THE R_ OGERS LOCOMO0TIVIE ALND INIACHUNE. =TOURS. pa~terson AN. J. I....... I.....' ~~~~~HI l' An 1'.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'i (:& iliil~j "J's d~,Brazilian".~~~~~~~~ 77-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~: i.......'' n.. AMIERICAN AND EUROPEAN RAILWAYY IBPRA(CT]ACE IN THE ECONOMICAL GENERATION OF STEAM, INCLUDING THE MATERIALS AND CONSTRUCTION OF COAL-BURNING BOILERS, COMBUSTION, THE VARIABLE BLAST, VAPORIZATION, CIRCULATION, SUPERHEATING, SUPPLYING AND HEATING FEED-WATER, ETC., AND THE ADAPTATION OF WOOD AND COKE-BURNING ENGINES TO COAL-BURNING; AND IN:PERIMANENT WAY INCLUDING ROAD-BED, SLEEPERS, RAILS, JOINT FASTENINGS, STREET RAILWAYS, ETC., ETC. BY ALEXANDER L.:HOLLEY, B.P. WITH SEVENTY-SEVEN PLATES, ENGRAVED BY J. BIEN. NEW YORK: D. VAN NOSTRAND, No. 192 BROADW,AY. SAMPSON LOW, SON & CO. 1867. entered according to Act of Congress, in the year 1860, by D. VAN NOSTRAND, In the Clerk's Office of the District Court of the United States for the Southern District of New York. JOHN F. TROW, PRINTER AND ELECTROTYPEB, 50 Greene St., New York. I(/~~~~~~~~~~~~~I jj~~~~~~~~~~~~~~~~~~BRR Hawick-f ~~~~-1~~~- -. (li~~~ahl t Eng~~~~~~~~~~~~~~~~~~~~~' N AvI aiekF~l Camic]KfeT ston D~~~~~~~~~~~~~~~~~~~~~~~~~~~~i:i Sta i E RFJ YA S~~~~~~~~~~~~~~~~~~~~~~~' BihoAucd IVM TM RA~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i D; jk le al ii ii on a~~~~~~TiskOG P ~ ~~r-I~WCA s\~~~~~~~~~~~~~W /i Fo~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~s0iT LF~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~tr ~~~~~~~jri; ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~Stlod V cfl~~~~~~~~~~~~~~~~~~~~~adn Oh R C ON TE TS. THE ECONOMICAL GENERATION OF STEAM. PART 1. BOILER-MAKING. CHAPTER PAGE CHAPTEB PAGE I. —INTRODUCTORY..... 15 5. Caulking.... 38 6. Furrows adjacent to Joints... 38 II.-~MATERIALS..... 17 7. Effects of Punching... 39 2. Use of Iron for. ire-pltes.. 8. Conical and Counter-sunk Rivet-holes. 39. of Iron for Fire-plates.. 19. Hammer, Machine.and Snap-riveting..39 3. Defectsof Iron forFire-plates.-Remedies. 18 Pi R Q9 10. Pitch of Ri vets.... 39 4. Advantages of Iron forFire-plates.. 19. Thick-edged Plates.... 40 5. Copper Fire-plates. 20 11. Thick-edged Plates. 40 5.JopperFire-plates....^ 2012. Welt and Hoop-joints... 40 6. Iron and Copper Combined. 20 13.Welding Plate.... 41 7. Conclusion as to Iron and Copper. 21 14. langing Plates.-Expansion Joints 8. Influence of Good Circulation on the Du- Firebo x Seams.... 43 rability of Fire-plates. 21. 15. Angle-iron Joints.... 4 15. Angle-iron Joints.... 44 90.tee... Steel.-S'.n.e's.urnaceE n-'16. Combustion-chambers and Mid-feathers. 44 10. Puddled Steel.-Spence's Furnace —Eng- 17. Steam-domes..... 46 lish Processes. 22.lish Proceansses.... 22-stey Iron-wSrks'18. Strength and Construction of Stayed Sur11. American Semi-steel.-Albany Iron-works' faces..... 46 Process. 25 19. Attachments to Boilers.... 54 12. Homogeneous Steel... 25 13. Bessemer Steel..... 13' Be's'emer Stee21. Testing Boilers, and the Safety-valve.. 57 14. Advantages of Steel... 26 15. Conclusion as to Materials... 29 23. General Conclusions as to Boiler-joints and ii.-~DIMENSIONS OF FIRE PLATES... 30 Construction.... 59 1. Thickness..... 30 V. —FLUES AND FLUE-SETTING.-WATER-TUBES. 61 2. Advantages of Thin Plates.. 30 1. Diversity of Materials employed. 61' 3. Staying Thin Fire-box Plates... 31 2. Conditions governing a choice of Material. 61 4. Corrugated, Plates................ 32 3. Defects and Remedies. —Materials con5. Area of Plates. 33 sidered.. 62 IV.-BoILER JOINTS AND CONSTRUCTION.. 34 4. Conclusion. —Advantages of Steel.. 64 1. Strength of Welt-joints.. 34 5. Flue-setting.... 65 2. Strength of Riveted and WVelded Joints. 35 6. Conclusions as to Flue-setting.. 68 3. General Inferences.. 37 7. Water-tubes.... 69 4. Grip of Rivet-heads... 38 PART II. COMBUSTION. I. —INTRODUCTORY.... 71 Proportions of Air and Gas necessary 1. The Economy of Coal as Fuel.. 71 to Combustion. 2. How to Estimate Economical Coal- Intimatemixture of Air and Gas necesburning.....73 sary to Combustion. 3. Adaptation of the present class of Engines Large surface contact of air and gas. to Coal-burning.... 74 Time. 4. Popular Errors with reference to Coal- Space. burning..... 74 The forcible mixture of air and gas.: II, T}E COBUSTION OF BITUMINOUS COAL. 77 An excess of air practically necessary.. Ge neral Principles.. The ignition of the mixture and the 1. GeneralPinciples....7X.^ 7 n T^.-f?~~ ^. maintenance of combustion. 2, Phenomena of Coal-burning... 77. ao_.~ i'~ * ~ * ~ n~ -~rn~ Intensity of combustion. 3. Experimental Examination of these Phe- t. om.. Heated air. nomlena. 7.... i- c8 4. Explanation of thlese phenomena.. 79. ecapitulation.... Chemistry of Coal-burning. Composition of Coal. III.-MEANS OF ArPLrI:Q TIE PRINCIPLES OF The Operation of Burning. o' COMBvSTION..... 91 Combustion and Explosion. Introductory. 44^ CONTENTS. OJAPTES PAGE CHAPTER PAGE 1. Mixture and Ignition of Air with Gas by By external steam pressure. the aid of the Draft. 91 Advantages of Clark's jet. Simple admission of Air above the Perfection of coal-burning. Fire; Combustion-chambers. 3. Grates.... Deflectors.-~Fire-bricks. Water-grates. Ignition by Incandescent Fuel. Intensifying Combustion, Small Grates, etc. By admitting air through the fire. Size of Openings. By jets of steam. Movable Grates. The downward draft. 4. Variable Draft... 100 Conclusion. 1. The Variable Blast. 2. The forcible mixture of Air with Gas at 2. M3ovable Chimneys. the surface of the fire. 103 5. The Steam Jet... 113 By the aid of the Draft, 6. Conclusions. 113 PART III. HEAT, WATER, AND STEAM. I-TIIE TRANSM3ISSION OF HEAT... 115 4. Conclusions as to independent feed.. 126 1. Experiments and Practice. 115 5. Feed-heaters.... 127 2. Movement of Air and Gases.. 115 3. Smoke-box arrangements for equalizing III. —VAPOIZATION.... 131 the draft.....116 1. Maximum rate of vaporization.. 131 Petticoat Pipes. 2. Water-spaces and circulation... 132 Diaphragms, 3. Movement of steam and water.. 132 Flues of different sizes. Circulating passages. 4. Current disturbers in flues.. 117 Arrangement of flues. 5. Smoke-box, Chimney, and Blast-pipe. 118 Rate of draft through flues. II.-SUPPLYING AND HEATING PEED-WATER.. 120 IV. —SEPARATION AND SUPEIIIEATING... 136 1. Importance of an Independent Feed. 120 1. Separation.... 136 2. The Giffard Injector.. 121 2. Superheated Steam.. 136 3. The Donkey Pump... 125 3. Conclusion... 141 PERMAN ENT WAY. CIHAPTER PAGE CIIAPTER' PAGE l.-INTRODUCTORY..... 145 Defects of the cross-sleeper system. II.-EARTHWORK. DR.AINAGE AND BALLAST.. 147 Formation level. Cast-iron Sleepers.. 163 a lopes| Longitudinal System of the Great WestRetaining works in cuttings. ern Railway. 165 Formation of Earthwork. The Sandwish system.. 166 Drainage.... 1i50 Dimpfel's longitudinal system.. 169 Drainage of Road-bed. Seaton's longitudinal system. 170 Drainage of Slopes. IV.-RAILS..... 172 Conclusion as to Earthwork and Iron and Manufacture... 172 Drainage. Weight of Rail... 173 Ballast.... 14 Form of Rails.... 174 Materials. Re-rolling Rails..... 7 Quantities. The Phoenix Iron Co's Process.. 178 Laying B3allast. The Tubular Rail... 178 Drainage by Ballast. The Continuous Rail... 179 Cost of Ballast.. Hard Rails.....1 0 Conclusion as to Ballast. Conclusion...... SO Elasticity of Permanent Way. 157 V.-RAIL JOINTS..... III. —SLEEPERS.... 160 teference to the Consideration of RailMaterials and Arrangement... 160 joints in the foregoing Chapters.. 181 Quality of Timber. Classification of joint fixtures... 1S' Preservation of Timber. The Fish JoinLt.... 1 Size and distribution of Sleepers.. 161 Conditions of successful joint-fastcening. The Longitudinal System... 1GS62 Conclusion. S6. APPENDIX. Street Railways..... 87 strength is o30 tons or G7,200 lbs. per sq, inch Adaptation of Machinery. —The Bissell Truck — of section..... 1 Elastic Wheels..... 189 Results of Experiments on the Comparative Tensile Wrood's Switch...'.. 189 Strength, &c, of various kinds of Steel and Dick's Frog..... 189 Iron Plates, by Miessrs. Robert Napier & Sons. 191 Burleigh's Switch..'... 190 Table of the principal dimensions of various standTable of the working pressure of Locomlotive Boil- ard English and American Locomotives. 192 ers made of plates' whose ultimnate tensile LIST OF PLATE S. 1. COAL-BURNING FREIGHT ENGINE. ROGERS L. & M. WORKS (Frontispiece). 2. RAILWAY MAP OF GREAT BRITAIN (Back of Title-page). 3. COAL-BURNING PASSENGER AND FREIGHT ENGINES. M. W. BALDWIN & CO. 3 figUreS (Facing Part 1.) Fig. 1. Passenger Engine, 4 drivers and truck. 2. Freight Engine, 6 drivers and truck. 3. Freight Engine, 4 drivers and 4-wheeled truck connected. See plate 51. 4. COAL-BURNING PASSENGER ENGINE. LONDON AND SOUTH-WESTERN RAILWAY (Facing Permanent Wayv). 5. W. B. ADAMS' SANDWICHI RAIL. 5 figures. Fig. 1. Cross section, 60-lb. rail,. full size. 2. Plan as applied with susperleld fish-joints,... " 3. Cross section, " "..... i " 4. Pla.n of track, ".... inch to I ft. 5. Plan of " stringers, breaking joints, and fish inclosed,.. " " 6. STREET RAILS. 17 figures. Fig. 1. Trenton Iron Co.'s Girder, 8S lbs., 7 inches high,...' size. 2 and 3. Bartholomew's reversible,....... i 4. Philadelphia 7-inch tram. 55 lbs.,.. -... ll 5. " 5-inch " 45".... 6. " 43-inch " 43.... 7. Pittsburg Locomotive, 81"... " 8. New York City (old), 63"......" 9. Cincinnati Locomotive, 68~ ".... 10. Boston " 45 ".... "I 11. Seaton's, by Trenton Iron Co. 45"... ~. 12. New Yorkl City (new), 5 6"...... full" 13 and 14. Montgomery's coping,........... - 15. Phoenix Iron Co.'s reversible tram girder, 75 lbs.,... - " 16. " " " compound " " 101"...' " 17. Reese's reversible. 7. LONGITUDINAL SYSTEM OF PERMIANENT WAY, GREAT WESTERN RAILWAY, EiNGLAND. 8 figures. Fig. 1. Cross section of double track,........ inch to 1 ft. 2. Plan of "... " " 3 and 4. Joint chair for double-headed rail,...... s ize. 5. Cross section of longitudinal timber, wide-footed rail and fish-joint, " 6. " " " with J_ rail,... 7 and 8. Bottom piece, ".. ~ i 6 LIST OF PLATES. S. IMPROVED AMERICAN RAILS AND JOINTS. 5 figures. Fig. 1. Reading rail and fish-joint. Fish 30 inches long, 4 bolts; nuts and keys, full size. 2. Southern Pacific Rail, 57 lbs.,........ " 3. Cleveland, Columbus, and Cincinnati Rail, 60 Ilbs.,... 3. Memphis & Ohio Rail, 68 lbs.,....... " 5. Pile of Memphis & Ohio Rail,....... 9. Dr. BOUCHERIE'S PLAN OF PRESERVING TIMBER. 4 figures. Sulphate of copper injected by hydrostatic pressure. Fig 1. Packing the " saw calf." 2. Inserting the rubber tube. 3. Packing the end of the stick for injection. 4. The process of injection. 10. APPARATUS FOR CREOSOTING SLEEPERS. 2 figures. Fig. 1. End elevation of Creosoting cylinder, engine, etc.,... *-inch to 1 fto 2. Plan of' " " " "... " 11. STANDARD EUROPEAN RAILS. 3 figures. Fig. 1. Avignon and Marseilles, 66 lbs.,. full size. 2. Leicester & Hitchin, with standard fish-joints, 92 lbs.,. " 3. London & North-western, 85 lbs.,.... " 12. OLD AMERICAN RAILS. 6 figures. Fig. 1. Haggins' 60 lbs. New England Roads, 20 years' service,. full size. 2. Western Railway (Mass) 1852, 60 lbs.,..... 3. Ebbw Vale, New England Roads, 20 years' service,.. 4. Old Camden & Amboy, ".... i 5. Boston & Lowell, 62 lbs., 2- years' service, bottom upwards,. " 6. Old Reading, 45 lbs. 20 years' service,..... " 13. L.B. TYNG'S WROUGHT-IRON SUSPENSION RAIL JOINT. 3 figures. Fig. 1. Cross section,....... full size. 2. Side elevation,.......... ~ " 3. Plan,........... 1i4. HIGH FLAT-FOOTED RAILS. 3 figures. Fig. 1. Northern and Western of France, Swiss and other lines, 74 lbs.,. full size. 2. Belvidere Delaware, 72 lbs.,......; 3. Austrian-common form, 62 lbs.,...... 15. IMPROVED AMERICAN RAILS. 3 figures. Fig. 1. 58 lbs., double-head, to be used without chairs and with bracket-joints, full size. 2. 58 Ibs. rail, designed by A. L. Holley, with Adams' bracket-joint, " 3. 58 lbs. Cleveland & Erie, for wood splices,..... " 15-. STEPHENS' TUBULAR RAIL; as made at the Crescent Iron Works, Wheeling, Va. 17 figures. Figs. I to 16. Various stages of the iron in the process of manufacture, from the pile (Fig. 1) to the finished rail,..... -- size. 17. 58 Ibs. tubular rail,......... full" 16. MISCELLANEOUS RAILS. 3 figures. Fig.:-1. Adams' Suspended Girder, 84 lbs., 7 inches high,.... full size. 2. Buffalo, Corning & New York R. R. Rail, 84 lbs., 4} inches high, selected as the worst example of form in American rails,. 3. Pacific R. R. of Missouri. New pattern, 50 lbs.,.... " 16-. THE PHOENIX IRON Co.'s PERMANENT WAY AND PROCESS. 3 figures. Fig. 1. Pile for making rail fig. 3,....... full size. 2. Chair and brackets as proposed for the Penn. Central Railway,. 3. Rail, " " a ". full " LIST OF PLATES. 7 17. ENGLISH RAIL JOINTS. 12 figures. Figs. 1 and 2. Adams & Richardson's fish-joints (Standard),... size. 3 and 4. Barlow & Woodhouse's fish and single-jaw chair,... 5 and 6. Samuel's fish-joint chair,...... - 7, 8 and 9. Adams' cast-iron bracket-joint,...... 10 and 11. Barlow's joint-chair,......... l 12. Wild's grooved fish........., 18. ENGLISH RAIL JOINTS. 8 figures. Fig. 1. Fowler's joint-chair,........ 2. Modification of Parsons' wedge-chair,.... i 3. Sinclair's suspending chair, Eastern Counties,...' 4. Adams' intermediate wooden bracket,......i 5. Barningham's vice-joint,....... 6 and 7. Parsons' wedge-chair. 8. WTedge of Parsons' chair. 19. ENGLISH RAILS AND RAIL JOINTS. 15 figures. Fig. 1. Adams' suspended girder; wrought-iron way, continuous brackets. London & N. Western,........ 2. Adams' suspended girder,..... 3. " " " cast-iron bracket sleepers, 2' ft. long, 2 ft. apart,. ~ c apart,.......... l 4. Adams' bridge-rail joint,........ I 5 and 8. " bracket-joint, wrought-iron, doubled-headed rail,.. 6. " " cast-iron; flat-footed rail,... i 7. " nut-fastenings for fish-joint bolts,..... 9. " " " ".. 10, 13 and 14. Permanent Way Co.'s cast-iron way with wooden cushions. 12. " " ~ i u. i 11. Barlow's Rail,......... 15. Ashcroft's cast-iron way, wooden cushions. 20. ENGLISH RAIL JOINTS. 17 figures. Figs. 1 to 4. Samuel's double-fish,..... 5 and 6. Barlow & Woodhouse's double-joint chair,.... 7. Adams' elastic chair,......... i 8 and 9. Samuel's cast fish-chair,........ 10 to 13. Sinclair's fluted spikes,......... 14 and 15. Tyler's key for fish-joint nuts,....... " 16. Richardson and Billup's right and left screw for fish-jointing,.. i " 17. Pole's tapped bolt, " " ~ 21. ENGLISH RAIL JOINTS. 12 figures. Fig. 1. Adams' drain rail, 2 to 5. " joint made of old rail, with gib and key,..... 6. " bridge-rail joint,........ i " 7 and 8. Modifications of Adams' suspended girder in wood. See also plates 5, 32 and 33,.......... 9. Ramsbottom's self-wedging joint chair,..... i " 10 and 11. De Bergues' joint chair,........ " 12. Barlow & Woodhouse's joint chair,...... " 21-. AMERICAN RAIL JOINTS. 12 figures. Fig. 1. Steele's combined joint, Reading R. P. Bottom plate, 33 inches; wood fish 36 inches, 2 bolts, iron fish 17 inches,.... 2. Cleveland, Columnbus & Cincinnati R. R., Albany Iron Works chair (Fig. 7, Plate 26) and single fish,...... " 3. Cleveland C. & C. R. R. chair and double fish,.... i " 4 and 5. Krausch's bottom-wedge joint, N. Y. & Erie R. R.,... i " 6 and 7. Howe's elastic splice (erroneously printed Payne's in the plate), New England and other railways,....... i " 8 LIST OF PLATES. 8. Baker's joint for battered rails, Illinois Central R. R.,.. i size. 9. " " " " " " (side view),. 10. Beers' wedge fish,......... 11. Modification of the Trimble wooden splice, Rock Island R. R.,. i " 12. Morley's chair, WVestern railways,..... i' 2:2. ENGLISH RAIL JOINTS. 4 figures. Figs. 1, 2 and 3. Burleigh's cast-iron key joint,....... " 4. Barlow's double screw-wedge joint,...... " 23. ENGLISH RAIL JOINTS AND WAY. 26 figures. Figs. 1 to 7. Dickson's permanent way. 8, 9 and 10. Bow's chairs and fastenings. 11 and 12. Permanent Way Co.'s bracket joint. 13, 14 and 15. Ashcroft's wooden wedge chair.:16. Permanent Way Co.'s spikes, 9 oz., 9 - oz., 8`- oz. and 9 oz. respectively, ~ c 17. Nickless' joint,........ 18. Modification of the ring-joint, with fishes,..... 19. Gregory's joint,....... 6 20. Rose's chair. 21. Turnbull's permanent way. 22. Compound rail. 23. Saxby's joint. 24, 25 and 26. Seaton's permanent way, London c& N. Western Railway,.. 23-. EUROPEAN RAIL JOINTS AND WAY. 20 figures. Figs. I and 2. Permanent way common on German lines,..... 3. Fish joint. German lines,...... 4. Middleton & Stent's dovetail chair,... 6 5. Ratchet wedge-joint,.... " 6. Tyler's locking plate,...... " 7. Triangular wedge to fasten fish-bolt nuts,..... " S. Strap fish-joint. 9 to 15. Adams' permanent way. 16, 17, 19 and 20. Samn.uels' wrought-iron permanent way. 18. Taylor, Worswick & Lovatt's wedge and fish-joint,... " 24:. Pr(ENIX IRON Co.'S AMACHINE rFO TESTING THE STRENGTII OF RAILS AND GIRDERS. 6 figures. Fig. 1. Side elevation,. inch to 1 foot. 2. End " showing the steelyard, " 3. The carriage holding-the girder, " 4. End rest for girder, 5. Diagram showing angle of deflection, 6. Scale for measuring deflection,...... size..$5 BURLEIG-H HIOLTOWT KV EY. 3 figures. Fig. 1. With Harleml rail, endc elevation,..... full " 2. Planll of chair,....-.... 3. Side elevation of key,...... 2G. PERIANENT WAY. M[ISCELLANEO US, 19 figures. Fig. 1. Winslowv's continuous rail, GO lbs.,..... full " 2. The Vibbard "........ 39... Old N. Y. Central "...... " 4. Liernur's "....... " 5. Phoenix Iron Go.'s continuous double-lipped wvrought chair. 6. N. Y. IR. i1. Chair Co.'s wrought chair, 7. Corning, Winslow & Co.'s continuous-lipped chair. 8. Flanders' vice-anvil -for mending rails. 9. Clochrane's proce.ss of compressing the heads of rails,.. i 10. Pile for making rails of Royal Swedish Railways. LIST OF PLATES. Pig. 11. Bagnall'a improved method of piling iron. 12. Adam^' horse-foot tire,.......' size. 13. Werr of tire without the horse-foot,...... " t 14. " with (.... ",, 15. Sandwich joint (suggested by MIr. Colburn),... " 16 and 17. " "....... 18. Hyllier's chair, N. Y. & Erie R.R.,...... - 19. Modification of Krausch's chair, N. Y. & Erie R.R',.. - " 27. DITAIT4. O0 FISH JOINTS, ROYAL SWEDISH RAILWAYS. 4 figures. Fig. 1. Side elevation,....... 2. Sectional plan,......... " 3 and 4. Elevation of Fishes. 28. STANDARD ENGLISI CHAIR. RANSOME & MAY'S PATTERN, G-T. NOTI-IERN Ry. 3 figures. Fig. 1. End elevation, Joint-chair 13x6 in. 39 lbs. 9 in. wooden key,. full size. 2. Plan "... 3. Plan of Intermediate chair 4. " 29. A. L. IHOLLEY'S PERMANENT-WAY AND JOINTS. 15 figures. Fig. 1. Cast bracket and wrought chair,..... ". 2. Penn. Central lRail; cast bracket, wroughbt chair,.. 3. t " plan.,. 4. Wrought braclets. Foot of each rail turned up for 7 in.,. o ~ " 5. Side elevation of Fig. 6,..... 6 6. 7-inch girder, suspended on wood stringers, 5x "X5-"X' ft.. 4 7. Adams' girder without bolts and nuts,..' 8. Side-elevation of hook B (Fig. 7)...... 9. Suspended-girder; wrought brackets, cast-iron sleeper,... 10. " " side-elevation,. " 11. Cleveland C. & Cincinnati Rail. Hook A welded on,... 12. Wrought-iron bracket, and wood-splice,.... 13. " " side-elevation. - 14. Suspended girder, wrought fish and brackets with key, 4 ( 15.; rr"'" plan,... 30. CAST-IRON SLEEPERS-JOINTS. 26 figures. Figs. 1 to 9. Dr. Bergue's cast-iron sleepers (16X20 inches in plan). 10 to 12. Per. Way Co.'s " Samuel's pla, timber-bedded,.. -' 13 to 16. c" " corrugated, timber-bedded,.... s 17 to 19. " " wood cushions ill pockets,... " 20 to 26. Thoimas Wright's Permanent Way. 31. SYSTEMS OF PERMANEN-T Ai;Y. Showing relative depths of support. 16 figures. 2 in. to I ft, Fig. 1. Old stone block system. 2. Late Great Western superstructure. Sill 10x10 inches. 3. Plan once used on South-eastern line. 4. Common cross-sleeper chair and wood kley. 5. Greaves' spheroidal cast-iron sleeper. 6. Former Great Western system. Sill 7X14 inches. 7. A cast-iron sleeper, by P. W. BarloW. 8. Cross-sleeper with Adams' bracklet-joint. 9. Reynolds' longitudinal system. 10. P. W. Barlow's cast-iron sleepers. 11. D. Bergue's sleepers. 12. Spencer's corrugatedl plate-bearing. 13. Burleighi's cast-iron sleeper and cast krey. 14.'W. H. Barlow's saddle-back rail. 15. McDonnell's rail. 16. W. B. Adams' suspended girder rail. 32. W. B. ADAMS' SUSPENDED GIRDER RAIL; See also plates 5, 21 and 33. 5 figures. Fig. 1. 8-inch girder, wooden bolt and keys, section througl fish plate,. - size. 10 LIST OF PLATES. Fig. 2. 8-inch girder, wooden bolt and keys. Plan,.... inch to 1 ft. 3. " " section,... e size. 4 and 5. " " fiat and round bolts,.. 33. W. B. ADAMS' SUSPENDED GIRDER. 3 figures. Fig. 1. Wooden web or keel for light rail. Continuous plank sleeper,.' 2. " " " plan,.. 3. Iron " "..... 34. BURLEIGH'S SWITCH (dimensions on engraving), 3 figures. Fig. 1. Inside elevation. 2. Plan of switch, open. 3. " (other side) shut. 35. BURLEIGH'S SWITCH (dimensions on engraving) 5 figures. Fig. 1. Section at A B. Fig. 34. 2.' of tongue-rail showing the bearing of the flange. 3. " at E F. Fig. 34. 4. Plan of hinge at X-X. 5. Section at A B, of hinge chair. 352. FROG AND SWITCH. 4 figures. Fig. 1. Dick's frog or crossing. Perspective. 2. " " Plan,.... 3. Wood's safety switch. 4. 36. AMERICAN RAIL-JOINTS AND WAY. IS figures. Figs. 1 and 2. Hilton's splice, N. Y. Central R. R.,...... i 3. " side elevation,... 4. The Trimble Joint-wood splice. 5. Davis's Elastic-chair-India-rubber cushion inch thick under the plate B. 6. and 7. Potters' wedge chairs,..... i 8. " " Details. 9 and 10. Dunning's Joint chair,...... i 11 and 12. Bayley's Z Rail with Adams' bracket-joint. 13. Dimpfel's Sandwich Permanent Way,.... i 14. Stephens' tubular rail and joint,...... 15 and 16. Bishop's joint-fastening. 17. M. Fisher's i' 18. Graham's joint chair,....... i 37. COLORED DIAGRAMS ILLUSTRATING COMBUSTION. 8 figuires. Figs 1, 2 and 3. Effects of air admission on the combustion of burning-gas. 4. Section of a candle-flame. 5. The combustion of Carburetted Hyldrogen. 6. Effects of exclusion of air from the fire-box. 7. " admnission of air in bulk to " 8. " " in fine streamns. 38. BEATTIE'S COAL-BURNING LOCOMOTIVE BOILER, LONDON & S. WESTERN RAILWAY; AND BEATTIE'S FEED-HEATER. 2 figures. Fig. 1. Cross-section through middle of combustion-chamber,.. in. to 1 ft. 2. Longitudinal vertical section, showing heater,.... " 39. BEATTIE'S COAL-BURNING BOILER, LONDON & S. WESTERN RAILWAY. DETAILS. 5 figures. Fig. 1. Fire-brick grate, I (Plate 38),....... size. 2. "' flue F (Plate 38),...... * 3. " " section,...... " 4. Cross-section through fire-box,....... inch to 1 ft. 5. Back end elevation of "...... " LIST OF PLATES. 11 39- BEATTIE'S NEW COAL-BURNING BOILER, LONDON & S. W. RAILWAY. 4 figures. Fig. 1. Cross vertical section through middle of fire-box,.. ~ inch to 1 ft. 2. Longitudinal vertical section (brick flues J in elevation),.. 3. Cross vertical section through brick flues,..... 4. Plan of crown-plate of combustion-chamber. Plan of fire-box under crown-plate,..... 40. CUDwORTH'S COAL-BUINING BOILER, SOUTII EASTERN RAILWAY. Longitudinal Vertical Section. 1 figure. 41. MCCONNELL'S COAL-BURNING ENGINE, LONDON & NORTH WESTERN RAILWAY. Longitudinal Vertical Section. 1 figure. 42. McCoNNELL's COAL-BURNING BOILER, L. &. N. W. RAILWAY. 4 figures. Fig 1. Half plan of combustion-chamber. Half plan of crown-plate,.. ~ inch to I ft. 2. Vertical section of shell,..... 4. Back end elevation,..... 4. Cross vertical section of fire-box,.. ~. 43. ENGLISH COAL-BURNING BOILERS. 2 figures. Fig. 1. Craig's; Manchester, Sheffield & Lincolnshire Railway; longitudinal vertical section of fire-box,..... 4 2. John Dewrance's coal-burning boiler, patented Oct. 1846,. 44~ McCoNNELL's COAL-BURNING BOILER, L. & N. W. RAILWAY. I figure. Longitudinal vertical section; see also plate 42,.... 441. S. H. HIIEAD'S COAL-BURNING BOILER. FITCHBURG RAILWAY. 4 figures. Fig. 1. Transverse section of fire-box,...... * size. 2. Longitudinal vertical section of fire-box,. 3. Back end elevation ". 4. Cross vertical section ".. " 45. EATON'S WOOD OR COAL-BURNING BOILER AND MARKS' SMOKE-STACK. 2 figures. Fig. 1. Longitudinal vertical section of boiler and smoke-stack,. inch to I ft. 2. Half vertical cross section through fire-box. Half back end elevation, " 451. EATON'S COAL-BURNING BOILER, SUPERHEATING, AND FEED-HEATING, GREAT WESTERN RAILWAY, OF CANADA. 3 figures. Fig. 3. Side elevation of boiler. Section of smoke-box showing superheater,' inch to 1 fto 4. Vertical cross section of smoke-box and chimney,... " 7. Section of air-tube to fire-box,....... - size. 46. ROGERS LOCOMOTIVE AND MACHINE WORKS' COAL-BURNING BOILER AND HUDSON'S FEEDHEATER AND DIAPHRAGM. 4 figures. Fig 1. Longitudinal vertical section of boiler and heater,.... inch to I ft. 2. Half vertical cross section of smoke-arch. Half end elevation of smokearch,.......... 3. Half vertical cross section of fire-box. Half vertical cross section of heater,........... 4. Perspective view of Hudson & Allen's grate,... 47. THE PHLEGER COAL-BURNING BOILER. Long. vertical section, ~ inch to 1 foot. 1 figure. 48. SEPTIMUS NORRIS' COAL-BURNING BOILER, AS BUILT BY RICHARD NORRI'IS & SON, PHILADELPHIA. 4 figures. Fig, 1. Longitudinal vertical section of boiler and movable bridge,.. ~ inch to 1 ft. 2. Half section at c d. Half section at a b,.... 3. Perspective view of water-grate and movable bridge. 4. Section of joint of movable bridge with crown-plate,... Q size. 49. H. BOARDMAN'S COAL-BURNING BOILER. 2 figures. Fig, 1. Longitudinal vertical section,....... -i-inch to 1 ft. 2. Vertical section at A B,......... " 12. LIST OF PLATES. 50. F. P. DIMiPFEL S COAL-BURNING BOILER. 3 figures. Fig. 1. Longitudinal vertical section,........ inch to I ft. 2. Transverse section, with crown-plate over the tubes removed,. " 3. Half cross vertical section of smoke-box. Half section of tubes and shell,.......... " 51. M. W. BALDWIN & Co.'s COAL-BURNING LOCOMOTIVE. Longitudinal vertical section, ~ inch to 1 foot. 1 figure. 52. H. YATES' COAL OR WOOD-BURNING BOILER, BUFFALO & LAKE HURON RAILWAY. 3 figureS. Fig. 1. Transverse section of fire-box,...... in. to 1 ft. 2. Longitudinal vertical section of fire-box,... 3. Half vertical cross section of fire-box. Half back end elevation of fire-box,.......... " 53. MILLHOLLAND'S ANTHRACITE COAL BURNING LOCOMOTIVE, READING RAILWAY. 4 figures. Fig. 1. Longitudinal vertical section of the boiler and locomotive (scale on the engraving). 2. Half cross section through the smoke-box. Half front end elevation,......... " 3. Half cross section through the fire-box. Half back end elevation. 4. Plan of top of chimney,..... 54. COAL-BURNING BOILERS AND APPARATUS. 12 figures. Fig. 1. The Delano grate, longitudinal vertical section,. in. to 1 foot. 2. 10 and 11. Fire-brick deflectors, by M. W. Baldwin & Co. 3. Jones' coal-burning grate, N. York Central R. R., long. section 2 in. to 1 foot. 4. " c" " plan.. " 5. Step grate, Northern Railway of France, long. section.. 6. Baker's curved deflector, long. section. 7. Air admission by MI. W. Baldwin & Co. 8. Rocking grate. 9. Perforated grate. 12. Grate of Baltimore & Ohio coal-burners. 55. ENGLISH COAL-BURNING BOILERS AND APPARATUS. 15 figures. Figs. 1 and 2. Stubbs & Gryll's combustion-chambers, patent sealed June 2, 1846. 3. Hawthorn's projection of air to the top of the fire, through the door. 4. Projection of air above the fire. Caledonian Railway. 5. Boiler by Gray & Chanter, 1837. 6. " " 1839. 7. Jenkins' cast-iron arch or deflector, Lancashire & Yorkshire Railway, 1857. 8. Parkes' split bridge and air admission, 1820. 9. Steam jet by Besnier de la Pontonerie, 1857. 15. Frodsham's steam jet and air admission, Eastern Counties Railway, long. section of fire-box. 14. " " " rear view of fire-bos. 10 to 13. " " details of door and air deflector. 56. ENGLISH COAL-BURNING BOILERS AND APPARATUS. 9 figures. Fig. 1. D. KI. Clark's steam-jet, for inducing currents and a mixture of air at the surface of the fuel. Longitudinal vertical section of fire-box. 2. D. K. Clark's steam jet. The air tube and jet pipe, half size. 3. " " Transverse section of fire-box through air tubes. 4. Evans' downward draft boiler. 5. Barran's cup-surface boiler. 6. Evans' method of supplying air. 7 and 8. Lees & Jaques' bridge and method of supplying air. 9. Air supply at the surface of the fuel by a projecting grate. 57. VARIABLE EXHAUST APPARATUS. 7 figures. Fig. 1. Patrick's; front elevation. 2. " section of exhaust-pipes and movable disc. 3.'" plan of movable disc and ratchet. LIST OF PLATES. 13 Fig. 4. Patrick's; view from beneath, showing attachment of ratchet, &c. 1. (of the upper figures) Baker's; longitudinal vertical section,. ~ size. 2. " " cross " " " 3. " " section at c d, fig. 2, showing gear, " 58. MISCELLANEOUS. 13 figures. Fig. 1 Spence's furnace for making puddled steel. Longitudinal section. 2. " " " Transverse " 3. A. F. Smith's telescope chimney, Hudson River Railroad (dimensions on engraving). 4. Fire-plate joint; Hudson River Railroad,.... size. 5. Parrott & Head's variable exhaust, plan with cap (fig. 7) removed. 6. " " " front elevation. 7. " " " plan of cap from beneath. 8. " " ( vertical section. 9. Lees' brick arch for coal burning. 10. Griggs' " " " 11 and 12. Slotted grate for coal burning. 13. A. F. Smith's superheating apparatus, Hudson River R. R. 59. BEATTIE'S FEED-WATER HEATER, LONDON & S. WESTERN RAILWAY. 2 figures. Fig. 1. Side elevation of engine and heater. Vertical section of condenser I in. to 1 ft. 2. Plan of " " Transverse " ", 60. FEED-WATER HEATERS. 10 figures. Fig. 1. Ebbert's; Longitudinal vertical section. 2. " Transverse " 3. " Front elevation. 4. Eaton's; Longitudinal section, J-' size. b and 6. Eaton's; Details. 7. Clark's; Vertical section (dimensions on engraving) 1st plan. 8. " " " 2d plan. 9. " " " 3d plan. 10. Wilder's; " Mine Hill and Schuylkill Haven R. R. -Ia size. 61. ENGLISH INDEPENDENT FEED OR DONKEY PUMP, BY BEYER, PEACOCK & Co. 9 figures. Figs. I and 2. Side elevations.. ~. (dimensions on engraving.) 3 and 4. Vertical sections.... 5, 6 and 8. Transverse sections. 7. Vertical section of air-chamber.. 9. Plan. 62. THE GIFFARD FEED INJECTOR, AS BUILT BY SHARP, STEWART & CO. (ENG.) 10 figures. Fig. 1. Longitudinal section, showing operation,. (dimensions on engraving.) 2. Transverse section at throat,.... " "': feed-water chamber,.. " 3. Application to locomotive boiler,... " 4, 5, 6 and 7. Modifications illustrative of its action. 8. Longitudinal section of the locomotive size (No. 6),. full size. 9. Excess of steam. 10. Excess of water. 63. TEE GIFFARD FEED INJECTOR, AS BUILT BY WILLIAM SELLERS & CO. (U. S.) 4 figures, Fig. 1. Elevation. 2. Longitudinal section. 3 and 4. Application to the locomotive. 64. BOILER MAKING. 44 figures. Figs. 1, 2 and 3. Rolling the same plate of copper to different thicknesses. 4. Expansion joint at bottom of fire-box. 5. Copper sheathing over an iron corner. 6,'32, 33 and 38. Countersunk and conical rivet holes. 7. Illustrating the effects of offsets on circulation. 8. Preventing large flues from collapsing. 9 to 17. Joints experimented upon at Woolwich (see Tables, page 36). 18. Double welt joint possessing 82 per cent. of strength of entire plate. 2 14 LIST OF PLATES. Fig. 19, 42 and 43. Construction of stay-bolts and tie-rods. 20 and 21. " rings in the water-space. 22. Prosser's expander for flue-setting, side elevation,. full size. 23. " " half longitudinal section, " 24. " " "' end elevation, " 25 to 29. Baldwin's tools for flue-setting, full size, except fig. 28. 31 and 44. Welt joints. 34 to 37. Showing the oblique strain on lap-joints. 39. " effect of caulking. 40 and 41. Bertram's scarf-weld. 65. BOILER-MAKING AND FLUE-SETTING. 19 figures. Figs. 1, 2 and 3. Millholland's water-grate,...... full size. 4. Corrugated fire-plate,......... " 5 and 18. Flue-setting. Iron pieced with copper and a thimble. Illinois Central, full" 6. " Copper ring outside the flue.. Millholland, " " 7. " Iron screwed joint,.. Beattie, " " 8. " Prosser's expanded joint,.... ". 9. " Outside copper or brass ring,... " 10. " Flue end swaged down to a shoulder,... " " 11 and 12. " Joint made by compression with screw. Fisher,. " " 13 and 14. " Common English plan,....... " " 15. " Thimble brazed in,....... " 16. Staying thin plates. Screw-washer,....... " 17. " Millholland's method,...... " 19. Tubular stay with thimble....... Betts,... " 66. MISCELLANEOUS. 17 figures. Fig. 1. Beattie's steel cylinder-lining with annular space for jacketing or superheating,........... X size. 2. Rogers L. & M. Works inside chimney (dimensions on engraving). 3, 4 and 5. D. K. Clark's superheating apparatus. 6 to 9. McConnell's current deflectors or heat-traps for flues. 10. Williams' Experiment on the transmission of heat in flues. 11 and 12. Worthington's Percussion gauge,....... l " 13. Griggs' chimney (dimensions on engraving). 14 and 15. Heat-traps or current deflectors. 16 and 17. Baker's feed-water heater,.. o... 6 " 67. MISCELLANEOUS. 12- figures. Figs. I and 2. Illustrations of the strength of the cylindrical parts of boilers. 3. Hawthorn's annular safety-valve. 4. Baillie's direct-action safety-valve, with volute springs. 5 and 6. Making and setting wrought-iron grates. 7. Worthington's steam auxiliary pump with hand valve-gear. 8 and 9. 0. W. Bayley's divided fire-box. 10. Bonnet for deflecting air upon the surface of the fire. 11. Davis' Grate, for cleaning and stirring coal fires, cross-section. 12. " " longitudinal section. 68. MISCELLANEOUS. 3 figures. Fig. 1. Worthington's duplex independent steam pump for locomotives. 8. Addison's variable rotary exhaust nozzle, front elevation,.. ~ size. 3. L" " " plan.... i " 69. THE BISSELI, TRUCK FOR LOCOMOTIVES. 7 figures. Fig. 1. Half side elevation. Half longitudinal vertical section,.. in. to I ft. 2. Half front " Half crpss "... " 3. Half plan,.......... 4. Half vertical section of 2-wheeled truck,....1 in. to ft. 5. Elevation of pedestal ",.... I 6. The Bissell truck on a curve of 300 ft. radius. 7. The common truckl,' " " WHITOLE NUMBER OF PLATES, 77; WHIOLE NUMBER OF FIGUIRES, 596. INT R O D U C T I ON. THE object of this book is to facilitate the economical working of railways, chiefly in the following particulars:FIRST. In the employment of coal as locomotive fuel, especially for the wood and coke-burning engines now in service. SECOND. In decreasing the cost of generating steam, by means of better boiler materials and construction, and of utilizing, as far as possible, all the heat generated from the fuel. THIRD. In the adaptation of a radically improved system of Permanent Way. The fact that " Engines and Working," that is'to say, fuel,' repairs, and attendance upon engines, and " 3Maintenance of Way," constitute one-half to five-eighths of the operating expenses of railways, truly.exhibits the importance of the subjects treated. As to the manner of treating these subjects, the author deems it important to make the following explanations:The usual basis on which new inventions are recommended, is that they save a certain definite percentage of mworking cost, in experimental trials: under what circumstances, is not stated with such notable precision. The fact that a certain " passenger train " is run for a few months with less fuel than a " passenger train " on another line, may be a test of either the skill of the engine-driver, the excellence of the running-machinery, or the smoothness of the permanent way. Or, this economy may result solely from the economical generation of steam; bad management, machinery, and track to the contrary. It is well known that an expert engine-driver can burn either one ton or two tons of fuel in doing the same work; at the same time, the condition of the ballast, drainage, sleepers, rails, rail-joints, wheels, journals, springs, valve-motion, rubbing-surfaces, fuel, combustion, feed-water, vaporization, and separation; the force and direction of the prevailing wind, the cleanliness of the rail, the weight and speed of the train, the frequency of stoppages, the gradients and alignment of the road, the temperature of the atmosphere and the protection of the boiler and cylinders against condensation, the " line " of the axles and working parts, and other variable and important elements of economy or of loss, may each so modify the general result, that to measure either by the general result, is wholly inadequate and unsatisfactory. Or if the difference in the permanent way and the running machinery of two lines is approximately equated, the economy in fuel, of one over the other, although a test of the functions of the respective coal-burning boilers, is not a measure of their ultimate economy: excessive cost of maintenance, not at first anticipated, has condemned a large amount of patented and promising machinery, in the long run. WVhile extravagant representations of 50 and 75 per cent. saving are still a trap for our counting-room officials, it is encouraging to observe what slight importance railway managers, as a class, attach to the premature "results," certificates, and estimates which usually herald new inventions: the day is evidently approaching when the records of the dynamometer and the indicator will take the place of all sorts of approximaton and guess-work. But while the comparison of general results, in which there is such a variety of elements, is an uncertain test of either element, it does not follow that there are no certain means of testing the individual features of the railway system. in fact, taking the case of combustion, for instance, we can go behind the measure of pounds of coal per mile run, to the visible phenomenon itself; and behind this to the Chemistry of Combustion, of which the laws are fixed and definite. If a boiler does not make provision for the operation of these laws-and there are many that do not-then it is quite unnecessary to insist, as many inventors do, in similar circumstances, that it will save so many pounds of coal per mile.' Should a saving occur, it will only prove the absurdity of this method of testing any'thing. The visible phenomenon, however, is more conclusive. We know, without reference to railway practice, that at least one quarter of the heating value of bituminous coal is in a gaseous form, and may be either utilized in flame or wasted in smoke. The presence of dense volumes of smoke, therefore, is a rather more direct test of the quality of combustion, than the cost of running the engine per mile. We also know, from reliable experiments, that the invisible gas (carbonic oxide) sometimes wasted from coal-burning furnaces, is due to an excessively thick fire; and long practice has established other less obvious facts of minor importance; so that, on the whole, the observation of the phenomenon of combustion in locomotive fire-boxes will lead us verynear to the exact truth; it is the best means yet provided, and if railway managers would make good use of the information derivable from these visible and obvious sources, alone, we should see great changes for the better. xiv INTRODUCTION. Good Combustion, however, is but one of the requirements of a boiler. It must be durable, in virtue of good material, construction, and circulation; the transmission of heat must be perfected by suitable draft apparatus; the vaporization must be facilitated by good circulation and the separation of steam and water; dry steam must be delivered to the cylinders, and all the heat once generated must be, as far as possible, utilized. The permanence, smoothness, and elasticity of road-bed and superstructure depend on as many different conditions. And all must be compatible with a practicable first cost. What are these conditions? It is certainly unfortunate that these great features of railway economy have not been accurately measured, once for all, by a system of what we may call experimental practice-by a course of observations which no private individual, nor even a single railway company, could afford to undertake, but which all enterprising companies should have undertaken, jointly, and may yet accomplish. B3ut we are not in the dark-we have more light on this question than we are disposed to make available. The same fundamental principles, and the same visible phenomena which lead us to important conclusions about combustion, although not as familiar and definite as we could wish, may be at least the stepping-stone to very valuable improvements in all the other features of the railway system. In short, while we have no perfectly accurate measurement of results, we are not dependent upon guess-work for farther progress, nor in estimating the economy of practice as it exists, and of novelties seeking introduction. The subjects mentioned have been treated in accordance with these principles. Such pertinent facts, general results, and authentic reports as the author has been able to gather, and the very full illustration, by the Plates appended, of the present practice, and of the improvements and novelties proposed, together with some comments which the principles and practice have suggested, are offered to the public, in the hope that they will stimulate original thought in the right direction, and thus hasten an improvement which has too long been postponed for the want of analytical and systematic reasoning on the part of railway managers. In the matter of adapting wood and coke-burning engines, at a comparatively small cost, to burning,' coal economically as to combustion and repairs, the facts furnished in the following chapters are founded on such well-attested principles, and are proved by such definite and unmistakable results, that the attention of railway managers is confidently and earnestly called to the subject. It will also appear, we believe, that the best method of constructing the cheapest permanent way that can be economically run upon-the best method of economizing iron, timber and ballast, of settling the vexed question of rail-joints, and of decreasing the cost of traction and maintenance, is to introduce, in renewals which are now in progress, a substantially different and improved system of superstructure, wherein the whole of the material is utilized and employed to the best advantage. While the European practice is, on the whole, more economical than our own, there are some features of the American system to which the attention of European practitioners is particularly invited; among others, the equalization of the draft by diaphragms, etc., in the smoke-box, intensifying combustion by the use of the deadplate, etc., the movable grate for cleaning the fire, the variable blast, and various features of boiler construction. There is little, if any thing, to copy, in the American system of permanent way, except its later adaptations to the streets of cities and suburbs. In treating of the Chemistry of Combustion and of Permanent Way, a portion of the matter of " European Railways," by Mr. Zerah Colburn and the author, has been employed;' and some of the plates of that work have been republished. The author is farther indebted to Mr. Colburn for important advice and assistance, here and abroad. In the consideration of boiler-construction, copious extracts have been made from Mr. D. K. Clark's excellent chapters on this subject, in " Recent Practice," and their author has, in various ways, facilitated the preparation of this work. Mr. J. K. Fisher, of New York, has rendered important aid in the preparation of the following chapters on the transmission of heat, vaporization, superheating, etc. In thanking these gentlemen for their valuable assistance, and the many locomotive superintendents and engineers, at home and abroad, for the information and the facilities for observation and experiment which they have furnished him, the author would express his belief, that if he has done his part as intelligently as they have done theirs, the work will not be without value to the general railway interest. NEW YORK, November, 1860. ..........................................................................................................................................................................~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ ~~~ ~~~~~~~~~~~:::::::::::::::::::::::::::::::::::::::::::::...........:.:::::,.'....:.:..::..... ll!::':J::::::s..................::: ~::::::::::::::: ~::.................::::::::::::::::.................................................:,...............:..~,......:,~,:.::.........~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........ 1-1 l' l. (.,I..... i.::::...... ~.......-......................~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:.:..:.~..:...:.:....:..-..... liil~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~llill~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~llllSB@SB^^^^^ 11,' - I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:!!t?:.<'-~.'-?:!:::......................................:.. l ~::"^'sl l l ^y ^ iK:-~.:^:.-~:-;~~ l.l.! l^ l." l........................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,.......... I..I............................................................................................................::::~::::::l I:::::::::::::: i.:~;;. ^:':::... l l..................................' ^: ^ ^ M..l ii i i ii. l.li.^y y l i i. i i i..............^:::u ^'~:.l ^.ll.~:.< I ^.:.::~.^.i: ^. ^... ^..........................-.^:.....:.;".........i..................................=."i..=;...=.:............................................:::::.:......lii I- ^'~-<.. ~......:..~.. - ^..:.....~^ ^ -s.^ -i~:......~~..^ ~l.^;:^.~.: ~ - ~" I o~ 1:.*:.:.::.:::...X^ ~.................... I'"III'I gllllllll'^111111111!l.........l'::*:I I ~ ^." " " ^'" " " ^''l...............................................................^''I~~~~~~~~~~~~...................... I..................'".......::' ~:.:.'.'.::..:.'.,'.:.::.:..:::.....-:-...:...-...:..:.:.::::::::::::::::::::::::::::::::::::::::::::::...................................................................................................~ ~~~ ~~~ ~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.., ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~......... =.=========== = ~:i~~~~~~~~~~~~~~~........... ^^'"~^-ilt~~~~~~~~~~~~~~~~~~~~~ll~~~~~~lllilliilli^ t~~~~~~~~~~~~~~~~~~~~~~~~~~~~ll'^:;^ ^.~ ~ i~~~~~~~~~~~-iiN......... -..^....i iili~ii~.^:_^^ -^.......;^ ^^......^ ^ ^ ^.'"".:..-.. -~..........:;~;lll:...................................................:...:..:.:"-il::-..-^ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~........... 11 ^^11111111111111111111^...^-^"^^.>~.:..>~.;...,^^-A-. lllllull^^.....-^11 ll~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~....... ~:l... llll:...l....~.;.-l-l~~~~~...........;:.-:,l:l ^ ll l l //......^........:^...^ ^,1:1 ^ W ~ ^ ll^::::::;::ll::^ ~^ 1.11~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~................. I'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..... ^^r~~~hT~ r ~'~ ~ —-~~'-~~-~^<'^~^'~^^^^^~~~~~^^^^~ ^^^^^ ~^ ^ ^^ ^^~~~~~~........ THE ECONOMICAL GENERATION OF STEAM. PART I. B OI LE-MAKIN G. CHAPTER I. INTRODUCTOPY. A STRONGER and more durable inaterial than iron, for boilers, and a proportionately sounder mode of construction, are especially desirable, chiefly for the following reasons: FIRST.-To decrease the cost of maintenance of boilers.* The substitution of coal for wood, as fuel, is materially increasing the wear of fire-boxes, flues, and grates.t SECOND.-To reduce the weight of locomotives. Without broadly discussing this policy, it is sufficient to mention that the professional opinion and the more enlightened practice of both Europe and America, are unmistakably in this direction.,For very high speeds, light-rolling stock is simply indispensable; and for common traffic, whatever weight for adhesion may be demanded the policy of employing tank-engines, and dispensing with tenders-the locomotive proper being proportionately lightened-will hardly be questioned. THIRD. — To meet the demand on the present class of machinery, for more power without a positive increase of weight, since sudden deviation from the standard patterns of locomotives, however desirable, is not to be expected. The maximum size and weight of locomotives appear to have been reached some time since; increasing business will, therefore, be accommodated by a higher steam pressure, at the risk of trouble from the slipping of driving whleels.$ * So few railway accounts subdivide the general head of " repairs of engines," that no exact average of boiler maintenance can be stated. On one bituminous coal-burning line, it is estimated at 8 per cent. of the total cost of repairs of engines; on another at 19 per cent., and on an anthracite coal-burning line, it is said to be 40 per cent. In the early stages of our coal-burning experience, the adaptation of boilers-especially of old wood-burning boilers —and the education of raw firemen, are extraordinary items of cost. Besides, an exact conclusion would hardly magnify the importance of a subject which is notoriously the head and front of difficulty in coal-burning. t The Locomotive Superintendent of the Pennsylvania Central Railway estimates as follows: The flues of a passenger engine, burning Pittsburg coal, will last 14 years, and burning wood, 25 years. Their cost is $555 —$17 per year extra cost for coal. A copper fire-box will last 6 years with coal, and deducting the value of old copper, will cost $680. An iron firebox, with wood, will last 12 years, and cost $480 —extra cost due to coal, $73 per year. A set of grate bars with coal, will last six months, and cost $14; with wood, one year~$14 extra cost for coal per year; total increased cost due to coal, one year, $104. The average mileage being 20,361, gives about one-half a cent per mile as the increased cost of repairing coalburners. The increased cost of repairs due to coal, on the Readcling Railway, is found to be 1} cents per mile run.: With the ordinary class of machinery, the standard pressure on the New Yorkl and Erie Railway has gradually risen 16 THE ECONOMICOAL GENAERATION OF STEA3M. It is believed that the extra cost of the occasionally deficient adhesion of light engines, is less than that of the extra repairs of way and equipment due to heavy engines. FOURTH.-By allowing thinner heating surfaces and greater uniformity of strength in boilers, to facilitate the economies of a higher steam-pressure, which are, chiefly, greater expansion, less priming, less boiler space, and less dead weight. FIFTH. — Another most important reason is, to increase the safety of boilers, since explosions are usually the result of insufficient strength of material, as well as of malconstruction. to 140 lbs. More than this is frequently carried. The standard pressure on the western division of the New York Central Railway, overcoming the heaviest grades of the line, has risen with the increase of traffic from 110 Ibs. and 120 Ibs. to 140 lbs. The same weight of engines has been preserved; Ten years since, the general maximum pressure was rarely above 100 lbs. It now averages 110, and is 120 on many American and European lines. CHAPTER II. MATERIALS. 1. IRON. —For all parts of boilers, except those exposed to the direct action of fire, the use of other materials than iron was not seriously contemplated till the introduction of cheap steel. Strength and lightness being the chief requirements, the only effort has been to secure iron plates which would bear the greatest tensile strain. In this country, the average of English plates has been found more uniform and sound than the average of American plates; this has rendered the former especially trustworthy for fire-plates, where these qualities, rendering them proof against blistering and cracking, are more important than mere tensile strength. The best American iron, however, is naturally both stronger and tougher than the best English, bearing an average tensile strain of 31 tons per square inch, while, according to Clark,* the best Yorkshire may be averaged at 25 tons, and the best Staffordshire at 20 tons.t And as the manufacture of American iron is constantly improving, it is coming into almost exclusive service, in this country, for purposes requiring strength chiefly, and into increasing use for fire-plates. It is now customary with the Paterson locomotive builders, to make special contracts, at a slight increase of price, for warranted plates. Should these show defects under working, they are returned to the maker, who exchanges them for new plates, and who also refunds the cost of all the labor that had been applied to the defective material. The experiments as to the strength of iron, with the grain and across the grain, essentially differ; those qualities which are most soundly worked, however, exhibit the least variation. The greatest average difference of strength of the same plates, with and across the grain, by Messrs. Napier's last experiments, was for iron 7,086 lbs., puddled steel, 22,621 lbs., and homogeneous metal, 790 lbs. The tendency of fibres to split apart, is obviously increased in a greater ratio than their tendency to pull apart longitudinally, by an excess of cinder, since the rupture in the former case is confined more to the cinder than to the iron. This shows the importance of employing sound and uniform iron, whatever its natural qualities may be, in all parts where the cross and lateral strains are similar; and it shows the propriety of placing the grain of iron in the direction of the greatest strain, wherever this is possible. 2. USE oF IRON FOR FIRE-PLATES.