to harass the fleet from the shore. In August he was engaged at the attack on and the destruction of the great fortress of Swealjorg. On the 11th of November, the fleets took their departure from the Gulf of Finland for England, leaving only frigates and cruisers to bring up the rear, and on the 18th, the "Geyser" had reached Faro Sound on the home- ward voyage.

During the winter of 1855-6, a number of gunboats were built ;it Birkenhead for the Government. The "Geyser" was sent to the l\rersey about the end of April, 1856, to superintend their fitting r»ut, and to bring them to Portsmouth ; but certain of them being found incomplete, Mr. Hills was left to superintend the completion of the " Blossom " and the " Gadfly." This delay led to his be- coming, on the 6th of May, a candidate for the post of Assistant Marine Surveyor for the port of Liverpool. Shortly he had to rejoin the " Geyser," and at Portsmouth he learned that the Liver- pool Dock Board had selected him for the post, and had applied to the Admiralty on the 28th of May for leave for him to accept the appointment. The leave was granted, and thus began his twenty- nine years' service under the board, now entitled the Mersey Docks and Harbour Board. At the commencement of his service, the superior authority was known as the Liverpool Dock Board, but upon taking over the control of the Birkenhead Docks by Act of Parliament in 1857, the title was extended to embrace both sides of the river.

On the 15th of November, 1866, Mr. Hills was promoted to Marine Surveyor, and became head of the department. The service extended from the Mersey Bridge at Warrington, to the whole of Liverpool Bay, and to the whole north coast of Wales, requiring in its chief an accurate knowledge of the shoals and channels always shifting, and a constant alertness to keep the sea-

[tIIE INST. C.E. VOL. XCV.] 2 D

402 GRAHAM HEWETT HILLS. [Obituary.

marks in serviceable positions. It included tlie control of the lighting of the coasts and signal stations, and of the lightships in the bay ; and constant readiness in all weathers, by night and by day, to proceed with steam-power, and often with gunpowder and dynamite, to the removal of wrecks and accidental obstructions occurring in the navigable channels ; and the organization of a portion of the life-boat service.

In 1867 Mr. Hills, with the whole of the naval officers of his grade, passed from the old fashioned title of Master, to the newly invented one of Navigating Lieutenant ; some time before his name had been placed on the Naval Eeserved List. On the 1st of August, 1869, he succeeded by seniority to the rank of Staff Com- mander with the courtesy title of Captain.

Early in his service at Liverpool, it was proposed to him to ask the leave of the " Dock Board " to give his assistance in selecting the landing-places for the Atlantic Cable. He objected to any attempt at " serving many masters," and declined to make the proposal. A matter in which he accepted emplojnnent, extra to his service with the Dock Board, was on the occasion of the construction of the Euncorn Bridge across the Mersey, by the London and North Western Railway. It was necessary in 1861, to have an exact record of the state of the bed of the river, before the construction of the bridge in reference to the widening of the water-way re- quired by the Admiralty of about 100 feet at Euncorn, and the effect afterwards of that widening, and of the construction of the bridge upon the river-bed, some 5 miles along the river. Captain Hills was applied to by the parties interested, and made the record required ; two or three copies of the chart being made in London from his instructions for deposit in the interested quarters.

A system of regular surveys of the Mersey had been set on foot by his predecessors in office. Captain Denham, E.N., and Lieutenant Lord, E.N. As a permanent basis for such surveys, Captain Hills, in 1860, made a complete survey, with a 100-foot chain, of both sides of the Mersey, from Warrington Bridge to the sea; he also levelled and fixed bench-marks on both banks of the river, and with the theodolite fixed by triangulation the position of all prominent objects near the Lancashire and Cheshire shores. This has been the basis for the periodical surveys and observation of the navigation of the river ever since. In 1861 Captain Hills made the first survey of the river-bed founded on that basis. A point which he very early took in hand was the form and construction of buoys. He found the variety, in form, in colours, and in

Obituary] GRAHAM HEWETT HILLS, 403

numbering confusing, tliougli intended to distinguish tlie buoys as sea-marks. The first necessity was to fix on a form which would not be run under water and lost to sight, as many were when most needed in view, by the strength of the currents and the violence of the sea. Another necessity was, that a buoy when sighted by a navigator, should tell him at once that he was inside the channel (where he wished to be), or was outside of it and in peril. The system established by Captain Hills for Liverpool, founded on that introduced by the first Marine Surveyor of Liver- pool, Admiral Denham, was ultimately adopted by the Corporation of Trinity House, when in 1882 a universal system was established for the kingdom.

The attention of Captain Hills wasearnestly directed to that part of the Manchester Ship Canal, which proposed to continue the cut from its union with the Mersey above Kuncorn down the tidal estuary by a deep low-water channel, as far down as a point near the middle of the river-bed, between the opposite towns of Frodsham and Garston. By his little book on the Hydrography of the Mersey, published in May 1858, it can be seen how early he began to study the causes of the formation of the water-channels of the river, and the changes in them, and that, like his predecessors in office, he then objected to any interference with their natural formation. He now strongly felt that such a low-water channel as was pro- posed to be formed in the estuary would certainly, in the course of years, entirely change the character of the inner estuary, and adversely afi'ect and impede, by the deposits it would occasion, the navigation of the outer estuary. The result of the canalization, viz., the silting up of the whole estuary of the Dee, had already been matter of observation with him.

The year 1883 saw the scheme of Mr. Leader Williams intro- duced into Parliament, and after a protracted examination at thirty- nine sittings of the Committee, the House of Commons passed the measure ; but in the House of Lords, the Committee, after ten sittings, decided, on the 10th of August, "that it is not expedient to proceed with this Bill in the present session of Parliament." Captain Hills was examined at great length before the House of Commons.

In 1884 the Parliamentary campaign was entered upon, with a much better knowledge on each side of the opposing views, and of the facts relied upon. The inquiry began in the House of Lords on the 12th of March. The opponents to the measure commenced their case on the 25th of April, opening with the case of the Mersey Docks and Harbour Board ; after the Solicitor to the Board had given evi-

2 D 2

404 GRAHAM HEWETT HILLS. [Obituary.

dence on legal matters, Captain Graham Hills was called, on the 26th of April, and was three days under examination. After forty sittings the Lords' Committee, by a narrow majority, passed the Bill, with some modifications of a monetary character. On the 7th of July the Bill came before the Hoi;se of Commons' Committee, and on the 18th of July Captain Graham Hills was again the second witness called for the opposition by the Mersey Docks and Harbour Board, and was for three daj's under examination. He pointed oiat how the instances of an artificial channel in the Tees, the Tyne, and the Clyde, had no analogy with the case of the Mersey, and that each of these cases, nevertheless, showed that silting up and the reclamation of land followed. This, and the circumstances of the Dee and the Seine, which were analogous to those of the Mersey, proved that the proposed low- water channel in the bed of the river must lead to silting up, to the contraction of the anchorages in the river, and loss of the scouring power of the water. As a practical seaman, Captain Hills pointed out some of the difficulties which must occiir in the navigation of the distance from Liverpool Bar to Manchester in a single tide. Powerful engineering evidence supported Captain Hills' scientific views, and on the 1st of Augiist the Committee unanimously threw out the Bill.

In 1885 the indefatigable promoters of the canal came to Parlia- ment a third time, but with almost a new scheme. The land course of the canal was much modified, and the canal between training-walls in the bed of the estuary, beginning a mile above Eimcorn and going down some 9 miles, was abandoned. In lieu of it, accepting the suggestion of the engineers called by the former opponents, Mr. Eads, Mr. Lyster, Mr. Law, and Mr. Stephenson, the canal was proposed to be carried from Euncorn to Eastham, along the Cheshire shore of the Mersey, terminating with locks there in deep water, and entering at once a large anchorage called the Sloyne. Again on the 23rd and 2-ith of March Captain Hills was in the witness chair before the House of Lords' Committee, con- tending for the least possible infringement on the area of the estuary, and against establishing an unnatural channel along the Cheshire shore. On the 7th of May the Lords' Committee reported in favour of the Bill. On the 15th of June it appeared before the House of Commons' Committee. On the 6th and 7 th of July Captain Hills was called to give evidence ; but the dangers of the present scheme were admittedly far less and more remote than those of its predecessors, and on the 5th of August the Bill was read a third time, thus terminating one of the most remarkable scientific con- troversies dealt with by Parliament.

Obituary.]

GRAHAM HEWETT HILLS.

405

Before this great question had reached its solution, Captain Hills' health had shown signs of weakness, which he had been warned would imperil his life, if he continued night service afloat another winter. In fact, in the year 1883, in the month of January, the wreck of the " City of Brussels," in one of the channels in the bay, about 10 miles outside of the Eock Lighthouse, had brought upon him such exertions and exposure during forty-eight hours, in the pressing necessity to remove the wreck, that a state of collapse followed, which left a permanent physical weakness behind. Before the third hearing of the Manchester Canal Bill, the Mersey Dock Board had consented to his retirement from office with a pension, and with a retaining fee to secure his consultative services. In the quiet of his retirement at Beckenham, Kent, he lived a little over three years, and then succumbed suddenly whilst on a visit to Folkestone, on the 16th of August, 1888. He was elected an Associate of the Institution on the 6th of March, 1866.

*^* The following deaths have occurred, or been made known, since the 3rd of September last, in addition to some of those included in the fores'oino; notices : —

Members

Bateman, Frederic Foster ; born 22

July, 1853 ; died February, 1889. Bell, Wilson ; died 27 November, 1888,

aged 49. DiBBLEE, Frederick Lewis; died 28

September, 1888, aged 50. Hamand, Arthur Samtjel ; died

September, 1888, aged 51. Lastarria, Victorino Aurelio ; born

22 November, 1813; died 27 July,

1888. Leslie, Frederick ; died 20 February,

1889, aged 61. Miller, Daniel; born January 1826;

died 28 September, 1888. MuiR, James; born 31 May, 1817;

died 4 January, 1889.

Associate Batten, "William ; died 18 July, 1888,

aged 65. Boothby, Alexander Cunningham ;

born 26 February, 1857; died 17

December, 1888. Crowley, Jonathan Sparrow; died

13 September, 1888, aged 62.

Murton, Frederic ; died 17 January

1889, aged 71. Parkes, William ; died 5 February,

1889, aged 66. Phipps, George Henry ; bm-n 27 March,

1807; died 11 December, 1888. PuRDON, Wellington; died 14 Feb- ruary, 1889, aged 73. Salter, Frank, B.Sc, Wh. Sc. ; born

19 October, 1848 ; died 31 December,

1888. Schmidt, Bernhard ; died 6 February,

1889, aged 60. Stephens, Frederick Cook; born 21

September, 1829; died 3 January,

1889.

Members. Phillips, Alfred; died 8 February,

1889, aged 44. Roberts, Peter ; born 3 April, 1846 ;

died 25 February, 1888. Tyndall, George Reaveley ; born 17

October, 1856; died 17 November,

1888.

406

GRAHAM HEWETT HILLS.

[Obituary.

Associates.

Hunt, Sir Henry AflTHrR, C.B. ; died 13 January, 1889, aged 78.

KoE, John Phentel ; died 8 Septem- ber, 1888, aged 73.

ViCKERS, George Naylor ; born 14 November, 1830 ; died 20 January, 1889.

Wakefoed, William ; died 9 Novem- ber, 1888, aged 64.

WiLLOCK, Captain Harry Borlase. R.E. ; born 6 March, 1854; died 7 February, 1889.

Wilson, Hugh; died 25 September, 1888, aged 60.

Information resiDecting the life and works, the career and leading characteristics, of any of the above, is solicited, to aid in the preparation of future Obituary Notices. — Sec. Inst. C.E., 22 February, 1889.

Abstracts.] NEW THEOEY OF FRICTION. 407

Sect. III.

ABSTEACTS OF PAPERS IN FOEEIGN TRANSACTIONS AND PERIODICALS.

Neiv Theory of Friction. By N. Pktroff.

(Neue Theorie der Reibung, von N. Petroff, niit Genehinigung des Verfiissers aus dcui russischeu iibersetzt vou L. Wurzel, &c. 8vo. Leipzig, 1887.')

The naphtha industry of Russia has attained enormous pro- portions. American crude oil yields 70 per cent, of petrolexxm ; the Russian only 30 per cent. Therefore it was early apparent to Russian manufacturers that they must find some use for the residue, and from it they made lubricants.

A lubricant, itself good, may be applied in unfavourable con- ditions. Numberless experiments have been made to determine coefficients of friction, but there is still wanted a thorough appreciation of the qualities and behaviour of lubricants. These contribute to the sum of friction their own internal fluid resistance ; also a superficial resistance to the relative motion of the opposed surfaces with which they are in contact. The Author recites the following propositions. For rigid bodies friction : —

(1) is approximately proportional to the normal pressure; (2) bears some proportion to the pressure per square inch of surface, and is greater for very light, or for very heavy loads, than for some intermediate load ; (3) diminishes as the velocity increases ; (4) depends on the quality of the materials in contact ; (5) depends on the surface condition of the materials in contact, the smoother the surfaces, the smaller the friction ; (6) dej^ends on some factor of importance as yet undetermined, which requires to be known.

Friction in fluids^ (1) is proportional to the relative velocities ; (2) is proportional to the wetted surfaces along which the relative motion occurs ; (3) is dependent on natural properties of the fluid ; (4) is independent of the pressure.

To these, recent observations require to be added : —

(5) An increase of temperature lessens friction.

(6) Friction varies with the remoteness of the flviid surfaces, and the coefficient of surface friction may be greater or less than that of friction within the fluid.

The second and third propositions are accepted by everybody.

• The original is in the Library of the Inst. C.E. ' The Author does not always discriminate fluids.

408 NEW THEORY OF FRICTION. [Foreign

The first, thougli generally accepted, is disputed. The Author examines in chronological sequence the data of research, criticises methods and deductions, and ela'.jorates equations. He finds nothing irreconcilahle with the statement that friction in fluids is proportional to the first power of the relative velocities, and concludes thus : —

(1) Newton's hypothesis is very nearly correct for water; there- fore probably for other fluids, unless they are dense.

(2) The internal friction of fluids is proportional to the tem- perature.

(3) The relation of the coefficient of friction to the temperature can be ascertained experimentally, and at dift'erent velocities, but the conditions should be such as to afford data to prove the results by integrating Navier's equations.

(4) Using Poiseuille's method : —

(a) The surface of the tubes must be even ; (b) and one of the factors vaiiable, either the diameter or the length of the tubes; and the velocity of discharge under varying heads, must be carefully noted.

(5) Equation (9) would probably then enable the inner, and perhaps the superficial friction to be determined, providing the direction of flow is sufficiently straight.

The Author supposes two vertical cylinders of infinite length, having the same axis of revolution, and the annular space between them filled with a homogeneous fluid, which adheres to the opposed faces in greater or less degree. It is convenient to consider the outer cylinder as stationary, and that the inner revolves always in one direction. However narrow the annular space may be, the liquid filling it has two surfaces and a thickness. He supposes that motion is communicated from the inner rigid surface to a concentric inner sheath of the fluid enveloping it ; from that to a next outer sheath, and finally to the opposite surface. The fluid particles, and the forces acting on them, will be sjnumetrically arranged about the common axis of revolution, and the jiarticles in each imaginary sheath will have a uniform velocity always less than that of the adjoining particles in the next inner cylindrical sheath, and always greater than the velocity of those in the next outer sheath. With these assumptions, and accepting Newton's hypothesis, the ratio of the flixid friction of adhesion and cohesion to the torsional moments on the inner cylinder can be worked out. If the annular thickness of the fluid is very small compared to the radii r^ r.^ of the concentric cylinders, he obtains equation (19).

Experiments should be made with cylinders having different circumferential velocity, and especially of difl^erent diameter, and they should be paired so that no eddies or currents are set up. The Author adverts to details in which the experimental apparatus would not represent the actual conditions of engineering practice, and points oiit that the thickness of the lubricant, and the ratios of the whole group of factors he has introduced, must be experi- mentally determined.

AbstractH.] NEW THEORY OF FRICTION. 409

Equations — q = volumo of li(iiu(l flowing thrnugli in

.74. A \ \ uuit time ; temperature coustaut.

^ TTl^nr / 4. ^-^ ) ] r = radius of tube.

S ixl \ Xr I A = weiglit uf a cubic liquid unit.

, {. I = uuits leugtli.

' "^^ Ahr'^ / 4 u\ ( h = head, units length.

V = J- ( 1 4" T~ )\ V = mean velocity of flow.

o fj. I \ Ar/j ^_ yjj j^ Qf surface-friction.

SUi

/x = unit of internal friction. F = sum of frictional resistance.

(19) F = / = coefficient of frictional resistance.

1^ /^i j^ A^-i /i = coefficient of resiutance at special pres-

^ A A^ sure.

1 ^ S = effective surface of inner cylinder.

TT U, = surface-velocity of inner cylinder.

(21) / = — ^1 = coefficient of surface-friction at inner

/ /Xj /Aj \ cylinder.

( « + r^ ~\~ ^ Jp X., = coefficient of surface-friction at outer

'^i ^2 ' cylinder.

TT (U, = coefficient of internal friction.

(2\a) f — ix -^ € = mean thickness of lubricant.

ep p — load on square unit of surface-bearing.

p^ — special pressure on square uuit of

/■orw f -e ^ / Pi surface-bearing.

P

M,U,

/Qo^f _ l^2^2 / P\ U2 and U3 = special circumferential-velocities.

(3-) ~7~ = Y\~ '\/ ^ /"2 and jUj = values of coefficient of internal

/i /^3 ^3 P friction.

Summary. — Equation (21) is justified, and other results brought in accord with equations (21rt), (30), and (32); the Author finds no instance of disagreement with equation (21a). Apparent differences are, no doubt, due to variation of temperature and want of adjustment in the bearings. Equation (30) is correct for rape oil under a load not exceeding S-i, and for W. Virginia oil 51 atmospheric pressures. The Author asserts that the coefficient of friction, for parts of machinery which are nicely adjusted and copiously lubricated, is —

(1) to the coefficients of internal friction of the fluid lubricant at the tem})erature of the film; (2) to the relative velocities of the rubbing parts directly measured ; (3) to the average thickness of the layer of lubricant interposed between the surfaces ; (4) in inverse proportion to the pressure exerted on a superficial unit of the bearing surfaces. (5) At a constant tenqierature the thickness of the couch of lubricant is inversely proportional to the square root of the relative load on the bearing surfaces.

(6) Therefore (4) and (5) : at a constant temperature the coefficient of friction is inversely proportional to the square root of the load on a superficial unit of the bearing surface.

(7) The temperature of the couch of lubricant depends on the character of the lubricant, on the velocity, on the load, on the conducting power of the surrounding bodies, and the temperature of the atmosphere at the time.

For parts of machines which are improperly or insufficiently

410 NEW THEORY OF FRICTION. [Foreign

lubricated, in wMcli the accuracy of adjustment is not greater, the coefficient of friction will he larger than in the foregoing cases, and on the whole larger as the influx of lubricant is less, and the want of adjustment greater, that is, if the surfaces are rough, or the shape distorted by external pressure. If the Author's coefficients are correct, his deductions elucidate some hitherto obscure pheno- mena ; but the incompleteness of available data makes new experi- ments indispensable. The Author indicates the direction which, in his opinion, such further experiments should take. . -p

On the Critical Extension of Bodies strained simultaneously in several directions. By — Wehage.

(Mittheilun£:9n aus den Koniglichen technischen Versuchsanstalten zu Berlin, 1888, p. 89.)

In investigating the strength of the walls of vessels and other bodies which are subject to stress in several directions simul- taneously, it is usual to take the greatest extension or com- pression in any one direction as a measiire of the stress on the body at any given point and to take no notice of the extension or compression parallel with the two remaining principal axes. According to the view represented by this method of treatment, the stress on a plate subject to tension in one direction woiald be reduced if it were, in addition, put under tension in a direction at right angles to the former, and would be increased by compression in this direction.

Experience, however, shows that in drawing wire a prismatic body can endure a much greater extension longitudinally without diminution of its strength, if it is sinniltaneously under compression transversely, than when the latter is not the case.

To obtain further insight into the subject, the Author was permitted, by the Eoyal Commission managing the Technical Ex- perimental Institute, to carry out bending tests with circular wrought-iron plates, and to determine, radially as well as tan- gentially, the extensions corresponding respectively to the limit of elasticity and the breaking stress ; this was done directly with the microscope.

For ascertaining the permanent extension on fracture it was found sufficient to draw small rectangles or squares with sides of from 1 to 1 • 5 millimetre and observe the alteration of these with the microscope by the help of an ocular micrometer. In order to measure the extension at the elastic limit, the method was finally adopted of drawing wdth a double scriber on the loaded plate a pair of parallel lines, and when the load was removed a second pair parallel with the first and so close to these that the corre- sponding lines with and without a load were well within the microscopic field.

As test-pieces three circular plates 420 millimetres (16 '54 inches) diameter of Kruj^p steel were employed cut out of the middle of

Abstracts.] EXTENSION OF BODIES UNDER COMPOUND STRAIN. 411

square plates about 10 millimetres (0-39 inch) thick. From the outer jjarts of each," of these plates, four test-pieces for tensile and bending tests were prepared. Two pieces out of the four were tested by tension to fracture, in order to be able to compare the extensions resulting from simple tension with those produced by compound stresses ; the remaining two were subjected to bending tests chiefly, to ascertain if the extensions were the same as under tension ; the material was, however, so good that when bent through 180'^ no fracture could be produced.

Two of the circular plates or disks were tested by being sup- ported against a concentric ring on one side and pressed in the centre by a die on the other side, in the Werder testing-machine ; the supporting ring had an inside diameter of 323 millimetres (12 '72 inches). The disks were jwlished on both sides. Each measurement was taken four times and the mean result of the four observations adopted. The dies employed had spherical surfaces of different radii.

From the results of his experiments, as far as they go, the Author draws the following conclusions: — (1) The extension at the elastic limit of a plate bent in the manner described is only about half that occurring with simple tension ; (2) when a wrought- iron disk is subject to equal stress simultaneously in two directions at right angles to each other, rupture takes place with an extension which is only about half that due to simple tension.

According to these results, the tensile stress to which a cylin- drical shell is subject, in a tangential direction, from the internal pressure of a fluid, is increased (not diminished) by the addition of a simultaneous axial tension. n -n -o

ijr. ±C. ±>.

Landslip at Zug, Switzerland, Jtdy 5th, 1887.

(Die Catastrophe von Zug, 5 Juli, 1887, &c., &c.' 8vo. Zurich, 1888.)

This Paper is an abstract of a pamphlet of sixty pages, published at Zurich in the early part of this year, containing : —

1. The opinions of a committee of experts, consisting of Pro- fessor Dr. A. Heim ; Chief-Engineer E. Moser ; and Dr. A. Biirkli- Ziegler.

2. Historical notes of the disasters at Zug in 1435 and 1887, by Anton Wickart.

3. Abstract from the registers, regarding the distribution of charity to the sufferers from the recent catastrophe, by Albert Keiser, Secretary.

From the above it would appear that, in past times, breaches and subsidence of the banks of the Lake of Zug have frequently taken place, and notably on the 4th March, 1435, when twenty-six houses, with about sixty inhabitants, were submerged ; and on the 7th March, 1594, when nine houses sunk into the lake. It is further

The original is in the Library of the Inst. G.E.

412 LANDSLIP AT ZUG, SAA1TZERLAND. [Foreign

stated that similar settlements or subsidences have occiirred in other parts of Switzerland, even in recent times, causing the destruction of a row of houses at Morcote, on Lake Lugano, in 1863, and of the quay at Vevey in 1877.

In view of the possible recurrence of such a disaster at Zug, every precaution was taken before commencing the building of the new quay wall in 1868, to ensure the stability of the site selected ; and the professional gentlemen who were called in to advise in the matter, after having made preliminary trial-borings, sunk shafts, driven in test-piles, &c., reported favourably on the site, and recom- mended a concrete foundation on a piled jilatform, the piles to be from 8 to 10 inches in diameter, and about 20 feet long; but it would appear that the piles were driven to a greater depth than this, which indicated a loose soil below a moderate depth of firm ground, of itself clearly an element of danger, which was, however, rendered more evident by the fact that the vil)ration caused by driving the piles j^roduced such serious cracks and openings in the walls of a hoiise in the immediate neighbourhood, that it had to be pulled down. It is also stated that in laying the foundation of the present steamer landing-stage there was great subsidence of the soil ; consequently, in order to prevent a recurrence of such settlement, a new method was adopted in the new quay j^roject.

This project comprised the construction of a strand road, about 50 feet wide, along the edge of the lake, and the allotting of the space between this and the Cham road (150 feet) for building purposes.

The work was commenced by driving in two rows of piles, 2^ feet apart, the piles themselves being at intervals of 5 feet. They were driven down to 12 inches below lowest water-level, and at intervals of about 20 feet they were tied to the bank and kept vertical. The space between the rows of piles was filled with ballasting uj) to within 6 inches of the pile-heads, then came a layer of concrete on which the wall was built.

In 1884, after some 260 yards of wall had been built, cracks were observed ; and as about the same length of wall had still to be constructed, a further examination of the nature of the soil was made, and it was decided to build the foundation on a layer of hurdles and fascines, which would settle down in the bed of the lake, the whole to be protected by piling ; the wall to have a slight batter in front, and to be biiilt up to an average of about 2^ feet above high-water level ; and the w^ork was carried out in accordance with this plan in the winter of 1885-86, and at the time of the disaster the whole length of quay was nearly completed.

The first warning of danger was noted at 3.20 p.m., when cracks were observed in the quay-wall, which rapidly increased, and within a quarter of an hour brought about the destruction of a portion of the wall and the sinking of some adjoining huts, with a loss of seven lives. Gradually the disaster extended until about 7 P.M., when it was found ^that many buildings near the Lake had sunk vertically to a depth of about 25 feet.

Abstracts.] LANDSLIP AT ZUG, SWITZERLAND. 413

The general opinion then formed — which was confirmed by sub- sequent investigation — was that the subsidence was caused by the squeezing out of the soft muddy stratum lying below the top soil, by the weight of the new quay-wall and the buildings adjoining ; and this mud-stream was pushed out from 200 to 300 yards into the Lake, carrying with it the foundation-piles of the wall, and which, when free again, stood up vertically out of the water, as testifi.ed by an eye-witness (Mr. Henggeler), who also observed a remarkable rise and fall of the Lake in front of the area affected. Measures were at once taken to protect the neighbouring houses, and by midnight of the 6th Jiily all hoiises within a fixed distance from the centre of disturbance were evacuated without accident of any kind. Small cracks were plastered up, and surveyors were stationed at certain points with levelling instruments, in order to note if further movements of the ground occurred.

After the panic had ceased, a thorough investigation of the disaster was made by Engineer Miiller, by means of trial shafts and borings, and a section of the Lake-bed was taken with extreme accuracy of detail by Engineers Hornlimann and Suter.

The result of these investigations is very clearly shown by sections to scale, and from these it is seen that the top layer of earth, sand, and gravel is only from 5 to 23 feet in depth, and then muddy quicksand is met with, for depths varying from 65 to 100 feet, of almost uniform consistency, being composed of mud and sand in nearly equal volumes.

At one or two points the borings showed chalk, and as similar slips of the bank had occurred at the Lakes of Zurich, Bret, &c., which were entirely caused by the washing out of the chalk, it was at first assumed that this was the cause of the Zug disaster ; but closer examination showed this to be a false assumption, and that the squeezing out of the stratum of muddy sand was the sole and true cause of the catastrophe.

On analysis this stratum yielded 38 per cent, of clean grey quartzose sand of angular grains, with limestone, mica, a little felspar and hornblende, and traces of organic remains ; the remain- ing 62 per cent, being composed of fine sand and stone-dust, without organic remains, and from this character the stratum is supposed to have been an alluvial deposit of the Lorze delta, the whole area from Baar to the Lake of Zug being, geologically speaking, a recent alluvium of the Lorze river.

The examination of the sub-soil water showed that it had fallen but very little, even in the immediate neighbourhood of the disturbed area, and stood everywhere from 6^ to 10 feet above lake level.

The survey of the bed of the lake (for which three thousand two hundred depths or soundings were taken) shows that the slip or subsidence of the bank caused a tearing-up of the bed, and the formation of a trench about 100 yards wide, gradually diminishing in depth till, at about 300 yards distant from the shore, it was on a level with the lake bottom, and from thence onwards, for nearly 800 yards more, there was a deposit of muddy sand, varying from

414 LANDSLIP AT ZUG, STSaTZERLAND. [Foreign

160 to 270 yards in width. The direction of this trench and deposit was not in a straight line perpendicular to the bank, but in an irregular curve trending to the left. The survey of the lake bed also shows that the first deposit (consequent on the first sub- sidence at 3.30 P.M.), commenced at about 135 yards, and terminated at 460 yards from the bank, is egg-shaped, and has an extreme breadth of 270 yards ; while the second deposit (caused by the subsidence at 6.50 p.m.) passed over this for a distance of about 1,100 yards from the shore, to a point where the lake was about 150 feet deep.

After having satisfied themselves that the primary caiise of the disaster on the 5th July, 1887, was the existence of a very deep layer of muddy quicksand under a comparatively thin stratum of earth and gravel soil, the qiiestion of the safety of the neighbour- hood was considered by the experts, for borings showed that the same formation underlies the whole Vorstadt or Strand, extends for a great distance towards Cham, and, in fact, is met with in all directions. The experts recommend that the destroyed houses should not be rebuilt, but the space utilized as a park or garden, and an efficient drainage of the soil should be carried out ; and for this purpose they advise masonry drains, 2 feet 7 inches wide, and about 5 feet high, or cement pipes, 1 foot in diameter, perforated in the upper half, and furnished with the necessary man-holes at the bends or jimctions ; such pipes not to be laid directly in the quicksand, but on a plank platform 2j inches thick, to prevent unequal settlement.

In the project proposed there are three main drains, each dividing into two branches, the total length being about 820 yards, and the cost estimated at £1,600. In addition to the above proposals for the safety of the place, it is recommended that all the houses between the Vorstadt road and the lake — from the Government buildings to the tannerj^ a distance of nearly 450 yards — should be pulled down, and that a protecting dam should be built along the base of the bank, on a foundation of trees and brushwood well embedded in the lake bottom, the crown of the dam to be about 33 feet wide, and 36 feet below mean level of the lake. The cost of this work is given at £28,000.

The proposed new plan of Zug provides in the first place for the restoration of the road to Cham, and as the project assumes the entire removal of the houses in the area above specified, this road will be correspondingly widened. Two new streets will also be laid out; one commencing at the junction of the railway-station and Cham roads, and terminating in the Baar road, at a point mid- way between the Post-office and the railway-station; the other crossing this at about the centre of its length, and extending from the railway-station to the jimction with the Cham road. The inter- sections of these roads furnishes four blocks for building-sites, but it is proposed to lay one block out as a public garden.

A new landing-stage for steamers will be built, about 30 yards from the site of the old one.

Abstracts.] LANDSLIP AT ZUG, SWITZERLAND. 415

As compensation to sufferers from the disaster, the sum of £27,630 was received, and £24,675 disbursed ; the balance of £2,955 being retained by the town of Zng for payment to the owners of houses which in the future may have to be pulled down as a measure of safety.

The pamphlet contains a plan of Zug and neighbourhood, showing the area disturbed ; longitudinal and cross-sections of the bed of the lake along the line of the mud and sand deposit ; the proi)osed new plan of Zug, and a photograph of the ruins.

W. H. E.

A Folding Levelling -Staff, By H. Bentabol.

(Revista Minera, Metalurgica y de Ingenieria, 1888, p. 310.)

By the aid of five illustrations, the Author describes a new form of staff adapted either for levelling or for tacheometry. The staff is 4 metres (13 feet 1;V inch) in length, and is divided into four sections, joined to each other by means of metal hinges. These lengths fold up in a zigzag manner, so that any required number may be used. When folded up, the staff is packed between a pair of boards that serve to protect the graduations. The boards are united by means of a pair of straps passing through metal guards, and one of the boards is provided with a strap, like that of a rifle, which the staff-holder can place over his shoulder, and in this way carry the staff with his hands free. The graduations of the staff are painted on a white ground, whilst the edges and the encasing boards are painted grey. The portions of the hinges visible when the staff is unfolded are painted grey at the edges and white at the face. At the side of the first length of the staff is suspended a plumb-line, which is held in its place by a hori- zontal ring that does not allow more movement than is absolutely necessary. For holding the staff, two handles project from the sides of the second length.

For short distances, the staff is graduated in centimetres with inverted figures, lines, and dots, in black throughout its entire length, each metre being numbered from 1 to 9. The second metre is distinguished by a black dot, the third by two dots, and the fourth by three. For great distances, the staff is graduated in double centimetres, the figures and dots being black in the lower half of the staff, and red in the upper.

B. H. B.

The Alignment of a Tunnel at Stuttgart. By — Widmann.

(Zeitschrift fiir Vermessungswesen, 1888, p. 520.)

In the summer of 1887, in connection with the Stuttgart water- works, a tunnel, 471 yards in length, and averaging 32 feet in depth, was driven from the Pfaffensee through compact sandstone.

416 ALIGNMENT OF A TUNNEL AT STUTTGART. [Foreign

When lined with water-tight masonry, the tunnel was 3-28 feet wide, and 63 feet high, from invert to crown, the section cut being 8 V feet high and o.y feet wide. The axis of the tunnel was not a straight line, bi;t altered its direction at three points, the transition from one direction to another being efiected by short curves 32 feet in radius and 13 to 32 feet in length. Along the tunnel axis shafts were sunk, 54 yards apart ; they had a depth of 30 to 40 feet ; a breadth, perpendicular to the centre line, of 6 feet, and a length of 10 feet. Of this length, 4 feet was taken up by timber and ladders, leaving only 6 feet as a base-line for setting out the tunnel. Nevertheless, with the aid of a simple instrument em- ployed by the Author, the subterranean alignment was effected with such accuracy that, on piercing through, the differences obtained, within a section 50 yards in length, never exceeded 0-19 inch.

Eanging-frames were erected, 6 yards apart, on both sides of each shaft, at right angles to the centre line of the tunnel. At some distance from the shaft, a theodolite was set up accurately in the line, and this was established on the frame by means of a fine cut with a saw. A string, 0 • 04 inch in diameter, was then strained in the cuts, and after it was found, b}" repeated observation, to he accTirately in the centre line, two })lummets of the Author's design were suspended, at some distance apart, from the stretched string down the shaft. The plumb-lines are attached to l)road hooks that enable them to slide along the string and to be xised for shafts of any length. Above the hook is fixed a sight, resembling that of the miner's dial. These sights are useful as a check in determining whether the plumb-line hangs accurately in the centre line, as the plumb-line is hidden by the string stretched across the ranging frame, and consequently cannot be directly sighted. The weight of the plummets attached to the lines causes the horizontal string to be drawn down vertically, but as this takes place in the direction of the vertical plane, it has no influence on the accuracy of the alignment.

B. H. B.

Methods of Testing the Resistance of Stones, Cements, and other Building Materials. By LtON Durand-Clayk.

(Annales des Fonts et Chaussees, 6th series, vol. xvi. 1888, p. 173, 1 plate and 20 woodcuts.)

Three methods are employed in these tests, namely, compression or crushing, tension or tearing asunder, and flexion. In the first method, the test-blocks are cubes or rectangular prisms with square ends, which are placed between two plates, to which the compres- sion is applied by a lever or hydraulic press. Tensional strains are aj^plied by gripping the two T-shaped ends of the test-block in jaws which are pulled in opposite directions, by means generally of levers. Lastly, flexional tests are made on prismatic bars, whose

Abstracts.] TESTING THE RESISTANCE OF STONES, CEMENTS, ETC. 417

ends rest on knife-edges, and have a strain imposed at the centre, either by loading- a plate suspended on a knife-edge, or by means of levers. Compression is mainly used in testing stones, and tension for limes, cements, and mortars ; whilst it is not usual to test by flexion in a regular manner. These three methods of testing do not give at all identical coefficients of resistance ; and the results by any one method vary according to the form of the test-blocks.

Compression. — The crushing pressure, P, under which a material under compression gives way, divided by the sectional area S of the block, gives the resistance to compression per unit of bearing surface. The block, however, is not wholly crushed ; but it fails by the formation of cracks, and the occuiTcnce of slidings in the plane of these cracks. When a i:)rism with a square base is employed, having a height more than double the length of a side of the base, a single crack is generally produced, with its plane following a slope of about ^ to 1, which accords with theory. The compression test, therefore, measures the resistance to shear- ing, which approximates to a quarter of the mean compressional breaking strain, in the materials under consideration. When the uiatei;ial is homogeneous, and the pressure is very uniformly distri- buted, the symmetry of the arrangement leads to the formation of two cracks, starting from the sides at the top, and converging towards the centre of the block, leaving a central wedge-shaped mass which tends to split down the lower part of the block into two parts, with a central vertical crack. In other cases, the two upper cracks extend down, forming a second wedge below pointing upwards, and dislocating the sides ; or, if the uniformity of the material and pressure is perfect, the wedges become prisms. When the blocks are cubes, the cracks, starting from top and bottom, sometimes form two truncated pyramids separated by a central horizontal crack ; but more often one crack follows the diagonal, and the two cracks from the two remaining corners, meeting it, divide the block into four triangular fragments. In the case of homogeneous material and uniformly distributed pressure, frag- ments fall off from the sides ; and the two oblique pyramids freed by the cracks slide along the diagonal crack. From the above

P

observations, it is evident that the resistance to crushing, -, varies

b

with the form given to the blocks, and that the results of experi- ments with different materials are not comparable unless the forms of the test-blocks are the same. The cube is the simplest and most easily pre^mred form.

Tension. — In the earlier experiments at the Ecole des Ponts et Chaussees, the T-shaped briquettes had their inner angles rounded off by curves, and a central section of 2^ square inches; but this was modified, about 1876, by substituting straight lines across the angles for the curves, increasing the solidity at this part, but retaining the same central section. At the present day, all operators in Europe adopt the German form of l)riquette. or similar types,

[the ixst. c.k. vol. xcv.] 2 E

418 TESTING THE RESISTANCE OF STONES, CEMENTS, ETC. [Foreign

widening ont to a rounded form at each end, and with a narrow contraction in the centre, having a sectional area of f square inch.

T

The ratio 77 is only the average tensional breaking strain, and not

o

the limit of resistance of the material ; for rupture occurs when the limit of resistance is reached at any one point of the section. The strain, indeed, is greater at the outside than in the centre of the section, as indicated experimentally in the laboratory of the Ecole des Fonts et Chaussees. Briquettes of india-rubber, with lines drawn on them perpendicular to the axis, were stretched ; and the curves assiimed by the lines showed plainly the greater strain towards the outside. Two series of similar briquettes were next made with the same mixture of cement, one set having a hole left in the centre of the section, and the other set with the hole at the side; and the breaking-strain averaged 570 lbs. per square inch in the first case, and 485 lbs. in the second. Lastly, briquettes were made 1*57 inch wide; and in one set, the central portion, 0 • 78 inch in width, was made of mortar, and the other portions of neat cement ; whilst in another set, the materials were reversed. Thus the sectional areas of the mortar and cement were eqiial in both cases, but their positions altered ; and the average breaking- strain of the first set was 301 lbs. per square inch, and of the second set 261 lbs. It is proved, analytically, that the larger briquettes have a smaller proportionate breaking-strain — a well-known fact, but not previously explained. The distribxition of the tensions also depends on the form of the briquette, as is readily perceived from an inspection of the different curves assumed by the lines on the india-rubber models under the same tension. The difiereuce in the curves of the same india-rubber model, when different jaws were used, shows that the form of the jaws also has some influence on the distrilmtion of pressures. It is clear, therefore, that com- parable results can only be obtained with briquettes of the same section and shape, clasped by similar jaws, which explains the frequent apparent discrepancies in results of experiments. The advantages of the German form of briquette are, that its small size enables several to be made with the same mixture of cement, and that rupture can only occiir in the contracted central portion ; whereas, in the other forms, rupture does not always take place in the minimum central section, in which case the conditions are somewhat modified.