-Iron was as exclusively used in America for fire-plates, before the introduction of coal fuel, as has been copper in Europe, with all kinds of fuel. At the present time, the use of iron is increasing in England for both coal and coke, while copper is largely employed here for bituminous coal, and by some practitioners for anthracite coal, while others pronounce it especially unsuited to anthracite coal. The experience of boiler-makers in both countries is represented as widely diverse, respecting the merits of the two materials; and this diversity is not always attributable to any special causes, such as sulphurous fuel, impure water, or unsound material. It rather appears to be due to such abnormal causes as insufficient waterspaces, bad construction and careless management. And here also the experience is not uniform. While one contends that (necessarily) thick copper will stand the harder treatment, another insists on the superior endurance of very thin iron. + The relative merits of the materials, excepting certain peculiar cases, are evidently quite as unsettled as any problem in locomotive economy can be. Were it not likely that both are to be superseded by steel, the subject would merit the most thorough investigation and discussion. The following considerations, however, will prove at least timely and suggestive. * "Recent Practice." t See also, Tables in the Appendix. $ Quarter-inch iron flue-plates were found more durable than any other metal, of any other thickness, on the Housatonic Railway. 3 18 THE ECONOMICAL GENERATION OF STEAM. 3. DEFECTS OF IRON FOR [FIRE-PLATES —REMEDIES.-There appear to be four principal defects in iron fire-plates: blistering and cracking, resulting from comparative want of homogeneity, corrosion, and comparatively small conducting power. The comparative want of homogeneity in iron, is both an indirect and a direct cause of its ultimate failure. It facilitates the formation of blisters in the manufacture of the plate, and it offers less resistance than copper to the trying effects of fire. The difference of temperature to which its respective sides are exposed, causes unequal expansion and contraction in even a I-inch plate. This, by constant repetition, sensibly affects the cohesion of the particles of any compound substance. Were iron more nearly pure did the disturbances of temperature act on simple molecules of metal, without the interposition of cinder, the deterioration would of course be less rapid. The failure of all very " raw " iron under alternate compressive and tensile strains illustrates this fact. Blistering is due chiefly to the imperfect welding of the component parts of the pile of blooms composing the plate, the imperfect weld being incident to the oxidation of the surfaces while heated. Many of the cheaper kinds of plate are made of a pile of three blooms; the outside strata are sound and reheated; the inside stratum is raw, poor iron. Such plates are, of course, very liable to blister. The experience with the cheaper varieties of American boiler-iron is, that very large plates are soundest at the edges, and especially weak and liable to blister in the centre. This is attributed to the inadequate heating of the central part of the pile, causing an imperfect weld. But when the pile is considerably flattened under the hammer, and then reheated, its centre is, of course, thoroughly softened, and the difficulty mentioned is not experienced. Indeed, such plates are found strongest and soundest in the middle, the edges being more strained by the process of rolling, and probably containing some cinder which has been worked from the inside of the pile.. Large plates may be made from single blooms, if the central part of the bloom is so softened by reheating and gradual flattening, as not to modify the texture and soundness of any part of the iron. Plates thus made, and avoiding the difficulty mentioned, are vastly superior in other particulars. No oxidation can' occur, except on the surface of the bloom, and no air or gases are confined, to reappear in the form of blisters. And the advantages of large plates, quality being the same, are too great to escape notice. Fewer joints are required, so that the general strength is increased. The absence of joints in the fire-box is important for another reason: the thicker parts, conducting heat less rapidly and expanding less uniformly, soon crack and disintegrate. Rivet-heads, especially, burn off, in an intense coal fire, before other parts are injured. Mr. Fairbairn long since suggested that the piles for forming rolled or forged masses, that are too large to be made from single blooms, should be composed of convex-sided pieces, touching each other in the middle, but not at the edges, so that the cinder should have a chance to work out from the centre. The hammering of a pile of bars on all sides, before the cinder is pretty thoroughly squeezed out, simply incloses or welds in the impurities of the metal, and the air or gases filling the interstices.* The " cold-shut" thus left in the mass, when it is rolled into plate, shows itself in the shape of blisters. These may be skilfully pricked and flattened, but there is no weld, and the fire will soon begin to destroy that part of the plate. Blisters will always show themselves in a plate. The strata of iron being thinner, they get red hot in a blacksmith's fire more quickly, and black more quickly when cooling. This also accounts for the rapid destruction of blistered plate. The fire side stratum does not come in conitact with water, but only with the inclosed air, which is a poor conductor. Hence it becomes red hot, if not white hot, while in use. The common practice of trying to keep plates flat while they are flanging, or of. straightening them cold, after they are flanged, tends to aggravate any incipient blisters and to create permanent strains, which a coal fire is liable to remove, by cracking or splitting the plate. The most successful boiler-makers heat the plates which the process of flanging has warped, before they are flattened; and on the Reading Railway, all the fire-plates are annealed by being raised to a red heat and cooled gradually in sand. Such plates, being of originally good material, stand the intense heat of coal nearly as well as copper. The blistering of iron plates is often a cause of their cracking, since it allows the fire side stra* Doing just the opposite of this, is the important function of Burden's squeezers, for working puddle-balls. MATERIALS. 19 turn to become weak by burning. But cracking is sometimes the first species of failure; it more frequently occurs when the plates are comparatively thick, and it is more likely to be associated with the earlier coal-burning experience of a railway than with its more mature practice. On most of the English lines, where iron fire-plates have been but experimentally used, their failure has been due to the excessive thickness of the metal. On the Eastern Counties, ~ inch Lowmoor plate cracked after a few months' use. With us, the same quality of iron, but 1 to 3ths thick, stands the more intense heat of anthracite coal. It does not, therefore, appear to be a remediless evil. The causes of cracking, except blistering, do not appear to be inherent in the material, since iron fire-plates stand the most intense heat of anthracite coal better than copper in some cases. In such cases, the plates are invariably of the best material in the market; they are usually thin-sometimes not above {-inch-and they are necessarily kept in contact with water by means of thorough circulation. Cracking is, of course, the result of a great and abnormal strain, caused by intense heat and expansion, which soft, homogeneous copper can better resist than hard and less pure iron. To remove the abnormal cause, by preventing undue heat through the means of perfect circulation, or by a construction of boiler which will allow some difference of expansion between its parts, would appear to be the best preventive of this species of failure. The parts in immediate contact with the fire, fail most rapidly; a dead-plate entirely around the grate, would prevent intense combustion in immediate contact with the fire-plates; this arrangement would also aid a bituminous coal fire, and could hardly be disadvantageous in case of anthracite.* Since the strength of plate required in the cylinder of the boiler is rendered unnecessary in the firebox, by reason of the stay-bolts, a purer and hence softer iron, more like copper in its nature, could probably be employed with advantage for fire-plates. The corrosion of iron plates appears to be a remediless evil. It would occur by simple chemical action, with' pure water, but is doubtless much aggravated by the impurities of the water on the one side, and by the character of the fuel on the other side; the decomposition of iron plates by the sulphur in coals, is the chief cause of external depreciation, while the corrosion of iron boilers is a serious evil. Mr. Zerah Colburn, in " Boiler Explosions," mentions the following facts: " In a boiler which recently exploded at Tipton, considerable breadths of the iron were iound to have been reduced in thickness to -T inch. In the case of the explosion of a boiler at the Clyde Grain Mills, at Glasgow, in April, 1856, extensive breadths of the iron were said to have been reduced to the thickness of a sixpence; and in the disastrous explosion which occurred in August of the same year, at Messrs. VVarb-rton & Holker's, at Bury, the evidence showed that the bottom plates had been reduced for a greater or less width to only -1 inch in thickness. In the year 1859, there were reported 44 cases of'dangerous,' and 153 cases of'unsatisfactory' corrosion, among the 1,618 boilers under the inspection of the Manchester Boiler Association. Thus there was nearly one case of corrosion in every 8 boilers, in a single year." Another comparative disadvantage of iron, aggravating the effects of want of homogeneity, is its inferior conducting power. f This affects the durability of fire-plates, in proportion to their thickness. Very thick iron plates give way sooner than those that are thinner, because the heat cannot pass through them rapidly enough to prevent either burning, or excessive expansion, on the fire side. The most obvious remedy, both for this defect and for the adverse results of want of homogeneity, is to make the plates very thin. More stay-bolts are required, but these will not necessarily impair the circulation.$ An objection to iron, however, simply because it is a less rapid conductor of heat than copper, would come with very bad grace from railway managers who allow every inch of heating surface to be plastered over with lime and other incrustation, frequently ~th inch thick, and often enveloping three or four rows of lower flues. However interesting the subject of conduction may be for discussion, it will never be of appreciable importance in practice, till pure water is employed, or till compositions for removing scale are adopted. 4. ADVANTAGES OF IRON FOR FIRE-PLATES. —The advantages of iron over copper are its * See also, the chapter on Combustion. t See also the section on Flues. t See also the sections on Circulation and on Thin Fire-plates. 20 THE ECONOMICAL GENERATION OF STEAM. superior strength, stiffness, and hardness. Its strength and stiffness allow the use of much thinner and lighter plates than would be safe in case of copper, since the latter metal, however thickly stayed it'may be in flat parts, must have considerable thickness for flanging and riveting. Its hardness renders it almost indispensable for anthracite coal,* the sharp flying particles of which soon cut out even the flat parts of copper-plates, as well as their exposed corners. 5. COPPER FIRE-PLATES.-The advantages and defects of copper fire-plates have already been indicated. The former may be enumerated as follows: Good copper is always sound; the process of its manufacture prevents the formation of blisters, the solid ingot being sufficiently large to make an entire plate.' There is a large amount of impure copper in the market, which does blister, crack, disintegrate, and harden under the action of hot water. It is probable that American railway companies are unwilling to pay for, or do not try to obtain such qualities as have lasted a mileage of above 300,000 miles on British lines. Few, if any, of our copper boxes for coal, have had time to endure this test, but the evidences are that they will not stand it. Ten to fifteen years is the life of a copper box, without essential repairs, on the Eastern Counties and other English lines, with coke fuel, which is more destructive both as to the mechanical action of flying particles and as to intensity of heat, than a raw coal fire. Copper does not materially suffer from oxidation or any other chemical action to which it is incident in the fire-plates. It is also a better conductor than iron. It is more uniform and homogeneous than iron, and will bear a greater degree of irregular expansion and contraction. It is softer, more ductile, and hence more easily worked than iron. It may be stretched to a greater extent, in intricate flanging, and may be rolled into one plate of several thicknesses. This is accomplished by laying a separate sheet on that part of the plate to be made thinnest, and passing both through the rolls together. (Plate 64, Figs. 1 and 2.) Or when a thick edge for riveting only is wanted, two plates are laid together as shown in fig. 3, and both are shaped as required.t This is a very simple and feasible method of adapting copper to the uses required, viz.: strength at the' joints and lightness in the body of the plate. Especially is it important in case of the flue-plate, of which the part perforated by the flues or by the rivets, may have the extra thickness of 7ths to I inch, while to continue the same thickness to the fire-plate part of the same sheet below, would be to unnecessarily load down the engine, at a considerable extra cost for materials, I or 3ths inch being thickness enough. On the Eastern Counties and other English lines, two thicknesses of one plate has been deemed so important as to warrant the expense of drawing down that part forming the leg of the fire-box, under the hammer, from 7ths to' inch. A joint between a thick and a thin plate, at this point, would be peculiarly liable to leakage from the intense heat and the great and unequal expansion due to it. The rivets of such joints are known to have been cut partially off, and to have worked loose. The chief disadvantage of copper, necessitating a considerable extra dead weight and first cost, is its comparative weakness, its tensile strength being but about 35,800 lbs. to the square inch. Copper grows constantly weaker with heat, and at 1,100~, it is weaker than lead. Its specific gravity is 8 9, while that of iron is 7'7. And the thickness of copper side and back plates is usually - inch, while that of iron is V'ths. The copper is therefore about 85 per cent. heavier than iron. A copper box would weigh 1,850 lbs., while an iron box of the same dimensions would weigh but 1,000 lbs. The cost of copper plates is 28 to 29 cents per pound in this market, while Lowmoor plate and the Albany semi-steel plate are but 8, cents, American fire-box plates being 6 to 6G cents. The old copper, however, is worth 24 cents per lb. when cut up, if not burned, while the worth of old iron plate is but 1 to 1} cents, when cut up. Cutting old fire-boxes to pieces costs about one cent per pound. The superior conducting power of copper is, to some extent, offset by the greater thickness required for strength. 6. IRON AND COPPER COMBINED. —A copper flue-plate is often put into an iron fire-box, because, by its superior expansion, it is supposed to hug the flues more tightly, and thus to prevent leakage. Under the latest practice of setting flues in iron plates, however, the use of cop* This is the fact on the Reading Railway. f This plan was suggested by Mr. Hudson, Engineer of the Rogers Locomotive Works. MATERIALS.:21 per, for this reason alone, is unnecessary. It may hug the flues more tightly at first, but being less elastic than iron, while its resistance to compression is but about 6,800 lbs. per square inch, it is found to relax its hold more readily. The softness of copper certainly affords a better opportunity for bad workmanship and rough usage in firing and repairs; but its superior expalnsion is likely to injure its joints with the iron plate, and has, in many cases, caused leaks around the rivets. The lower joint of the copper plate, when made as at Fig. 4, Plate 64, is a sort of expansion joint, and sensibly remedies this difficulty. For both coal and wood burning, iron flues are preferred to copper or brass, in this country; in Europe, brass is the favorite material, although iron is much used. In either case, the flue is harder than the flue-plate, if the latter is of copper, hence the flue-holes often become oval and irregular during' the process of setting. A copper flue-plate in a colbustion-chalber, being a mere perforated sieve, with little solid margin, is found to crush and expand unequally. And since the heat is less intense than in case of a plane fire-box, the peculiar advantages of copper are not required. Iron is better, because it is stiffer. The conclusion from the comparatively incomplete practice of this country is that the whole of the fire-box (technically called inAmerica, the "inside fire-box") should be of copper, if this metal is employed at all. A copper flue-plate with iron side and back plates is rarely, if ever, used abroad.; Indeed, the standard objection to a copper box with an iron casing (outside box) is that its superior expansion is likely to loosen and break the stay-bolts. Wider water-spaces somewhat remedy this evil. The same result, in less degree, occurs with a fire-box wholly of iron, since the inside plate is most heated. Wide water-spaces are equally valuable in any case. 7. CONCLUSION AS TO IRON AND COPPER.-AS between iron and copper, the results do not show such decided advantages on either side, as to settle the question finally. It has been mentioned that on English lines, there are large numbers of copper boxes which have, without material repairs, run 300,000 miles during the last 10 to 12 years, and are yet in good condition. In view of thie rapid introduction of coal fuel with its sulphur and intense heat, it would not be safe to say that iron, well as it may have stood a short test, will prove the better material in the long run, however thin, light, and cheap it may be. Nothing, so far, in the useof steel plates, indicates their final failure, so that both copper and iron are likely to be superseded. Meanwhile the test is going on, and results are more carefully analvyzed every year. For present purposes, it is sufficient to say that the' most successful practice of this country favors thin iron for fire-plates, in preference to copper. And the current practice in this matter, unlike that in many other matters where bad results do not so speedily, palpably, and expensively exhibit themselves, is to be relied on as a pretty accurate exposition of the real facts in the case. 8. INFLUENCE OF GOOD CIRCULATION ON THE DURABILITY OF FIRE-PLATES.-Finally, there is one important principle which appears to underlie this whole mater: The chief cause of the failure of a fire-plate, is the direct effect of intense heat. Now if the heat applied to it is conducted by it to the water, through the medium of thorough circulation, as fast as the water will receive it, then the temperature of the fire side of the plate can hardly be much higher than that of the water, or some 330~ for the ordinary locomotive pressure. The author has been unable to learn of any experiments directly bearing on this matter; the general fact, however, that boilers with wide water-spaces last longer than those with narrow spaces, appears to establish the theory. The plates of boilers designed for wood, having only 2~ to 2f-inch spaces, when subjected to the more intense heat of coal, invariably fail faster than those with 3 and 4-inch spaces. But we have no reason to believe that even 3 or 4-inch spaces allow thorouyh circulation, As far as the practice teaches any thing, it teaches that the life of fire-plates is proportionate to the circulation. The water-tubes of the Dimpfel boiler (Plate 50), where they turn up towards the crown sheet, never burn out, although subjected to the most intense heat of the fire; the circulation through them is full and rapid. The lead'rivets used by Fairbairn in the crown-sheets of fire-boxes, and the fusible plugs, consisting mostly of lead, commonly used in American engines, last for years. The heat applied to them is less intense than that adjacent to the legs of the boiler, and they are better conductors, but they are much thicker than the plates. '.S THE ECONOMICAL GENERATION OF STEAM. It is probable that the temperature which all fire-plates sustain, is much greater than is generally supposed. The author has repeatedly found " blue steam " at the lower gauge of a wood-burning engine while working at its full power, while there was steam and water at the upper gauge. The water-spaces were 21 inches, and the lower gauge 1 inch above the crown sheet, and the engine did not materiallydiffer from the ordinary pattern. Firemen and engine-drivers have reported that the fire-plates of their wood-burning boilers, when coal was burned in them, have been found visibly red hot, and that these plates failed rapidly. The defective circulation in these boilers was further proved by their excessive priming. Since the strength of iron plates remains unchanged at all temperatures from 0~ to 400~,* the safety of the boiler is not impaired by the heat of steam or water which, under 200 lbs. steam pressure, is but 381". But when steam and water are so repelled by heat, in consequence of insufficient circulation, as to allow the plates to acquire a low red heat-probably a daily occurrence in boilers with narrow water-spaces, changed from wood-burners to coal-burners-the strength of the plates is reduced one quarter. Next to good circulation, thin fire-plates will obviously promote the durability of boilers, while they will also relieve them of much dead weight. This fact appropriately brings us to the consideration of steel for fire-plates. 9. STEEL. —Semi or puddled steel, and homogeneous metal or cast-steel, are rapidly coming into service for fire-plates. The regular use of these metals commenced in England some three years since. In this country, Clay's steel had been applied to parts of small boilers a year ago. A new stimulus was quite recently given to the general introduction of the improved material, by the advertisement of Messrs. Corning, Winslow & Co., of the Albany Iron Works, to furnish semi-steel of a warranted tensile strength of 90,000 lbs. to the square inch, at the-price of Lowmoor plate. This establishment had, during the two previous years, perfected their process, and introduced the necessary machinery. Meanwhile Howell's homogeneous metal and Firth's puddled steel, as well as the Bessemer steel, have begun' to be regularly imported, and it is probable that locomotive boiler-making is at this time in the early stages of a thorough and radical improvement, which will surely and palpably increase the safety and economy of locomotive maintenance and working. Indeed the whole process of iron and steel manufacture seems to be in a transition state. The remarkable results of the Bessemer experiments at their commencement in 1856, stimulated new enterprises, and revived old and half-perfected schemes in this direction. One feature of the re. form promises an unprecedented measure of success; the dicta of the unaided senses, derived from long observance of the phenomena of metals under treatment, are giving place to the profound teachings of chemistry, out of the ultimate laws of the construction of matter. Each succeeding year will be likely to clear the horizon of much uncertainty and, error, and to elicit and apply new discoveries to such an extent, that the discussion, experiment, and practice of the present day, however comprehensively detailed, would soon be found incomplete and unsatisfactory. A brief mention of the manufacture of cheap steel for boilers, etc., will not be unimportant to our present purpose. 10. PUDDLED STEEL.-The process of making puddled steel t may be described in general ~ Fairbairn on the tensile strength of wrought iron at various temperatures. t This is probably the process adopted by all the makers of puddled steel. The history of the improvement, however, has developed the following facts. The plan of making steel in a puddling furnace appears to have been first advanced by Mr. Riepie, a German. Riepie's proposed method, upon which the proposed method of Clay and Benzon (patent dated Aug. 20, 1858) is an improvement, is substantially as follows: The iron, as in other processes of puddling, is decarbonized by being heated in contact with the air, the oxygen of the latter uniting with the carbon of the iron, and passing off in the form of carbonic acid gas and carbonic oxide. It is then recarbonized, during the same process, chiefly by the addition of fresh carbonaceous iron, and is then prevented from going into the state of iron by a reduction of the heat. A charge of say 280 lbs. is placed in the common reverberatory puddling furnace, and after melting has commenced, the fire is tempered by the damper; a quantity of cinder and scale from the rolls is then added, and the mass is melted down together. The whole is then puddled, with the addition of a small quantity of the black oxide of manganese, common salt, and dry clay, previously ground together. After puddling for someminutes, the damper is raised and about 40 lbs of pig iron are put into the furnace near the fire bridge, upon beds of cinder raised for the purpose. This iron is used to recarbonize, just to the right extent, all of the iron just decarbonized by puddling. When this additional quantity of pig iron begins to trickle down, and the mass on the bottom of the furnace begins to boil and throw out from the surface, the well-known blue jets of flame, the additional quantity of pig iron MATERIALS. 23 terms, as follows: Cast-iron contains from 3 to 5 per cent. of carbon; steel contains from - to 1 per cent. of carbon, while wrought iron contains but a trace. In changing from cast to wrought iron, in a puddling furnace, the pig-metal passes through the state of steel, that is to say, it is steel before it is wrought iron. Now, making puddled steel is simply stopping the common puddling process just at the moment when the decarbonizing mass under treatment is in the state of steel. Plate 58 shows sections of a furnace invented for this purpose, by Mr. James Spence, of Liverpool, and designed to allow the decarbonizing process to be instantly and completely stopped at the proper point-otherwise the steel will rapidly pass into the state of wrought iron. Now, as the principal agent in the process of decarbonization is the oxygen contained in the air passing into and through the furnace (which, having a chemical affinity for the carbon, combines therewith, and removes it in a gaseous form), it follows that in the first stage of this process a full supply of oxygen is required, and that in the second stage none should reach the metal. In other words, in steel puddling, the first stage requires heat with oxygen, and the second heat without oxygen. * To effect objects thus differing and opposed, requires an apparatus having different modes of action. The common puddling furnace hitherto used in this manufacture has only one mode of action, and in consequence, steel has been made therein with difficulty and uncertainty by the aid of large quantities of cinder or other materials,'and by the closing of the damper at the latter stage of the process, and thus the uniformity of the product depended greatly on the judgment and dexterity of the puddler, and moreover, the closing of the damper caused a reduction of the temperature, which had an injurious effect upon the working of the metal in the after processes of collecting, welding, and balling. In order to obviate these difficulties and to obtain a more uniform product, Mr. Spence constructs and works the puddling furnace in the following manner: The puddling furnace is built with two grates (one behind the other), divided by a bridge, the ash-pit of each grate being entirely inclosed, and having a door or doors in the front or side. The grates being both of the same length (about 3 ft. 3 in.), the inner one is made about 2 ft. 4 in. in width, and the outer or end grate about 1 ft. 9 in. in width. The fire-bars of the first or outer grate are also placed about 4 in. below the level of the bars of the second or inner grate. It will be seen that by this arrangement when the door of the inner or second grate is closed, no air can pass through that fire, while the air which has escaped combustion in passing through the first or outer grate before it can reach the metal must pass over the incandescent fuel on the second or inner grate, and will be thereby deprived of its remaining oxygen. The fire-bridge of the ordinary puddling furnace is a low wall of fire-bricks extending acrc's the furnace behind the grate. The first bridge, or that which divides the two grates, is here constructed in the ordinary manner, but the second, or that which divides the inner grate from the carefully watched, as the puddling proceeds, the heat being never allowed to rise above cherry redness, or the welding heat of shear steel. The blue jets of flame are gradually extinguished, while the formation of steel grains continues, these fusing together so as to bring the whole into a waxy state. Except with great care, the mass would pass more or less into iron, the object being to retain and more uniformly diffuse, or in other words, to chemically combine all the carbon contained in the additional charge of iron. The mass being now converted into steel, a part of it in the shape of a puddle ball, is taken to the hammer, the fire being in the mean time stirred to preserve the heat for the remainder of the charge, the damper being entirely shut, and the rest of the charge left in the furnace being always kept covered with cinder slack, until the whole has been taken out to be worked under the hammer. If very carbonaceous sparry iron be used, 20 lbs. only are sufficient for the carbonization of the original charge of 280 lbs., the object being to retain, as nearly as may be, from i to 1 per cent. of carbon, uniformly combined in the whole mass of the product., Clay's -improvement consists chiefly in charging all the materials at the same time, instead of opening the furnace frequently, and thus lowering the heat and interfering with the operation. Together with the charge of say 280 lbs. of pig iron, a nearly equal quantity of slag or cinder is charged at the same time, and from 20 to 60 lbs. of the iron, according to the proportion of carbon it contains, are then so arranged in the furnace, as to remain covered by the cinder, and thereby protected from the action of the heat, until the remainer of the charge is thoroughly melted. The reserved portion of iron is then uncovered and drawn forward into the molten mass. This produces a more thorough rising and boiling of the charge, after which the process is completed as before. 24 THE ECONOMICAL GENERATION OF STEAM. tube of refractory clay is fixed, closed at one end and connected at the other with a supply of air which may be given under pressure. The side of this tube next to the metal is furnished with a longitudinal opening, through which such air will pass directly on to the metal. Fig. 1 represents a longitudinal section of a puddling furnace constructed according to the principles of this invention, and Fig. 2 is a horizontal section of the same. A, A, are the exterior walls of the furnace; E, E, the dome; C, C, the sole; and D, D, the chimney. E is the first or outer grate with its ash-pit F, which is inclosed on all sides and furnished with a door or doors; and G is the second or inner grate with its respective ash-pit I, also inclosed and furnished with a door or doors; J is the first bridge which divides the two grates; and K is the second bridge, the upper course of which is formed of a tube of refractory clay L, L, closed at one end (see also detached view), the inner side of which, or that next the sole of the furnace, is furnished with a narrow longitudinal opening M, M; N is a pipe in connection with a supply of air. In operating with this furnace, the puddler charges it with about 480 Ibs. of pig or cast iron. Whilst this is melting, the doors of both the ash-pits F and H are kept open, or mainly open, in order to bring both fires into a state of activity. As soon as the iron is melted, the'object is to remove the excess of carbon, and in order to effect this, the door of the outer ash-pit F is closed, and that of the inner ash-pit H is kept quite open, and the supply of air is admitted through the hollow bridge L on to the surface of the' metal. The fluid metal is puddled in the usual manner, every portion of it being thoroughly exposed to the decarbonizing action of the oxygen. When the metal begins to solidify (technically called " coming to nature ") the carbon is reduced to the proportion constituting steel, and as steel is less fusible than cast iron, it passes from the fluid to the solid state. The commencement of this transition is indicated by the appearance of grains above the fluid surface. On observing these grains, the supply of air through the hollow bridge L is shut off, the door of the outer ash-pit F is opened, and the door of the inner ash-pit H is closed. This being done, the necessary heat is maintained by the outer fire for the further work of collecting, welding, and balling the metal, whilst the air which may have passed unconsumed through that fire, having to traverse over the incandescent fuel on the inner grate, is thereby deprived of the remainder of its oxygen, and rendered inert. It is found that great advantage arises from the use of this method, as it substitutes mechanical contrivance for the very uncertain skill of the puddler, and thus greatly reduces the difficulty and uncertainty of the manufacture. When the puddle-balls are formed, they are taken to the hammer, and worked out as usual. In place of making the outer grate to draw or open to the front, it may be built to open at the end, to suit the convenience of the works. It is also not absolutely essential to inclose the ash-pit of this fire, but in practice the patentee finds it desirable, as it gives a more perfect control over the process. The complete inclosure of the inner ash-pit, as described, is absolutely necessary, and forms an essential feature of the invention. The dimensions above given for the grates are those which are preferred, but they should be regulated to suit the particular description of coal used as fuel. When steel of a full degree of hardness is required, it is found expedient to use but one sort of iron; but when a soft steel is desired, it is preferable to work a mixture of different sorts of iron. It is the common practice to introduce fluxes or salts into the furnace to assist in this manufacture, but their use is not absolutely necessary or essential to the production of steel in the furnace described, and their benefit is more or less uncertain and contingent on.their equal and thorough diffusion throughout the entire mass. When this equal diffusion can be depended upon, the puddler may use as a flux an alkaline base, which will unite with and abstract the impurities of the iron. If the qualityof the steel to be produced be the first consideration, the patentee prefers to employ the muriate of ammonia as a flux, and he considers this the best, the next best being the sesquicarbonate of soda; but when economy of cost is the leading consideration, the chloride of sodium (common salt) may be used with some advantage. The precise flux to be used beneficially can only be determined by the character and quality of the particular kind of iron or mixture of iron with which it is to be employed, as no two kinds 3 MATERIALS. 25 of iron contain exactly the same impurities. In all cases, the flux is reduced to a fine powder, and introduced into the furnace, and thoroughly incorporated and diffused throughout the mass of metal at the commencement of the boiling. 1I. AMERICAN SEMI-STEEL. —The method of making Semi-steel at the Albany Iron Works is as follows: The furnace employed is substantially the ordinary boiling furnace, that is, a puddling furnace adapted to'a higher degree of heat than is used in the ordinary process. The pig-iron, broken into small pieces, is placed on a trough-shaped hearth, removed from the solid fuel. The flame, as usual, passes over a bridge wall and is deflected upon the iron. For a charge of 336 Ibs. of pigs, about two barrow-loads of cinder and scales from the forge (oxide of iron) and other fluxes are added to the cinder previously melted on the hearth, the whole forming a bath in which the iron is heated under a constantly increasing temperature.. The cinder and fluxes boil from the escape of the gases caused by the oxidation of the carbon in the iron, and of the iron itself. To prevent the too rapid decarbonization of the iron, a much larger quantity of cinder is charged than in case of making iron, and a much higher temperature is employed-the highest that can be obtained. The metal " comes to nature," or parts with its carbon sooner if the heat is kept com-aratively low. The cinder bath is composed largely of the slag from the boiling furnaces which are employed in making iron, and in which more cinder and a higher heat are employed than in the common puddling furnace. A solvent of manganese and its earthy bases, prepared in a manner which is not made public, for obvious reasons, is charged with the iron, and forms a part of the bath, all of which, under the high heat, melts as thin as water, and covers the molten mass of iron. The presence of the manganese is found to produce great uniformity in the product, and to prevent, in a considerable degree, the blistering of boiler-plate rolled from the blooms thus obtained. As in the ordinary puddling process, the operator breaks up the lumps of iron, turns them over to expose all parts to the heat, and when they become pasty, works them into puddle-balls. When 336 lbs. of iron are charged, five heats per day are made; with a charge of 280 lbs. six heats per day are made in each furnace. The time of stopping the process is decided by the operator, from the appearance and consistency of the mass, and with reference to the quantity of iron used. A longer time is required to work the pig metal into steel than into iron. The same time is required to convert 336 lbs. of pig-iron into steel, as would be necessary to work 448 lbs. into wrought iron, in.a boiling furnace. A higher heat being employed, more coal is consumed, a ton of steel requiring 30 cwt. and a ton of iron requiring but 16 cwt. The puddle balls are hammered two together into a slab, under a three-ton hammer; the slab is heated and hammered twice at a welding heat; at a fourth heat it is rolled into plate. The long celebrated Salisbury iron is exclusively used at this establishment for steel making, and its peculiar qualities promise a higher success for the product, than is likely to be obtained from almost any other variety of raw material. Salisbury is a neutral iron, being neither red-short nor cold-short, from the presence of either phosphorus or sulphur; it is naturally extremely tough, and stands a tensile strain equal, at least, to that borne by the best irons in the market. The Albany steel thus far proves more uniform than any of the puddled steels that have been imported. Its chief defect, in the shape of plates, is a tendency to blister in the rolls. A remedy for this, however, appears to be perfecting. All plates are tested by hammering their entire surface. A quarter to a third of them are found to have small blisters. Of those sent out, however, all have proved perfectly sound. A larger proportion of imported plates are found to blister. All puddled steel which is free from blisters, is more homogeneous and uniform than iron. 12. HoMOGENEOous STEEL. —This may be generally defined as simply cast-steel —wrought iron melted in a crucible, with carbonaceous matter —and sometimes with ingredients that combine with its impurities. These ingredients are as various as the natures of the irons under treatment. J. B. Howell's homo-metalj made by Shortridge & Howell, and extensively used for fire-plates, is made as follows:~Swedish iron is cut up into small bits, melted in a crucible with charcoal, 4 26 THE ECONOMICAL GENERATION OF STEAM. and cast into ingots which are rolled into plates. For every 40 lbs. of iron, 6 ounces of charcoal are used. The crucibles stand but two heats, and are then useless. The product is very soft steel. It cannot be filed if cooled ill the air at a white heat, and it hardens in water to a less degree than ordinary cast steel. But when annealed, it is as soft as iron, and at a red heat it works like copper. Mr. Howell patented the use of iron scales to be put into the crucible. They are not used in practice. If the contents of a large number of crucibles be poured together into a single mass or bloom, the plates drawn from it are more likely to be sound and proof against blistering than when made from several slabs with strata of oxide and cinder between. And the carbon being thoroughly fused with the iron, is more uniformly distributed than is possible in case of puddled steel. But the strength of this steel is not as great, on the average, a! that of puddled steel. 13. BESSEMER STEEL.-The Bessemer process of steel and iron making is too widely understood to require more than a passing notice. In it, however, appear to lie at least the germs of an improvement which will ultimately supersede all the roundabout and expensive methods now employed. The pig-metal is drawn from the smelting furnace into a converting vessel, through tuyeres in the sides of which air is injected. The oxygen of the air immediately combines with the carbon in the iron and with the silicumn, causing the whole mass to boil violently. In from five to six minutes, when the proper amount of air has entered (as measured by a meter, in cubic feet) to leave, say one per cent. of carbon in the iron, the process is stopped, and the mass poured into a souznd ingot of steel, which is worked up under the hammer or rolls. If iron is required, the decarbonizing process is still farther continued. For steel, 15 to 20 per cent., and for iron, 20 to 25 per cent. of material is lost. This, however, may be recovered. The steel thus made is not only homogeneous, but stronger than the average of puddled or cast-steel. It is believed that tre perfection of this process will lead to the manufacture of a material capable of resisting a strain of more than 100 tons to the square inch, and at a considerably less cost than that of wrought iron. TNo extensive trials of the Bessemer steel for boilers, have been made. The manufacture is mostly confined, for commercial reasons, to the higher classes of tool and cutlery steel, which pay better than boiler-plate. So far, the metal appears excellently adapted to fire-plates as well as to parts requiring strength merely. Its toughness a most important feature in boiler-plates, which are constantly expanding and contracting —has been proved by remarkable and extraordinary tests. The casting of very large ingots, and the consequent production of very large.plates at a cheap rate, will be of the highest advantage in boiler making-the joints now being the most imperfect parts of the structure. The Lowmoor Iron Company demanded ~22 per ton for plates weighing 2. cwt. each, but if the weight was increased to 5 cwt., ~37 per ton were demanded. By the Bessemer process, plates of 10 to 20 cwt. are made with less waste of material, and at a less general expense, than smaller plates. 14. ADVANTAGES OF STEEL.-The advantages of steel for all parts of boilers except fire-plates, are too evident to require special discussion. The leading fact is the superior strength of steel. The result of 150 recent experiments by Messrs. Robert Napier & Sons, on the standard kinds of iron plates, showed a mean strength of 49,215 lbs. to the square inch, while 80 experiments made by the same engineers, at the same time, on steel plates, showed a mean strength of 85,275 lbs. The difference in the weight of iron and steel plates of the same dimensions, isnot great enough to be of practical importance. Other things being equal, therefore, a steel boiler is 73 per cent. stronger than an iron boiler. Or the weight of the steel boiler may be decreased in the same proportion, without reducing its strength below that of the iron boiler, or a smaller increase of strength and decrease of weight maybe obtained by adopting a medium thickness of plate. Steel plates of the thicknesses commonly used for iron, are frequently employed in prac* Bars, 3 inches square, have been bent cold under the hammer into a close fold, without the smallest perceptible rupture. A difference in length of 94 inches was thus made by compression on one side and extension on the other, as compared with the formerly parallel sides of the bar. MATERIALS. 27 tically testing the new boiler material, and it is as frequently asked, " What is the advantage of stronger material, if we cannot use less of it? " Without discussing the policy of being on the safe side, especially in new enterprises, it may be answered that the economy of a higher steam pressure and an earlier cut-off, or the occasional safe increase of pressure in emergencies, on the one hand, and the reduction of repair expenses on the other hand; or, to some extent, both these advantages, will result from the use of stronger boilers. The history of steel boiler-plate is too short to throw much light on the subject of repairs, save that it has required no repairs during its limited service of three or four years. With the exception of fire-plates, the conditions of steel in boilers are so similar to its conditions under other uses, that we may safely predict for it that superior resistance to strain, oxidation, crystallization and general "wear and tear," which have characterized it under other circumstances. The forces tending to destroy locomotive boilers are, first and chiefly, the steam pressure; second, the sudden and excessive strains caused by the lurching and pitching of the locomotive, the boiler being the backbone of the whole structure, stiffening the frame and ultimately receiving all strains not neutralized by the springs; third, the continuous jarring of the vehicle; fourth, the chemical action of water, or oxidation; and fifth, the mechanical action of water and its impurities. To these may be added the weakening of the material —first, its crystallization by reason of the strain and jarring, an occasional result, and second, the " fatigue of the metal," * or the gradual loss of strength under permanent strain. It is not proposed to discuss these latter conditions, since nothing mathematically definite can be said about them. Explosions'cannot be definitely traced to these causes, since the original quality of the metal is rarely known; old boilers are usually removed and broken up, not because the iron is weak, for it is often up to the average strength,j but because they have so worn and decomposed and blistered as to be too thin for service. O1 the contrary, many striking instances have been adduced to establish these causes of failure. But since there is no authenticated remedy+ for these defects in iron, if they are really serious and universal, the employment of steel is the most promising method of avoiding them. Steel is already crystalline in structure, and therefore cannot become weak by becoming crystalline. General practice has not revealed the fact that steel becomes weak from continued use. But general practice has so few definite statistics to show, that nothing besides the more notorious peculiarities of metals is absolutely known. If steel is like iron, in this respect, however, its greater positive strength will the longer resist the strains which cause its deterioration. To return to the more definite causes of failure, their enumeration at once shows the superiority of sKeel. It resists steam pressure better than iron, by reason simply of its strength-that distinct quality shown by the testing machine. It resists the sudden strains and the jarring incident to locomotive boilers, by superior elasticity. It better resists the attrition of the circulating water and its impurities, ~ by its superior hardness. That steel corrodes less rapidly than iron, is a notorious fact-how much less in case of boiler steel is not known. It is evident then, that in case no more steam pressure is carried, the repair expenses of steel boilers, as compared with iron of equal section, will be decreased, not only in proportion to their superior strengths but in a greater proportion, by reason of their elasticity, hardness, granular construction, and resistance to corrosion. And if proportionately higher steam pressure is carried, so that the relation of strength to strain is the same as in iron boilers, the repair expenses will still be decreased by reason of the last-named qualities of steel. What is true as to the expenses of maintenance is true as to safety. Recent discussions and recently compiled facts on the subject of boiler explosions, show quite conclusively that the larger proportion of these casualties result simply from the want of proper strength in the boiler. Steel, like iron, is liable to one very serious defect, which, however, improved manufacturing * See Mr. F. Braithwaite's paper on this subject, before the Inst. of Civil Engineers. t Mr. S. J. Iayes, while Locomotive Superintendent of the Baltimore &; Ohio R. R., found the strength of the plates of an exploded boiler which had run 15 years, to be 60,000 lbs. per square inch.; Iron is believed to return from a crystalline to a fibrous state by being annealed; axles, for instance, may be so treated, but to anneal a boiler without destroying it, is evidently impossible. ~ The scraping and chipping of boilers to remove scale, so long as the wretched practice of allowing scale to collect is tolerated, will be less destructive in case of a harder metal. 28 THE ECONOMICAL GENERATION OF STEAM. processes-the perfection of the Bessemer process-will be likely to remove. It is not uniform; its strength can never be predicted from the quality of the raw material. If the nominal tensile strength of a certain quality be called 80,000 lbs., a large margin of extra cost and weight must be left for plates that will only stand 70,000, for there are sure to be such in every large quantity made. According to the experiments of Messrs. Robert Napier & Sons, before mentioned, the greatest difference between the highest and lowest strength of the same make of iron plates was 20,215 lbs.-of puddled steel ship plates, 36,370 lbs., and of Howell's homogeneous steel 23,250 lbs. The range of strength of 150 specimens of iron plates, was from 32,450 to 62,544 lbs. Mr. Fairbairn found the strength of a broken plate from an exploded boiler, to'be only 10,438 lbs. per square inch, or but one-fifth the proper average. The greatest average difference of the same plates, with and across the grain, by Messrs. Napier's experiments, was, for iron, 7,086 lbs., for puddled steel, 22,621 lbs., and for homogeneous steel, 790 lbs. The greatest actual difference between the highest strength lengthwise and the lowest crosswise, was, for iron, 20,643 lbs., for puddled steel, 25,621 lbs., and for homogeneous steel, 24,689 lbs. per square inch. These considerations especially refer to all boiler plates except those in contact with the fire. It has been before observed that strength is not so essential a feature of fire-plates as homogeneity and soundness. The advantages of iron, in comparison with copper, apply in a greater degree to steel, while its defects are modified. The puddled or semi-steel plates that do not blister in the rolls, do not so far appear to blister in service. Whatever may be the want of uniformity in large numbers of plates, as to tensile strengthl, those that do not show defects in manufacture and working, are pronounced more homogeneous and uniform than the best iron plate. The testimony of boiler-makers in this direction, is singularly harmonious. That there may be imperfect welds and incipient blisters between the strata of steel plate, is probable from the fact that so many blisters occur while it is passing through the rolls. Steel would be used thinner than iron, however, and would conduct heat more rapidly, so that the blisters would not be so likely to develop themselves. The cast-steel, before-mentioned, possesses most of the advantages of copper for fire-plates, in addition to advantages of its own, rwhile it avoids the excessive weight, softness and extra cost of copper. Being rolled from a single ingot, the plates have no laminae, nor strata of oxide or cinder, nor other causes of blistering, but are solid and homogeneous. Their service, thus far, is extremely satisfactory. Mr. Alexander Allan, locomotive superintendent of the Scottish Central Railway, one of the earliest users of steel fire-plates, has five fire boxes of Shortridge & Howell's metal, and is constructing several more. He says of the plates that they are " as uniform, mild, and ductile as copper, and much easier worked," and that they do not harden with hot water or get brittle, after three years' service. This material is quite generally regarded with equal favor, partly because it has not as yet shown symptoms of failure, partly because it is so easily worked; and chiefly on account of the superiority of steel as demonstrated in general practice, in default of the special facts which another ten years' experience alone can furnish. The greater strength and elasticity of steel plates is likely to prevent their cracking, although ~ood circulation would be the best preventive of this defect, in any case. The stiffness of steel allows the use of much lighter plates than would be admissible in case of copper or even of iron. The same distance between the stay-bolts being preserved, {-inch steel is considered suffi.ciently stiff. It has been remarked that iron has taken the place of copper on the Reading Railway, chiefly because the softer material was cut away by the flying particles of anthracite coal. Steel is evidently superior to iron in this respect. What has been said as to the inferior corrosion of steel, applies with equal force to fire-plates. As to conducting power, the most homogeneous and close grained metals are always the best conductors. In practice, this is never made a test question, durability being the chief desideratum; and the use of a good conductor covered with scale has been before mentioned. It has been noticed on several English lines, that steel decomposes less rapidly than iron, under the action of sulphur in coal. Should this prove true in longer practice, it will render steel fireplates indispensable in some localities. MATERIALS. 