Flexion. — Tests by flexion are rarely resorted to, though the materials are often subjected to such strains. This is due to the necessity of interpreting the results by help of calculations, instead of obtaining them by simple observation as in the preceding methods. With a rectangiilar prism of height h, breadth h, and length/ between its siipports, loaded with a weight W at its centre, and having a weight w per unit of length, the maxima tension and compression on the bottom and top fibres respectively, are repre- sented l)y the formula, R =.— 7 r-f — '. This formula is, however,

Ahstracts.] TESTING THE RESISTANCE OF STONES, CEMENTS, ETC. 419

only applicable for materials which elongate proportionally to the tensions up to the moment of rupture. This is true for the top fibres in compression of the materials under consideration ; but the bottom fibres, being in tension, give way first ; and for some of these fibres, which have passed the limit of elasticity at the moment of rupture, the formula is no longer applicable. The modulus of elasticity of various building materials, subjected to compression, has been determined in the laboratory of the Ecole des Fonts et Chaussees ; but the tensional breaking-weight is so small that it is impossible to ascertain directly the corresponding elongations. Diagrams, however, are given, showing the results of the measure- ment of the deflection of prisms of the materials under increasing loads, where the loads are taken as the abscissas, and the deflections as the ordinates. As long as the limit of elasticity is not passed, the locus is a straight line, but becomes a curve concave upwards beyond this limit ; and it is evident from the diagrams that, for the materials under consideration, the limit is reached considerably before the breaking-strain. It follows that the resistance deduced from the formula at the moment of rupture is greater than the real limit of resistance, so that the results derived by the ordinary formula of elasticity from the flexional tests are too high. The coefficient for deducing the real limit of resistance from the calcu- lated limit, can be found by investigating the curves of flexion ; but the calculations are complicated, and the operations and measurements are delicate. Nevertheless, by reducing the dimen- sions of the test-bars, and merely measuring the breaking-load, comparable results are obtained with bars of exactly the same form ; and this method of testing is as reliable as tension-tests, and more simple, owing to the greater simplicity of the manufacture of the test-bars, and of the machinery employed for testing. The Author has made a large number of experiments with bars, 0*78 inch square, of varioiis cements, placed on supports 3 • 94 inches apart, which gave as regular results as briquettes subject to tension. Any one of the three methods may be employed for comparing the resistances of the several materials ; but none of the methods give the absolute resistance of the materials. In experimenting upon moulded materials, such as limes, cements, and mortars, it is essen- tial that the specimens should be prepared under identical con- ditions in respect of proportion of water, duration and extent of mixing, temperature, compression in the moulds, &c. ; so that really few experiments are comparable, except those made in the same laboratory, by the same persons, and with the same precau- tions. Tabulated results of the Author's experiments by the three methods are appended to the article, together with further calcula- tions and investigations with reference to flexional strains.

L. V. H.

2 K 2

420 THE TESTING OF PAPER. [Foreign

On the Testing of Paper. By N. Haselkoos.

(Znjiisky Iinperatorskavo Russkaro Teclmitchezskavo Obstchestva, 1888, No. 11, p. IG.)

After making mention of the investigations of Professors Hart- ing and Goier, and of the establishment of a series of standards (Xormalien) by Mr. Martens, at the Berlin testing-oiEce, the Aiithor enumerates the tests now employed : 1st, for the strength and stretching power ; 2nd, for the resistance to softening and crumpling ; 3rd, for the measurement of the thickness of paper ; 4th, for the determination of ash ; 5th, microscopical investigation ; Gth, determination of free chlorine and acid ; 7th, determination of the kind and qiiality of sizes.

1. The absolute strength of a paper is measured by its resistance to tearing. In machine-made paper the strength and stretching power vary according as the force acts lengthways or across ; in hand-made paper there is little difference. The Author found that in the former the diflerence was in the projiortion of 2 : 3 accord- ing to the direction of the tearing force. The stretching-power acts inversely to the strength, i.e., is greater across than length- ways. The Author gives a description of Horak's and the Harting- Resch testing-machines, the latter giving in his opinion the best

T

results. He proposes the formula a; = — Q, where T = the length

of the test-paper, P = its Aveight, and Q = the tearing-weight, for finding the absolute strength of a paper.

2. In order to test the resistance of a paper to the most varied mechanical wear, it is crumpled and kneaded between the hands. After siich treatment a weak paper will be full of holes, a strong paper will assume a leathery texture. This test also gives a rough insight into the comj^osition of a paper, much dust showing the presence of earthy impurities, while breaking up of the paper shows over-bleaching.

3. The thickness of a paper is ascertained either by measuring the thickness of a certain number of sheets, or by taking that of a single sheet by means of a micrometer or " piknometer," where the paper is placed between two rules, one fixed and the other movable, acting on a pointer showing the thickness of the paper on a dial.

4. Over 3 per cent, of ash shows the presence of clay, kaolin, heavy spar, gypsum, &c.

5. Microscopical investigation of paper aims at determining the kind and equality of fibre. For this a magnifying power of 150 to 300 diameters suffices, when, by colouring the paper with a solution of iodine, a yellow coloration shows the presence of wood fibre ; a brown coloration that of linen, cotton, or flax ; and no coloration that of cellulose.

6. The test for free chlorine and acid is rarely necessary.

7. The determination af the kind and quality of size may be

Abstracts.] THE TESTING OF PAPER. 421

made by boiling in distilled water and adding a concentrated solution of tannic acid, when a flocculent precipitate shows the presence of animal size, and by heating in absolute alcohol and adding distilled water, when a precipitate shows the presence of vegetable size. If well sized a drop of Fe2Cly on one side and a drop of tannin on the other will not permeate through the paper and form ink. The amount of moisture ought also to be determined. The Paper is fully illustrated by drawings and diagrams.

G. K.

Yield of Hydraulic Mortars. By — Bonnami.

(Annales des Pouts et Chaussdes, Gth series, vol. xvi. 1888, p. 99.)

The yield of a lime in jiowder is defined as the volume of paste obtained with 1,000 kilograms of powder. The yield of a lime or cement varies with the proportion of water used in mixing, which may be expressed in a formula, such as R4gf| = 0*84, indicating that 1,000 kilograms of the substance, treated with 460 litres of water, gave 840 litres of paste. The yield is obtained by treating 1 kilogram of the substance with water in a graduated vessel and multiplying the volume by 1,000. When water is gradually added in measured quantities, the moistened powder swells at first ; it next contracts on the addition of more water, the volume of paste reaching a minimum, called the minimum yield ; after which the increase in volume is equal to the volume of water added. The paste attains a firm consistency, representing its maximum strength, only slightly before the minimum yield is reached. Accordingly, by finding the minimum yield, the value of the paste for mortar is ascertained. If e is the percentage of water corresponding to the minimum yield, and E, in respect of the weight of the ])owder, the j^ercentage of water in the paste of a mortar, the inferiority of this paste will increase with the difi'erence E — e ; and for the same substance R„ — e = a constant which is the absolute volume of the powder minus the volume of water absorbed during the trial. The yield of a lime or cement, to fill the interstices in the sand, is the foundation of all important mortar making ; for the impermeability and compactness of a mortar is at least as important for durability as its strength. The consistency of a mortar, made with a given sand, depends upon the quantity of the paste and its fluidity. With little paste, great fluidity is necessary to give the mortar proper consistency, and cohesion is rapidly reduced by increasing the fluidity of the paste ; whilst, for a given consistency of mortar, the fliiidity of the paste is reduced by increasing the amount of powder. Thus, a lime requiring less water than another, even though the strongest of the two in a pure state, may form the weaker mortar, owing to the greater excess of water required to be added to it to give diie consistency to the mortar. For a given substance, the suitable

422 YIELD OF HYDRAULIC MORTARS. [Foreign

proportion of water increases with the fineness of the powder, and, therefore, with the amount of free lime, which is always in very fine powder. Beyond a certain fineness, the ultimate resistance of neat Portland cement does not augment with the fineness of the grinding, whereas, with the mortars made from the cement, it augments very perceptibly. Inert matters in the lime should be excluded from the yield, and regarded as merely acting like sand ; and their presence explains Avhy some j^owders give good results when used neat, and only furnish moderate mortars with the best sand. The yield of hydraulic mortars varies between 0 • 50 and 1-10.

L. V. H.

Pulverization of Claij and its Application at the Works of the Societe Arnaud Etienne and Co. By C. Bidois.

(Bulletin de la Societe Suientifique Industrielle de Marseille, 1888, p. 38.)

In the manufacture of ordinary terra-cotta, bricks, tiles, &c., the first operation is preparation of the raw material. The clay should be freed from all foreign matter which would be injurious to the production of the finished articles. The subject of this Pajier, the pulverization of clay, is not quite novel ; it has long been employed in the preparation of the materials used in the finer kinds of pottery, porcelain, Dutch tiles, and similar manufactures. But in these cases, the quantity of material dealt with is compara- tively small, the process here described is in use where very large quantities of material are used. In late years of keen competition many means have been used to cheapen the cost of production, such as the use of inferior clays, and the pushing to the utmost of the machines used in producing the finished articles. But the i^rocess in use at the above-mentioned company's works, admits of the use of clays hitherto held as unfit for the production of good sound finished articles, and thus lowers the cost of production by enabling cheaper raw material to be used, and in addition to this imjDortant point, it also reduces the quantity of faulty articles turned out. In the branch of the ceramic art which has for its object the pro- duction of bricks, tiles, terra-cotta, and other artificial building materials, the low cost of production necessary renders it impera- tive that all processes be as simple as possible. The regularity of the chemical composition of the clay does not seem to be of great importance, for although all clays are very similar in chemical composition, they do not always appear alike, perhaps the differ- ence is molecular. It is very important for the clay to be worked up into a perfectly homogeneous paste, otherwise it is impossible to turn out sound finished products. With respect to clay prej^ared for the manul'acture of liricks, &c., plasticity and homogeneity may be considered as equivalent terms. The object of the Paper here quoted is to show that by pulverizing the clay in a dry state, an

Abstracts.] PULVERIZATION OF CLAY, 423

almost perfectly homogeneous paste can be made. If the moulded articles become defective in drying, it shows that in the piece there are parts of different density, which cause unequal contraction. Failures in burning show the presence of empty spaces or cavities in the piece. A method is then described of preparing the clay by passing it well moistened through a series of rollers ; this machine has seven rollers, three pairs placed vertically over each other, and the remaining roller engaging with one of the lower pair ; the rolls are 19f inches long and 11^ inches in diameter. The upper pair revolve at sixty revolutions per minute, and their surfaces are about y\v inch apart ; the second pair revolve at from eighty to eighty-five revolutions per minute ; their siirfaces are about j\v inch apart ; the other three rolls receive the clay and force it through a perforated steel plate ; it is stated that the action of this machine is to give the worked up clay a laminated structure. With clay prepared by exposure to frost during the winter, and then passed through this machine, fairly good results have been obtained ; but the process is slow and costly. The idea which led to the use of dry pulverization was to utilize beds of clay containing stones, chalk, &c. The effect of chalk is not deleterious if the quantity in the clay does not exceed 13 to 15 per cent. Such clay, thoroughly pulverized so as to disseminate the chalk equally through the mass, yields good results. The pulverization is effected very successfully by edge-runners ; the pressure must not be too heavy, as clay is always slightly unctuous. The edge-runners weigh 2,204 lbs. each, and run equidistant on both sides of a vertical spindle ; the pan in. which they revolve has rectangular holes in the bottom, in which perforated plates are fitted, and scrapers are provided to turn the clay into the track of the rollers ; the powder, as it falls through the plates, is taken by a screw conveyor to a polygonal reel covered with No. 40 wire cloth, and the overtails are returned to the pan by an elevator ; the pulverized clay is then damped by an automatic arrangement, and joasses on to a kneading machine, from which it passes out ready for use.

Some of the principal advantages of pulverization are stated to be as follows : —

Eaw material of less value can be successfully used.

The bricks, tiles, &c., made from pulverized clay are stronger than those made in the ordinary way ; experiment shows the total resistance to rupture of ordinary plain tiles to be from 210 to 280 lbs. when made in the usual way, and 343 lbs. when made from pulverized clay ; by reason of the extra strength fewer tiles are broken in transport.

The appearance of the tiles, &c., from pulverized clay is better than those made in the ordinary way.

Owing to the uniformity of the prepared clay by pulverization, especially when pure, the ground product may be passed through comparatively coarse sieves, so that the purer the clay the larger the yield of the grinding machinery. Some remarks are made on the cracking of terra-cotta by frost. The Union Ceramique ap-

424 PULVERIZATION OF CLAY. fForeign

pointed a committee in 1886 to determine the causes of frost-cracks in terra-cotta ; they came to the conclusion that terra-cotta does not crack by the action of frost, if the burning is jDushed just to the point of fusion or vitrification, pores may be formed, but they are then impenetrable to water. Again, it Avill not crack by frost if the pores have communication with the surface by sufficient channels. Granular structure is favoiirable, but laminated structure is not. For these reasons fabrics made by comjjression are best, and pulverization of the clay favours the granular structure.

H. H. P. P.

Stone-cutting and Quarrying hij Wire.

(L'Inhistrie Moderue, 1S88, i)p. 203, 218.)

Al)out ten years ago a Belgian company was formed to work the old Eoman marble quarries of Schemton in Tunis. Though the marble, of various colours and structure, was estimated at more than 253,165,800 cubic feet, working was discontinued on account of the expense. Lately the comjiany has been reorganized to work the quarry by the " Helicoidal-wire " system, by which not only can the blocks be subdivided, but also the marble extracted from the mountain side.

Power from a 60-HP. engine is transmitted by teledynamic cable to the highest point of the quarry, whence it is distributed to the sevei'al working jilaces by three helicoidal cords, each com- posed of three ^iteel wires, twisted spirally, and running at the rate of 14 feet 9 inches per minute. The cord cuts the marble into slabs by penetrating into the rock at the rate of from 5 to 5^ inches per hour for hard marble, sand and water being allowed to flow constantly into the groove. By changing the direction of the cord, by means of pulleys with adjustable axes, their bearings being fed down as the stone is penetrated, the same cord can be made to serve several working places. The marble, cut to the required dimensions, without being touched by the chisel, is brought down in tramways to a workshop, where the blocks may be still further subdivided by the helicoidal wire so as to be reduced to the required dimensions. The Avorkshops are connected with the Bona-Guelma Railway by a tramway 2.V miles long made by the comjiany.

The installation of the Societe Anon^aue Internationale du Fil Heliq'oidal in the grounds of the Brussels Exhibition of 1888, exemplifies the principal applications of this new method of work- ing quarries. The endless wire cord is sent by the driving pulley to a tension truck at the end of the yard, and, guided by pulleys with universal joints, is diverted at given points for sawing a mass of concrete and a block of marble, while there are also the following appliances : — A frame, in which the usual blades are rejdaced by cords i'or sawing slabs ; a finishing ajjparatus ; and a drill, driven ly teledynamic rojic, for sinking the shafts by which

Abstracts.] STONE-CUTTING AND QUARRYING BY WIRE. 425

the cord carriers are introduced, the whole being driven by a 14-HP. engine.

In most quarries, especially those of marble, it is less important to exti'act the greatest quantity of stone, than to obtain blocks of the form and size desired with as little waste as possible ; and this is ac- complished in a high degree by the helicoidal cord ; while, manual labour being superseded by a regular mechanical operation, there is no need for skilled workmen, but only a few boys to tend the appa- ratus. A still further saving of lalwur is effected by the mass being subdivided into blocks of the desired size on the spot where it is quarried.

The rapidity of the operation naturally depends on the hardness of the stone ; but it may be put roughly at ten times as great as that by old methods, while concrete, and such rocks as cannot otherwise be worked, yield to the helicoidal cords. At the Exhi- bition, the same cord which sawed a block of marble also cut simultaneously a mass of concrete composed of quartz and flint peljbles.

Quarries in France, Algeria, Tunis, Italy, Spain, Germany, Eiissia, and Finland, have been provided with the new apparatus, while it is exclusively used in the marble qi;arry of Traigneaiix, near Philippeville, Belgium. Here the trench 60 centimetres, or nearly 2 feet, wide, which was formerly, as it is still generally in other quarries, made by hand, is superseded by vertical cuts with the helicoidal cord on all faces not detached, and a horizontal cut underneath the mass to be extracted. If the mass be not detached on any side, it is necessary to run two cuts 2 feet apart along one of the faces.

In order to permit the cord to descend, it is also necessary to sink shafts at all the angles of the mass where not detached, in order to receive the pulley carriers ; and this work is now per- formed mechanically by the drill invented by Mr. Thonar, at the same time preserving the cores for use as columns. It is usual to make three contiguous shafts, and break down the intervening angles ; but the number and size of the shafts may be made subservient to the diameter of columns most in demand. The drill, driven by teledynamic cable, requires from 3 to 3\ HP., and descends at the rate of about 10 centimetres (4 inches) per hour in Belgian marble.

The endless helicoidal cord, composed of three steel wires, varies from 100 to 300 metres in length, and receives its longitudinal motion from a fixed engine, the requisite tension being preserved by a weighted truck on an incline. The downward feed is given by screws in the pulley carriers, turned either automatically or by hand ; and the helical twist of the cord causes the rotary motion, which is demonstrated by the even wear of the wires. The cord serves as a vehicle for conveying the sand and water, the former of which is the real agent in cutting the stone.

The diameter of cord found most suitable for quarrying is 5\ to 6 millimetres, or less than a ([luirter of an inch, running at a speed

420 STONE-CUTTING AND QUARRYING BY WIRE. [Foreign

of 4 metres a second, while smaller diameters and quicker speeds are adopted for siiLdividing the masses. A cut of 10 to 12 centi- metres, or more than 4 inches per hour, is obtained for lengths of 3 or 4 metres in Belgian marble. In Quenast porphyry, which it had not before been found possible to saw, a cut of ci or 4 centi- metres, or from 1 to 1 .V inch per hour, is obtained.

For quarrying, 2 HP. is found sufficient. If the cord should break, it is readily spliced; and a cord of average (150 metres) length will produce -from 40 to 50 square metres of sawn surface, before wearing out, when it may be used for fencing. The sawn surface, plane if not smooth, is readily finished by the application of an amalgam of emery with lead, tin and antimony, used in a machine like that for polishing glass.

The latter of the original articles is fully illustrated, and further Papers relating to the system have been presented to the Institution, and can be consulted in the Library\

J. W. P.

The Theory of Jointed Bow-Girders. By E. A. Werner.

(Journal of the Franklin Institute, 1888, ]\Iay to October, pp. 387 et seq.)

Under this title the Author presents a series of mathematical Papers, having for their object a complete investigation of that class of structure which is commonly known as the jointed arch or jointed arched rib ; and throughout the Paj^er the terminology employed is often different from that which is familiar to English engineers.

However, the conditions assumed are, practically, that the structure shall be hinged at the two abutments A and B, and also at some intermediate point C, which may be conceived to have any position between A and B, but is assumed by the Author to be situated at the centre of the span. Under such conditions a siispended system might be included as well as an upright arch ; but the Paper deals only with the upright form, in which the height of the central hinge C above the chord line A B is denoted

A certain curve passing through the three jioints A, C, and B, is designated the "line of thrust," and is defined as the geometrical locus of the points of application of thrust in the structure. In the Paj^er it is treated as the curve about which moments are reckoned ; and the position of any j^oint m in the curve is defined by the rectangular co-oi-dinates a:,,, and y„„ the point A being taken as the origin of co-ordinates.

At the joints A, C, and B, the bending moment is always zero ; but between the points A and C, the moment may have any value between -|- a and — a, according as the curve is raised high above or sagged deeply below the chord line. If v/ = -j- a , then M = — a, and vice versa. -

Abstracts.] THE THEORY OF JOINTED BOW-GIRDERS. 427

Continuing the cliord line BC until it intersects, a vertical line AD erected at A, the triangle ADC is called the " deciding triangle," and it is shown that whatever may be the manner of loading, the bending moment can only be zero when the curve lies within the deciding triangle. It is also shown that, under the dead load, the moment reaches its maximiim at the point where the curve runs parallel to the chord AC ; and the point may be found by drawing a tangent to the curve parallel to AC.

By the aid of these and similar theorems, the Author examines the effect of varying the form of the rib, or of the " line of thrust," and illustrates it by hypothetical examples, in which the curve is supposed to have various imaginable shapes.

Another branch of the enquiry relates to the effects produced by varying the position of the load ; and in discussing the moments and stresses in the half-rib AC, the deciding triangle again comes into requisition ; the ordinates of the line AC being denoted by p, and those of the line DC by p'.

The load is divided into three portions, of which

G^ = the load between the points A and m, G-2 = „ „ „ m „ C,

while (/,, f/2 and ^3 represent the abscissae to the points of applica- tion of Gi, G2 and G3 resj^ectively, and I rejjresents the total span AB.

Then, by analytical methods, it is shown that the moment at the j)oint m will be —

M,.= G,j7,(^'

p - y

2/ + G3 G - 173) (^)^^ (i«)

The second term is positive or negative according to the value of rt,, and it becomes zero when q., = n,,, = ^ ^ — ^ . The moment can be expressed by —

M^ =

1 X

— {g,n - x) Gi g, + G2 {g,n - g.,) — +

_y»i Urn

2 g,„ — l\ X

+ Gs(l- Os) i^-^'f^)

fjoij

(Ih)

and the expressions (la) and (16) are regarded as the pivot equations of the whole theory.

In structures whose " lines of thrust " lie outside of the deciding triangle, or coincide with either side of that triangle, the maximum moment occurs when the whole structure is loaded. But when the line of thrust lies inside the triangle ADC, the maximum positive

428 THE THEORY OF JOINTED BOW-GIRDERS. [Foreigu

moment is attained by applying the whole of Gj and a portion of G., extending up to a certain point where (/., = g,„ ; while the gTcatest negative moment is obtained by removing these loads and applying all the load that can come upon the remainder of the span. The critical point is found by drawing a straight line through A and m, which being continued will intersect the line DC at the point in question.

The method pursued is analytical throughout, and the geometrical ex]iression of each conclusion arrived at comes out as an incidental corollary.

By similar methods it is shown that the greatest shearing force occurs when the load extends to a certain point whose abscissa is (J., = r/s ; and gs may be found by drawing through A a line })arallel to the tangent of the curve at the point x,„ «/„„ which line will intersect DC at a point whose abscissa is rj^.

The theory is applied to arched ribs consisting of an upper and lower member converging together at the joints A and C, and united by diagonal bracing. The horizontal and vertical reactions of the abutments are determined in the usual form, and the stresses in the principal chords and in the bracing are worked out in detail for certain examples.

T. C. F.

Experiments on a Neiv Form of Strut. By C. L. Strobel, M. Am. Soc. C.E.

(Transactions of the American Society of Civil Engineers, vol. xviii., 1888, p. 103.)

In the new railway bridge over the Mississippi at Kansas city, a length of 1,545 feet consists of a double track carried on iron trestlework, at an average height of 45 feet ; and for all the columns and struts in this part of the structure, the Author designed a new form of cross-section which is believed to combine many practical advantages. The strut may be described as consist- ing of two flanges, or legs, united by a single central web of lattice bracing ; while each flange consists of a pair of Z irons riveted foot to foot with the lattice bars sandwiched between them.

To ascertain experimentally the strength of this form of con- struction, fifteen columns were tested at the joint expense of the Eailway Company and the Keystone Bridge Company, by means of the hydraulic testing-machine at the works of the latter company, and under independent supervision.

The Z irons employed in the structure were 3 inches by 5 inches by 3 inches, with a thickness of -}! inch ; but the tests were made with columns of smaller dimensions, the Z irons being 2 '; inches by 3 inches by 2.^ inches with a thickness of -{^^ inch. The columns were placed horizontally in the testing-machine, with the lattice bars in a vertical jjlane, and the weight of the column was counterbalanced by an upward pull applied at the centre. In

Abstracts.] EXPERIMENTS ON A NEW FORM OF STRUT.

420

every case, however, the strut gave way by flexure in the lateral direction, which was theoretically the plane of easiest flexure. In this direction the radius of gyration is said to he 2 • 05, the sectional area varying, however, from 9'IG to 10*10 square inches in the several struts tested.

The breaking weight per square inch, as shown by the tests in the case of the longer columns, is greater than that given by

Itaukme s lormula, viz., }) = —

1 +

L^

and to express it, the

46,000 - 125 -, as

36,000 r-

Anthor proposes the empirical formula p

I

applicable to struts whose length is more than ninety times the radiiis of gyration ; while for all shorter lengths he takes a con- stant strength of 35,000 lbs. per square inch.

Taking the average of the two or three experiments made with square-ended columns of each length, the results were as follows : —

Breaking Weight in Pounds per Square Inch.

Length.

By Experiment.

By Rankine'i Formula.

By Author's Formula,.

Feet Ins.
10 Hi	64	35,700	32,300
15 0	88	35,600	29,600	35,000
19 Of	112	33,750	26,700	32,200
22 0	i 129	30,300	24,600	29,900
25 0	t 146	28,170	22,600	27,750
28 0	■ 164	27,770	20,600	25,500

The Author considers that the customary factor of safety, in bridge compression-members, is on the average about 4*35; and adopting a slightly higher factor, he proposes the following ex- pression for the allowed stress in square-ended Z-iron columns, viz.,

10,600—30 - , for lengths exceeding 90 radii; and 8,000 lbs. per

square inch for struts whose length is equal to or less than 90 radii.

As compared with a strut composed of two channels with two planes of lattice bracing, the new form of construction has the advantage of saving half of the lattice bars and half of the riveting, while the lattice, being in a more protected position, is less liable to damage in handling, and the form of cross-section offers very great facilities for connection with other members of the structure. In addition to this, the strut is not weakened by rivet-holes at the outer edges, where the material is most serviceable for resisting flexure.

When this form of section was first adopted by the Author, Z irons were not rolled in the country, and it was necessary to

430 EXPEKEMENTS ON A NEW FORM OF STRUT. [Foreign

prepare rolls specially for the purpose. The section was rolled without any dilficnlty in the same way as an ordinary angle-iron, the line of rolling contact being diagonal to the section ; and, notwithstanding the cost of preparing the rolls, the section was furnished by the makers at a lower rate per ton than channel iron. The material was manufactured by Brown, Bonnell and Co., of Youngstown, Ohio.

T. C. F.

Sighivay Bridges of Iron and Steel.

(Journal of the Association of Engineering Societies (U.S.) 1888, p. 451.)

The manner in which highway bridges, of iron or steel, are designed and contracted for in America, formed the principal subject of a discussion which was held by the members of the Engineers' Club of Kansas City ; and which followed upon a Paper by Mr. J. A. L. Waddell, Assoc. M. Inst. C.E., of which an abstract only is given in the Journal.

The members were unanimously in accord with the writer of the Paper as to the urgent need of reform in the present methods of designing highway bridges ; whose frequent imperfections were attributed, in great measure, to the system under which bridges are let to competing manufacturers, and contracts taken upon their own designs.

Under the system which, by law, is comiDulsory^ in the State of Missouri, the construction of a new county bridge is let by public auction, or " outcry," to the lowest bidder ; the purchasers being represented by certain highway officials, who are not supposed to have any technical knowledge of bridge designs, and are unassisted in general by any engineering adviser.

The abuses of such a system, and the mutually opposed intrigues of buyer and seller, are adverted to in detail ; while the inadequacy of the means which are supposed to guarantee the safety of the bridge, is demonstrated by reference to the actual results of the system. Numerous failures of highway bridges are quoted (and they are said to occur every month), while it is also stated that the iron bridges built to-day, in the West, are often more unsafe than those built five years ago.

Under the stress of a competition, stimulated on both sides, the strength of the bridge sufiers in many ways. The design is prepared with a theoretical strength lower than it ought to be — the scantlings of all members that do not appear in the strain- sheet are cut down to the lowest possible figure — and when the design has been accepted in this faulty form, the strength is further reduced in execution by paring down those portions of the ironwork which are not exposed to view.

For the evil results of the system, the Paper suggests the following alternative remedies : —

Abstracts.] HIGHWAY BRIDGES OF IRON AND STEEL. 431

1. State inspection.

2. The formation of an association of highway bridge-builders pledged to the adoption of a standard specification.

3. The employment of a bridge engineer to decide l)etween competing designs, and to inspect the bridge after completion.

4. The employment of a bridge engineer to prepare the design. The second alternative is examined at considerable length and

elaborated in detail ; and the Paper contains a complete set of specifications, which are intended to cover the whole gronnd of highway bridge designing, and are suggested as forming the standard that is required for the object in view.

The salient points of the specification, as described, refer chiefly to those questions which commonly form the main subject of American bridge specifications, the first being the live load which the bridge shall be designed to carry.

For this purpose highway bridges are divided into four classes, according to the locality and the kind of traffic that is expected ; the floor load varying from 65 lbs. to 100 lbs. per square foot of area, while the load on each truss is never to be taken at less than a certain figure, varying from 800 lbs. to 1,800 lbs. per lineal foot in the different classes. Certain concentrated wheel-loads are also specified as affecting the joists.

The working stress per unit of sectional area is next treated, and is made to vary, not according to any general formula, Ijut according to a specified list of members, which appears to have been drawn with special reference to those types of girder-bridge that are most commonly used in America, and comprises such members as " hip verticals," "beam-hangers," &c.

The discussion evinced a general approval of the proposed specifications, qualified by some criticism directed to the small details above mentioned, but was chiefly concerned with the broader question of finding some means by which the general safety of highway bridges might be adequately guaranteed. The highway commissioners were powerless to obtain the requisite guarantee, for they would probably be satisfied with a factor of safety of two, if it were explained to mean that the bridge would be strong enough to carry a load twice as great as any that would come upon it.

The discussion, however, did not evince any general confidence in the proposed association of bridge-builders, the preference being given either for the employment of an independent engineer, or for State intervention ; and it was urged that whatever measures were taken, they should apply to railway bridges also, which stood equally in need of a thorough inspection ; while it was further stated that the question had recently been widely discussed among the engineering societies of the several States, and that a memorial to the Legislature would be generally supported.

Incidentally it was pointed out that the question of danger is sometimes only a comparative one ; for among the sparse com- munities of the West, the choice sometimes lies between having a

432 HIGHWAY BRIDGES OF IRON AND STEEL. [Toroigii

very cheap l)ri(l(2;e, and having none at all ; and in siich cases it is only a question I'ctween the danger of crossing a weak structure, and the danger of fording the stream.

T. C. F.

InsjJedion and Maintenance of Railway Structures.

(Transactions of the Aniericnn Society of Civil Engineers, vol. xvii., 1887, p. 259.)

At the request of a number of niemhers of the above Society, a disciission upon the Inspection and Maintenance of Eaihvay Struc- tures was announced, and a circular embodying seme leading points that had been suggested was issued to the members, with a request for discussion.

The questions suggested were as follows : —

1 . What measures, legal or other, can be taken to insure a proper inspection of railway-bridges ?

2. What is a j^roper bridge-inspection ?

3. Should there not be a standard sj^ecified rolling load, much heavier than as now generally used, and a siiecified engine wheel- base for rolling loads ?

4. Is it not expedient to adopt a standard bridge-floor?

5. Should not bridges of small span lie made strong enough for a buckle-plate floor and a continuous coat of ballast on the bridge? and if so, up to what span should this apply ?

6. Should not a safety-giiard (Latimer) be iised at all openings over a certain width?

7. Should there not be required either overhead crossings, or, in their place, interlocking ajiparatus with derailing switches ?

8. Is legislation as to any of these points, or as to any other you may suggest, expedient ? and if so, what sort of legislation ?

In reply to these questions, wi-itten communications were received from fifteen members, including many well-known bridge- engineers, in addition to an oral discussion by some others.

In regard to the first and eighth questions, it was the general opinion that an inspection of all existing railway-bridges is de- sirable, with the object of ascertaining whether the bridge is adapted to the purpose for which it is being used, and esjiecially for the continually-increasing rolling load, and whether it is being properly maintained. It was also repeatedly urged, and on no part contravened, that the inspection, to be efiective, must be made by a skilled engineer who is an expert or specialist in bridge-con- struction, and that a periodical inspection by practical foremen unskilled in the principles of design, is insufficient ; one member stating that the Ashtabula Bridge was habitually inspected in such a manner, and uniformly reported to be " all right," up to the date of its collapse, although any bridge-specialist would have at once detected that it was " all wrong."

There was, however, some difference of opinion as to the authority under which the inspection should be made. Of the fifteen cor-

Abstracts.] INSPECTION OF KAILWAY STRUCTUEES. 433

respondents, eight were in favonr of invoking legislative control in one form or another; some recommending a Government Com- mission to deal with the whole question, while others advised that the inspection should be compelled by legislative authority, but not undertaken by it. Two or three of the remaining members assumed that the insi^ection would be taken in hand by the railway companies, one of them being prepared to recommend Government interference in the event of their failing to do so. Three cor- respondents were expressly opposed to any legislative interference, chiefly on the ground that it would entail a divided responsibility, and that the companies should be made directly and solely re- sponsible for all bridge accidents. One member was of opinion that the desired object might be attained by the united action of the engineering societies.

It was suggested that the inspection should include all designs of bridges about to be erected, and highway- as well as railway-bridges.

As touching upon the second question, it was repeatedly urged that the detailed plans, calculations, and strain-sheets for every bridge should be filed on record, and mamerous suggestions were made as to the particulars and form of the official inspector's report.

Upon the third question, it was generally admitted that the prevailing tendency is to increase the weight of engines and rolling- stock ; while it was generally recognized that a standard which would be applicable to trunk lines, would be considerably beyond the present requirements of many branch lines. In view of this difficulty, the balance of expressed opinion was in favour of adopting several different standards for different classes of road, some members recommending that each line should be licensed only to carry certain defined loads corresponding with the safe capacity of its bridges. There were, however, one or two opinions in favour of applying a universal standard, and one or two against apjilying any fixed standard at all ; while one member would apply it only to trunk lines of railway.

As regards the fourth question, there was much diversity of opinion, some advocating a uniform type and standard of floor- system, while others would admit several types, and some were opposed to any standard at all.

A greater unanimity prevailed in regard to the employment of a buckle-plate floor and covering of ballast, eight or nine members being distinctly opposed to it, while only two or three favoured its adoption in small bridges ; some, however, being of opinion that the dead-weight of the ballast is useful in small bridges, but that it is better to carry it upon a planked and creosoted platform.

In conclusion, there was an entire concurrence of opinion as to the necessity of using rail-guards on all excejDt small-span bridges ; some members preferring an inside guard-rail brought to a point at each end of the Ijridge, while others preferred a high timber guard outside the track, and some advised the employment of both methods. The only objection to a high guard was that, if placed within 18 inches of the rail, it would be fouled by the snow-])lough,

[the INST. C.E. VOL. XCV.] 2 F

434 INSPECTION OF RAILWAY STRUCTURES, [Foreign

and if placed at a wider distance it would be comparatively ineffective. To meet this, it was suggested that the plough might he provided with an aiitomatic lifting apparatus, so as to clear the rail-gaiards on bridges.

The remaining questions and remarks had reference to matters lying outside the scope of the main enquiry.

T. C F,

The Garahit Viaduct} By G Eiffel.

(Compte rendu de la Society des Ingenieurs Civils, July 1888, p. 55.)

This is an account of the calculations of strains, and of the sections of the principal parts of the structure. Beginning at the Marvejols side, and ending at the Neussargiies side, the dimensions are as follows : —

ahcdefgC 70-09 I 51-80 I 55-50 | 55-50 | 55-50 | 51-80 | 24-64 ] 12-32;

C h i h I

: 12-32 I 24-64 | 51-80 | 51-80 | 45-91 metres.

To the left of a, and to the right of /, there are masonry viaducts ; h, c . . . .k are iron piers. Between a and / is a continuous girder of five spans ; between / and i a continuous girder of three spans ; and between i and I a continuous girder of two spans. Between e and /i is a moon-shaped parabolic arch, of 165 metres (541 feet) span, and 56-850 metres (186-5 feet) rise in the neutral fibre, so that the piers, /, g, h, i, stand on the arch, C being its centre ; piers g and h are therefore very short, / and i a little higher, and e and k are 60-736 metres (200 feet) high. The platform is for a single line of railway, and the width of the arch at the crown is 6 • 28 metres (10-6 feet), at the base, however, it is 20-00 metres (65-6 feet), and it is anchored down to the masonry by two bolts of 90 milli- metres (3i inches) diameter at each of the four supports. The calculation of strains is divided into four parts, viz., 1, the calcula- tion of the continuous girders ; 2, the calculation of the piers ; 3, the calculation of the arch under the vertical loads, and under the change of temperature ; and 4, the calculation of the arch under the wind-pressure.

1. Calculation of the Continuous Girders. — For the calculation of the bending-moments, and therefore also of the strains in the flanges, the moving load is applied as one unbroken train of the length of either one span or of two spans, the weight of the train being assumed as evenly distributed, viz., 4,800 kilogTams per lineal metre = 1-44 ton per lineal foot. Clapeyron's formula is used for the calculation of the moments over the piers, and

* A general account of the Garabit Viaduct will be found in tLe Minutes of Proceedings, vol. Ixxvii. p. 398r

Abstracts.] THE GARABIT VIADUCT. 435

diagrams show the parabolas of moments derived from them under the assumptions as stated. The increase of the moments occurring over the piers and in the middle of the spans, when more than one train of the stated length is applied, is not mentioned, as also the increase occurring between the pier and the middle of the S2:)an when these trains are broken up in parts. The eifect of the change of height of the tall piers upon the moments is also not stated. In the calculation of strains in the trellis-web, the actual train on wheels has served for the construction of the polygon of maximum shearing-forces, assuming at first that the girder is not continuous, and the correction for the continuity is then made by adding a uniform quantity derived in the well-known manner from the moments over the piers, which, however, were calculated with the evenly distributed load. It is mentioned that this is not mathematically correct.

Each flange of a girder is composed of a vertical plate 600 x 15 millimetres,^ two angle-irons 100 x 100 X 12 millimetres, and several flange-plates 500 millimetres wide ; the depth of the girder between the angle-irons is 5*16 metres. In the case, for example, of an aggregate thickness of flange-plates of 42 millimetres, the moment of resistance of the section of the girder (without having regard to rivet-holes) is stated to be 0-224126 metre^ and the corresponding bending moments calculated as above, 1,296,478 metre-kilogram. This gives, as stated, a strain of 5 '77 kilo- grams per square millimetre (3*66 tons per square inch). A panel of the web consists of two diagonals. The strain in each diagonal is stated, for example, to be 82,500 kilograms, the gross section of the tie, consisting of a web-plate, two angle-irons, and a flange- plate, to be 15,200, the nett section 14,257 sqiiare millimetres, and the strain in the tie accordingly 5*78 kilograms per square milli- metre (3*68 tons per square inch). The gross section of the striat is, in the same panel, 14,168 square millimetres, and the strain accordingly 5 • 85 kilograms per square millimetre (3 • 72 tons per square inch).

In the end verticals of the girders it is noticeable that liners are calculated as parts of the sectional area. In this way one end- vertical has a gross sectional area of 47,960 square millimetres. The stated pressure is 255,390 kilograms, and the strain 5 • 32 kilo- grams per square millimetre (3*39 tons per square inch). The admissible strain is stated to be 6 kilograms per square millimetre (3 • 82 tons per square inch).

The cross-girders, which have a support in the centre by means of a cross-bracing between the main girders, are strained to 4-08 kilograms per square millimetre (2*53 tons per square inch), the moment of resistance of the gross section, inclusive of the web, being taken; the rail-bearers, which are also plate-girders, have in the same way a strain of 5*76 kilograms per square millimetre (3 '66 tons per square inch).

> The millimetre = 0-039 mch.

2 F 2

436 THE GAKABIT VIADUCT. [Foreign

2. Tlie Calculation of the Piers. — The height of the tallest piers is about 60 metres, and their width transversely 5 metres at the top, and 18 -SI metres at the hottom. The wind-pressure is taken at 150 kilograms per superficial metre on the structure with the train, and 270 kilogi'ams withoiit the train. The surface of the girders exjiosed to the wind is calcialated according to Nordling's method, by taking one top and bottom flange and two webs, the wind blowing horizontally ; to this is added the thickness of the platform structure, and so much of the body of the carriages as is not protected by the top flanges, the spaces between the carriages, however, being ignored ; this gives for the train 2-20 — 0* 60 = 1-60 square metre per lineal metre; 3 '70 square metres per lineal metre being the surface of the girders and platform. The surface of the piers is calculated by taking the surface shown in the elevation of the bridge twice.

The calciilation of the strains from these forces is too simple to be referred to here. The gross sectional area of the four inclined posts, having a tubular section, varies between 33,000 and 49,200 square millimetres each ; the greatest corresponding pressures are 182,373 and 288,153 kilograms. This gives the strain jier square millimetre, viz., 5 • 53 and 5 • 85 kilograms (3 "52 to 3 • 72 tons per square inch). The strains in the diagonals are, in the same way, 4 • 84 and 3 • 08 kilograms per square millimetre (3*08 to 1 • 89 tons per square inch). The posts terminate at the bottom in a bed- plate one metre square, and 15 millimetres thick, resting on a stone 1 • 60 metre square, and 0 • 5 metre thick. The compressive strain for the maximum pressure between bed-jDlate and stone is

288 742 therefore ' = 29 kilograms per square centimetre (26*49

tons per square foot), and that between stone and masonry beneath 11 kilograms per square centimetre (10-09 tons per square foot).

3. TJie Calculation of tlie Arch under the vertical Loads. — Four cases are calculated, viz., 1, when there is no train-load on the arch ; 2, when the load extends between the piers e and h ; 3, when it extends between / and i ; and 4, when it extends between e and the centre. The pressures on these joiers are taken from the previous calculation of the continuous girders, and the loads thus applied to the arch in the four points, /", g, h, i, are stated in a Table.