29 Steel welds as readily as iron, and will therefore be equally adapted to that most desirable, but not yet fully developed improvement in boiler making-dispensing with rivets.. Careful experiments show that the comparative strength of joints increases as the thicikness of the plates decreases' so that the joints of thin steel, if riveted, would be stronger than those of iron of the same total strength of plate. That all steel boiler-plates work in the boiler shop "like copper," and at a lower heat than would be required for iron, is universally admitted. They have been submitted to severe tests by being repeatedly doubled together in opposite directions, and are invariably found tougher and stronger than iron. There are few locomotive superintendents in England, if any, who are not already using at least one test plate of homogeneous or of puddled steel. The puddled steel is more frequently unsound, but when sound is often pronounced equal to the homogeneous in every respect. The highest strength per square inch recorded in Messrs. Robert Napier & Sons' experiments, was of the Mersey Co.'s puddled steel-108,906 lbs. The next was of Howell's homogeneous steel, 108,900 lbs. The average of Howell's first quality was 96,675 lbs., while the average of the Mersey Co.'s was 93,209 lbs. The highest strength of the Albany semi-steel is 104,000 lbs. It is guaranteed to stand a strain of 90,000 lbs. In this country, the practical test of the Albany semi-steel has been quite largely commenced. But both here and in England, the use of steel is confined almost exclusively to fire-plates. The author is not aware that any one but Mr. Eaton, Locomotive Superintendent of the Great Western Railway of Canada, has made an entire locomotive boiler of steel. However the new material may stand the great test of fire, it is quite unnecessary to postpone its known advantages for the outside plates of boilers. It is certain that too much importance is attached to the first cost of steel plate. Too many counting,-room engineers are ready to reject it altogether, if it is not as cheap as ordinary iron. Fortunately, the Albany semi-steel and the English puddled steels are as cheap for ordinary sizes as first-class iron plate, the present price being 83 cents per pound. Homogeneous steel, selling in England for about 9 cents, is held at 14 cents here. But a greater demand would, of course, reduce the price. If no more strength of boiler is required, and steel plates are made lighter in proportion to their strength, the average cost of a semi-steel boiler, with a cast-steel fire-box, would not exceed the cost of a good iron boiler. And if the same thickness of plates is preserved, the increased strength, allowing high steam pressure, and promoting safety, is fully worth the additional cost, not to speak of the other advantages of steel which have been already mentioned. The treatment of steel plates and rivets will be noticed in another chapter. 15. CONCLUSION AS TO MATERIALS.-In conclusion, after reviewing the conditions of locomotive boilers, and the properties of the materials used in their construction, and in view of the daily improvement in the manufacture of steel, and of the changes mentioned in boiler construction (which will be farther considered), it appears extremely probable that this material will gradually come into exclusive service, not only increasing the safety and decreasing the repair expenses of boilers, but promoting the economy of steam generation and of railway working generally. * Mr. James Tucker, master smith at the Washington navy yard, in a report on the official test of the Albany steel, says: " This material was found to bear a heat amply sufficient to enable us to weld it without the aid of a flux." t See section on Boiler Joints. CHAPTER III. DIMENSIONS OF FIRE-PLATES. 1. THICKNESS.-The practice of different builders and in different countries, in this respect, is very variable. For a 48-inch shell, intended for a pressure of 120 lbs. to the square inch, and sometimes carrying 160 Ibs., the plates of American boilers are from I to 1y6 inch thick; of French boilers, 15 millimeters, or, inch (nearly -} inch), and of English boilers, from 3 to - inch. The thickness varies somewhat in proportion to the strength of the iron used in these countries, the American iron being the strongest. 2. ADVANTAGES OF THIN PLATES.~-AS far as resistance to pressure is concerned, all the solid parts of plates have a large excess of strength over riveted joints. This subject will be fully considered in another section. It is stated by Clark,* on the basis of Brunel's experiments, and recent experiments at Woolwich, on riveted joints, that 8-inch riveted plates are practically as strong as -1 or I-inch riveted plates. This is the case with longitudinal seams, but the conditions of circular seams are different, and the strength increases with the thickness of the plate. Fig. 37, Plate 64, shows the ultimate effect of oblique strains on lap joints. The plates being originally parallel, as in fig. 34, would tend to assume the shape shown in fig. 37. But in case of a " telescope" joint —one tube entering another-the edges of the inner tube would have to be compressed, and those of the outer tube expanded, in order to produce the result shown in fig. 37. The circular joints, however, receive only one-half the strain of the steam pressure borne by the longitudinal joints, and they are further-strengthened by the flues. The strength of cylindrical parts of boilers, then, is primarily due to the strength of the joints, and this is due more to the quality than to the thickness of plates. Thin iron is likely to be more thoroughly compressed in the rolls, and therefore more pure, sound and strong than thick iron. A l -inch bar, having 56 per cent. more section, has only 40 per cent. more strength than a 1-inch bar, while iron that will stand 30 tons in inch bars, has a strength of 40 tons in wire, The same is believed to be true of plate. The thickness of fire-plates, with reference to their durability, it being for the present assumed that their resistance to steam pressure is provided for, has an important bearing on the first cost and expenses of maintaining boilers. No definite experiments have been made in this direction, but the general fact that ]-inch Lowmoor and other iron fire-plates are successfully used in America, while i-inch Lowmoor plates fail after a few months' use in England, plainly indicates the advantage of thin plates. But the principles involved are more conclusive. Fire-plates fail chiefly from blistering and cracking; their wear and decomposition are of very little comparative account. Blistering and cracking occur only by reason of the intense heat of the fire. If this intense heat is conducted to the water as fast as it comes in contact with the plate, the temperature of no part of the plate will exceed that of the water, which at 120 lbs. pressure is 350~. Now, the outside fire-box plates, in contact with the same water, do not blister nor crack, and iron or steel heated over a fire to this temperature have never been known to exhibit these symptoms of failure. It therefore appears that the plates fail because they do not conduct heat with sufficient rapidity. If the plates were infinitely thin, they would conduct heat infinitely fast. On the contrary, a piece of iron two feet long may be white hot at one end, and not hot enough to * "Recent practice." . MATERIALS. 31 make steam at the other end. An approximation to these extremes gives proportionate results. Water may be boiled in a piece of paper, over as intense a heat as the flame in a fire-box, but the inflammable paper is not destroyed-it conducts the heat practically as fast as it receives it. But a quarter-inch iron plate is so much hotter on the fire side than on the water side, that it is often injured, while a half-inch iron plate is soon destroyed. Joints, or a double thickness of metal, exposed to the most intense heat of the fire, give way sooner than thinner parts, and stay and rivet heads burn off for the same reason. The durability of fire-plates, then, must be chiefly in proportion to their thinness. If the necessary strength can be preserved, ~ inch thickness, or less even, would be desirable. STAYING THIN FIRE-BOX PLATES.-The strength of fire-boxes and the flat parts of boilers is dependent chiefly on the strength of the stay-bolts and on their hold upon the plates. The latter condition is not necessarily dependent on the thickness of the iron, although stays cannot be screwed into very thin plates. A plan employed by Mr. James Millholland, Locomotive Superintendent of the Reading Railway, for fastening stay-bolts into thin iron, is shown by fig. 17 of plate 65. The hole is made smaller than the bolt, and the metal heated and upset with a drift, so as to leave a collar on the inside, which adds to the length of the hold of the screw-bolt. The fire side of the bolt is riveted over, is nearly flush with the plate, and will not rapidly burn away. Fig. 16, plate 65, shows a simpler method-a washer, through which the bolt is screwed, as well as through the sheet. Mr. Betts, boiler-maker of the Jersey City Locomotive Works, uses a tubular stay (fig. 19), around which, abutting against the inside of the opposite plates, is a thimble, to increase the stiffness and to prevent the sheets from bulging inward awhile the tube is setting. Tubular stays, set by Prosser's expander (fig. 22, plate 64), and making the joint shown at fig. 8, plate 65, prevent the necessity of the thimble. Short pieces of old iron flues, cut out of the sound parts which would otherwise be useless, may be employed for this purpose. The flues, holding together the flue-sheets of a boiler, are examples of hollow stays. The sectional area of a -inch stay is.4417; the sectional area of a 2-incil tube, ~ inch thick, is.7364, while that of a'-inch tube, I inch thick, is.5400. Thus an excess of strength may be readily left for the. flange, should this be required. If the flange is carefully turned, its strength will not be materif ally reduced. Short tubes may be readily heated and nearly expanded while hot, which will prevent any extraordinary strain on the iron. A 2-inch iron tube has been expanded to a diameter of 2~ inches, while cold, without visibly injuring the structure of the metal. Such tubes are not likely to be injured in setting. Copper stay-bolts of J inch thickness are commonly used in England, for copper fire-boxes. If the tensile strength of such bolts is sufficient, that of 2 inch copper tubes would be much greater, and copper tubes would stand flanging if iron could not be trusted. But steel tubing promises the best results for hollow stays, especially since it can be worked while hot. Hollow stays of 11 inch diameter may be placed { inch further apart than solid stays of I inch diameter, so that practically, no useful heating surface or circulating room will be abstracted by large stays. Besides, the hollow stays themselves afford a considerable heating surface. Thus tubular stays would appear to be not only practicable, but cheaper than solid stays, since they require no screw cutting, but are rapidly put in with an expander. Mr. Beattie, Locomotive Superintendent of the London and South-western Railway, has for some time employed hollow screw stays in his fire-boxes, with plates of ordinary thickness. Tubular stays near the surface of the fire, are quite necessary to promote the combustion of bituminous coal. (See Clark's jet, plate 56.) The ends of those not required for air tubes, may be readily stopped by discs of metal. A little leakage of air will do no damage. It may be remarked that stays of large diameter will be more likely to break from repeated bending, caused by the superior expansion and contraction of the inside plates. This would be true in case the plates were too thick. There can be no practical difference in the temperature of the outer and inner plates, if the circulation is good and the heat of the fire is rapidly conducted to the water. Besides, thin plates would be, in themselves, a sort of expansion joint, and would yield to an excess of temperature between the stays. If plates are made too thin to be caulked in the usual way, a strip of copper between them will make a tight joint. This method is efrpont, -, employed. 32 THE ECONOMICAL GENERATION OF STEAM. But the disadvantage of thin fire-plates, as commonly constructed, is their want of stiffness between the stay-bolts. Not a mere tensile strain, but both a tensile and a compressive strain, as in the case of a rail or girder, are brought to bear upon them. The part of a plate lying between two stay-bolts is a beam or girder, of which the stiffness is as the cube of the thickness. A piece of iron 4 inches long, an inch wide, and -J- inch thick, which one could bend with his hands, has a tensile strength of 3,000 lbs. The stiffness of steel is far superior to that of iron, but it is quite insufficient for these purposes. Therefore, if plates are to be flat, the stays must be very near together. But this would impede circulation and promote incrustation, while the small conducting power of the large amount of -metal thus employed, is the very thing designed to be avoided. Cracking usually begins next to the stays for this reason. 4. CORRUGATED PLATES.-The only feasible plan of using very thin fire-plates, therefore, appears to be such an arrangement of the metal that its tensile strength shall come into use, and this may be accomplished by employing corrugated, or cupped fire-plates. If these cups or corrugations are respectively halves of spheres, or halves of circles in section, and project into the fire, the tensile strength of the metal alone resists the steam pressure; if they project into the waterspace, their resistance to compression is alone brought into service-the stability of the arch is employed. And their resistance to the steam pressure is in proportion to their approach to these forms. The cup-surface boiler of Mr. Barran, shown by fig. 5 of plate 56, to some extent illustrates this principle; the cups are, in some cases, separately formed and riveted into the plate, for the sim.pie purpose of increasing the heating surface; and a uniform thinness of metal is not preserved. Corrugated plates have been used on the Great Western Railway of England, for bridges or midfeathers.* But it would be almost impossible to join wholly corrugated fire-box plates, meeting each other at right angles. Vertical corrugations, terminating in a flat plate above and below the more intense heat of the fire, the stay-bolts entering the iron in the hollows and placed at the usual distances, without complication or farther deviation from the common practice, would probably allow the manifold advantages of a thin fire-box. (Plate 65, fig. 4.) The tensile strength of the metal would give it even greater resistance than that of flat plates; the stays being farther removed from the fire would not rapidly burn out; the greatest mass of metal would be farthest removed from the fire; the rigidity and stiffness of the fire-box, as a whole, would be enhanced; the width of water-space would be greatest where there was most heat-indeed there would be ample room for a distinct up current of steam and hot water, and for a downward current of cooler water; the heating surface would be materially increased (the increase of heating surface due to the cups, in Barran's boilers, is 58 per cent.); a less number of horizontal rows of stays would be required on account of the stiffness due to the corrugations; a larger amount of the heat of the fire would be conducted to the water; the metal would expand, if necessary, between the corrugations, preserving the stays; and the plates would last very much longer than the thick plates now employed. The metal would be thus placed in the shape to which the steam pressure would tend to force it, and its tensile strength would greatly assist its stiffness, in resisting the outward strain. The hollow stays could be set by an expander in the ordinary manner. It is probable that the plan shown by fig. 4, plate 65, would be better in this case, since the stay and not the plate would receive any irnjury firom the fire, while the whole would be relmoved from its most intense action. There is no practical difficulty in so corrugating thin steel plates at a red heat plates which are invariably said to stretch and flange as readily as copper, and which, in experiments, are frequently corrugated, cupped, doubled at short corners and much more severely'treated without injury, than would benecessary for the purpose mentioned. After the metal had been rolled to nearly the required gauge, it would be passed hot through several sets of rolls with increasing corrugations along a part of their surfaces. The flat edges of the sheets would then be trimmed in the ordinary manner. The gusset and back sheets of every boiler that is made without angle iron (all American boilers are so made), are subjected, in flanging, to a far greater strain than this corrugation would cause, generally without injury. The strong probability of preserving fire-plates, * These are illustrated in Clark's " Railway Machinery." DIMENSIONS OF FIRE-PLATES. 33 decreasing repairs and increasing the useful effect of the heat, would render some experiments in this direction reasonable and timely. The conditions of the service of the outside shell of the fire-box, are essentially different. Clark" very truly remarks: " A boiler may be abundantly strong but insufficiently stiff; whereas, in a locomotive boiler, above all others, identity of form is of great importance, as, besides the ordinary contingencies of overstrained joints and leakage, resulting from change of form, there are, unavoidably, connections and attachments to be made here and there, which can only be maintained in good order under superior conditions of stability of parts." But it does not follow, as farther implied by Clark, that the fire-plates must be thick, as well as the outer plates, since no attachments of framing, etc., are made to them, while an extra thickness of outer shell would afford a superior hold for such attachments, besides preserving all the required rigidity of the boiler. It is probable that many American fire-boxes are defective in this respect-the outer plates are so thin as to promote leakage and fracture. It is customary on many English lines, to rivet on an extra thickness of plate where frame-stays are attached to the boiler. All attempts to increase the life of fire-plates, however, by better materials and construction, are secondary to, if not useless without better circulation of water and less incrustation of heating surfaces, than are common in the present practice. 5. AREA OF PLATES.-There would be a very great advantage in using much larger plates than are at present made, chiefly to reduce the number and to save the cost of joints, which are at once the weakest and the most expensive parts of the structure. A double thickness of plates at the joints, hooping the cylinder and bracketing the flat parts of the boiler, does not add any useful stiffness to it, since the parts where seams occur are already stiff enough, by reason of their shape. The lap of plates, irrespective of loss of strength by riveting, is an element of weakness. Joints in the fire-plates so rapidly burn out, that they are avoided at all parts, except at the corners of the fire-box, by the use of the largest plates that can be obtained. The leakage and the difficulty and extra cost of making attachments, due to numerous joints, and indeed every practical consideration, favor the use of large plates. It may be said that one joint (or weak part) is practically as bad as a dozen joints, since the same limit of pressure must be preserved in either case. But it must be remembered that double the money could be expended on half the joints without increasing the total cost of the boiler. If a single plate large enough for the cylinder of a boiler could be obtained, the expense of welding it into a single solid tube, would not be greater than that of the present riveting, even at the present cost of the former process. A cheap process of weldingt would render it practicable to make the entire boiler into a single plate. But the excessive cost of manufacturing large plates, and indeed the impossibility of making them large enough by the processes in use, render any great deviation from present practice out of the question. It has been already remarked that the Lowmoor Iron Company demanded 22 per ton for, plates weiglhing 2- cwt., and ~37 per ton for plates weighing 5 cwt., or about double the size. Very large blooms, from which such plates would be rolled, being piled fromt smaller blooms and these from puddle-balls, would contain a greater number of strata of oxide cinder and incipient blisters, in proportionto the number of original pieces. Larger puddle-balls could not be conveniently worked together or handled in the furnace. But there is no such objection in case of cast-steel slabs, which are entirely homogeneous, however large. Heavier machinery would be required to roll such blooms, but this would be equally necessary in case of large cast-steel slabs, and is not impracticable. Mr. Bessemer proposes to roll large and small plates at the same rate per pound. It has also been suggested that cast steel may be passed directly from the Bessemer converting vessel into the rolls, and that steel sheets may be made by the furlong, like paper. It certainly appears reasonable that very large masses can only be made sound and homogeneous from the liquid metal, that is to say, from steel. Another advantage of steel, in this connection, is the fact that a slab or bloom of given weight can be flattened into nearly twice the area of plate, that would be equally strong in case of iron. * "Recent Practice." t See chapter on Boiler Joints. 5 CHAPTER IV. BOILER JOINTS AND CONSTRUCTION. THE single-riveted lap-joints by which the plates of boilers are commonly fastened together, have but 56 per cent. of the strength of the entire plate. Thus not much above one-half the steam pressure which the plates might safely resist, can be carried; while the cost of a proportionate amount of worse than useless dead-weight is incurred. On the other hand, the scarf-welded joint possesses the total strength of the entire plate. Welding boilers, however, by the present methods, is an expensive, and to some extent an uncertain process, although its great importance has led to extensive and promising experiments in this direction. Such are the leading facts about boiler joints; but it will appear that considerable improvements in construction may be at once and safely introduced into ordinary practice, with immediately profitable results. 1. STRENGTH or WELT JOINTS.-AS early as 1838, Mr. Fairbairn found, by experiments, that the tensile strength of the single-riveted joint was 56 per cent., and of the double-riveted joint 70 per cent. of the entire plate. Some years since Mr. Brunel experimented on riveted-joints, on a large scale; the plates were 4 inch thick and 20 inches wide, butt-jointed, with welts or fish-plates 9 or 10 inches deep and 8 inch thick on both sides, a aat A, fig. 18, plate 64. The rivets were arranged thus' on each side of the joint in the plate. Two pairs of plates, precise duplicates of each other, were thus jointed with — inch rivets, 10 through each plate, or 20 in all. The first pair failed at 153 tons, the rivets in one plate having been cut through, and the plate extracted whole. The second pair failed at 164 tons, one of the plates being torn through the outer line of rivet-holes. Hence it was inferred that the strength of the rivets, on the whole, balanced that of the plates; the mean strength of the joint being 158 5 tons. The total sectional area of 10 — inch rivets is 3'75 square inches, and twice this area, or 7 5 square inches, is the amount of shorn surface, each rivet having been cut at two places. The fractured plate had 10 square inches of solid section, the section of the rivet-holes being less than this by 1'75 square tion.,iThe area of shorn surface of the rivets is thus square inch, or about 9 per cent. less than that of the punched section of the plate. The two other duplicates, the same as the first in all respects, except that they had 43-inch rivets, were also tested, and failed in each case through the outer line of rivet-holes in one of the plates, at the strain of 167 and 147 tons, respectively, or at a mean of 157 tons breaking weight. The total sectional area of 10 4-inch rivets was 4'5 square inches, which, doubled, amounts to 9 square inches of section to resist shearing. The sectional area of the plate, in the line of the rivet-holes, was 8 125 square inches, so that the area to resist shearing was 10 per cent. more than that of the punched plate. Practically, then, the two riveted joints made with -'' and 1-inch rivets, respectively, were found to be equally strong, as the mean breaking weights were respectively 158'5 and 157 tons. In the former, the sectional area of the rivets to resist shearing was 9 per cent. less than that of the punched plate; in the latter it was 10 per cent. more. It may be thence inferred that the maximum strength is obtained when the united sectional area of the rivets to resist shearing is equal to the sectional area of the plate through the line of the rivet-holes; or, in other words, when the united sectional area of simply the rivets is equal to half the sectional area of the punched plate. BOILER JOINTS AND CONSTRUCTION. 3b In the same series of experiments, a number of joints with 1-inch rivets and similar plates and fishes, arranged in a zigzag course ('...'. ), were tested. Four joints were broken with 158, 160, 161, and 168 tons, respectively, the mean being 162 tons. The fracture of the plate in one case followed the zigzag course of the rivets, and in two cases the rivets partly failed, showing, on the whole, well-balanced joints, and proving that a double row of rivets, alternating, or zigzag, makes a rather stonger joint than a double row of rivets in pairs. A third variety of joint, with the same size and quality of plate and fishing, was tested, having four rows of i-inch rivets, three in a row, as in plate 64, fig. 18. Two of these joints broke at 171 and 176 tons respectively, or at a mean of 173-5, proving this to be decidedly the strongest of the three kinds. The 12 rivets had 10.8 square inches of shearing section, and the plate 8-5 square inches of section in the line of the rivet-holes. There is no doubt that the reduction of the number of rivets in the row from 5 to 4, compensated by an extra row behind, accounts for the extra strength of this joint. Solid plates of the same quality were at the same time tested —the plates being E inch by 12 to 16 inches wide, and similarly broken-and their breaking weight varied from 19'4 to 22 tons per square inch, or 20'6 mean. The mean breaking weight of the zigzag riveted joint was 20 tons per square inch of section, its strength being 79 per cent. of that-of the solid plate; and of that with four rows of rivets, three in a row, was 20'4 tons, its strength being 84 per cent. of that of the solid plate, showing that the strength of the material was practically unimpaired by the punching, and that the strength of the well-balanced riveted joints was simply in proportion to the transverse sectional area of solid metal in the main line of rivets. The double-welt or a fishing plate on each side of the butt-joint, such as was employed in the experiments mentioned, doubles the effective resistance of rivets to shearing, as compared with their resistance in case of single welts or lap-joints; and with double the resistance, half the number of rivets would give the same strength to the joint, and would afford a large remainder of solid section of plate in the line of the rivet-holes. And it is probable that some increase of distance apart of the rivets, would farther add to the strength of the joint. In ordinary double-riveted lap-joints, 30 per cent. of the perforated section is occupied by rivets; in the double-welt joint there will not be, according to these experiments, more than 19 per cent. deducted for rivet-holes. Such joints (plate 64, fig. 18), for the longitudinal seams of the cylindrical parts of boilers —joints which require twice the strength of the circular seams-are quite practicable. As compared with the single-riveted joint, they are 24 per cent. stronger, since that has but 60 per cent. of the strength of the entire plate. 2. STRENGTH OF RIVETED AND WELDED JOINTS. —An extensive series of experiments with riveted and welded joints was not long since made at Woolwich. Good Staffordshire plates were carefully selected for the trials; of three thicknesses, ~ inch, X7 inch, and- inch, made up into specimens 4 inches broad and 24 inches long, the rivets being w-inch, at 2 inches pitch. The joints are illustrated by figs. 11 to 17, plate 64, as follows: fig. 11, single-riveted by hand; fig. 12, single-riveted by hand, snap-headed; fig. 13, single-riveted by machine; fig. 14, single-riveted with countersunk head; fig. 15, double-riveted, snap-headed; fig. 16, double-riveted, countersunk and snap-headed; fig. 17, double-riveted with single welt, countersunk and snap-headed. The welded joints were Bertram's scarf-weld, fig. 9, and shown as prepared and as finished, full size, by figs. 40 and 41; and Bertram's lap-weld, fig. 10, with lI-inch lap. Three specimens of each variety of joint, for each thickness of plate, were tested, and the results were averaged for each set of three specimens. The tensile strengths of the solid plates were tested with very uniform results, showing a breaking weight of 20 tons per square inch of section, for all the thicknesses.' The fractures occurred in nearly all cases in the line of rivet-holes in one of the riveted plates. In a few cases, the rivets were shorn across. The first table on the following page gives the results of these experiments. The second table, of the ultimate and working tensile strengths of boiler-plates and joints, deduced from experiments with C-inch plates, and the explanations following it, from Clark's " Recent Practice," will be valuable for reference: 36 THE ECONOMICAL GENERATION OF STEAM. TABLE NO. I. SHOWING THE RESULTS OF THE EXPERIMENTS ON RIVETED AND WELDED JOINTS. Ultimate Tensile Strength of Joint, that of the entire Plate being 100. Figure, Plate VARIETY OF JOINT. 64.. F -inch — inch 1-inch Mean Strength Plate. Plate. Plate. of 3 Thicknesses. Entire plate,....100 100 100 100 9 Scarf-welded joint,.. faulty 106 102 104 10 Lap-welded joint,...50 69 66 62 11 Single-riveted joint, by hand,. 40 50 60 50 12 Single-riveted joint by hand, snapheaded,.. 50 52 56 53 13 Single-riveted joint, by machine. 40 54 52 49 14 Single-riveted joint, with countersunk head,...... 44 50 52 49 15 Double-riveted-joint, snap-headed, 59 70 72 67 16- Double-riveted joint, countersunk and snap-headed,.. 53 72 69 65 17 Double-riveted joint, with single welt,.countersunk and snap-headed. 52 60 65 59 General averages of all the lap-joints, 48 60 62 57 TABLE NO. II.-SHOWING THE ULTIMATE AND WORKING STRENGTHS OF BOILER-PLATES AND JOINTS. TENSILE STRENGTH. Percentage of Tensile Quality of Plate and Nature of Joint. Strength, that Working Strength of the Ultimate Strength per Square inch of the entire entire Plate per Square Inch of the entire Section of Plate, being = 100. Section of Plate. taken at one-fifth of the Ultimate Strength. BEST LOWMOOR PLATE.. Per cent. Tons per sq. inch. Lbs. per sq. inch. Tons per sq. inch. Lbs. per sq. inch. 1. Entire plate,.... 100 25 56,000 5 11,200 2. Scarf-welded joint,... 100 25 56,000 5 11,200 3. Lap-welded joint,... 66 165 36,960 -3 7,392 4. Double-riveted double-welt joint, 80 20 44,800 4 8,960 5. Double-riveted single-welt joint,. 65 16-25 36,400 3-25 7,280 6. Double-riveted lap-joint,.. 72 18 40,320 3-6 8,064 7. Single-riveted lap-joint,.. 60 15 33,600 3 65720 BEST STAFFORDSHIRE PLATE. 1. Entire plate,... 100 20 44,800 4 8,960 2. Scarf-welded joint,... 100 20 44,800 4 8,960 3. Lap-welded joint,... 66 13 29,120. 26 5,824 4. Double-riveted double-welt joint, 80 16 35,840 3'2 7,168 5. Double-riveted single-welt joint,. 65 13 29,120 2'6 5,824 6. Double-riveted lap-joint,.. 72 14'5 32,480 2'9 6,496 7. Single-riveted lap.joint,... 60 12 26,880 2'4 5,376 NOTE. —1. For the strengths of the joints of American best plates, allow one-half more than.for best Staffordshire plates; for ordinary American plate, one-third more; and for cast-steel plate, double. 2. The contents of the table are correct for 1-inch plates, and for thinner plates; but they are altogether too high for thicker plates. In the order of strength the joints range thus: — 1. Scarf-welded joint,........... 100 2. Double-riveted double-welt joint,....... 80 per cent. 3. Double-riveted lap-joint,..........72 " 4. Lap-welded joint,........... 66 5. Double-riveted single-welt joint,........ 65 " 6. Single-riveted lap-joint,......... 60 " BOILER JOINTS AND CONSTRUCTION. 87 These percentages are to be accepted as for plates not more than - inch thick, with straight joints. But, with circular joints, they no doubt hold good for all thicknesses; excepting that the welded lap-joint would bear a much higher ratio, and rank next to the welded scarf-joint. In round numbers, the working strengths of best boiler-plates are thus:Yorkshire plates, per square inch of entire section,... 11,000 lbs. Staffordshire plates, do. do........ 9,000 " American plates, do. do........ 14,000 " Do. (ordinary), do. do........ 12,000 " Cast-steel plates, do. do........ 18,000 " In round numbers, the working strengths of joints are thus:Best Yorkshire. Ic st Sdtitordshirc. 1. Scarf-welded joint, per square inch of entire section,. 11,00. lbs. 9,000 lbs. 2. Double-riveted double-welt joint,... 9,000 " 7,000'" 3. Double riveted lap-joint,...... 8,000 " 6,500 " 4. Lap-welded joint,........ 7,400 6,000 " 5. Double-riveted single-welt joint,..... 7,3 6000 6, 6. Single-riveted lap-joint,....... 6,700 " 5,400 " 3. GENERAL INFERENCES.-The following important inferences may be drawn from this table: First, the scarf-wvelded joint is fully as strong! as the entire plate; and referring to the last column of the first table, which contains averages of the three thicknesses, selected for the trials, the lapwelded joint has I or 62 per cent. of the strength of the entire plate. On account of the shortness of the lap, however, which will be again referred to, this cannot be considered a fair trial of the lap-welded joint. The varieties of single-riveted joints average equally strong with each other, and they have only one-half the strength of the entire plate, excepting the snap-headed, which is rather stronger. Of the double-riveted joints, the ordinary lap is the strongest, having twothirds of the entire strength of the plate. The single-welt joint is the weakest, while the doublewelt (fig. 18), as tested by Brunel, had 84 per cent. of the entire strength of the plate. The trials show that there is no superiority in the strength of machine riveting over that done by hand, and that for — inch plates, it is inferior. The 1-inch lap-welded joint has two-thirds of the strength of the entire plate, and the ~-inch has only one-half, showing that the absolute strength of the two joints are the same. The lapweld is stronger than the single-riveted joint, but not so strong as the double-riveted, the relative percentages for — inch joints being 60, 66, 72. The lap-weld is strikingly weaker than the body of the plate, notwithstanding the solidity of the joint, which has not been weakened by rivet-holes, and still farther weakened by punching these holes. The weakness arises from the indirectness of the lap. The plate, though solid, is not straight. (Fig. 34, plate 64.) The same is true of the riveted joint shown by fig. 35. The ultimate effects of the oblique strains operating with a leverage, due to lap-joints, are shown by figs. 36 and 37. It appears from these facts that the lap is essentially an element of weakness, irrespective of the loss of strength by rivet-holes. The thicker the plate, the greater the leverage and transverse strain, so much so that the ~-inch lap-welded plate was not stronger than the 8-inch weld. On this principle the practically equal strength of a joint with countersunk rivets may be accounted for, as compared with one having external rivet-heads, although the former sustains a greater loss of strength from the greater content of the rivet-hole. The leverage is shorter, and may be measured from the cylindrical part in the line a a, fig. 38, plate 64, towards the inner side of the plate. The superiority of the double-riveted to the single-riveted joint, may be partly ascribed to the greater extent of the lap, in virtue of which, though the leverage may be the same, the angularity of the strain is diminished. And if the lap is quite long, its power of resisting distortion is increased. The lap in this case was short~-l1 inch hence it failed under a comparatively less strain. But the double-welt for riveted joints (fig. 18), and the scarf-weld, have no such objections. The experiments further show what has already been referred to in speaking of thin platesthat the - and 8-inch lap-jointed plates are equally strong, and that they are above one-fourth stronger-than for the — inch plate, relatively to the thickness of the plate, proving in fact that a boiler of 8-inch plates is as strong as one of — inch plates, if not stronger, when riveted in the ordinary way. 38 THE ECONOMICAL GENERATION OF STEAM. 4. GRIP OF RIVET-HEADS.-It has been contended that the grip of the rivet-heads on the plates supplies an important element of strength to the joints, by offering frictional resistance to the sliding of one plate on another. Mr. Edwin Clark found the friction due to the shrinkage of one 8-inch rivet, by riveting one plate between two or more plates, varying the number and thicknesses of the plates, to try the effect of length of rivet on the grip. The frictional resistances were as follows: Shank of the rivet 10-8 inch long-Frictional resistance 4- tons. do. 15-8 do. do. do. do. 52 do. do. 23-8 do. do. do. do. 8 do. The grip increased with the length of the rivet, but this would soon become excessive, since with 6 or 8 inch rivets, the heads are often drawn off and the rivet would always fail when the strain by shrinkage exceeded the tensile strength of the metal. But the strength of the rivets to resist shearing strain is manifestly decreased in the proportion of increase of grip, since the shrinkage tends to draw the particles of the iron asunder. Hence the total strength of the joint is not likely to be practically increased by the shrinkage of the rivet. 5. CAULKING. —The caulking of seams is probably a cause of weakness in riveted joints. Fig. 39, plate 64, is an exaggerated representation of the way in which this process tends to buckle and separate the plates and to strain the rivets. A very unnecessary strain is doubtless put upon the seams in the process of caulking. It is well known that a prick-punch or the smallest coldchisel, with a blunted edge,'under very light hammering, will stop leaks in seams. The steam pressure on the microscopic area of the little flange thus formed, is too slight to require much strength, and the rest of the seam will soon fill up with rust and sediment, if the plates are so stiff as not to absolutely open and shut between the rivets. There is, at least, no necessity for upsetting the whole edge with a broad caulking tool and a heavy hammer.* 6. FURROWS ADJACENT TO JOINTS.-It is well known that continuous furrows, apparently cut like a groove, close to the lap-joints, are generally found in boilers after some service, and that these furrows are often the cause of failure and sometimes of explosion.t Indeed it is usually observed that boilers first show defects and require repairs at these points. The primary cause of this defect is probably the alternate bending and unbending of the plates, due to the lap-joint, with the alternate getting up and letting down of steam, as illustrated by figs. 35 and 37, plate 64. As before mentioned, the lap-welded joint, of which the weld itself is perfectly sound, has but 62 per cent. of the strength of the entire plate, and no other cause than this is apparent. The same action, less in degree, but repeated daily, must gradually loosen the texture of the metal and lay it open to corrosion. Indeed, the corrosive action is so great on the corners of iron sheets, flanged while hot, in the usual manner, that a thin plate of copper is sometimes laid over the bend to exclude the water. (Fig. 5, plate 64.) Another active agency in this furrowing process is the unequal vibration of the plates at the joints, by reason of a greater thickness of metal.+ The fact that such results are not so frequently observed in America, where thin plates are used, as in England, where thick plate is employed, is another proof of the destructiveness of the oblique strain due to the lap. The relative vibration of the plate and the lap is manifestly unchanged by the thickness of plates. The galvanic action of iron and copper, in very impure water-a weak battery —doubtless aggravates this evil. Lastly, the chisels of " cheap" boiler makers, who trim the edges of lapped plates, are very likely to leave an unmistakable furrow. * See also the section on " The Pitch of Rivets." + As an example, the engine Vulcan, which exploded on the Buffalo & Erie Railway in 1856, after a very short service, gave way on the outer crown-sheet of the fire-box. The boiler had been built with unusual care, of first-class material, But a flaw 24 inches long was discovered in the plate, close to the riveted joint. $ " As riveted joints destroy the elastic homogeneousness of the boiler, the waves of expansion, contraction and vibration are arrested there by the greater rigidity of theriveted double thickness of metal, which tends to localize the fatigue sustained by the iron near these points, and it also appears to increase the susceptibility to corrosive action, since the furrow.s generally take the line of that fatigue, and are often deeper than the spots on thle plates." —Report of the Board oy' Trade on Railwzay Accidents, 1855. BOILER JOINTS AND CONSTRUCTION. 39 7. EFFECTS OF PUNCHING.-Punching rivet-hloles, according to Mr. Fairbairn's experiments, is in itself a cause of weakness. Not only is the section of the plate in the line of strain reduced by the area of the holes, but the plate between the holes is not so strong per square inch as the solid plate. The excessive strain of the punch appears to disturb the molecular arrangement of the metal, and to start fractures which, in case of stay-bolts, often radiate in every direction, allowing corrosion to take place, and ultimately causing the bolts to pull out of the plate. In eight experiments by Mr. Fairbairn, the highest strength of plate experimented upon was 61,579 lbs., and the lowest 43,805 lbs. per square inch; but with the same plates after punching, the strength per square inch varied between 45,743 Ibs. and 36,606 lbs. The average of the two experiments therefore showed a loss of 10,896 lbs. per square inch, due to the jar and strain of punching, in addition to the loss of section through the holes. Getting plates together by " drifts," when the holes do not coincide, aggravates the effects of punching. Reaming the holes is practised on the Caledonian and other lines, and is found worth the trouble. Mr. Scott Russell uses a plan of rivet illustrated by fig. 33, plate 64, except that the holes taper rather more than here shown. The die is considerably larger than the punch, and it is found that a regular cone is thus produced. It is probable that the iron is somewhat less weakened as the taper is increased, since there is less lateral strain on the plate. In fact all punched holes are more or less conical. 8. CONICAL AND. COUNTERSUNK RIVET-HOLES. —The advantage of the conical holes thus produced over drilled holes, which are cylindrical, without the extra cost of countersinking is undertaken, is that the rivet becomes a double cone, converging the lines of strain, and remedying the defect shown in fig. 37 of plate 64. The countersunk rivet-hole, fig. 32, tends to remedy the same defect in the same manner. In both these cases, however, there is a greater loss of plate between the rivets, than with cylindrical holes, for a given minimum sectional area of rivet. A conical or countersunk hole is likely to allow a sounder rivet-head, since the head has a greater depth of metal and does not join the shank at a right angle. Countersunk heads in the fire-box are useful; if the projection a (fig. 6, plate 64) burns off, there is still a head left. 9. HAMMER, MACHINE AND SNAP-RIVETING. iIt appeared from the experiments mentioned, that machine-riveting had no advantages as to strength; in general practice it is not found to have appreciable disadvantages. A rivet finished at a stroke, while hot, is likely to fill the hole better than one which is not fully completed till it begins to cool. But if the holes in two plates do not coincide, the rivet will fill both, and be partially cut off in the centre. Many boiler-makers prefer hand-riveting on this account. But the neck of the machine rivet is certainly sounder; hand rivet-heads are rarely finished till the iron is cool enough to crystallize or crack under the head by the heavy blows of the hammer, and if the material be not of superior quality, will frequently snap off under rough usage. The great advantage of machine-riveting is its quickness and cheapness. In Mr. Fairbairn's works it is employed for all accessible parts of locomotive boilers, 12 rivets being set by machine to 1 by hand. Snap-riveting is similar to machine-riveting in its results. A concave " swage" is placed over the projecting end of the rivet and struck several blows with a heavy hammer, the head being formed in the concavity. The process is rarely if ever used in American locomotive works, but is frequently employed in those of England.t The jarring and disturbance of the metal are greater than in the case of hammer-riveting, so much so that the latter process is preferred for rivets adjacent to such rivets already set as might be loosened by the concussion. Snap-riveting was proved stronger than hammer-riveting, by the experiments mentioned. It also has the advantage of quickness, without extra cost for machinery, 3 to 2 and often 3 to 1 being the time occupied, as compared with hammer-riveting. 10. PITCH OF RIVETS.-Rivets are commonly from 3 to i inch diameter, and are pitched at 14 to 2 inch centres. For 8 and ~ inch plates, f-inch rivets at 2 inch centres were found by trials * See also the section on the Treatment of Steel. t In those of Beyer, Peacock & Co., and of Sharp, Stewart & Co., for example. 40 THE ECONOMICAL GENERATION OF STEAM. to be as strong as the plates. Steel rivets, the material being stronger, may be placed farther apart, allowing less section of plate to be cut away by the holes. On the North London road t-inch steel rivets pitched at l-1 inches we're found to make a stronger joint than i-sinch iron rivets similarly pitched. But the thinner the plates, and the greater the pitch of the rivets, the less is the important feature of tightness secured to the joint, as ordinarily made.* Strips of copper are often placed between iron plates, the copper only being caulked, as in fig. 4, plate 58, which would remedy the defect, but it would also increase the diagonal strain as shown in fig. 37, plate 64. The true remedy would be to roll thick edges upon the plates, if welding is impracticable. 11. THICK-EDGED PLATES.-Many schemes have been proposed and several attempts made to produce such plates; a few have been used in England; but none are, as yet, regularly manufactured. By a process lately patented,+ it is proposed to form the original slab somewhat in the shape of an I rail, the thick parts being rolled into the edges of the plate. There appear to be some difficulties in te pres in the present state of te art, in roin rollingt uch plates, but should welding prove too expensive or unreliable,t the will doubtless be overcome. Were the projections rolled only on one side, the sheets would be lapped so that the flat parts of the plate would touch each other, the projections being under the heads of the rivets on either side. Thus the oblique strain would be due not to the thickness of the edge, but to the thickness of the body of the plate. The edges should be sufficiently thick, 1st, to allow so many rivets that their' total resistance to shearing would be equal to the tensile strength of the body f the body of the plate, and 2d, so that the tensile strength left after punching these rivet-holes should be equal to that of the rest of the plate. For I-inch steel plate, with a tensile strength of 80,000 lbs. to he esquare inch, and rivets 1 i inch apart centres, the diameter of the steel rivet thus required, estimating its resistance to shearing as practically equally equal to its tensile strength, per square inch, would be inc would be inch, and the edges of the plate would require to be to be inch thick, to make good the loss of section through the line of rivetholes. The oblique strain might be relieved in the manner already adopted on the Great Western Railway of England, and shown at fig. 36, plate 64, by bending the edges to the shape theyg swould tend to acquire. Thick edges would not be required for any parts of a boiler except the longitudinal joints of the cylinder. The circular joints are strong enough, with single riveting.4 They would indeed be useful where right-angled flanges were turned, if right-angled flanges were necessary. 12. WELT AND HOOPED JOINTS.-The double-welt joint (Fig. 18, Plate 64) having 82 per cent. of the strength of the plate, is probably sufficiently strong for practical purposes. The remaining 18 per cent. of the strength of the entire plate, is likely to-be taken away by attachments-stays, braces, etc. This joint, applied to the longitudinal seams of the cylinder of the boiler, and so constructed that the welts shall break joints with, or overlap the cylindrical section of the boiler, will serve another important purpose. This part of the boiler is, to a great extent, a girder-the back-bone of the whole fabric; leakage and failure due to strain, first show themselves at the bottom. Three thicknesses of plates, at the bottom and at the top, if sheets long enough to go entirely round, cannot be obtained, will constitute the flanges of a girder, and largely promote the stiffness and durability of the structure. The double-welt joint is only necessary at these longitudinal seams. 1; The joint shown by fig. 31, plate 64, as used on the London and South-western Railway, involves the principles of the double-welt. The rivets of the outer rows are 3 inches apart, and those of the middle row, 1 inch. On the Caledonian and other Railways, the joint shown at fig. 44, plate 64, is somewhat used. The plates are t-inch thick, and the ring 8 inch thick and 5 inches wide. The abutting * See also the section on Caulking. - f G-eo. Alton and Jno. Fernie, Eng. May 25, 1858. t See sections on Strength of the Boiler, etc., and on Welding Plates. ~ Mr. Richard Eaton, Loc. Supt. of the Great Western Railway of Canada, finds that the thin American boilers, althongh of better material, fail at this point sooner than the thicker English boilers. II See the latter part of tlie section on Welding Plates. BOILER JOINTS AND CONSTRUCTION. 