The arch has thirteen verticals (thirteen and a half panels), between abutment and crown, and its dead weight is considered as acting in those thirteen verticals. Then the theory of the arch is explained. As it is pivoted at each springing one eqiiation suffices for the calculation of the horizontal strain ; one of the pivots is imagined to move freely horizontally under the loads when there is, of course, no horizontal strain ; the movement Ax of the pivot is cal- culated ; then the horizontal strain Q is considered as acting alone, withoiit the loads, the corresjionding movement A.r is then calculated and added to the former As?, and the total is put equal to nil, because, in fact, there is no movement. From this equation Q is found. The

Abstracts.] THE GARABIT VIADUCT. 437

members in the equation constituting the portions of which Ace is composed, are of the usual form, and consist of three kinds, viz., 1, horizontal movements due to forces acting in the neutral axis of the arch ; 2, the same due to forces acting at right-angles there- with ; these forces prodxice the strains in the diagonals, and the expressions therefore contain, besides the angle of the tangent on the neutral axis with the horizon, a, also functions of the angle of the diagonals with the neutral axis, /3 ; 3, the same due to the bending moments. The first and the second kind of members contain the sectional areas of the booms and the diagonals respectively, while the third kind contains the moments of inertia of sections through the whole arch.

In the calculation of the strains from a change of temperature (30 CentigTades above and below the mean), the sum of the members with Q is simply put equal to the elongation of a bar as long as the chord of the arch. The strains in the booms on this account vary between 0 and 0 • 7-1: kilogram per square millimetre (0 • 47 ton per square inch).

4. The Calculation of the Arch under the Wind-pressure. — The acting forces are here computed in the same manner as they are in the calculation of the high piers, but with regard to those acting on the arch itself, it is assumed that the leeward booms are protected by the windward booms, while both trellis-webs are exposed. In this way the forces acting upon the thirteen and a half jDanels of the half arch vary between 22*30 and 35*38 square metres, multiplied with 150 or 270 kilograms, as the case may be.

In the case of the symmetrical distribution of the loads and the wind-pressure upon the structure — the only case considered — there is at the crown no torsional moment, but only a bending moment, i.e., one acting in a horizontal j^lane at right-angles with the sectional plane through the crown, m being this moment. The arch can now be cut at the crown, and one half can be removed if this moment and a horizontal force is put in its place. In all sections between the springing and the crown act moments in the plane of the sections as well as at right angles with them ; they are derived from the acting forces which have been ascertained before, but have to be supplemented by the moment m at the crown. The various moments produce elastic rotations of the sectional planes, which can be resolved into those round a horizontal axis and those round a vertical axis. As the sectional plane at the crown is the free end of the half arch, the rotation of -that plane will be the sum of all rotations between the springing and the crown. The rotation of the section through the crown round a vertical axis is nil, on account of the symmetrical position of the forces ; conse- quently the corresponding sum of rotation can be put = 0. This equation only contains one unknown quantity, viz. m, and this can therefore be determined ; with it all moments acting in the various sections can be found, as also the rotations round the horizontal axes, but the latter are not required for the calculation of strains.

438 THE GARABIT VIADUCT. [Foreign

The results of the calculations of the arch are stated in twenty- two Tables, and it can be seen that the strains in the booms of the arch from all causes, excepting the change of temperature, do not exceed 6 kilogTams per square millimetre (3 • 81 tons per square inch) of the gross sectional area, and those in the parts of the web do not exceed 5 kilograms per square millimetre (3-17 tons per square inch).

M. A. E.

The Bridge over the Po at Casalmaggiore for the Parma-Brescia Bailway.

(L'Ingegneria Civile e le Arti Industriali, 1888, p. 129.)

This bridge consists of seventeen spans, of which those at each end are 180 feet, and the other fifteen are 213 feet, the total length between the abutments being 3,560 feet. The j^iers and abutments were founded by means of compressed air at dejiths varying from 65 to 85 feet.

The bridge is for a single line of way, the load being on the bottom flange. It is formed of two parallel continuous lattice girders fixed upon the eighth pier, and supported on rollers on the other piers and abutments. The river, when not in flood, is 1,312 feet wide, and a temporary bridge of timber was built over this vridth for the carriage of materials. Near one of the abutments shops were erected for the engines, air-compressors, pumps, dynamos, repairs, smithies, stores and offices. After the first ten supports had been b^^ilt, these shops and their contents were transferred to the other side of the river. The air-compressors were driven by two semi-fixed engines of 35 nominal HP. each, and one of 10 HP., the latter being used for shallow depths.

The whole of the compressed-air foundations were completed in fourteen months, during two of which work was susjiended on account of the cold. The material excavated, which was almost entirely of a sandy nature, was thrown by the workmen into a chest about 1 foot 8 inches sqaiare, and 2 feet 8 inches high, into which water was pumj^ed through a pipe so as to mix with the sand. The mixture was then forced out through another pipe, in an almost continuous stream, by the pressure of the compressed-air in the chamber. By this means a volume of aljout 130 cubic yards, in the j^roportion of one-third of sand and two-thirds of water, was forced out in twenty-four hours. The working chambers were lighted by electric lamps continuously, and during the night the temporary bridge and the whole of the shops were also lighted, there being in all fifty Edison lamps of 6-candle power each, driven by a 4-HP. (nominal) portable engine. The total depth of founda- tion amounted to 1,312 feet, and was executed in three hundred

Abstracts.] BRIDGE OVER THE PO AT CASALMAGGIORE. 439

and eighty-three working days of twenty-four hours. There was no considerable flood in the river during this period.

The calculations for expansion and contraction were based on the assumption that the temperature would range from 14° to lO-i'^ Fahrenheit, and that the difference in temperature between the top and bottom flanges, owing to the top being in sun and the bottom in shade might amount to 22^. This difference might be disregarded in a single-span bridge, but in the case of eight or nine continuous spans it must be considered, owing to the fact that it will have a tendency to throw the uprights out of the vertical, and strains will be brought upon the flanges in resisting this. In the present case, the elongation of the upper over the lower flange, between the eighth pier and the abutment, would amount to 2'8 inches, or to 0"36 inch in one span. The upper flange would therefore be compressed to the extent of 0*18 inch, and the lower flange extended to the same amount, and the force necessary to produce these alterations in length was allowed for in the calcula- tions for the flanges. The permanent load was taken at 2,460, and the moving load at 4,000 kilograms per lineal metre of bridge (j ton and 1} ton per lineal foot respectively). In calculating the wind-bracing the wind-j^ressure was taken at 51^ lbs. per square foot (250 kilograms per square metre). The working loads were taken at 6 kilograms per square millimetre (3 "SI tons per square inch) for plates, 5 kilograms (3-175 tons per square inch) for lattice-bars, and 3-5 kilograms (2 "222 tons per square inch) for parts subject to crushing over the supports. The tests for the iron were : — 1st, to bear a tensile strain of 15 kilograms per square millimetre (9-22 tons per square inch), without the least i:)ermanent set ; 2nd, to show no sign of ruptiire with a less strain than 35 kilograms per square millimetre (22 • 22 tons per square inch), and to show an extension of not less than 8 per cent, if intended to bear tensile strain in the work ; 3rd, to show no sign of rupture with a less strain than 32 kilograms (20*32 tons per square inch), with elongation not less than 8 per cent, if the iron is to be subject to compression in the work. There were three hundred pieces tested, and they all satisfied the first test, and the breaking-load varied from 36*6 to 38*98 kilograms per square millimetre (23*24 to 25 * 4 tons per square inch).

The work was commenced in March 1885, and finished in March 1887. The bridge was tested first by allowing five engines and three tenders to stand on each of the seventeen spans suc- cessively, then by allowing ten engines and six tenders to stand on each pair of girders. A train consisting of six engines and tenders was then run over the bridge at 30 miles an hour, and afterwards two engines and tenders. The gTcatest deflection, both with stationary and moving loads, was 1*18 inch, the calculated deflection being 1 * 30 inch. The greatest lateral deflection was 0 * 088 inch. The maximum strain on the iron (taken by Castigliano's multiplying micrometer) was 2 * 80 tons per square inch on the flanges, and 2 * 54 tons on the lattice bars. The tests of the various

440 BKIDGE OVER THE PO AT CASALMAGGIOEE. [Foreign

spans gave very uniform results, showing that the material was homogeneous, and the workmanship accurate.

W. H. T.

Note. — There is a mucli fuller description of this bridge, giving in detail the preliminary calculations and the j-esults of the proof-loads, in the " Giomale del Genio Civile" for February, March, and April 1888. — W. H. T.

Erection of the Large Girders of the Machinery Hall at the Paris Exhibition of 1889.^ By Eugene Henard.

(Le G^nie Civil, vol. xiii. 1888, pj). 211 and 321, 2 plates and 7 woodcuts.)

A detailed description of ithe work of erection is given on the eve of its completion, with several drawings in ilhistration, sup- plementing the summary description previously given shortly after the commencement of the work.^ The huge iron structure, 1,378 feet long, and 337 feet wide, was to be completed in September 1888, six months only after its commencement. The method adopted by the Fives-Lille Company, of riveting together the portions of the four sections of each rib on the ground, and then lifting the four sections into place by aid of scaffoldings and w^inches, so as to leave very little riveting to be done on the scaffolding, is first described in detail. The lifting of the large central portion requires about five hours. The weight of each of the ordinary ribs is about 193 tons, and of the end rib 236 tons; and the total w^eight of the ribs, with their purlins and framework, for half the building, is about 3,640 tons. Out of 32,000 rivets in each ordinary rib, 19,600 were riveted up at the shops, 10,300 on the gToiind at the works, and only 2,100 on the scaflbldings. The first rib, with its accessories, was erected in twenty-three days, the second in sixteen days, the third in twelve days, and the remaining ribs, on an average, in aboiit 10 days each.

The second portion of the article describes in detail the method of erection, for the other half of the ribs of the biiilding, adopted by the Anciens Etablissement Cail Company, consisting in raising the pieces of the ribs separately, not exceeding 3 tons each, and riveting them together on a single scaffolding following as nearly as practicable the intrados of the arch. The scaffolding consists of five large upright stagings 52 j, 59, and 65 1 feet long, and 26 j feet wide, connected together at a height of 33 feet above the ground by braces, and at the top l)y two timlier platforms, one in steps following the curve of the arched rib which it is made to support, 16 jf feet wide, and the other running horizontally at a height of 115 feet alongside the first, and touching it in two i:)oints where they are at the same level. Two cranes run along this latter plat- form, on a line of 8.1 feet gauge, for erecting the girders. Each

' Minutes of Proceedings lust. C.E., vol. xciv. p. 372.

Abstracts.] MACHINEBY HALL, PARIS EXHIBITION, 1889. 441

crane consists of a braced iron staging, 39. \ feet high, supporting at the top a wronght-iron double-webbed girder, 32t feet long, overhanging the stage on each side, carrying a movable winch for lifting the several pieces. The scaffolding, which is borne on twelve rollers under each of the upright stagings resting on rails running lengthways along the building, can be shifted, after the erection of one rib, to the site of the next, a distance of 70\ feet, in 1^ hour at most, by the aid of ropes, pulleys, and winches, though the total length of the scaffolding is SS-ij. Special precautions "were taken in adjusting the rails, and in equalising the rate of progress of the five stagings. The piers at each side are erected first by means of the large travelling cranes on the ground, and then these proceed to the next rib ; whilst the scaffolding and its travelling cranes are employed for the erection of the remainder. In this case, out of the 32,000 rivets in each rib, only 4,000 were put in at the workshops, 8,000 on the site, and 20,000 from the scaffolding. The number of men employed on this half of the work was 215 i)er day, as compared with 250 men for the other system of erection. The first rib was completed on the 2-ith of May, together with its intermediate parts, the second and third were finished in thirteen days each, the fourth and fifth in twelve days each, and the remainder in ten days each.

L. V. H.

Reports of the French Delegates on the Proceedings of the Second International Inland Navigation Congress, held at Vienna in 1886.

(Annales des Pouts et Chaussues, 6th series, vol. xv. 1888, p. 85(3, 2 plates.)

After giving a brief summary of the origin of these international inland navigation congresses, and of the questions submitted, and resolutions adopted at the first of these congresses, held at Brussels in 1885, the programme of the second congress, held at Vienna in 1886, is given, followed by a short reference to the proceedings of the congress drawn up by Mr. Voisin Bey, the President of the French delegation. Eeports of the proceedings of the four sections, into which the subjects to be considered by the congress at Vienna were divided, are then given. The study of navigable waterways from the economic point of view formed the subject allotted for the deliberations of the first section, reported on by two of the French delegates, Messrs. Boule and Hirsch. The congress arrived at the resolution, on this point, that the economic importance of artificial navigable waterways for trade is so great, that it is advantageous to construct them in suitable localities, even where there are railways, and to equip them with the appliances for working requisite for modern traffic ; whilst acces-iory advantages, such especially as drainage and irrigation, would often facilitate their construction. It was considered that a more complete and practical collection of statistics is very necessary, to enable the economic

442 INTERNATIONAL INLAND NAVIGATION CONGRESS. [Foreign

vahie of navigable inland "n'aterAvays to be fully determined. Tlie question submitted to the second section was the normal sections of canals, and the dimensions of works of construction relating to artificial navigable inland waterways ; and the proceedings of this section were reported on by Messrs. Holtz and Carlier. The pre- liminary consideration of this subject had been entrusted to one of the French delegates, Mr. Holtz, and to a German professor, Mr. Schlichting ; and their reports formed the basis for the delibera- tions of the congress. Mr. Holtz proposed that the minimum dimensions adopted for the principal French waterways, under the law of 1879, namely, a depth of 6 feet 7 inches of water, width of locks 17 feet, available length 126_^ feet, and a clear height of 12i^ feet under bridges, should be extended to all the waterways in direct communication with the French watero^ays. Mr. Schlichting, however, whilst recommending the same depth of 6 feet 7 inches as on the French canals, proposed wider sections and larger locks. Thus he proposed that the wetted section of the canal should be four times the immersed section of the boats, instead of only three times as on French canals ; that the minimum width of the locks should be 23 feet, and their available length 1 885 feet ; and a clear height of 14j feet under the bridges, and in tunnels, above the water-level. Owing to the superior number of German members, the proposals of Mr. Schlichting were adopted by the majority. The congress also expressed the opinion that the existing canals, where directly in communication with the canals of adjoining countries having larger dimensions, should be as soon as possible enlarged. The third section of the congress devoted their attention to the methods of working the navigable waterways ; and a rejiort of the proposals brought forward, the discussion, and the resolu- tions arrived at, is given by Mr. Denys of Epinal. The congress decided that a regular service of traction should be organised on canals ; that on rivers, it is neither necessary or desirable to impose any restrictions on the liberty of towage as actually exist- ing ; that it appears expedient to establish on all waterways, by side of public traction and private navigation, associations of boatmen for the transjwrt of goods, which are not generally sent in large quantities, but nevertheless for long distances ; that a rapid extension of warehouses and sheds is very important for the progress of inland navigation, provided with the best mechanical appliances for handling the goods ; that a uniform classification of grains would greatly facilitate inland navigation, and enable European agTiculturists to compete more advantageously with the markets of the world ; and, lastly, that the formation of public havens of refuge for vessels in the winter is an absolute necessity for inland navigation. The labours of this section of the congress, bear the marks of being influenced by local considerations, for the three first resolutions are only strictly applicable to a system of navigable waterways of very large size, and very large traffic, such as Northern Germany is constructing; whilst the three other resolutions relate particularly to the special conditions and needs

Abstracts.] INTEENATIONAL INLAND NAVIGATION CONGRESS. 443

of the Danube navigation. The fourth section had to discuss the constrixction and utility of ship-canals ; hut the conclusions arrived at by the members of this section were too general for the congress to adopt any definite conclusions, and the subject was referred to the Frankfort Congress of 1888. The report of the proceedings of this section was drawn up by Mr. Barlatier de Mas.

The French delegates have added to their reports of the proceed- ings at the Vienna Congress, notices of navigations and works visited by them on their way to the congress, and after its termination. Mr. Barlatier de Mas, in his notice on the navigation of the Ehine in 1885, gives particulars with reference to the navigability of the river in that year, as compared with 1882, which is affected by the lowness of the water, by floods, and by ice, and therefore varies from year to year, so that the actual benefits from the important works in jirogress cannot be precisely deter- mined. Statistics are next given of the tonnage of goods conveyed along the Ehine in 1885, which amounted to over 12 million tons, an increase of 21 per cent, on the tonnage in 1882, which was mainly due to the increase at the German ports, as the tonnage at the Dutch and Belgian ports showed an advance of less than 5 per cent. The port of Mannheim-Ludwigshafen is the real terminus of the Ehine navigation, for the traffic is quite small beyond. Between 1876 and 1885, the traffic of this port had trebled in tonnage ; whilst the port of Duisburg had doubled its tonnage, and the port of Euhrort had increased its tonnage by one half. Tables are given showing the cost of transport between Eotterdam and the several German ports. The notice concludes with some references to the boats employed, and traction. The ample supply of means of transport on the Ehine, and the active competition, have produced a constant reduction in the freightage ; and the reduction in the price of towage has been considerable, having amounted, for one company, to 20 per cent, between 1882 and 1885. A regular direct service was started between Cologne and London by the BadenNavigationCompany in 1885, with one steamer; and the results were so satisfactory that a second steamer was put on in 1886.

A detailed account of the works for the canalization of the river between Frankfort and Mainz, which were in progress when visited by the French delegates in 1886, is given by Mr. Boule ; a description of them, subsequent to their completion, will appear in a Paper on " Some Canal, Eiver, and other Works, in France, Belgium, and Germany" to be published in the Minutes of Pro- ceedings, vol. xcvi.

The last notice ajipended to the reports is an account of the excursion made by the members of the congress down the Danube, from Eegensburg to Turnseverin, by Mr. Hirsch. A summary description is given in this notice of the portion of the Danube traversed, and of the works being carried out for the improvement of the navigation. L. V. H.

444 MEASUREMENTS OF THE FLOW OF THE ELBE IN SAXONY. [Foreign

Measurements of the FIoiv of the Elbe in Saxony, 1886 and 1887. By A. EixGEL.

(Civilingenieur, vol. sxxiv. 1888, p. 505.)

The measiirements of the flow of the Elbe forming the subject of this commnnication were undertaken by the Author in conjunction with the Engineers, Messrs. M. Eingel and F. Lindig, in the year 1886, at the instance of the Eoyal Hydraulic Department (Koenig- lichen Wasserbau Direction).

In Saxony there are three points on the Elbe arranged for taking measurements. For about 500 metres both above and below these points the sections of the stream previously taken are indicated by blocks of sandstone, the tops of which are furnished with bolts.

Current meters were employed for measuring the velocity of the water, which were made to slide up or down on vertical rods firmly secured in the bed of the river. Every revolution of the screw was transmitted by electricity to a counter or chronograph placed above water. The current-meters were supported by wire ropes wound upon drums attached to the rods before referred to.

In addition to the usual method of measuring by current-meters, Professor Harlacher's system of so-called integration was employed. The observations were conducted from a floating platform carried by two strong barges connected with each other in siich a way that between the barges there was a clear space of 6 metres (19 feet 8-22 inches).

The rod on which the current-meter slides was supported at the upper end by an iron fork carried on a tripod, at the top of which was a roller over which the rope passed for hoisting up the rod and meter. This tripod was hinged at one of its legs so that it could be laid down flat on the platform.

The allocation of the apparatus at a previously determined point of the section in which a series of observations at various depths were to be made, was effected by means of a wire rope stretched across the stream, on which the points were indicated by small bands tied on.

Every observation lasted two minutes, and in each position at least three measurements were taken.

Professor Harlacher's method of so-called integration is carried out as follows : — The current-meter is hoisted by means of the drum about 0 • 5 metre above the surface of the water, and then allowed to slide slowly down the rod by its own weight until the screw touches the bottom ; at the moment when the axis of the meter passes below the surface, all the recording instruments in connection with it are put into gear, and as soon as the bottom is reached thrown out again. The recording instruments consist of a stop clock, a counter, and three styles belonging to the chrono- graph.

Abstracts.] MEASUREMENT OF THE FLOW OF THE ELBE IN SAXONY. 445

On the moving paper band driven by the chronograph are recorded : —

l) The number of revohitions of the meter.

[2) The depth of the axis of the meter below water.

(3) The time in quarter seconds.

By taking into account these three quantities the mean velocity can be approximately arrived at.

In order subsequently to obtain some mathematical connection between the surface gradient and the other observed quantities, the levels of the surface were observed twice each day in that portion of the river where the measurements were carried out, piles having been previously driven in and graduated for this purpose.

After reviewing the varioiis theories as to the distribution of the velocity of flow according to the depth, the Axithor states that he and his colleagues arrived at the conclusion that the experi- mental results agreed best with the assumption that the maximum velocity was at the surface and the minimum velocity at the bottom, and that the variations of the velocity with the depth were most accurately represented by a parabolic curve of the wth order, of the form ?/" = p x, where x is the height of any point above the bottom, and y proportional to the velocity of flow at that point ; the curve has its apex at the bottom and the axis of abscissa coincides with the axis of the curve which is vertical.

The value of n varied at difierent points of the section from 3 • 53 to 5*65 in one case, from 3*43 to 8* 01 in another, and from 2-69 to 4-58 in a third. The original is accompanied by diagrams in the text.

G. E. B.

Begulation of the Isar according to Wolfs method. By E. IszKOwsKi.

(Wochcnschrift des osterreichischen Ingenieur- und Architekten Vereines, 1888, pp. 74 et seq.)

The Author states the advantages of Wolfs method over the systems hitherto employed to be, not only that it is a more speedy and effective method of regulating rivers, but much more eco- nomical, and he grounds his statement on the results obtained on a section of the Eiver Isar, 44 miles in length, extending from the boundary between Upper and Lower Bavaria (7^ miles above Landshut) down to the village of Plattling.

The Isar, throughout its course of 170 miles from its source in the Karwiindel moimtains to its junction with the Danube, has the character of a mountain stream, and has a drainage area or catch- ment basin of 3,416 square miles. The average fall at low- water level, in its coitrse of 111 miles through Upper Bavaria, is 1 in

446 KEGHLATION OF THE ISAE. [Foreign

600, and for the remaining 58 miles 1 in 1000, bnt owing to large deposits of gravel and boulders here and there, the fall varies in places from 1 in 200 to 1 in 1,500.

Its discharge in ordinary floods may he taken at 900 cubic yards per second.

The dejith varies very considerably, and at mean-water level is as follows :

At Mittenwald (near the soui-ce) . . . 1-65 foot per second.

At Tolz 3-28 feet „

From Municli downwards 6"56 „ „ „

but in places where the concave banks are specially exposed to the current, the depths vary' from 1 3 to 23 feet.

The normal breadth in the lower section of the river is about 75 yards.

Wolf's system aims at leading a river at the smallest possible cost in a normal channel ; that is, with breadth and depth con- sistent with its proper regimen or stability ; it also effects a uniformity of fall, the reclamation of old channels, the stopping of branches, and the proper maintenance of the regulated section of the river. This is attained by the employment of floating spurs (^Gelidnge) or layers of fascines (attached to piles driven in the bed of the river), the action of which increases the velocity of the water in the parts of the river to be improved or deepened, and reduces it in the part to be reclaimed, causing respectively an erosion in the former case, and a dej^osit in the latter, and this much more quickly and completely than by the works usually constructed.

The piles are from 8 to 10 inches in diameter, and are driven in a line parallel to the regulation trace to a depth of from 10 to 16 feet below the river-bed, and at distances of about 8 feet apart. To these the fascines are attached with their thick ends towards the stream, that is, with their brush-like ends towards the line of bank to be formed, and the distance of the row of piles from this bank is dependent on the depth of the water. At places where specially strong currents are met \\ith, a second row of piles is driven in behind the first row, and both are secured by cross stays oblique to the axis of the river. When the deposit of gTavel or boulders behind the fascines has settled down to its natural slope, the fascines are sunk so as to cover this slope, and are weighted down with stones so as to form a complete revetment ; the top of the bank is protected in a similar manner.

According to Mr. Wolfs experience a river can be trained or regulated in this way in a remarkably short time, but the speed is dependent in great measiire on the nature of the water-levels. Strong continuous mean average levels accomplish more in a week in this respect than low water-levels in twenty times that period.

As a result of the success of this method, it is stated that in a section of the Isar, about 15 miles long, regulated by this method, no place was found were the action of the floating spurs did not produce a deepening of the" channel in front of them, and the re-

4

Abstracts.] REGULATION OF THE ISAR. 447

quired deposit of gravel behind extending tip to the surface of the water or even above it. The cost of these works is as follows per lineal yard : One row of piles 6i to 8 feet apart, including fas- cines, &c., 2s. 9f/. to 3s. 8d. ; two rows of piles, front row 8 feet apart and back row 6i to 13 feet apart, distance between the rows 6 J to 10 feet, inclusive of fascines, cross-stays, &c., from 4s. Id. to 6s. bd. ; and in exceptional cases, where an unusually strong current is met with, the price may be taken at from 6s. bd. to 7s. 4(i. per lineal yard. The Paper is illustrated by twelve examples, showing the arrangement of the piles and floating spurs to meet particular cases during the regulation of the Isar.

W. H. E.

On the Imi^rovement of the River Moldau at Prague, and the Construction of a Port there.

(Wochenschrift des osterreichischen Ingenieur- und Architekten Vereines, 1888, p. 2U2.)

The general project prepared by the Government engineers was referred to a Sub-Committee, and by it again referred to a Technical Committee for report.

This report, drawn up by Mr. Kaftan, is divided into two parts. The first part treats of the economic importance of waterways in general, and of the great benefit accruing to Prague (and Bohemia generally) by making the Moldau navigable there, and the shipment and trans-shipment of goods.

The second part is of a purely technical character, and lays down a programme for carrying out the details of the scheme. This programme was strongly opposed by Messrs. Olwein and Umrath, on the ground that the fixed weirs, which it was proposed to retain, should be removed and replaced by movable weirs with a view to lower the existing flood-levels, and thus prevent the con- tinually recurring inundation of the low-lying parts of the city, and to admit of its proper drainage and sewerage.

The lowering of the water surface in floods being of the greatest economic importance, as then all the quay- and wharf-walls might be built at least 1 metre less in height, and the drainage and sewerage of the city being matters of great urgency, the proposal of the Technical Committee to retain the fixed weirs was re- jected by the Sub-Committee ; but the project for the Holesovic harbour and wharf for trans-shipment of goods, 3j miles distant from Prague, was accepted and highly approved.

This scheme provides for the berthing of about three hundred boats, and of a convenient approach for the Austro-Hungarian Government Eailways and the Buschtehrad Railway. The Techni- cal Committee also recommended an enlargement of the Carolinenthal harbour, where at present there is only accommodation for thirty boats ; biit as it was feared that this work might interfere with

448 IMPROVEMENT OF THE RIVER MOLDAU AT PRAGUE. [Foreign

the immediate execution of the Holesovic harbour works, which were urgently necessary, it was declined by the Sub-Committee on that ground.

A proposal by Chief Inspector Olwein to include in the scheme a goods wharf and landing-stage on the right bank of the Moldau, between Lieben and Carolinenthal, on the ground of its easy and cheap connection with the railways on that bank of the river, as well as with the Holesovic harbour, was accepted by the Sub- Committee on condition that they did not prevent the Holesovic harbour works being first taken in hand, together with the drainage of Prague, and that they did not interrupt the free navigation of the river.

At a subseqiient meeting of the Sub-Committee the scheme for the enlargement of the Carolinenthal harbour was again considered, and a proposal to carry oiit this work before any of the other works was passed by a majority of the members.

W. H. E.

Special Plant for Blasting binder Water at the Panama Canal Works. By Max de Naxsouty.

(Le Geuie Civil, vol. xiii, 1888, p. 324, 1 woodcut.)

A wooden raft, 82 feet by 46 feet, kept above water by forty- eight floats, and covered with planking, in which holes 8 inches square and 8i feet apart have been made, serves as a stage, from which the blasting of boulders under water on the site of the canal at Mindi, near Colon, can be carried on across the whole bottom width of the canal. Tubes 1-38 inch in diameter, and 16| feet long, which can be screwed together, j^rovide bars of variable length, by which blasting holes can be driven to a depth of 29.5 feet without too great toil. These bars are passed through the holes in the planking, and are kept vertical by passing them through the meshes of a sort of trellis roof erected over the raft. An iron pipe, about 3 2 inches in diameter, resting on the bottom, and passing at the top through one of the holes in the planking, prevents the blasting hole being lost when the bar is removed, or in charging the hole ; and it also serves for guiding the bar. The pipe is raised, when the charge is in place, by tackle from the superstructure. The explosive is inserted through a tube of 2\ inches inside diameter. The several charges are connected by wires, so that 35 to 40 mines can be fired simultaneously by electricity.

L. V. H.

Abstracts.] THE EMBANKMENT OF THE PO AT TURIN. 449

The Emhankment of the Po at Turin. By Tommaso Prinetti.

(Gioriiale del Genio Civile, June 1888, p. 314.)

In prolonging the embankment of the Po, it was decided, instead of constructing an ordinary retaining-wall, to adopt the system of forming a series of chambers with a roadway over them. The principal reasons for this decision were, first, that the height from water-level to ground-surface was so great that it would take a long time for the earth filling behind the wall to consolidate; secondly, it was considered advantageous to have a number of underground chambers which could be used for stores, workshops, laundries, boat-houses, and so on ; thirdly, that the additional cost would be compensated by the rents ; fourthly, that as houses would be built along the roadway on the further side from the river, and these houses would have certainly one and probably two stories of cellars, the height of the earth pressing against the back of the wall would be reduced by the height of these cellars.

Several types of wall were adopted, and drawings of one of them are given. In this case there is next to the river a towing- path wall founded on piles and concrete, and built of rubble faced with ashlar. Its height is 8 feet 6 inches above water-level. The towing-path is 26 feet 3 inches wide, and slopes 1 in 20 towards the river. At the side of the path is the fa9ade to the chambers, which is 29 feet high to top of cornice, and 33 feet 3 inches to top of parapet. It is 3 feet thick, and consists of a series of piers 8 feet 10 inches wide, and openings 6 feet 10 inches wide, 11 feet 6 inches high to springing, with semi-circular arches over them. Above each of these openings is a square window. The face of the wall is hammer-dressed coursed rubble with rusti- cated quoins. The underground chambers are behind this wall, their width being 37 feet 9 inches. At the back is a solid wall 3 feet thick carried up to road-level, and forming the foundation of the front wall of the houses facing the road. The span of the vaults is 13 feet, and the walls separating them are formed of a central pier 3;^ feet wide, and two arched openings of 15 feet 6 inches with semicircular arches. Flights of stairs are formed at intervals, leading from the roadway to the towing-path and from the latter to the water. The vaulting ai )hes are 1 feet 7 inches thick at the springing and 1 foot 3 inches at the crown; they and the spandrel wall are covered with a layer of hydraulic cement and another of asphalte. As chimneys will be required in some of the chambers, they have been provided. Where houses are to be built along the roadway the chimneys will be carried up the cross walls, and where this is not the case they will be left in the face-wall and ornamental iron columns will be provided as chimneys.

The cost of the wall as described above was about £68 per lineal metre. A simple retaining-wall would have cost about £48. Some

[the INST. C.E. VOL. XCV.] 2 U

450 THE EMBANKMENT OF THE PO AT TTTEIN. [Foreign

of the chambers are retained bj^ the Municipality, others have been made over to owners of frontage, others are let at a rental of Is. 9f^ per square metre, or about £l Is. per lineal metre.

W. H. T.

Jandin's Compressed- Air Dredger. By M. Boulle.

(Annales des Fonts et Chaussees, 6th series, vol. xv. 1888, p. 1034, 1 woodcut.)

This dredger accomplishes the removal of sand or silt by an injection of comj^ressed air, instead of the well-known method of suction. It consists of a tube going down under water to the bottom to be dredged, and a compressed-air injector placed at the bottom of another pipe at right angles to it, and surrounding the main tube, into which the compressed air from the pipe is injected from a great number of little mouthpieces directing the current upwards. This injection caixses the w-ater to rise in the tube, drawing in water from outside, which eflects the dredging at the bottom ; and a mixture of water, silt, and air, flows up the tube. This dredger has been tried in the Loire at Saumur, and on the Seine at Havre. At Saumur the tube, 4 inches in diameter, dredged pure sand at a depth of 13 feet, w'hich was lifted 5 J feet above the water-level, and transported to a distance of 50 feet. A compressor of 15 HP., drawing in 3*53 cubic feet of air per second, enabled the dredger to raise 130 cubic yards of water -per hour charged with sand, which composed three to four-tenths of the whole volume. In the Eure Dock at Havre the silt w^as dredged, in a depth of 26 to 30 feet, by a tube 9 inches in diameter; it w^as lifted 5 feet above the water-level, and discharged into a barge. Using a compressor of the same power as at Saumur, 390 to 520 cubic yards of silt and water were lifted per hour, the silt forming one quarter of the whole volume. This dredger is most efficient in soft silt, sand, or gravel ; but, nevertheless, at the foundations of the Palma Bridge, stones weighing 22 lbs. w-ere removed with a tube 9 inches in diameter. This machine might be advantageoiisly used for pumping out water charged with gravel, sand, or debris.

L. V. H.

Renewal of the Water in the Hague Canals. By M. E. v. Pichler.

(Woctienschrift des Osterreichische Ingenieur- und Architeckten Yereines, 1888,

p. 118.)

In this Paper the Author first gives a sketch of the special physical feature of Holland, namely, that a great j^art of the country lies many feet below sea-level, and requires for its pro- tection from the sea and from rivers, the construction of strong dykes and embankments ;- and as regards the neighbourhood of the

Abstracts.] RENEWAL OF THE WATER IN THE HAGUE CANALS. 45l

Hagiie, it is pointed out that between tlie years 1713 and 1863, the natural bulwarks of the coast (the Dunes or chains of quartzose- sand hills), had been so washed away by the action of the waves that the shore bad advanced in some places from 580 to 020 yards, which necessitated the construction of massive stone dykes, the so- called Delft Iwofde.

An account is also given of the formation of the Zuyder See in the year 1300 by the bursting of the natural dam connecting North Holland with Friesland, and of the disastrous inundations in South Holland in the fifteenth century by the breaching of a dyke. This locality is now an archipelago called Biesbosch.

The drainage of the Haarlem Lake (commenced in 1840, finished in 1853) is also noticed, and a description given of the pumping- engines emjiloyed which are in use to this day.

A large portion of the Paper is devoted to a detailed account of the river systems of Holland, the Rhine, the Maas and their tributaries, and an account is given of the locks at Katwyk at the mouth of the Rhine, constructed in 1808 with a view of giving it a free outlet into the sea, and of the pumping-engines that were erected there in 1880.

The chief feature of the drainage system of Holland is the high- level canal or reservoir called the Boezem, the level of which is carefully regulated by pumping water up from the low-lying parts when that level is too low, and by leading it off into other canals and thence into the sea, or by utilizing it for working water-wheels in times of flood. The particular canal which affects the Hague and Delft districts is the Schie Boezem, so called from the River Schie, and the ratio of its water-surface to that of the land draining into it is given as 1 : 73*2, which is said to be much too small from the fact that to this day a large extent of country is under water in wet seasons because the Boezem cannot carry it off. It is pointed out in the Paper that while cities like Amsterdam and Rotterdam, situated on rivers subject to tidal influences, can easily have their canals renewed and flushed, Hague is under the disadvantage of being only indirectly connected with the sea by its high-level canal or Boezem, and the consequence is that in summer the canals (^gradden) which intersect the town in every direction, are most insalubrious and offensive. The town archives show that such has been their condition for centuries, and that although from time to time suggestions have been made for the renewal of the water in them, nothing was really done till 1883, when Chief Engineer J. Van der Vegt submitted to the district council a plan and estimate of cost of the works.

Under this scheme 88,000 gallons of water per minute were to be taken from the Schie Boezem and led to the Hague through the Trekvaart, and after passing through the town canals woiild be lifted or pumped into a higher canal furnished with flood gates, and thence led into the sea. This project was opposed by the Delft district authorities because it did not favour their interests sufficiently, and Mr. Van der Vegt (in conjunction with Messrs.

2 G 2

ip

452 EENEWAL OF THE WATER IN THE HAGUE CANALS. [Foreign

Malsen and Siccama) prepared anotlier scheme, in which the water was also to be taken from the Vaart and pumped into a high-level canal with sill 6^ feet above the standard gauge, and discharged into the sea at the " dunes " by means of three siphons each 5 feet in diameter. This project was approved, and work was commenced in 1886, but the working expenses were so high and the season so un- favourable owing to floods, that the first project was again taken up, and being subsidized by a gTant of £14,500, was accepted by the council. The carrying out of the entire scheme was entrusted to Chief Engineer J. Van der Vegt.

The site for the outlet of the canal into the sea was fixed at a point 1^ mile south of the celebrated bathing place Scheveningen, so as not to injuriously affect that station, and it was decided that a velocity of only about 8 inches per second could be allowed during the day, so as not to render navigation, and especially towage or haulage, difficult ; but at night when there was no traffic a higher velocity was permissible.

Taking the experience of Amsterdam as a guide, it was calcu- lated that nothing less than 44,000,000 g-allons should be supplied daily for the renewal of the water in the Hague Canals alone, biit for the irrigation of the " polders " in the whole of the Delft district, and the maintenance of the i)roper level of its Boezem canal, twice that volume would be necessary.

The water would be taken from the River Maas at Vyfsluizen, then along the Schie and through the " polders " to Delft, and thence through the Trekvaart to the Hague, and 10 hours a day was the period fixed for admitting the water.

As regards the fall to be given, this was fixed at 15^ inches in the 1 0-mile section from Vyfsluizen to the Hague, or about 1 inch in a mile.

From the Hague to the sea the canal is divided into two sections, viz., from the gas-works to the pumping-station and lock adjoining, and from thence to the sea. The length of canal or reservoir between the lock near the pumping-station and sea-lock near the canal mouth, is about 1 mile, and the pumps have to deliver 88,000 gallons per minute into this reservoir. The lock adjoining the pumping-station is provided with double gates, which are made to slide or turn, and to act as flood-gates in the event of the sea-lock- gates failing to act when reqiiired.

This sea-lock is built on concrete, and has two openings, each about 13 feet wide, instead of one large one, in order the more efiectually to resist the occasional terrible attacks of the sea, for which purpose also wave-breakers or fenders formed of strong baulks, are fixed in proper positions. The cost of the project complete is given as £79,000.

Plans and cross-sections of the locks and longitudinal section of the country adjoining them, as well as a plan of the Hague and the canals referred to, accompany the Paper.

W. H. E.

Abstracts.] NEW YORK AND BROOKLYN BRIDGE. 453

The Cable Bailwaij on the Neiv York and Broohlyn Bridge. By G. Leverich,

(Transactions of the American Society of Civil Engineers, vol. xviii., 1888, p. 67.)

The New York and Brooklyn Bridge is 601 7 '33 feet in length. The railway is double line, laid to the 4 feet 8.V inches gauge, with steel rails weighing 52 lbs. per yard. The inclines vary from level to over 3j per cent., or 1 in 26.V. The total length of cable within the rails, for actual train service, amounts to 2^ miles. It is driven continuously in one direction, by one of two stationary steam-engines at the Brooklyn terminus. The cars are each connected to the cable by a roller-grip, and they are run singly or in trains as may be required. They are moved and placed by locomotives at each terminus. The cable is driven by means of winding-drums and their accessories.

Steam is generated in six water-tube boilers, each of which has fifty-four water-tubes, 4 inches in diameter, 18 feet long, with a 3-foot drum overhead. Each fire-grate is 3 feet 10 inches wide, 7 feet long, and consists of revolving self-stoking fire-bars. The chimney is 129 feet 4 inches high, 5^ feet square at the top. Steam is supplied for three electric-lighting engines besides the cable service.

The two steam-engines are horizontal, having each a 26-inch cylinder, with a stroke of 4 feet. They are connected to the outer ends of the main driving-shaft by clutches, working one at a time, making fifty-seven revolutions, or 456 feet of piston per minute.

The rope passes over two large drums, 12 feet in diameter, on axes nearly horizontal, in each of which there are four semi-cir- cular grooves for the rope, j inch deep. The drums are 17 feet apart, slightly inclined to each other, the shafts making angles of 1 in 128 with the horizontal, so that the several rope-grooves of one drum may be directly opposite to those of the other. The drums are placed one at each side of the main-shaft, 5 feet clear of each other. A smaller drum 5 feet in diameter, occupies this interval, and runs loose on the shaft. Beside it, a toothed wheel 5 feet in diameter is keyed on the shaft, and gears into a 12 feet toothed wheel keyed on the shaft of the driving drum. Thus the drums are kept exactly at the proper distance apart, and are driven by the engine through reducing sjiTir-gear, making 23^ revolutions per minute ; and moving the rope at a speed of 895 "35 feet per minute, or 10*17 miles per hoTir. Allowing for wear and tear of grooves and slipping, the speed is usually taken as 880 feet per minute, or 10 miles per hour. The cable is of crucible-steel, 11,500 feet in length. It is, when new, 1^ inch in diameter, and it consists of 114 wires, nearly -/j^- inch thick, laid in six strands around a central strand of hemp. It weighs 3i lbs. per foot of length. Tension- cars are employed to prevent the cable from slipping, and to take up slack. Between the rails, the cable is supported on pulleys

454 NEW YORK AND BROOKLYN BRIDGE. fForeign

placed 30 feet apart, in which the groove is packed with leather and india-rubber beltings laid radially.