41 edges of the plates are turned. With snap-riveting, the boiler costs $25 more than if the common lap-joint is used. Messrs. Sharp, Stewart & Co. make another style of hoop-joint. The entire cylinder is round, and of equal diameter, the plates abutting at the seams and having a longitudinal welt inside and flush countersunk rivets, the ring being slipped over the circular joint, and riveted on. Clark suggests that hoops should, in any case, be shrunk on at various intervals, to bind the boiler together, as similar hoops are shrunk on the cylinders of hydraulic presses. These hoops act like a series of abutments for the intermediate sections of the boiler, just as the fluesheet binds and strengthens the front end of the boiler. Hoops of the plan shown at fig. 8, plate 64, are sometimes used in large flues, to prevent collapse. They are not sufficient, of themselves, for combustion-chambers and other flat-topped flues, but may be used to assist stays, when the latter cannot be conveniently attached. They might be useful for staying the sides of the chamber shown at plate 49. 13. WELDING PLATES.-The extreme desirableness of a cheap and reliable process of welding boiler joints, has been fully appreciated, since the earliest introduction of the locomotive. Indeed, some of the first engines by Bury had welded joints. Exceptional boilers have been welded by a variety of processes; the longitudinal seams of the cylindrical part are frequently welded by Beyer, Peacock & Co., and other English builders; this is done with an ordinary blacksmith's fire and apparatus. In most of the railway boiler-shops where the process has been attempted, however, it has not been deemed sufficiently convenient, economical or reliable, to warrant its general adoption. The plates buckle, twist and cool, leaving unsound parts in the seam. To make a weld as strong as an equal section of solid metal-to raise long and thin sheets to a melting point without burning one part and unduly cooling another part-to prevent the oxidation of small and delicate edges as well as their injury by sulphur, and the consequent " cold shuts"-and to compress them while hot with an equable and instantaneous force, manifestly cannot be the mere modification of an old process; it must be, substantially, a new art, to be practised with new tools and apparatus. It is, therefore, unreasonable to contend, as many boiler-makers do, that because the wrong means have not been successful, no means can ever beemployed to weld plates with advantage, and that the scheme should be given up without farther waste of money. The difficulties that have been indicated are not insurmountable. A sufficient temperature call be maintained with hardly a degree of variation, and at the exact points required, by a simple modification of the blow-pipe; flame, by enveloping a substance, protects it against oxidation, as in case of the candle flame and wick —the flame of purified coal gas or of wood, contains no sulphur, and instantaneous pressure under a drop or steam-hammer, so arranged as to strike the moment the iron is removed from the flame or the flame from the iron, as the case may be-all these operations are simple, and for all parts of boiler work where they are required, can hardly be costly, since it fortunately occurs that the seams of such parts are, with hardly an exception, straight seams, easily got at and adjusted. Gas-welding was long since proposed by Mr. W. B. Adams, of London. The experiments of Mr. Bertram (whose process will be described) show that heavy hammering is not required to weld parts which fitted before heating; hammering is only necessary to bend the parts into contact-an operation which should precede the heating. All that smooth-fitting and clean surfaces require, after being raised to the welding heat, is a moderate pressure to secure contact, when welding occurs as a matter of course, as much so as when the contents of two crucibles are poured together into a mould. Mr. J. C. Gooke, of Middletown, Conn., U. S., has recently patented the employment of the oxy-hydrogen blow-pipe, in connection with a portable apparatus, consisting of hammers or rollers, to be clamped to the plates, and both the fire and compressing machinery to move along over the joint, completing the operation as they move. This process is intended chiefly for iron ship-building. The most promising plan of welding boiler-plates appears to be that of Mr. William Bertram,a of Woolwich, Eng. The patent specifies various modes of preparing the plates by lapping them * Patented Dec. 21, 1854. No. 2,692. 42 THE ECONOMICAL GENERATION OF STEAM. or by thickening them at the edges, and then scarfing them. Figs. 40 and 41, plate 64, show the scarf-joint before and after welding, as commonly made; very little compression of the plates is required. The edges to be welded are placed in contact between jets of flame issuing from two furnaces attached to cranes or cars, one on each side, after which the furnaces are removed, and the compression is done by hand-hammers, or by steam-hammers so fixed to the same or other cranes or cars, that they can be instantly brought into service. The following quotation from the specification, explains the process more in detail: " These forges, or furnaces, each consists of a vessel or chamber to contain fuel, having an opening on one side to receive the end of a blowing pipe, and on the opposite side, another opening, so that when the blast of air is sustained, and the opening has the sheets or parts resting before it, the flame and heat will be projected and concentrated against the surfaces of the iron, and as the other similar blast-furnace is held and used on the opposite side of the iron, the two pieces in contact (which are to be welded together) will become heated to a welding heat so as to be welded together, by hammermen striking opposite to each other, or the one pressing and the other striking; or, in place of hammermen, when required, I propose to employ two steam-hammers suitably arranged and supported to act opposite to each other, or one steam-hammer and an anvil to resist its action whilst welding the parts together." X * e " Each of these forges or furnaces consists of a closed vessel or chamber (to contain coke or"other fuel), lined with suitable fire-brick or material to stand the requisite heat." The relative strength of the scarf and the lap-joints has been referred to in a preceding table, that of the scarf-weld being equal to that of the entire plate. The author has examined a number of plates welded together by this process-the joints having been dissected in various ways-and has found them entirely solid. Very thin plates can be joined with only slight hammering. It has been remarked that most of the seams of locomotive boilers which require welding, are straight and accessible. Since the welded joint is practically twice as strong as the riveted joint, and since twice as much steam pressure is exerted on the longitudinal seams of the cylinder of a boiler as on its circular seams, the right proportion of strength would be preserved by welding the former and riveting the latter. The same is true of the wagon-top or outer crown-sheet. The flues, as stays, give superabundant strength to the flue-sheets; the back sheet of the boiler, above the crown of the furnace, requires exactly the same staying per square foot, as the fire-box, to prevent its bulging, however its joints may be made; and the fire-box plates have above twice as much strength at the seams as the stay-bolts give them at other points. In short, the whole fire-box has an excess of strength whenever it has rivets enough to keep it stiff and tight. When made as at fig. 4, plate 58, the abutting of the plates resists the steam pressure. The joint shown, however, is easily welded. The remaining joints are those of the gusset or part connecting the fire-box and the cylinder of the boiler. Those that are longitudinal can generally be welded, while a short double-welt joint would cost and weigh very little. As far as mere resistance to steam pressure is concerned, then, only the longitudinal joints of the cylindrical parts require that of the entire plate. Considering the boiler as a girder, however, the circular seams of the gusset, being nearly in the centre, are subjected to other and severe strains which frequently cause leakage and failure. They therefore require the strength of the entire plate. But welding all boiler seams would be of almost equal advantage, in various ways: 1st. It would cheapen the process of construction, by saving much of the time occupied in riveting, and all that consumed in caulking. The interest on the cost of machinery for dressing the plates, etc., in any large shop, could hardly offset this saving. The process is not yet far enough settled to estimate its economy closely. 2d. The full strength of plates being preserved, less boiler iron would be paid for and transported as dead weight; or 3d, double the pressure would be carried without increased weight of boiler. 4th. There would be no double thicknesses of plate to promote unequal vibrations. 5th. Where the greatest strain would occur, there would be no lap-joints and oblique strain. (Fig. 35, Plate 64.) 6th. There would be no leakage. Greater simplicity of construction would facilitate any improved method of making joints, especially welding, since applying this process to crooked and intricate joints would require a great variety of special tools. BOILER JOINTS AND CONSTRUCTION. 43 14. FLANGING PLATES.-EXPANSION JOINTS.-FIRE-BOX SEAMS.-It has already been re" marked, that trying to keep plates flat during the process of flanging, is likely to develop any incipient blisters. They should be allowed to buckle, if they tend to do so, and should then be heated before they are flattened. It is customary in this country to flange the back outer fire-box sheet to a radius of frolm 4 to 8 inches (see Plates 46, 50 and 51); the vertical corners of the outer shell of the fire-box, and the gusset, are similarly flanged, with a round corner. The chief object intended is to preserve the strength of the plate, which will not so readily crack as if bent at a short corner. But the metal is more stretched in the former case-4 to 6 inches are considered to be the proper limit. Indeed, very short corners in plates are more unsafe than angle-iron corners. Round corners form a sort of expansion-joint, similar to thtose placed in the steam-pipes of American river-boat engines, which adapts itself, without undue strain, to the alternations of temperature. The whole gusset of the boiler, if the diameter of the cylindrical part is less than the width of the fire-box, is a continuous expansion-joint, and visibly compensates for the superior expansion of the flues. A corrugation has for this purpose been put into the long crown-sheets of some combustion-chamber engines. But this mode of construction is expensive. The greatest possible simplicity is desirable, for the purpose of saving first cost, and to allow the welding of joints to be performed with facility. A large number of boilers have been made perfectly straight on the top, without raised outer crown-sheets. Although they have not been found to " carry water" as well as those with raised crown-sheets, no especial difference as to durability has been reported. The latter variety of boilers is more easily examined internally; but the chief object sought is a plenty of steam room. If drying the steam by a separate process* should prove feasible, it will hardly be necessary to continue this style of building, simply to facilitate expansion. And it is probable that no practical difference of expansion will occur, if the circulation of water is perfect, and if the shell and flues are of the same material. The effect of corrugated fire-plates, in regard to expansion, has already been mentioned. The bottom of the fire-box, when made as at fig. 4, plate 64, operates as an expansion joint.t It is not found best in practice to corrugate the inner plate quite as much as shown, since it requires larger plates, and because an intermediate ring can be fitted more easily to the plates than the latter to each other, and because it is difficult to get the inside box in place, if ~uilt as shown in fig. 4. A solid ring stiffens the whole structure, but this advantage is offset by its greater dead weight, and by the cost and trouble of long rivets. A plate ring, as at fig. 20, plate 64, both for the bottom of the fire-box and for fire-doors, is deemed sufficiently light, elastic, and durable to be worth its extra cost. It also offers the least obstruction to circulation. Fire-doors are made on the London and South-western Railway, as at fig. 21. The thickness of the metal exposed toan intense coal fire, and the length of the rivets, frequently cause leakage, if the door is made with a solid ring. The copper crown (fire) plates on the North London Railway are flanged to a radius of 14 inch, to prevent their cracking. The back vertical corners of the fire-box might be bent to a large radius, with advantage. But the flue-plate corners could not be so rounded, since this would prevent setting the flues near the edges of the plate. Its top corner is sometimes rounded to' a radius of 2 inches. A chance for this or the opposite flue-plate to give a little, is very desirable, to compensate for the superior expansion of the flues, but it would be better to let the forward or smoke-box plate give, since it would not be so severely tried by fire. In case of a round smoke-box, this might be done by increasing the diameter of the box and riveting the flueplate to it instead of to the boiler proper. In some English engines, the front flue-plate is not flanged, but is put in somewhat after the manner shown by fig. 2, plate 39~, which forms a firstrate expansion-joint. The joint of the fire-plates inside the water-space (fig. 4, plate 58) was in more general favor a few years since than at present. If it got to leaking there was no way to stop it but to take out the fire-box; putting in a new flue-sheet required the removal of the whole furnace. The caulking is done on the fire side of the furnaces of the udson River R. R. engines, by the use of * See chapter on Superheating. t See the section on the Strength of the Boiler, etc. 44 THE ECONOMICAL GENERATION OF STEAM. a strip of copper between the plates, as at fig. 4, plate 58. Its advantage is in removing the rivets from the fire. There is an advantage in having a boiler so constructed that the fire-box can be removed and new fire-plates put in, without disturbing all the attachments of the framing and engines to the cylindrical part.* These attachments cannot easily be adjusted again, therefore one circular seam of rivets might be left, in case all the rest of the boiler were welded. The gusset, however, being a very weak point, should be as strongly fastened as the longitudinal seams. Two or more hangers or long stay-bars are sometimes riveted in the water-legs, near the top of the outer plate and near the bottom of the inner plate, to sustain the weight of the inner firebox; the crown-stays obviously do not hold up the side plates firmly. These hangers have been found to prevent the leaking and cracking of the bottom joints of the legs, in case of very large and heavy fire-boxes and wide water-spaces. The crown and side-plates are made in one piece, -inch copper, with round top corners, in the Eastern Counties shops. The crown-plate and one half of each side plate are so made on some American engines, which is bad, since a riveted seam very near the fire is certain to leak and burn away. The lap-seams in the fire-boxes of M. W. Baldwin & Co.'s engines are trimmed close to the rivets to prevent cracking. Countersunk rivets will stand longest in the fire. Bolt-heads projecting 21 inches into the fire, soon burn down to 1 inch, as ascertained by Mr. Dewrance. Copper plates burn down much faster from ~ to t inch than from - to -, at which point they remain without much additional burning. Double-riveted joints and wide laps in the fire-box, are therefore worse than useless. In some English fire-boxes the copper plates are thus doubleriveted, the alternate holes of the inside row having stay-bolts instead of rivets. Patches 12 by 14 inches are riveted to the inside of many English boilers, where the framing is attached, to receive and distribute its strain on the boiler. The Dimpfel boiler (Plate 50) possesses various advantages as to durability. The furnace crown-plate is the tube-plate also, and is removed from the intensity of the fire, as the current of gases does not strike it. This plate is hollowed so as to bring its tensile strength into service. The water-tubes B bend independently where they curve upwards, while expanding or contracting, and do not strain themselves or any part of the boiler. The long crown-plate over the tubes, however, is not so good a feature. As a whole, the boiler stands as well as other boilers, and has advantages with reference to circulation and combustion. Its principle of operation is good, and its construction is likely to be improved. The Boardman boiler (Plate 49) having considerable advantages as to combustion, has so little durability that it has been removed from several engines, and is not likely to be further employed for locomotives. The flues, indeed, being short, and removed from the fire, stand well, but the box containing them, having flat sides and very long stay-bars, cannot be kept tight. The boiler is nearly all made up of stayed surfaces, and is expensive and heavy. And there being a very sluggish circulation in the tube-box, it soon filled up with deposit, and could not be cleaned. 15. ANGLE-IRON JOINTS. — These are never used in American boilers, and have been abandoned by the French. The English continue to use them, but not as much as formerly. It is probable that making a double row of rivets and holes does not cost more than flanging the sheets. But the latter is always likely to make the soundest job. The real resistance of angle-iron to longitudinal cracking can never be known, since the process of rolling it subjects the metal to the severest strain, induces longitudinal cracks, and gives it a reedy structure. 16. COMBUSTION CHAMBERS AND MID-FEATHERS.~ The complication of fire-boxes, and the increase of stayed surfaces, are always expensive, and likely to increase leakage and repairs. The * Mr. Betts, of the Jersey City Locomotive Works, cut off one circular seam of rivets from the boiler of an engine on the New Jersey R. R., removed the whole back end of the boiler to his shop, put in new fire-plates, and then joined the parts together again without removing any attachments except the expanision-joint on the framing, and without disturbing the wheels or other parts. BOILER JOINTS AND CONSTRUCTION. 45 hanging bridges D and F of Beattie's boiler (Plate 38), the bridges of the " Phlleger" boiler (Plate 47), and nearly all kinds of mid-feathers, are going out of use. The cross bridges, once so common in English fire-boxes, are not reproduced in new engines, and the long longitudinal midfeather, similar to that in McConnell's engine (Plate 44), formerly used on the Hudson River line, is now abandoned. Not a single combustion-chamber and but one mid-feather are shown in Clark's " Recent Practice." The small mid-feather A in Baldwin's engines (Plate 5'1) is not found particularly objectionable. There are two reasons why such structures are generally unnecessary: 1st. As has been stated, they are not durable; unequal expansion and contraction shows itself in the shape of leaks at the joints or weakest parts. 2d. They abstract much of the heat from bituminous coal, which could be better employed, at first, in setting fire to the mixed air and gases; their action is directly the reverse of that of fire-bricks, which retain the heat instead of lowering the general temperature of the furnace. It is not impossible that the importance of midfeathers to alternate firing, as in the case of Head's boiler, (Plate 441,) may justify their use. But they should then be made wide-Clark recommends not less than six inches-and provided with the most ample facilities for circulation. Mid-feathers extending to the crown-plate of the furnace undoubtedly support it, and the cross mid-feathers of the Great Western (Eng.) boilers, are said to be very useful as stays; they are often strongly recommended for this reason; but if this is all the good they do, one-half their cost would accomplish the object in the form of stay-rods. The movable bridge of the Norris boiler, (Plate 48,) together with the water-grate, is a decided improvement over the fixed bridge of the " Phleger " boiler. For anthracite coal, the larger the heat-abstracting surface; the better, while solid grates will not stand at all. The circulation is kept up in Norris's bridge by means of the grate, and the parts can expand and contract independently of the rest of the boiler. A large combustion-chamber is formed without extra costmerely continuing the rectangular fire-box-and no pendent fire-box is required, so that the wheels can be placed wherever the stability and easy traction of the vehicle may dictate.* On the whole, it is a cheap and durable arrangement. The waist of the boiler, however, might be improved. The plain combustion-chamber, though not so absolutely essential to good combustion as has been supposed, is still deemed sufficiently important to it and to other economical functions of the boiler, to warrant its application to new engines, the increased cost over the plain boiler being some $200.t The opinion of the best practitioners is, however, that it will not pay to put a combustion-chamber into an old wood-burling boiler for the purpose of burning coal, except in case the fire-plates and flues need renewal. The cost of such a change is rarely less than $1,000, and often exceeds $1,500, since the whole boiler must ill many cases be raised to allow room for the gusset. The manner of constructing the combustion-chamber in the Baldwin engines, as shiown on plate 51, is a very simple and substantial job. That employed in the Rogers engines (Plate 46) has several advantages. The "telescopic " or conical construction of the cylinder of the boiler prevents the necessity of any offset in the gusset, since the cylinder so widens as to embrace the flue-plate. The conical construction also gives more room for circulation, allows the sedimentary matter to run back into the legs, where it can be removed, and widens the mouth and increases the capacity of the combustion-chamber. The ring E of Beattie's boiler (Plate 38) forms an expansion joint which tends to prevent leakage. What has been said in a preceding chapter about thin and corrugated plates for fire-boxes, applies with equal force to combustion-chambers and mid-feathers. Such a method of construction would be as important to durability by providing for expansion and contraction,as by preventing the burning of plates. A combustion-chamber, built upon the inner fire-box, prevents putting the latter in place from beneath, in the manner that straight fire-boxes are put in. Therefore a part of the outer box must be left unfinished till the inner box and chamber are adjusted; which part depends on the particular shapes and dimensions of the work. It is generally found easier to drop in the box from above, * This is also the case with Allan's boiler, Scottish Central Railway. f See chapter on 0ombustion1 46 THE ECONOMICAL GENERATION OF STEAM. and then to put on the outer crown, than to put it in from behind and then fit and attach the back sheet. 17. STEAM-DOMES.-Although perforated pipes appear to carry dry steam to the cylinders,'when the perforations or slots are so proportioned that an equal draft is made upon the whole water surface, this simple method of taking steam has not been largely adopted here or in Europe. The steam-dome is preferred-indeed, two steam-domes are placed upon the best classes of American engines-such as the Rogers, the Baldwin, and the Jersey City Works engines. The best practice shows the necessity of a large amount of steam-room; but as has been before remarked, and will be further considered, superheating, or at least drying the steam, after it.is taken'up, will be likely to reduce the amount of steam-room required, and to rid the boiler of those objectionable features, as far as strength and cost are concerned-wagon-tops, and more especially steam-domes. The weakest part of a boiler (because it is a circle of the largest radius) is the wagon-top, or upper crown-plate. This is often still further weakened by cutting a hole for the dome in it, half as large as the diameter of the cylinder of the boiler. A single-riveted dome, as ordinarily made, does not restore much above half the strength thus taken away. Nor will cutting an opening smaller than the dome into the boiler," decrease the strength only in proportion to the opening, since there is equal pressure on both sides of the part of the shell of the boiler that is enclosed within the dome, so that its slight resistance to bending from the curve it occupies to a straight line, is all the resistance it offers to the steam pressure. Transverse stays across the opening are generally in the way of the steam pipe, but they can be placed in front and back of the steam, pipe with advantage. According to Mr. Colburn,j locomotive boilers frequently burst through the plates to which the dome is attached, or through plates immediately adjoining. A smaller diameter and greater height of dome would doubtless accomplish the required purpose better, and weaken the boiler less than those commonly used. In England, domes are frequently round-topped and welded up from a single plate; a separate man-hole must then be put in. Cast-iron rings turned to receive a flat plate, are certainly cheaper,-and heavier. The ring should be welded in before turning, otherwise it will spring out of shape during the process of riveting. An abundance of steam-room, better situated than a dome, to carry dry steam, and allowing thin material, on account of the small diameter of the parts, is shown in the engraving of Evans' boiler, plate 56, fig. 4. This, in connection with the retreating fire-box of Millholland, plate 53, which requires no crown-bars nor braces, would make, probably, as strong and light a boiler as the necessities of good separation of steam would allow. A very little increase in the diameter of the cylinder of the boiler, will give as much steam room as very capacious domes; for instance, an increase of the diameter from 44 to 46 inches, in a cylinder 101 feet long, is equal to two domes, 22 by 24 inches each. And the increase of diameter would hardly weaken the same joints and plates more than two domes would weaken them. But mere steam-room is not the only condition of dry steam; the mouth of the steampipe, especially in case of a locomotive on a rough road, must be considerably removed from the water level. Hence Evans' arrangement is good. 18. STRENGTH AND CONSTRUCTION OF STAYED SURFACES. —This subject, has been before alluded to with reference to staying thin fire-plates. In any case, a fiat plate situated like the fire-plates of a boiler, under steam pressure, is a girder, of which the stiffness to resist that pressure is as the cube of the thickness of the plate. Therefore the stiffness of the plate is of no practical account, and the stay-bolts and tie-rods alone resist the bursting pressure of the steam. The best Lowmoor bars are found to have the following ultimate strengths: I inch diameter, 24- tons, or 31 tons per square inch. 1~ do. 34~ tons, or 28 do. do. * The 30-inch domes of some of the Baltimore & Ohio R. R, boilers have but a 15-inch opening into the boiler. + Boiler Explosions. BOILER JOINTS AND CONSTRUCTION. 47 The lI-inch bar, having 56 per cent. more section, has but 40 per cent. more strength than the 1-inch bar; or, otherwise, per square inch of section, the I~-inch bar has 28 tons ultimate strength, and the 1-inch bar 31 tons. Hence much of the anomaly that arises in experimentalizing, when a bar is turned down to a smaller diameter, and thus only the core is tested for the measure of the strength of the whole bar. It is a general fact, that iron is improved in quality and tenacity by repeated rolling; and that thus smaller bars are stronger than larger bars:wire, the extreme, being strongest of all, having 35 to 40 tons per square inch of ultimate strength., Rivet-iron has from 24 to 28 tons per square inch ultimate strength. It may be concluded that the tensile strength of rods per inch of section varies materially with the diameter, being greater as the diameter is less: that is to say, the entire strength of the bolt does not increase as rapidly as the sectional area. Tie-rods used in boilers are usually about I inch diameter, varying from -i inch to 1 inch; and, within these limits of size, the ultimate tensile strength of good Staffordshire bars is 25 tons per square inch; of best Staffordshire bars, 28 tons per inch; of best Yorkshire bars, 30 tons per square inch. When screwed at the ends, within the original diameter, the strength is reduced by two tons per square inch, as will subsequently be shown; and the reduced strengths of the rods are, respectively, 23 tons, 26 tons, and 28 tons per square inch of section. When the screw is formed upon a sufficiently enlarged diameter, the tensile strength of the bolt remains unimpaired. Estimating the working strength of tie-rods at one-fifth of the ultimate strength, the table contains the strengths of tie-rods of the diameters used in practice. ULTIMATE AND WORKING STRENGTHS OF TIE RODS. ULTIMATE STRENGTH. \ _ WORKING STRENGTH QUALITY. Diameter. Per taken at one-fifth Square Inch Whole Section. of Section. Inches. Tons.* Tons. Lbs. Tons. Lbs. Good Staffordshire, 1 2 5 1 9 44,000 4 8,800 (Clark,).25 56,000 5 11,200 Average Pennsylvania I 16-8 37,700 3'4 7,500 Charcoal, (Govern- 128 282 49,300 44 9,800 ment test, 1859,).1 hi_____28 62,700 5'6 12,500 Average Salisbury 18 40,400 3-6 8,100 (Government test, 1 30 23-5. 52,800 4'7 10,500 1859,)...\ 1 30 67,200 6 13,400 * 2,240 lbs. The values in this table are correct where the full section of the rod is maintained, in whatever way it is fixed. If, screwed within the original diameter, one-tenth of the tabulated strength must be deducted. According to experiments made by Mr. Brunel, screwed bolts and nuts, of Coalbrookdale iron, 1; inch diameter, bore an average breaking weight, or tensile strain, of 29 tons, applied between the head and the nut, equivalent to 23'2 tons, or, say, 23 tons per square inch of section of the body of the bolt. Bolts and nuts of smaller diameter were found to increase in strength, per square inch of section, as the diameter was less. Thus, — Diameter. Breaking Weight. Strain per Square Inch. 14 inch....... 29 tons....... 23 tons. "...... 20 "...... 25 " 1 *....... ". 25 " 1 "...... 12...... 27 " t " "....... 32' * By successively reheating and reworking puddled iron, Mr. William Clay found that whilst its original tensile strength was 43,904 lbs. per square inch, its strength at the sixth reheating was 61,824 lbs. Subsequent workings reduced the strength, until at the twelfth reheating it again stood at 43,904 lbs. 48 THE ECONOMICAL GENERATION OF STEAM. In most of these instances, the bolt snapped at the base of the screwed part, the bolt being 16 inches long between the head and the nut, as sketched, fig. 43, plate 64, the length of the screwed. part 31 inches. It is remarkable that the relative strength of the bolts increases as the size is reduced. Thus, 8-inch bolts, having only one-fourth of the sectional area of l1-inch bolts, are proved to have more than one-third the breaking weight; otherwise, the strength rises from 23 tons per inch of section for 1-inch bolts, to 32 tons per inch for -inch bolts. It is probable that the smaller rods are proportionally the stronger, for the same reason that bars are stronger than boiler-plates-because they are more thoroughly rolled down. That the screwing of a bolt should reduce its tensile strength, seems certain. The bolts were not proved to show this; and an estimate can only be formed upon the results of four bolts, fig. 42, plate 64, of which the shank was 1 inch diameter, and the screwed part 1 inch. The mean breaking weight was 311 tons, the bolts failing in the shank: equal to 25 2 tons per inch of section. Thus, the extra size of the screwed part added 2 tons per inch to the strength of the bolt; and, inversely, it may be inferred that the screwing of 1~-inch bolts deducts 8 per cent. of the strength of the entire bolt. The deduction may amount to 10 or 12 per cent. for smaller bolts. The heads of the li-inch bolts were 1 inch thick, and stood fast during all the trials. The nuts of these bolts varied from 1: inch to i inch in depth; at 1 inch deep, they stood well; at inch, the threads were strained; and at inch, the thread was stripped from the bolt:showing that the nut was rather shallow where its depth was three-fifths of the diameter. It is nevertheless established in ordinary practice, that the depth of the head of a bolt, and of the nut, is sufficient when it is half the diameter. Full and closely-fitting threads are, of course, essential with such a proportion. In general, the thickness of the head may be half the diameter of the bolt, and that of the nut five-eighths of the diameter. From experiments on the strength of screws let into boiler-plate and copper-plate, in the style of fire-box stay-bolts, by Mr. Fairbairn, it was found, 1st, That a.-inch copper bolt, with an enlarged end, screwed into a 8-inch copper plate, and riveted on the end, as in fig. 19, plate 64, broke with 7 2 tons tensile strain. 2d. That a i-inch iron bolt, screwed and riveted into a i-inch copper plate, failed with a load of 10'7 tons, the rivet-head being broken off, and the bolt drawn out of the plate, stripping the copper thread. 3d. That a'-inch iron bolt screwed, without riveting, into a 8-inch copper plate, was drawn out of the plate with 8-1 tons, stripping the copper thread. 4th. That a 1-inch iron bolt, screwed and riveted into a 8-inch iron plate, failed by fracture across the shank, with 12'5 tons, the, screw and plate remaining uninjured. These results, then, exhibit as follows:1. Copper into copper, screwed and riveted,.... 7'2 tons breaking weight. 2. Iron into copper, do. do...... 107 do. 3. Iron into copper, screwed only,...... 8'1 do. 4. Iron into iron, screwed and riveted,.... 12'5 do. Touching the last of these results, iron into iron, the brealking weight is practically the same as that of the i-inch screwed bolts, tested by Mr. Brunel; the 1-inch depth of screw in the plate, supplemented by the riveting, being sufficiently strong for the bolt, and equivalent to a nut f inchdeep. The first and second results, showing the strengths of copper and iron bolts screwed and riveted into copper plates, are those which directly concern the question of fire-box stays-the iron stay being 50 per cent. stronger thanthe copper stay. Fire-box stays have been tested still more directly, by Mr. Fairbairn. The experiment, however, did not measure the effects of the superior expansion of the fire-plate and the consequent weakening of the stay-bolts. He constructed two flat boxes, 22 inches square, with top and bottom plates of }-inch copper and -1-inch iron respectively, inclosing a 21-inch water-space, with il-inch iron stays, having enlarged ends screwed and riveted into the plates, to represent the BOILER JOINTS AND CONSTRUCTION. 49 conditions of an ordinary fire-box. The first box had the stays placed at 5-inch intervals. On the application of water-pressure, the sides commenced to bulge out or swell, between the stays, under 455 lbs. pressure per square inch; and, with 815 lbs. per inch, the box burst by drawing the head of one of the stays through the copper plate. In the second box, the stays were placed at 4-inch centres; the swelling commenced under 515 Ibs. pressure, and amounted to onethird of an inch under 1600 lbs.; at 1625 lbs. per square inch, the box failed, by one of the stays drawing through the iron plate, stripping the thread in the plate. As each stay, in the first case, bore the pressure on an area of 5 X 5 -= 25 square inches; and in the second, an area of 4 X 4 = 16 square inches; the total strains borne by the stays were, for the first, 815 lbs. X 25 inches = 9 tons on each stay; for the second, 1625 lbs. X 16 inches = 11 tons, nearly, on each stay. These strains are less than the tensile strength of the stays, which, according to Mr. Brunel's results, would be about 14 tons. It appears that the grip of the stays is superior to the strength of the plates to resist bulging, even at 4-inch centres; that the tensile strength of the iron stay-bolt is at least equal to the grip in the plate; and that that of the copper bolt is less. In any case, the distortion of the plate is the first symptom of weakness. Thus:At 5-inch centres, distortion commenced at a pressure of.. 455 lbs. per square inch. At 4-inch do. do do. 515 lbs. do. Again, distributing the test-strengths of the bolts over surfaces of 5 inches and 4 inches square, they failed with the following pressures:5-inch Surface. 4-inch Surface. 1. Copper into copper, screwed and riveted, per square inch, 645 lbs. 1,008 lbs. 2. Iron and copper, screwed and riveted, do. 959 lbs. 1,498 lbs. 3. Iron into copper, screwed only, do. 726 lbs. 1,134 lbs. 4. Iron into iron, screwed and riveted, do. 1,120 lbs. 1,749 lbs. Diameter of bolts, - inch, clear of thread. The total working strength of copper and iron bolts, -inch diameter at the base of the thread, screwed and riveted into w-inch copper-plates, taken at one-fifth of the rupturing strain, are, for copper bolts, 3,200 Ibs., and for iron bolts, 4,800 lbs. For:-inch iron bolts, in — inch iron plates, 5,600 lbs. Iron is commonly believed to be the best material for stay-bolts, on account of its superior strength. It is not a question of strength, however, for, as proved by the experiments mentioned, copper is strong enough; but a question of durability. The drift of the evidence is in favor of copper bolts, even in case of iron plates. Iron bolts not only corrode, but crystallize and break off. Some English makers place iron stays in the two or three upper rows, and copper in the remainder, where the corrosive influences are stronger. But the upper bolts are found to break most frequently, from the superior expansion of the inner plate; hence the material that will endure the most bending should be employed for them especially. The entire pressure on the roof of the fire-box, that is, on the horizontal area of the base of the fire-box, not transferred to the outer crown-plate by " braces " or hangers, must be resisted by the stay-bolts and the mediums of junction at the base; and it is not impossible that the downward pressure considerably strains these lateral overhung supports; besides, the rates of expansion of iron and copper under varieties of temperature are different. A locomotive-boiler is observed to expand -13-inch in a length of 15 feet, or, say, 1 in 1,000, in rising from an ordinary temperature of 62~ to 365~ —the temperature of steam of 150 lbs. pressure per inch. Again, according to ordinary unauthenticated tables, copper expands by heat half as much again as iron, and, taking the mean temperature of the copper of the fire-box at twice as much as that of the shell —an assumption which, we suppose, is something much below the fact —the vertical expansion of the fire-box would be, upon the whole, three times as much as that of the shell, and the difference of expansion would be twice that of the iron, or at the rate of 1 in 500. On a fire-box 5 feet 3 inches high, the difference of expansion would, at this rate, amount to i inch. That is to say, the 7 50 THE ECONOMICAL GENERATION OF STEAM. upper stay-bolts would be deflected I inch from their normal position, when under the power of high-pressed steam. On a length of 3 inches, a deflection of - inch is immoderate; and, considering the alternate expansion and contraction, bending and relaxing, attendant upon getting up steam and letting it down, it is reasonable to conclude that the same cause of degradation is at work with the stay-bolts as that already suggested for boiler-plate at tle rivet-joints the alternation of strain, tension, and relaxation, which loosens the texture, and ultimately overpowers the cohesion of the material so treated-incurring partial fracture and accelerated corrosion. On this arguments the failure of stay-bolts should be, as in fact it is, localized at or near'their junctions with the plates, which are the points of maximum strain, similarly to the localization of furrows near rivet-joints. Occasionally, entire rows of rivets are found to have snapped across, close to the plate, independently of corrosive action; suggesting a cause of failure precisely the same as that which breaks axles-an alternating lateral strain and relaxation, beyond the limits of enduring elasticity. If this be the cause of the failure of stay-bolts, it follows that, as copper is more pliable than iron, copper is the more suitable material for stay-bolts. The reasons above advanced, afford an explanation of the fact, that fire-boxes with narrow water-spaces are more subject to leakage than those with wider spaces-the stays being shorter and less flexible in the former case, and more likely to fail. For the same reasons, solid stay-bolts of smaller diameter, sufficiently strong, are preferable to others of larger diameter. They are more elastic, and yield to unequal expansion more readily than thicker stays, and are, therefore, likely to be more durable. The oblique strain to be resisted is similqr in character to that which tells upon weak wrought-iron wheel-spokes, under excessive shrinkage of tyres, converting them into serpentine or ogee forms. As the duration of wheel-spokes, under such circumstances, is increased by expanding them towards their junction with the nave and with the rim, and rounding them in; so, that of stay-bolts is improved by turning off the thread on the middle part, now sometimes done in England, and frequently in America. Probably the application of the principle might be advantageously extended; and a superior iron stay-bolt may be made in the form shown at fig. 19, plate 64, in which a -inch screw is turned down to E inch or A inch diameter at the middle. It is said that the bubbles of steam and the active circulation of impure water, close to the plates, tend to cut off the stay-bolts at this point, rendering a thick end desirable. When the brackets by which the fire-box shell is supported upon the frame are small, the adjoining stay-bolts are liable to be overstrained, and leakage is incurred. To avoid this, and to distribute the strain, the brackets should be co-extensive with the fire-box shell. Tubular stays have been discussed in the chapter on Thin Fire-plates. They are sometimes employed with ordinary plates, and were used on the Liverpool and Manchester Railway above 20 years ago. Since bituminous coal-burning has been practised, large hollow stays have been employed as air tubes. Mr. Beattie places l~-inch hollow stays, exclusively, in his new fire-boxes. A few are made thin, to show defects earlier. Mr. Betts, of the Jersey City Loc. Works, sometimes uses a hollow stay made of "hydraulic" iron pipe, 7-inch diameter, with a bore. It is riveted as usual, and then expanded like a flue, with a steel pin or mandrel. This makes a very tight job. Tubular stays, in addition to their importance to thin plates, burn out less rapidly than solid stays in any case. Stay-bolts, whether solid or tubular, should be made of soft material, in order to rivet soundly; when hard, they crumble in the thread. Cast-steel stayT-bolts, with a spring temper, would be likely to stand the effects of expansion and contraction better than any other material, since their small diameter and great elasticity would allow the highest flexibility. Their ends might be so large as to prevent the necessity of riveting them. The stay-bolt employed by Mr. Millhlolland, in the Reading engines (Fig. 17, Plate 65), gives a long thread, and renders much riveting unnecessary. Fire-box roof-stays or crown-bars are variously constructed and braced. They consist of flat bars set on edge, singly or in pairs, extending between the front and back plates on most English engines, and between the side plates on most American engines, with bolts let down through them, or screwed from below ilnto theml, to sustain the crown-sheet of the fire-box. Mr. Edwin Clark BOILER JOINTS AND CONSTRUCTION. 51 found, on testing ~wrought-iron bars for their transverse strength, that the deflection became excessively long before the bars broke, and that, practically, bars should be tested transversely, not for strength, but for stiffness. He found that bytpreviously straining and straightening a wroughtiron bar, its stiffness and practical strength are greatly increased. According to an account of experimental test of fire-box roof-stays recently published by the English Board of Trade, it appears that a stay made of two plates i inch thick, 5 inches deep, and placed 1 inch apart, riveted together by 7-inch rivets, and made to span a fire-box 3 feet 6 inches long, bore a load of 103 tons in the centre, the deflection under which was not mentioned, but it acquired a permanent set of ~ inch. The lower rivet-holes were barely 1 inch clear of the lower edges of the stay, and deducted probably one-third from the strength of the solid bar. Allowing, then, one-half more than the testing load, we estimate that the strength of the stay, had it been solid-composed of two entire bars 5 X- inch-would have been equal to a load of 13 tons at the centre. By Mr. E. Clark's formula, the working strength of a stay 42 inches long, consisting of two solid plates 5 x inch, would be 5 I 5I X15'3-=9'1 tons, which is only about i the strength estimated directly from experiment. Another stay, similarly made, of two'-inch plates riveted together, but reduced to 31 inches deep in the middle, was tested in the same way, and broke across, under a load of 71 tons at the centre. By the formula, the working strength of this stay would have been 3 85 tons at the centre, or one-half the breaking strength. But in this it is assumed that there was no reduction of strength by rivets near the middle of its length. In connection with the same subject, it is reported that a boiler was tested by water-pressure, in which roof-stays of the same construction were applied to a fire-box 42 inches long, fastened to the roof in the usual way, by bolts at short intervals, there being a clear space of 1 inch or so between the roof-plates and the stay, except at the bolts, and the plate being, say e inch thick. The stays were placed at intervals of 5 inches between centres. The two central stays were only 31 inches deep. The boiler was proved by water-pressure; at 469 lbs. per square inch, four of the roof-stays failed —the two central stays and the stay next to each; they cracked across the lower half of each stay, and deflected 1I inch for the central stays, and 18 inch for the two others. The failure was accompanied by two successive reports at a few seconds' interval, indicating that the central stays failed first, then the others. The total pressure borne by each stay previous to failure acted on an area of 42 X 5 inches=:210 square inches, and amounted to 210 X469 lbs.= 98,490 lbs., or 44 tons, uniformly distributed, or 22 tons at the centre. This, then, should be accepted as the ultimate strength of the two central stays, in their connection with the fire-box. Stays of the same dimensions, proved separately, bore only 71 tons of breaking weight —just onethird of the breaking strength when fixed to the fire-box. To account for this striking difference, without reference to the absolute value of the tests, it may be suggested that the crown-plate of the fire-box, being intimately united to the stay, became a part of it, and bore its share of the strain tensilely, as a lower flange. There was, in fact, for each stay, a lower flange of copper, 5 inches broad, - inch thick, presenting 2 square inches of section, fixed, moreover, at the extremities. It would be idle to speculate on the precise amount of strength thus added, or on the amount of lateral support derived from the neighboring stays, in the absence of necessary particulars; but, at least, it is apparent, that an important accession of strength is gained to a roof-stay by intimately uniting it to the roof-plate. The strength of the smaller stays, then, presents itself thus:Tested alone, failed with.. e....... 7- tons on the centre. Tested in its place, failed with........ 22 do. do. Working strength, by formula, treating it as a solid plate, unaffected by rivet-holes, 3'85 do. do. Again, the strength of the larger stay appears thus:Tested alone, acquired a permanent set of R inch, with.... 101- tons on the centre. Corresponding strength of solid stay alone, as unaffected by rivet-holes, esti- 3 do doI mated from the test,.... o o..... Tested in its place, failed after the smaller stays gave away, and bore at least,. 22 do. do. Working strength, by formula, treating it as a solid plate,. o.. 13 do. do. 52 THE ECONOMICAL GENERATION OF STEAM. As to,the smaller stay, it appears that, in its place, its ultimate strength was between five and six times the working strength assigned to it by the formula, treating it as a solid plate; and the larger stay bore, in its place, at least two-thirds more than the working strength as by formula. It must have borne much more than that proportion, had it been placed beside stays of equal strength. Upon the whole, therefore, Mr. E. Clark's formula for the transverse strength of iron bars, as applied to the roofs of fire-boxes, and fixed in the ordinary manner, appears to be correct. It may be remarked tlat, inasmuch as the absolute strength varies only as the square of the depth, and as the length inversely; and the stiffness varies as the cube of the depth, and inversely as the cube of the length, the proportion of deflection to span cannot be the same in all cases. Should it be found, in practice generally, that there is an increase of strength by straining and straightening iron bars, already mentioned on Mr. E. Clark's authority, the process may, of course, be beneficially applied to strengthen crown bars. The perfect fitting of the roof-stays to the superficies of the fire-box, is of essential importance; not only at the intermediate points of contact, where the holding bolts are applied, but particularly at the ends, so as to direct the pressure vertically upon the walls of the fire-box.'Proper and well-spread fitting is the more needful at the ends, as the resisting power of the copper of the firebox-if that metal is used-to compression is very low-only 3 tons per square inch, according to Mr. Fairbairn; whereas, a roof-stay carrying 12 tons of pressure, throws 6 tons upon each end, for which a well-spread bearing is essential. An ordinary roof-stay, 1 inch thick, with a bearing at the ends 2 inches long, has 2 square inches of bearing surface, which, to carry 6 tons, must support a load of 3 tons per square inch, up to the limits of resistance to compression. Fairbairn'judiciously spreads the end of the roof-stays, increasing the bearing to about 5 square inches at each end; others, desirous apparently of clearing the rivet-joint, bear only upon the extreme corner of the vertical plate. Additional provision should be made to extend the bearing of the ends of roof-stays, beyond what is generally provided, considering the want of firmness of copper plate. T1he method of flanging the front and back plates, and lapping them under the roof-plate, affords probably the strongest junction for resistance to pressure, when the roof-stays are accu-rately fitted to, and made to bear upon, and embrace the upright plates. It is otherwise if they are not so. The American practice of placing the roof-stays transversely, causing their ends to bear directly upon the edges of the side plates, brought up square on the outside for the purpose, possesses the advantages of a direct and square bearing, of relieving the tube-plate of the downward pressure, and of adjusting the span of the stays to the width of the fire-box, instead of its length. It is likely that the pressure of roof-stays resting on the tube-plate, has occasionally been a cause of leakage of flues, and of failure of the flue-plate; the side-plates are better adapted to bear the pressure, as the stay-bolts in the sides are more numerous and shorter than in the flue-plate. Transverse stays would obstruct the insertion of longitudinal tie-rods, except above their level; but gussets could be inserted to stay the ends of the boiler.'Applied transversely, the roof-stays could be extended to the sides of the fire-box shell, and riveted to them, thereby entirely relieving the side stay-bolts from downward strains; further, they would stiffen the sides of the boiler, and their strength as girders would be doubled, bearing the same relation to ordinary roof-stays, that a beam fixed at both ends bears to a beam laid upon props. On this plan, explosions of the fire-box would be practically impossible. On the ordinary plan, the linking df the three or four central roof-stays, or, say, half the number, to the crown of the shell, is of great utility in strengthening the boiler: the central roof-stays are the most in need of such aid, for, as was shown n the test by water-pressure, already quoted, out of eight roof-stays, the four in the middle were the first and only ones to give way. The " braces," or, more properly, hangers, which connect the crown-bars with the furnace, are intended, not so much to stiffen the bars, as to hold the entire inner fire-box in place. Thle bars simply prevent the roof from bulging the braces help to resist the total steam-pressure on the crown-sheet. These braces are more numerous in American than in English boilers, although in the latter they are often larger and fitted with more care. Some of the most reckless designing, and a large amount of the poorest workmanship ever met BOILER JOINTS AND CONSTRUCTION. 53 with, is found in the crown-supports of locomotive furnaces. Old J'L rails, 3 inches deep, riveted to the crown-plate, have been employed as stay-bars, with no other end supports than the plate itself. Braces of various lengths, and therefore expanding differently and bringing undue strain on the shorter ones-braces converging at the centre of the upper crown-plate instead of diverging front the bars and radiating towards the outer crown, thereby securing the greatest possible strength ~braces put in with sole reference to convenience in getting into the boiler and examining itall these varieties of construction are too frequently seen. A very substantial method of arranging the braces is shown in the plate of the Rogers boiler (No. 46). Mr. Eaton, Loc. Supt. of the Great Western of Canada, employs a novel and obviously excellent plan of braces, consisting of thin plates extending the whole length of the crown. (See Plates 45 and 451.) It will be seen that Millholland's boiler (Plate 53) has no crown-bars, the crownplates being stayed together like the sides of the fire-box. The same feature is adopted by Winans.* It could probably be carried out with advantage for roofing all' fire-boxes and combustion-chambers. Simple tension rods are certainly much cheaper, and are likely to be stronger than a system of girders arranged and tied to the upper crown-plate according to the various caprices of boiler-makers. The steam-room is somewhat reduced, but if there must be domes even with the wagon-top, they have only to be made higher in the former case, to deliver dry stealm. Gusset-stays are applied in the angles of the fire-box shell, to stay the flat surfaces, front and back, to the cylindrical portions; superseding the more ordinary through tie-rods, which they do effectually. They may be most solidly and simply welded into their places by Bertram's process, of which there are excellent examples; and are often put in, as in Yates' boiler, (Plate 52.) As constructed in this country, the sides of combustion-chambers are often flat for some distance, and require stay-bolts like the fire-box. The gusset of a boiler with a raised crown-plate, but without a combustion-chamber, is a little oval. and therefore not as strong as the cylindrical part. Cross-stays between the plates on either side, and extending over the flues, are sometimes applied in such cases. Regarding the locomotive-boiler as a cylinder with flat ends, the greatest strain falls necessarily upon the longitudinal seams, and the least upon the transverse circular seams at, and between the ends of the boiler. The distinction may be concisely illustrated thus:-Let the circle (Fig. 1, Plate 67) be a cross section of a boiler, then the area of the circle is a measure of the longitudinal pressure, exerted by the steam upon the end of the boiler, to be resisted by the transverse joints. Again, let the parallelogram (Fig. 2) be a longitudinal section of the boiler, then the area of the figure is a measure of the transverse pressure of the steam to be borne by the longitudinal joints. To find the proportional strains upon equal sections of the circular and longitudinal joints, set off the interval^a a upon the circular seam, and- an equal interval c c upon the longitudinal seam; draw the diameters a b, a b, and the parallels c d, c d; then the areas of pressure to be resisted in the two cases, are, for the circular seams, the two triangular spaces a b; and, for the longitudinal seams, the rectangle, c d. The united areas of the two triangles are obviously but one-half of the area of the rectangle, and therefore the total steanm-pressure to be resisted by the two lengths of circular joint, a a, b b, is only one-half of that to be borne by the two equal lengths c c, d d. The same proportion holds good, all round the circle: and it follows, generally, that the strain per unit of length, on transverse circular joints, is only half of that upon longitudinal joints. The longitudinal seams, therefore, slould be the strongest, and there, setting aside meantime the welded joint, the double-welt double-riveted joint should be applied; and for the circular seams, bearing only half the strain upon the others, the single-riveted lap-joint is sufficient, being, in fact, proportionally stronger in its place than the other; —proportionally stronger, that is to say, with respect to strains'arising from steam-pressure; and when the boiler is employed exclusively as a generator of steam, and is placed freely on the frame. But, where it is rigidly bound to the frames, or employed as a responsible fastening for the steam-cylinders,'double-riveted joints had better, no doubt, be employed to resist the extra strains thus incurred, at the gusset or junction of the barrel with the fire-box shell and at the smoke-box. * It was first patented larch 8, 1848, by John ]IcConochie and L. J. Claude, of England. 54! THE ECONOMICAL GENERATION OF STEAM. The working strength of the double-welt double-riveted joint was found to be 9,000 lbs. per square inch of section, for best Lowmoor plate; and, with that joint applied to the longitudinal seams, the boiler becomes qualified to work under a maximum strain of 9,000 lbs. per square inch of section. The distinction here recognized, is of practical importance:-the tensile strength transversely of the cylindrical elements of the boiler, is the measure of its strength, because the longitudinal seams of cylindrical parts are strained twice as much as the other seams; and a boiler united by double-welt double-riveted longitudinal joints in the cylindrical portions, and everywhere else bysingle-rivet lap-joints, is practically of equal strength throughout. Take an example: A boiler is 48 inches in diameter, of'-inch plate, with 120 Ibs. steam:-to find the strain on the metal. The two sides are, together, - inch thick, and one square inch of section is developed in a length of 1 ~ I = 1i inch, measured along the boiler; because, inversely, X 1 -- 1 square inch. Now 48 X 1 =- 64 square inches, the diametrical area of pressure (which is represented by the rectangle c d, in fig. 2, recently considered), sustained by the square inch section of metal; and 64 X 120= 7,680 Ibs. per square inch, is the strain of the steam-pressure on the longitudinal section. The longitudinal strain per square inch of transverse section, due to the pressure on the entire end of the boiler, would be, as already shown, half of the transverse strain; that is, 7,680 - 2 = 3,840 lbs. per square inch. It may be otherwise proved arithmetically:~48 inches diameter yields a circumference of 151 inclhes, and an area of 1809 5 square inches, and 1809'5 X 120 = 217,140 lbs. total pressure on the end; to bear this pressure, there is 151 inches circumference of i-inch plate, equal in area to 56'6 square inches, and 217,140 ~ 56'6 3,840 lbs., the longitudinal strain per square inch of transverse section of the boiler, as already found. At the maximum working strain, 9,000 Ibs. per square inch of longitudinal section of metal, the longitudinal strain is 4,500 lbs. per square inch of transverse section, which may be abundantly provided for by single-rivet lap-joints; and the steam-pressure due to this strain, would amount to 9,000' 64 =- 140 lbs. per square inch of surface. As a matter of fact, the ends of locomotive-boilers are not so exposed as has just been assumed for illustration. For the flues not only subtract by as much as their total sectional area, from the smoke-box end, but they are in themselves very efficient stays-obviously so, for they entirely support the upper part of the fire-box flue-plate. The greater part of the back end of the boiler, also, is stayed to the fire-box, and they are mutually supporting. There are, in short, only the upper segments of the end plates requiring extra stays; and the question is, then, simply, whether they should be stayed by through tie-rods or by gussets? and that is not entirely a matter of convenience, since the long stays may expand more than the rest of the boiler. For example, in a 48-inch boiler, the circular segment above the flues, may be, say 20 inches high, and may be effectually guarded by a few 1-inch tie-rods, at 6 to 8 inches apart. If each rod be charged with the pressure of 120 lbs. steam, on a surface averaging 7 inches square, the total would be about 6,000 lbs.; opposed to this, aninch bar of best iron has a working strength of 10,000 Ibs., equal to upwards of 200 Ibs. per square inch. 19. ATTACHMENTS TO BOILERS.-A locomotive boiler must evidently possess other features of strength than those required in a mere steam generator. I-Iowever strongly and independently the frames of the engines and of the vehicle may be constructed, the simple holding of the boiler in place upon them, necessitates considerable extra stiffness in the latter. But in the most approved designs of locomotives, the boiler forms an important part of the framing, and it is likely to come into still greater service as a girder, for the following reasons:-that part' of the framing lying between the driving-shaft and the cylinder is not an essential part of the locomotive, and has been omitted in some cases; the boiler answers as the framing, and not only stiffens the structure, preventing side or lateral flexure, but sustains the entire fore and aft strain of the engines, that is, all the power developed in the cylinders. Since the centre line of the boiler is so far nbove that of the cylinder, giving the latter so much leverage, this strain tends to pry the boiler asunder at the gusset, and is very objectionable. The other extreme is to make the framing so * In the engines of the "' Corsair" class on the Great Western Railway of England. BOILER JOINTS AND CONSTRUCTION. 55 enormously heavy as to stand not only the thrust of the piston, but all other strains to which the engine and the vehicle are subjected. Now a 24 to 3-inch rod, merely as large as the piston-rod, running between the driving-jaws and the cylinder, is all that the fore and aft thrust of the steam requires. But frames of four times this area have not been stiff enough of themselves, not to buckle and tremble and bend with the oblique thrust of the connecting-rod; not because these strains are very severe, but because metal arranged primarily as a tension-rod, is least effective as a girder. At a convenient distance from this tension or passive-rod, lies the boiler-a girder of perhaps a hundred times its stiffness, and under such conditions of stealm pressure that the lateral and vertical strains of the engines and the vehicle would fall upon its circular seams, which are the strongest. Therefore, a passive-rod, not much larger than the piston-rod, frequently stayed to the boiler by expansion joints, would make the stiffest, lightest, and cheapest framing possible. It is evident that the stiffness of the boiler is excessive, compared with any strains that might be thus transmitted to it. No increased thickness of plate, and no expenses of construction not warranted by other considerations, would be required. To whatever extent the boiler may be employed as a girder, all the attachments which connect it with the rest of the structure can be so made as to prevent weakness and leakage. The cylinders of most outside-connected American engines, are secured from vertical and lateral motion in a very simple and substantial manner. The method used at the Rogers Works is partially shown by fig. 3, plate 46. The smoke-box is of — inch iron, farther stiffened by y —inch patches on the inside. It has perpendicular sides, and a flat bottom 1-inch thick, extending beyond the smoke-box and resting on the frames. A ring of say 4 by 1 -inch iron is riveted into tile front of the box. Two diagonal braces sometimes run from the lower back sheet of the box to the cylinder of the boiler. A 4-inch flat flange on the cylinder rests on the projecting bottomplate; and the cylinder-flange, the plate and the frame are bolted together by six lI —inch bolts in reamed holes. Another flange on the cylinder turns,up for some 15 inches, resting against and bolted to the side of the smoke-box. With the round smoke-box, wide cylinder flanges, or an intermediate cast-iron saddle between the cylinders, are bolted to the smoke-box, which thus stiffens the engines in a similar manner. The braces from the framing to the cylinder of the boiler, when attached to the latter by broad, spreading " palms " fastened by from six to eight rivets or bolts, do not cause perceptible weakness or leakage, if they are set with one edge, rather than the side, outward, or if being set side outward, they have slotted holes where they are bolted to the frame, so that the boiler can expand without straining them. These palms are best fastened to the boilers by bolts and nuts, but when the flues are set before the boiler is put in place, bolts and nuts cannot be put in, and they are fastened by tapped bolts. Quarter-inch plate does not give the thread a very good hold, and therefore the English practice before mentioned, of putting patches inside such parts of the boiler, is important; with very thin steel plates, it will be indispensable. The larger the palms, and the more numerous the braces, the less is the boiler weakened by them. The cross-brace between the frames, or "belly-brace," a deep - or 1-inch plate attached to 18 or 20 inches of the circumference of the boiler by a flange or an angle-iron, allows sufficient expansion and takes a large hold on the boiler. Sometimes all the braces are triangular plates, standing with one edge outward, one corner of the triangle resting on the framing and the opposite side joining the boiler with a long flange. The part of the boiler most liable to strains from without, as mentioned in another section, is the gusset. When a strong brace between the frame and the cylinder of the boiler is placed close to it, and the bearing of the fire-box on the frame is as long as possible, the strains upon it are much diminished. The expansion-joint, or bearing of the fire-box upon the framing, is frequently the cause of fracture and leakage. In many English engines, no provision is made for the superior expansion of the boiler, which is about R inch, as measured. Thus no part of the engine can be kept exactly in line, and the whole boiler, especially tlie gusset, is unnecessarily and severely strained. In some American engines with inside cylinders (which are becoming obsolete), the expansion joint is at the smoke-box end. The most approved plan inll case of the bar or square frame, is to bolt long angle-iron brackets to the entire sides of the fire-box. These rest and slide on the frame. 56 THE ECONOMICAL GENERATION OF STEAM. If these angle-irons are narrow, and bolted to the outside plate only, or if they are insufficiently bolted-for an excess of positive strength is required to give proper stiffness-they are sure to work loose and leak. So many boilers are leaky at this point, that the necessity of a good job should be pretty well understood by this time. Similar angle-irons forming a lower expansion joint on the stay between the pedestals (or pairs of jaws) distributes the strain over a larger surface of the fire-box. With the slab frame, the fastening used at the Jersey City Works is simple, and distributes the strain over a still larger area of the fire-box. Each frame passes through two broad clasps of thick plate bent to receive them, each plate resting against the firebox, to which it is bolted with two palms, say 10 inches square. The screw-bolts which fasten the angle-irons or clasps, usually pass through the water-space and are secured in the fire-plate, like a stay-bolt. In some cases, a nut with a concave face and a packing of canvas and red lead is placed upon the bolt while it is being passed through the water-space, and screwed tightly against the inside of the outer fire-box plate. The advantages of comparatively thick outer firebox plates have been alluded to: their value to a good expansion joint is obvious. When an equalizing beam centre or pin, forged on a flat plate, is fastened to the side of the fire-box, care should be taken to secure the outer end of the pin to some part of the framing, or to the fire-box, by a broad brace, otherwise the rivets holding the beam-pin, as well as the fire-box, will be unduly strained. Were it not for the very heavy plates employed, the fastening of frames to the corners of the fire-box, as is done in England frequently, would cause serious leakage and failure. There is some trouble at present from this cause. Care should be taken not to fill up too much circulating space in the water-legs by these attachments. Large collections of scale have been known to form upon four stay-bolts and a screw-bolt between them equidistant from each, filling up some 48 cubic inches of water-space. There are so manly clumsy cast-iron attachments made to American boilers, that any saving of weight by stronger material and better construction, is not likely to be appreciated, so long as they are tolerated. The author has seen a cast-iron dome-cover in use weighing 350 lbs.; castiron smoke-stack saddles often weigh 200 Ilbs.; saddles for securing the cylinders to round smokeboxes sometimes weigh a ton or more. There are sand-box saddles weighing 200 lbs., and castiron dome-saddles and covers weighing 600 lbs. for the two domes, and dome-cover saddles weighing 300 lbs. = 3,650 lbs. in all, etc., etc. Three-quarters of all this dead weight is unnecessary..20. TREATMENT OF STEEL.-The principles of working steel are very easily applied, in practice; indeed, it works at a lower heat, and is bent and compressed more easily than iron. Nor is it necessary to spoil a single plate or rivet, in getting used to it. The author has seen steel plates of at least three different makes, returned as " rotten," one end having been burned by a careless operative; the other end had been flanged in the most intricate manner after the plate was returned. There are two cardinal rules as to working boiler steel: 1st. Never heat the plates above a medium shade of red, not approaching white, for flanging, etc. For welding, a lower heat than that required for iron will answer, and the plates may be kept from oxidizing by applying sand to them, which forms a glassy covering of silicate of iron, in the usual manner, or by heating them in a gas flame, as described in the section on welding plates. 2d. Never condense or upset steel, as in the case of rivets, while it is cold. This process hardens the metal, making it brittle. Steel, containing more carbon than iron, oxidizes more easily, and will thus burn away rapidly in contact with air, at a heat which iron will bear safely. Boiler steel, however, is invariably found softer than iron when red hot. Borax, which by forming a cinder enables dry or much-worked iron to weld readily, is not required in welding the semi-steel.* Borax, or some similar flux, is used in welding Howell's cast steel. This leaves the welded part much harder than the rest. The softness is restored by heating the part to a low red, and then cooling it in pine-wood sawdust, which is better than ashes, the material generally used for annealing. * Report of James Tucker, Master Smith, on Government Test of Corning, Winslow & Co.'s Semi-steel, March 30th, 1859. BOILER JOINTS AND CONSTRUCTION. 57 Steel rivets are so much smaller than iron rivets, that they can be more quickly headed. They invariably crack if hammered cold. They can be set with sufficient rapidity with a riveting hammer, but snap-heading is obviously better. 21. TESTING BOILERS,AND THE SAFETY-VALVE.-No method of testing boilers, which does not absolutely burst them, will reveal their ultimate strength, and no method will indicate their strength beyond the testing pressure. The hydrostatic test is believed to injure boilers,* because its increase of pressure is not uniform, but strains the boiler without showing the effect of each additional pound, and because any pressure beyond that employed in working, is liable to aggravate defects that might not otherwise develop themselves. It does not show the strength of boilers when they have been unequally expanded by working temperature, and is likely to be especially incorrect with reference to the crown-braces, &c. Besides, the test does not, of course, show the deterioration of boilers by the oblique strain of lap-joints, by unequal vibration of thick and thin parts, by corrosion, and by other effects of use. The fact that a new boiler will stand a higher cold-water pressure than that required for working, is by itself of comparatively small account. The quality of the material as developed in working, and the proportion of sectional area of plate and stays to the designed pressure, is more conclusive. But a very valuable method of testing-not the ultimate strength, but the workmanship of boilers as well as their actual strength up to the testing point-is that adopted some two years since by Dr. Joule, of Manchester. He fills the boiler entirely full of water, and then builds a fire on the grate. When the water has been warmed from 50~ to 90", the safety-valve is loaded, up to the point desired. The rise of the pressure is then carefully noted by the ordinary steamgauge; if the progress of the pointer be constant and uniform, without stoppage or retardation, it is inferred that the boiler has withstood it without undue strain or incipient rupture. The starting or springing of any bad work or weak joints is certain to be indicated. The expan, sion of the water by the heat is so rapid, that the pressure has been known to rise from 0 to 62 lbs. in five minutes. It is probable that a considerable number of explosions occur from defective safety-valves. The common puppet or conical valve, with a steep taper, is likely to stick in its seat, and must rise entirely out of its seat to allow the steam a free passage. But the worst feature of the common valve is the lever, with so little movement at the long arm where the spring is attached, that the valve, which is close to the fulcrum, cannot rise far enough in any case to blow off all the steam the boiler can make while standing. On a boiler with two ordinary safety-valves, set at 100 lbs. pressure per square inch, the author has observed 140 lbs. pressure accumulated while the engine was standing, both valves meantime blowing off all they could. It is probable that some locomotive boilers in use, might be exploded by accidentally leaving the ash-pan damper open, and the steam jet turned on-the safety-valve being in order. I)irect-action safety-valves are by far the best. Mr. Corliss, of the Corliss Steam Engine Co., Providence, uses a common half-elliptic spring, secured in the middle to the top of the boiler, and pressing at each end upon two safety-valves. Hawthorn's annular valve, plate 67, fig. 3, is a ring with two concentric edges for the escape of steam; thus with a given diameter, the actual opening is nearly twice that of the disc-valve. It is reported by Clark, that whereas, with the ordinary mitre valve, with lever, the pressure rose 15 per cent. above that at which the valves were set, with the annular valve, the surplus pressure did not exceed 31 per cent. Baillie's safetyvalve —a large direct-action valve held down by volute springs, is shown at plate 67, fig.'4. Mr. Baillie has described the results of comparative trials of an ordinary lever safety-valve arid direct-action saftety-valve, on the same boiler; from which it appears that the trials were made on a locomotive-boiler, having 890 square feet of heating surface. The engine was not worked, and the draft was created by a ~-inch jet in the chimney, led through a small tube from the boiler. The two safety-vales were equally weighted, to 64 lbs. per square inch, and compared with the actual pressure in the boiler, by means of a manometer; the small valve was 3'6 inches in diam* Mr. Colburn mentions, if "Boiler Explosions," that a new locomotive boiler exploded in England, probably from having been injured by a previous test of 130 lbs. per square inch. 8 58 THE ECONOMICAL GENERATION OF STEAIM. eter, weighted with the usual lever and spring balance, and the large valve was 12 inches diameter, weighted with seven volute-springs, as in fig. 4. The large and the small valves were then set fast alternately, and the blow-pipe started in the chimney; the surplus steam escaped through the free valve in two successive trials, and it was found that under the stimulus of the w-inch blast, the pressure, with the small valve in action, rose in four minutes from 64 to 105 lbs. per square inch, through 41 lbs. per inch. The valve was in good order, and rose 12 inch from its seat; the experiment was interrupted at 105 lbs., as the pressure continued to rise. With the large valve in action, and the blast worked as before, the pressure rose in four minutes; from 64 lbs. to 76 Ibs., through 12 Ibs. per square inch; the fire was kept up for half an hour longer, but the pressure remained stationary at 76 lbs., the valve having risen'T inch-sufficiently high to let off all the steam that was generated. The opening of the large valve, -- inch, multiplied by the circumference, was 1'57 square inch area; add the area of the'-inch blast-pipe,'196 square inch, and the sum, 1'766 square, inch was the total area of escape. The opening of the small valve, -L- inch high, was only'942 square inch; and the total opening, with the jet, was 1'138 square inch. Thus the total opening for escape with the small valve, was about two-thirds of the opening with the large valve. In a subsequent trial to ascertain the evaporation, the firing was continued for an hour, the steam blowing off at 76 lbs. through the large valve; 80 cubic feet or 21 tons of water were evaporated during the hour. These facts corroborate the opinions of Clark as set forth in " Railway Machinery." 22. GRATES.-The offices of the grate, with reference to combustion, especially in America, where it is employed to stir the fire and remove clinker, are numerous and important, and will be considered in another chapter. Under the head of Boiler-making, a few facts and suggestions may be useful. A set of cast-iron grate-bars in bituminous coal burning boilers, is estimated to last six months.* They frequently last above a year, and have been known to fail on the first trip. The life of wrought-iron grate-bars, with coal and coke, on English. lines, is above one year. Considering the first cost and value when burned out, of the two kinds, there is a decided economy in using the wrought-iron bar. The causes of the more rapid failure of the cast.bar, are evident:-cast-iron melts and oxidizes more rapidly, by reason of its excess of carbon, while wrought iron will not melt; wrought iron is stronger, and will stand the weight of coal better when hot; the wrought bar is generally about half the thickness of the cast bar, for thne reason last named, and offers twice the area per square inch of section to the air, and is thus kept comparatively cool by the draft. Wrought bars are very cheaply made in some of the English shops. The piece A, fig. 5, plate 67, is simply cut out of a flat plane bar, by a heavy punching machine, and the grate-bars thus formed are set in a slotted angle-iron, as shown at fig. 6. The best movable grate in use, probably, for bituminous coal burning, is that of Hudson and Allen, shown with the Rogers engine, Plate 46. This must evidently be of cast iron. In fact, nearly all the rocking or movable grates for freeing the fire from clinker, would be very expensive in wrought iron. The life of cast-iron grates may be considerably lengthened by leaving a recess in the top of the bar, which, filling with ashes-a non-conductor mitigates the action of the fire; and the clinker will not adhere to the ashes, and may be readily removed. The bottom edge may be made very thin, if the bar is deep, and will thus last better. Grates burn out most rapidly when the damper is closed, or when there is no blast, since no cool air can then pass through them. It is customary on many lines to make the ash-pan water-tight, and to flood it with water to a depth of one or two inches, which, especially where no air is entering, tends, by its evaporation, to keep the grates cool. For anthracite coal burning, as has been before mentioned, the water-grate is deemed indispensable, since cast iron, at least, will not stand its intense and concentrated heat. The water* Report of Experiments on Coal-burning Locomotives; Pennsylvania Railway, 1859. BOILER JOINTS AND CONSTRUCTION. 59 grate employed by Mr. Millholland, on the Reading Engines, is shown with the engraving of his boiler, on plate 53, and a portion of it is shown full size by figs. 1, 2 and 3, plte 65. The lower bar of each group, is merely a thimble passing over a rod which may be withdrawn to drop the fire; this is so far below the more intense heat, as to stand well. The other tubes communicate with the forward and after water-legs, and a good circulation is preserved in them by their inclination. The water-grate in the " Phleger " boiler (Plate 47), is similar to that in Norris' boiler (Plate 48), where it is explained in detail, together with Norris' movable bridge-wall, which has already been mentioned. This grate is not so simple or easily repaired as Millholland's, but in connection with the movable bridge, and the other advantages of Norris' arrangement, it is a good feature. Yates' water-grate (Plate 52) is similar to, and, indeed, a part of the legs of the fire-box, and will stand as well, if the side and bottom water-spaces are large enough to insure good circulation. For coal sufficiently clean not to require a shaking grate, this is doubtless a durable arrangement. 23. GENERAL CONCLUSIONS AS TO BOILER JOINTS AND CONSTRUCTION.~The general facts and conclusions about the strength of locomotive boilers, are as follows: — 1st. The resistance of the plates of the cylindrical part to steam-pressure, in the direction of the longitudinal seams, is about one-half that in the direction of the circular seams. The strength of the plates, therefore, must be proportioned to the diameter of the cylindrical part and to the intended pressure. 2d. The resistance of the longitudinal joints to the steam-pressure is about one-half that of circular joints. The following rule, by Clark, is given to find the thickness of plate due to a given diameter of boiler, quality of joint and working-pressure:-Multiply the working-pressure in pounds per square inch by the diameter in inches; and divide the product by the working strength of the longitudinal joint in pounds-and by 2. The final quotient is the required thickness of plate in inches.* 3d. Since the circular joint requires about one-half the strength of the entire plate, while the single-riveted lap-joint possesses about one-half the strength of the entire plate, this joint is strong enough to resist the steam-pressure only, whatever may be the strength of the longitudinal joints.t But the boiler is, in some measure, a girder, requiring a higher degree of strength than that due to the steam-pressure, and the circular joints, practically, require greater strength than singleriveting would give them, in case the longitudinal joints are made as strong or nearly as strong as the plate, and a proportionate increase of steam-pressure is carried. The exact proportion of strength of the circular joints to the longitudinal joints would vary with the length, diameter, and weight of the boiler and the material used. Considering the advantages of stiffness to the general durability of the structure, and of hooping the boiler to its longitudinal strength, a doubleriveted hoop (single-welt) circular joint, or a double-welt joint, would probably increase the life of the boiler sufficiently to warrant its extra cost. The welded joint, if practicable, being as strong as the entire plate, should be applied to the longitudinal seams; if inconvenient or impraticable, the double-welt joint, which maybe relied in as possessing at least 80 per cent. of the strength of the entire plate, should be employed. To continue to use the single-riveted lap-joint, which is little more than half as strong as the plate, in a part of the boiler which requires the entire strength of the plate, is simply absurd. 4th. The resistance of the plates and joints of the flat parts of the boiler to steam-pressure, is, practically, a question of the strength of stays. The plate is thick enough to resist the pressure, and, with the proper mode of fastening, to hold the stays, whenever it is thick enough to furnish the required stiffness —to stand the strain and jar of the engines and the vehicle, without undue deterioration. As has been shown in another chapter, there will be great advantage in making the fire-plates thin. It is probable that the outer plates of the fire-box should not be less than X inchl'thick for iron, and { inch for steel, to furnish the required stiffness. The follow* The rule applies to every kind of joint, when the thickness of plate does not exceed ]-inch. The table of working' strength of joints, whichis 5 their ultimate strength, will be found on page 36. f The drawing of the plates under the hammer, in riveting, is a cause of tight circular joints. Mr. Betts, of the Jersey City Locomotive Works, mentions the fact that he has thus made the circular joints tight enough by mere riveting, not to require caulking. 60 THE ECONOMICAL GENERATION OF STEAM. ing rule, by Clark, applies to the strength of the flat parts:-To find the pitch of stay-bolts, due to a given working pressure. Divide the working strength of the stay-bolts in pounds, by the working pressure in pounds per square inch-and find the square root of the quotient. The result is the pitch in inches. The following are the working strengths of stay-bolts 4 inch solid diameter, screwed and riveted into plates:Copper stay-bolt in copper plate,. 4. 3,200 lbs. Iron do. do. do....... 4,800 lbs. Iron do. in iron plate,........ 5,600 lbs. The construction of the crowns of furnaces similar to that of the water-legs,* appears to be cheaper and less liable to malconstruction and injury from unequal expansion, than the wagon-top and crown-bar construction. Hollow stay-bolts are well adapted to thin fire-plates, and are useful in some parts as air tubes; they will burn away less rapidly than solid stays, and can be more strongly set, especially if made of copper or steel. But if the circulation is not sufficient to prevent the overheating and superior expansion of fire-plates, they will be more likely to break than solid bolts of smaller diameter. Copper stay-bolts are undoubtedly much superior to iron, for the reason last named. It is probable that cast-steel stay-bolts would be superior to those of any other material; they would not crystallize, but would spring without injury, with the expansion of the fire-plate, and they would occupy much less water-space, being stronger and therefore smaller. Steel rivets occupy a less section of plate than iron, and make a proportionately stronger joint, with a given thickness of plate. In pitching rivets, the necessity of a tight joint, as well as a strong one, must be considered. Conical and countersunk holes make the best rivet-heads and decrease the oblique strain due to the lap. Snap-rivets are quite as strong as those headed by hammers, and are set about twice as fast. The lap of joints is an element of weakness, and hence, the lapjoints of {-inch plates are not practically stronger than those of I-inch plates. Gas-welding, by Bertram's process, appears likely to become cheap and reliable; if not, thickedged plates will make a sufficiently strong joint. Round corners on flanged plates, acting as expansion-joints and straining the iron less severely than square corners, are desirable. Water-legs, mid-feathers, and other complications of fire-surfaces, are going out of use. The combustion-chamber (in new engines) is deemed sufficiently important to combustion, to warrant its extra cost. The'unnecessary weight of the cast-iron parts and attachments of boilers, is so great as to offset the advantages of stronger material as to lightness. The other considerations with reference to boiler construction, which have been treated, can hardly be briefly recapitulated. On the whole, it appears that the common style of boilers is defective, with reference to strength, chiefly in the longitudinal joints of the cylindrical part, and, to some extent, in the staying of the crown-plate. As to durability, the fire-box is generally defective, and the entire boiler is susceptible of considerable improvement, at a practicable extra cost. * The " Rocket," by Stephenson, 1829, was the first example of this construction. CHAPTER V. FLUES AND FLUE-SETTING.-WATER TUBES. 1. DIVERSITY OF MATERIALS EMPLOYED.-There is a great difference of opinion among locomotive superintendents, especially inAmerica, with reference to the best material for flues. In England, brass flues having always been employed for coke, with fair success, were very naturally continued in use for raw coal, which is, on the whole, less trying than coke. But in America, copper flues, which had been as widely employed for wood, began to fail under the mechanical action of flying particles of coal. Iron flues, moreover, were cheaper; and when set so as to remain tight, gave promise of greater durability than copper or brassythe latter material having been taken up as a compromise between the softness of copper, to set well, and the hardness of iron, to wear well. At the same time, the use of long thimbles in copper flues, has increased their durability. In fact, either of the three materials is quite satisfactory if well put in and carefully used, and each has its advocates. The author has seen copper flues cut out and replaced by iron flues on one line, and new iron flues cut out and pieced with copper, that they might be soundly set, on an adjacent line-both to adapt wood-burning engines to coal-burning. Meanwhile, steel flues are coming into use. This state of affairs is explained by the fact that our coal-burning experience has been too limited to establish a standard of practice. The following considerations may be useful in giving direction to farther experiments. 2. THE CONDITIONS GOVERNING A CHOICE OF MATERIAL.-The office of a flue is to conduct heat to the surrounding water, at the least cost; the items of cost being, 1st, waste heat, 2d, maintenance of the flue. Assuming, for the present, that the best conducting flue is the least durable, and that the poorest conducting flue is the most durable, the question is-by avoiding which species of expense, shall we attain to the highest economy? A glance at the general results of practice, furnishes an answer. The cost of recaulking, resetting, and renewing all kinds of flues, is an appreciable amount. The first cost of a set of iron flues, according to the Pennsylvania Railway experiments, is $555 for 14 years, or about $40 per year, not to speak of repairs. That of copper flues, even supposing them to last as long as iron, would be above $70 per year. But no practice nor working experiments, with full-sized machinery, as far as the author is aware, have detected any saving of fuel due to the superior conducting power of copper flues. The relative thermal resistances of iron and copper are as 96 to 40.' Dr. Ure has made some small-scale experiments which show the decided advantage of copper, under the particular circumstances of his experiments, as a conducting material for boilers. On the contrary, the same locomotives have repeatedly done the same work with the same fuel, as nearly as could be measured, first with copper flues, and afterwards with iron flues. There is, doubtless, a saving due to better conduction, but so long as that saving cannot be detected in ordinary practice-so long, indeed, as the cost of fuel due to poor conduction is less than that of repairs due to the perishableness of the material, the question of conduction is not in order; if there were no difference in the durability of metals, it would be the only question.+ It has been assumed that the best conducting flue is the least durable, and vice versa. This * Despretz. + As has been before remarked, if a thickness of scale equal to that of the flue is allowed to collect and remain outside the latter, the question of conduction is merely one of abstract scientific interest, without practical importance. 62 THE ECONOMICAL GENERATION OF STEAM. may be true of the materials from which flues are made, but it is not necessarily true of the flues. The thickness of the material, the length of the flue, and the amount of heated gas which strikes its surface, when taken together, may affect the ultimate result to a degree in comparison with which the mere conducting power of the metal is of no practical account. It has already been found advisable to shorten and to decrease the diameters of flues in many cases' Too great a length of flue, of whatever material, abstracts so much heat from the gases, that the smoke-box end ceases to be heating surface, the steam and the gases being of equal temperature. Too large a flue allows a large body of heated gas to pass through its centre, without touching its sides at all. Heat-traps and current deflectorst are employed with great success to prevent the latter waste of fuel. The thickness of flues, within practicable limits, has quite as much influence as the simple conducting power of the metal on the rapid transmission of heat;+ and it may yet appear that the thinnest flue is the most durable. As far as the materialof flues is concerned, then, the only practical consideration is that of ducrability; what material, with reference to the joints and the body of the flue, will require the least cost for maintenance? 3. DEFECTS AND REiMEDIES.-MATERIALS CONSIDERED.-The principal failures of flues are as follows:-Wearing, burning, breaking, corrosion, leakage, sagging, and collapse; with reference to which the various materials will be briefly considered. The wearing of flues is caused chiefly by the attrition of the sharp, flying particles of coal or coke. In case of perfect combustion, this cause would not exist; as combustion is improved, it will be less formidable. Wearing occurs principally at the extreme fire-box end; the flange by which the flue is set is often cut through. The resistance of flues is manifestly due entirely to their hardness; the materials then range in the following order: steel, iron, brass, copper. Brass sometimes softens with use, and is cut like copper. In case of steel and iron, wearing is not a serious defect; with brass or copper, without iron thimbles, it would ordinarily be the most serious. But thimbles (Figs. 5 and 18, Plate 65) abstract the best part of the heating surface, in proportion to their length. Thimbles 9 inches long decidedly injure the steam-generating capacity of a boiler. ~ It may be said that the remainder of the flue would take up any heat that passed the first 9 inches; so it may be said that if bricks are put next the fire-plates, to preserve them from the intensity of the fire, the heat which is not taken up by the water at this point, will be, at another point. Practically it is not so, [1 probably because much of the heat is not again brought into actual contact with the metal, in the absence of deflectors and heat-traps. Thimbles also decrease the area of escape for the gases,requiring a sharper blast,T[ without promoting their contact with the flue-surface, and they interfere with cleansing the surfaces from soot. But cast-iron thimbles tend to keep the flue tight.*', With iron and steel flues, properly set, thimbles are entirely unnecessary; with copper, if not with brass, they are indispensable. The burning of flues is entirely due to a contracted water-space, and bad circulation between them;tt it is an abnormal condition of the boiler, and should not be contemplated, much less discussed. If a 1-inch iron plate will stand the immediate contact of white-hot coals, flues less than one eighth as thick, would hardly be expected to burn up in the less intense heat of flame. Very thick flues of impure iron, may be injured, however, by a normal amount of heat. It would hardly be economical to carry a great weight of water between widely-spaced flues, in * Mr. Mihllholland is removing the 14 feet by 2 inch flues, from the Winans engines on the Reading Railway, and substituting flues 12 feet long by 1- inches diameter. t See the chapter on " Transmission of Heat." I By substituting 16 wire gauge, copper for 13, for the flues (2 in. diam.) of a stationary boiler, in a case mentioned by Mr. Betts, the steam-generating capacity was increased about one-quarter. At the ends of the flues, \ inch copper was employed to make a tight joint. ~ Mr. Morris Miller, Loc. Supt. of the Harlem Railway, arrived at this result by experiment. II Experiments on the North Pennsylvania Railway. 1T In some cases two-thirds of the power of the blast has been required to get the gases through thimbled flues. —larck. ** Cast iron permanently expands by the heat, while flues tend to contract slightly, after the mandrel or expander is withdrawn. - Hence cast-iron thimbles were (uite necessary to keep tubes tight, with the old-fashioned mode of setting. They were first applied by Mr. W. S. Hudson, in 1850. i~t See also the chapter on " Cirlculation." FLUES AND FLUE-SETTING. —-WATER-TUBES. 63 the whole front and middle of the boiler, in order to have enough to take up all the heat transmitted to the first few inches of the flues. Narrow spaces at the fire-box end would be unnecessarily wide at the smoke-box end. An intense heat must be sustained at the fire-box end, and the metal must be not only homogeneous, but thin, to stand it. The advantage of steel, then, is evident. Brass is the first material to fail from accidental overheating. The breaking of flues usually occurs close to the fire tube-plate, and is caused by the unequal expansion of the flues and the shell of the boiler. Copper will stand this action better than the harder materials, but it has more to stand,by reason of its superior expansion. Setting iron flues by Prosser's Expander (Figs. 22, 23 and 24, Plate 64), if the metal be not of superior quality and well annealed, sometimes starts ruptures, and aggravates the effects of expansion and contraction. Short flues which also allow a chamber and better combustion, expand less, of course, and therefore stand better. Bad circulation not only allows the flues to become weak by overheating, but it is the chief cause of their superior expansion, since the metal is too thin to get much hotter than the water, if the water is in actual contact with it. Corrosion and wearing, by reducing the substance of the flue, also aggravate this defect. A thimble fitting the flues tightly for an inch, is useful. in case the flues have a tendency to break, from any cause.* An expansion joint in the smoke-box tube-plate t is a very important feature, if the superior expansion of flues cannot be prevented. It is said that some forms of flue-setting allow the flue to slip back and forth; the leakage due to any such arrangement, and the manner of keeping the tube-plates from bulging, are not mentioned. Theoretically, considering the superior expansion as well as the greater ductility and softness of copper, the inferior expansion and greater positive strength of iron, the Intermediate qualities of brass, and the aggravating circumstances common to each, there is not much difference in these materials, with reference to breaking. And practically, as far as the author has learned, there is not much difference, except that hard iron flues without safe ends, and overstrained in setting, are most frequently broken. And badly-welded iron flues are likely to split during or after setting. Steel flues, however, have the good qualities of copper, due to homogeneousness, without its superior expansion, in addition to the strength of iron. Their history is yet too short to warrant positive conclusions, but it will be very surprising if they do not stand better than any other material. As to corrosion, copper and brass are quite superior to iron, resisting both the action of the water and the sulphur in coal. Steel approaches the excellence of copper in both these particulars. The corrosion of flues, however, is not an important defect, except at the ends, and these may be protected by an extra thickness of metal. The leakage of flues is the result of defective setting, and will be considered in a following section, where it will perhaps appear that this species of failure is not dependent to a great extent on the material of the flue. The sagging of flues is dependent on the softness qf the metal and on the length and diameter of the flue and its consequent stiffness. And in case of brass and copper which have been at all overheated, the elongation of the flue causes it to buckle and sag. Brass, though harder than copper at first, often becomes so soft as to sag in the middle. As a result, the water-spaces being very much diminished, and the flues perhaps touching each other, they are soon burned out. Sagging also causes leaky joints. This defect does not often exhibit itself in flues that are not either too long or too small for the best transmission of heat. With iron it would rarely occur within any practicable dimensions, and with steel it would never occur, since a steel flue is both stiffer and lighter than an iron flue.: Collapse is dependent on the diameter, length ~ and thickness of flues, and on the stiffness of * Mr. Dyer Williams, Loc. Supt. on the New York Central Railway, has found flues that were entirely broken off, held in place and kept from leaking by the thimble only. t See section on Flanging Plates and Expansion Joints. t Very thin flues will float in water if their ends are plugged; hence steel, as far as its weight is concerned, would tend to rise rather than to sag. It is an interesting'fact, having at least a remote bearing on locomotive-boiler construction, that steamship boilers weighing from 20 to 40 tons have been plugged at the safety-valve and pipe attachments, and then dropped from a crane into the water and floated some distance, as the easiest mode of transportation. ~ Mr. Fairbairn's experiments on large tubes established the fact that the resistance to collapse is inversely as the length. 64 THE ECONOMICAL GENERATION OF STEAM. the metal. It is also dependent on the steam pressure. It does not often occur to locomotive flues,* because they have an excess of thickness in proportion to the causes of collapse. But this thickness is adverse to the rapid transmission of heat, and might be somewhat reduced in short-flue engines. Above all, the flue might be so strengthened by current deflectors placed from one to two feet apart, along its length (Plate 66, Figs. 14 and 15) as to allow a great reduc. tion in its thickness. The only objection that could be raised against permanent thimbles or heat-traps of any kind in flues, is that in case of imperfect combustion they would be likely to stop up with fine coal. They would somewhat interfere, also, with cleansing the surfaces from soot. Reducing the diameter of the flues of wood-burning engines, except at the fire-box end, either by thimbles or piecing, is almost sure to cause stoppage, since chips and coals of the full diameter of the flue are drawn into it by the blast till they reach an obstruction, at which point they form a permanent dam. Coal burning, however, which is defective to this extent, should be abandoned entirely. If the draft is very uneven through the flues, and especially if they leak a "little, even very fine coal may lodge, and in time stop them up. But it is not possible with the equable draft of petticoat pipes or diaphragms,t that a few spiral deflectors should cause any stoppage; as far as they have been used there has been no such result. And if there was, the remedy would be to regulate the degree of deflection by the shape and number of the spirals, so as to decrease the velocity of the current where it would tend to be strongest, and thus equalize it through all the flues. The spirals or heat-traps, then, would have three important functions:-the equalization of the draft; the bringing of all parts of the current of hot gases into contact with the metal; and the strengthening of the flue against collapse. As a result of the last office of the spirals, the flue might be as thin as paper, especially if made of steel; the setting, of course, being facilitated by safe-ends. As far as materials are concerned, the hardest and stiffest are the least liable to collapse. Steel in any case has a decided advantage. A perfectly cylindrical flue of course offers the greatest resistance to collapse; copper flues lapped and brazed are not perfectly cylindrical; solid brass and copper flues may be much thinner.1 4. CONCLUSION-ADVANTAGES OF STEEL.