The cars are 48 feet 10 inches long over all ; the body is 39^ feet long, 9 feet 7 inches wide, and 13 feet 8 inches high above the level of the rails ; on a four-wheel bogie at each end. The wheels are of paper, steel-tired, 30 inches in diameter. The car, without load, weighs 16^ tons. The entrances at the ends, are 3 feet 5 inches wide, with double sliding-doors ; and at the middle of each side, a 2 feet 8 inches opening, with a single sliding-door. Forty barred seats are placed in line along each side, with an aisle 4 feet 8 inches wide. When fully occupied — seats and standing room — the car has carried 150 persons. The car is lighted by eight argand lamps, burning mineral oil. In cold weather, the car is heated by a hot-water heater with ten lines of Ij-inch wrought- iron pipes under each line of seats.

The grip is so devised as to maintain automatically the relative positions of the pieces subject to wear. It is fully described in the Paper. The brakes are applied to every car-wheel, by hand or by vacuum. Five locomotives are employed on the terminal service, having cylinders 11 inches and 12 inches in diameter, with a 16-inch stroke. Four are in constant use. The automatic car- coupler is applied to the whole of the wheeled stock.

The indicator power exerted by the driving engine was observed during the entire working day, 19^ hours, April 26, 1886. The power ranged from 303 • 1 HP. to 12-9 HP. minimum and negative ; the average for the whole day was 96*2 HP. ; for five hours of the busiest time, the average was 150-5 HP. To drive the plant without trains, the power was 47 • 7 HP.

The steel wire of which the cable is constructed has a breaking stress of from 140,000 lbs. to 190,000 lbs. per square inch. The first cable was in use 3 years 43 days ; during which time it hauled 226,273 miles, 837,895 cars, and 48,960,000 passengers, making a total weight of above 12,000,000 tons ; of which 9,000,000 tons of cars were hauled -j^ mile, and 3,000,000 tons of passengers were hauled 1 ^V mile ; making a total service of nearly 22,000,000 ton-miles (exactly 21,777,710 ton-miles). The Paper is illustrated by twenty-seven plates. D. K. C.

The Neiv Harbour Worlis at La Bochelle.

Keport of Committee commisBioned by the Italian Government to inspect the works.

(Giornale del Genio Civile, June 1888, p. 281.)

The port of Eochelle is protected from Atlantic storms by the islands of Oleron and Ee, which form two great natural break- waters, between which and the mainland shij^s can at all times ride in safety. The tidal range is 9 feet 6 inches at neaps, 16 feet 9 inches at springs, and 21,^ feet 6 inches at equinoctial springs. The

Abstracta.] NEW HARBOUK WORKS AT LA ROCHELLE. 455

highest seas (the maximum wave height being about 13 feet) are pro- duced by west and south-west winds. The present harbour being deficient both in area and depth, new works are now being constructed at a point about three miles west of the town (Porto della Pallice), it having been found impracticable to enlarge the existing port.

The position of Pallice is very favourable ; the five-metre line (below equinoctial low water) runs parallel with the shore at a distance of 1,150 feet, so that by running out piers of moderate length a sufficient depth can be obtained to enable small vessels to enter at all times, while at high water there will be a depth of 33 feet ; there is no danger of silting up, and there is close by an area of low-lying land which can, with advantage, be filled up from the excavations. The ground consists of oolitic limestone in horizontal layers alternating with dry beds of clayey marl.

The works now being carried out consist of a tidal harbour with a depth of 10^ feet at low water of ecpiinoctials, two locks available for the largest vessels, and an inner basin having a minimum depth of 26 feet. The area occupied, exclusive of the tidal harbour, is 1 48 acres, of which the water area of the inner basin is about 30 acres. This basin is 700 metres long; the first length of 400 metres is 200 metres wide, the remaining 300 being 120 metres wide. The northern side is straight, the reduction in width being obtained by an ofifeett on the south side. The total length of quay is 1,800 metres ; but of this, only 1,650 metres is available for ships to lie alongside. It is calculated that with a proper equipment of cranes and sheds, 700,00iJ tons of shipping per annum can be accommodated in the basin. The excavation for this basin amounts to 1,200,000 cubic metres, of which a large portion consists of rock requiring blasting. The plant includes 9 locomotives, 250 wagons, about 10 miles of railway, and 3 centrifugal pumps driven by a 40 HP. engine, and capable of discharging 300 cubic feet per minute. Owing to the character of the ground, the walls consist simply of a lining of masonry 1 metre thick with counterforts 2 metres deep at intervals of 15 metres. Bollards and mooring- rings are provided alternately at each counterfort. The face of the wall has a batter of 1 in 10.

There are two graving-docks, which are placed at an angle of 35^ with the basin to facilitate the entry of vessels. The larger is 541 feet long, 72 feet wide at entrance, and has a depth on the side of 32 feet at spring tides, and 28 V feet at neaps. The smaller dock is 328 feet long, 49 feet wide at entrance, and the same depth as the larger. The larger dock is sufficient for any merchant steamer now afloat, except the "Great Eastern" and the "City of Eome." The quay space round the basin will be 100 metres wide, and can be extended to 200. There will be a railway station on the northern quay, and lines of rails will be laid round the basin. The corners will be tiarned liy means of curves, turn-tables not being permitted. The basin will be entered from the tidal harbour by two locks, each having a depth of 16', feet at the harbour end, and 13 feet at the basin end, at equinoctial low water,

456 NEW HARBOUR WORKS AT LA ROCHELLE. [Foreign

their lengtlis being 541 and 475 feet, and breadth 72 and 49 feet respectively. There are intermediate gates in each, so as to save water in passing through small vessels. At neap tides vessels of 23 feet draught can enter at all times. There will be a swing- bridge across each lock.

In designing the tidal harbour, it was decided that sufficient funds were not available for a construction which would admit large vessels at all times, but that during rough weather these must remain in the roads, which they could do with perfect safety, while smaller vessels would be able to enter at all times. The depth was fixed at 16^ feet below equinoctial low- water, which would admit ships of 31^ feet draught at high-water neap tides, and 35i feet at ordinar\^ springs. The harbour is formed by two breakwaters. The direction of the shore-line is north-north-east. The soiithern breakwater runs from the shore in a west-north- westerly direction, and the northern breakwater runs from the shore towards the south-west, the distance between the two at the entrance being 295 feet. The depth of water at the end of the southern breakwater is 16^ feet at equinoctial low-water. The harbour forms approximately a trapezium, the lesser base of which (at the entrance) is 295 feet ; the larger base (at the shore) 1,475 feet, and the height (distance from shore to entrance) 1,640 feet. Means have been adopted to insure the tranquillity of the water in the harliour, so as to facilitate the entrance into the locks leading to the inner basin, but these means cannot well be described without reference to a plan. A very large amount of excavation (about 850,000 cubic metres) was necessary to obtain the required depth. In order to excavate the bulk of this in the drv% the greater part of the area was enclosed by a temporary masonry dam, about 300 metres long, built across the harbour from breakwater to break- water, and foimded at mean low-water. The excavation was carried on at the rate of 1,800 to 2,000 cubic metres per day.

The breakwaters are biiilt of concrete, faced above low-water with masonry. The work, both excavation and wall below low-water, is executed by compressed air in caissons, of construction similar to those described in the Minutes of Proceedings Inst. C.E., vol. xciii. p. 522. The walls are built in a series of piers, each pier built in the caisson, which is 72 feet long and 33 feet wide. Arches are turned over the spaces between the piers, and later on the spaces are also filled with concrete, which is executed by siDCcial means described in the Paper. When the breakwaters are finished, a temporary masonrj- dam will be built across the entrance to the harbour, and the remaining excavation wall then be completed in the dry. The cost of the compressed-air plant is given as : —

£. Two large caissons, 72 feet by 33 feet, of wrought iron, at £4,800 9,600 „ small „ 33 „ 26 „ „ £iOO 800

Four steam engines (total HP. 90) 4,000

Air-compressors, pipes, sheds, repairing-shop 17,600

Total 32,000

Abstracts.] NEW HAKBOUR WORKS AT LA ROCHELLE. 457

There are four air-compressors, capable of furnishing in all 2,400 cubic metres of air at the pressure of 1 atmosphere, and one giving a pressure of 4 atmospheres, for driving a machine for raising and lowering materials in the caissons.

The work to be done by this installation comprises about 23,000 cubic metres of walling, and 5,500 cubic metres of excavation in rock ; the total value of this work being about £120,000.

In front of the entrance to the harbour some rock has to be excavated under water. This is to be done partly with compressed- air caissons, partly by means of drills worked from a pontoon, which, by means of adjustable legs, can be fixed in any required position. In both cases the blasting-agent will be dynamite. The broken rock will be brought to the surface by a Priestman dredger.

It should be noted that the caissons used for the foundations have frequently been exposed to storms of considerable violence (waves of 3 or 4 metres in height), and though it is not safe for men to "work in them under such circumstances, the caissons with- stand the force of the waves, and completely protect the work inside them.

The cost of the works is : —

£.

Inner basin 230,000

Tidal liarboiu- 352,000

Land and accessories 198,000

Railway and sheds 220,000

Total 1,000,000

In conclusion, the Commissioners speak in very high terms of the design and execution of the works, and especially in regard to the Messrs. Zschokke and Terrier's compressed-air system.

W. H. T.

Tlie Qualities of Potable Waters.

(Les Annales des Travaiix Publics, 1888, p. 2156.)

Water should be fresh, aerated, free from smell, having a faint but agreeable taste, and a residue not exceeding 42 grains per gallon on analysis ; and the health of the popiilation drinking it affords a test of its quality. The Consulting Committee of Hygiene in France classes water according to its degTee of hardness after boiling from two to over twenty minutes ; but the quality of water cannot be determined merely by the extent of its hardness, though where water contains more than 1 grain of lime per gallon it must be regarded as tin wholesome. The hygienic committee also classes water according to the chlorine it contains, regarding as potable those waters containing less than 3j grains per gallon, and as bad where the amount exceeds 7 grains per gallon. The same body states that potable waters should not contain more than 2 grains

458 THE QUALITIES OF POTABLE WATERS. [Foreign

of sulpliiiric acid per gallon ; whilst waters are bad which contain over 7 grains. The estimation of organic matters in water has hitherto been very imperfectly effected, for these matters are estimated all together ; whereas some kinds are far more injurious than others, and a regular classification is needed in this respect. The municipal laboratory of Paris has fixed 0-35 grain of organic matter per gallon as the proper limit, which none of the waters consumed in Paris comply with ; and the hygienic committee lays down 1 grain of organic matter per gallon as allowable for potable water, with a limit of 1 • 6 grain ; whilst water containing over 2 • 8 gTains must be absolutely rejected. The gas contained in the water should be 3 j per cent, by volume, of whi(3h one-tenth should be carbonic acid, and three-tenths oxj^gen. Water charged with microbes loses its oxygen rapidly, whilst water charged -wdth algce absorbs oxygen. Navigation and wind, by stirring up the lower layers of water, and bringing them into contact with the oxygen of the air, which oxidises the organic matter, improve the quality of the water. Potable water should not contain any sulphuretted hydrogen, or carburetted hydrogen, and not more than traces of ammonia and other nitrogen compounds ; and the amount of iron should be limited to -1- gTain per gallon. Microscopical investiga- tion supplements chemical analysis ; but the two methods often lead to discordant results. It is possible, however, that the innocuous microbes destroy the hiirtful ones to a great extent. There are four sources of water-supply available for Paris and its neighbourhood, namely, spring water, such as the Dhuys and the Vanne, near their sources ; the waters of the Seine, the Oise, and the Marne ; artesian wells, and shallow wells. The spring waters are the best, but costly to obtain owing to the distance of convey- ance. The waters of the Seine, Oise, and the Marne are about equally good. The water from artesian wells is small in quantity, and, if drawn from a gTeat dejith, is too warm and little aerated. The water from shallow private wells is still largely used ; but it is gradiially being abandoned, owing to its liability to contamination. A Table gives the degree of hardness, the residixe, and the organic matter resulting from analyses of the waters supplied to the princiiial towns of Europe, affording an idea of their quality.

L. V. H.

Water-Supply in the Kingdom of Wurtemherg. By J. R.

(Wochenschrift des osterreichischen Ingenieur- und Architekten Vereines, 1888, pp. 210-211.)

The paper opens with a description of the physical features of the country referred to, viz., the Eauhe Alb, an elevated, bleak, and sterile plateau north of the Danube. The works have been in progress for the last twenty years, and have recently been com- pleted. The general project consists iu the formation of nine

Abstracts.] WATEE-SUPPLY IN THE KINGDOM OF WURTEMBERG. 459

independent water-supply " groups " or districts, embracing an area of 700 square miles and a population of 40,000 inhabitants. The requirements are estimated at 16 j gallons per head per diem; but provision is made for supplying nearly twice this amoixnt, if necessary.

From the nine pumping stations, 1,100,000 gallons of water are lifted daily and discharged into sixty-two reservoirs, which are situated on heights not too far distant from the stations, and which contain from 65,000 to 308,000 gallons, and have a storage capacity equal to from six to ten days' supply.

The cost of the Eauhe Alb water-sujiply, inclusive of all charges, amounts to about £280,000, and the cost of individual "grou})s" varies from £6,750 to £60,000, according to population supplied, and the share per head ranges accordingly from £-1 10s. to £10.

The benefits of the scheme, it is stated, are already evident in the better sanitation of the houses, the less frequent occurrence of cattle disease, and in the provision for extinguishing fires ; in short (it is added), that the Alb water-supply in the kingdom of Wurtemberg may be generally regarded as a model for imitation under similar conditions.

W. H. E.

Facts in Relation to Friction, Waste and Loss of Water in Mains. By C. B. Brush, M. Am. Soc. C.E.

(Transactions of the American Society of Civil Engineers, vol. xix., 1888, p. 89.)

The first part of this Paper contains some experiments on the loss of head at different velocities in the main, supplying the City of Hoboken. The main on which the observations were made is 20 inches diameter, 75,000 feet long. The quantity flowing appears to have been obtained by noting the number of pump-strokes, and allowing 5 per cent, for slip. The heads were determined by pressure gauges. The friction at velocities of 2 to 3 feet per second is a good deal less than that given by some ordinary formulae. The results agree best with the following —

« = 203-3 ro-ea^gi-sss . . (Lampe) V = 111-0 ^(rs) . . . (Darcy).

In 1882, Hoboken had a population of thirty- three thoiisand, and the water consumption was 4,000,000 gallons per day. The Author was convinced that half this was wasted. By prompt action, partly by extending the use of meters, partly by inspections, about 750,000 gallons per day were saved by preventing waste in the mains, and 750,000 gallons by preventing waste by consumers. Since the waste was checked, a pressure 25 per cent, higher has been maintained in the mains. In 1886, the Water Company was obliged to establish a low minimum rate of charge depending ou

460 FRICTION IN WATER-MAINS. [Foreign

the size of meter attached to the service pipe, in order to check a tendency to extreme economy on the part of some consumers. Experience shows that the meter system is advantageous both to the consumer and the Company. Also that the true policy of a Water Board is to furnish meters and keep them in repair free of expense to the consumer, except in the case of misuse or freezing.

Some data of measurement of the waste in mains are given. The following is specially interesting: — A new 24-inch main, 11 miles long, was laid by the Hackensack Company. At the time of test it was supplied with gates, air-cocks, and blow-offs, but no hydrants. The main had been carefully examined, and there was no evidence of leakage ; about midway it was connected to an iron tank at the summit of the whole system. The test was made by opening the gates and filling the tank under a pressure of 107 lbs. per square inch ; the gates were then closed. The loss was found by tank measurement in several trials to be 70,000 gallons per day. The main had good lead joints, well caulked and laid by experienced men under careful inspection.

w. c. u.

Tlie East Orange Sewage- Works.

(Scientific American, 17 Nov. 1888, p. 307.)

The townshijD of East Orange, a typical suburban community of New York, has recently been provided with sewage-works, designed by Mr. C. P. Bassett, associated with Mr. E. Hering, M. Inst. C.E. The pipes run from 8 to 2-i inches in diameter, and the sewer- gradients vary from 1 in 30 to 1 in 800. Many difficulties were encountered in laj-ing the sewer-jiipes, owing to the presence of quicksands and rock, which latter necessitated tunnelling. The outfall works provide for a chemical treatment combined with intermittent filtration.

The precipitants employed are lime and sulphate of alumina, which are mixed with the sewage-water in a trench 100 feet in length, furnished with baffle-plates. The sewage is received in tanks, where the suspended matter subsides, and the effluent is clarified by intermittent downward filtration, being drawn oif from the sludge by a floating swing-pipe. The sludge is conveyed into vacuum-pans, and forced from thence into filter-presses, where it is formed into cakes in the usual way. The works have been five months in operation. No details are given of the cost, or of the amount of the sewage-flow, nor is the area of the farm stated. The article is accompanied by illustrations showing the works and plant.

G. E. E.

Abstracts.] DISINFECTING ACTION OF SUPEKHEATED STEAM. 461

On the Disinfecting Action of a Current of Suiierlieated Steam. By Professor Max Gruber.

(Gesiuidheits-Ingenieur, Oct. 15, 1888, p. 674.)

The Author alhides to his previous observations on this subject, and to the conditions laid down for successful disinfection, namely, that the steam must be as nearly pure as possible, that is, free from admixture with atmospheric air ; and, second, that the steam must become condensed upon the object under treatment by means of dripping water, if the germs are to be rapidly destroyed. He also calls attention to a statement in the ' Zentralblatt fiir Bak- teriologie und Parasitenkunde,'^ that " The xase of moist steam is preferable to dry superheated steam for purposes of disinfection." A simple form of apparatus devised by Dr. Von Esmarch for testing the action of steam is described, and the results of experiments with steam at 100^ Centigrade, and with steam superheated to 110°, from 120^ to 123°, and lastly from 140° to 150°, are described. With moist steam at 100°, the spores of anthrax were killed, without any exception, in from five to ten minutes. In super- heated steam at 110°, complete sterilization of the spores was not attained in less than twenty miniites ; and as much as half an hoiir's exposure was required when the temperature of the steam rose to 120° to 123°. Even under the action of steam at 150°, ten minutes failed to suffice for the destruction of the spores, and complete success was only attained by still further prolonging the ex]ieri- ments. Similar tests with the spores of the garden-mould bacillus gave corresponding results. The drier the steam, the less energetic was it in its action. Steam superheated to from 120° to 130° acts more slowly than steam at 110°, and it is not until a temperature of from 140^ to 150' is reached that a destructive action, due to the high temperature, is set up, similar to that produced by hot air.

For the purpose of these trials, the test-objects were merely enclosed in an envelope of filter-paper, and it was to be expected that, in the case of trials upon a large scale with test-spores con- tained in clothing, &c., the results would be more favourable to superheated steam, as the amount of steam which would penetrate to the centre of the bundles would be small, and moreover, it woTxld be cooled down to condensation-point. This was found in actual practice to be the case, and the results of three trials made by Dr. v. Esmarch with a Henneberg's disinfecting apparatiis are given. The spores were enclosed in rolls of blankets, which had been provided with a signal thermometer adapted to give an alarm as soon as a temperature of 100° Centigrade was attained in the centre of the roll. In each case the germs were sterilized, though, similar test-objects on the outside of the bundles were only in part

> Vol. iii., p. 638.

462 DISINFECTING ACTION OF SUPERHEATED STEAM. [Foreign

deprived of vitality. The Author draws attention, also, to Yon Esmarch's statement respecting the necessity for a strong current or jet of steam, as simple steaming had a less destructive effect on the spores.

G. E. E.

Elucidation of the Bisinfecting-Pou-er of Steam. By A. Walz, of Dusseldorf.

(Gesundheits-Ingenieur, Nov. 1st, 1888, p. 697.)

In the case of some previous experiments conducted by the Author, which led him to form an opinion that the disinfecting action of the steam was not due to its jiropulsion as a jet, but to the extent to which it was generated under 2:)ressure, his views were called in question by Professor Gruber. He now draws atten- tion to the impossibility of caiising a jet of steam to jiass through a bundle of clothing in a disinfecting chamber, unless the chamber is filled with steam under pressure ; the matter to be decided is whether the steam passes into, or out of, the space to be disinfected as a jet at 100^ Centigrade. He examines the statements of Pro- fessor Gruber in the ' Gesundheits-Ingenieur,' No. 20,' and shows that the results of the experiments of Dr. von Esmarch, on which Professor Gruber relies, might with miich probability be referred to the action of steam under pressure ; and he states that the steam, moreover, was in each case mixed with atmospheric air.

G. E. E.

Comparative Trials of various Gas-Burners. By S. Lamaxsky.

(Journal fiir Gasbeleuchtung und Wasserversorgung, 1888, p. 629.)

During the exhibition of gas-lighting apparatus held in St. Petersburg by the Eussian Technical Society, an extensive investi- gation of various gas-burners was made. All the tests were made with the same gas and by the same observers, thus rendering it possible to compare the different types of burners with each other and also the burners of different makers ; the illuminating-powers were also determined with varying rates of consumption to ascer- tain the most favourable conditions for each. Batswing, argand, incandescent, and inverted regenerative burners were tested, a Bunsen photometer being employed with a standard argand burner, the illuminating-power of which was verified every evening with English standard candles. The illuminating-power of the argand burner, with a consumption of 5-29 cubic feet per hour varied between 14-5 and 14*75 standard candles. An angle photometer

' Ante, p. 461.

Abstracts.] COMPARATIVE TRIALS OF VARIOUS GAS-BURNERS. 463

was Tised for determining the illuminating-powers of the inverted regenerative burners.

The batswing burners tried were the so-called " hollow-top " burners, the surfaces of the flames being parallel to the photometer disk ; with five descriptions of batswing-burners, and consumptions varying between 5 '07 and 11*29 cubic feet per hour, the highest results were obtained with Bray's 80-candle-power burner, con- suming 9 • 32 cubic feet per hour, when it gave 2 • 25 candles illuminating power per cubic foot of gas consumed, while the lowest was the ordinary hollow-top 7 cubic feet per hour burner, which gave 1 "81 candle per cubic foot per hour.

Eighteen varieties of argand burners were tried, the highest results being obtained with the Rotsiper burner with heated-air supply, consuming 8-34 cubic feet per hour, which yielded 3*75 candles illuminating- power per cubic foot of gas per hour ; and the lowest was the Parisian union burner, consuming 7*37 cubic feet per hour, with a duty of only 1 • 15 candle per cubic foot per hour.

Of incandescent biirners only two, the Sellon-Lewis and the Welsbach, were exi:)erimented with. Of these, the Welsbach, con- siiming 3 • 28 cubic feet per hour, gave the best result, namely, 2 • 93 candles per cubic foot, the light from it being whiter and more agreeable than that obtained from the Sellon-Lewis burner. Tliese incandescent burners are, however, not much more eco- nomical than argand burners, and require more careful treatment.

The inverted regenerative burners, of which nine descriptions were tested, were all, with the exception of the small Cromartie Ijurner, tried at angles of 30'^, 45^, 60^, and 75', the best results l)eing obtained with the No. 11 Siemens burner, which, with a consumption of 42 '95 cubic feet per hour, gave 6*72 candles per cnl)ic foot, while the lowest was the Butzke lamp, consuming 15*17 cubic feet per hour and yielding only 3* 18 candles per cubic foot, these results being the average of the observations at the four angles.

Comparing the lowest consumption of gas per candle of illumi- nating power for the various ty2:)es of biirners, it appears that the batswing, argand, and regenerative burners stand in the relation of 3 : 2*1 : 1, or that the regenerative burners are three times more economical than batswing,^ and nearly twice as economical as argand burners ; to determine their absolute economical value the cost of fitting and maintaining must however be considered.

CO.

' This comparison between the Argand and incandescent burners does not agi-ee with the preceding figures. With the Argand giving 3-75 candles illuminating-power pev cubic foot, and the incandescent burner 6 "72 candles per cubic foot, the proportion would be 1 '79 : 1 instead of 2 : 1, but the figures given are those in the original. — C. G.

464 WILMSMANN*S SMOKE- CONSUMING FURNACE. [Foreign

Wilmsmanns Smohe-Consuming Furnace.

By — SeilER, of Mannheim.

(Journal fiir Gasbeleuchtung und Wasserversorgung, 1888, p. 135.)

The progress ruade in various industries has attracted considerable attention to the question of smoke-consumption, and numerous suggestions have been made for its accomplishment by many who have not realized the difficulties which surround the problem. Although generator firing has been extensively applied in gas manufacture, smoke-consuming furnaces are rarely used for steam boilers, on account of the costly nature of the arrangements required, and the difficulty of applying them to existing boiler settings. A simple and cheap smoke-consuming and fuel economis- ing arrangement is to be found in the furnace of Mr. Wilmsmann (Mining-Director of Hagen, Westphalia), which can be easily applied to all descriptions of boiler and other furnaces. The arrange- ment consists, mainly, in converting the front part of the furnace into a generator chamber, by forming a fire-brick shield and division wall above the fire-bars, near to the bridge of the furnace, to keep the gases and smoke from going direct into the flues ; the fuel being put into the furnace so that it slopes up against the shield, and the gases and smoke being compelled to pass through a red hot layer of fuel, thus ensuring their perfect combustion. In starting the furnace, the fuel is at first burnt for three-quarters of an hour in the same manner as with an ordinary furnace, but with the difference that, the incandescent layer of fuel is gradually increased in depth until there is sufficient to form an inclined heap against the baffle, and thus close the combustion-chamber from the fire-space. As the coal burns, carburetted-hydrogen is liberated, and combines with the air entering through the fire-bars, while the coal, partly converted, into coke, burns off with a further supply of air. The gases formed on the upper surface of the red-hot coals in the combustion chamber are conveyed through channels at the top of the furnace into the back chamber, where they require a secondary air-supply for complete combustion, for which heated air is supplied through air- channels regulated by dampers.

The formation of smoke, which occurs with ordinary furnaces after supplying fresh fuel, is entirely avoided with the Wilmsmann furnace, if the gas-chamber is separated from the fire-space and the flues, by the glowing layer of fuel. The principle of this system of firing is similar to generator furnaces for gas-retorts in which the hydro-carbons from the fresh coal have to pass a glowing layer of coke. In both cases carbonic acid is formed at the com- mencement of the combustion, Avhich, by passing through the o-lowing coke, takes up a further quantity of carbon, and is reduced to carbonic oxide, and this, if sufficient air is present, is burnt to carbonic acid.

^ta

Abstracts.] WILMSMANn's SMOKE-CONSUMING FURNACE. 465

Many unsuccessful attempts have been made to apply generator firing to steam-boilers, partly on account of the cost and partly from other difficulties, and also because, although, with inter- mittent working, the smoke could be consumed, there were other pecuniary disadvantages connected with it, but with the present extremely simple system smoke consumption must succeed ; in fact, about 600 Wilmsmann furnaces have been put to work within the last few years for steam-boilers, smelting, puddling, and other furnaces in the Ehenish provinces, Westphalia, and the South German States, and most favourable testimonials have been received, not only with regard to the prevention of smoke, but also as to economy of fuel, varying from 15 to 30 and even 40 per cent, as compared with ordinary furnaces.

Experiments as to steam-production have been made with a horizontal Cornish boiler, fired with an ordinary and also with a Wilmsmann furnace. With an ordinary furnace a trial of eleven hours was made, during which the consumption of coal was 1,109 lbs., and the water evaporated 7,0-45 lbs., the average feed-water temperature being 104° Fahrenheit, and the average steam-pressure 4^ atmospheres. This gives 6 • 35 lbs. of water evaporated per lb. of coal.

With the Wilmsmann furnace, fitted to the same boiler, a trial was made for seven days of eleven hours per day. The total coal consumed was 4,916 lbs., and the water evaporated 46,751 lbs. ; the average temperature of the feed-water was 140° Fahrenheit, and the average pressure of steam 4^ atmospheres, which gives 9 • 5 lbs. of water evaporated per lb. of coal. The coal used in both cases was Ruhr nut coal.

Mr. Vogt, Chief Engineer of the Berg Steam Boiler Examination Society, has also spoken very favourably of the Wilmsmann furnaces as being sound in principle, of great simplicity, and small in cost for fitting.

C. G.

Raising the Steamer " Ferndale," sunk in the Entrance Channel of the Port of St. Nazaire.

By Messrs. Kerviler and Preverez,

(Annales des Fonts et Chaussees, 6th series, vol. sv. 1888, p. 1030, 1 plate.)

On the 4th of February, 1888, the English steamer "Ferndale," of 604 tons, loaded with coal and tar refuse, was run down by the English steamer "Dowlais," and sunk almost immediately, only giving time to draw it alongside the northern jetty before its bow plunged into the silt. The hinder part was readily unloaded, as it remained afloat for two days ; but the front part of the hold was only out of water just at low-tide. The rent caused by the bow of the "Dowlais" was about 13 feet long, and, at 6i feet below the water-line, it suddenly enlarged from a width of 4.V inches to 2 feet

[the INST. C.E. vol. XCV.] 2 II

466 RAISING THE STEAMER " FEKNDALE." [Foreign

for a height of about 5 feet, having been torn open by the anchor of the " Dowlais " which hung across its bow. The narrow parts were closed by poplar wedges with packing. A large wrought- iron plate, ^ inch thick, lined round its edges with two large tresses of greased cotton, was placed over the large hole, and was fixed in place by being bolted to two very strong T irons put across the opening inside. Eolls of tow were interposed where the plate was not close to the torn plate-iron skin, and the whole was covered with a large felted sheet. Two rotary pumps, capable of discharging 000 tons of water each per hour, were placed upon pontoons along- side the sunken vessel. One of the pumps discharging water from the front hold enabled the unloading from it to proceed for two or three hours each tide. On the 13th, the two pumps pumped the water out of the holds and engine-room, so that the stern soon rose slightly; and at half-tide the bow was raised from the bottom. The vessel, accompanied by the pontoons, was then taken into the dock ; the cargo was further removed from the front hold, and on the 14th, the vessel was placed in the graving-dock, which was pumped dry the next day.

L. V. H.

Consolidation of Earthworks on the Railway from Gien to Auxerre. By Messrs. Lethier and Joyan.

(Annales des Fonts et Chaussdes, 6th series, vol. xvi. 1888, p. 5, 2 plates.)

The portion of the Gien- Auxerre Eailway between the Toucy- Moulins station and the Auxerre junction with the line from Laroche to Nevers has a length of about 18^ miles, and passes almost entirely over the clayey greensands and eocene strata. Seventeen cuttings were excavated in treacherous clays, having a total length of about b\ miles, with slopes of 3 to 2, except in four cases, where the slopes were made 2 to 1. The slopes were covered with moiild, drained, sown and planted ; and trenches were carried along the foot of the slopes, lined with dry pitching, or occasionally laid in mortar. Water-channels with a sharp fall were formed on the top of the cuttings. The wet clayey portions of the formation were drained by longitudinal and transverse rubble drains. Bad slips occurred at only three places, owing to the presence of springs ; and the rej^airs of the slips increased the cost of the consolidation works from the original price of 25s. 3^cZ. per lineal yard for the works described above, up to 27s. for the repairs of the slips as well. Nine of the embankments experienced serious slips directly after their completion, which necessitated large and costly works of consolidation ; and the Fritton's embankment has been selected for description as being the most seriously damaged, and giving rise to the most complete works of reparation. This embankment, about 550 yards long, crosses a small valley about 36 feet deep; it was constructed in 1880-81, with a top width of

Abstracts.] CONSOLIDATION OF EAKTHWORKS. 467

19| feet and slopes of 3 to 2. Slips soon occurred in the middle, and towards the base of the slopes. In order to drain oif the water inside the embankment, which caused the slips, rubble-drains were made in the slopes parallel to the railway, restin^^ uj^on the solid ground, about 10 feet high and from 2 to 2| feet wide, and propped up on the outer side by a bank of punned earthwork, having a benching 8j feet wide at the top. The water collected by the rubble-drain on each slope was discharged beyond the embankment by trans- verse rubble-drains 50 feet apart. A sum of £800 had been spent on these repairs when the constant rains of the summer and autumn of 1882 disintegrated all the higher parts of the embankment, causing a settlement of over 2 feet at the top, and slips which, commencing at the top, reduced the formation-width to 10 feet in some places. The previous repairs stood well for the most part ; but the large slips over the bench threatened to carry it away. The consolidation and drainage was effected by spurs of rubble- stone, perpendicular to the line of railway, 6 ^ feet wide and 39. \ feet apart, cutting into the embankment from top to bottom, right up the slopes, and into the formation width on each side to the extent of the slij)s, nearly vertical on the inner face, and following the line of the slope on the outer face. The spur rests at the ground- level on a watertight layer of concrete ; and a passage, 1 foot by 8 inches, formed through its base, serves to discharge the water and to supply air to the interior of the mass, which is very favourable to its drainage. To repair and prevent slips between the spurs, transverse rubble walls of an inverted V shape, 3j feet wide, were formed from spur to spur near the top of the slopes ; they were founded upon laj^ers of concrete carried down at least 2 feet below the base of the slip, and passages formed through the base of these walls or drains convey the water collected by them to the adjacent spurs. These works were completed by the beginning of 1884 ; they absorbed 11,120 cubic yards of rubble and broken stone, and cost £6,720 on a length of 262 yards repaired, or £25 10s. per lineal yard, the high price being due to the extent of the work and the dearness of materials. Including the earlier repairs, the cost amounts to £28 10s. per lineal yard, and 2s. O^d. per cubic yard of earthwork. Details are also given of the cost of the repairs of the eight other embankments which experienced serious slips. The sj)urs and the top of the embankments settled gradually till the end of 1 885, the greatest amount of settlement at the highest embankment of Fritton's being 17 inches. This was made up with rocky excavations before the opening of the line on 30th December, 1885, since which period the spurs have not shown any further signs of settlement ; and though the embankment between the spurs has settled a little, no further work has been necessary beyond the ordinary maintenance of the road.

The practical conclusions to be drawn from the experience on these embankments, as well as from those on the Yonne and Cher lines, are as follows: (1) In treacherous clay embankments, con- solidations by benches at the top, with longitudinal and transverse

2 11 2

468 CONSOLIDATION OF EAKTHWOEKS. [Foreign

drains, are unable generally to stop slips ; (2) The best plan consists of nibble-spurs about 33 to 40 feet apart, resting on an impermeable floor, supplemented, where necessary, by inverted V-shaped rubble-walls carried down into the undisturbed portion of the embankment ; (3) It is expedient to make these spurs 6^ feet wide when the embankments exceed 20 feet in height and the slips are large ; (4) It is always possible to construct these rubble-spurs and trenches, like those at Fritton's, in an embank- ment which has slipped, with safety, and without excessive expense ; (5) It is assumed in this method of consolidation, that the injured embankment rests upon a fairly solid soil ; but it is nevertheless wise to provide against possible settlement, by giving the top of the consolidated embankment an excess of width of one- twentieth to one-tenth of the height of the embankment.

L. V. H.

The Laon Steep-Gh'adient Raihvay. By A. Braxcher.

(Le Genie Civil, vol. xiii. 1888, pp. 75 and 169, 4 woodcuts.)

The tramway laid down along the high road to connect Laon with the station of the Northern Railway Company, has a gauge of 1 foot 11§ inches; its rails are of steel, weighing 19i lbs. per lineal yard, carried on steel sleepers, 3 feet 3^ inches long, hollow underneath, and closed at the ends on Captain Pechot's system. The line, which surmounts a difference of level of over 328 feet in about l;j mile, has necessarily steep gradients, reaching to 1 in 15 "4, with curves having a minimiim radius of 1^ chain. The traction is effected by a compound locomotive, on the Mallet system, weighing 8f tons when empty, and lOf tons when loaded. This engine has four axles, to make the load borne by each as little as practicable ; all its wheels are driving-wheels, to utilise the whole weight of adhesion ; and the front set of wheels are made convergent, so that the engine can go readily round curves of 1 chain. The heating-surface is 254 square feet ; the hinder, high pressure, cylinders have a diameter of 6j inches, and 10 J inches stroke; whilst the front cylinders, low pressure, have a diameter of 10 inches, and the same stroke ; and the wheels are 1 foot llf inches in diameter. The total wheel-base is 9 feet 2 inches ; but the distance apart of each pair of axles is only 2 feet Qh inches. The locomotive has a tractive-force of 1"43 ton at the tires of the wheels when running compound, but by the admission of live steam into the four cylinders, this force can be raised to 2 "07 tons. This latter method of working, however, is only exceptionally used, to overcome excessive temporary resis- tances. The carriages are on bogies, and have seats for twenty- four passengers, and standing-room for eight more ; so that the ordinary train of three carriages can convey ninety-six passengers, amounting, when filled, -to a load of 12 to 13 tons. The laying of

Abstracts.] THE LAON STEEP-GRADIENT RAILWAY, 469

the line was effected by fifty workmen in three days with satis- factory results. The locomotive, designed by Mr. A. Mallet, is of a quite novel type, and is able to go up a gradient of 1 in 14-3, on a curve of 1^ chain, at a speed of 19 J miles an hour, drawing a load of 10 tons. The first locomotive manufactured with this object, by Messrs. Decauville at their Petit-Bourg works, was the " Lilliput " locomotive of 1 ton weight, exhibited at Compiegne in 1877, where it drew sixty passengers at a time over a portable line of 2-foot gauge and 11-lb. rails. Subsequently, the same firm constructed locomotives of 2i, 3, 5 and 6 tons, before manufacturing the locomotive described al)ove, which appears destined to prove very useful in extending traffic on railways where heavy rolling-stock is unsuitable.

L. V. H.

Cost Prices on Railways.^ By G. Eicour.

(Annales des Fonts et Chaussees, 6th series, vol. xv. 1888, p. 534, 2 diagrams.)

The present article is an extension of the Author's previous article, and contains a discussion of the objections raised by Mr. Noblemaire,^ manager of the Paris, Lyons and Mediterranean Eailway, against the principles laid down by the Author in his first article. The relations between the cost price of the unit of traffic, the amount of traffic, and the gradients, are first considered, the unit of traffic being taken as one passenger, or 1 ton of goods, carried 1 kilometre. The cost price per unit of traffic of various systems of railways can be no more compared directly than their coefficients of working. After obtaining the amount of traffic for each system, the corresponding normal cost price must be deter- mined, either by the Author's diagram or formula ; and then this normal cost price must be compared with the real cost price. It is the ratio of these prices, which the Author terms the typical ratio, which is the measure of the economy effected in working. The following Table gives the respective results of the working, in 1884, for the large railway companies, and the State railways.

Large Railway State

Companies. Railways.

Average traffic 712 252

£. s. d. £. s. d.

Normal cost of 1,000 units of traffic, per tou-mile 2 10 7 2 19 3

Real cost „ 1 18 3 2 10 5

Typical ratios 0-98 0-85

Coefficients of working 0 • 53 0 • 80

In comparing the two systems of railways according to their coefficients of working, instead of by their typical ratios, the

Minutes of Proceedings Inst. C.E., vol. xci. p. 501. Ihid., vol. xciii. p. 517.

470

COST PRICES ON RAILWAYS.

[Foreign

important iBfluence of traffic is entirely neglected ; and, in this way, the State railways were sui^posed to be worked expensively ; whereas, in reality, they manifest a notable economy over the large private lines.

When it is desired to limit the comparison to the cost prices of the rolling stock and traction, it must be based on the cost price of the ton per kilometre of the gross load of goods trains or passenger trains. These cost prices, determined carefully by taking into account the changes in price of the materials consumed, cannot be compared together directly ; but the characteristic gradient must be found for each system of railways ; and then the normal cost price of the ton per kilometre of the gross load, corresponding to this gradient, must be ascertained, either from the Author's diagram or table. It is this normal cost price, for each system of railways, with which the real cost price must be compared. The ratio between these two prices, which is termed the typical ratio, affords the measure of the economy in respect of the working of the rolling stock and traction. A comparison of these various data for the Paris, Lyons and Mediterranean Eailways on the one hand, and for the State railways on the other, as gathered from the results of working in 1884 and 1885, are given in the following Table.

Paris, Lyons and Jledi- terranean Kailways.

State Railways.

Characteristic gi'adients .

Normal cost of 1,000 ton-miles ofl gross load of goods trains . . /

Real cost of 1,000 ton-miles ofl gross load of goods trains . . /

Typical ratios

152-67 155-04

d. 11

s. d. 4 8

2 1

0-92

1 19 9

0-89

96-43

£. s. d. 3 0 7

2 9 2

0-81

93-72

£. s. d. 3 1 10

2 9 1 0-79

L. V. H.

Signalling-Apparatus on the St. Gothard Railway. By — Cox.

(Zeitschrift des Vereines deutscher Ingenieure, 1888, p. 1020.)

The difficulty experienced by the brakesmen on the goods trains in understanding signals given from the engine, while traversing the long tunnels, has led to the introduction of an electric arrange- ment devised by the Author. The wire is connected with each brakesman's lantern (there being a brakesman to every three or four triTcks), and Ijy pressure on a small pin a slide is made to rise or fall, so as to show a red;^ green, or white light. By repetition or

Abstracts.] SIGNALLING- APPARATUS ON THE ST. GOTHARD RAILWAY. 471

combination of flashes of these lights a proi:)er code of signals is established. The red light signal can be given by any of the brakesmen ; the white or green light by the engine-driver only. If the couplings break, the red light is shown automatically.

P. W. B.

Diminution of Earth- Temperature in Deep Mines.

(Oesterreichische Zeitschrift fur Berg- und Hiittenwesen, 1888, p. 199.)

At Pribram meteorological and magnetic observations have been carried on for some years at the thirtieth level of the • Adalbert mine, at a depth of 1,000 metres from the surface, or 4G5 metres below the sea-level in the Adriatic. The thermometer used for taking the temperature of the rock can be read to one-tenth of a degree Centigrade, and is protected against draughts in an enclosed space, where the air-temjierature has remained unchanged at 2-i° C. for five years. The observed temj)eratures of the rock were : —

In 1883 24-5° Centigrade.

At the end of 1885 24-3°

And at the end of 1887 24-1° „

Showing that the rock has been cooling at the rate of one-tenth of a degree j^er annum since the observations were commenced.

H. B.

An Apparatus for Measuring Earth-Pressure Underground.

(Zeitschrift fur das Berg- Hiitten- und Salinenwesen, 1888, p. 244.)

In order to measure the vertical changes due to the thrust of tlio unsupported ground in levels, the following contrivance has been adopted in the Prussian Government salt mine at Stassfurt. A wooden plug about 1 metre long is driven into a hole in the potash salt bed forming the floor of the level, and a second similar one in the salt clay of the roof vertically above the first. A length of wrought-iron tube, somewhat shorter than half the height of the level, is screwed to the end of each of the plugs, so that the tubes do not exactly meet, but a connection between them is made by a guide-pin attached to the lower one, which slides freely up and down within the upper one. A board placed on one side of the tubes is attached by two clamps to one wall of the level, which keeps it in a vertical position, but without touching either the roof or the floor. From the outside of the board project two light metal frames, forming the centres for a pair of unequally ariued levers, one of which is connected with either tube on its shorter side, while the longer one moves over the face of one of two graduated scales drawn on the board. If, in couseqiTence of downward pressure

472 MEASURING EARTH-PRESSURE UNDERGROUND. [Foreign

in the salt rock the roof is depressed, the pointer on the upper scale rises, and as the arms of the levers are on the ratio of 1 to 10, a depression of 1 millimetre alters the position of the pointer- arm by 1 centimetre. With the lower tube and scale, a rise in the floor by upward thrust is indicated in a similar manner by the movement of the lower pointer in the opposite direction.

The first trial of this apparatus was commenced on the 23rd of May 1886, when it was set up in the end of No. 26 level, driving north, the pointer being adjusted to zero. Between that date and June 4th, 1887, or in a period of nearly a year, the upper pointer had risen 8j centimetres and the lower one fallen 11^ centimetres, corresponding to a sinking in the roof of 8f millimetres, and a rise in the floor of 11^ millimetres. In another trial in the north end of No. 16 level, the roof sank 26 millimetres, and the floor rose 1 millimetre between April 24th, 1886, and June 4th, 1887.

H. B.

Differences of Level in the Mines of Austria and Hungary. By F. E. M. VON Friese.

(Oesterreichische Zeitschrift fiir Berg- und Hiittenwesen, 1888, p. 321.)

In a memoir on the determination of altitudes in mines, which reviews the methods adopted of fixing datum points in the different mining districts and notices the fact that no general authoritative datum has as yet been adopted, the Author gives a complete list of the depths of the principal workings in the Austro-Hungarian dominions. Some of the more important conclusions are given in the following abstract. The highest point at which mines are worked is at the Goldzeche, on the frontier line between Carinthia and Tyrol, 2,925 metres above the sea-level, and the lowest, the sump of the Maria shaft at Pribram, 537 metres below the sea, or a total range of 3,462 metres.

The greatest depths below the sea in the different districts are —

Mines. Metres.

Pribram, Bohemia. Silver lead 537

Ostrau-Karwin, Moravia. Coal 225

Kladno, Bohemia. „ 162

Rossitz, Moravia. „ 71

Wieliczka, Poland. Salt 14

The brown-coal mines of Briix and Dux-Osseg, in Bohemia, and the salt mines of Maros-Ujvar in Transylvania, are nearly down to the sea-level.

The vertical range of metallic-ore mining is, as given above, 3,462 metres, from 2,295 metres above to 537 metres below sea- level ; that of salt mining, 1,951 metres, from 1,928 metres above to 23 metres below sea-level ; that of Alpine (tertiary) coal mining, 431 metres, from 867 to 436 metres above the sea ; and that of the coal-measure districts in Bohemia and Moravia, 634 metres, from 409 metres above to 225 metres below the sea.

Abstracts.] LEVELS OF MINES IN AUSTRIA AND HUNGARY. 473

Some of the deepest individual shafts are —

Shaft. Metres.

Pribram. Maria 1080-88

„ Adalbert 10G9-36

„ Anna 9-12-98

„ Franz Joseph 880 • 68

„ Procopi 879*52

Joachimsthal. Einigkeit 533

Schemnitz. Amalia 540 "IT

Elisabeth 448-37

„ Andreas 432-82

Kladno. Mayrau 520

Miroschau. Libusehin 477

Barre' 446

Wieliczka. Joseph 262-99

Elizabeth 259-61

The Hallstatt salt mines, which are worked entirely by solution, have a range of 503 metres, from 1,233 to 730 metres above the sea-level. The Schneeberg zinc mines in T^toI, also worked by levels, are at present worked between 2,522 and 2,2-16 metres above the sea.

H. B.

On the Relations hetiveen Seismic and Atmospheric Disturbances

and the Disengagement of Fire-Damp.

By G. Chesnau.

(Annales des Mines, vol. xiii. 1888, p. 389.)

After referring to previous researches bearing upon the subject, the Author proceeds to describe an extended series of observations made at the Herin pit of the Anzin Coal Company. The micro- seismic movements of the earth's crust, as he calls them, in dis- tinction to sensible tremors or earthquakes, were observed by a tromometer similar to that employed at the Douai School of Mines, ^ consisting of a pendulum 1 • 50 metre (4 feet 1 1 inches) siispended from an iron pivot fixed in a thick wall, its oscillations being read by a microscope. The wall was a very old one, of brick, 2 feet thick, running in a direction from N.E. to S.W., and completely sheltered from the sun. The most violent winds are from the S.W., and a series of observations extending over several months appears to establish the fact that the wind exercises no appreciable influence upon the apparatus, but that there are certain coincidences between seismic and barometric movements which Bertilli, of Florence, has named "baroseismic squalls." Local vibrations, such as those produced by passing vehicles, are also inappreciable, being too rapid to affect the weight or bob of a pendulum of this length, although they may cause nodes and curves in the sus- pension-rod. A shorter pendulum, while indicating these vibrations,

' The Douai tromometer was fully described and illustrated in the " Annales des Mines," vol. ix. p. 241.

474 ATMOSPHEKIC DISTURBANCES AND FmE-DAMP. [Foreign

would fail to record microseismic movements having a period of oscillation greater than one second ; and this is probably the reason why pendulums of from 1 to 4 metres length (39 to 157 inches), with a period of oscillation of from 1 to 2 seconds, are found to give the best results for this purpose.

Observations, taken simultaneously at Herin and Douai, during the months of October, November, and December, 1887, show a close correspondence, except on November 22nd and 23rd, when the violent seismic disturbances in the Mediterranean basin, although distinctly transmitted to the department of the Nord, appear to have escaped the Douai observers ; and, generally, the Herin in- strument would appear to be far more sensitive than that at Douai.

A similar concordance has been noted between the monthly averages recorded at the observations of Bologna, Florence, and Velletri.

A series of barometric observations taken at bank, at the pit bottom (558 feet deep), and in the return air-course, were found to correspond so closely as to render it immaterial which was adopted as the standard of comparison ; while from a similar series of observations, at diiferent hours of the day, as to the disengage- ment of fire-damp (which it was found could be estimated very closely by means of the cap or aureole on the flame of the Pieler safety-lamp),^ it was found that the results obtained at 6 a.m., half an hour after the miners had assembled, but before they had fairly commenced work, gave a very close average for the whole day, and was unaffected by the greater or smaller amount of coal actually got from the faces. Erom the percentage of fire-damp found in the return air-course, the daily amount given ofi" was found by reducing the quantity of air actually passed through the j)it to a given unit.

A diagram is given for the eleven months, February to De- cember, 1887, showing (with certain interruptions from various causes, by which the daily observations are reduced to 180 out of a possible 230) (a) the intensity of the seismic movements ; (b) the barometric readings ; (c) the amount of fire-damp given ofi" ; and the following general deductions are offered : — ■

(1) Seismic disturbances and flow of fire-damp —

Coucordant 81 clays.

Discordant 46 „

Independent 51 ,,

178 „

(2) Barometric readings and flow of fire-damp — -

Concordant 75 days.

Discordant 51 „

Independent 54 ,,

180 „

Minutes of rrocccdinj:;s lust. C.E., vol. xc. p. 170.

Abstracts.] ATMOSPHERIC DISTURBANCES AND FIRE-DAMP. 475

But in order to eliminate such sources of error as might arise from falls of roof, &c., &c., the value of which cannot correctly be estimated, it is considered best to take into account only the more marked variations of the curves as plotted. Taking the tro- mometer curve as the basis, there are 43 concordances to 17 dis- cordances; or, inversely, if the fire-damp curve is taken, 17 con- cordances to 7 discordances ; or in each case, nearly as 5 to 2.

Taking the barometer curve as the basis, there are 1 1 concordances to 9 discordances; or, inversely, 14 to 15.

It would thus appear that the connection between seismic dis- turbances and disengagements of fire-damp is fairly demonstrated ; but that the influence of barometric variations is not so well established, although the concordances outnumber the discordances ; and it is interesting to remark that whenever a large increase of fire-damp has been maintained for several days above the average, including the very exceptional outburst on December 8th and 9th (of which some interesting particulars are given), there is a very remarkable coincidence between all the three curves.

The Paper contains references to numerous other investigations on the same subject, and especially to an article in Vol. xxxvii. of Proceedings of the North of England Institute of Mining and Me- chanical Engineers, containing the first report of the committee appointed to investigate this subject ; and to one in the Zeitschrift fiir das Berg-, Hiitten-, und Salinen-Wesen, 1887, pp. 277, 279, con- taining the result of Mr. Nasse's careful observations. In this case one district of a mine, which had been walled off to isolate a fire, afforded a convenient gas-holder of known area and capacity ; and his experiments show that a diminution of atmospheric pressure, as observed on the spot, was decidedly favourable to the disengage- ment of fire-damp.

W. S. II.

Shaft-Sinking hy Haase's Method.

(Zeitschrift fiir das Berg- Hiitten- und Salinenwesen, 1888, p. 225.)

At the brown-coal mine of Guerini, near Cottbus, a seam of lignite of an average thickness of 7 metres with a roof of bituminous clay 1 -20 metre thick, is covered with about 26* 65 metres of sand very full of water. A pumping-engine shaft commenced with ordinary timbering, reached the water-level at 3 metres, and the sinking could only be carried 3 metres deeper before the flow of water and sand became so great as to endanger the work. It was then determined to continue the sinking by Haase's method, which consists in driving down a series of parallel wrought-iron tubes side by side to form an impermeable wall, the ground being at the same time partially drained by the flow of water from the tubes. The shaft is rectangular in form, 3-30 x 2 • GO metres within the original timbering, and 2 • 90 x 2-20 within the wall of tubes. The

476

SHAFT-SINKINa BY HAASE S METHOD,

[Foreign

tubes are supplied in 4-metre lengths and are 5 millimetres thick with an internal diameter of 105 millimetres. In order to keep them straight during the driving, carefully-planed wooden guides were screwed to the timbering of the shaft for a depth of 3 metres with cast-iron cross-bars at the top and bottom fitting the curvatxire of the tubes. The depth of ground to the coal was 21-85 metres, requiring six sets of 4 metres in length of the patent tubes. The work was begun on May 14th, 1887. On the 25th of October following the clay above the coal was reached, and the tubes could not be driven down further by pressing. Eighteen rows of tubes were used on the longer and fourteen on the shorter sides of the shaft, or sixty-four in all. The actual time required in driving them was two hundred and twenty-five shifts of twelve hours. The chief items of cost were as follows : —

	11' „=. •,„„^„A■, Lineal Metres Cost per ^I^^^aV K^^ 1 of Tube put Lineal two Mechanics. ! ^^^.^^ ^^4^^
Set No. I. 24 May— 8 June . . . II. 9 June— 22 June . . III. 23 June— 7 July . . . IV. 8 July— 13 Aug. . . V. 14 Aug.— 18 Sept. . . VI. 19 Sept.— 25 Oct. . .	£. «. d. 21 16 5 18 15 0 33 11 9 18 6 9 39 0 3 36 11 9	209-96 237-81 236-84 242-57 243-28 183-58	£. s. d. 0 2 1 0 1 7 0 2 11 0 16 0 3 3 0 4 0
Total for wages . . .	167 17 10 53 17 3 1,676 1 6 204 14 9	1,354-04	0 2 3
Supervision .... Cost of 1,440 metres of tubes and tools Coals for boiler firing
Total cost of lining .	2,102 11 4	1 11 1

When the tubular lining was finished, the removal of the sand from the interior of the shaft and the erection of a permanent iron lining were commenced. Up to the end of November 1887, 9-50 metres had been completed, down to which depth no alteration in the position of the tubes had been observed. The flow of water in the sinking amounted to 1,350 litres per minute, which was kept down by two j^ulsometers. The filtration of water through the tube-wall is only apparent to a height of 3-5 metres above the bottom of the sinking, the ujiper beds having been drained perfectly dry.

^ KB.

Cast-iron Tubbing for Lining Levels.

(Zeitschrift fiir das Berg- Hiitten- und Salinenwesen, 1888, p. 230.)

At the Eschweiler-Eeserve coal mines, near Diiren, cast-iron tubbing, put together in segments in the same way as is usual in

Abstracts.] CAST-IRON TUBBING FOR LINING LEVELS. 477

sinking shafts, has been successfially employed in driving a level through ground broken up by faults.

The tubbing made by Messrs. Haniel and Lueg, of Grafenberg, is 2 • 33 metres outside and 2 • 05 metres inside diameter, the breadth of the rings to the back of the flanges is 140 millimetres and the thickness 40 millimetres. The segments, | of the circumference, are 75 centimetres long, and weigh 320 kilograms, each exclusive of the connecting bolts, which are 32 millimetres in diameter. In the fixing of these heavy segments, some of which required to be lifted nearly 8 feet overhead, a special erecting carriage was used of the following description : —

A horizontal shaft, parallel to the axis of the gallery, is attached to a pair of standards fixed upon a wooden wagon frame, which runs upon a railway laid upon the three floor segments of the tubbing. One end of the shaft carries a large worm-wheel and an arm terminating in a fork fitting into the back of the segment within the flanges, which can be moved radially b}^ a sliding motion. When the segment is taken up by the fork it can be turned radially by the screw into which the worm-wheel gears, the motion being com- municated by a ratchet lever, which moves the screw through one- sixth of a revolution at each stroke, and producing a corresponding smaller movement of the forked arm. When the segment is turned into position, it is finally placed in its seat by sliding the arm outward until the flanges are brought into contact with those of the segment last laid. By this simple contrivance the work was rapidly done, and without accident of any kind. The apparatus is ilhistrated by figures, but no details as to the length of ground lined are given,

H. B.

A New Modification of the Bessemer Process.

(Oesterreichische Zeitschrift fiir Berg- und Hiittenwesen, 1888, p. 142.)

The following modification of the Bessemer process has been introduced by Carlsson of Ulfshytte, in Sweden. The pig-iron treated contains, silicon 1*5 — 2, manganese O'l — 0'15, and carbon 4 per cent., 3*9 per cent, of the latter being graphitic. The charge, tapped directly from the blast furnace, is blown for five or six minutes, biit as soon as the blue flame of burning carbonic oxide appears, the converter is turned and a proportion of the metal is emptied into a ladle attached to a weighing machine. This is known as reducing metal, and contains on an average, carbon 4*15, silicon O'Oo, and manganese 0*07 per cent. The quantity so removed depends upon the weight of the charge and the temper of metal required in the ingots.

The converter is then raised, and the blowing goes on until the remainder of the carbon is oxidized and soft iron is produced ; this usually contains, manganese 0 • 03, carbon 0 • 05, sulphur (maximum)

478 A NEW MODIFICATION OF THE BESSEMER PROCESS. [Foreign

0 • 02 per cent, with traces of silicon, and is generally red almost. High, ferro-manganese is then added, and when the reaction is finished, the final tempering is effected by adding the desilicized reducing metal in such proportion as may he required to produce ingots of the proper degree of hardness. The silicon in the finished metal is usually one-tenth of the carhon.

The advantages claimed for this modification over the ordinary process are as follows : —

1. The reqiiired proportions of carbon, silicon, and manganese are more easily obtained.

2. Steel poor in silicon, with any required proportion of carbon, may be made from the most siliciferous pig-iron.

3. Sound steel products, with silicon and manganese in any required proportion, are easily and cheaply made.

4. A considerable variation in the hardness of the finished steel may be obtained without altering the working conditions of the blast furnace.

5. The formation of pipes in the ingots is more easily prevented.

H. B.

Electrolytic Beduction of Antimony from Ores. By W. Borchers.

(Dingler's Polytechnische Journal, vol. 266, 1888, p. 283.)

Sulphide of antimony, even in very poor ores, as well as all other antimony compounds which are readily soluble in sodium sulphide, may be reduced electrolytically with advantage. For each molecule of antimony trisulphide, three molecules of sodium sulphide should be present in the solution whose density should not exceed 12^ Baume when cold, or 9^ to 10^ when hot; about 3 per cent, (calculated on the total quantity of the solution) of common salt is added, which not only clarifies the liquid and separates dissolved sulphide of iron, but serves to diminish the resistance during electrolysis.

The decomposing cells are iron vessels, which also serve as cathodes, and according to their form the surface of the latter is increased hj the insertion of plates or concentric cylinders of sheet iron, insulated lead plates being placed between them as anodes. The latter are collectively connected with the positive pole and the decomposing cells and iron plates with the negative pole of the battery. An electromotive force of 2 to 2h volts per cell is sufficient to effect the decomposition, the reduced metal appearing eilher in brilliant scales, or in powder, which partly adheres to the iron, and can be easily brushed off", and partly goes to the bottom of the cell.

The reduced metal, after being washed successively with water containing a little sodium sulphide and caustic soda or ammonia, clean water, water acidified with nitric acid, and finally with clean water again, is melted with a little antimony glass, giving a very

Abstracts.] ELECTROLYTIC REDUCTION OP ANTIMONY FROM ORES. 479

pure product. The residual liquor may be utilized as a source of sodium hyposulphite, the dissolved chloride sodium separating in the final evaporation. The presence of hydrosixlphide and poly- sulphides of soda is not injurious to the process as long as the pro- portion of sulphur to sodium in the solution is so regulated by the addition of caustic soda that for each atom of oxidizable sulphur one atom of sodium is present for its neutralization. With a higher proportion of sulphur, or less of sodium, a separation of sulphur and consequent disturbance of the process ensues, while on the other hand more sodium and less sulphur work detrimentally by increas- ing the electrical resistance of the solution.

According to Borchers, sodium hydrosulphide, sodium disulphide, and sodium hyposulphide, are developed on the layers by the initial decomposition of three molecules of water by the current, thus

3H2 0 = 6H-f30

whence the following changes result —

At the cathode Sb., S.^-f 3 NagS + 6H = Sb-f-6NaHS anode 6NaHS-f30 = 3 H2O -f- 3 Na2 83

Probably, therefore, three molecules of water are decomposed to one of trisulphide of antimony, and by the development of three molecules of water, and two of sulphantimoniate of soda, two mole- cules of antimony are set free, thus

2 (Nag S&S,)4-2NaH04-6H =

2 S b -h 2 Na2 S2 + 4 Na H S -f 2 H., 0.

H. B.

Eledrolytie Copper -Befining in Hungary. By A. Soltz.

(C. A. M. Balling's Electro-Metallurgie,' p. 80.)

At Stefanshiitte, in Upper Hungary, copper smelted from anti- monial grey ores is reduced electrolytically from blister copper, which, after partial refining in a reverberatory furnace, contains 6 per cent, of antimony. The copper plates used as cathodes are either made at the works from electrolytic copper, or of commercial refined sheets. In the latter case they are first placed in the bath as anodes until they become perfectly bright. The liquor in the bath contains 200 grammes of copper sialphate, and 20 grammes of sulphuric acid per litre. The dynamo, Siemens and Halske's type Cg gives a current of 100 to 120 amperes and 2 volts electromotive force at 600-650 revolutions per minute. The internal resistance is 0 • 00529 ohm. The external resistance 0*011 ohm. The cathode surface per bath is 2-45 metres, and the density of current 50 amperes per square metre. The six baths in use are each 1 metre

The original is in the Library Inst. C.E.

480 ELECTROLYTIC COPPER-BEFINING IN HUNaARY. [Foreign

in length and depth, and 0 • 5 metre broad. The anodes and cathodes are left in for about two months. The mud deposited in the bath contains 3 per cent, silver, 60 per cent, of antimony, and more than 6 per cent, of bismuth. A new extension of the works is j^lanned to contain forty-nine baths 1 ' 4 metre long, 0 • 6 metre broad, and 1*10 metre deep, to hold eight anodes and seven cathodes each. The cathode surface per bath will be 5 • 6 metres, corresponding to a density of current of 43 amperes per square metre. The dynamo, Siemens and Halske's CH^ gives a current of 240 amperes at 25 volts, when driven at 1,100 revolutions per minute. As, how- ever, with an external resistance of 0*070 ohm, only 17 volts are required, a further number of fourteen baths may be worked by the same power. The productive capacity of the w^orks, based on a deposit of 6 • 9 kilograms of copper per bath daily, will be about 180 tons annually.

^ H. B.

The Smelting of Gold and Silver Ores in Eastern Hungary and Transylvania. By Dr. Schnabel.

(Zeitschrift fiir das Berg- Hiitten- imd Salinenwesens, 1888, p. 128.)

In the aiitumn of 1887 the Author visited the smelting works in the district of Nagybanya in Eastern Hixngary, and has described the methods carried out in the principal establishments in the present memoir. The more important of these establishments are at Fernezely, 3J miles east of Nagybanya, and at Kapnik. At Nagybanya the ores occur in lodes in greenstone trachyte, and contain native gold and silver, sulj^hide of silver, ruby silver, copper pyrites, galena, blende, and pyrites. More compact vein stuff is subjected at the mines to the Hungarian process of amalgamation by stamping it fine and passing it through mills containing from 18 to 25 kilograms of mercury, when the free gold is collected partly by amalgamation and partly by subsidence in the mercury. After passing two series of mills, the stamped stuff is subjected to various processes of washing on frames, in order to collect the pyrites and galena in the form of slime. The amalgam from the mills, after pressing through leather, is distilled in a cast-iron retort, and the sponge metal, when melted, is sent to the gold refinery, at the Kremnitz Mint. The loss of mercury is from 50 to 100 per cent, of the w^eight of gold collected. For instance, in one case 155 tons gave 1"957 kilogram of gold with a loss of 1*9 kilogram of mercury, while in another 754^ tons gave 5*736 kilo- grams of gold with a loss of 2 * 9 kilograms of mercury.

The Fernezely Works. — The ores treated at these works, in addition to the slimes from the amalgamation mills, consist of : — (1.) Dry or quartzose ores, with native gold and silver, averaging 46 ozs. of silver per ton with 9i to 19 dwts. of gold per lb. As a rule, the gold contents diminish with increase in the silver, and the

Abstracts.] SMELTING OF GOLD AND SILVER ORES. 481

pyritic ores are riclier in the former metal than the qnartzose ores. (2.) Pyritic ores in lumps, containing native gold and rn1)y silver ore, with 15 to 40 ozs. of silver with 2-i dwts. per lb. of gold. (3.) Pyritic slimes, the richest gold material, containing 10 to 50 ozs. of silver with 48 dwts. per lb. of gold. (4.) Lead ores in lumps, with 20 to 60 per cent., and slimes with 50 to 60 per cent, of lead, the silver being about 26 ozs. with 24 dwts. per lb. of gold.

The reduction of the ores is now entirely effected by means of lead, the pyritic minerals being first concentrated by fusion into a coarse metal. The lead, when containing 120 ozs. and upwards of silver per ton, goes to the refinery at once, but when poorer it is subjected to a preliminary desilverization by zinc. The copper contents are ultimately concentrated in a regulus, which is con- verted into blister copper and refined at Felsobanya. It contains 20 ozs. of silver per ton which is not extracted. Formerly a portion of the ores was at different times treated by Kiss's ex- traction process, by barrel amalgamation and by Desiguolle's process, but these have now been entirely given up. The first of these (Kiss's process) was developed on the works, and for some time it was thought that the use of hyposulphite of calcium was advan- tageous in possessing greater solvent powers for gold in chloridized materials than the corresponding sodium salt, but later experi- ments have proved that the advantage is, if anything, rather with the latter. The Designolle process, substituting a solution of bichloride of mercury, and iron, for mercury in amalgamation, although recovering 90 per cent, of the silver value of the ore, only saves 60 per cent, of the gold, and causes very large losses of mercury, partly as undecomposed bichloride, partly as calomel, and partty as finely-divided metal which could not be collected. It has been abandoned since 1887. As regards the application of barrel amalgamation to these ores, the results of experiments at Schemnitz show that it is about on a level with smelting with ores containing 32 ozs. per ton, but for those with less, silver amalga- mation has the advantage, the most favourable results being obtained with 16 oz. ores.

The smelting processes now followed consist in a concentration of the auriferous pyrites into a coarse metal, which, when roasted, is mixed with calcined and fritted lead ores and smelted for argenti- ferous lead, and a regulus, which is sulijected to a further series of fusions, usually three, to desilverize it and concentrate the con- tained copper to a workable point. The chief points of interest in the processes are the use of the shelf furnaces of Ollivier and Perret, of Chessy, for the roasting of the auriferous pyrites, which are now taking the place of the open heaps formerly in general use. At the time of the Author's visit twenty of these furnaces were in course of construction at Fernezely.

The desilverizing of the furnace-lead is effected in 10-ton cast- iron pots with three zincings. The desilverized lead is freed from zinc by adding a mixture of sulphate of lead and common salt, and from antimony by poling.

[the INST. C.E. VOL. XCV.] 2 I

482 SMELTING OF GOLD AND SILVER ORES. [Foroifcn

The zinc-silver-lead alloy is liquated to remove part of the lead, and then treated hy soaking into a hath of clean lead on the refining-hearth, the zinciferous dross produced being added in the reguhis fusion. This method, however, not only wastes all the zinc, but causes loss of silver by volatilization, so that it is intended to treat it by distillation in fixture.

In 1886 6,935*6 tons of ores were smelted at Fernezely, pro- ducing—

Fine gold 230 kilograms.

Fine silver 3,375 „

Market litharge 132-9 tons.

Soft lead 550

Copper 4-5 „

The Kapnih Smelting WorJcs. — At these works both furnace and wet extraction processes are in use, about 1,800 tons of ores being treated annualty by the former, and 1,200 tons by the latter method, which is reserved for dry and pjTitic ores, while those containing lead in sufficient quantity are passed through the blast-furnace.

Wet Extraction. — This is a combination of the Augustin and Von

Patera processes. It is applied to dressed pjTitic and hand-picked

ores, the former containing about 70 j^er cent, of suljihides, including

blende 16 to 30, copper 1^, and lead 2 to 3 per cent., besides 10 to

11 ozs. of auriferous silver, with 8 to 10 thousandths of gold. The

picked ores are quartzose, and free from lead, with 22 per cent, of

sulphides, half of which is blende, their silver contents ranging

from 16 to 24 ozs. j^er ton, and the gold from 2 to 3 thousandths.

Both classes are mixed as nearly as possible in equal proportions

with 8 per cent, of salt, and subjected to a chloridizing roasting in a

four-storied shelf furnace, about 4 per cent, more salt being added

to the charge when it reaches the third shelf. From 36 to 38 cwt.

are roasted in twenty-four hours. "When the pyritic ores form half

the charge, the heat produced by the burning sulphur is sufficient

to effect the roasting without any additional fuel, but with a

lesser proportion, the furnace must be fired in the ordinary way.

The roasted ore is sifted ; the lumps, about 30 per cent, of the

whole, are ground and roasted with 3 per cent, more salt in a long

reverberatory furnace for six hours. The roasted material from

both kinds of furnaces is next lixiviated in wooden vats, with

perforated false bottoms covered with linen. From 50 to 60 cwt.

are treated at one time, first with 22 to 25 per cent, of brine heated

by steam to 28^ Centigrade, and afterwards for two days with a

cold Avatery solution of hj^posulphite of sodium of 3° to 5*^ Baume

density. By this combined method of extraction, 90 per cent, of

the silver and 80 per cent, of the gold in the ore are recovered,

namely, 60 per cent, of the former metal by the salt solution, and

the remaining 30, with nearly the whole of the gold, by the hypo-

sul})hite. The residues from the extraction are not subjected to

further treatment. The silver is recovered from the brine by

passing it over copper plates in two series of vats, the liquor being

Abstracts.] SMELTING OF GOLD AND SILVER ORES. 483

heated to 28° Centigrade, and the dissolved copper is in its turn precipitated by iron. The liquor after settling- is returned to the extraction process, and continues in use for three years. The gold and silver from the hyposulphite liquors are precipitated as sul- phides by sulphide of sodium, which is made by boiling flowers of sulphur with caustic soda in a cast-iron pan. The residual liquor after the separation of the sulphides is exposed to the air, when the excess sodium sulphide is converted into hyposuljihite, which goes back to the extraction works.

The cement silver, and the gold and silver sulphides, are treated by soaking in a bath of metallic lead heated to redness in a cast- iron pot. This gives a lead with 200 ozs. of silver, which is refined, and an argentiferous dross which is treated with lead.

The smelting processes at Kapnik are generally similar to those at Fernezely, the richer lead passing through the refinery at once, while the poorer is first desilverized with zinc. There is, however, no preliminary concentration of the jiyritic ores to coarse metal.

The production of Kapnik in 1886 was 1,520 kilograms of silver, with 74 kilograms of gold ; 185 tons of lead, and 12 tons of copper. The whole of the litharge produced is used to supply the lead required for collecting the silver in the blast-furnace.

The Author also describes in some detail the smelting operations followed at Zalathna, in Transylvania, which do not differ from those in previously published accounts.^ The ores treated are essentially iron pyrites containing gold and silver, and some tellurium minerals. They are classified into —

Ozs. per tfm of Auriferous Silver.

Poor pyrites 3 J to 9 J ozs.

Medium rich pyrites 10 to 16 „

Rich pyrites above 16 ,,

The telluride ores, which however only occur in very small quantities, are divided into —

First class 490 to 1,630

Second class 60 to 320

The pyrites are concentrated by fusion to a coarse metal, part of which is roasted in the shelf furnaces, while the remainder is treated with sulphuric acid to remove the bulk of the sulphide of iron. The roasted metal retains about 6 per cent, of sulj^hur, while that from the acid treatment, in addition to 8 • 85 per cent., retains a large quantity of sulphuric acid in the form of Ijasic sulphates, which give considerable trouble in the subsequent smelting with lead. This latter operation is performed in small blast-furnaces, litharge and refining-hearth bottoms being added in sufficient quantity to produce a lead containing 1 part in 2-iO of gold and silver. The tellurium ores are added to the furnace charge without

' Minutes of Proceedings Inst. C.E., vol. Ixxxii., p. 447.

2 I 2

484 SMELTING OF GOLD AND SILVER ORES. [Foreign

previous calcination. The average assay value of the lead is, silver 172 ozs., gold 120 ozs. per ton; it is refined to cake silver in quantities of 90 cwts. in a German refinery fired with wood. The operation lasts two hours. The resulting silver is 975 fine and contains 50 • 4 per cent, of gold.

About 28 per cent, of the ore furnace charge is obtained as regiilus, which is treated first with acid and then smelted with lead to desilverize it. These operations are repeated until the residual regulus contains 30 per cent, of copper, which is not subjected to further treatment on the spot.

In 1886 1,186 tons of pyrites, containing 241*52 kilograms of gold, and 307*94 kilograms of silver, were smelted at Zalathna.

Gold Amalgamation at Zalathna. — A very considerable quantity of gold is obtained by battery and mill amalgamation in the Zalathna district, nine thousand nine hundred and ninety-four heads of stamps, one hundred and eleven percussion tables, and one hundred and sixty Hungarian mills having been at work in 1886. The most successful of these appears to be a Californian twenty-stamp battery at Vulkoy, belonging to a French Company, which treats in twenty-four hours 35 tons of ore containing 18^ dwts. of gold per ton, of which 82 per cent, is recovered by com- bined battery and copper-plate amalgamation, the pyrites being saved in Frue vanners. The total gold production of Transylvania in 1886 was 1,222-08 kilograms.

H. B.

A Winding-Engine with Sjpiral Balance-Drum. By K, Habermann and J. von Hauer.

(Berg- und hiittenmannisches Jahrbuch, 1888, p. 120.)

At the No. 1 pit of the Camphausen Colliery at Saarbriicken, which is 500 metres deep, the winding-engines, of 1,000 HP., have two horizontal cylinders of 48 inches diameter, and 79 inches stroke, working direct upon a cylindrical drum of 26 feet diameter, and 11 feet breadth of face. The balance arrangement consists of a shaft somewhat smaller than that of the drum, which is coupled to the latter by a crank, so as to move at the same sjieed, but in an opposite direction. This shaft carries a double spiral drxam, with a rope-track for 28 revolutions. The rope, made fast to the smallest diameter of the drum at one side, passes first over a guide-pulley at the surface, then down a balance-pit 262 feet deep, round a loose pulley at the bottom, carrying the balance- weight, and returns to the surface passing over a second guide- pulley back to the drum, where the opposite end is fixed to the smallest diameter on the other rope path. When the engine is started with the loaded cage at the bottom, as the balance- rope unwinds from the larger and winds up in the smaller part of the drum, the counterpoise in the shaft falls, but with a gradually

Abstracts.] WINDING-ENGINE WITH SPIRAX BALANCE-DRUM. 485

diminisliing speed, until the cages meet in the shaft, when it is stationary, after which the weight rises, the rope being taken up on the larger, and unwound from the smaller coils. The balance- pit has a framing of cast-iron pillars, with diagonal struts, carrying two horizontal cross-girders for the guide-pulleys, which are 13 feet diameter, the loose pulley in the pit being of the same size. The axis of the j^it is 144 feet distant from the main winding-drum.

The largest diameter of the spiral track on the drum is 33 feet, and the smallest 10 feet; the breadth of the drum is 11 feet 5 inches. It weighs 28 tons, and is mounted on a shaft 16 inches in diameter, and 18 feet long, weighing 6 tons. The movable counterpoise in the shaft is an old boiler shell, loaded with scrap- iron, of a total weight of 15 tons.

The winding-cages have three decks, carrying two tubs on each, the total weight being : —

Tons.

Empty cage 3

Six tubs 2

Coal in tubs 3

Total 8

Eound cast-steel wire ropes are used in both pits ; those on the cages are tapered from 2*09 to 1 '93 millimetres diameter ; that on the balance-drum is 1 • Go inch throughout.

H. B.

On the Beer System of Wire Ropeways. By Charles Eaoult, Engineer to the Beer Engineering and Foundry Company.

(Revue Universelle des Mines, 3rd series, vol. iii. 1888, p. 49.)

This Paper describes improvements in the method of disposing of slag at the Seraing furnaces of the Esperance-Longdoz Company.

The slag, as it floAvs from the furnace, is conveyed by a channel, some few yards long, to a cast-iron gutter with a semi-circular bottom of 8 inches radius, supplied with a stream of water under slight pressure, with a volume of from 200 to 260 gallons per minute, by which it is granulated and carried into a basin, from which it is elevated by a chain of buckets, driven by an in- dependent engine of 6 or 7 HP., into a circular wrought-iron receiver, 16 feet 4 inches high by 14 feet diameter, formed with a conical base so as to discharge from different shoots round its circumference, and containing about 98 cubic yards.

The ropeway will transport 130 tons per day of ten hours to a distance of 300 yards. The starting-point is 11 feet 6 inches above the ground-level, and the point of delivery 160 feet above the starting-point.

The carrying-rope for the full road is Ij inch diameter, composed of nineteen wires, each ^j inch diameter, arranged, one in the

486 BEER SYSTEM OF WTRE EOPEWAYS. [Foreign

centre, six intermediate, and twelve exterior. Its weight is 2 If lbs. per fathom, and its theoretical hreaking-strain 37 tons, the actual breaking-strain being sensibly less. It is strained by a counterjjoise of 5 tons 18 cwt.

The carrying-rope for the empty road is l^V inch diameter, composed of nineteen wires, each -j^ inch full diameter, similarly arranged. Its weight is I2.j lbs. per fathom, and its theoretical breaking-strain 23 tons. The counterpoise is 3 tons 18 cwt.

The hauling-rope, |^ inch diameter, is composed of a hemp core surrounded by six strands, each of twelve wires, ^\ inch diameter, and weighs 4^ lbs. per fathom, with a theoretical breaking-strain of 14 tons 18 cwt. The counterpoise is 1 ton 19 cwt.

The joints of the carrying-roi:)es, which are made in convenient lengths, are usually formed by inserting each end into a slightly conical sleeve, slightly separating the wires, and brazing them to the sleeve with a special solder. The larger, or adjacent, ends of each pair of sleeves are tapped with a right- and left-hand thread respectively, and coupled with a corresponding right-and-left screT\ ed plug.

In the Beer system, however, instead of soldering, the wires, after being separated, are wedged into the sleeve, first by three curved wedges forming conjointly a feather-edged tube or ferrule between the outer and intermediate layers of wires, and next by a smaller solid conical ferrule between the intermediate layer and the central wire, which last is screwed at the end and secured by a nut.

A series of tests of this coupling, made on a length of 3 feet 3 inches, gave, for the larger rope of 1 j inch diameter, the following results • —

Load.		Elongation.
4'9 tons.		m\.
9-8 „		0-03 inch.
14-8 „		0-04 „
19-7 „		o-io „
24-6 „		0-62 „
29-5 „		1-21 „
30-1 „	Eup	ure of all the wires

The elongation was partly due to the wedges taking up their bearings inside the sleeves.

Kone of the wires were drawn out of the sleeves, but were all broken externally, and the joints were uninjured.

The hauling-rope is endless, the two ends being sjiliced together. In cases where the gradients are slight, the carrjdng-skeps may be attached to the hauling-roj^e at any point by a simple friction-clip, easily engaged and disengaged ; biit where the gTadients are more severe, as in the present case, thimbles must be fixed on the hauling-rope, to engage with the clips on the skeps. These thimbles have hitherto been made solid, necessitating the cutting and splicing of the rope at each point where one had to be fixed; but in the Beer system they are made in halves, dovetailed together, so as to be slij^jied on anywhere, and are secured by a

Abstracts.] BEER SYSTEM OF WIRE ROPEWAYS. 487

small rivet with countersunk heads, thus avoiding the injurious effect of solder on the rope. They are Ij inch external diameter and 1^ inch long, and are fixed 38 fathoms apart. When loaded with a weight of 2 tons, and tested by repeated blows with a hammer, one of these thiml)les has failed to show any sensible displacement. It is found desirable to change the position of the thimbles from time to time, so as to equalize the wear on the rope.

The hauling-rope is driven by a 9-HP. vertical engine, placed under the platform at the loading or starting station. A pinion 8 inches diameter on the crank-shaft, which makes 120 revolutions per minute, gears into a spur-wheel 7 feet G inches diameter, keyed on the same shaft as the driving-drum, which has two grooves lagged with wood. The rope passes twice round this pulley, and once round a single-grooved idle pulley placed above it in the same vertical plane, and is led away horizontally over two guide- pulleys. The return-pulley at the discharging station is movable, and weighted with a counterbalance of 1 ton 19 cwt. to keep the rope taut. At each station thei'e is a fixed rail, on to which the skeps are shunted, so as to be passed in the one case round the return-jiulley, and in the other round the receiving hopper, so that they may be filled from any of the shoots before described. Movable switches at the starting station allow the skeps to be removed for repairs, &c. The travelling-sjieed is about 2j miles per hoiir, and it is noticed that the hauling-rope constantly revolves on its own axis, and always in the same direction.

The discharging station consists of a platform 66 feet high, carried on a light but very substantial framing, steadied by guy- rojies. It stands on an old spoil-heap.