-Durability being the governing consideration in the choice of material, it would appear that the chief normal causes of failure are wearing, breaking and leakage, the latter being more a matter of defective setting than of defective material. Wearing is a very serious defect in soft flues. Breaking (except by reason of bad setting, which is an abnormal cause) is due chiefly to the superior expansion of the flue, and is remedied, as far as the flues are concerned, by using a homogeneous material, which can endure the strain. As to wearing and breaking, then, steel is far superior to all other materials. The practice (with all its diversity) bears out these conclusions: Iron flues are always preferred by those who know how to set them soundly, because they are stiffest and hardest. Brass flues being more easily set than iron, by the old-fashioned methods, are gaining in favor, as colllpared with copper, simply because they are harder. Steel in the form of flues, as was before remarked, has a very short history, but so far, those who have used it, believe that it will supersede all other materials. A large number of steel flues are on their trial; Neilson of Glasgow has put Russell & Howell's homogeneous (cast-steel) flues, 1- inch outside dianeter, into several locomotive boilers with a P-inch steel flue-plate. The ordinary setting —expanding with a drift, and a steel thimble an inch long at the fire-box end-is employed. A safe-end 1 inch thick is preferred, but is not absolutely necessary. It is certain that steel would stand the strain of Prosser's Expander, since good, soft iron will stand it, steel being also soft and much more tenacious and tough. The present cost of Russell & Howell's steel tubes, is as follows: * Several flues 5 ft. long by 1H diameter, and 51 inch thick, have collapsed in one of the New York Steam Fire-engine boilers. f See the section on the Smoke-box, &c. t A new process of making copper flues is to pour the melted metal into a hollow iron drum which revolves with great rapidity; the centrifugal force throws the liquid mass into the shape of a thick hollow cylinder, in which shape it cools; it is then drawn down over a mandrel, between rolls, to the required length and thickness. Copper cannot be cast like brass, into a hollow ingot, with facility. FLUES AND FLUE-SETTING.-WATER-TUBES. 65 Diameter. Thickness. Weight per Foot. Price in New York. Price in England. Il- inch. 17 or 18 wire gauge. Average, &l1b. 35 cents per foot. 24 cents per foot. 1I I" 16 or 17 " " 1 " 45 " c 32 " " 2 " 15 or 16 " c 1" cc 55 " " 40 "'" It being a commercial fact, that a large demand decreases prices, we may expect to get steel tubes at about the price of iron, ultimately. The material, indeed, is much more expensive, but less of it is employed, and the process of manufacture is the ordinary lap-welding process, the cost of which would be much the same in either case. But at present rates, steel is cheaper than brass or copper; and its weight is considerably less, even with the present thicknesses, as will be observed in the following table, which is probably a correct average. Material. Number. Length. Diameter. Thickness. Weight. Price per Foot. Total Cost. Steel, 140 11 ft. 2 in. 16 W. G. 1,925 lbs. 55 cents. $847 Br 5 140 11 ft. 2 in. 14 W. G. 2,770 lbs. 56 cents, $862 00 8 ^ 140 thimbles at 10 cents, 14 00 Copper' 4140 i1 ft. 2 in. 13 W. G. 2,800 lbs. 60 cents, 924 00 B 938 140 thimbles at 10 cents, 14 00 TI1'I $ 140 11 ft. 2 in. 13 W. G. 3,080 lbs. 28 cents, 431 20 476 ron, 280 safe-ends at 16 cents,; 44 80 The duratiori of brass flues is from 2- to 3 years on the Eastern Counties and other English lines; that of iron flues on the Pennsylvania railway, is said to be 14 years; copper flues with coal, can hardly be averaged above li years; the duration of steel flues at present can only be estimated, in accordance with the facts mentioned. Old copper flues are worth about 38 to 40 cents per foot, if not burned, and can be pieced so that parts of them will last above 25 years with wvood. Brass flues can be similarly pieced, but with more difficulty, and as old metal are worth about 34 to 36 cents per foot. Old iron flues are worth about 2 cents per foot, and old steel would not be much more valuable. But with coal, it may be assumed, in the absence of complete data, that the cost offrequently piecing and renewing brass and copper, offsets this disadvantage of iron and steel. The transmission of heat through a layer of metal, is proportional, as far as thickness is concerned, inversely the thickness.* Steel can be made thinner than any other material, because of its hardness and strength, and its consequent resistance to depreciation and fracture. Therefore the superior conducting power of copper may be assumed to be offset by its superior thickness. But it is probable that deflectors in flues, arranged to bring more heat into contact with them, is of greater importance than either thickness or conducting power. Durability being the first consideration, the advantages of steel are evident and decided. They may be briefly enumerated as follows:-Hardness to resist wearing; toughness to resist breakage, and to facilitate sound setting; stiffness to allow thinness and rapid transmission of heat without risk of collapse; a satisfactory degree of resistance to corrosion; greater homogeneity as well as thinness and consequent resistance to undue heat —in short, superior durability, to which may be added lightness, as compared with either of the other flue materials. The consideration of flues with reference to their diameter, length, distance apart, and respecTransmission of Heat, and Circulation. Good circulation, as has been observed, is especially important to the durability of flues. 5. ELUE-SETTING.-A permanently tight joint of the flue in the flue-plate, is chiefly important to prevent leakage. It is also necessary to the durability of the flue and to the strength of the boiler, since the flues act as tie-rods between the tube-plates. Flue-setting is done by various essentially distinct methods: —by expansion with a plane mandrel, and subsequent caulking; by similar expansion, the tightness being preserved by a thimble; by expansion and corrugation or beading, as by Prosser's method, which preserves both the strength * Prof. Rankine's " Steam Engine and other Prime Mfovers." 66G TTHE ECONOMICAL GENERATION OF STEAMI. and the tightness of the joint; by compression, as by driving or forcing the flue into the plate; by wedging or external thimbles; and by screwing the flue into the plate. There are also combinanations of these methods. The setting of copper flues for wood-burning was a very simple process-expanding the flue to fill the hole in the plate with a round mandrel, and then caulking the end by upsetting a flange on it, while the mandrel was in the flue. But the soft copper soon began to yield under the difference of expansion between the flue and the shell of the boiler, and under the effects of imperfect workmanship. The flue also had a tendency to contract, when the mandrel was removed. Thimbles were therefore introduced-first wrought-iron, and then cast-iron, which were found to expand permanently with the heat. Steel resembles cast iron in this particular, on account of its excess of carbon. In England, substantially the same setting was employed for brass flues. Figs. 13 and 14, plate 65, illustrate the plan commonly adopted in England. The flue-brass, 10 wiregauge-is put in from the siloke-box end, for which purpose it is enlarged for 31 inches of its length to 17 inch diameter and 13 wire gauge, the other end being 1 inch, compressed in a ring or dressed in the lathe, to fit tightly. The holes in both plates are bored without taper. The flue is driven into the fire-box end, and a round mandrel, tapering 3 — inch in 1 inch, is driven into the flue, after which the projecting edge is slightly turned over with a caulking tool, and a wrought-iron or steel thimble I- inch thick and from I to 1 inch long is driven in. The smoke-box end is treated in the same manner, but has no thimble. When each flue is ferruled as it is put in, the caulking does not disturb the flues already set, and does not tend to make the plate concave. Wroughtiron thimbles made in a die, with the proper taper, are furnished among railway supplies. Messrs. Beyer, Peacock & Co. use steel thimbles, 10 wire gauge, turned to 11 wire-gauge, and tapering -J inch in 1~ inch. On the Lancashire and Yorkshire line, the thimble is driven tightly enough to stop leakage and to prevent the sheet from pulling off from the flue; and the flue is not riveted over and caulked, but is very slightly turned over. The smoke-box end is caulked, and 20 per cent. of the flues are thimbled at this end. At Fairbairn's works, flues are expanded at the smoke-box end, the joint somewhat resembling fig. 8, plate 65, and are thimbled at the other end. Mr. Allan, Locomotive Superintendent of the Scottish Central and formerly of the London and North-western line, has expanded flues at both ends, as shown at fig. 8, plate 65. On the Caledonian line, an expander quite like Prosser's (Fig. 22, Plate 64) is employed for the front end, the thimble being put in as usual at the fire-box end. On the Great Western railway of England, 2-inch flues were at one time compressed at the fire-box end to 1i inch for a foot of their length. But brass hard enough not to wear out rapidly would not stand the strain. The plan was excellent in one particular-it increased the water-space where there was most heat. It is probable that the users of steel and iron flues would do well to try it. These flues are of brass, and there is, with rare exceptions, very little difference in the method of setting them, all over Great Britain. Tlhis is the practice on the Continent, also, as far as the author has been able to ascertain. In this country, the introduction of coal fuel, with its flying particles and intense heat, has necessitated a considerable change as compared with the wood-burning practice. The result has been a great variety of plans; some of the most promising will be briefly noticed. Prosser's Expander,: which makes the joint shown at d, fig. 23, plate 64, or the one shown at fig. 8, plate 65, is illustrated in detail by figs. 22 (side.elevation), 23 (longitudinal section), and 24 (end view), plate 64, and is variously modified to suit the work required. The triangular pieces b are forced apart by the conical mandrel or wedge a, which is driven in by hammers, after the instrument has been placed in the flue. The ring y, with the projecting pins/, is merely a guide for the expanding sections b. A ring of india-rubber e, holds the expanding sections b close to the mandrel. The joint shown, requires the use of another tool called a chambering-tool, to cut the recess in the plate. Three beads, one in the chamber and one on each side of the flue-plate, are sometimes made. In cases where the joint is not accessible to the form of expander shown, another form, which is operated by turning a nut on a screw cut on the small end of the mandrel, is employed. The mandrel may be so long as to make a joint several feet distant from the flue* This was invented by Mr. Thomas Prosser, of New Torlk, who manufactures tle expander and other flue-setting tools. FLUES AND FLUE-SETTING.-WATER-TUBES. 67 plate. The most intricate boiler and condenser work is thus accomplished. It has been mentioned that iron flues are frequently strained by the severity of this process of setting; good material, annealed in charcoal dust, makes a very sound joint. A hard flue is likely to make the hole in the plate oval. A safe-end-a ring of annealed iron with the grain running around the flue-is an important protection against splitting or other fracture during the process of expansion.* Sometimes a ring of copper is put between the flue and the plate, and the expander is then employed. If the flues are not truly round, or if the holes have been worn oval by previous setting, this makes a tight joint. Countersinking or rounding the edges of the holes is important for this form of setting. Prosser's method is used in many of the locomotive works and repair shops of this country, almost exclusively for iron flues, and is in many cases deemed indispensable, especially for smoke-box joints. It is prematurely condemned on some lines, because, by reason of hard flues or some other abnormal circumstance, it has made very bad joints. It has one disadvantage; the flues cannot be taken out without either cutting them off with a special tool, or crushing them together, which spoils several inches of the flue. This is also true of any species of expanded, flanged or upset joints; but it is not true of compressed joints. The advantages of Prosser's method for hollow stays, have already been mentioned. The facility with which soft steel can be expanded and beaded, is too evident to require comment. The plan of flue-setting employed by Messrs. M. W. Baldwin & Co. in their locomotive works, has been practised for some years with great success. The joint made is substantially like Prosser's, fig. 8, plate 65. The tools are shown on plate 64. The flue projects - inch beyond the plate, and is first expanded by the triangular mandrel (fig. 25) which increases the diameter not by mere stretching, but by compressing and drawing the metal; after which a similar mandrel with wide corners turned to the diameter of the flue, is employed. The beading tool, fig. 28 (full-sized section fig. 29) is then put in and opened by a flat wedge driven by the hammer; it expands a small are of the flue at a time, and is turned a little after each blow, till the whole bead is completed. A round mandrel is then driven in, and while it is in, the end of the' flue is caulked or upset by the tool shown at fig. 26, or in case the flue is too close to the side plates, the tool fig. 27 is used alone. The thickness of the flue-plate is ~ inch. The fire-box end of the flue is sometimes swaged down to a smaller diameter while hot, thus leaving a sound shoulder, and is then set as at fig. 10, plate 65. Again, the same form of joint is obtained by brazing a short piece of iron flue into a copper flue, and setting as at fig. 15, plate 65. Iron flues are often pieced at the ends with copper, to facilitate good setting, and are then provided with long thimbles which cover the whole of the copper end and protect it from the action of the coal. Two forms of this joint, as used on the Illinois Central Railway, are shown at figs. 5 and 18, plate 65. In fig. 5, the copper is set by the, Prosser Expander and then flanged, the whole of the flange being protected from wearing, by the iron thimble. In fig. 18, the copper end is swaged down, and the smoke-box end is set by Prosser's Expander. Upsetting the ends of copper and brass flues by caulking them, is found to weaken the flange considerably, and on the New York Central and other lines, the flues are not caulked while the expanding mandrel is in them, but are turned over and but slightly upset by a " thumb tool," the action of which resembles that of the one shown by fig. 27, plate 64. On the New York Central, Mr. Jones, Loc. Supt. at Albany, uses a — inch tube-plate and a 1 -inch cast-iron thimble -3 inch thick, which, with good workmanship, makes a sound joint. Mr. Eaton, Loc. Supt. of the Great Western of Canada, uses a 2-inch brass flue, the fire-box end of which is compressed by being driven into a conical hole before setting. The hole in the flue-plate is slightly tapered and the flue is driven in. Much importance is attached to driving the flue: it is thus under permanent compression and always tends to fill the hole. A cast-iron thimble 14 inch long is then put in for, say 1l inch; the projecting end protects the brass, which is not flanged with a caulking tool, but merely turned over a little. The short /-inch iron flues oSBeattie's boiler (Plate 38) are set,on the South-western and the Great Western Railways, as shown by fig. 7, plate 65. The ferrule or safe-end is welded upon * Copper flues are now made with thick ends 1hv the process mentioned on page 20 and illustrated by fig. 3, plate 64. G8 THE ECONOMICAL GENERATION OF STEAM. the flue at the manufactory; the screw is cut in the lathe (14 threads to the inch) with a slight taper. The holes in the plate (iron - inch) are tapped from the inside, and the flue is screwed in against the shoulder, shown, from the smoke-box end, by a tapering mandrel with flat sides and sharp corners, which is driven tightly into it. A ring is temporarily put over the projecting end of the flue at the smoke-box end, which prevents the mandrel, or screw-driver, from splitting it; after which the end of the flue is sawed off and set as usual. The method of setting iron flues, first'adopted on the Reading Railway some years since, and shown at fig. 6, plate 65, has, within a year or two, been largely adopted at various locomotive works and on coal-burning lines. A modification of it is shown at fig. 9. The holes in fire-box plate (Fig. 9) are reamed to taper, say 4 inch to the foot, and are made slightly larger than the flue; between the flue and the plate, a copper or brass ferrule is driven as shown. This compresses the flue, fills the annular space perfectly, and appears to give every advantage of a copper or brass flue as to setting, while preserving the decided advantage of the iron fnlue as to weariTg. The flue is turned outward a little in fig. 6, to keep the ferrule in place, but is not strained at all, either by beading, expanding, or caulking, and the draught is not obstructed by thimbles. The joint is simple, durable, and cheap. The great feature of a compressed joint is that both the flue and the flue-pllate are under permanent compression, which constantly tends to keep them in contact, whlile the elasticity of an expanded flue tends to relax it when the expander is removed. To take the greatest advantage of the principle just mentioned, and to avoid at the same timue, botlh the loosening, and destructive effects of hammering, and the weakening of the flue by beading, Mr. J. KI. Fisher, of New York, quite recently schemed the plan of flue-setting shown by figs. 11 and 12, plate 65. It has been successfully tried at one or two of the Paterson locomotive works, and is believed to be feasible by experts who have examined it. The ends of the flues are dressed in a lathe extreme end to an angle of 60~ so as not to cut the plate; they are also tapered at the shoulder to the same angle. The hole is slightly tapered so as not to strip the flue. The flue is then pressed in by a screw passing through it and holding the other end of it by a cup-head. The flue may be turned over a little, as at fig. 12, if it is deemed advisable, but this would hardly be necessary, for the joint cannot fail to be both tight and strong, since the flue is pressed in by a greater force than the steam-pressure would exert to loosen it, and since the shoulder is jammed into the sheet far enough to secure perfect contact. The plate is slightly stretched and the flue is slightly compressed. All this is done without even jarring the adjacent flues and joints, or dishing the flue-plate. This method certainly requires good workmanship. Tihe flues and holes must be dressed exactly to gauge, which is probably quite as cheap, in the long r-in, as making up deficiencies by extra expansion. 8. CONCLUSIONS AS TO FLUE-SETTING.-The whole subject of flue-setting may appear trivial to those who are unacquainted with the driving and repairs of locomotives. The fact is, however, that leaky flues are an unmitigated evil, not so much in their waste of heated water, as in their inteTruption of the process of combustion, by allowing steam to generate in the fire, the steam thus taking the place of the gases, and cooling them and the heating surfaces. The first condition of permanently tight flues is good circulation of water among them, to receive their heat as fast as the fire transmits it to them, and to keep their temperature low, since excessive and unnecessary heat is the great dclestroyer of the heating surfaces. 2d. Thinness of flues and plates, as far as is consistent with strength enough to bear the steam-pressure, is important to durability as well as to the rapid transmission of heat. 3d. Sound and durable joints are not to a great extent a question of the material of flues, since the use of thimbles will preserve soft copper and brass from wearing, while external ferrules of copper or brass will be sure to make.a tight fit between the hard iron or steel flue and plate, although such a fit may be made without the use of a soft intervening material. 4th. But thimbles, while they make a compressed joint, give extra thickness, abstract heating surface, and interrupt the draught. They are never used with iron or steel flues. 5th. Caulking flues by upsetting their ends, weakens the material, and is not necessary. Hard brass is more injured than either copper, iron, or steel, by caulking, and if it is not hard, it vill wear out like copper. 6th. Compressing both the flue and the plate is likely to FLUES AND FLUE-SETTING. — VATER-TUBES. 69 make a permanently sounder joint than expanding them, since the tendency, in service, is, in the former case, to tighten, and in the latter, to relax. 7th. Round holes and round flues of the same size, render any extra straining of the flues by expansion, unnecessary, and they are certain to make tight joints. 8th. Prosser's Expander seems quite important for the smoke-box end of all iron or steel flues that are put in from this end, since the flue-holes here are necessarily a little larger than the flues. The elasticity of iron, at least, would not allow a simple round mandrel to make a tight joint. The practice seems to indicate quite clearly, that iron flues, compressed and caulked by an external copper thimble at the fire-box end and expanded by Prosser's Expander at the smokebox end, are the best job in every particular. Steel, however, is likely to supersede iron for reasons already set forth; and being more homogeneous than iron, and equally hard, has all its advantages as to wearing and the advantages of the softer materials as to setting. Screwed joints in the fire-box plates (Fig. 7, Plate 65) might be found to warrant;their extra cost in general practice, and should be tried, Mr. Fisher's method of pressing in the fire-box end of the flue (Fig. 11, Plate 65) also deserves attention. 7. WATER-TUBES.~-Water-tube boilers will be chiefly considered under the heads of Con-s bustion and Circula-tion. There are, however, a few facts of importance which should be stated under the head of boiler-making. Water-tubes may be made excessively thin, since the pressure is internal and only the tensile strength of the metal is employed. Thus the great advantages of rapid transmission of heat and superior durability are secured. A flue 1 inch diameter, of steel which would stand 90,000 lbs. to the inch, would require to be only 2-o inch thick (working strength) for 120 lbs. working pressure. The ends could, of course, be thickened to allow the strain of setting. In any probable arrangement of tubes, the joints would be removed from the more intense action of the fire; this is the case with Dimpfel's boiler (Plate 50) and with Montgomery's marine boiler, which is largely and very successfully used. Neither the shell nor the'tubes of Dimpfel's boiler are unduly strained by the superior expansion of the tubes, since they bend in their curved portion. The joints thus last remarkably well. There would be some difficulty in removing tubes thus arranged, were not this emergency especially provided for; at the space E, fig. 2, plate 50, any of the tubes may be removed after the joints have been cut off. Considering the better contact of the gases with the heating surfaces of water-tubes, the thorough circulation in them, and their durable features, they appear to possess decided advantages over flues. But they require a conformation of boiler which is not quite unobjectionable. The flue boiler requires but one shell; the water-tube boiler has two shells. PART II. COMB S T I ON CHAPTER I. INTRODUCTORY. I. THE ECONOMY OF COAL AS FUEL.-The heating value of a ton of average bituminous coal may be assumed to be equal to that of at least 1 cords of average wood. The recent experiments on the Pennsylvania Railway show that a ton of Pittsburg coal is equal to 1~ cords of selected oak wood. A series of trials on this line, in 1851, showed a ton of Alleghany coal, which is now pronounced inferior to the Pittsburg, for locomotive purposes, to be equal to 21 cords of mixed hard and soft wood. The results on other lines have been quite as various, establishing the fact that the proportion assumed is at least not too favorable to coal. The estimation of the relative values of wood and coal depends largely on the thoroughness of the combustion in either case; the use of a combustion-chamber is universally admitted by those who have occasionally employed wood in engines thus adapted to coal, to give equally favorable results with either fuel; and the recent and quite general employment of small grates, thin fires, and large fire-boxes (which are, practically, combustion-chambers) for wood burning, has led to a very decided economy.* The heating power of wood by bulk also varies greatly, not only with the kind of wood, but with its dryness and state of preservation. The comparative value of dry wood and of wood in its ordinary state, by weight, (the moisture varying from 20 to 50 per cent. in different kinds of wood,) is as 35 to 26;t and-the comparative value of dry over wet wood, by bulk, is probably quite as great, since wet wood causes' slow and imperfect combustion; and wastes a considerable amount of heat in the evaporation of its water, while the steam thus generated from it occupies the place of hot gases in the flues. Partially decayed wood, having about the bulk of sound wood, has also lost by a process of slow combustion (decay) a part of its heating value. The term "a cord of wood," then, has no exact calorific value. " A ton of coal" is a more exact term, as to heating value, but this quality is also largely de, pendent onthe kind of coal and on the thoroughness of its combustion. Professor Johnson's experiments on American coals showed an extreme variation of 3'16 lbs. of water evaporated from 78~ per pound of coal. The "Report of Coals suited to the Royal Navy," showed an extreme variation of 1'47 lb. of water evaporated per pound of coal. And as to combustion, estimating a ton of coal to contain 100 lbs. of hydrogen, which is in itself equal to 300 lbs. of charcoal, and * On the Eastern division of the New York Central Railway, the following were the results of a change of grate, in ease of the Engine No. 76: Size of grate. Size of exhaust orifice. Mdiles run to 1 cord of wood. 1872 sq. in. 2- inches diameter. 35'15. 405 sq. in. 21 do. do. 47'42. t Ure's Dictionaly. 72 ITHE ECONOMICAL GENERATION OF STEAM..whichl also abstracts some 300 lbs. of gaseous carbon from the coal, when the carburetted hydrogen, or what is popularly called " smoke," is entirely unconsumed, it appears that in burning the solid carbon only,of a ton of bituminous coal, 600 lbs. of what is equivalent to good coke or carbon, are wasted. The general practice would seem to establish the fact, that a ton of bituminous coal is equal to at least 1-f cords of wood. From this datum and from the comparative cost of wood and coal in different sections of the country, the economy of introducing coal as fuel may be readily estimated. In case of old and ordinary boilers, which do not embrace the improvements considered in the foregoing chapters, a deduction of probably I cent per mile must be made from the saving thus estimated, for the extra cost of repairs due to coal. But this deduction is believed to be fully offset by other elements of expense due to wood fuel; viz.: the cost of woodsheds at numerous points on the line; the gradual, but almost unavoidable decay of wood; the proportion of loss of wood, buildings, and bridges, by conflagrations-there are few lines that have not suffered seriously from this cause; the extra transportation of wood-it is often carried twice over most of the line, first on wood trains and then on the tender;* the cost of stopping heavy through trains for wood and the cost of long stoppages; the irregularity of steam generation and hence of running time, due to the irregular quality of wood; and lastly, the gradual depletion of an almost continuous and unguarded wood-pile, in the night, from obvious causes. The locomotive superintendent of a certain line having made a system of observations in this direction, concludes, that what would be equivalent to one-quarter of the population within half a mile of his wood-sheds, were supplied with fuel at the company's cost, till the introduction of cl, vlwhich is now carried from the seaboard chiefly on tenders, and is therefore mostly employed for generating steam. Again, wood is constantly becoming more scarce and expensive, while the opening of new coal mines, the adaptation of mining machinery, the improvements in transportation, and the general results of an increased demand-the supply being inexhaustible-will be likely to cheapen the latter fuel. There are probably few, if any railways in the Northern States, whereon wood will be found to be the cheaper fuel, all things considered. The general rush of the Northern railways into the substitution of coal for wood-a remarkable feature in the railway practice of the last year-especially of the New England railways, by reason of the high price of wood, and of the North-western railways, by reason of the cheapness of coal, together with the notorious economy due to this change, appears to establish this conclusion. The United States is peculiarly a bituminous coal country. The anthracite coal-field of Pennsylvania, according to Overman, has anl area of 437 square miles. The bituminous coal-fields of the entire country are considerably over 150,000 square miles in extent. Ohio alone, which is the ffth in rank as a coal-bearing State, contains more of that fuel than all Great Britain. The bituminous coal of America is deposited in fields of greater or less extent from the great basin of Nova Scotia on the east, to the upper part of California and to Washington Territory on the west; while it ranges also from Michigan on the north to Central Alabama and Central Texas on the south. The principal bituminous coal-field, including also the semi-bitumninous varieties, extends from near the southern boundary of the State of New York to Central Alabama, having the following approximate areas respectively in nine States: Pennsylvania,........ 15,000 square miles. Ohio,.... 11,900 " Iaryland,............ 550 * Even in case where coal is hauled in a tender to a greater distance than wood,. the farther transportation of the former is believed to be met, as follows: " 1st. By the greater oeight which has to be hauled a less distance, where wood is used, since about 2~ lbs. of wood are required to do the work of 1 lb. of coal. "' 2d. By the fact, that in ascertaining the cost of woool delivered on the tender at the different stations (as stated in the tables), only the ordinary train expenses were included, for the large portion which requires to be hauled; and the cost of repairs to engines and cars, and of maintenance of track due to such transportation of wood, were not counted. In computing the price of coal, all the items that enter into the cost of transportation werle included." —Report on the Pen.nsylvania, R. experiments, INTRODUCTORY. 73 Virginia,........... 21,195 square miles. Kentucky,............ 6,000 " Tennessee,....'. 4,300 " North Carolina,......... 250 " Georgia,....... 150 " Alabama,........ 3,400 " Total,. 62,745 This vast field contains several varieties of coal among the best on the globe. At the extreme north, near Towanda, Pa., and at the threshold of the great canal system of New York, are extensive beds of very pure coal, lately opened by the Barclay Coal Company. The Pittsburg and Briar Hill coals in the same great formation, are of well-known excellence. The semi-bituminous coal of the Cumberland region is everywhere celebrated. Seventeen years since, Professor Johnson proved by a most elaborate series of experiments, that the Cumberland coal possessed greater excellence, in nearly all respects, than ally other fuel then competing for the supply of the National Steam Navy. Subsequent experience, including other coals since brought into market, has confirmed this preference. For seventeen years, this coal has been the principal fuel burned in the engines of the Baltimore and Ohio railway, and it is now indispensable to the success of nearly all the modern patented boilers. With the exception of locomotives running immediately within the anthracite region, nearly all coal-burning engines employed near the seaboard, burn the coal of the Cumberland region, exclusively. The excellence of this coal, the large scale on which it is worked, and its convenience to market, must, as coal-burning is developed in railway locomotives, make it the great staple fuel for many of the roads in the North-eastern and Middle States. The characteristics of the Barclay, Pittsburg, Cumberland, and other of the choice varieties found in the great American coal-field, are these:-They ignite readily, burn with a free, clear flame, contain a high percentage of combustible matter, have little sulphur, little sand and mineral oxides forming clinker, and they leave but comparatively little ashes. No fuel could possess qualities more desirable for its adaptation to railway locomotives. The great Illinois coal-field, extending into Indiana and Kentucky, has an extent of upwards of 50,000 square miles. As we go west, the coal strata are thinner, however, and less numerous, and the coal deteriorates in quality. So in Illinois, although time will doubtless develop many locations affording better coal. Iowa and Missouri are well supplied with coal. In the Bear River Valley, west of the South Pass and near the summit of the proposed South Pass Pacific Railway route, coal is reported as abundant. It would be impossible, in view of the many variable and accidental conditions mentioned, to estimate, with mathematical, exactness, the advantage due to the substitution of coal for wood. On each railway the results would be somewhat different, and each railway can best make its own estimates —the foregoing suggestions have been intended merely to give direction to such inquiry. In any locality where a ton of coal does not cost more than 1 cords of wood on the tender, it may always be safely concluded that coal is the more economical fuel. And in special cases, where wood is poor andcoal is peculiarly good, a much larger difference of cost would still render coal the more economical fuel. Since so large a proportion of American railways is committed to the use of coal, the more important question is, not the economy of using coal in place of wood, but how to bmrn coal economically; and this will form the subject of the following chapters. 2. How TO ESTIMATE ECONOMICAL COAL BURNING. — The reasons for not adopting the results of those special experiments on coal burning, so frequently paraded, wherein perhaps half the elements of accurate conclusions are wanting, and the propriety of substituting, as a basis of comparing different schemes, the visible phenomena and the known principles involved, have been treated in the introduction to this work. Economical coal burning, or, in fact, economical steam generation, measured by this standard, can only consist: — 10 74 THE ECONOMICAL GENERATION OF STEAM. 1st. In the thorough combustion of the solid carbon of the coal and of the whole of the gases evolved from the fuel, by the proper arrangement of grate, furnace and draft apparatw s, and by a regulated supply of air to the whole of the fuel at the right point. There are t wc reasons why the combustion of the gases is necessary. lsto If not consumed, they form smoke which is a nuisance; and it should be prohibited by law, as it is in Great Britain. 2d. If tho gases are not burned, from I to i of the fuel, depending on its character, is wasted. 2d. In promoting the durability of the boiler: —lst. By preventing the destructive effects of the more intense heat of coal, by thorough circulation, 2d. By resisting the increased wear due to the harder material, by the use of harder flues and fire-plates. 3d. By preventing ine accumulation of scale. 4th. By sounder boiler construction generally. 3d. In facilitating the most thorough transmission of heat by heat-traps and an equable oraIt. 4th. In promoting the highest possible ultimate vaporization, by good circulation; jby utilizing all the heat generated; by drying the steam; by a supply of feed-water and of air for combustion, independent of the working of the engines; and by preventing the accumulation of scale. 5th. In so encouraging and rewarding the skill and faithfulness of enginemen and firemen, that the conditions of the economy indicated shall be complied with. It is probable that ignorance and carelessness in working locomotives, can offset all the benefits which might be derived from improved apparatus and machinery. And these conditions do not apply entirely to new locomotives. 3. ADAPTATION OF THE PRESENT CLASS OF ENGINES TO COAL BURNING.-It is estimated that at least 9,000 locomotives are in service on the railways of the United States, and that above 5,000 locomotives are employed in Great Britain. Probably not above a hundredth part of these 14,000 engines, which represent not a market value, but a worth to their owners, of perhaps $100,000,000., were especially designed for the use of bituminous coal. In England, they were designed for coke, which, as far as the effects of heat and wear are concerned, is more trying than coal; but they were not originally adapted to the combustion of gases. In America, thewood-burning engines are still less adapted, primarily, to the use of coal, since they have smaller water-spaces, softer flues, limited combustion-room, and no means of supplying air to the gases. Supposing 5,000 coal-burning engines to be required, under these circumstances, what is the most economical method of supplying them? 1st. This number of new engines should cost at least $50,000,000. There are at the same time two distinct methods of adapting old engines to coal-burning. 2d. The structural alteration of old boilers, that is, a combustion-chamber, a new and better fire-box with wider water-spaces, and new and harder flues, which is substantially the best arrangement, as far as the boiler, strictly, is concerned, for burning coal. This would cost from $2,000 to $3,000 for each boiler. 3d. Compelling good combustion to take place in limited combustion room, by the forcible mixture of air with the gases at a point where they wouldignite; promoting good vaporization and the durability of heating surfaces by inducing good circulation in a manner not required in boilers specially adapted to coal; and gradually supplying structural alteration and better material in the ordinary course of repairs. The entire cost of the first adaptation to coal by thevarious plans to be considered under this head, will not exceed, say $250. Such being the methods of producing coal-burning engines, one other consideration is nvolved:~What are the comparative results as to combustion, vaporization and repairs, of the three methods mentioned? It is believed that these results will appear so nearly alike as not to wvarrant the purchase of new engines except where more power is required; or the structural alteration of old engines, except where extensive renewals of boilers are necessary. 4. POPULAR ERRORS WITH REFERENCE TO COAL BURNING. —It was formerly believed that the efficiency of coal was in proportion to the carbon in it; that the gas from it did not yield more than enough heat to distil it. And since there was no economy in its use as compared with coke, INTRODUCTORY. 75 but a great disadvantage in the nuisance of its smoke, no decided effort was made to introduce it as locomotive fuel. A considerable number of inventors, however, are to this day trying to burn smoke-not combustible gas but incombustible smoke-by returning it to the fire, and in other ways which are too absurd to deserve notice. Thirty years ago a distinguished chemist, Mr. Goldsworthy Gurney, testified that he had, with great difficulty, succeeded in burning smoke. He passed it through sand mixed with quicklime, by which the carbonic acid was absorbed, and the remaining carbon rendered combustible. He knew of no mode likely to succeed in practice; nor did he conceive it possible to burn coals so as to consume all the volatilized matter; the combustion of smoke was prevented principally by the particles of carbon being enveloped in carbonic acid; and he did not think its combustion could be effected by any quantity of air. This appears to have been the theory among engineers, and the whole theory. It is true; but not the whole truth: smoke once made cannot be burned, by practicable means; but gas can be burned without making smoke. Singularly enough, there are "Practical' Railroad men" at the present day, whose conclusions are no farther advanced than those of Mr. Goldsworthy Gurney; who say that "smoke is smoke," that they can see' it rise from the coal when thrown upon the fire, pass through the flues and come out at the chimney top; that the use of coal'necessitates the toleration of smoke; that the admission of air above the: fire (which they have heard about) is absurd, not only because "air would not burn smoke," but that since " coal can be burned without air above the fire, it cannot be burned with it." There are not many, indeed, who say as much as this, but there are hundreds of cars running at this day, on American and English railways, which are insufferably foul with the' smoke and unconsumed gases of bituminous coal. "' Actions speak louder than words." A second variety of unbelievers admit the possibility of "burning smoke" by letting air into it-they have tried it; but it cooled down the boiler so much that it did not pay. It turns out that their experiments consisted in keeping the fire-door open all the time. The air was admitted in bulk and not in the right place. A third class of practitioners believe in the necessity of admitting air in comminuted streams, so that it shall mix thoroughly with the gases, but they have not heard of the third element in the combustion of gases-heat to set the mixture on fire; so they admit air wherever it happens to be convenient-in the end of the combustion-chamber, or at the top of the fire-box; they are disappointed at the result. They would be hardly surprised, however, if no flame should result from simply opening a gas cock without applying fire to the burner —here is perfect mixture of air and gas. Either the mixture must be forced to occur, in a locomotive furnace, at the surface of the fire, or heat enough to ignite it (say heated fire-bricks) must be placed where the mixture tends naturally to occur. Several kinds of patented locomotives have been made in defiance of these principles; and they make so much smoke as to be a nuisance. A fourth mistake is to attempt to get air enough (i. e. the oxygen in it) through the fire to form a complete mixture with the gases. This is not done, however, in practice, with the most shallow fire that will give heat enough for the necessary steam generation. And there can be little economy in it; the most air is required above the fire when the fire is the thickest —when fresh coal has been added; that is to say, when least air can get through it. And if the fire is very thin, air that is cooler than the plates will pass into the fire-box. Frequent firing of a morsel at a time would reduce and equalize the distillation of gas so much that the air coming through the grate or leaking into the fire-box would be sufficient for it; but when the fire-door is thus almost constantly open, great volumes of cold air are rushing through the flues. Admitting a regulated supply of air above the fire is far easier as well as more economical. A fifth mistake is to make too thick a fire. As shown by experiments in France, mentioned by Clark, fires from 18 to 40 inches thick generated a considerable amount of carbonic oxide, which, in the absence of air above the fire, was a waste of fuel. A sixth class of experimenters are enthusiastic on the subject of time and room for the thorough mixture of air and gases. When this mixture is not forcible —when strata of air and gases move in one general direction by the action of the blast, too much importance cannot be attached to the time occupied'by the movement of the currents, before they get beyond the reach of ignition into the flues; and this certainly requires room —either open space in combustion-chlambers or distance 76 THE ECONOMICAL GEONERATION OF STEAM. developed in passing around deflectors. Since very long flues have disadvantages as to transmission of heat and durability, a combustion-chamber is not unimportant; and it is more likely to insure combustion, however the air and gases may have been mixed. But all this may be carried to extremes, as has been done by Beattie (Plate 391), and would be done in the shape of bridges and water-walls and chambers innumerable, by many others, could they have their own way. If any simple means of forcing the mixture of air and gases, at a point where they will ignite, aided by the careful and skilful firing and working of the engine, will produce good combustion, it is not economical to complicate boiler work and masonry to such an extent that imperfect combus. tion cannot occur under any management-at least until thin steel is supplanted by thick iron, for fire-plates and flues. A seventh class of "practical men " think, on the contrary, that coal burning is the most " simple " of problems. All they require is " to let the men get used to it," so that they shall not allow the grate to become clogged nor the bed of coals matted and heavy. " A little " smoke occurs, of course, but they burn coal with decided economy over wood, and that is the conclusion of the whole mater. There is a great deal of truth in this view of the case, but not the whole truth. It is a nuisance to have to ride after this kind of coal burning; and because coal is cheaper than wood, it does not follow that coal gas must be distilled and then thrown away, especially since its combustion is a very simple problem. Lastly, a considerable number of locomotive superintendents are willing and anxious to derive all possible advantage from the improvements which are developed from time to time. They have not quite faith enough in the published reports to risk a novelty without some experiments —so they " try it." Several locomotive superintendents have tried Clark's jet (Plate 56), which is the forcible induction and mixture of air with the gases, at the surface of the fire, by jets of steam; and they have abandoned it as inefficient. Instead of approximating to the capacity and position of the apparatus used by Clark, which is at least six 2-inch hollow stay-bolts at the surface of the fire, for admitting air, one of these experimenters had put three w-inch stay-bolts immediately under the door, with steam jets blowing air through them straight into the flues. Another, who had heard that air above the fire is a " good thing," put in four hollow stay-bolts right under the crown-plate, on each side. Their diameter was { inch each. And having tried it, he " did not think there was much in it." In this manner, many railway companies are deprived of important benefits, and the whole subject of railway improvement is brought into popular discredit, just as it is when counting-room officials, personally ignorant of. the requirements of the case, are led into expensive failures by the splendid promises of crazy or dishonest patent-owners. It has been deemed appropriate to premise a detailed discussion of the subject of coal burning, by the mention of these leading obstacles in the way of its rapid and satisfactory progress, and by the other introductory matter of this chapter, for the purpose of reminding those who are not attempting to improve their present practice, that there are great improvements within their reach, and those who are waiting to see what is the best plan, that their time has come-not by the ultimate perfection of the process, but by its adaptability and feasibility under any circumstances where a ton of coal is cheaper than 1I cords of wood. CHAPTER II. THE COMBUSTION OF BITUMINOUS COAL, 1. GENERAL PRINCIPLES.-No substance in nature is combustible of itself, to whatever degree of heat it may be exposed. 1st. It can be ignited only when in presence of or in mechanical mixture with air, or its vital element, oxygen. 2d. Combustion is continuous ignition, and can only be made to exist by maintaining in the combustible mixture, the heat necessary to ignite it. Coal is composed, as far as combustion is concerned, of solid carbon and a gas consisting of hydrogen and carbon. When the coal is heated, it first discharges its gas; the solid carbon left then ignites in presence of oxygen, and will retain the temperature necessary to combustion so long as oxygen is supplied. The combustion of the gas is illustrated by the flame of a candle; the flame, which is the same gas and oxygen in the process of ignition, furnishes an igniting temperature to the succeeding mixture of gas and air, which in turn ignite, furnishing heat to the next atoms of the mixture, and so on. But this self-generating succession of flame, as in case of the candle, requires regularity and continuity in the supply of gas and in the mixture of air; when the mixture occurs away from the influence of the igniting temperature which distilled the gas, or when the supply of either air or gas becomes so irregular that the combustion temporarily stops, an ignitingtemperature must be again supplied. As this will appear to be the usual state of affairs in the firebox of a locomotive, the mere distillation of the gas cannot be depended upon to keep the combustible mixture ignited, and other means must be supplied. Since this third and essential element of combustion-the maintenance of an igniting temperature-has been frequently overlooked in the theory and practice of coal burning, it has been deemed important to call attention to it thus early in the discussion of the subject. 2. PHENOMENA OF COAL BURNING. —The principles to be applied in designing boilers may be deduced from careful observation of the manner in which coal is burned in ordinary wood furnaces, the observer being at the same time aided by a moderate knowledge of chemistry. With a fire already lighted, and steam raised in an ordinary wood-burning engine, we place 500 pounds of bituminous coal in the fire-box. The engine being in motion, we observe some or all of the following results: — For the first few minutes, the flame in the fire-box is darkened or nearly extinguished, the pressure of steam goes down, and a dense cloud of smoke and fine sparks comes from the chimney. If the coal is very sulphurous, this smoke has the stench of rotten eggs. After some minutes, the fire becomes bright, the escape of smoke diminishes and the steam pressure rises to its maximum point. There is now a strong local heat in the furnace, the temperature being greater than with wood. After some further time, the grate becomes, perhaps, clogged with clinker, and the surface of the fire, which at first had swelled as the coal was ignited, sinks in a crust of greater or less thickness. Allowing the fire to bumrn down and bringing the engine to a stop, we find considerable fine coal in the smoke-box and that the flues are coated with soot. The evaporation of water, for each pound of coal burned, will not have exceeded 5 or 58 pounds. Continued use of the coal may result in burning out the grates and furnace, and in the leakage of the flues. 78 THE ECONOMICAL GENERATION OF STEAM. 3. EXPERIMENTAL EXAMINATION OF THESE PHENOMENA.