There are three intermediate supjiorts, consisting of wrought- iron lattice posts of elegant design, bolted to masonry foundations. The tallest is 72 feet high. Each is provided with two cross-bars for supporting the carrying- and hauling-ropes, which are one above the other in the same vertical plane. The hauling-rope is simply carried on grooved pulleys, but the support of the carrying- ropes is a more complicated problem, as it is found that, owing to variations of temperature and in the positions of the loaded skeps, they have an endlong movement to and fro of 10 inches or more. If the motion of the two ropes is in the same direction, it tends to overturn the supporting posts ; if in opposite directions, to twist them. If the ropes are arranged simply to slide on their supports, they soon get set fast, no matter how well greased ; if carried on a simple pulley, they soon show signs of wear, from want of a more extended bearing ; if on a block carried on small wheels, it soon works itself to one end or other of its track, and there sticks. In the Beer system the blocks are carried on pro])erly-formed blocks, slung from pendulum-rods, allowing a perfectly free motion end- ways, but restrained from side oscillations by qtiadrant guides. A quarter turn-over is given to the carrying-roj)es from time to time, so that all sides may be eqiially Avorn.

Five persons only are employed ; one to attend to the engines

488 BEER SYSTEM OF A^TEE ROPEWAYS. [Foreign

and macliiner\% a filler and a hooker-on at the starting point, and a boy to tip the skeps and a hooker-on at the delivery point. The clii:)S which engage with the thimbles are antoniatically released by coming in contact with a fixed trijjper-bar at each end of their travel.

It is claimed that a saving of 66 per cent, is eifected, as compared with the system previously employed.

The Paper is fully illustrated with engravings of the general arrangement and details described.

W. S. H.

Desrozier's New Dish-Bynamo. By E. Meylan.

(La Lumifere Electrique, toI. xsix. 1888, p. 401.)

This Paper, which is well illustrated by diagrams, describes what is probably the greatest novelty that has been produced in the design of dynamo-machines for several years. The magnetic field is multipolar, the poles being arranged similarly to those of the Siemens alternate-current machines, with their opposing faces brought nearer together, as the space between is that required for two layers of the wire forming the armature ; this latter is formed of a series of radial wires, with junctions alternating between the external and internal ends ; these jimctions are not, however, between neighbouring radii, but between those that are almost sjnnmetrically placed with respect to the magnet poles ; that they are not qidte spnmetrical is due to the fact that after one progression roimd the circumference the next series of radial wires is reached, and thus the whole area is filled up, and the arma- ture wires connected in one series. By an ingenious arrangement of the connections on the anterior and posterior faces of two annular disks of insiilating material, they are made without any difiiculty, and at the same time without crossing of the wires. Adopting for simplicity's sake a six-pole machine, the disk arma- ture may be likened to the geometric projection on two parallel planes of the contour lines of the teeth of a circular cutter, this cutter having a number of parallel teeth helically arranged, three teeth of deep gnillet forming one circumference ; the front and back edges of the teeth would represent the radial conductors, and the junction-pieces would be formed by the gullet and external face of the teeth respectively ; the two parallel planes being neces- sitated by the fact that in half the length of the imaginary helix the front and back edges of the teeth woiild be in one line. In such a machine each elementaiy coil, corresponding to one com- plete tooth in the imaginary cutter, is connected to three plates on the commutator, each 120^ apart, these connections being also arranged with great mechanical simplicity ; there are thus only two brushes, and all sparking at the commutator is eliminated. The advantages to be gained by this form of machine result from

Abstracts.] DESROZIER's NEW DISK-DYNAMO. 489

the ease with which the three elementary requirements, that make for power, can he secured, viz., intensity of magnetic field, density of current, and linear velocity ; while the absence of iron in the armature, and its open construction, aflbrd effective means for pro- curing a very good commercial efficiency ; the only drawback — which obtains in all multipolar machines — being the division of the magnets, with the consequent extra expenditure in excitation. Four of these dynamos have been suj^plied to the " Formidable " (French ship), each driven direct by a Bregxiet steam-engine, and furnishing 12,250 watts, with a commercial efficiency of 79 per cent.

F.J.

Gadot Accumulators, pattern 1888. By J. Laffargue.

(L'Electricien, 1888, p. 562.)

The new accumulator of the Gadot type tested recently by the Author consisted of ten negative and nine positive plates of the following weights and dimensions : —

Lb.

Negative plate 1 • 89

Positive „ 1-88

Plate, without oxide 0-97

Active material iu the negative plate 0 • 924

„ „ „ positive „ 0'915

Inches.

Total thickness 2-28

Height 59-00

Breadth 59-00

Mean distance between the plates in the liquid 2-36

The total weight of the nineteen plates is 35-87 lbs. The accumulator gave a useful discharge of fifteen hours, when the potential, which was 2 • 0-i volts at the beginning, had fallen to 1 • 75 volt. The initial current strength was 15 • 9 amperes, and 13-8 at the end of the fifteen hours ; the maximum internal resistance was 0 • 003 ohm, and the total useful capacity in ampere-hours was 225*7 or G-29 per lb. of plates. A diagram is given in the Paper showing the results of the tests in the form of curves.

J. J. W.

Account of a Series of Experiments made on Hessner's Cell. By W. Chukoloff.

(Electrichestavo, St. Petersburg, 1888, p. 63.)

The Author first tested the adaptability of this cell to inter- mittent electric lighting, but it proved to be quite unsuited for this purpose. For telegraphic or telephonic purposes, however, the cell was found to be most efficient, especially in cold climates.

490 SERIES OF EXPERIMENTS MADE ON HESSNEr's CELL. [Foreign

The battery was connected with an electric bell of 12 ohms resis- tance, and periodically rung from January 26 to February 28, with an average weakening of 3 • 8 per cent, of the electromotive force of the cell during the period of the experiment. The longer the circuit be closed the greater is the fall of the electromotive force, and the fall is most sudden at the beginning of the action ; thus, during the first hour the fall in volt-power was 0 • 04 volt, while the average fall per hour in twelve hoiirs was 0-01 volt. Also the longer the cell be kept at rest, the greater is the rise in its electro- motive force, and the rise is more sudden during the first hour of rest than the succeeding hours, so the average rise for one hour was 0-01 volt, and that for seventeen succeeding hours was 0-02 volt. The rise is slower the nearer it approaches the normal electromotive force of the cell.

To test the behaviour of the cell under conditions of heat and cold, the following experiments were made. The cell was cooled from -1- 20^ Centigrade when its electromotive force was 1-32 volt to — lo"^ Centigrade, when the electromotive force was found to be 1*28 volt. After being exposed for five hours at -(-oO^ CentigTade when the electromotive force was 1 -34 it was cooled to -f 20° Centi- gTade, when the electromotive force was found to be 1 • 32 volt, and its internal resistance 0"o5 ohm, and the cell was then ex- posed to a frost of — 17"^ Centigrade for twenty-four hours, and — 15° Centigrade, when its electromotive force was 1*28 volt and its internal resistance 0-65 ohm. After being exposed five days to a Irost of —17° Centigrade, which went as low as —26° Centi- grade, the electromotive force was 1*27 and the internal resis- tance 0-7 ohm. Then after being three days in a room at -f-20° Centigrade, the electromotive force had risen to 1*3 volt, the internal resistance reduced to 0-55 volt. Thus the electromotive force during these experiments had only decreased by 4 per cent., and the internal resistance risen 21 per cent., or the general weakening of the cell m.ay be taken at 30 per cent.

The results of these experiments, together with the facts that the cell contains no liquid, and that the chemicals do not require lenewal for over a period of two years, lead the Author to conclude that this cell is one most fitted for telegraphic and similar purposes.

G. K.

0)1 the Measurement of the Resistance of Suhmarine Cahles.

By A. EOUILLARD,

(L'Electricien, 1888, p. 707.)

In testing the conducting resistance of a subaqueous cable by means of the " repioduced deflection " method, the line (with its distant end earthed) is connected through a shunted galvanometer and reversing key to earth, the battery being joined to the key in the usual way. As, however, in this arrangement, the battery

Abstracts.] RESISTANCE OF SUBMARINE CABLES. 491

is not in circuit when readings of the earth currents c are observed, its resistance is not taken into account ; hence an error is intro- duced, which, in the case of strong currents, may be a considerable one. The strength of the observed earth current is

e

'° " E'

e being the electromotive force and R the conducting resistance of the cable. When the deflection d is noted, the battery (of resist- ance r and electromotive force Ej is in circuit and the current strength is

E + g

E

The value actually required is I' = — , since the current of

strength I' is reproduced through a resistance R plus r, or that of the battery. Therefore, in eliminating i^, the expression becomes

E-f e

which differs from I' by a quantity c representing the error in question.

If € = I' - (I - i) =

e then, as — = ^_,,

r e

R (R 4- r)'

R-hr /R^^

and as the deflections are proportional to the strength of the currents producing them, the formula

d^— d — c -{•

r

gives the deflection d^, which has to be rejiroduced by means of the battery through a resistance R, the value of which will then be equal to that of the cable under test.

The value of this correction e is important, especially when a fault has to be localized. For instance, assuming the earth current to vary from — 20 to -f- 20 divisions, then with a battery of 40 Callaud cells (each of 9a) resistance) r = 40 x 9 = 360 ohms, and a cable having a resistance of 4,000 ohms, the correction expressed in divisions of the scale is

492 BESISTAIsCE OF SUBMAEINE CABLES. [Foreign

and if, with the constants chosen, one division represents 20 ohms, the error has a value of 33 ohms, equivalent to about 10 miles of cable.

By aid, however, of a resistance r (equal to that of the battery) inserted between the testing-key and earth, the necessity of applying this correction is avoided. This supplementary resistance can be short-circuited by means of a key, so that the readings of the earth currents can be taken, with r in circuit, by pressing down this key, and when the deflection d is observed r is cut out by leaving the key in its normal position. For the rapid readings essential in this method of testing, the Author has found a Desprez d'Arsonval galvanometer very suitable. The tests re- ferred to in the Paper were made with an instrument of this class having a resistance of 200 ohms, whilst that of the shunt was 1 to 2 ohms, according to the resistance under measurement.

J. J. W.

Pliilippart' s Electrical Tramcars in Paris.

(L'Electricien, 1888, p. 617.)

The motive-power is supplied by a battery of one hundred and forty-four Faure-Sellon-Volckmar twin-plate accumulators, of a storage capacity of 150 ampere-hours, at a normal rate of discharge of 25 amperes. Taking 1 • 8 volt as the mean useful electromotive force per accumulator, the total cajiacity is 37,000 ^ watt-hours, equal to 50 ^ HP. hours, or say an effective energy of 40 HP.-hours, sufficient to drive a car for six hours without re-charging the battery.

The motor consists of a Siemens dynamo, coupled to a counter- shaft by means of a Eaffard endless rope. On the shaft, which is in two lengths joined together by a differential coupling, there are two pinions geared to the rear wheels of the car by chains of the Gall type. The driving-wheels work independently of each other, and on the straight parts of the track the speed of the two pinions, and consequently that of the wheels, is equal, but on curves the action of the differential coupling causes the two pinions to revolve at different velocities, so that curves of very small radius can be passed over withoixt difficulty. The cars are running between the Place de I'Etoile and Porte Maillot ; the weight of each vehicle, inclusive of accumulators, motor, transmission gearing, and fifty passengers, is about 9 tons.

J. J. W.

These figures should be by calculation 39,000 and 52 respectively.

Abstracts.] THE TELEPHONE-EQUATION. 493

On the Telei^hone-Equation. By C. L. Madsen.

(Elektrotechnische Zeitschrift, 1888, p. 462.)

The experiments forming the subject of this Paper were nntler- taken Toy the Author with a view to the further development of Preece's equation for the limiting distance of telephonic speech, viz. :

K

K being the total resistance of the line in ohms, C its electrostatic capacity in microfarads, and K a coefficient. The value of K, as determined by various experimenters, ranges from 2,000 to 15,000; it is therefore probable that some uncertainty exists respecting the actual meaning of the above expression. In general, trials appear to have been made simply to find out through what length of line of a certain type conversation could be carried on, but scarcely any data have been put together relative to the electrical values of such line, number of offices, construction of the apparatus, &c. Then, as regards the transmission of speech, it is important to know whether the adopted standard expressed a condition which enabled every word to be distinctly heard without repetition, or merely one that just admitted of an articulation being sent through the line. In order to remove all ambiguity from the factor ?/, the Aiithor and his coadjutors adopted the figure 100 to express a conversation of a certain uniformity, as regards the strength and clearness of the sound, whilst other interchanges of speech were valued according to their prevailing quality, whether better or worse than that of the standard. The experiments were made by means of the latest improved form of Bell-Blake instruments, and in the Exchange Offices the electro-magnets were connected in shunt, so as to avoid, as far as possible, the weakening effect of self-induction. Electrical measurements were obtained of the ajiparatus in the central stations and subscribers' offices ; some of the main lines were also tested, and from the various data thus acquired the exact value of K for the telephone lines in Denmark was determined. The equation now has the following definite form : —

T ""

K . C

or for a telephone circuit, having a " figure of merit " equal to the standard

T = 100 = :- J^=jr, ,

K • C l^ • re

where Ir is STibstituted for E and Ic for C ; r and c being respec- tively the resistance and capacity per given unit of length I.

The Table in the Paper embodies the results obtained on

494 THE TELEPHONE-EQUATION. [Foreign

nineteen telephone lines, which, were connected with Copenhagen. Sixteen of these circuits are in work, and the other three, Aarhiis, Odense, and Malmo, were employed for the elucidation of some special technical details. As an example of the enormons gain that has been obtained from the nse of hard drawn copper wire, or bronze wire 2 millimetres in diameter, in place of iron or steel wire, the Author mentions the Helsingor-Copenhagen line, on which telephonic conversation is now uniformly loud and clear, whereas with the latter kind of wire the transmission was very imperfect. Hard-drawn copper wire is practically free from self- induction, which weakens speech almost to the same extent as doubling the resistance of the line. The experimental results give 300,000 as the absolute vahie of K for the telephone system in Seeland. The equation is then

_ 300,000

and if T is foimd to be equal to 100, when the actual values of R and C for any particular line are inserted (R being doubled in the case of iron or steel wire), then the line is a standard one as regards its power of transmitting speech.

The tabiilated results are, in some instances, of general interest, as proving that the excellence of a telephone line depends far more on the kind of wire used and the number of Exchange offices in circuit than on the actual distance separating the two end stations. For instance, the line Copenhagen-Ringsted, 67*9 kilometres in length, consists of 1 kilometre of cable, 61*5 kilo- metres of copper or bronze wire, and 5*4 kilometres of iron or steel wire. When worked with only two intermediate Exchanges in circuit, this line has " a figure of merit " of 302, whereas the Copenhagen-Kjoge circuit, which is 20 kilometres shorter, but is composed of 46-9 kilometres of iron or steel wire and 1 kilometre of cable, is only 199. Again, in the case of a second circuit between Copenhagen-Ringsted, the lengths in kilometres being 2*5 of cable, and 72*1 of iron wire, with four intermediate offices in circuit, the "figure of merit" fell to 76.

J. J. W.

The Telephone Line hetiveen Paris and Marseilles.

(L'Electricien, 1888, p. 647.)

The line, which is almost an aerial one throughout, consists of two bronze wires, each 4^ millimetres in diameter, the cross- section being 15-9 square millimetres, the conducting resistance 1 ohm per kilometre, conductivity 97 per cent, of that of pure copper, weight 146 kilograms per kilometre, and breaking strain 45 kilograms per square millimetre. The distance between the ter- minal stations is 900 kilometres, consequently 1,800 kilometres of

Abstracts.] TELEPHONE LINE BETWEEN PARIS AND MARSEILLES. 495

wire are used for the complete circuit. The suliterranean cables are of the Fortin-Hermann type ; each conductor consists of a strand of seven copper wires each 0*7 millimetre in diameter. Beads of paraffined wood are strung at intervals on the strand, and six such conductors, each having a resistance of 7 ohms per kilometre, are cabled together helically and encased in a leaden tube. Two of these conductors are used for the Paris-Marseilles telephone line. As the dielectric is, for the most part, air, the insulation is as high as 4,000 to 5,000 megohms per kilometre, and the electrostatic capacity 0*053 microfarad. The only parts of the line where cables are used are from the Exchange, in Paris, to the railway station at Yincennes, viz., 3 kilometres ; two lengths of 100 metres each in tunnels, one near Paris, and the other not far from Lyons ; also one length of 800 metres in the Saint-Louis tunnel near Marseilles.

In order to minimize inductive disturbances, the wires are attached to the posts in such wise as causes a reversal of their relative positions at each kilometre. The ordinary telegraph wires are also fixed to the posts, the principal circuits being worked with Van Eyssell^erghe anti-induction apparatus. At each ter- minal station there is a set of microphone transmitters with induction bobbin and the usual telephone receivers, whereas at Lyons there are two such sets. Ordinarily, Paris and Marseilles work on separate circuits to Lyons, so that each terminal town can be in communication at the same moment with the intermediate station ; but whenever direct terminal working is required, the operator at Lyons effects the necessary connection of the main wires. The battery at each end station consists of six Lalande and Chaperon cells, three in series and two in parallel, and at Lyons each set of apparati;s is worked by three similar cells in series. The electromotive force of each cell is 0 ■ 8 volt, and its internal resistance 0 • 1 ohm. The transmitters and receivers are of the D'Arsonval pattern.

J. J. W.

The Begulation of Are Lamps. By E. Hospitalier.

(L'Electricien, 1888, p. 276.)

Arc lamps in parallel are more easily regulated when the difference of potential is in excess of that actually required. For instance, with 60 volts as the standard, a nearly constant arc can be maintained with 60 volts, and a still steadier one with 70 volts. If this method be extended, it is found that with 100 volts a single arc can be kept perfectly constant when the surplus electromotive force is absorbed by a resistance of about 5 ohms, the normal current strength being 5 amperes. Assuming that an arc has to be maintained with a current strength I of 10 amperes, the difference of potential e at the terminals being 50 volts, and the useful power

496' THE REGULATION OP ARC LAMIPS. [Foreign

Pu (= el) in the arc, 500 watts. A knoAvledge of the variations of these three elements, when the apparent resistance of the normal arc of 5 ohms fluctuates between 4 and 6 ohms, is easily obtained by- means of the graphical method, which also furnishes convenient data for determining the kind of regulation that should be employed for each jiarticular case.

The cur\'es shown in the Paper are numbered respectively 1, 2, 3, and have reference to three different conditions, viz. :

1. An electric generator having a constant electromotive force of 100 volts; the total resistance of the circuit, exclusive of that of the lamp, being 5 ohms.

2. Generator of 70 volts ; external resistance, 20 ohms, irrespec- tive of that of the lamp.

3. Generator of 50 volts ; resistance inappreciable Tas in the ease of a battery of accumulators) ; resistance inserted in the circuit theoretically null.

A comparison of the cur\-es demonstrates that all the conditions of good regulation are at hand, when working with an electromotive force of 100 volts, and are absent when fifty volts only are avail- able. It can also be seen that a regulator which maintains el constant is not ai:)plicable to the first case, but gives good results when applied to the third. The regulating power of the electro- magnet ordinarily applied is a function of I, whereas the action of an electro-d;^Tiamometer is proportional to I'-. The use of the latter accessory would therefore increase the sensibility of the regulating system, and although the lamp could not be run satisfactorily with 50 volts, it would work well with 60, instead of 70, which is the voltage almost universally adopted at present for mixed distribu- tions having arc and glow lamj:)s simultaneously in action.

J. J. W.

TJie Electric Lighting of the City of Geneva. By EoGER Chavaxnts.

(La Lumifere Electrique, vol. sxis., 1888, p. 451.)

In the Central Station there are three Piccard high-pressure turbines, each of 200 HP., with horizontal axes, carrjdng a central cro^^^^ of vanes ; the water enters from the outer circumference by means of a distributor, and the flow is so regulated by a centrifugal governor that the variations of speed are within 1 per cent. The governor consists of a self-acting hydraulic motor ; it operates a cylindrical slide-valve which admits the water at high- pressure into a vertical piston that adjusts the orifice of the dis- tributor. The motion of the valve is efl'ected by means of a lever, having at its fulcrum a rod jointed to the vertical piston. The cover of the governor forms the movable point of this lever. By the action of the regulator the valve is opened, bixt the motion of the piston tends to close it ; these opposite effects are so balanced

AbstractB.] ELECTRIC LIGHTING OF THE CITY OF GENEVA. 497

as to give a regulating system of extreme sensibility. The water enters the turbines at a constant pressure of 13^ atmospheres ; constancy is maintained by means of a reservoir situated about 135 metres above the level of the lake. Each turbine drives two Thury hexagonal-shaped dynamos, each giving a current of GOO amperes at a jDotential of 110 volts, and speed of 350 revolutions a minute. The three-wire system of distribution is employed, and the mains consist of Siemens concentric lead-cased cables. At present, the circiiits are confined to a portion of the town on the left bank of the river. The main cables or feeders are joined at five diff"erent points to intermediate cables by means of junction-boxes, and the service-wires are led from distributing-boxes into the hoiises. Each pair of shunt- wound dynamos is connected in series, each machine being joined to the distributing conductors on the switch-board by means of three cables ; that is, two for the main, and one for the shunt-circuit. The lower part of the board is fitted with the rheostats, or current-regulators, which are in circuit with the field magnets. The rheostats are arranged in pairs, and can be mani- pulated either singly or altogether, by means of suitable gearing. In the main circuit of each group of machines two cut-outs are inserted ; one of them acts automatically, and causes a disconnection in case of any accidental reversal of the current. The switch-board is also provided with voltmeters and ammeters. In the houses which are lighted, the consumption is recorded by Aubert's meter when the lamps are few ; by Aron's single-coil meter in the case of a medium supply, and by a double-coil meter of this type when forty lamps or more are in action.

J. J. W.

Self-Begulating Electric Search-Light. By W. E. Eein.

(Centralblatt fiir Elektroteclinik, 1888, p. 5G4.)

In an electric lamp devised for military and naval purposes, the chief requirements are, that the arc should remain in the focus of the reflector, and the regulation be eff"ective at any angle of inclina- tion. When the motion of the carbons is efiected by means of the superior weight of the upper holder, the lamp only works well in a vertical position ; hence hand-adjustment has been introduced. The Author has constructed an automatic lamp which can be used in any position.

The carbons are carried by two square rods working horizontally between friction rollers. The frame containing the rollers for the upper, or positive carbon-holder, is fixed to an upright which is screwed to one end of a cast-iron base-plate ; the negative holder works in a guide, which with its rollers is suspended in four rails beneath the upper holder, and is entirely insulated from the latter. At the ojiposite end of the base-plate is a second upright, to which is fixed an electro-magnet E, having its armature in front of the

[the INST. C.E. VOL. XCV.l 2 l<

498 SELF-EEGULATING ELECTRIC SEAECH-LIGHT. [Foreign

movable framing formed by the rails. The central roller of each system is kept in contact with its carbon-holder by means of two spiral springs ; each of these two rollers is fitted with a ratchet- wheel, the pawls being joined together and to the armature of an electromagnet e by a steel bar. As soon as the current actuates the armature of e, the pawls are drawn downwards ; the armature is then automatically released, springs back to its former position, and causes the friction-rollers to revolve simultaneously in opposite directions, whereb}^ the carbon-holders approach each other.

The electromagnet E is in the main circuit, and directly the current traverses its coils, the armature is attracted, causing a separation of the carbons and formation of the arc ; when the distance between the carbons becomes too great, the current in the shunt circuit of the electromagnet e increases until the armature is actuated, and the carbons drawn together by the motion of the friction-rollers, as described above.

The horizontal position of the carbon-holders enables one of them to be placed in the axis of the reflector, so that only a single slot is required for the reception of the other holder. All the rays from the crater of the positive carbon are thrown radially on the inner surface of the paraboloidal reflector, and the projection outwards is more complete than in the case of lamps with vertical holders.

The case containing the mechanism is so mounted as to admit of motion in any desired direction.

J. J. W.

Electric- Light Installation on the Arniour-elad Cruiser " Admiral Nahimof\" By Lieutenant Kolokoltzoff.

(Morskoi Sbornik, St. Petersburg, January 1888, p. 41.)

This is the first Russian armour-clad lighted throughout by electricity. The installation was set tip by the Jablochkoif Com- pany, and consists of four comjDound-wound Gramme dynamos, designed for an output each of 140 amperes at 65 volts, driven by four separate engines, and feeding three hundred and twenty glow- lamps (8-candle power and 50 volts electromotive force), and two Mangin search-lights, placed at the ends of the fore-bridges. The two dynamos for feeding the search-lights are placed amidships on the gun-deck, and the other two, for lighting the decks, are placed amidships on the main-deck. The engines working each pair of dynamos are connected by separate steam-pipes to both the main and auxiliary boilers. Each dynamo can be switched on to any circuit at will, and each pair of dynamos can be connected in parallel arc. The switching of any dynamo on to the deck- or search-lights is done by means of two three-way commutators attached to each dynamo.

The construction of the cruiser is as follows : Above the pro- tected deck, along the whole length of the vessel, extend the gun- j

Abstracts.] ELECTRIC LIGHT ON THE " ADMIRAL NAKIMOFF." 499

and main-decks, and under the protected deck, in the central part of the vessel, are placed the engines, and in the liows and stern the ammunition magazines and provision-stores. The conducting mains form three closed double circuits, all connected to the main switchboard. Two of these double circuits supply the gun- and main-decks, one pair extending along the port side of the gun-deck, descending at the bows to the main-deck, thence led aft along the whole length of the port side of the main-deck, and ascended at the stern to the gun-deck, where the respective conductors join, and thus form each a complete circuit. The second pair of conductors travels along a parallel path, but on the starboard side of the vessel. The third pair of conductors supply the engine-rooms, stokeholes, and magazines. Owing to the uneven distribution of the lamps along the decks, the conductors are not of uniform section throughout, one half being 20 square millimetres (0"31 square inch), and the other half 40 square millimetres (0*62 square inch.) The cor- responding main conductors of the port and starboard circuits on the gun- and main-decks can be connected together at the fore and after part of the ship by means of switchboards, to which they are connected by supplementary condiictors. The lights on the fore and after orlop decks are regulated by auxiliary switchboards, so that they can be fed direct from the dynamos, or, in cases of need, from one hundred and fifty accumulators on each deck.

The advantages claimed by the Author for this system of lighting on war-ships are, that the laying of the conductors on the gun- and main-decks in an annular j)ath prevents the extinction of any lamp, in the case of a single injury to one of the conductors by an enemy's shell ; as main conductors will be still connected at the ends to the dynamo, and the circuit will not be broken ; also by joining the circuits at the fore and aft switchboards, even the total severa- tion of one pair of the conductors leading to the main switchboard would not cause the extinction of any lamp; and thirdly, that the orlop decks, where the chief engines of war are situated, being supplied either by accumulators or the dynamos, their illumination can be cut off only by the destruction of the decks themselves.

The main switchboard consists of a marble slab, on which are placed all the commutators necessary for regulating the lighting of the whole vessel. The commutators are of three kinds — plug-com- mutators, ordinary friction-commutators, and a special form of double pole friction-commutator. The switchboard is so arranged that the positive and negative poles of the dynamos are connected respectively with commutators situated at the two sides of the slab, and for each dynamo in a separate horizontal line. The main conductors are similarly connected to the board, biit in vertical lines. The dynamo commutators are placed in serial order ; thtas the upper commutator is connected with the first dynamo ; the next lower commutator with the second dynamo, and so on. So also with the circuits : the left-hand commutator is connected with the port circuit ; the next commutator with the starboard circuit ; the third commutator with the engine-room circuit, &c. The

2 K 2

500 ELECTRIC LIGHT ON THE " ADMIEAL NAKIMOFF." [Foreign

double-pole friction commtitators allow the current to be measured at any moment without interrupting the circuit.

The Paper is furnished with drawings of the various com- mutators, lamps, and other appliances, and gives detailed descrip- tions of the management of the main and auxiliaiy switchboards.

G. K.

On Siemens and Hahhes Electric Winding-Engine at Neu Stassfurt.

(Oesterreichische Zeitschrift fiir Berg und Hiittenwesen, 1888, p. 105.)

In 1885 an inclined shaft was commenced at the bottom of the Keu Stassfurt mine, in order to explore the continuation of the potash salt bed in depth. The winding in this shaft, which had a slope of 40 degrees, is effected by a geared engine driven by electricity from a primary djmamo and steam engine at the sur- face, which has already been in use for some time as the motive power of an electric railway, about 1,500 metres long, on the 300-metre level of the same mine. The engine at the surface being 155 metres from the shaft bank, the depth to the bottom of the mine 360 metres, and the distance from the shaft bottom to the driving dynamo 40 metres, a total length of 555 metres of double conductor was required to convey the cuirent to and from the engine. The conductors at the surface are of bright copper wire, while in the pit and underground a covered cable laid in a wooden trough is used. The work to be performed consisted of the lifting of one loaded tub weighing 1,200 kilograms gross, and the lowering of an empty one of 400 kilograms through a height of 100 metres, on an inclined plane 155 metres long, once in four minutes, or, having regard to the time lost in hooking on and landing, in an actual working time of three minutes, or an average working speed of 3 feet per second. This requires an effective power of 431 kilogrammetres per second for lifting, and 21*2 kilogTammetres for overcoming the frictional resistance; or, in round numbers, 450 kilogrammetres i^er second as the useful work of the secondary dynamo ; or, taking 40 per cent, as the effect realized, about 15 HP. is required at the steam-engine. The useful electrical effect realized is at least 53 per cent., about 25 per cent, being lost by the driving gear of the rope-drum.

The primary dynamo is of the Makers' Dg type, giving a current of 22 amperes and 370 volts, the loss in the conductors being only 5 or 6 per cent. The secondary dynamo, or electromotor, is of the D, type, and converts 75 per cent, of the electric energy into power. It makes one thousand revolutions per minute, and drives an intermediate shaft by means of a belt 160 millimetres (6j inches) broad at about one-third of that speed. This motion is again reduced by means of spur gearing to 13*3 revolutions on the rope- drum, which is 1,240 niillimetres (4 feet 1 inch) in circumfer-

II

M

Abstracts.] SIEMENS AND HALSKE's ELECTRIC WINDING-ENGINE. 501

ence, and moves at the rate of about 3 feet per second. The axis of the drum is horizontal, and the direction of the windng-ropes is reversed by guide joulleys 1 metre in diameter, at a distance of 4 metres from the drum. The whole of the machinery is contained in a gallery 3*G metres high and 18-5 metres long. The transmis- sion of the power by a intermediate shaft driven by a belt, against which some objections were raised, has proved to be advantageous in practice, as the load is started and stopped without shock, and great regularity of motion is obtained.

In order to prevent injury from heating by sudden changes in the current when starting or stopping, a series of resistance elements are attached to the motor, which can be put in or out of the circuit simultaneously with the action of the reversing belt, in order to regulate the current to the motor. The reversal of the current is effected by lifting or lowering one or other of a pair of brushes, placed diametrically opposite to each other in regard to the commutator.

The engine was in continual use, without accident of any kind, from the beginning of November, 1885, to the end of March, 1887, during which period the sinking was carried to a depth of 132 metres, and several trial drifts were made at the bottom. The purpose for which it was undertaken having been accomplished the work was then stopped. The makers observe, in conclusion, that with their im})roved forms of dynamo since introduced a nmch higher effect may be realized from a similar plant.

H. B.

0)1 the Connecting of Lir/htning-Concluetors ivitli Water- and Gas-Pipes. By L. Weber.

(Elektrotechnische Zeitschrift, 1888, p. 285.)

In this Paper the Author has embodied many of the considera- tions which led the sub-committee (appointed by the Berlin Electrotechnical Society to investigate the question of lightning- conductors) to pass the following resolutions : —

" The joining of lightning-conductors to the gas- and water- pipes does not cause injury to the latter. These pipes are as much endangered when no such junction exists as they would be by the absence of a conductor. It is therefore imperative that lightning- rods shoiild be in permanent metallic connection with the system of pipes within a building. This connection should be made at a suitable part of the pipes, and at an external point as regards the principal meter."

The following is an outline of the data on which these reso- lutions were based : —

1. Danger to buildings due to the promotion of electrical dis- charge by water- and gas-pipes.

As an underground system of pi})cs is usually in intimate

502 CONNECTING LIGHTNING-CONDUCTOES [Foreign

connection with the conducting mass of the earth, it forms a path of least resistance, and in buildings where the upper branches of the system are carried to the topmost floors, there is no doubt that the discharge finds its way to these branches through the roof or •walls.

2. Danger to the gas- and water-pipes.

The nature of the damage to the pipes depends upon their position and the way in which they are struck.

(a) If the lightning strikes any part of a pipe which is out of doors and exposed to the air, a slight fusion may be caused, result- ing (in the case of thin gas-tubesj in an explosion of gas. When the point which is struck lies either in water, earth, or a wall, considerable mechanical destruction often takes place.

(6) If the discharge passes to pipes caulked with a badly con- ducting material, they are in danger of being cracked, but the gas is not likely to be exploded when the caulked joints are in the groxmd, as even explosive mixtures cannot be fired by a spark unless they are collected in some large hollow space.

3. The manner in which the dangers cited under §§1 and 2 are affected by a lightning-rod having no connection with the water- and gas-pipes.

Lightning-conductors are joined to ground-plates which, as regards surface and close contact with the earth, are quite in- effective in comparison with the large network of underground pipes. Consequently, whenever any distant in-door branch of this network is situated near a lightning-rod, the discharge has a tendency to spring from the rod to the pipe, and thence pass to the mains below. If the two metallic discharging paths are separated by several metres of air-space, in which conducting substances would not be likely to stand, even temporarily, the springing across of the lightning might not occur ; but in dwelling- houses such a condition is never fulfilled, because each bell-wire, gilt cornice-pole, &c., constitutes a conducting bridge.

4. By means of a metallic junction between the lightning-rod and pipes the dangers above specified can be avoided.

When the connection is established, there is, in the majority of cases, immunity from danger, and in no instance can such a junction be productive of an increase of risk.

5. It is necessary that the connection be made with the two systems of pipes.

It is not advisable to rely on making a " good earth " with one system of mains only, but both should be joined to the lightning- rod, in order to avoid lateral inductive discharges. Owing to the close proximity of the two sets of pipes, either in buildings or undergTound, such discharges and their attendant sparks are very common, when only one set is connected with the rod.

6. Objections to a metallic junction between the lightning- conductor and pipes.

During repairs to the pipes, the metallic continuity is sometimes intermitted at certain places, and it is contended that, at these

Abstracts.] WITH WATER- AND GAS-PIPES. 503

parts, not only the pipes, but the workmen engaged in repairing them, are exposed to danger while a thunderstorm is in progress. There is, however, ample evidence to show that this risk is slight, and can, if necessary, be easily prevented by bridging over the gap with a wire rope. Any j^ermanent lack of continuity at the joints can be avoided by making the use of lead caulking compulsory.

7. Is a special earth-plate required for the lightning-rod when the latter is joined to the mains ?

The system of pipes with which a rod is connected may form a metallic network of considerable extent around a building, and as this network may have gaps in it, it is best to join the lightning- conductor to a separate earth-plate.

8. The kind of junction to be employed, and in what jiart of the conducting system it should be made.

The metallic rope or band used for the connection should have the same conductivity as that of the lightning-rod, and be well soldered to the latter. The junction between the metal surfaces is, in this case, easily made, but greater i^recaution is required in respect of the connection with an iron pipe. The rope or band should therefore be soldered to an annular clip of large internal surface ; the part of the pipe where the joint has to be made is scraped clean and a piece of sheet-lead placed around it ; the clip is then put over the lead and tightened up. The joints should be made in such places as are easy of inspection, and the junction with the iron pipe ought to be on that side of the principal meter which is nearest to the mains. A supplementary connection between the lightning-rod and branch-pipes in the topmost storeys is also usefiil, but in this case the water- and gas-meters should be bridged by a metallic by-pass to guard against any want of con- tinuity incidental to their construction. The joints in the pipes must also conduct well, or be bridged over in the same manner.

In order to avoid any damage to the mains that might arise from galvanic action set up between the copper earth-plates and iron pipes, the ground-connection of the lightning-conductor can also be effected by means of an iron j^late.

9. Precautions to be observed when a house has no lightning- rod, but is fitted with gas- and water-pipes.

The parts of the pipes nearest to the outer walls and roof should be connected with strong wires which are led to the outside of the house. The joints of the pipes and the meters must then be protected as explained above. Such precautions, are, however, only a makeshift, and must not be regarded as an efficient substi- tute for good lightning-conductors.

J. J. W.

504 DECOMPOSITION OF SALT BY ELECTKOLYSIS. [Foreign

The Decomposition of Salt hy Electrohjsis. By N. N. Beketoff,

(Zapisky Iniperatorskavo Eusskavo Technicheskavo Obstchestva, 1888, p. 25.)

After pointing out the great importance of common salt to the arts and mamifactures, and the abundant and widespread oc- currence of this suhstance in Russia, the Author proceeds to dis- cuss the best methods of manufacturing soda in that country ; and he advocates the decomposition of salt into its component parts by electrolysis when in a fused state, confirming his view by the following calculation : — •

The heat evolved in the combination of sodium with chlorine being 96 '7 calories, the electromotive force required for the decomposition of salt is 4 • 5 volts. The conductivity of fused salt is 8660 ; that of mercury, being 100 or 150 millions ; that of silver being taken as 100 millions; the temj^erature of the fused salt being 500'' Centigrade in each case. The number of amperes required to decompose 1830 lbs. (50 poods) of salt, giving 732 lbs. (20 poods) of metallic sodium and 1098 lbs. (30 poods) of chlorine, is 16,000 amperes per twenty-four hours; or, at 5 volts, the work necessary would be 80,000 volts, or about 120 HP. Assuming that 120 HP. i^er day is eqtiivalent to the combustion of 13,176 lbs. of coal (360 poods), and that the fusion of the salt requires 1,464 lbs. of coal (40 poods), there will be required 14,640 lbs. of coal at 7s. 6d. per ton (6 kopeks per pood), or £2 8s., and the cost of 1,830 lbs. of salt at 15s. per ton, or 12s.; the manufacture of 732 lbs. of metallic sodium and 1,098 lbs. of chlorine comes to £3. So that taking the average value of these products at £9 6s. per ton (1 rouble 50 kopeks per pood) and IBs. loss, there results a clear profit of £3 12s.

The Author goes on to mention the best ways of employing the chlorine and metallic sodium.

G. K.

The Wimshurst Machine. By E. Dieudonne.

(La Lumiere Electrique, vol. xxix., 1888, p. 613.)

Dr. Vigouroux has made some improvements in this machine in order to render it more serviceable in electro-therapeutics. The metallic sectors or carriers, instead of being flat throughout, have a raised boss along their vertical axis, and the brushes are so fixed that they only make contact with the ^projections thus formed, whilst the remaining parts of the rotating plates are untouched. When the sectors are flat the brushes sweep across the entire surface of the disks and gradually wear away the varnish. In place of combs, collectors made of flat metal tubing are j^rovided and fixed as near as possible to the plates. The knobs which carry these collectors are mounted direct on glass pillars ; the socket and discharging-rods are dispensed with.

Abstracts.] THE WIMSHUEST MACHINE. 505

According to Dr. Vigoiiroiix, the action of tlie Wimshurst machine may be explained as follows : Assume that a sector on the front plate is in contact with one of the fine copper wire brushes, say, a, whereby a difference of potential is produced and the sector charged with a small quantity of positive electricity, which is retained as it moves away from the brush. As soon as the charged carrier reaches the collector B, towards which the rotation of the plate is bringing it, the initial positive charge is parted with ; the carrier now approaches the other brush b, but while doing so it remains influenced by the positively charged collector B, and in such wise as causes the upper half of the sector to be electrified negatively by induction, and its lower half posi- tively. Directly contact is made with the uninsulated brush b, this distribution is disturbed, and the sector, when receding from b, is in the negative state. It parts with its negative electricity to the second collector A, which likewise acts inductively on the sector until the latter again makes contact with the brush a, and the cycle of effects is repeated. At the second contact with a no initial charge is required, because the carrier is already infiuenced or polarized by the negative electricity on A. All the sectors undergo the same series of changes, and the electrical condition of one-half of the circular plate is s;^^umetrical as regards the other half, but opposite in sign. The same applies to the second plate, but as it turns in the contrary direction, the distribution on the halves of this plate is the reverse of that on the other one. Hence static induction is also exerted between the two sets of carriers, and augments the total effect. When the machine is once started, the principal factor in the continuity of the action is the inductive influence of the collectors. If the insulation of either of them be diminished, thereby causing a corresponding loss of charge, the output of the machine is lessened. This is an effect peculiar to the Wimshurst machine, and distinguishes it from other influence- machines.

J. J. W.

TJie Gldser Influence Machine.

(Elektrotechnische Zeitschrift, 1888, p. 452.)