-On opening the door, soon after the coal has been placed on the fire, we may observe that the smoke will be diminished inquantity and intensity, although the steam pressure will go down, showing that the furnace is cooled by the admission of air. The smoke which escapes from the chimney will deposit soot on any surface exposed to it. A sheet of white paper would, in a short time, become blackened if held in the smoke. But the dense vapors formed in the fire-box will not, as they leave the coal, deposit any soot. In an open parlor grate, just after fresh bituminous coal has been thrown on, the dark products of distillation, by many called " smoke," will not blacken the whitest cambric handkerchief. Hence, whatever may be the visible resemblance between the dark cloud, as distilled from te coal, and the smoke at the chimney-top, they are. evidently different in at least one respect. Besides, the dark'vapor in the furnace ignites and gives out flame as the air gains access to it, while the smoF,P, once formed, cannot be burned by any practicable process-by heating, mixing with air, or otherwise, But while the vapor, which we shall now call gas, may be burned in the fire-place, it is not of itself combustible. If this gas be caught in a jar, without intermixture with air, a lighted taper, thrust into it, will be instantly extinguished. And if enclosed in a tight vessel, it might be exposed for any length of time in a furnace at white heat, without taking fire. In ordinary gas works, the same kind of gas is expelled from coal in cast-iron vessels heated to redness, and no combustion takes place. It will burn only in contact with air. Again, while this gas burns regularly and silently when. left to come into gradual mixture with air-precisely as the same kind of gas burns in an ordinary illuminating burner-the same gas, previously mixed with about nine times its weight of air, will, when subsequently ignited, explode like gunpowder. By " shutting off" the steam from the engine, or closing the damper, the smoke is much increased. Yet the oniy direct effect of these manipulations is the exclusion of air from the fire. If a close jar be filled with the coal gas, and another be filled with common air-while neither can be ignited, by itself, by any degree of heat-a jet of air led into the gas-jar can be ignited and burned as it escapes; Whvlile precisely the same flame may be also produced by a jet of gas led into and ignited in the air-jar. The combustion of either is simply its chemical combination with the other. But this chemical combination will not occur without the maintenance of the requisite heat. If we place so much fresh.coal upon the fire that the gas distilling below is cooled by passing through the cold mass above, and is also thus protected from the heat of the coals below, no combustion will occur by the mere admission of air above the fire. While the gas is being expelled from the coal, the latter remains at a low heat-no particle of solid coal (which is then becoming coke). can burn while gas is issuing from it. This, of course, refers to coal in its respective particles, as a lump of coal may be giving out gas in one place while all has been expelled from another, and the remaining coke already ignited. This precedence in the burning of the gas is proved in making coke or charcoal. By admitting just sufficient air for the combustion of the gas in the raw fuel, this gas only is burned and the coke or charcoal pure carbon —is' left behind. Indeed, but for this distinction in nature, we could not manufacture either coke or charcoal, excepting by distilling the gas from the raw fuel in close vessels, and by the consumption of a considerable quantity of additional fuel in producing a high external heat. As coke does not produce smoke in burning, the smoke issuing from our experimental engine must be generated from the gas, in the combinations which it forms before reaching the chimney. The duration of smoke, therefore, measures the time during"which the gas is distilling from the coal. If we provide means for watching the course of the flame, we find it never enters a flue of ordinary size (1:F to 2 inches), for more than a few inches from its mouth. However near the flues are placed to the surface of the fire, the flame is extinguished immediately on entering them. Hence, all the combustible matter contained in the flame, at the moment of its extinction, is lost, as it can only impart its full heat by its complete combustion, - The intensity of heat from coal, or any other fuel burned in air, is a fixed and inevitable degree for each kind, and depends on the nature of its constituents. By the perfect combustion of seasoped wood (which, however, contains about 20 per cent. of water), the resulting heat is 2867~. Green turf gives 2732~. Bituminous coal, of average composition, 4082~. Anthracite coal, 4170~. Dry coke or charcoal, 4352.~ THE COMBUSTION OF BITUMINOUS COAL. 79 The effect of clinker and of crust on the surface of the fire is to prevent the passage of air. The clinker is mainly sand or clay, vitrified or reduced by heat, and sometimes mixed with mineral oxides, which by their fusion still further increase its amount and strength. The quantity of clinker depends entirely on the cleanliness of the coal. The only specific against its effects is to keep it well broken and cleared out. Thus, by these observations and by reference to a few well-established and familiar facts, we find: 1st. That the gas (distilled from the coal) and the smoke in the chimney are not the same. 2d. That the gas is expelled from the coal before the latter (or its cole) commences to burn. 3d. That the production of the smoke commences and ceases with that of the gas. The smoke is, therefore, a product of the gas in the combinations which it forms before reaching the chimney.* 4th. The gas, by itself, is not inflammable. Neither is the coke by itself (without air) combustible. 5th. The smoke is in no case inflammable. 6th. The presence of air enables the gas to ignite and burn, and, to a corresponding extent, diminishes smoke. So, too, it enables the coke to burn after the gas has been expelled. 7th. Heat enough to ignite the mixture of air and gases must be maintained in the mix. ture. 8th. Small tubes effectually extinguish flame. 9th. The degree of heat, evolved in the combustion of any kind of fuel, is fixed and inevitable for that kind, dependent only on its constituents, and independent of the form, dimensions or material of the furnace. 10th. Clinker is mainly sand, (or its base, silicon), and its amount depends on the cleanliness of the coal. We may note, also, that the volatilization into gas abstracts considerable heat from the burning coal-the steam generally falling on the application of fresh coal to the fire. 4. EXPLANATION Or THESE PHENOMENA-,CHEMISTRY OF COAL BURNING.-The chemical principles bearing' on the use of coal in locomotives, although few in number, cover the entire ground, leaving no room for speculation. Wherever they have been substantially embodied in the plan of locomotive boilers, the most bituminous quality of coal has been.burned without smoke, sparks, or offensive gas, and with a complete control of the steam pressure -and a high rate of evaporation in proportion to the fuel consumed. And in the few cases where such a result has been approximately attained without reference to these principles, they have nevertheless been approximately embodied, and were the only cause of the results produced, the theories of the inventors to the contrary notwithstanding., Composition of Coal.-Coal is a compound substance, which may be decomposed by heat into several distinct elements-generally five or six at least. So far as relates to combustion we are concerned principally with but two of these, viz.:-Solid carbon, represented by coke, and hydrogen, generally known under the indefinite term of "gas." These two elements contain practically the full heating qualities of the coal. The carbon, so long as it remains as such, is always solid and visible. The hydrogen, when driven from the coal by heat, carries with it a portion of carbon, the gaseous compound being known as carburetted hydrogen. A ton of 2,000 pounds of average bituminous coal contains, say 1,600 pounds, or 80 per cent., of carbon, 100 pounds, or 5 per cent. of hydrogen, and 300 pounds, or 15 per cent. of oxygen, nitrogen, sulphur, sand and ashes. But if this coal be coked, the 100 pounds of hydrogen driven off by heat will carry about 300 pounds Of carbon in combination with it, making 400 pounds, or nearly 10,000 cubic feet of carburetted hydrogen gas. Thus but 1,300 pounds of carbon (65 per ~ In case of the excessive distillation of gas by thick firing and an intense draft, some gas may pass out of the chimney as well ai smoke. But a combustible mixture in the smoke-box is of no importance; besides, when the gas gets into the flues it is stripped of the air necessary to combustion; it cannot be burned simply for the want of room. THE ECONOMICAL GENERATION OF STEAM. cent. of the original coal) will be left, and, with the earthy matter, ashes, sulphur, etc., retained with it, the coke will weigh but about 1,350 or l,400 pounds,~671 to 70 per cent. of the original coal. Thus for every 2,000 pounds of coal, we have, say 1,300 pounds of solid carbon, and 400 pounds of carburetted hydrogen to be burned —the remaining 300 pounds being waste-partly gaseous and partly solid. The Operation of Burning.- The visible phenomenon of burning (combustion), although familiar to the senses, is, by no means, generally understood. For instance, we say coal, thrown upon a fire, commences immediately to burn. Now, before any burning can possibly commence, the coal must suffer the preparatory process of decomposition. Its constituent elements must be separated, after which a regular order of precedence maintains in their combustion. The burning which then occurs is this: the gas, which, having been distilled, invariably burns first, does so merely by its chemical union with the invisible oxygen of the air. Since the process is simply that of combination, the oxygen is burned precisely the same as the gas, and the combustion is entirely a mutual process. The gas having been burned, the coke takes its turn, and burns in exactly the same manner, by combination with air. In this case, also, the air (or its oxygen) is burned equally with its combining element-the coke-and in every respect, both in the activity of the fire and the heat evolved, plays an equally important part in the process. This is proved, also, by its presence in the resulting product, which, with coke, is carbonic acid-an invisible gas, containing definite proportions only of carbon and oxygen. That the combustion of the elements of coal in air is a mutual process, is proved by the fact that neither of these, by itself, can possibly be burned. No heat can burn hydrogen excluded from contact with air-as is seen in the generation of gas from coal in close red-hot retorts. That air does n6t burn, excepting with some mutual supporter of combustion, is sufficiently known. Combustion and Explosion.-But if the entire atmosphere surrounding the globe could be once saturated with one-thirty-sixth of its weight of hydrogen, the first spark would produce an Instantaneous and universal explosion. The same gas escaping from gas-pipes and mixing with air in a close room, has often produced disastrous explosions, throwing down buildings and crushing the occupants. A familiar phenomenon, occasionally attending the lighting of fires in stoves, shows the same results. If when a fire has been lighted at the bottom of a stove, the upper part be packed full of pine shavings, the draft through the stove will be stopped-the heat below will distil the gas from the shavings, which gas will saturate the confined air above, and when a spark of the fire finally gets access to the mixture, an explosion of greater or less energy occurs. The same gas, evolved in coal-pits, and mixing with the air entering the shafts, is the fatal "firedamp " of the miners. And it is precisely the same gas that is generated at the rate of 10,000 cubic feet from every ton of ordinary bituminous coal burned in a locomotive furnace. If the gas, expelled from half a ton of strong coal, were to combine with its saturating equivalent of air on the instant when the two were presented to each other in the furnace, their explosion would dash the engine from the rails. Nature, however, has provided against such dangerous combinanations, by giving to the separate fluids very different densities, and such moderate affinities for each other, as will induce mutual mixture only where the two are confined in close contact, as in a glass jar, a close room, or in a coal-pit, and in these situations only after a considerable time has elapsed. Flame is only the continuous explosion of successively combining atoms of the gas and air, and which, had they been previously mixed in any considerable quantity, would detonate in the furnace like gunpowder. It is simply because this very mixture requires TIME for its accomplishment, that the explosoin can only go on by successive atoms, thus forming continuous fame. Chemists have sometimes found it necessary to give from four to twenty-four hours for * The late Dr. Drake, of Philadelphia, designed an engine in which coal-gas, mixed with nine times its weight of air, was exploded alternately in each end of the cylinder, and the power thus developed was applied through the ordinary piston and connecting rods to a shaft. The initial explosive power of the gas was stated to be between 150 and 200 pounds per square inch. The engine was built on a large scale by a prominent builder of Philadelphia, and was publicly exhibited at the time. THE COMBUSTION OF BITUMINOUS COAL, 81 the complete atomic mixture of gases already placed together in the same vessel. Having seen the violent consequences attending the ignition of gas and air after their mutual saturation, we can comprehend the phenomenon of flame which attends their ignition daring the same process of mixture. It is only in the condition of flame, in the natural process of combustion, that the gas and air can develop their useful heating power. Proportions of Air and Gas necessary to Combustion.-We have already seen that the burning or combustion of this matter is simply its chemical combination with air, or, more strictly, with the oxygen of the air, since the other element of air, nitrogen, passes through the fire uncombined, and without playing any other part in the process of inflammation than that of an abstractant of heat. Here we meet one of the great laws of nature —that chemical combinations can occur only in fixed and determinate proportions of their constituents. The carbon and carburetted hydrogen to be now burned, will each combine with air in fixed proportions only-these proportions being known as chemical equivalents. We have now to observe these proportions and provide the due equivalent of air, with the same care and exactness as we would that of the coal. As well might we expect half a ton of coal to do the work of a ton, as that a given bulk of air could be made to do the work of double that bulk. If 1,300 pounds of carbon require 200,000 cubic feet of air for their combustion, we might as soon hope tp develop the heat of the whole by burning 650 pounds of carbon, as by burning 100,000 cubic feet of air. We provide the coal, indeed, at a cost, in cash, and through the labor of placing it on the grate, while the air is generally ever present, and ready to combine with the coal almost without cost or labor on our part. What we have then to do is to permit and not to obstruct the access of the air to the coal, this access of air being permitted under such circumstances as shall favor its complete combination; or when the arrangement of the fire-box happens to be unfavorable to this combination, we must facilitate it by forcible means. Diagram fig. 5, plate 37, shows exactlythe substances with which we have to deal in burning the gaseous matter of coal, and the combinations which are formed in their combustion. The elements and their mixtures are represented by colored surfaces, of different sizes.* The left hand column of squares shows the relative bulks of the gas and air to be burned. The air is seen to have ten times the volume of the gas. The figures show the relative weight of each, whereby the required air will be seen to have eighteen times the weight of the gas. The middle column shows the gas and air separated or decomposed into their elements. The hydrogen alone now forms twice the original bulk of the whole gas, while the carbon vapor also forms one-half of the same total bulk. The elements of the air occupy separately the same room as when combined. In the third or right hand column of squares we have the products of combustion..The carbon has united with twice its bulk of oxygen, forming carbonic acid, equal in bulk only to that of the oxygen alone. The hydrogen has taken up one-half of its bulk (eight times its weight, however) of oxygen and formed steam, equal in bulk only to the hydrogen. The nitrogen of the air, forming four-fifths of its bulk, has passed through uncombined, and has had nothing to do with the process of combustion, excepting to abstract heat for itself. The specific gravities of the substances just considered, show the relations between their volumes and weight, both as simples and compounds: Atmospheric air (oxygen, I bulk; nitrogen, 4 bulks),..... 1'000 Oxygen,............... 1104 Nitrogen,..............'972 Carburetted hydrogen (carbon vapor, 1 bulk; hydrogen, 4 bulks),....'555 Hydrogen (the lightest gas known),.........'070 Steam (oxygen, 1 bulk; hydrogen, 2 bulks),........'624 Carbonic, acid (carbon, 1 bulk; oxygen, 2 bulks),....... 1524 * In some compound gases the relative bulks or densities of their elements are different from the bulks or densities of the same gases when separated. That is to say, two cubic feet of hydrogen will unite with one of oxygen, and the product, instead of being three, will be but twoo cubic feet of steam. The bulk of atmospheric air is the aggregate of the bulk of its constituents. 11 82 THE ECONOMICAL GENERATION OF STEAM. The only proportions in which carbon and hydrogen combine with air in combustion, are these: For every pound of carbon (pure coke) twelve pounds, equal to 159ff cubic feet, of air, are required to be combined intimately with it. For every pound of hydrogen, thirty-six pounds, equal to 4781 cubic feet, of air, are required to be similarly combined. Thus for every pound of carburetted hydrogen gas, being one-fourth pound of hydrogen and three-fourths of a pound of carbon, eighteen pounds, equal to 239i'cubic feet, of air, are required to be combined with it. These are the elements, and their combining proportions, as we have to deal with them in a locomotive furnace. For every 2,000 pounds of coal burned, the 400 pounds of carburetted hydrogen-the " gas "-requires 95,700 cubic feet of atmospheric air, at ordinary temperature, and the 1,300 pounds of solid carbon, require 207,350 cubic feet of air. Practically, the " gas " from a ton of ordinary bituminous coal, requires 100,000 cubic feet of air for its combustion, while the remaining coke requires 200,000 feet. Thus the gaseous matter of the coal, requires one-half as much air as is taken up by the solid coke. The heating value of any combustible is exactly proportional to the quantity of air with which it will combine in combustion. Hence hydrogen, which combines with three times the quantity of air (oxygen) which would be taken up by carbon, has, for equal weights, three times the heating value. Thus, the 100 pounds of pure hydrogen, in a ton of coal, have the same heating efficacy as that due to 300 pounds of the remaining carbon or pure coke. Arrived at this stage, we now have something like an exact knowledge of the elements to be burned, and of their relative combinations. Where these combinations are completed, combustion will be perfect. A given quantity of gas, completely burned, cannot produce smoke, since smoke contains a quantity of unburnt matter, and is in itself a proof of incomplete combustion. The products of perfect combustion are invisible, being, for carbon and oxygen, carbonic acid; and for hydrogen and oxygen, invisible steamr, which condenses into water. It is especially at this stage that we realize the true functions of a coal-burning furnace. It is simply the vessel in which given elements are to be chemically combined in definite proportions. So far as the structure of the furnace may aid these chemical combinations (that is to say, combustion), by giving room, direction, and time, for mechanical mixture and ignition, so far will the furnace be efficient in burning coal completely, which is the same as burning it economically. Intimate Mixture of Air and Gas necessary to Combustion.-Among the most essential provisions in the arrangement of the furnace, are those which operate to induce or solicit the mingling of the coal gas and air, since both these are to be burned, and, as we have seen, neither can burn without the other-they can only burn together. Thus, wherever the air is induced to enter by the draft, whether through the grates or through openings in the door, the admission spaces should lbe small and numerous, so as to present as many surfaces of contact as possible. Or when it is supplied by a force separate from the draft, in large bulks, it should be comminuted and mixed with the gases by the same force, after entering. The difference in the sizes of air-jets is precisely the.same, in principle and in effect, as the difference between fine and coarse gunpowder. Dr. Ure says of this compound, " That nitre does not detonate until in contact with inflammable matter: whence the whole detonation will be more speedy the more numerous the surfaces of contact." So in the gradual explosion (combustion) of gases; the detonation is superficial only, and can go on no faster than the respective atoms of the gases can combine. Large Surface-contact of Air and Gas. —In ordinary gas-burners, the " bat's-wmng," or slotted nozzle, gives more surface-contact for air than that with round holes simply, and hence gives a better light for the same quantity of escaping gas. And when we say a better light, we are to understand only that more of the escaping gas is burned in the bat's-wing than in the perforated burner —as in the latter a quantity of gas passes through without being burned. Argand, by makling a hollow wick, doubled the contact-surface of air and gas in the common lamnp-flame, and thus doubled the illuminating power of the lamp. Since the date of his invention, in 1780, it has been universally adopted in all large oil lamps, and still bears the inventor's name. THE COMBUSTION OF BITUMINOUS COAL. 83 Charles Wye Williams applied the same principle of increasing the contact-surface of air and gas in an ordinary furnace, thus greatly increasing its heating power. His plan, first brought out in 1839, is known as the Argand Furnace, and, to this hour, forms one of the most essential features, so far as it is used, in every successful coal-burning locomotive boiler. Figs. 1, 2, and 3, Plate 37, show different conditions of flame as due to different modes of mixture with air. Each represents a tin apparatus with a glass chimney. The gas is admitted the same way in all three, the air only is admitted differently. In all cases the quantity of both gas and air are the same. In Fig. 1, no air is admitted from below; and the gas, consequently, does not meet with any until it reaches the top of the glass, where it is ignited, producing a dark, smoky flame. In Fig. 2, air is admitted from below and rises through the air-orifice, concurrently with the gas issuing through the gas-orifice. On being ignited, one long flame is produced, of a dark color, and ending in a smoky top. In Fig. 3, the air is introduced from below, and into the tin chamber, from which it issues through a perforated plate, like the nose of a watering-pot; thus producing immediate mixture with the gas. On being ignited, a short, clear and brilliant flame is produced, as in the ordinary Argand gas-burner. In an experiment made to test the heating powers of these flames, a vessel containing a given quantity of cold water was placed over each. When over fig. 2, it required 14 minutes to raise the water to 200~, whereas over fig. 3 it reached 200~ in 9 minutes. The above illustration is taken substantially from C. W. Williams' " Combustion of Coal." The superficial character of flame may be readily perceived in a common candlelight. Immediately surrounding the wick (see Fig. 4, Plate 37), is a body of dense carburetted hydrogen, just distilled from the melted tallow which is carried up the wick by capillary attraction. This dark gas corresponds with the dark brown vapor distilled from bituminous coal, just after being thrown on an open fire. Next surrounding this gas, within the flame of a candle, is seen an envelop of the same gas, but highly heated and waiting only access to the air to be ignited. Next is the flame itself, where the air and gas, having reached each other, unite in vigorous combustion —the actual combustion, however, extending through a very small thickness of gas and air atoms. Pure hydrogen gives but a faint flame, the bright flame of carburetted hydrogen gas being caused by the luminosity of the: atoms of qcrbon which form three-fourths of the total weight (although but one-fifth of the bulk) of this gas. Outside of the actual flame is a thin covering, barely visible to the eye, of carbonic acid. This has a faint purple color, and can be seen by looking closely across the,candle flame. This carbonic acid is the compound product formed by the carbon of the carburetted hydrogen in its union with the oxygen of the air, and its presence outside only of the actual ~flame of the candle shows that it is burned only after the hydrogen is consumed. Finally, radiating from the flame in all directions are rays of invisible steam, the product of the union of hydrogen and oxygen.* Time. —But, as has been already intimated, the mixture necessary to combustion, in case the gas and air are merely placed in contact, requires considerable time for its completion. Carburetted hydrogen has but little more than one-half the weight of air —the specific gravity of air being taken as 1, that of the gas is.555. Placed together in a close vessel, the two would mingle but slowly. Violent agitation would be required in order to mix them suddenly. Mr. Graham, an English chemist, in a treatise quoted by C. Wye Williams, in " Combustion of Coal," in one case says: " The receiver was filled with 75 volumes of hydrogen and 75 of defiant gas, agitated, and ~ This steamrn is condensed, however, in the atmosphere, and can be readily collected by passing flame through a tube surrounded with cold water, ice, or other agents of condensation. If a large gas-flame be directed into a tube, kept quite cool, the inside will be immediately covered with drops of condensed steam, and if the tube be inclined downward, a constant stream of water will be discharged, appearing to come (as in reality it does) directly from the flame. The water is formed in combustion by the union of hydrogen and oxygen (which are the only elements of the pure water of nature). A pint of oil, when burned, will produce a pint and a quarter of water. An Argand gas burner,'in a closed shop window, will produce in four hours two pints and a half of water, to condense or not upon the glass or the goods according to circumstances. Hydrogen unites with eight times its weight of oxygen, and in no other proportion, in forming water. Hence, a ton of coal, containing 100 pounds of pure hydrogen, will make in combustion 900 pounds, or nearly half a ton of water. 84- THE ECONOMICAL GENERATION OF STEAM. allowed to stand over water for twenty-four hours, that the mixture might be as perfect as possible." In general, he allowed four hours to elapse before he considered the gases adequately mixed. In a furnace, not four seconds, nor perhaps one, are given for the mixture of the coal gas with the air. It is true that the commotion within the furnace is favorable to mixture, while the affinity of hydrogen for the oxygen of the air is probably greater than for the olefiant gas, in Mr. Graham's experiments. In any case, however, there must be a great difference between the degree of intermixture effected respectively in the locomotive furnace and in the chemical laboratory. Professor Faraday says also on the same point:-"Although in making mixtures of gases, they will become uniform without agitation, if szcient time be allowed, the period required will be very long, extending even to hours, in narrow vessels. If hydrogen be thrown up into a wzide jar, full of oxygen, so as to fill it, and no further agitation given, the mixture, after tHie lapse of several minuztes, will still be of different composition above and below." The very length of flame proves how much time is required, in case of ordinary draft, to complete the mixture of the combustible gas and air. Under a stationary boiler, the flame from soft coal has extended thirty feet from the grate. Flame is entirely superficial, enclosing a core or central body of gas, waiting its turn to come into combustion. This gas, having a progressive motion in the flues, or anll ascensional power when discharged in the open air, must follow the draft, while by the continual combustion of its external atoms, its thickness is being steadily reduced, until, if not violently extinguished, it filially burns out to a point, and the flame terminates. Thus, if flame extends for 30 feet, the combining gas and air are mixing for the whole distance, and may not even be fully mixed when the flame ceases —what remains passing off unconsumed. The want of time is experienced with the gaseous portions only of the fuel, as these, on the moment they are distilled, are on their flight to the flues, giving but a fraction of a second for mixture and consequent combustion. The solid portion remains quietly upon the grate and takes its own time for combustion. In England, before the thorough combustion of coal was understood, a pound of bituminous coal would seldom yield an evaporation greater than five pounds of water-just about what was due to the solid carbon alone, the gaseous matter being nearly or quite lost in black smoke. So with many ocean and river vessels; the Lake Erie steamers, lately laid up because they could not earn expenses, always wasted tons of valuable fuel in a cloud of black smoke, perhaps twenty miles long-a circumstance in itself a reproach upon the practical and scientific knowledge of the country. Space.-In a common locomotive boiler, flame is instantly extinguished in the flues. This may be proved repeatedly by holding a glass tube of small bore over a clear flame. Flame being, as we have seen, superficial, it is stripyed, so to speak, of its air in entering a small flue, and hence the core or central body of gas (which by its bulk gave diameter to the flame) passes through unconsumed; the tube already filled with carbonic acid, steam, and nitrogen, containing no air for the combustion of the gas. No flame whatever can pass thLroutgh a' tube. An unignited compound, known as carbonic oxide, may pass through, and having a low igniting temperature, may afterwards take fire and burn, on coming to the air at the chimney top, just as the gas in fig. 1, Plate 37, burns at the top of the chimney where it also meets the air. The blue flame attending the conversion of carbonic oxide into carbonic acid is not, however, the flame with which we have to deal in the furnace. So far as we are concerned in burning carburetted hydrogen in the furnace, we can do so only by having the process of its combustion completed before it reaches the flues. If the ordinary fire-box does not afford room, we must obtain it by an ample combustion-chamber in the barrel of the boiler, and even there we must aid the requisite mixture by deflecting the currents of gas and air into as intimate contact as possible. We cannot control the volume of the gas after the coal is on the grate, but we can divide the air in numberless thin streams, and by means of fire-brick, or iron bridges, or plates, or by mid-feathers, we can defect the gas into mixture with the streams of air. These arrangements apply to the combustion of gaseous fuels mainly, and not to coke, nor, to a great extent, to anthracite coal. The forcible Mixtuzre of Air and Gas. —The foregoing considerations refer chiefly to the mechanical mixture (which must always precede chemical mixture or combustion) of air and gas THE COMBUSTION OF BITUMINOUS COAL. 85 which would result merely from their contact as they move together by the force of the draft. But if while the gase's are drawn upward from the surface of the fire by the draft, fine currents of air can be driven with an equal force, laterally, so as to intercept the currents of gas, and downwardl so as to force an intimate mixture, the bridges and mid-feathers mentioned, will not be required. But the room afforded by a combustion-chamber, which is not in itself an objectionable feature, is valuable in any case. _Anz Excessof Airpractically Necessary.-7-We have seen how indispensable is air in sufficient quantity, meaning by sufficient, the exact chemical equivalent for the coal burned. This quantity, which is fixed and determinate, must be thoroughly combined with gas and coal. Suppose the quantity of air admitted is insufficient to yield oxygen for burning all of the hydrogen of the carburetted hydrogen gas: then there would be burned only so much hydrogen as could obtain a full equivalent (eight times its own weight) of oxygen from the air then lresent. The rest of the hydrogen passes off and is lost, as does all of its contained carbon, which, by condensation as soot, would be partly deposited as a non-conducting lining in the flues, while the rest would form the coloring matter of the escaping smoke. If the admission of air to the furnace was now increased so as barely to furnish oxygen sufficient for all the hydrogen, these two gases would combine to the exclusion of the carbon, which would still pass off, to be condensed as soot in the flues, or in suspension in the smoke. The hydrogen of carburetted hydrogen gas, although forming but onefourth of the total weight, takes to itself oze-half of the total weight of air required for the gas. The remaining three-fourths'(by weight) of carbon, take up the other half of the whole quantity of air required for the combustion of the gas. Taking by itself that portion of the air required only for the carbon contained in the gas-if less than one-half of this quantity of air be present, the carbon will not combine with any of it, but will pass off entirely unconsumed, to be condensed in soot. If more than one-half, but less than the whole required quantity be present, the carbon will combine with that one-half only, forming a wasteful compound called carbonic oxide. This, unless subsequently mixed with as much more air and then ignited, carries all of the carbon to waste, just the same as if no air had been present. It is only by supplying and mixing the full equivalent of air, that the carbon will be combined so as to give off its full heating power, andcl without the production of soot and its consequent smoke. We observe, then, the consequences of a deficiency of air, in the loss of heating matter, in the deposit of soot in the flues, and in the formation of smoke, and how essential is the complete mixture of the gas and air, by the aid either of time and room or of forced currents, even where the air is present in sufficient gross quantity or bulk. The best chemical authorities tell us that, in the most careful laboratory practice, a considerable time, often a whole day, is necessary for the complete mixture of certain gases; while those now under consideration can only completely mix after many minutes, and even then with an excess of oxygen, or, what is the same, an excess of air. That is, mnore air(ttrh the combining equivalent must be present in order that that combining equivalent may be taken up-although the excess of air, above this equivalent, is not burned but passes through uncombined. Professor Daniell, of England, stated that for the complete combustion of bi-carburetted lhydrogen, or olefiant gas, it must be mixed with five times its volume of oxygen, although but three only are consumed. If less be used, part of the carbon escapes combination, and is deposited as a black powder. Professor Daniell added, "that which takes place when the mixture is designedly made in the most perfect manner must, undoubtedly, arise in the common processes of combustion, where the mixture is fortuitous, and much less intimate. Any method of ensuring the complete combustion of fuel, consisting partly of the volatile hydro-carbon's (compounds of.carbon and hydrogen), must be founded upon the principle of producing an intimate mixture with them of atmospheric air in excess in that part of the furnace to which they naturally rise." Practically in a furnace, nearly twice the air must be present that is actually required for the combustion of the gaseous matter. And as coal, in burning, does not give off its gas uniformly and continuously, but principally soon after being thrown upon the fire, we must have such control over the admission of air as will enable us to admit the right quantity according to the variable conditions within the furnace. 86 THE ECONOMICAL GENERATION OF STEAM. The Ignition of the Mixture and tie Maintenance of Combustion. —We have now to consider the ignition of the inflammable compounds in the furnace, and the mode by which the continuity of flame is maintained. Carburetted hydrogen gas ignites at about the temperature of iron heated to a " cherry red." Pure hydrogen will ignite at a lower temperature than when mixed with carbon; and as the proportion of carbon in combination is increased, the igniting temperature of the compound becomes considerably higher. An iron wire - of an inch in diameter, will, when heated to a cherry red, inflame pure hydrogen, but when heated to a white heat, it will not inflame carburetted hydrogen. Something depends, however, on the size of the wire itself, since a wire - inch in diameter, heated only to a cherry red, will inflame the most highly carburetted burning gas. If the gas to be burned was previously thoroughly mixed with its combining equivalent of air, an igniting temperature presented to the mixture at a single point would explode the whole; the ignition being communicated from atom to atom, just as that of gunpowder extends from grain to grain. No heat would be required beyond what would be generated in the violence of combination, as this would extend throughout the largest as well as the smallest quantity of the explosive compound. In continuous flame, also, the successively combining atoms of gas and air are ignited by contact, the, process being described as a " self-generating succession," so long as both the elements are supplied. The heat under which the gas itself distils will always ignite it, if the due admixture of air is immediately obtained. But if the access or mixture of air is delayed until the gas has risen beyond the reach of an 1initing temerature, it will then pass away unburned. 1st. The supply of air to the gas, at a point where there is heat enozgh to ig(nite the mnixture. It is by far the more rational plan to effect the immediate mixture with air, while the gas is ready to burn, than, neglecting this mixture, to endeavor to recover lost time by heating the gas afterwards, when it may be supposed to have fortuitously taken up its equivalent of air. If the necessary heat cannot be maintained in the mixture on account both of the variableness of the supply of gas and the variableness of the temperature, at the point where the gas is distilled, the heat should be supplied from other sources, but practition. ers should avoid the plausible popular error that is involved in the notion of burning gas, simply by passing it through or over a " hot fire "-incandescent fuel. No heat whatever can possibly burn carburetted hydrogen or either of its separate elements, when unmixed with air; while, if mixed, the heat at which the gas itself distils will always ignite it, provided the nixture is within the influence of that heat. Here, then, may be a real advantage in passing, not the gas, but the nixture, over incandescent fuel —to preserve the igniting temperature at times when fresh fuel would shield the mixture from the action of the fire, or when the rapidity of the draft would forcibly remove a, part of the air and gas before mixture, too far from the distilling fire to be ignited by it. And since the incandescent fuel which distils the gas is at first in immediate contact with the gas, nothing can be more reasonable than to mix the air-the other element of combustion-at this point, by the forcibly inducted currents before alluded to. And then, whenever gas is distilling, we have gas, air, and maintained heat —the three elements of combustion —in a constant supply. 2d. The sugply of an igniting temperature to the mixture, supposing it to be impracticable to saturate the nascent gas so as to prepare it for ignition, wLile it retains its own igniting temperature. In the case of cannel coal, which contains a considerable quantity of oxygen, this, distilling along with the carburetted hydrogen, materially assists its ignition, and gives to that coal its characteristic white flame. If the full chemical equivalent of the oxygen, due to the complete combustion of hydrogen and carbon, existed along with them in the coal (whereby the coal would require to have 80 per cent. of oxygen, 15 only of carbon, and 5 of hydrogen), such coal would, onf the instant of its volatilization, enter into complete combustion, without making any demand uponlthe atmosphere. If, then, the gas, while it isbeing evolved, or in its nascent state, cannot be saturated with its due equivalent of air while in range of a common igniting temperature, we must provide such temperature at the point to which the gas and air naturally rise, and where they naturally mix. At the same time, we must induce the commencement of such mixture at the earliest practicable moment, so as to afford all the time possible for combustion. If the evolution THE COMBUSTION OF BITUMINOUS COAL. 87, and saturation of gas were performed uninterruptedly, the requisite igniting temperature might be a spark merely, or at most, a fragment of fire-brick heated quite hot. But with every variation ~ofthe relative volumes of gas and air, their mixture is disturbed and the flame thereby interrupted. The fire-box of a locomotive, being made of highly conducting material and surrounded by water, affords no igniting points in its own surface. Hence, with combustible mixtures forming at various times, in various parts of the furnace, igniting points may be advantageously placed where they will intercept these mixtures at the proper time. These "igniting points" may be fire-bricks merely which become highly heated, and which may be made to serve several purpose, viz.: — that of deflecting the combustible gases and air into mixture, that of igniting the mixture when in progress, that of equalizing the temperature of the furnaces, and that of protecting the ends of the flues from the direct action of the heat. Intensity of Combustion. —While Cornish boilers burn coal at the low rate of 4 lbs. per square foot of grate, per hour, and stationary boilers from 10 to 20, locomotive boilers consume 50 to 150 lbs. This offsets the necessity in the locomotive for the great amount of combustion-room required by the stationary boiler referred to, where the flame ran 30 feet from the grate. The solid carbon burns faster, and thus imparts a higher temperature to the adjacent mixture of air and gases, and is more likely to impart an igniting temperature to a larger amount of the mixture, than if the combustion of the solid carbon were slow. The same fixed amount of heat is produced by the union of each equivalent of air and gas, but forcing more air and gas to unite in the same place, gives a higher temperature to that place, just as the heat in a smelting furnace is more intense than that in a parlor grate. Thus a small amount of coal, burning rapidly, will maintain an igniting, temperature in the gases generated from it, when a large amount burning slowly will not; and a given total amount of heat being required, the amount of fuel will be less in the former than in the latter case. This shows the importance of a small grate, although it by no means renders a large fire-box unnecessary. Large grates have been found to allow the best combustion of the gas when no air was admitted above thefire, simply because a large amount of free air thus came through the grate and facilitated the formation of a combustible mixture. There is already a very limited space for perfect combustion, in the largest locomotive fire-box, however combustion may be facilitated by an intense heat on the grate to ignite the gaseous mixture above. The mistake made in practice has been to make a small fire-box as well as a small grate, which is another matter altogether. It will be shown that more recent practice has demonstrated a small grate together with a large combustion-chamber, to be peculiarly effective. Heated Air.-The admission of heated air into the fire-box, can only be valuable for the reasons mentioned above. The combustible mixture of air and gas will not ignite till the temperature has reached 800~, nor is any chemical change made in it by heat, till this temperature is reached. Burning-gas will ignite when this temperature is applied to it, in winter and summer, and will always give off the same amount of heat by its combustion. The admission of air heated to 800~, would obviously be impracticable. Hence any heat which can be given to the air will only be valuable in keeping up the general temperature of the fire-box, since this heat must come in any case from the fuel; it should hardly be abstracted from the water —it will have been difficult enough to get it into the water by means of thorough transmission. The idea of hot air in this case has been always traced to that of the seemingly similar application in the hot-blast furnace. There, however, the case is totally different. The iron lies among the coal, and the entering cold air, before it las combined with the carbon, chills the iron, thus rendering more coal necessary than if the air had been heated. If, however, the air could combine with the coal before it reached the iron, its temperature (that of the air) would be quite a matter of indifference. In the boiler this combination-takes place, or should do so, before the air reaches the plates, and the resulting heat is that generated directly within the air and coal, and cannot be derived from any outside source. Explosive compounds will detonate at the lowest temperature. Coal, thrown on cold, absorbs heat in the volatilization of its gas, by the conversion of heat from the sensible to the latent state. Air, however, already in the gaseous form, cannot increase its latent heat. There are certain practical objections to heating air. First, for every 480~ of added heat, it is enlarged by the amount of its original volumne, being doubled at 530~, trebled at 1,010~, and so 88 THE ECONOMICAL GENERATION OF STEAM. on-so that at 3,000~, the heat of the interior of the furnace, it has six times its original volume. This makes it more unmanageable, and as its contained oxygen remains the same in weight, its mixture with the gas becomes more difficult, while when mixed, it can do only the same work as before. What is wanted is rather to condense the air than thus mischievously to expand it. If air is heated by passing through flame or over burning coal, the air will be robbed of a greater or less part of its vital oxygen. This is a positive loss. When ignition has extended by atomic communication, to the last series of combining gas and air atoms, the combustion of the gaseous matter is complete, and here its chemical consideration naturally terminates. The chemical laws and facts to which the subject under examination has been so freely referred, are established upon the most unquestionable authority, and may be proved by experiment-in many cases by the simplest observation of every-day phenomena, or by means of exceedingly simple apparatus. The chemistry of gas-lighting is in many respects identical with that of the combustion of coal-perhaps, indeed, less complicated-and yet while no respectable gas company would consent to be deprived of the services of either an acting or a consulting chemist, the railway interest, with perhaps ten thousand times the pecuniary stake in the matter, has not only neglected but repelled the application of the simplest chemical principles to a subject so exclusively chemical in its nature as that of burning soft coal in locomotives. 5. RECAPITULATION.-The facts to which the examination already made leads us, are these: Coal contains both solid and gaseous combustibles. The gaseous portion invariably burns. first. Combustion is chemical combination,-in the case before us, with atmospheric oxygen. No element in nature is combustible by itself, hence in all considerations of combustion, it must be regarded as a mutual process. Gaseous combustion is instantaneous or'gradual, according as the combustibles are ignited afteror duzring atomic mixture. With solids, burned with oxygen, as the burning of carbon, combustion is always gradual. Different combustibles of whatever kind, combine in fixed and unalterable proportions only (oxygen being considered as a combustible). Combustion will be perfect or imperfect according to the accuracy of the proportions of the elements present and to the intimacy of their mixture. A ton of average soft coal requires intimate combination with 100,000 cubic feet of air at ordinary temperature for its gaseous portion, and 200,000 cubic feet for its solid coke. The production of smoke cannot attend perfect combustion, since smoke contains and owes its existence to unburnt matter. Gaseous combustibles to be so saturated with each other as to be prepared for combustion, require either time or forcible mixture to effect this combination. Separate gases cannot suddenly mix in bulk, but only by the affinity between their superficial atoms, and hence gradually. But these bulks can be made very smnall and more numerous by forcible means. Thus flame is always superficial, or hollow, until its internal gas is burned to the last atom, when the flame ceases, or " goes out." Gaseous combustibles have an ascensional power sufficient to carry them rapidly beyond reach unless burned in their flight. Solid combustibles, as coke, lie quietly on the grate, awaiting whatever access of air we may choose to give them, and if we shut this off entirely they still remain for future ignition. Streaks or bulks of flame are merely masses of gas in bulk, burning on their surface only, where they are in contact with air. Hence if these are directed into tubes of a diameter less than that of their own bulk or volume, they are stripped of their superficial air, and consequent flame. Coal gas will ignite at the heat at which and by which it is distilled, if its due equivalent of air can be mixed with it at the moment of its evolution, or in its nascent state. If the due intermixture cannot be provided until the gas has risen in its flight to beyond its own igniting temperature, the mixture must be ignited at the earliest possible moment by external heat, as by fire THE COMBUSTION OF BITUMINOUS COAL. 89 bricks, which after a fire has been once brought into action, will take up, and retain heat sufficient to ignite the combustible mixture formed on the application of fresh charges of coal. Ignition requires only to be initial, or at a single point in an explosive compound formed or forming. Each atom in combustion communicates an igniting temperature to those adjoining, which in turn inflame others, just as the ignition of a grain of gunpowder will explode a magazine. This self-generating succession of igniting temperature may be interrupted by the variable supply of gas, or by the insufficient supply or the suppression of air. If the igniting temperature of this self-generating succession is interrupted for an instant, whatever may be the perfection of mixture of air and gas at the next instant, combustion will cease. Gases which will not burn without exposure to continued heat, cannot and will not while in the same state, bum witL such exposure. The air is the only element entering into combustion, th'e constant admission of which we are now able to control, hence the whole question of the combustion of gaseous fuels.turns upon the admission and mixture of AIR, each atom of which must intercept its destined atom of gas, in the flight from the surface of the coal to the flues. The form and dimensions of furnaces can have no influence whatever in the process of conmbustion, further thanas they can afford space, direction and time for the chemical union of the combustible elements. These only are to be considered, being limited simply by the quantity of coal to be burned in a given time, and by the available room on the engine. The phenomena presented on the trial of bituminous coal, in an. ordinary wood-burning engine, can be now easily comprehended. The gas distilled from the fresh coal could not obtain its due admixture of air, and was therefore carried off not only unburned, but in an incombustible condition. Its condensible matter was partly precipitated as soot in the flues, and partly carried off in suspension in the smoke, of which it formed the coloring matter. To volatilise gas from coal, much heat must be converted from the sensible into the latent state, just as heat is so converted in forming steam from water. This loss of available heat can only be restored by the immediate combustion of the gas, and hence, ih our experimental engine, the furnace was cooled and the steam went down. Under a strong draft, sparks and even fragments of coal of considerable size, were forcibly drawn through, as there were no deflectors in the fire-box, or deflecting currents separate from the draft to throw them down for subsequent combustion. As already shown in previous chapters, with anthracite coal especially, which decrepitates rapidly under sudden heat, the mere mechanical action of these sharp fragments upon the flue-ends and furnace-plates, is sufficient to abrade them rapidly, until the metal is finally cut through. The sulphur of the coal unites with its associate hydrogen, forming sulphuretted hydrogen gas, too readily recognized by its offensive smell. As the gas must be distilled always before the solid portion of the coal can burn, it is necessarily some timle before the ignition of the latter is effected. As this coke burns without loss, the furnace then recovers its heat and the steam pressure again rises. The temperature generated in the burning of the solid coke is naturally greater than that obtained from wood. The furnace-plates need only ample circulation of water upon them at all times to prevent their being burned, and in default of this, thin plates of homogeneous material are necessary. The clinker is formed by the fusion of impurities in the coal. The production of this obstruction is entirely a matter of the cleanliness of the fuel, and can in no way be lessened by any arrangement of the furnace, although means may be provided for breaking and clearing it out after it has formed. The coal swells while its gas is distilling, from the expansive force of the gas itself. But two-thirds of the heating power of the coal having been utilized, its evaporation, for each pound has not exceeded five or five and a half pounds of water. 