This machine, which has an exceedingly high efficiency, consists of a fixed horizontal steel rod mounted between two iron uprights, the latter being screwed to a rectangular wooden frame forming the base of the apparatus. On this rod there are two ebonite tubes or axes, each carrying a small pulley and an ebonite drum. The pulleys are keyed on the ends of the axes near the iiprights, and to the other ends are bolted the drums, one of which is of such dimensions as enable it to revolve within the other. The driving gear is arranged to work between the lower part of the two standards ; it consists of two axles WjW.^, each fitted with a large ]uille3^ I'J^ transmitting the motion, l)y means of belts, to the small

506 THE GLASER INFLUENCE MACHINE. [Foreign

pulleys. The rotation of the drums in ojiposite directions is secured by aid of toothed wheels on WjWo. Midway between the uprights are two glass rods fixed on opposite sides of the wooden base ; each rod carries at the top a brass knob to which is screwed a vertical brass rod ending in a second metal knob. In the upper knobs the discharging rods work, and to the lower ones are fixed the collecting combs for the outer drum. Near the middle of the steel rod is a short vertical bar, which carries the collecting combs for the inner drum, these combs being so placed as to form a right angle with the outer ones. Eigidity is given to the upper part of the apparatus by an ebonite tie-rod secured to the upright, and a bar of the same material is screwed crosswise to this rod and serves to hold the ui3j)er jDortion of the discharging ajoi^liance firmly in position.

The initial excitation is imparted by means of a strip of rubbed ebonite held just above the outer drum. The machine can be charged whether the handle be turned clockwise or in the contrary direction, and the electrical effect is good even in damp places. J. J. W.

Banger-Indicator for the Prevention of Collisions at Sea. By P. Marcillac.

(La Lumifere Electrique, vol. sxix., 1888, p. 516.)

This apparatus, which has been devised by Messrs. Orecchioni and Cavalieri, of the French military marine, consists of a sub- marine boat shaped like a fish torpedo, and divided into three separate compartments. The fore part is fitted with the danger- signalling-gear ; the central chamber contains the electric motors, and the one aft, the submersion mechanism and leading-in boxes for the cables. At the bow is a conical contact-j^iece, riveted to a rod, which passes through a stuffing-box into the front part of the boat. When the cone strikes against any obstruction (such as a submerged rock, the hull of a passing vessel, &c.), it is forced inwards, and an arm at the rear end of the rod actuates a contact lever, whereby the circuit of a battery and alarm bell on board the protected ship is closed. The range of action is amplified by means of jointed arms fixed to the cone, and, according to the part of these arms which comes into collision with the obstacle, the rod is either drawn out or pushed in, closing in each case the alarm circuit. The signal boat is driven ahead of the vessel it protects by means of an electric generator, the current being conveyed through two conducting cables to a series of motors mounted on the propeller shaft of the boat. The hydraulic gear for effecting the submersion and maintenance of the boat at a certain depth consists of a three-chambered cylinder, in which three pistons work. The rod carrying the topmost one ^3, ends in a circular disk, which butts against the inner face of a ciirved lever, uaovable on an axis that is carried liy a support on the outer wall of the

Abstracts.] PREVENTION OF COLLISIONS AT SEA. 507

cylinder. The other end of the lever is jointed to a rod having a thread on its lower end, which passes through a bracket screwed to the oiitside of the cylinder. The rod can he raised or lowered by means of a wing-nut, in order to regulate the pressure of the curved lever on the upper piston. The other pistons pj and p., ^i"^ carried by a rod that works in a stuffing-box fitted in the central chamber of the cylinder. Between the faces of p^ and ^3 is a strong helical spring, which transmits the opposing force to the lowest piston _p,. Attached to its rod by a crank-pin is the first arm of a system of levers in connection with the submersion-blades, and as soon as the boat is propelled forwards the pressure of the surrounding water drives the piston p^^ inwards, the length of the stroke and consequent action of the levers being dependent on the amount of the opposing pressure, which has been calculated before- hand for the required depth of submersion. Normally, the position of the submerged boat is in a straight line with the longitudinal axis of the vessel it precedes, but when the latter veers, say, to starboard, the cable on that side is pulled taut, whereas the one on the port side is slackened. At the stern of the boat is a steering- bar, which projects 1^ metre on each side. As the cables are attached to this bar, it partakes of their motion, and the boat is thereby speedily swung into its initial position. In each cable is run a leading-wire for the alarm circuit. The weight of the boat is 900 kilograms for the large pattern and 500 for a smaller size. The cables are 280 to 400 metres long, according to the dimensions of the vessel and its protecting signal-boat.

J. J. W.

Studies on the Gas-Thermometer, and Comparison of tlie Mercury- TJiermometer therewith.

By P. Chappuis.

(Archives des Sciences et Naturelles. Geneva. July-September, 1888, pp. 1 ct scq.^

The Author states that the air-thermometer does not always give uniform results, and that it is necessary to select a definite gas, deviating least from Boyle's law. The Author was commissioned by the International Committee of Weights and Measures to make experiments upon nitrogen, carbonic acid, and hydrogen gases, to determine which was most suital)le for the purpose of a standard gas-thermometer. The apparatus used in the experiments is very fully described, and all possible corrections were made and sources of error eliminated. Standard mercury thermometers were prepared for comparison. The gas-thermometer reservoir used consisted of a cylinder of platinum iridium, and its capacity exceeded 1 litre (63 '4 cubic inches). The coefficient of the expansion of nitrogen gas was found to be 0 • 00367466 for 1° Centigrade (1-8° Fahrenheit) between 0^ and 100° Centigrade (32° and 212° Fahrenheit), and the nitrogen gas-thermometer was found at 40" Centigrade (104°

508 STUDIES ON THE GAS-THERMOMETER. [Foreign

Fahrenheit), to read 0*097° Centigrade lower than the mercury- themiometer.

Two values of the coefficient of expansion of carbonic acid were found at diflerent pressures ; at the higher pressure it was 0-00372477, and at the lower jiressure 0-0037163-i for 1° Centi- grade (1'8° Fahrenheit), between 0" and lOO^ Centigrade.

The carbonic acid gas-thermometer, at 40° CentigTade (104° Fahrenheit), read 0*048 less than the mercury. In the hydrogen experiments the Author foimd the coefficient of expansion of hydrogen to be 0*00366254 for 1° Centigrade (1*8° Fahrenheit), and the hydrogen gas-thermometer at 40° Centigrade (104° Fahren- heit), read 0* 107° CentigTade less than the mercury.

The Author fully discusses the divergences of the expansion of the three gases, and states that the more perfect the gas, the greater the divergence from the curve of the exj^ansion of mercury.

Tables are given to readily correct readings on a mercury-ther- mometer to nitrogen, carbonic acid, or hj^drogen gas-thermometers.

D. C.

The Foisonoiis Action of Water-Gas. By Heinrich Schiller of Zurich,

(Zeitschrift fiir Hygiene, 1888, p. 440.)

A short time ago the workmen in a hat manufactory, in the neighbourhood of Zurich, were attacked by various forms of illness, which manifested itself in headaches, dizziness, vomiting, debility, &c. In the preparation of the hats, iron moulds were used to j)ress the felt, and these moulds were heated by numerous small jets of water-gas, of the description furnished by a Dowson gene- rator. The factory inspector satisfied himself that the illness was probably caused by the diffusion of the unconsumed gas which passed into the atmosphere, in consequence of the liability of many of the small burners to become extingiiished. The Author, who was charged to investigate the behaviour of the Dowson gas, gives a sketch of its composition and preparation, and an account of the use of water-gas in America, where, in ten gasworks and one hundred towns, it is employed for ilhiminating purposes.

AVater-gas is only now used for lighting and heating at one jilace in Switzerland, in Messrs. Sulzer's works at Winterthur, but Dowson gas has been introdiiced in several different localities. The points the Author had to determine are formulated in a series of propositions, eleven in number, which deal with the effects of a mixture of the gas with the atmosphere on animals, the least quantity actually fatal to life, the percentage harmful, the toxic agent in the gas, the sjnnptoms of Dowson gas and Avater-gas poison- ing, &c. An apparatus of special design was contrived to prepare uiixtures of Dowson gas in various proportions, an illiistration of which is given. A number of ex]icriuicnts were conducted with

Abstracts.] THE POISONOUS ACTION OF WATER-GAS. 509

cats, rabbits, guinea-pigs, mice, and frogs, the results of which are set forth in tables, and similar experiments were carried out with water-gas. The conclusions arrived at are as follow : — As the gases are devoid of smell, it is only when evolved from fuel con- taining sulphur compounds that an odour of sulphuretted hydrogen is perceptible. Both gases are strongly poisonous. The symptoms are identical with those of poisoning by means of carbonic oxide gas. A fatally poisonous mixture is an atmosphere with aboiit 1 per cent, in the case of water-gas, and 1 • 5 per cent, in the case of Dowson gas. Symjitoms of jioisoning are produced by abotit 1 per mille with the former, and three parts per mille with the latter gas. In non-fatal cases a rapid recovery, as a rule, ensues. The poisonous principle is the carbonic oxide gas.

G. R. R.

Heat of Comhustion of Coal of the North of France. By — Scheurer-Kestner.

(Annales de Chimie et de Physique, 1888, p. 262.)

Thfe Author has determined the heat of combustion of twenty- one samples of coal from the Bassin du Nord, and the coal-basins of Charleroi and the Pas-de-Calais. A Favre and Silbermann-calo- rimeter was used, and the oxygen in which the combustion was conducted was moist, in order to avoid error by carrying away water of combustion. Chemical analyses of the samples were made, and the whole results are tabulated in the Paper.

In the Tables the actually observed heat of combustion is given, and compared with values calculated by different methods from the analysis.

The maximTim observed heat of combustion was 9,030 calories (35,831 English heat-units), and the minimum 8,5-15 calories (33,906 English heat-units). The nature of the coke produced from each sample is also given.

D. C.

( 511 )

INDEX

TO THE

MINUTES OF PKOCEEDINGS,

1888-1889.— Part I.

Accident to tlie Czar's train at Borki, Oct. 1888, 358.

Accumulators, Gadot, pattern 1888, 489.

Adams, Gr. F., admitted student, 112.

, T., Paper on the friction of the slide-valve and its appendages, 167, 18G,

189.

" Admiral " class of British -war-vessels. See Ships of War.

" Admiral Nakimotf," armour-clad cruiser, electric light installation on the, 498

Alexander, Prof., surveyor-general of Hawaii, and Lyons, C, district maps of Windward Hawaii, 196.

, Prof. T., and A. W. Thomson. — Correspondence on Friction-Brake

Dynamometers : Etficiency of the Ajipold brake if properly used, 70.

Allen slide-valve, experiments on the friction of the, 173.

" Alpine Engineering" L. F. Vernon-Harcourt (S.), 237. — Alpine passes and roads, 237. — Semmering railway, 241. — Brenner railway, 245. — Mont Conis Fell railway, 249. — Considerations affecting main lines through the Alps, 253. — Mont Cenis railway, 255. — Mont Cenis tunnel, 257. — Influences of the Mont Cenis railway, 261. — St. Gothard railway, 262. — St. Gothard tunnel, 265. — Arlberg railway, 268. — Arlberg tunnel, 269. — Proposed Mont Blanc railway and tunnel, 271. — Proposed Great St. Bernard railway and tunnel, 272. — Pro- posed Simplon railway and tunnel, 274. — Comjiarison between the three Alpine schemes, 276. — Appendix : Data concerning Alpine railways, 278.

Amos, C. E. See Appold.

Anderson, A., elected member, 113.

Andrews, W. W., elected associate member, 113.

Annealing, influence of, on the strength of steel, 129 et seq.

" Anson," H.M.S., speed trials of, 325 et seq.

Antimony, electrolytic reduction of, from ores, 478.

Appleby, 0. J. — Correspondence on the Strength of Bessemer-Steel Tires : Experi- ments with oil-tempered axles of Bessemer steel, 160. — Mistaken impression that Bessemer steel would not harden, 161. — Tempering a more proiicr term than oil-hardening, 162.

Appleton, E., transferred member, 112.

Appold, W., and Amos, C. E., compensating brake-dynamometer of, 7 et seq. — Ditto, experiments with, 71.

512 INDEX.

Aqueduct. " The Fnihxre of the Kali Nad i Aqueduct on the Lmcer Ganges Canal" abstracted by W. H. Thehvall (S.), 283.— Original design of the aqueduct, 283. — Revised design, 284. — Partial destruction of the aqueduct, 285. — Complete destruction, 286. — Design for new aqueduct, 287. Arc lamps, the regulation of, 495.

Arnold, J. O. " On the Influence of Chemical Composition on the Strength of Bessemer-Steel Tires" 115. — Discussion on ditto: Various and occasionally mutually destructive criticism of his Paper, 155. — Molecular change versus " fatigue," 156. — Steel identical in composition capable of varying largely on the machine, 156. — Alleged change wrought by annealing in the state in which carbon exists in steel, 156. — Brittleness of unannealed steel castings, 157. — ■ Quality of open-hearth steel tires giving the best results, 157. — Chemical com- position of the struts used in his tests, 158. — Alleged rarity of tire-breakages of late years, 158. — Fractures of tires on German railways, 158. — The practice of annealing, 159. — Analysis of Belgian steel used for the tires on Victorian Government railways, 159. — Simple means of bringing his experiments into accord with those of others, 159.

Ashworth, J., memoir of, 394.

Aspinall, J. A. P. Discussion on the Strength of Bessemer- Steel Tires : Crystallization of steel after long work, 140. — Annealing of mild steel, 141. — The suggestion to trace the life of a tire from day to day throughout its life impracticable, 141. — Wear of the tire often the measure of the period during which an engine could be kept out of the shops, 141. " TJie Friction of Locomotive Slide- Valves" 167. — Discussion on ditto: Further exijlanation of the experiments, 179_ — Earity of broken valve-spindles, 191. — Use of thick glycerine in the hvdraulic cylinder of his apparatus, 191. — Lubrication of slide-valves, 192. — Diaphragms versus cup-leathers for the indicator, 192. — Lubrication, 192. — Mode of applying the weights, 193. — Very slight wear of cast-iron eccentric-straps, 193. — Substitution of cast-iron for brass valves on the North-London railway, 193.

Atkinson, L. B., A.K.C., elected associate member, 113.

Atmospheric and seismic disturbances and the disengagement of firedamp, 473.

Attard, W., elected associate member, 113.

Austin, C. E., rapid survey made for, in Asia Minor, 209.

Aveling and Porter, Messrs., mode adopted by, for lubricating the axles of traction-engines, 187.

Ayres, P. P., elected associate member, 113.

Ayrtou, Prof. W. E. — Discussion on Friction-Brahe Dynamometers : Transmission and absorption-dynamometers for electro-motors, 51. — Dynamometer-couplings for electro-motors, 53. — Practice of varying the arc of contact to neutralize the variation of the coefficient of friction in compensating absorption-dynamo- meters, 54. — Kuotted-belt absorption-dynamometer, 55.

Bagnall, H., M.A., B.E., transferred member, 112.

Balk's compensating brake-dynamometer, 10.

Bamford, C. F., admitted student, 112.

Banks, J., elected associate member, 113.

Barker, — , mode of testing steel tires adopted by, for the Great Indian Peninsula

Railway, 135. Barnes, C. A. A., elected associate member, 113. Barr, Prof. A. Discussion on Friction-Brahe Dynamometers : Question whether a

friction-brake could or could not give a scientifically accurate measure of the

INDEX. 513

work done by a motor, 29. — Different forms of dynamometer, 30. — Hia rope- brake dynamometer, 31. — Spring-balances v. weights for dynamometers, 31. — Rope-brake attributed to Prof James Thomson more probably due to Sir William Thomson, 33. — Froude's brake at the Owens College, 33. — Use of the dash-pot, 34. — Prof. James Thomson's mode of regulating the action of his brake, 34. — Using of Eankine's formula in regard to friction-brakes, 35. — Apix)ld's compensating-levers, 36. — Brake-trials of the Royal Agricultural Society, 36.

Barratt, S. H. H., admitted student, 112.

Bates, O., elected associate member, 113.

Batterbee, R. C, memoir of, 382.

Bayne, T. J., admitted student, 112.

Beattie, W. G., paper by, on a balanced slide-valve for locomotive engines, 167, 186.

Beaumont, W. W. " Friction-Brake Dynamometers," 1. — Discussion on ditto : Ayrton and Perry's brake, 29. — Controversy with regard to compensating- lever brakes, 29. — Brake-trials for the Royal Agricultural Society, 68. — Com- parison of indicators and friction-dynamometers, 68. — Rise and fall of the load on a dynamometer, 68.— Use of springs instead of weights, 69. — Use of the dash-pot, 69. — Alleged abuse of mathematical method in the calculation of the maximum tension on the brake-straps, 69. — Coope's friction-brake, 70. — Use of leather-covered brake-blocks, 70.

Beer system, the, of wire ropeways, 485.

Beketoff, N. N., the decomposition of salt by electrolysis, 504.

Bell, J. McK., elected member, 113.

Bentabol, H., a folding levelling-staff, 415.

Berkley, G., Vice President. — Discussion on the Strength of Bessemer Steel Tires ; Importance of chemical knowledge in regard to steel, 148. — Alleged ten- dency among engineers to specify high tensile steel for tires, 149. — His practice in regard to testing tires, 149. — Effects of cliromium and other elements on steel, 150. — Experiments by Professor Kennedy and Dr. Riley in 1882, 150. — Practical impossibility of tracing the life of a tire day by day, 151. — Objec- tions to chemical analyses, 151. — His practice in mechanical testing, 152.

Bessemer-steel tires. See Steel.

process, a new modification of the, 477.

Betchworth tunnel, Dorking. See Tunnel.

Beveridge, J., elected associate member, 113.

Bidois, C, pulverization of clay, and its application at the works of the Societe Arnaud, Etienne & Cie., 422.

Binet, E. P., elected associate member, 113.

Biss, C. H., admitted student, 112.

Blackshaw, W., elected associate member, 113.

Blasting under water at the Panama canal works, special plant for, 448.

Bloys van Treslong, C. — Correspondence on the Withani Out/all Improvement Works: Three works of different character simultaneously executed on the Witham, 108.— Grand Sluice, 109.— Cut through the Scal^j, 109.— Cutting of the Hoek van Holland, 109.— The case of the Donge and the Oude Maasje analogous to that of the Witham and the Welland, 109. — More advantageous on the score of cost to bring the outlet of the Welland to Clayhole than to Lynn Deeps, 109. — Small width of the Witham channel to Boston, 110.^ Means of providing a deep-water outlet for the Welland, 110. [the INST. C.E. VOL. XCV.] 2 L

514 INDEX.

Bonnami, — , yield of li}-draulic motors, 421.

Borchers, W., electroh'tic reduction of antimony from ores 478.

Boston Dock Cut, the, 97.

Bottle, F. ^y., admitted student, 112.

Boulle, M., Jandin's comi^ressed-air dredger, 450.

Bow-gii'ders. See Girders.

Bradley, A. "W., admitted student, 112.

Bragg, J. W., B.A., admitted student, 112.

Brakes, friction-. See Dynamometers.

Brancher, A., the Laon steep-gradient railway, 468.

Brauer, Prof. E., Paper on brake-dynamometers by, 28.

Brenchley, J. V., elected associate member, 113.

Bridge, New York and Brooklyn, the cable railway on the, 453.

over the Po at Casalmaggiore for the Parma-Brescia railway, 438.

Bridges, highway, of iron and steel, 430.

, railway, inspection and maintenance of, 432.

Brown, J., memoir of, 361.

, W. A., memoir of, 363.

Brownlow, Col., E.E., and the design for the Kali Nadi aqueduct, 289. Bruce, Sir G. B., President — Bimussion on the Witliam Outfall Improvement Works : Question of Mr. Duif Bruce as to the mode of handling the material dredged from the river, 104.

, W. D. — Discussion on the Withmn Outfall Improvement Works : Question

as to the mode of handling the material dredged from the river, 104. Brush, C. B., facts in relation to friction, waste and loss of water in mains, 459. Brustlein, H. A. — Correspondence on the Strength of Bessemer-steel Tires: The metal experimented uj^on by Mr. Arnold a manganese-steel with a small percentage of chromium, 162. — Testing-jjractice at the Unieux works, Francci 162. — Superiority of chromium to manganese as an alloy for iron, 162. Bucknall, H., admitted student, 112. Budge, 0., elected member, 113.

Building materials, methods of testing the resistance of, 416.

Bunuugal, parish of, county of Eipon, Victoria, survey of, 215.

Burgess, S. E., elected associate member, 113.

, W. E., admitted student, 112.

Burmau, E. S., elected associate member, 113.

Burstall, H. R. J., Wh.Sc, elected associate member, 113.

Burton, C, elected associate member, 113.

Butler, W. R., elected associate member, 113.

Button, F. S., B.E., elected associate member, 113.

Buyers, W. L., elected member, 113.

Byng, H. A., his improvements in Balk's compensating brake dynamometer, 10.

Cable railway. See Railway.

Cables, submarine, on the measurement of the resistance of, 490.

Caland, P. — Coirespondence on the Witham Outfall Improvement Works : Efficient

nature of this work of river-improvement, 105. " Camperdown," H.M.S., sj^eed-trials of, 325 et seq. Canal, Ganges. " The Failure of the Kali Nadi Aqueduct on the Lower Ganges

Canal," abstracted by W. H. Thelwall (S.), 283. , Panama, special plant for blasting under water at the works of the, 448.

I

INDEX. 515

Canals, Hague, the, renewal of water in the, 450.

Capper, D. S., M.A., " The Speed-Trials of the latest additions to the Admiral class of British War-Vessels," 325.

Carbon, influence of, on the strength of steel, 115 et seq.

Carr, H., memoir of, 364.

Casalmaggiore bridge. See Bridge.

Cell, Ressner's, account of a series of experiments made on, 489.

Cements, methods of testing, 416.

Chapman, G. J., elected associate member, 114.

Chappuis, P., studies of the gas-thermometer, and comparison of the mercury- thermometer therewith, 507.

Chavannes, R., the electric lighting of the City of Geneva, 496.

Chesuau, G., on the relation between seismic and atmospheric disturbances and the disengagement of firedamp, 473.

Chromium, influence of, on the strength of steel, 119 e< seq.

Chukoloff, W., account of a series of experiments made on Hessner'a cell, 489.

Clark, E. G., elected associate member, 114.

Clarke, W. J., elected associate member, 114.

Clay, pulverization of, and its application at the works of the Societe Arnaud Etienne & Cie., 422.

Cleaver, H. L., elected associate member, 114.

Clifton, C. T., elected associate member, 114.

Clyde. See Eiver Clyde.

Coal of the North of France, heat of combustion of the, 509.

Cochrane, J. H., admitted student, 112.

Cockerill and Co., J., Seraiug, steel tires supplied by, for the Victorian Government railways, 142.

Coey, R., assistance rendered by, in Mr. Aspinall's experiment on the friction of locomotive slide-valves, 178.

Collisions at sea, danger-indicator for the prevention of, 506.

Compensating-brake dynamometers. See Dynamometers.

Congress on inland navigation, proceedings of the second, at Vienna, 1886, report of the French delegates on the, 440.

Connor, Major A. S. W., memoir of, 396.

Coomber, W. H., elected associate member, 114.

Coope's dynamometer-brake, 49.

Cope, W. A., admitted student, 112.

Copper-refining, electrolytic, in Hungary, 479.

Cost prices on railways, 469.

Cowan, D., elected member, 113.

Cowper, E. A. — Discussion on the Strength of Bessemer-Steel Tires : Falling-weight test of no avail to alter the molecular arrangement of a steel tire, 139. — Experiments of the London and North Western Railway on axles, 140. — Alleged brittlcncss of steel castings prior to annealing, 140. — Influence of the amount of work put upon steel, 140. — Hardness in a tire not necessarily provocative of fracture, 140. — Discussion on the Friction of Locomotive Slide- Vahes: Less power required for east-iron valves than for those of other material, 190. — Balanced slide-valve, 190. — Reasons why a large valve required more power to work with a short stroke than with a long one, 191. — Breakages of slide-rods in marine engines, 191.

Cox, F. N., admitted student, 112.

2 L 2

516 INDEX.

Cox, — , signalling-apparatus on the St. Gothard Railway, 470.

Crawford, R., elected associate member, 114.

Cripps, F. S., elected associate member, 114.

Crook, C. R. E., elected associate member, 114.

Cross, W. — Discussion on the Friction of Locomotive Slide-valves : Subject very little understood, 179. — Fracture of valve-spindles, 179. — Apparatus con- structed by Mr. Marshall and himself for obser\-ing the friction of marine- engine slide-valves, 180. — Doubtful advantage of cup-leathers in Mr. Aspiuall's apparatus, 181. — Paper by Messrs. Marshall and Weighton on high-sijeed engines, 182.

Cubitt, Sir W., recommendations of, in respect to the River Witham, 81.

D'Aeth, J., elected associate member, 114.

Danger-indicator for the prevention of collisions at sea, 506.

Dare, H. H., B.E., admitted student, 112.

Dashpot, use of the, with friction-brakes, 34, 69.

Davidson, W., elected member, 113.

Davies, J. T. L., admitted student, 112.

, M. "W., elected associate member, 114.

, W. A., elected associate member, 114.

Dawney, A. E., admitted student, 112.

Dawson, J. S., admitted student, 112.

Deas, J., his Paper on the River Clyde, 279.

Debenham, F. B., admitted student, 112.

Denny, R., memoir of, 369.

Deprez's compensating brake, 4, 60.

Desrozier's new disk-dynamo, 488.

Dickinson, T. R., elected associate member, 114.

Diesclhorst, W., elected associate member, 114.

Dieudonne', E., the Wimshurst machine, 504.

Dimier, P., A.K.C., elected associate member, 114.

Disinfecting power of steam. See Steam.

Donkin, B., jun. Correspondence on Friction-Brake Dynamometers : Experiments in 1876 with an Appokl brake fitted with Amos adjusting-levers, 72.

Dredger, compressed-air, Jandin's, 450.

Dudley, Dr. P. H., his experiments on the chemical composition of steel rails, 133.

Durand-Claye, L., methods of testing the resistance of stones, cements, and other building-materials, 416.

Dynamo, disk-, Desrozier's new, 488.

Dynamometers. " Friction-Brahe Dynamometers" "W. "W. Beaumont, 1. — Mea- surement of power with friction-dynamometers, a subject of dispute, 1. — Prony brake and modifications, 2. — Deprez brake, 4. — Imray brake, 6. — Amos and Appold's brake, 7. — Balk's compensating brake, 10. — Water-cooled brake of Messrs. Richard Garrett and Sous, 12. — Druitt Halpin's brake, 15. — Professor James Thomson's brake, 16. — Professor Kennedy's brake, 17. — Proportions and dimensions of brakes, 17. — The Appold friction-brake dyna- mometers, 22. — Results obtained with a water-cooled brake by Messrs. J. and H. McLaren and by the Engineers to the Royal Agricultural Society, 27. — Discussion : ^Y. W. Beaumont, 29, 68.— Prof A. Barr, 29 ; Dr. E. Hopkinson. 37; R. E. Froude, 39; P. W. Willans, 47; Prof. A. B. W. Kennedy, 49; G. Kapp, 50; Prof. W. E.^Ayrton, 51 ; J. Imray, 55; Prof R. H. Smith, 56;

INDEX. 517

D. Halpin, 62; J. Goodman, 67. Correspondence: Prof. T. Alexander and A. W. Thomson, 70; B. Donkin, jun., 71 ; F. Garrett, 73; Prof. A. Jamieson, 74; W. Schonheyder, 76; J. E. Sweet, 77.

Earth-pressure underground, an apparatus for measuring, 471.

temperature, diminution of, in deep mines, 471.

Earthwork. ^^ Hurst's Triangular Prismatic Formula /w EartMcorh compared luith the Prismoidal Fonnida" J. W. Smith (S.), 229.

Earthworks on the railway from Gien to Auxerre, consolidation of the, 466.

Eddy, E. M. G., elected associate, 114.

Edwards, L., designer of the Grand Sluice, River Witham, 83, 100, 105.

Eiffel, G., the Garabit viaduct, 434.

Electric light installation on the armour-clad cruiser " Admiral Nakimoff," 498.

. Regulation of arc lamps, the, 495.

. Self-regulating electric search-light, 497.

lighting of the City of Geneva, 496.

— machine, Wimshurst, 504 ; Glaser, 505.

winding-engine at Neu Stassfurt, 500.

Electrical tramcars in Paris, Philippart's, 492.

Electrolysis, the decomposition of salt by, 504.

Electrolytic copper-refining in Hungary, 479.

reduction of antimony from ores, 478.

Ellis, E., elected associate member, 114.

Ely, T. N., elected member, 113.

Embankment of the Po, at Turin, the, 449.

Emdin, A. R., admitted student, 112.

Emery testing-machine, application of the principle of the, to slide-valve in- dicators, 185.

English, Col., experiments of, on the increase of tenacity in steels of various chemical composition, 137.

Eversley, Viscount, memoir of, 360.

Examination of surveyors in the Australian colonies, 215.

Extension, critical, of bodies strained simultaneously in several directions, 410.

Falling- weight test v. other modes of determining the strength of steel, 166 et seq.

Fein, W. E., self-regulating electric search-light, 497.

Fell, J. B., his centre-rail railway over Mont Cenis, 249.

" Ferndale," steamer, sunk in the entrance channel of the Port of St. Nazaire, raising the, 465.

Firedamp, on the relations between seismic and atmospheric disturbances and the disengagement of, 473.

Fitz, N., B.E., admitted student, 112.

FitzGerald, M. F. — Discussion on the Witham Oidfall Improvement Worhs: Importance in works of river-improvement of securing a free outfall at the lowest possible level, 166. — Periodical excess of the flow of a river over the observed rainfall as instanced by the Shannon, 106. — Effects of the new Witham outfall channel on the discharge-channel of the V/elland, 108.

Flood of the river Wytham, October, 1883, 87, 101.

Flower, J. J. A., memoir of, 384.

Floyer, G. W., elected asssociate member, 114.

518 INDEX.

Foster, J. F., admitted student, 112. Fowler, C. P., admitted stiideut, 112.

, J., memoir of, 371.

, P., transferred member, 112.

Francis, W., memoir of, 374. Fraser, A., admitted student, 112.

, P. A., transferred member, 112.

Frere, F. H., A.K.C., admitted student, 112. Friction, new theory of, 407.

brake dynamometers. See Dynamometers.

of water in mains, 459.

, slide-valve. — " The Friction of Locomotive Slide-valves." J. A. F.

Aspinall, 167. Friederichs, H. F., admitted student, 112. Friese, F. K. M. von, diiferences of level, in the mines of Austria and Hungary,

472. Froude, R. E., Discussion on Friction-Brake Dynamometera : Variations of the

frictional resistance in respect of the accuracy of the record, 39. — Turbine

dynamometer invented by Mr. William Froude, 45. , W., his turbine djniamometer-brake, 33 et seq. — His mode of determining

the internal resistances of the engines of war- vessels, 330. — His modification of

the Griffiths form of screw-propeller, 326. Fulton, A. R. AV., transferred member, 112.

, J. E., transferred member, 112.

Furnace, smoke-consuming, Wilmsmann's, 464.

Gadot accumulators, pattern 1888, 489.

Ganga Ram, Rai Bahadur, transferred member, 112.

Garabit viaduct. See Viaduct.

Garrett, F. — Correspondence on Friction-Bralic Dynamometers : Exaggerated im- portance sometimes attached to the construction of friction-brake dynamo- meters, 73. — His experiment with the water-cooled brake referred to in Mr. Beaumont's Paper, 73.

, R. and Sons, water-cooled brake used by, 12, 18, 21.

Gas, oil-. " Tlie Mmuifacture of Oil-Gas on the Pintsch System, and its application to the Lighting of Railway-Carriages" G. M. Hunter (S.), 218.— Manufac- ture, 218. — Compression, 222. — Application to the lighting of railway-carriages, 223. Appendix : Experiments to test the candle-power of gas, 228.

Gas-burners, comparative trials of various, 462.

Gas-thermometer. See Thermometers.

, water-, the poisonous action of, 508.

Geneste, F. A. B., memoir of, 375.

Geneva, electric lighting of. See Electric light.

Gerrard, A. S., transferred member, 112.

Gibson, A. S., elected member, 113.

Giddiugs, C. M., paper on the friction of slide-valves, 182.

Gilbert, C. H., Wh.Sc, admitted student, 112.

Giles, A., M.P. — Discussion on the Witham Outfall Tmpjrovement Worhs : General effects of the new works, 96. — Boston Ocean Dock bill, 97. — Moral obligation of all the persons whose land was improved by the works to contribute to their cost, 97

INDEX. 519

Girders, bow-, jointed, the theory of, 426.

of the machinery hall of the Paris Exhibition of 1889, erection of the,

440.

Glaser influence machine, the, 505.

Gloyne, E. M., elected associate member, 114.

Gold ores, the smelting of, in Eastern Hungary and Transylvania, 480.

Goliath rail, Sandberg's, 354.

Good, G. L., elected associate member, 114.

Goodman, J. — Discussion on Friction-Brake Dynamometers : Use of leather-covered brake-blocks, 67. — Discussion on the Friction of Locomotive Slide-Valves: Diajahragm indicator on the princiiole of the Emery testing-machine, 185. — Friction of the glands of slide-valve spindles, 186. — Friction of balanced slide- valves, 186. — Mode of experimenting adopted by Mr. T. Adams and Mr. W. G. Beattie, 186. — Lubricants for slide-valves, 187. — Mode of lubricating trac- tion-engine axles adopted by Messrs. Aveling and Porter, 187.

Gradients, parabolic, for railways, principles of, 165.

Grand Sluice, river Witham. See Sluice.

Gray, J., of the Manchester and Liverpool railway, the first to adopt balanced slide-valves for locomotives, 191.

Gribble, T. G. " Preliminary Survey in Neio Countries, as Exemplified in the Survey of Windivard Hawaii," (S.), 195.

Grimshaw, J. W., elected associate member, 114.

Gruber, Prof. M., on the disinfecting action of a current of superheated steam, 461.

Haase's method of shaft-sinking, 475.

Habermann, R. and J. von Hauer, a winding-engine with spiral balance-drum, 484.

Hadfield, R. A. — Discussion on the Strength of Bessemer-Steel Tires : Experiments of Mr. Dudley in the United States, 133. — Tests with varying proportions of manganese, 133. — Tests by the Terre Noire Company in 1878, 134. — Effect of the work put upon steel, 134. — Mode of testing steel tires adopted by Mr. Barker of the G. I. P. railway, 135. — Microscopic examination of steel, 135. — Work of Mr. Howe of Boston, U.S.A., 135. — Influence of temperature at working on the future characteristics of the steel, 136. — A good pyrometer a desideratum, 136.

Hague canals, the, renewal of the water in the, 450.

Hallett, H., admitted student, 112.

Halpin, D., his water-cooled brake-dynamometer, 15. — Discussion on Friction- Brake Dynamometers : Alleged inability of the dynamometer to give scientifi- cally accurate measurements of work done, 62. — Brake-tests for the Royal Agricultural Society at Newcastle-on-Tyne, 62. — Use of the Moscrop recorder in brake trials, 63. — Mode of conducting the Newcastle trials, 64. — Discussion on the Friction of Locomotive Slide- Valves : Paper by Mr. C. M. Giddings in the Transactions of the American Society of Mechanical Engineers, 182. — Results of experiments with Mr. Giddings' indicator, 184. — Causes of the variation of the resistance during the stroke, 185.

Hamilton, W. L., admitted student, 112.

Harbord, F. W. — Corresjyondence on the Strenr/th of Bessemer-Steel Tires: Con- clusion that high tensile steel can only be obtained at the expense of ductility and general reliability, 163.

520 INDEX.

Harbour works, La Rochelle, the new, 454.

Hardening, influence of, on the strength of steel, 129 et seq.

Harpur, S., memoir of, 385.

, TV., transferred member, 112.

Hartley, Sir C. A., K.C.M.G. — Discussion on the Witham Out/aM Improvement

Works : Threefold results of the improvement, 93. Har\-ey, F. J., admitted student, 112. Haselkoos, N., on the testing of paper, 420. Hauer, J. von. See Habermann. Hawaii, survey of. See Survey. Hawkins, G., memoir of, 397.

, H., admitted student, 112.

Hawkshaw, Sir J., recommendations of, in respect of the river Witham, 81, 92

et seq. , J. C., Discussion on the Witham Outfall Improvement Works : Reason

why the improvement had been so long deferred, 92. — Condition of the district

before the completion of the works, 92. Henard, E., erection of the large girders of the machinery hall at the Paris

Exhibition of 18S9, 440. Hessner's cell, account of a series of experiments made on, 489. Hewitt, J. E., elected member, 113. Higgins, G., elected associate member, 114. Highway bridges. See Bridges. Hills, Staff-Commander G. H., memoir of, 398. Hobbs, G., elected associate member, 114. Hobhole sluice, river "Witham. See Sluice. Hobley, C. W., admitted student, 112. Hodgkinson, A. J., elected associate member, 114. Hoek van Holland, cutting of the, 109. Holliday, J., admitted student, 112. Holmes, E., elected associated member, 114. Hopkinson, Dr. E. — Discussion on Friction-Brake Dynamometers : Froude's

hydraulic brake, as improved by Prof. Osborne Reynolds and used at the

Owens College, 37. — Disadvantages of the Prony brake, 38. — Brakes made for

the triple-expansion engines in the Whitworth laboratory at the Owen

College, 39. Hospitaller, E., the regulation of arc-lamps, 495. Howe, — ., his work on the " Metallurgj' of Steel," 135. Hughes, "VV., M.E., transferred member, 112. Hull, P. W. admitted student, 112. Humphrys, Tennant «fe Co., machinery of H.M.S. " Anson " constructed by, 325

et seq. Hunter, G. M. " The Manufacture of Oil-Gas on the Pintsch System, and its

application to the Lighting of Bailway-Carriages," (S.), 218. Hurst's prismatic formula for earthwork, 229 et seq. Hurtzig, A. C. — Discussion on the Witham Outfall Improvement Works: Probable

effect of the larger volumes of water now passing in and out of the river, 103.

— Adoption of sandstone for the hollow quoins of the lock, 108. — Eifect of

onshore gales in promoting silting up, 103. Hydraulic mortars. See Mortars. slide-valve indicator, Aspinall's, 168 et seq.

INDEX. 521

Ikin, A. J., elected associate member, 114.

Imray's self-adjusting brake dynamometer, 6.

, J. — Discussion on Friction-Brake Dynamometers : Investigation by the

late Mr. W. Fronde and himself of the conditions of the frictional hold of

belts on pulleys, 55. — Description of the brake used, 56. Influence-machine, the Glaser, 505.

^ tlie Wimshurst, 504.

Inland navigation congress, second, at Vienna in 1886, reports of the French

delegates on the proceedings of the, 441. Instruments used in surveying, 202 et seq. Iron highway-bridges, 430. Itizkowski, R., regulation of the Isar according to Wolfs method, 445.

J. R., water-supply in the Kingdom of Wurtemberg, 458. James, A., B.A., admitted student, 112.

Jamieson, Prof. A. — Correspondence on Friction-Brahe Dynamometers : Dynamo- metrical tests of a Griffin gas-engine and forms of brake used, 74. Jandiu's compressed-air dredger, 450. Jaques, W. H., Lieut. U.S.N., elected associate, 114. Jasper, N. P., elected associate member, 114. Jenkin, C. F., B.A., admitted student, 112. Jenkins, W. J., elected associate member, 114. Jones, H. H., admitted student, 112.

, H. S., elected associate member, 114.

Joyan, — . See Lethier.

Kali Nadi Aqueduct. See Aqueduct.

Kapp, G. — Discussion on Friction- Brake Dynamometers : Rope-brake for testing electro-motors, 50.

Keith, J., elected associate member, 114.

Kennedy, Professor A. B. W. — Discussion on Friction-Brake Dynammneters : Brakes tested in motor-trials undertaken for the Society of Arts by himself, Dr. Hopkinson, and Mr. Tower, 49.

and Riley, Dr., experiments by, on tire-steels of

varying chemical composition, 150.

Kershaw, S., information as to the manufacture of oil-gas on Pintsch's system, 227.

Kerviler, — , and — Prcverez, raising the steamer "Ferndale," sunk in the entrance-channel of the Port of St. Nazaire, 465.

Killou, H. B., admitted student, 112.

King, J. W. — Discussion on the Strength of Bessemer-Steel Tires : Great advance in steel manufacture since the date of Mr. Arnold's experiments, 163.

Kirkaldy, W. G. — Discussion on the Strength of Bessenner- Steel Tires : Sufficiency of mechanical tests for steel, 146. — Increase of hardness not necessarily co- incident with increase of tensile strength, 146. — Importance of the mode of preparing the test-piece, 147.— Tests of steel by Messrs. D. Kirkaldy and Son, 148.

Kolokoltzoff", Lieutenant, electric-light installation on the armour-elad eraiser " Admiral Nakimoflf," 498.

Lafi'arguc, J., Gadot accumulators, pattern 1888, 489. Lamansky, S., comparative trials of various gas-buruers, 462,

I

522 INDEX.

Lambert, J., admitted student, 112.

Landon, W. H. F., admitted student, 112.

Landslip at Zug, Switzerland, July 5, 1887, 411.

Lang, S. A., admitted student, 112.

Langley, H. W., admitted student, 113.

Lawton, R. J., elected associate member. 111.

Lee, A. G. V., admitted student, 113.

Lethier, — , and Joyan, — , consolidation of earthworks on the railway from

Gien to Auxerre, 466. Levelling-staif, folding, a, 415. Levels, lining, cast-iron tubbing for, 476.

Leverich, G., the cable railway on the New York and Brooklyn bridge, 453. Lewis, W. B. — Discussioti on the Strength of Bessemer-Steel Tires : Quality of the

steel in a contract for tires let by the Victorian Government to Messrs.

Cockerill and Co., 142.— Similar tires since made by three English firms, 143.

— Qualities of the steel in the foregoing tire-contracts, 143. — Mr. J. T. Smith's

observations of the wear of soft steel rails on the Furness railway, 143. Libbis, G. H., elected associate member, 114.

Lightning-conductors, on the connecting of, with gas- and water-pipes, 501. Lindsey, E. S., admitted student, 112. Littlejohn, H., elected associate member, 114. Locomotive slide-valves. See Slide-valves. Lopes, G. " The Reparation of the Betchicorth Tunnel, Dorlcing, on the London,

Brighton and South Coast Eailway" (S.), 291. Lorden, F. L., admitted student, 112. Loss of water in mains, 459. Lovegrove, E., admitted student, 113. Lyons, A. 0., elected member, 113. C. See Alexander.

Macalister, D.— " The River Clyde " (S.) (Abstract), 279.

Macandrew, H., admitted student, 113.

Macdonald, D. G., elected associate member, 114.

Machinery Hall of the Paris Exhibition of 1889, erection of the large girders of

the, 440. Maclaren, J. W. B., elected associate member, 114. MacLean, L. F., elected member, 113. McCallum, T. S., elected associate member, 114. McDonnell, A. — Discussion on the Strength of Bessemer-Steel Tires: Practice of

annealing tires, 155. McKenzie, L. S., admitted student, 113.

McLaren, J. and H., form of friction-brake used by, 15, 18, 21, 24. Madsen, C. L., on the telephone equation, 493. Magee, W. S. T., elected associate member, 114.

Mains, facts in relation to friction, waste, and loss of water in mains, 459. Mair, J., Wh. Sc, admitted student, 113. Maitland, General, his exi:)erieuce of the effect of working steel to varying

degrees, 154. — Ditto ditto oil-hardening, 161. Manganese, influence of, on the strength of steel, 115 et seq. Manning, 11., his observation that the flow of a river sometimes exceeds the

observed rainfall on its drainage area, 106.

INDEX. 523

Marcillac, P., danger-indicator for tlie prevention of collisions at sea, 506.

Markham, C, memoir of, 377.

Marks, E. C. R., admitted student, 113.

Marshall, F. C, apparatus for measuring the actual strains on a slide-valve, 181.

, F. C, and Weighton, R. L., Paper on high-speed engines, 182.

Marten, H. J. — Discussion 07i the Witham Outfall Improvement Works : Nature of the works and mode of execution, 97. — Influence of the new Witham outfall on the channel of the Welland, 98. — Works at the Grand Sluice, 98.

Martin, E. M., admitted student, 113.

, H. W., elected member, 113.

Martindale, W. B. H., elected associate member, 114.

Martin-Leake, S., elected associate member, 114.

Mason College, Birmingham, friction-brakes used at the, 59.

Materials, strength of, modes of testing the, 416.

Maud Foster Sluice, river Witham. See Sluice.

Maudslay Sons and Field, machinery of H.M.S. " Campcrdown," constructed by, 325 et seq.

Mercury-thermometer. See Thermometers.

Merewether, E. A. M., elected associate member, 114.

Meylan, E., Desrozier's new disk-dynamo, 488.

Miles, H. P., admitted student, 113.

Milues, G. P., elected associate member, 114.

Mines, deep, diminution of earth-temiJerature in, 471.

of Austria and Hungary, differences of level in the, 472.

Mittelhausen, C. J. A., admitted student, 113.

Molesworth's formula for earthwork, 229 et seq.

Montresor, C. E. C, elected associate member, 114.

Moore, C. J. A. P., admitted student, 113.

Morant, E. F., memoir of, 387.

Moriarty, A. D., admitted student, 113.

Mortars, hydraulic, yield of, 421.

Nansouty, Max de, special plant for blasting under water at the Panama Canal

works, 448. Napoleon I., Alpine roads constructed by, 238 et seq. Nevill, P., elected associate member, 114. Nisbet, T., elected associate member, 114. Norman, C. E., transferred member, 112.

Oakes, C. S., admitted student, 113.

Ocean Dock (Boston) scheme, the, 97.

O'Hara, J. G. M., admitted student, 113.

Oil-gas. See Gas.

Osborne, F. C, admitted student, 113.

Outfall channel (new) of the river Witham. Sec Eiver Witham.

Outram, F. D., admitted student, 113.

Owens College, Manchester, Froude turbine-dynamometers at the, 39 et seq.

Paper, on the testing of, 420.

Park, J. C. — Correspondence on the Friction of Locomotive Slide-Valves : Adoption of cast-iron instead of brass for slide-valves on the North Loudon railway, 194.

524 DIDEX.

Partridge, W. A. M., elected associate member, 114.

Passes, Alpine, 237 et seq.

Paterson, P. J., admitted student, 113.

Pazzani, J., memoir of, 379.

Pedder, D. P., admitted student, 113.

Peutland, A. T., elected associate member, 114.

Perceval, R. D., elected associate member, 114.

Permanent-way. " The Permanent-way of some Railways in Germany and in Austria-Hungary" translated and abstracted by W. B. Worthington (S.), 303. — Table of dimensions, &c., of rails, 303. — Alsace-Lorraine State Railways, 306. — Bavarian State railways, 310. — Saxon State railways, 312. — Rhenish Bavarian (Palatinate) railways, 313. — Hessian Ludwig's railway, 315. — Austrian State railways, 316. — Baden State railways, 318.— Austrian Southern railway, 320. — Hungarian State railways, 321.

Perry, Prof J. and Prof. W. E. Ayrton, experiments of, with transmission and absorption-dynamometers, 51.

, W. A., elected associate, 114.

Petroff, N., new theory of friction, 407.

Philip, A., analysis by, of tire-steel for the Indian Government Railways, 163.

, W. M., elected associate member, 114.

Philippart's, electrical tramcars in Paris, 492.

Phosphorus, influence of, on the strength of steel, 115.

Pichler, M. R. von, renewal of water in the Hague canals, 450.

Pinchin, R., memoir of, 388.

Piutsch's oil-gas as used on the Caledonian railway, 218 et seq.

Pilot, C. L. E., admitted student, 113.

Playfair, W., admitted student, 113.

Potable waters. See Waters.

Port, construction of a, at Prague, 447.

Portsmouth, J., admitted student, 113.

Powell, A., elected associate member, 114.

Prague, construction of a port at, 447.

Prescott, H. E., elected associate member, 114.

Presentation of medals and premiums, 1.

Pressure, earth-, underground, apparatus for measuring, 471.

Preverez. See Kerviler.

Price, J., jun., B.E., elected member, 113.

Prinetti, T., the embankment of the Po at Turin, 449.

Prinsep, R. S., elected associate member, 114.

Prismatic formula for earthwork. See Earthwork.

Prismoidal formula for earthwork. See Earthwork.

Prony brake, principles of the, 2 et seq.

Pulverization of clay. See Clay.

Pyrometer, trustworthy, want of a, 136.

Quarrying by wire, 424.

Rails. " On the Use of Heavier Rails for Safety and Economy in Railway Traffic," C. P. Sandberg, Abstract (S.), 354.— Goliath rail, 354.— Experiments at the Domnarfvet works, Sweden, 355. — English permanent- way, 356. — Supplement : Accident to the Czar's train at Borki, October 1888, 358.

INDEX. 525

Railton, A., admitted student, 113. Railway, Arlberg, 268.

, Austrian North Western, permanent-way of the, 303, 308.

, Austrian Southern, permanent-way of the, 304, 320.

, Brenner, 245.

, Caledonian, lighting of carriages by oil-gas, on the, 218 et seq.

, cable, the, on the New York and Brooklyn bridge, 453.

carriages, lighting of, by compressed gas, 218.

, Gien and Auxerre, consolidation of earthworks on the, 464.

, Great St. Bernard (proposed), 272.

, Hessian Ludwigs, ijermanent-way of the, 304, 315.

, London, Brighton and South Coast, reparation of Betchworth tunnel on

the, 291 et seq.

, Mont Blanc (proposed), 271.

, Mont Cenis, 255.

, , (Fell), 249.

, St. Gothard, 262. — Signalling-apparatus on the, 470.

, Semmering, 241.

, Simplon (proposed), 274.

, steep-gradient at Laon, 468.

structures, inspection and maintenance of, 432.

Railways, Alsace-Lorraine, permanent-way of the, 303, 306.

, Austrian State, permanent-way of the, 304, 316.

, Bavarian State, permanent-way of the, 304, 310.

, cost prices on, 469.

, Hungarian State, permanent-way of the, 304, 321.

, Rheinish Bavarian (Palatinate), permanent-way of the, 304, 313.

, Saxon State, permanent- way of the, 312.

Rainfall at the Grand Sluice, Boston, 90, 91.

in the Witham Drainage District, October 1883, October 1885, 87, 101.

Raising the steamer " Ferndale," sunk in the entrance channel of the port of St. Nazaire, 465.

Ranken, A. W., A.K.C., admitted 8tudent,'113.

Ransomes, Sims, & Jeflferies, Messrs., forms of friction-brake dynamometer used by, 10, 18, 21.

Raoult, C, the Beer system of wire ropeways, 485.

Reid, R. N. H., admitted student, 113.

Rendel, Sir A., K.C.I.E. — Correspondence on the Strength of Bessemer-steel Tires : Testing tires for the Indian State railways, 163. — Analysis of tire-steel for the same, 163.

, W. S., transferred member, 112.

Rendell, A. W., elected member, 113.

Rennie, Sir J., recommendations of, in respect of the river Witham, 81.

Reynolds, E. — Discussion on the Strength of Bessemer-Steel Tires: Hardness not the only quality essential for long wear, 131. — Steel tire made by Naylor Vickers & Co. for the Great Western Railway in 1868, 131. — " Body " iu steel, 131. — The so-called molecular change iu steel under certain conditions, 132. — Quality alone to be relied upon if high results were expected, 133.

Richards, R. W., elected associate member, 114.

Ricour, G., cost prices on railways, 469.

Rilev, Dr. Sec Kennedy.

526 INDEX.

Ringel, A., measurements of the flow of the Elbe in Saxony, 1886 and 1887, 444.

River Clyde. " The Eiver Clyde," D. Macalister (S.) (Abstract), 279.— Changes in the navigable and subsidiary channels from Dumbarton to Greenock between 1860 and 1880, 279.— Old channel, 280.— Volume of water passing Garvcl Point, 281. — Works necessary for maintaining the channel, 282.

Donge (Holland), affected by the Oude Maasje, 109.

Elbe in Saxony, measurements of the flow of the, in 1886 and 1887, 444.

Isar, regulation of the, according to Wolf's method, 445.

Moldau, improvement of the, at Prague, and the construction of a port

there, 447.

Nadi, aqueduct carrying the Lower Ganges Canal over the, 283.

Oude Maasje (Holland), influence of the, on the Donge, 109.

Po, embankment of the, at Turin, 449.

Shannon, flow of the, 106.

Welland, influence of the Witham new outfall on the, 95 et seq.

Witham. " The Witham New Outfall Channel and Improvement Works,"

J. E. Williams, 78. — Course of the river from its rise, 78. — Grand Sluice, 79. — Unfavourable nature of the river previous to 1878 for drainage and for navigation, 79. — Recommendations of the General Commissioners of the Witham in 1879, 81. — River Witham Outfall Improvement Act, 1880, 81. — Execution of the works : Excavation, 81. — Embankment, 82. — New channel, 83. — Enlargement of the Grand Sluice, 83. — New locks, 84. — Widening between Grand Sluice and Tattenhall Bridge, 85. — Cost of the works, 86. — Appendixes : I. Flood report, 1883, 87. — H. Flood report, 1885, 88. — in. Tidal curves at Boston before and after completion of the new channel, 89. — IV. Rainfall at the Grand Sluice, Boston, 90. — V. Greatest rainfall in one day at the Grand Sluice, Boston, 91. — Discussion: J. E. Wil- liams, 92, 104 ; J. C. Hawkshaw, 92 ; Sir C. A. Hartley, 93 ; L. F. Vernon- Harcourt, 94 ; A. Giles, M.P., 96 ; H. J. Marten, 97 ; W. Shelford, 99 ; J. G. Symons, 101; A. C. Hurtzig, 103; F. Wentworth-Sheilds, 103; Sir G. B. Bruce, 104. — Correspondence: P. Caland, 105; M.F.Fitzgerald, 106; C. Bloys van Treslong, 108.

Roads, Alpine, 237 et seq.

Roberts- Austen, Prof. W. C. — Discussion on the Strength of Bessemer- Steel Tires : Mode of existence of the carbon in hard and in soft steels, 136. — Influence of added impurity in metals probably governed by a law, 137.

Robins, W. H., elected associate member, 114.

Robinson, J. P., admitted student, 113.

Robson, R. O., elected member, 113.

Rochelle Harbour. See Harbour.

Rochfort, J., elected associate member, 114.

Rogers, R. B., M.A., elected associate member, 114.

, W., memoir of, 380.

Roome, G. W., admitted student, 113.

Ropeways, wire-, the Beer system of, 485.

Rose, F., jun., elected associate member, 114.

Ross, G. H., memoir of, 382.

Rouillard, A., on the measurement of the resistance of submarine cables, 490.

Rowlandson, C. A., elected member, 113.

Roy, N. W., elected associate member, 114.

Royal Agricultural Society, dynamometers used by the, 18 et seq.

INDEX. 527

Salt, decomposition of, by clectrolyBis, 504.

Sampson, J., elected associate member, 114.

Sandberg, C. P. " On the Use of Heavier Rails for Safety and Economy in Railway Traffic (S.) (Abstract), 354.

Scheurer-Kestner, — , heat of combustion of the coal of the North of France, 509.

Schiller, H., the poisonous action of water-gas, 508.

Schnabel, Dr., the smelting of gold and silver ores in Eastern Hungary and Transylvania, 480.

Schiinheyder, W. — Corresijondence on Friction-Brake Dynamometers : Type of friction-brake most suitable for general engine-testing, 76. — Alleged difficulty of adjusting brakes fitted with compensating-levers, 76. — Efficiency of the water-cooling arrangement for brake-wheels, 76.

Schwartz, A., admitted student, 113.

Search-light, electric, self-regulating, 497.

Seller, — , Wilmsmann's smoke-consuming furnace, 464.

Sewage-works, East Orange, the, 460.

Shaft-sinking by Haase's method, 475.

Shaw, J. J., elected associate member, 114.

Sheilds, F. Wentworth-. — Discussion on the Witham Outfall Improvement Works : Question as to the direction of the flood-tide in the Witham, 103.

Shelford, W. Discussion on the Witham Outfall Improveinent Works : Physical aspect of the Fen rivers, 99. — Omissions of tidal observations and tidal diagrams from the Papers of Mr. Williams and Mr. Wheeler, 99. — Keasons for adopting a slope of 4 to 1 in the new channel, 99. — The designer of the Grand Sluice, 100. — Effect of its construction, 100. — Question of self-maintenance of the new channel, 101.

Shenton, H. 0. H., admitted student, 113.

Ships of war. " The Speed-Trials of the latest additions to the Admiral Class of British War-Vessels," D. S. Capper (S.), 325.— H.M.SS. " Camperdown," and " Anson," 325. — Engines of the " Camperdown," by Messrs. Maudslay Sons and Field, 325. — Engines of the "Anson," by Messrs. Humphrys, Tennant and Co., 327. — Mode of conducting Admiralty official trials, 328. — Trial of the " Cam- perdown," 330. — Trial of the " Anson," 332. — Special conditions influencing the design of the machinery of war- vessels, 334. — Differences between the engines of the " Camperdown " and the " Anson," 337. — Appendix, Tables : I. Abstract of mean results obtained at trials of H.M.SS. "Anson" and "Camperdown" in 1887, 343. — II.-V. Half-hourly records of trials with open and closed stokeholds, 845. — VI-VII. Temperatures of engine-rooms and stokeholds during trials of the " Anson," 349. — VIII. Particulars of machinery, 357. — IX. Ditto, ditto, ratios and coefficients, 352. — X. Comparative table of weights of engines and boilers, 353. — XI. Data for twisting moments, 353.

Siemens and Halske's electric winding-engine at Neu Stassfurt, 500.

Signalling-apparatus on the St. Gothard railway, 470.

Silicon, influence of, on the strength of steel, 115 ei seq.

Sillcm, W., admitted student, 113.

Silver ores, the smelting of, in Eastern Hungary and Transylvania, 480.

Skelton, K., A.K.C., elected associate member, 114.

Sketchley, H. G., transferred member, 112.

Slide-valves. " The Friction of Locomotive Slide-Valves," J. A. F. Aspinall, 167. — No trustworthy data extant of the friction of slide-valves under steam, 1 67. — Mechanism devise<l to give a diagram of the exact force required to move the

528 INDEX.

valve at each point of the stroke, 168. — Table I. Experiments on the friction of the apparatus, 169. — Method of dealing with the diagrams obtained, 170. — Table II. Experiments with the valve pulling and pushing, 171. — Discussion of the results, 172. — Table III. Pressure on valve and valve-resistance at mid- stroke, 176. — Table IV. Percentage of power lost in friction of valves and eccentrics, 177. Discussion: J. A. F. Aspinall, 179, 191; W. Cross, 179; D. Halpin, 182; J. Goodman, 185; B. Tower, 188; W. Stroudley, 188; E. A. Cowper, 190 ; E. Woods, Past President, 191. — Correspondence: J. C. Park, 194.

Slow-cooling, influence of, on the strength of steel, 120.

Sluice, Grand, river Witham, enlargement of the, 79 et seq.

, Hobhole, river Witham, 79 et seq.

, Maud Foster, 79 et seq.

Smeaton, J., usually credited with the design of the Grand Sluice on the river Witham, 100, 105.

Smijth, E. C. B., transferred member, 112.

Smith, J. W., " HxirsVs Triangular Prismatic Formula for Earthwork compared with the Prismoidal Formula" (S.), 229.

, Prof. R. H. — Discussion on Friction- Brake Dynamometers : Description of

his transmission-dynamometer, 56. — Friction absorption-dynamometers, 57. — Type of brake used at Mason College, Birmingham, 59. — Deprez's brake, 60.

Soltz, A., electrolytic copper-refining in Hungary, 479.

Sorby, Dr. H. C. — Discussion on the Strength of Bessemer- Steel Tires: Microscopic structure of steel, 144. — Phenomenon of hardening as observed under the microscope, 145. — Necessity for prolonged individual study in regard to steel, 146.

Sowerby, W. — Correspondence on the Strength of Bessemer-Steel Tires : Shocks to tires in testing by the falling weight not of the same nature as those occurring in actual work, 164. — Parabolic gradients on railways, 165. — Use of manganese and of chromium by early Indian and Sjianish steel-makers, 165. — Mode of working steel for swords and guns in India, 165.

Speed-trials. — " The Speed-Trials of the latest additions to the Admiral Class of British War-Vessels." D. S. Capper (S.), 325.

Spencer, C. T., memoir of, 391.

Steam, elucidation of the disinfecting power of, 462.

superheated, on the disinfecting action of a current of, 461.

Steel. " On the Influence of Chemical Composition on the Strength of Bessemer- Steel Tires." J. O. Arnold, 115. — Importance of the subject, 115. — Tendency of railway engineers to specify high tensile steel for tires, 115. — Proper limits of carbon in Bessemer tire-steel, 116. — Eifect of the varying proportions of the different elements usually present in tire-steel, 117. — Effect of chromium, 119. — Influence of slow cooling on the arrangement of the molecules, 120. — Influence of the amount of work put on steel, 122. — Views of the engineer on the mode of testing as distinct from those of the manufacturer, 127. — How far the makers' tests indicate the fitness of tires for the work to which they will be subjected, 128. — Influence of hardening and of annealing, 129. — Dis- cussion: E. Reynolds, 131; R. A. Hadfield, 133; Prof. W. C. Roberts-Austen, 136; W. M.Williams, 137; E. A. Cowper, 139 ; J. A. F. Aspinall, 140; W. B. Lewis, 142; Dr. H. Clifton Sorby, 144; W. G. Kirkaldy, 146; G. Berkley, 148; T. E. Vickers, 152; W. Stroudley, 154; A. McDonnell, 155; J. O. Arnold, 155. Correspondence: C. J. Appleby, 160'; H. A. Brustlein, 162; J. W. King, 163; Sir A. Rendel, 163; W. Sowerby, 164; B. W. Winder, 165.

I

INDEX. 529

steel, Bessemer-, i^rocess, a new modificatiou of the, 477.

, highway bridges, 430.

Steep-gradient railway. See Kailway.

Stephens, F. C, elected member, 112.

, H. F., admitted student, 113.

Stephenson, E. P., elected associate member, 114.

Stewart, S., photometric tests of oil-gas, 227.

Stone-cutting and quarrying by wire, 424.

Stones, methods of testing the resistance of, 416.

Stoney, E. D., admitted student, 113.

Stothert, P. K., elected associate member, 114.

Strain. On the critical extension of bodies strained simultaneously in several directions, 410.

Streatfeild, H. S., admitted student, 113.

Strobel, C. L., experiments on a new form of strut, 428.

Stroudley, W. — Discussion on the Strength of Steel Tires: Experience on the Brighton Railway with tires of various compositions of steel, 154. — Discussion on the Friction of Locomotive Slide-Valves: Mode of lessening the friction by plugs of tin inserted in the face of the valve, 188. — Slide-valves of the B, C and D classes of engines on the Brighton railway,. 189. — Reasons for discarding balanced slide-valves, 189. — Experiments made under other than the working conditions not of much value, 190.

Structures, railway, inspection and maintenance of, 432.

Strut, new form of, experiments on a, 428.

Sulphur, influence of, on the strength of steel, 115.

Surveying. " Preliminary Survey in Neic Countries, as exemplified in the Survey of Windward Haicaii." T. Gr. Gribble (S.), 195. — Physical features of the island, 195. — Proposed railway to convey produce from the interior to the coast, 195. — Mode of surveying determined upon, 196. — Optical work at the gulches, 198. — Field-book, 199. — Appendixes, I. : List and description of instru- ments, 202. — II. Field-book, 204. — III. General principles of telemetry and telemeters, 207.

, '■'Rapid Surveying" F. D. Topham (S.), 209. — Route survey in

Asia Minor, 209. — Mode of working with the prismatic compass, 209.

The Practice of Surveying in the Australasian Colonies." S. K.

Vickery (S.), 211. — Organization of the several Government departments employing surveyors, 211. — Surveying in the colony of Victoria, 212. — Appen- dixes, I. : Examinationof surveyors, 215. — II. : Survey of block with check-line, 215.

Sweet, J. E. — Correspondence on Friction-Brake Dynamometers : Application of the platform weighing-machine for producing resistance, 77.

Sykcs, C. M., elected associate member, 114.

Symons, G. J. — Discussion on the Witham Outfall Improvement Worhs : Rainfall of September, 1883, in the Witham Drainage District, 101.

Syson, R. C, elected associate member, 114.

Tait, W. A. P., B.Sc, admitted student, 113. Taylor, F. M. S., admitted student, 113. Telegi-aph cables. See Cables. Telemeters. See Telemetry. Telemetry, general principles of, 207.

[the INST. C.E. VOL. XCV.] 2 M

530 INDEX.

Telephone equation, on the, 493.

Telephone line between Paris and Marseilles, 494.

Terre Xoire Company's, the, tests in 1878 of steels of varying chemical compo- sition, 134.

Testing of building materials. See Materials.

of paper. See Paper.

of steel. See Steel.

Thelwall, W. H. — " Hie Failure of the Kali Nodi Aqueduct on the Lower Ganges Canal" (S.), 283.

Theodolites, modem, 202.

Thermometers. Studies on the gas-thermometer, and comparison of the mercury thermometer therewith, 507.

Thomas, A. D., elected associate member, 114.

Thompson, C. W., admitted student, 113.

, J., admitted student. 113.

Thomson, A. W. See Alexander.

, Professor J., rope dynamometer, proposed of, 16 et seq. — Variation of

the arc of contact in friction-brakes, due to, 54.

Thurston, Professor K. H., water-cooled brake-dynamometer described by, 18, 20.

Tires. — " On the Influence of Chemical Composition on the Strength of Bessemer- Steel Tires" J. O. Arnold, 115.

Topham, F. 'D.—^'Bapid Surveying" (S.), 209.

Tower, B. — Discussion on the Friction of Locomotive Slide-Valves: Question as to the mode of a^iplying the weights in Mr. Aspiuall's experiments, 188. — Defect of the frictional indicators, 188. — Mr. Cross's diaphi-agm indicator, 188.

Tramcar, electrical, Philijipart's, in Paris, 492.

Trickett, J., memoir of, 392.

Tubbing, cast-iron, for lining levels, 476.

Tunnel, Arlberg, 269.

, Betchworth. " T/ie Beparation of Betch worth Tunnel, Dorking, on the

London, Brighton and South Coast Bailioay," G. Lopes (S.), 291. — Circum- stances attending the failure of the old tunnel, 291. — Works of reparation : timbering, 292. — New brickwork, 297. — Connecting the new work with the old, 300.— Cost of the work, 302.

, Great St. Bernard (proposed), 272.

, Mont Blanc (proposed), 271.

, Mont Cenis, 257.

, St. Gothard, 265.

-, Simplon (proposed), 274.

, Stuttgart, alignment of, 415.

Tumbull, N. K., Wh.Sc, admitted student, 113.

Vernon-Harcourt, L. F. — Discussion on the Witham Outfall Lnprovtment Worhs : Difficulties in connection with the removal of the Grand Sluice, 94. — Recom- mendation of Sir John Hawkshaw, in 1877, for dealing with the floods, 94. — Proposed enlargement of the Grand Sluice, 95. — Conflicting interests of the rivers Witham and Welland, 95. — Probable insufficient width of the new outfall channel, 96. — District above Lincoln not yet secured from floods in very wet winters, 96. — " Alpine Engineering " (S.), 237.

Viaduct, Garabit, the, 434.

IKDEX. 531

Vicars, J., B.E., admitted student, 113.

Vickers, T. E. — Discussion on the Strength of Bessemer-Steel Tires : Mr. Arnold's mode of experimenting varying from the received methods, 152. — Composition of a good steel tire, 153. — Alleged molecular changes in steel of high temper, 153. — Chromium as an alloy of steel, 154. — Reheating of tires, 154.

Vickery, S. K. — " The Practice of Surveying in the Australasian Colonies " (S.), 211.

Victoria, practice of surveying in, 211.

Wakeford, J., memoir of, 393.

Walker, A., elected associate member, 114.

, C. L., elected associate member, 114.

Walz, A., elucidation of the disinfecting power of steam, 462.

Ward, T. H., elected associate member, 114.

Waring, H. F., memoir of, 394.

Warner, W., elected associate member, 114.

Wash, the, and the Fen rivers, 78 et seq.

Water-cooled brake-dynamometers, 13 et seq.

Water-gas. See Gas.

, loss of, in mains, facts in relation to friction, waste, and, 459.

, renewal of, in the Hague canals, 450.

supply of the kingdom of Wurtemberg, 458.

Waters, potable, the qualities of, 457.

, waste of, in mains, facts in relation to, 459.

Wearing, W., elected associate member, 114.

Weber, L., on the connecting of lightning-conductors with gas- and water-pipes, 501.

Wehage, — , on the critical extension of bodies strained simultaneously in several directions, 410.

Weightman, W. J., elected associate member, 114.

Weightou, R. L. See Marshall.

Werner, E. A., the theory of jointed bow-girders, 426.

Wheeler, W. H., proposal of, for treating the Fen rivers, 82, 99.

Whitaker, J., Wh. Sc, admitted student, 113.

White, H. T., admitted student, 113.

Widmann, — , the alignment of a tunnel at Stuttgart, 415.

Willans, P. W. — Discussion on Friction-Brahe Dynamometers : Trials of an Appold dynamometer, 47. — Fronde's turbine dynamometer, 47. — Coope's brake, 49.

Willcox, B., elected associate, 114.

Williams, J. E. — " Tlie Witham New Outfall Channel and Improvement Woi-hs," 78. — Discussion on ditto : Effects of the new channel, 92. — Effect of concen- trating the scouring action of the tide on the new outfall channel, 104. — Scour of the Witham not always advantageous to the Welland, 104. — Provisions of the River Witham Improvement Act in regard to the Welland, 104. — Area benefited by the present works, 105. — Reasons for adopting a slope of 4 to 1 for the banks, 105. — The designer of the Grand Sluice, 105. — Mr. Symons's rainfall observations, 105. — Material of the hollow quoins of the lock, 105. — Disposal of the dredging, and cost, 105.

, W. M. — Discussion on the Strength of Bessemer-Steel Tires : Alleged

tendency of railway engineers to specify for high tensile steel tires, 137. —

532 IKDEX.

Col. Englisli's experiments at Sir John Brown's Atlas Works, 137. — Inability of steel to resist shock when under vibrating strain, 138. — Influence of man- ganese on steel, 138. — Admiralty specification of 1878 for ship steel, 189.

Wilmsmann's smoke-consuming furnace, 464.

Wilson, R., his balanced slide-valve, 190.

Wimshurst machine, the, 505.

Winder, B. W. — Correspondence on the Strength of Bessemer Steel Tires : Mode of occurrence of carbon in high-grade steels, 165. — Importance of co-operation of the diflerent persons concerned in producing and using steel if the question of molecular change was to be solved, 466.

Winding-engine, a, with spiral balance-drum, 484.

, electric, Siemens and Halske's, at Neu Stassfurt, 500.

Wire, stone-cutting and quarrying by, 424.

Wire-ropeways. See Eopeways.

Wise, B. D., elected member, 113.

WolFs method, regulation of the Isar according to, 445.

Woods, E., Past President. — Discussion on the Friction of Locomotive Slide-Valves : Mr. John Gray, of the Manchester and Liverpool railway, the first to use balanced slide-valves, 191.

Worthiugton, W. B. — " The Permanent-way of some Bailicays in Germany and in Austria-Hungary" 303.

Wrench, F. H., admitted student, 113.

Wyatt-Smith, A., admitted student, 113.

Yonge, M. E., admitted student, 113.

Zug, Switzerland, landslip at, July 5, 1887, 411.

i

lOKDON: PKINIED BT WILLIAM CLOWES AND SUNS, LIMITED, STAJIFOBD STREET AND CHAIUSG CROSS.

f

l^T.ATE .1.

lAT SAND ; GRANOV; S L "^ °'' ^^^ ^°^^ ON A.B

ITUDINAL

3= KELL 80 SON, IITH, 40 KING S- CCr»ffiMT GARDEN.

SECTION FROM GRAND SLUICE TO CLAYHOLE .

l<Iinui,-rE .A- Pro. sodiums of THs Tiwuiiit.ioii „CC.vil P.n^uioers Vol. XCV Kession. 1888^89. Part 1

PLATE . 2.

Fiq. 3

.LEN VALVE,

dings of The Insti

THO? KELL & SON, LITE. 40,10X0 ST COVENT GARDEN.

1F"1F1U'S''J^[1€)1?'J '-S'lF' lL'0(£(e)lj^'>S)YU'y E 5>lLlJl£!)t£- V-^iLV£S

Scaley, % Indi to I Fool

Fin. 4.

Scale. T'2 Lick to I Foot.

PLAIN VALVE .

ALLEN VALVE.

A :.kM'\Ut'\.\. IjKI,'

Mmutes of Pi-oceedings of The lustitunon of Civil Engineers. _ Vol. XCV. _ Session 1888-89 Ru-t I.

THO» KKLi. * SOS. UTH -taSTSLi S» >.VT«JH

F!',/S: 0'.

/■y

/'//■■

/ ^^'

AjP. nis aftj,-.

;)2 a . iSG-.^ai -niu;

End ,.r«,,m,llr iT,.l I/

:'cV VJ . / :w .^1,-1	- n ii:
KiuJ kI' .sftiiuUf	m.-s : II,

En/l iif HprndJU W>^ • Ij

rilli-'l Uk f>uU rii VuJA^e

i'ii 7 > / :%■ / .■767	:i 7 :i'i
^r!)l>. 2i;r.j	111(11 1
Hjul ill' Nj'itlttlr	s,-,-i c
	I'M 11 Ihi

PULLING END OF HYDRAULIC CYLINDER-

SLIDE VALVE DIAGRAMS FPOM AM EXPRESS CN«IMC WTtm M*€m

>iajle for Stftjjn I'ylinjljr ili/itfmmj$ tt Hrtile for Vah-e if

Hodo P,-et:«un 160 lbs

a I J I : ;'.v"7

Erit/ ot\'<fHitfllf

u::' a its Anmt .•« IoA»

A.K A£HI11AJ>7-,DELT

PUSHING END OF HVORAtttlC CrVl«»li

Mmutes of Pi-m-eearngs o:' The lustitunon ct' v'-.v-.'. Euj;"iii-

r«lVt DIAGRAMS FROM AN EXPRESS ENGINE WITH ALLEN VALVES.

/v//... /^^,

Srul ofA'pttiflle 254 ■ (>'

/-W7 I /U tlu7,«t on \UI\-.

Fu) : n.

48 5 - 1-16'Sei - eei^ t;e72. 28 z: = isse .

En J ufspmiUe 3)0 a

1S7S-6 Ibn thrjjitt on- Hih-r

■ I v.v

D ■••--.

71

/ : / V

■ '■': B Iht thnijtt mi Kih-e.

62 75 ■ 2H-27 - 1773 ,1

Eiui of spindle 2S4 t>

■Ul r,.-, 12(7. .7 CI - .^71; 7 ■717 7 • 2H27 - 1(702 !)

EnJ „rxpuitlU 2.'t4 (7

l/i4,1 H 11,.- Ihru^i „„

PUSHING END OF HYDRAULIC CYLINDER.

JijjtiUtoon of Cml En^neers Vol. Xcf . Session , 1 888 8.? J'ari/ 1

B,<il,;Prm«iirr li;(> lh.-<

X.Y. .SV,,,„, fhr.sl J);

^'°^- .78-8 ■ 121! ' .76V " 7» 7 7:>7-2H27 - 2253 1

K,„l „r.sp„„ll,-

APcJ^lis

-^2' sa «.;»»•-..-&/ - 7:ij

73 4 k2827 = 2t>7oO

End cfspmJU StO-ri

176*6 lis dtnut <•> lu^>

.V;'A

671-2827 - ..x.-'b.*

End ot'spuuile 310^

IS»i4 Mc Humat tm HAm

l:ll2 (7 ll^- tJtiiisI im I'li/lf

PUSHING END OF HYDRAULIC CYLINDER.

SLIDE VALVE DIAGRAMS FROM A GOODS ENGINE WITH ORDINARY SRAS& ^AL^ES .

A«/<. fl»

/>,/••>• /■?

A.RJMjLMi«

I'UJSII

Hcala for fiteam Cyijii/J/r- cUa^rornji u ScaZo for VaJyei Jt

r.^.

13'5 ■ 126^3&i — i-^-M 2233x2651 - e07 8

End afspinJU 190 I

IS :i > I2U - .7 67 :W fe'ff 2S(ili ■ i'S.^I To-'' r

/;■«// or.-riiuUe is:i I

:u:) I /A.v /«,// ,./' li/Ai

^0£/«r 3essure_m}_lb;,

■St,, Hit -r/„.v/ /)v»\s'»;> /;'/ //«

^° 'lH4o. I^e*5m - 4145 41 46 '2827 — 1171 i

End cfaputdL, 119 S

I P3ie u.

Sd2 0 Ihn thniat on Vklu

r.i i: , 1 -^df^Gl = 46-4Z I.-,.I2> 28-21 - 1284

And afapimUe ISB-3

Jtm 1 /i.s /J,n,.-i „„ i;,h:

eiexzs-zi - mtioi

End ofspind/i

250- n leeOo Uk.- Ihnt<st oil Fahr

- A y ASPISALL. DELT

Mmutes of ft-oceedings ot' The Insutunou a£ Oml Engineer. Vol. XCV S^ssue :f»*..S?

;i»»l«S rSOM * 5000S ENGINE WITH ORDINARY BRASS VALVES .

I'lqs: 15

Fifls: 16

Figs: 11 .

Bode,- Bvsxure 1401/,

-Af^

-4 :i,-, 12C f JJbl - 4S dH 4S89 , 26S1 - 1Z80-2

End of spindle 255-0

ir,y>o lbs-pull

£oiL;- B-cssiu-e Ufl lbs

<:a-95.l2C-t5ei " 3314 :i3-14x 28-27 — 2650 0

End atspindle 271 -2

2378 S //,.-■ lliniM ,m liilu

Sfnlrn -Chest F,

4Sti \I2H^5-Gt = eS I eS I . iG-Sl - IS05 3

End cfspuuUe- 262 9

2i>6S 2 Iks fiuU .»! fi/iv

ic IflTHnnoD of Ci»ii En^n.'iern . Vol. XCJ V. Seseaon

raO" KEU, * SOS, UTH •W-KTSa S' ,-UVEXT Iv^HMS

PLATE. 5

lAILWAY-CARf

JF'i

kf

ThFtniia/se Gnat^e-

Tb^SzrpinyGvtufC'

SECTION CD. IT ELEVATION. HALF BACK ELEVATION.

h.-^=4!^^^jj

1,1888-89. Part THD?-KEli & son, .TXTH. OO king- S- CCvTSKT GAPOiEN.

11,,,/.

''Ilk

I

SCH SYSTEM, AND ITS APPLI C AT I O N TO TH E LIGHTING OF RAILWAY - CAR R I AGE S .

C M EVBTEP.. DtL

MinuWB of ProosainlB of Ths Incuuilinn of CivU En|-moois Vol XCV Seeaion, 1888^89. Piun. 1 .

liSON.imi. «; los? 5« ccnnaiT GAK»

ALp]?]i i?]a]?]i2/j],i§.

. I '/EHBOB-HARCOUIIT. CELT.

Mmutes >,f ft>x«^l-.r

A L p J ?] £ 1 ?j a J ?] 1 :i /! J ?] a

l-!.,:i;.

,

ALT } i-i % :s ?j a ] ?] 1 1 ji] ?] a =

A L P !1 jM E 2 Kl 6 J M g E Ifi Q Kl 5

Ifloq ^jee^ «lm,vt' Spa y-vi'l.,

.-'liiif^'; of Tlw Inotitiition of Cr/il En^meers.Vol: XCV. Session 1888-89. Part I.

PLATE . 8 .

/

THOSKEU, 8cS0jr,LITH.40,mN-G- SP. COVENT GAHUEN.

)if'gSi0-iri«iii5!iiL!§ 101? isKDvasKi Wii^a- yggisiLs.

_ (i„A„r»j„z, „w.„z„;,„j p,.,.,r.

•^'...t.„/r ( 1 > t f

. S. CAPPER. DEL?

Minuton of Piocoedinga of Tile InntAuuon of CmLEngineeis. Vol. XCV. Se88ion"l688-89. Part I.

R A I L W A Y S .

i'linutes ot' Procfledings at The Lasutution c4" Civil iinguwars

« A I I. w A Y s . siVyll

o; Cr^l En^maors. Vol ZCV Skekjh 1S88-83 Part

TA Institution of Civil

1 Engineers, London

1^79 Minutes of proceedings

V.95

pt.l

Engin.

PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY

			(M
; —	O
:^	CO		. —
izz	G)	<	.;r-
—	(M	z	—
I		I
;rr —		o
::	00	o	T— -^
■^ —
'-Z—		in
•		o
E	N	•<*•
fe	C\J	O
h—		z
1=	(D	m
:n	CM	^
1 —		O	o —
^	LO		"T-
EE	(M	©
t—
	^
	CM
_	CO		^==

"CM

CM "CM

E CM

O "C\j

.0) 00

T-

CD

G)

00

N

rn.

cj 1	1 1 1 60INCH 2	TTTT 3	TTTT 4	1 1	1 1 5	TTTT 6	1 1 Ml 7
''- ''i\ ^D^,': ef^^^^^^^^^^^^^^^^^^^^H	^^^^^H
	^gmnji	^^^IH^^^^^^^I^^^^^^I

	— o -
::=. — 'Zi	3 i:^
— to		3=3
N3
—		■=\
— .		— i. ZZZ
■zzr"	i )	~^
—	<"> — 1
::=■ —		3 =3
—	l\j — .
^
at		00 — =f

r'.

IV)

00

-_^

ui — ^

0)-

-sl-

00 — ^

CD-

O"

cn

fNO"

00'

9

lO

fji

g^sii&iij^^i^Hia^