1o 90 THE ECONOMICAL GENERATION OF STEAM. 6. CONCLUSIONS.-The solid carbon of bituminous coal, after the gases are distilled, will burn on the grate by the admission of air to it; it will not waste but will simply not burn, if air is not admitted; and it will, when burning, distil the gases from the fresh coal added to it. The practical facts attending the burning of the gaseous parts of the coal are:1st. The gas will be carried beyond the reach of the conditions of combustion and lost, by the action of the draft, if not burned immediately after its distillation. 2d. A combustible mixture can be formed only by bringing an excess of air into intimate contact with the gas. -3d. This mixture will be carried beyond the reach of the conditions of combustion, if not immediately brought within the influence of an igniting temperature. 4th. An igniting temperature may be furnished to the mixture by the heat which distilled the gas. 5th. Wlhen the supply of gas is uninterrupted and regular, and the conditions of mixture with air are fulfilled, the combustion once started will continue without farther aid. 6th. The supply of gas is almost entirely variable and intermittent, and requires the presence of a constant igniting temperature. 7th. The greater part of the mixture of air and gas is likely to occur, out of the reach of the igniting temperature which distilled the gas, by reason of the action of the draft. 8th. If the mixture of air and gas is forced to occur in the presence of the igniting temperature that distilled the gas, that is, the fire beneath it, the combustion will be complete. 9th. An igniting temperature may be furnished to the mixture by other incandescent fuel or by heated fire-bricks or plates of metal so placed as to intercept the mixture before it reaches the flues. The first and grand condition imposed by these facts, therefore, is:The imgmediate and thorough intermixture of a plentiful but regulated supply of air AT THE SURFACE OF THE FUEL. It is true that the combustible mixture may be and is made and ignited at other points besides the surface of the fire. Here, then, start two legitimate, but as far as apparatus is concerned, distinct methods of coal-burning, of which nearly all others are modifications; the one is represented by Clark's Jet (plate 56), which, so to speak, makes the combustible mixture where there is heat; the other by Beattie's chamber and bricks (Plates 38 and 391), which makes the heat where there is the mixture-both with substantially similar results. Which is the simpler of the two methods can be inferred, before the results of practice are mentioned. CHAPTER III. MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 1. INTRODUCTORY.-A bituminous coal-burning boiler, in addition to its adaptation to burning the solid carbon of the coal, must be provided with appliances, 1st, for mixing air with the gas of the coal the instant it is distilled, and 2d, for maintaining an igniting temperature in the combustible mixture of air and gas thus formed. There are substantially two ways of accomplishing these results. Ist. By giving the air time and room to mix with the gas, by means of deflectors and combustion-chambers, wherein are provided the means of preserving an igniting temperature, such as fire-bricks, iron plates, or incandescent fuel. Tliis system evidently cannot be applied to old engines without expensive structural alteration. 2d. By forcing a mixture of air and gas in the presence of the incandescent fuel which distilled the gas. It will appear that this process, together with the necessary means of improving the circulation in boilers with narrow water-spaces, mway be so simply and cheaply applied to old engines as to effectually solve the great problem of modern railway enyineering —how to burn coal economically and without smzoke in wood or coke-burniny enyines. And in the construction of new boilers and in the renewal of old boilers, it will appear that these methods may be advantageously modified and combined. It is proposed to consider the practical working of the various coal-burning boilers and appliances with reference to the general principles thus indicated. 2. MIXTURE AND IGNITION or AIR AND GAS BY THE AID OF THE DRAFT. Simple admission of Sir above the Fire.-In 1785, James WaVtt promulgated the idea of a separate supply of fresh air above the fire. In 1820, Josiah Parkes patented the arrangement shown at fig. 8, plate 55-a split bridge to mingle a broad stream of air with the gas above the fire. Perhaps one-half the stationary boilers of 1860 would be improved by the application of Parkes' device. In 1838, Mr. F. P. Dimpfel (inventor of the Dimpfel water-tube boiler, Plate 50), used hollow stays for admitting air above the fire in stationary boilers. In 1838, C. Wye Williams patented the Argand furnace, which was similar to Parkes', save in one important particular the air was admitted in fine jets through the bridge, with excellent results, because there was a thorough mixture as well as an igniting temperature. The admission of air through hollow stays, and through a perforated door, into the ordinary fire-box, is largely practised. Fromn 12 to 20 C-inch or — inch stays, immediately under the crown-plate, are the entire structural adaptation of many boilers in service, to coal burning. The limited measure of success which has, in any case, attended their operations is, therefore, due entirely to the careful manipulations of the firmmen. These boilers make smoke so long as gas is distilled. It is probable that the accidental entrance of some air through the thin, loose fire, is much more effective than the trivial attempt at air admission above the fire. Perforated doors are more common because they are cheaper; and when the door is high above the fire, the effects of the homeopathlic dose of oxygen thus administered, is not startling. This is called air admission, and because it does not prevent smoke, smoke is pronounced inevitable. As well might we expect steam enough for 16-inch cylinders from a teapot. Four hundred 1-inch holes in two doors, and 24 w-inch hollow stays, as in Beattie's 92 THE ECONOMICAL GENERATION OF STEAM. engines, are of some appreciable advantage. If air is to be drawn in by the draft only, perhaps 4,000 minute holes through the whole back of the fire-box would be a better arrangement. The proximity of the holes to the surface of the fire is very important, so as to admit the air where there is heat enough to ignite the mixture. The air-sieve C, plate 46, of the Rogers engine, is found to prevent smoke with careful firing, save at the time of supplying fresh fuel; then the distillation of gas is so great as to require more air and a better mixture. The air deflector used in Baldwin's engines, fig. 7, plate 54, tends to promote a.good mixture. Leaving the fire-door open a little is found to prevent smoke, except at the time of firing; but it admits cold air straight to the flues and cools them, as the air enters in bulk. Nor is it in the right place, for there is not always an igniting temperature at the level of the fire-door. On the Hudson River Railway, air was at one time supplied to the top of the fire behind the grates, which were two inches shorter than the fire-box. The place of admission was obviously right, but the air came up in bulk, and it cooled the plates to some extent. Yates' grate (Plate 52), and the grate shown at fig. 9, plate 56, are better; * the air enters at the surface of the fire and in the centre of the bed of coals in small streams, at a point where the mixture is sure to find an igniting temperature. It is regulated by a damper, as shown. The latter is a very simple plan, and could not fail to be effective. Those who are not prepared to adopt any more thorough plan will do well to employ this. The tubes shown at fig. 3, plate 56, even without the steam jet, since they admit a large amount of air over the surface of the whole fire, and close to it, are alone sufficient to prevent smoke, save at the time of'firing; but their delivery' can in no way be so well regulated as by the steam jet. On the Great Western Railway of Canada, air is admitted as shown at D, fig. 2, plate 45, through a large tube, in which is a sieve shown at fig. 7, plate 451. On the Great Western of England, sixty 1-inch air-holes and a perforated door were put into some 50 coke fire-boxes, thus adapting them to burning coal without much smoke, except at firing. In other boilers, 18 bellmouthed air-tubes, taking the place of as many flues, and running through the cylinder of the boiler from the front of the smoke-box, have given moderately good results. But the steam pressure is reduced as much as the air is heated, and when the air is heated it is not chemically changed. It would require 800", or above twice the heat the steam could give it, to be hot enough to ignite the combustible mixture. The plan is good in one respect; the air is driven back through the forward current of gas, facilitating mixture; and it does not, therefore, so soon enter the flues. The absence of convenient means for regulating the area of air-holes has rendered them practically useless on a large number of American boilers. Air enough above the fire, when gas is distilling from a fresh supply of coal, is too mucL when the gas is all distilled and the coke only is burning on the grate; too much,simply because it lowers the temperature of the fire-box without replacing any heat by its own combustion, as would be the case if carburetted hydrogen were distilling. Air: enough to fill the fire-box with white flame at one time, is so injurious at another time as to visibly lower the steam-pressure. Now, on a large number of engines fairly supplied with air-holes, the author has observed, either the absence of a damper for regulating the supply, or, one so inconveniently arranged and clumnsily constructed that adjusting it was a very difficult and serious undertaking. In the former case the holes were frequently plugged because the firemen preferred smoke to low pressure, and in the latter case the damper was invariably kept shut because no fireman would exert himself to adjust it. It is a safe rule, that dampers, variable exhausts, independent feed pumps, and all movables that require frequent adjustment, will be let severely alone by engine men, if any unusual amount of exertion and inconvenience are necessary to operate them. On the whole, the admission of a plenty of air-say what will be drawn by the blast through 50 hollow stays, of 1 inch diameter (about 40 square inches area) —close to the top of the coals, will be found to prevent smoke, with careful firing, except while the fire is being replenished or stirred. This is the practical effect of 14 scant 2-inch tubular stays in boilers fitted with Clark's jet (the jet not being applied), and having about the same area. These openings must be regulated byconvenient dampers, or they will admit too much air after the gas has been mostly distilled. They may take the place of one row of ordinary stays all around the fire-box, and thus * This was suggested by an anonymous correspondent of the London Engineer. MEANS OF APPLYING THE PRINCIPLES OF ~COMBUSTION. 93 not interfere with good circulation; otherwise, more and smaller holes would be more effective. But it does not follow that a better apparatus cannot be as cheaply applied. Comnbustion-clZambers.-The combustion-chamber, as now used, and in the two forms recently patented here by McConnell, was patented in England, June 2, 1846, by Stubbs & Gryll. (Figs. 1 and 2, Plate 55.) The forms adopted by McConnell, are illustrated by plates 41 and 44. In 1846, Mr. John Dewrance patented the boiler shown on plate 43. The air was admitted to the combustion-chamber in fine streams through the pipe c. The baffle-plate d (so frequently reinvented) deflected the air and gas into thorough mixture. The results were rather favorable, but combustion was not perfect. Indeed the mixture could hardly have been more completebut the igniting temperature was not maintained. Mr. A. F. Smith, then superintendent of the Cumberland Valley Railway, specified the combustion-chamber in 1851. McConnell commenced its use in 1852. The same device, for substantially the same purpose-room and time for combustion-but modified to suit other features of boilers, have been since applied by Beattie, Craig, Allan and others in England, and by Millholland, Boardman, Dimpfel, Phleger, Yates, and locomotive builders quite generally, in America. The simple combustion-chamber has been proved, by long and general practice, to be an excellent feature of the boiler, with wood in any case, and with coal in case of proper manipulation; it has not been deemed very advantageous by those who have supposed that gas will ignite of itselfin a chamber. In addition to its functions in the matter of combustion, it shortens the flues and removes them from the more intense action of the fire. Its chief object is to allow room and time for the thorough mixture, of air and gas; and as far as this is concerned, the largest size which the best transmission of heat will allow, would appear to be the best. It must be remembered, however, that the combustible mixture cools in the very act of expansion; it is suggested that Beattie's new boilers (Plate 39-), which are nearly all combustion-chamber, require more fire-bricks than those with smaller clambers, to maintain an igniting temperature. Another trouble has been experienced with long combustion-chaimbers, on American lines generally; fine coal is deposited in the chamber, sometimes stopping one or two rows of lower flues. It is probable, however, that an equable draft through the flues, by means of spirals or better smoke-box arrangements, would be a complete remedy. Combustion-chambers have frequently been made so large as to reduce the heating surface so much as to waste a considerable amount of heat. Mr. D. K. Clark has stated,* that the waste heat in the smoke-boxes of boilers like McConnell's (Plate 44), had been found to be as high as 1,100~. The correct proportions for combustion-chambers are not settled. Those of Beattie's new engines (Plate 391), are 71 feet long with 1 foot 10 inch flues; McConnell has made them 7 feet long; Dimpfel's (Plate 50) extend through the whole of the boiler; the best coal burners on the Hudson River line, have chambers 5- feet long with 61 feet flues; Baldwin's (Plate 51) are from 4 to 5 feet long; and chambers 18 inches long have been built at Paterson. Deflectors in chambers practically increase their capacity. Room for forming a mixture is evidently of limited importance, when one of the elements of the mixture is wanting. The practice on sorne lines, however, is to furnish combustion-chambers at the cost of rebuilding old boilers, wherein the gas is expected to burn without air. The admission of sufficient air at the surface of the fire, is equally important in any case. In fact, boilers with chambers and without adequate air admission, make more smoke than contracted fire-boxes well supplied with air. The chamber is by no means a substitute for oxygen, as many practitioners seem to think; it does not even modify the conditions of air admission. The chambers once used by Millholland, in the central part of the cylinder of the boiler, between sets of short flues, did not prevent smoke, nor does air admitted into the chambers of McConnell's or of Boardman's boilers (a and D, Plate 49) prevent it. The author has frequently observed McConnell's 7-feet chamber engines smoking to an extent which indicated that none of the gas was burned. The air admitted in a thin film under the disc a in the Dimpfel boiler (Plate 50), is well mixed; but that supplied at J, in the extreme forward end, is not suree to meet an igniting * Paper read before the Institution of Civil Engineers, 1853. 94 THE ECONOMICAL GENERATION OF STEAM. temperature. The chamber facilitates combustion merely by giving air and gas a better chance to mix before reaching the flues; now, if the air were not supplied till the gas had got close to the flues, the chamber would be of little value; but some air will get through a thin, loose fire —careful manipulation can do a great deal to modify the effects of bad designing —and, therefore, the chamber is always important. However thoroughly the mixture may be made in the fire-box, the draught instantly carries it into the flues, where combustion is impossible; the chamber, in this case, almost insures combustion. It is desirable for new engines, where its extra cost is not excessive. But in view of the results of the forcible mixture of air and gas, yet to be mentioned, it is probable that the reconstruction of the boiler at a cost of from $1,000 to $2,000, is too great a price for this insurance. Deflectors-Fire-bricks.-In 1837, Robert L. Stevens, of New York, put a water-bridge into the fire-box of the steamboat " Independence," for the purpose of deflecting the currents of gas and promoting their combustion. The deflector of Mr. John Dewrance, before mentioned, and shown on plate 43, was intended to deflect the air and gas into mixing currents, and to keep them for a longer time out of the flues. The arrangement was patented October, 1846. Fire-brick deflectors were used by Mr. Joseph Beattie in the fire-box of the engine "Britannia," on the London and South-western Railway in May, 1854. Mr. Geo. S. Griggs employed the fire-brick deflector (Fig. 10, Plate 58) in a manner not employed till afterwards by Mr. Beattiea bridge or arch extending across the fire-box, under. the flues, and tending to make a combustion-chamber of the top of the fire-box. This plan was adopted by Mr. Thomas Yarrow in 1857, on the Aberdeen Railway; it is similar to that of Mr. Lees (Fig. 9, Plate 58), employed on the East Lancashire Railway in 1858. Cast-iron deflectors were used in this country by Mr. Samuel J. Hayes, on the Illinois Central Railway; by Mr. W. S. Hudson, of the Rogers Locomotive Works, and by Mr. H. Uhry, of the New Jersey Locomotive Works, about the time they were adopted by Mr. William Jenkins (1857) on the Lancashire & Yorkshire Railway (Fig. 7, Plate 55). A great variety of deflectors is now employed for the purposes indicated, in English and American engines. Fire-brick deflectors and air-pipes have been patented in great numbers, especially in England, since the results of Beattie's engines have been made public. The ignition of the mixture, as well as its formation, is an important function of deflectors. The sides of the fire-box, and all water-bridges and mid-feathers, are less than half hot enough to ignite the air and gas, 800" being required, and from 350~ to 360~ being provided; they rather cool it, and prevent combustion. For this reason, walls of fire-brick have been placed against the fire-plates; the combustion was improved, but the vaporization was reduced; probably because the heated products of combustion did not all-come in contact with conducting surfaces in the absence of heat-traps in the flues. Fire-brick walls were first employed by Dr. Nott about 1825. Water-bridges may be valuable for the purposes of alternate firing, as in case of Head's boiler (Plate 44}), or to form an end to the grate, or a bulkhead for the coals, as in case of the rNorris and the Phleger boilers (Plates 48 and 47), but because they cool rather than ignite the combustible mixture, while they are notoriously expensive in construction and repairs, they are not considered the best deflectors. They considerably extend the heating surface, but the best modern practice does not furnish many examples of their use for this purpose. Sur-.face enough to utilize all the heat generated may be placed in a cheaper and more durable shape in large fire-boxes. Bulk of material as well as the right position, is necessary to the maintenance of an igniting temperature. The curved deflectors, shown at fig. 6, plate 54, although well shaped to mix the air and gas, do not afford it an igniting temperature; they are hollow and take warmed air from condensing tubes C, which run through the barrel of the boiler. The deflector of Dewrance (Plate 43) had not material enough to retain sufficient heat to ignite the mixture; it cooled too quickly. All similarly light devices have failed in this particular; but they have not been abandoned for this reason, but because they rapidly burned out. On the Illinois Central, a cast-iron box 3- inches deep, and extending 3 feet back from the flue-plate, was, during two or three trips, ruined by the heat of the fire, the metal being only - inch thick. Air passed into the box from seven 2-inch hollow stays, and was distributed through smaller holes. Heavier masses of MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 95 iron or brick are now employed on various lines; they retain heat, and last longer. The cast-iron deflectors (Fig. 7, Plate 55) used on the Lancashire & Yorkshire line, weigh 200 lbs., and last from ten to twelve weeks, the cost of renewing them being about 25 cents per week. Wrought-iron deflectors of the same weight last nine months. The greater the mass, however, the longer time is required to heat it to an igniting temperature. It is said, on good authority, that the 800 Ibs. of fire-bricks in Beattie's engines do not all reach an igniting temperature on short runs, and that firing must be commenced much earlier than is usual, in order to get them into the proper condition. Fire-bricks, for a given weight, are more durable than cast-iron in the furnace of a locomotive, and are certainly cheaper. How long either will last, is very variously stated by practitioners. It can be only concluded safely, that so far as these deflectors are absolutely necessary to good combustion, the cost of renewing them is found to bear a very small proportion to that of wasting the gas, without them. But when they are so extensive and complicated as those in Beattie's new boiler (plate 39-), the general belief is that they do not save enough over a simpler arrangement to pay for their cost and renewal. The results of practice are quite clear on this point. Fire-bricks, then, are evidently the best material for deflectors, being cheap, readily moulded to any required form, and easily removed. How large an amount of this material is necessary to good combustion, is not an independent consideration. The supply of air and the management of the fire are much more important elements in good combustion; deflecting the mixture and promoting its ignition are subsequent operations which, in fact, can be performed without deflectors. In case the fire is carelessly managed, the simple brick arch produces hardly appreciable results; but when the other conditions of combustion are fulfilled, its advantages are distinct and decided. On the other hand, the combustion in Beattie's boiler, plate 38, when the bricks are once heated, is always practically perfect, so long as air enough is admitted, in bulk or otherwise, and however carelessly the fire may be managed. Here, then, start two distinct methods of using brick; the first, Beattie's, perfects and ignites the mixture of air and gas; the second, represented by Griggs' arch, can be only auxiliary to good firing, and to a mixture of air in fine jets at the surface of the fire, in the perfection of combustion. Although the simple arch has never been employed with the best and most thorough method of air admission, it is safe to conclude that the results in the cases mentioned, would not be very different; that is to say, either would prevent smoke. The question then lies between the extra ultimate cost of Beattie's boiler, on the one hand, and the risk of bad management where good combustion is so very dependent on the skill of the fireman, on the other hand. An intermediate method of using fire-brick-Eaton's-is shown on plate 45, and will be again referred to. The construction of Beattie's standard pattern of boiler is as follows (see plate 38): The firebox is divided transversely into two compartments, by the inclined water-space, B. The two compartments have each separate grates, ash-pans, and dampers, and separate doors, I', C'. A combustion chamber F, extends into the, barrel of the boiler, with a man-hole and cover at G. From this chamber extend the flues, I, I, I, as in ordinary boilers, excepting that these are very small. A "hanging-bridge," ), depends from the roof of the furnace, and another, F', from the roof of the combustion chamber. A grating of fire-bricks, 1, 1, shown also in end section, plate 39, spans across from the top of the, water-space, B, to the back of the furnace, just between the two doors. An arch of fire-brick segments is also sprung across as shown, just in front of the mouth of the combustion-chamber. From the top of this arch, another set of fire-brick grates reach up to the side of the hanging-bridge, D. In the combustion-chamber, a faggot of fire-brick tubes, 2, 2, 2 (shown also in detail plate 39), is placed in the arrangement shown in fig. 2, plate 33, which is a section through the combustion-chamber. These tubes are perforated as shown. The furnace has been made both of iron and copper —the latter, however, being now preferred. The combustion-chamber is of iron. The flues have been made of brass, but iron is now preferred. The grates are of wrought-iron, laid stationary, or without any arrangement for rocking. The doors have each 200 holes, - inch in diameter in the inner plate, with large slots, covered by a * This arch is a recent application, having been made subsequent to the use of the same by Mr. G. S. Griggs, of the Boston & Providence railway, as before mentioned. 96 THE ECONOMICAL GENERATION OF STEAM. sliding cover, in the outer plate. Ten hollow stays of 4 inch bore, open, in addition, into the lower or rear compartment, and 10 others of the same size, into the upper or forward compartment. The exhaust is single, or through but one pipe, with a fixed nozzle of invariable size. There is nothing peculiar in the arrangement of the smoke-box-no deflectors, spark-boxes, wire-nettings, or other obstructions, the chimney being straight and entirely open. There is a steam-jet or blower, to take steam up the chimney, to maintain the draft when the engine is standing. The working of this boiler is as follows: Fires of bituminous coal are lighted on both grates, and when steam is raised to its working pressure, the forward damper is closed perfectly tight, not to be again opened while the engine is running. The working-fire is thus in the compartment, A, behind the partition, B, while that in the front is burning but slowly. The gas arising from the coal in the rear compartment, is deflected backward and upward by the inclined waterspace, B, and thus into mixture with the air entering through the 200 small holes. in the lower door. Mixture having commenced, the whole passes through the heated fire-brick grating, 1, 1, beyond which it takes an additional infusion of air through the upper door. The whole is deflected downward by the " hanging-bridge," _D, upon the surface of the front fire, where all the fine coal and sparks, carried over, are slowly burned. The gas and air, already in the course of mixture, and partly ignited, now pass through the fire-brick tubes, 2, 2, the temperature of ignition being sustained for such currents of gas and air as may not have become finally mixed until they reach the combustion-chamber. These arrangements are so far sufficient, that, by the time the gas would otherwise have reached the flues, it is burned out in clear flame, hot carbonic acid and steam only passing through the flues, neither of which, separately, or together, can form soot or smoke. The entire arrangement is thus such as to mingle the gas with air, while the former is rising from the fire, then passing them together over hot surfaces in order to effect their ignition, and by deflecting and dividing the currents, to develop space and time for their complete combustion, before entering the flues. The engine has a 15-inch cylinder; 21-inch stroke; one pair of 6 —feet driving-wheels. Inside length of furnace (through partition, B, Plate 38), 4 ft. 11 in.; width, at grate, 3 ft. 6 in.; at top, 3 ft. 41) in.; inside depth at back, 5 ft. 2 in.; at front, or at mouth of combustionchamber, 4 ft. 2 in. The whole furnace made of e-inch copper. The combustion-chamnber extends 4 ft. 2 in. into the boiler, and is 3 ft. 6 in. in diameter. There are 373 iron flues, 6 feet long and 1l inch outside diameter, set without thimbles. The fire-box surface is 107 square feet; combustion-chamber 37 square feet, and flues 625 square feet; or 769 square feet in all. The whole grate is 16 square feet, of which but 8 feet are occupied by the active fire. There are about 800 pounds of fire-brick, presenting from 80 to 100 square feet of surface to the gas. Beattie's new boiler, plate 39-, is similarly worked; its construction embodies simply an exaggeration of the principles already described. The perforated fire-brick arch, C, is substituted for the grating of fire-brick between E, C, and D, plate 38. The combustion-chamber, plate 39, is extended to within 1 foot 10 inches of the smoke-box and is filled with faggots of fire-brick tubes, J, shown in section by fig. 3. A mid-feather extends nearly the whole length of the chamber. The heating surface is also considerably reduced, varying 100 to 200 square feet in diferent classes of engines. Clark's conclusion as to the boiler, plate 38, after an elaborate series of experiments, is, that by means of the fire-bricks (80 square feet, 5r cwt.) the consumption of coal was reduced 14 per cent. as compared with the boiler without them, with heated feed water in both cases. The consumption of coal was 25 per cent. less than that of coke for the same. duty, with fire-bricks and hot water. But with cold water, the consumption of coal was about the same, with or without fire-bricks. In " European Railways," Mr. Colburn and the author gave the results of other experiments on Beattie's boiler, and of their own observations and experiments. Since a statement of the pounds of coal per mile, and of the evaporation of the different'coal-burning engines, is not contemplated in this work, for reasons already mentioned (the want of data for accurate comparison), and since MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 97 it may always be assumed that the observed quality of combustion and vaporization and the repairs of the boiler are a good measure of economy, the detailed accounts given in " European Rail. ways" need not be repeated. The great fact of importance to the present inquiry is, that when the fire-bricks in Beattie's boiler (Plate 38) are once heated to 800" or more, the combustion is practically perfect. There is entire uniformity in the testimony on this point. The fact that the bricks were not a cause of economy when the feed-water was cold, proves nothing against the system, because the boiler was then in an abnormal state-the dampers were left open to allow the necessary steam generation, and the peculiar functions of the boiler were not called into action. The combustion in the new boiler (Plate 39-) is also practically perfect, as far as the author's observation has extended. In October, 1859, the author made several trips over the London & South-western Railway on an engine constructed as shown in the plate. No smoke was made at any time either at stopping, starting, firing, or working the fire, the steam jet in the chimney, of course, being turned on when the steam was shut off from the cylinders. And at no time did the steam pressure fall below 120 lbs., but was usually 140 lbs., except when purposely let down at the time of approaching stations where long stops were required. The feed-water was heated bythe apparatus shown on plate 59, to a temperature varying between 180" and 200.~ It is reported that, as a result of general practice, these boilers do not make the comlplement of steam, and that they consume 35 lbs. of coal per mile run in doing the work performed by those shown on plate 38, with 15 to 18 lbs. per mile run. But while the combustion is perfect in Beattie's boilers, their repairs are so great as to offset much of the saving thus made-so great that the London and South-western Company have determined to have no more engines built at their own shops, which is understood to mean that a simpler and more durable plan of boiler will be adopted. While the cost of renewing a simple fire-brick arch four or five times a year is comparatively trivial, that of renewing the masonry shown in plate 391, and of keeping the mid-feather and bridges D and B in repair, together with the extra dead weight, must be sufficient to offset much of the saving due to perfect combustion. Without going largely into the question of repairs, it is reasonable and sufficient to conclude from the data already before us, that if perfect combustion can be produced with a simpler construction of boiler, the general result must and will be more economical. It would be useless to attempt to prove what is self-evident. The simple brick arch in a plane fire-box, however, does not cause perfect combustion. TIhe author's observations in this particular are corroborated by those of observers generally. The deflection of the air and gas by the arch, backward and upward, giving them time and room to mix; and the ignition of such strata of the mixture as pass within the influence of its temperature -all this is excellent as far it goes. But where the distillation of gas from fresh coal is excesssive, and especially where the air is not admitted at the very surface of the fire and spread over the fire by forcible means, the gaseous contents of the fire-box are so suddenly drawn into the flues by the draft, that perfect mixture, ignition and combustion, are impossible. The deflectors used by Mr. Richard Eaton in the new Great Western (Canada) engines are shown on plate 45. The water-bridge B is inclined, as shown in Fig. 2, so that the circulatior in it may be facilitated. As a result, it stands better than water bridges as ordinarily made. Resting upon this, and abutting against the back plate of the fire-box is a grating C of fire-brick or cast-iron bars, between which all the currents of air and gas are compressed into mixture, and by which the mixture is at the same time ignited.@ The result is more satisfactory, as might be inferred, than if the simple arch were employed. With careful firing, and a plentiful but regulated admission of air at the door, no smoke is produced except when the distillation of gas is. excessive. The bricks touch each other at the ends, and at the two lugs shown at C, and thus form a continuous and substantial structure; and they are not liable to be broken like a wide arch, by careless firing. This arrangement embodies much of the excellence of Beattie's as to * The use of a fire-brick grating thus arranged was:suggestedl and a drawing was furnished by the author to Mr. Morris Miller of the Harlem Railway, in 1858. The bricks were made, but were accidentally broken up, and the experiment was niever carried out. 13 98 THE ECONOMICAL GENERATION OF STEAM. combustion, with very little more weight than that of the simple arch. For plain boilers, this is probably the best plan of deflectors. The faggot of fire-brick tubes in Yates' boiler, plate 52, is similar to that in Beattie's, and its effect is good. The same arrangement was put into one of the Rogers engines by Mr. Hudson, for the Illinois Central Railway. The coal being very impure and the firing unskilful, as it was about the earliest coal-burning practice on that line, the brick tubes became stopped with cinder, and were abandoned. Mr. Alexander Allan of Perth has patented several varieties of deflectors. One of his boilers is constructed very like that of Norris's (Plate 48), without a pendant fire-box. A brick bridge occupies the place of Norris' water bridge, and three hanging bridges like that in Beattie's boiler (B Plate 391) depend from the crown sheet. In another boiler constructed similarly to Norris's, Allan employs merely a fire-brick diaphragm, through a moderated opening in which, all the air and gas is brought into approximately close contact. A similar diaphragm is specified by Messrs. W. M. Baldwin & Co., as shown at fig. 2, plate 54. A transverse diaphragm by the same builders, especialy adapted to mixing the air and gas, and giving them all possible time and room, and an igniting temperature, is shown by figs. 10 and 11, plate 54. In conclusion it appears that the thoroughness of the mixture and ignition of air and gas by deflectors, depends on their extent of surface, on the proportion of the air and gas which comes in contact with them, and on their ability to maintain an igniting temperature, or in other words, their bulk. When all these conditions are so thoroughly carried out as they are in Beattie's boiler, combustion is perfect-and repairs and renewals are excessive. A mere brick arch is only an aid to skilful firing and proper air admission, in the perfection of combustion, and is, comparatively of lminor importance-and its repair expenses are of no practical account. There are various intermediate methods of arranging deflectors, which give intermediate results, and as in case of Eaton's plan, they are, on the whole, superior to the extremes which have been mentioned. A more minute statement of results would be important, were it not probable that another system of promoting combustion, yet to be described, is still more simple and effective-a system to which fire-brick or iron deflectors may be merely auxiliary, if useful at all. 4. Ignitior n by incandescent Fzuel. —Alternate firiny..In 1837, Gray & Chanter put an engine called the " Liver " on the Liverpool & Manchester Railway, of which the double fire-box is shown by fig. 5, plate 55. A grate a, was formed in the water-space dividing the fire-boxes, on which coke was burned. Coal was burned on the lower grate 6, and on a supplementary grate c, which caught any coal that dropped from above. Air was admitted to both chambers by numerous perforations in the hollow fire-doors; and air was admitted by hollow stay-bolts above the coal fire. A steam jet was employed in the chimney, as at the present day. The igniting temperature was given to the air and gas, by passing them through the coke fire. The engine ran without smoke, under careful management. In 1839, Gray & Chanter applied the arrangement shown at fig. 6, plate 55, to an engine on the Liverpool & Manchester Railway. Coal was burned on the back grate and coke on the front grate —two-thlirds coal and one-third coke. The gas was deflected by the diaphragm, over the incandescent coke fire, and air was admitted above, but too near the flues to allow thorough combustion, except with good management of the fire. In this was the germ of Beattie's first arrangement, whichl is more practicable. Gray & Chanter's fire-boxes had too little room for combustion, in default of a forced mixture of'air and gas. In 1851, the step grate (Fig. 4, Plate 54) was largely employed on the Northern Railway of France for coal burning, the gas from the top of the steps passing over the incandescent fuel below. This plan, so far as it is important to combustion — -that is to say, the inclined grate-was farther carried out by Cudworthl, and by Craig, of England with results which will be mentioned. In 1852, Mr. J. E. McConnell, of the London & North-western Railway, modified Stubbs & Gryll's combustion-chamber engine, by placing a longitudinal mid-feather in it, as shown at plates 42 and 44, and thus adapting it to alternate firing. The working of this boiler is as follows: After the fires are well under way, all subsequent firing goes on alternately in the two divisions of the fire-box —say first in the right hand compartment, and then in the left. The gas MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 99 from the fresh coal in either compartment, passing to the combustion-chamber, mingles with the air introduced at the inner ends of sheet-iron air-boxes or casings, and thence passing on, is joined at the forward end of the water-space by the full flame from the other compartment. Whatever gas may remain unignited should now go into combustion-the products passing off through the flues. On stopping at stations, the steam-jet must be turned up the chimney. But it will be seen that the air and gas are made to mix while going in the same direction, and that should there not be sufficient temperature to ignite them at the moment of mixture, this must come from the other compartment, just as all the gaseous matter is entering the flues. The combustion, with careful firing, however, is good, though by no means perfect. In 1853, Mr. Beattie placed a small supplementary fire-box on the back of a common fire-box, in which he burned coal; the gases passing over a fire-brick deflector and then over the coke fire in the forward fire-box, together with air, were consumed. This led to his boiler of 1855, illustrated by plate 38, in which the passing of the mixture over the partially coked coal in the forward compartment C, aided the bricks in perfecting combustion. This feature gradually became of less importance, as will be seen, by comparing plates 38 and 391. In 1855, Mr. O. W. Bayley of the Manchester Locomotive Works, U. S., brought out a novel plan of burning the gases of coal by means of passing them with air over the incandescent fuel. The sketches, figs. 8 and 9, plate 67, illustrate this fire-box. An inclined water-space having a square opening into each lower compartment, these openings being alternately closed by dampers, extended upward and backward from the lower flues, completely dividing the fire-box transversely. The lower portion of the fire-box was again subdivided by a longitudinal mid-feather, in which, near the forward fire-box plate, was a square opening, which was always kept open. Each lower compartment had its ash-pan and fire-door. The mode of working was as follows:-one of the dampers, say the right-hand one, was closed; coal was fed to the right-hand grate. The gases were then forced to pass into the left-hand compartment over the incandescent fuel in it, and thence through the transverse water-space into the upper compartment, and thus into the flues. The object was to burn the gases by direct heat only. When air enough entered to allow this to be done, the combustion was good. But the complicated water spaces soon failed, and the plan was abandoned. In 1856, Mr. Edward Jeffries introduced a step grate which turned bodily forward on a swivel, for the purpose of sliding the incandescent fuel towards the flue-plate so as to make room for fresh coal behind, the gas of which, passing with air over the live coals, was burned. The mixture, however, could not have been perfect, although the plan gave moderately good results, with "favorable kinds of coal" on the Shrewsbury Railway. In 1856, Mr. W. G. Craig introduced the long inclined grate shown at plate 43, on the Manchester, Sheffield & Lincolnshire Railway. Tlis plan has not been successful, according to Clark, who accounts for it by the injurious restriction of the air-passages and mixing-spaces, and the excess of cooling surface. The mere grate, however, was well designed to develop the principle which was afterwards quite successfully carried out by Mr. J. J. Cudwortlh, on the South-eastern line, in the boiler shown at plate 40. In Cudworth's boiler, not only is the gas from fresh coals at the top of the inclined grate made to pass over the incandescent fuel below, but alternate firing is also employed, and an igniting temperature is furnished to the fresh products of one fire by the heat of the other. But as far as alternate firing is concerned, the same mistake as that alluded to in McConnell's engines is made; the space between the front end of the mid-feather and the flue-plate is too short to allow thorough mixture or ignition. A shorter mid-feather or a combustion-chamber would not fail to give better results, with this arrangement. The combustion in Cudworth's boiler is good. On an experimental trip, the author did not observe any black smoke, at firing or stirring the fire, when the coals were carefully laid on at the extreme back end of the grate and the gas allowed to distil regularly. Throwing coals upon the front of the fire caused a large amount of black smoke. Quite recently, these principles have been better carried out on this line, and, on the whole, the plan is quite satisfactory. In 1859, Mr. S. H. Head, of Boston, patented the boiler shown at plate 44~, and applied it on the Fitchburg Railway. A mid-feather A extends from the bottom to the roof of the fire-box, 100 THE ECONOMICAL GENERATION OF STEAM. joining the front plate, and extending to within 6 inches of the back plate. A damper K alter. nately covers the entrance of each compartment of the fire-box into the combustion-chamber. Thus when fresh coal is put into the right-hand compartment, the damper being closed over the end of that compartment as shown by fig. 1, the whole of the gas is compelled to pass back towards the doors where it may receive an abundant supply of air, around the back end of the mid-feather, over the whole of the incandescent fuel in the left-hand compartment and through the combustion-chamber. The run of the gas is very long, and the opportunity of mixing with air, and of ignition, is particularly good. The results in practice, as to combustion, are very satisfactory. The mid-feather is wide, giving good circulation. This is the most thorough of all the plans of this character Wvhich have been introduced. The only objection to it, is its want of simplicity as compared with the plane fire-box and combustion-chamber. It is highly probable, however, that the use of thin plates, especially of corrugated plates, as suggested on page 32, would prevent the extra repairs which mid-feathers, as ordinarily made, invariably necessitate. The results of practice as well as the evident principles embodied, would indicate, then, that the only feasible methods of igniting the air and gas by passing them over the incandescent fuel or through the heat of a coke fire, by the aid of the draught, and at the same time of giving them time and room for combustion, are that of Cudworth and that of Head. 5. By adnzittiny Air through the Fire.-This method of admitting air for combustion has the merit of antiquity, which accounts for the persistence with which it is adhered to in some quarters. Its impracticability for locomotives has been already alluded to. The thinnest fire that will furnish heat enough for the necessary steam generation, is not, in practice, thin enough to admit the requisite quantity of free air for thorough combustion. Except at firing, indeed, there is generally no smoke, if the fire can be kept sufficiently thin and loose; that is to say, if the management of the engine is remarkably good, and if no clinker forms, stopping up the fine grate necessary for this kind of firing, the chances are, that a thin fire will admit air enough to insure pretty good combustion, except where the distillation of gases is very irregular as at the time of firing or stirring the fire. The greatest amount of air for the gas is required when fresh coal has been fed, or just at the time when the fire being thickest, the least can get through. And when no air is wanted above the fire, the most can get through. The air supply cannot be regulated, and as far as it regulates itself, it is exactly opposite to the requirements of the case. Or, if the distillation of gas is equalized by feeding a handful of coal at a time, each moment, there must be enough extra heating surface to compensate for the condensation of all the air rushing straight. through the flues, that can enter a constantly open door. This method also necessitates a large grate, which is less advantageous than a small grate in promoting intense combustion and in maintaining an igniting temperature in the fire-box. All the requirements of the case can be better and very much more simply met by allowing air to enter in regulated jets, close to the surface of the fire. To attempt, then, in spite of all the risks and inconveniences mentioned, to get air enough through the fire, and to persistently exclude it from above, is as absurd as it would be to make the exhaust steam move a steam-engine and the engine work a pump, and the pump create a vacuum in the smoke-arch, instead of accomplishing the result at a stroke, by the simple blastpipe. This by no means proves that thin fires, and grates adapted to carrying them are unimportant. A thick fire, as before mentioned, allows carbonic oxide to form, which, in default of its subsequent mixture with air and ignition, is a waste of fuel. A thin fire prevents this loss. But with the common wide grate bars and spaces, it will be fomund impossible, even with careful manipulation, to keep a thin fire regular and equally loose and intense in alloits parts. There will be strong local draughts and holes in the bed of coals. The perforated grate used by Mr. Bullock of the Old Colony Railway,will carry a regular fire from 4 to 6 inches thick. It consists of a castiron plate 48 X 36 inches in plan, and 1 inch thick, in which are as many 1-inch holes, i-j- inch apart centres (-T5 inch apart outside), as there is room for. Some air, of course, enters through this grate, and 30 to 40 hollow stays-bolts supply a larger amount. Bullock's grate would, of course, be stopped with cinder in all cases where the coal is so impure as to require the MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 101 shaking grate. With the best coal, it has proved valuable in the particulars which have been mentioned.'Figs. 11 and 12, plate 58, show a plan of slotted grate employed on the Great Western Railway of England, and on Boydell's traction engines, for similar purposes. Fig. 9, plate 54, illustrates still another variety, used in England. Such grates are also better preserved than rectangular bars, having a larger surface contact with air. 6. By Jets of Steam.-The induction of air by steam jets is not referred to under this head, being another matter altogether. In 1838, Mr. M. W. Ivison applied a jet of steam, blowing across the fire, to one'of the engines of the Edinburgh & Glasgow Railway. The plan did not prevent smoke, but deadened the fire. In 1857, Besnier de la Pontonerie, of Paris, applied the arrangement shown by fig. 9, plate 55, to several boilers on the Eastern Counties and North London Railways. Steam from the boiler was blown in a superheated state, through small jets into the gases. Although smoke was, to a considerable extent, prevented, the fire was dampened, a large amount of extra fuel was required, the steam-pipe burned out, and, on the whole, the plan was not satisfactory. Steam thrown upon the fire, no doubt, tends to prevent smoke, partly in the following manner: the oxygen of it, unites with a portion of the solid carbon on the grate and forms carbonic acid, and the hydrogen goes to lighten the heavier hydro-carbons, reducing them to forms better able to exist at high temperatures, holding their carbon in suspension. And the action of steam is partly mechanical, mixing the gas with what air may leak into the fire-box. But as far as the author has personally observed, during several trips over the Eastern Counties line, ~on engines fitted with Frodshalm's apparatus (Figs. 10 to 15, Plate 55), the admission of steam made no perceptible difference in the amount of smoke. This plan consists in forcing air already directed down upon the fire, into mixture with the gases, by jets of steam, as shown by the dotted lines at A A, fig. 14. For experimental purposes, however, the air was repeatedly shut off after firing, the steam-jets alone being turned on. In fact, no perceptible change occurred in the amount of smoke at the chllimney-top by turning on the steam, so long as gas was distilling. There are other considerations of interest, about the agency of steam, however, which, are mentionedl by Clark in " Recent Practice," as follows: —" There is one little feature, of considerable importance, common to all (locomotives) —the steam blow-pipe or auxiliary jet in the chimneywhich is made use of for the purpose of continuing the draught of the furnace, when the powerful artificial stimulus of the blast is at intervals suspended. The action of this smoke-annihilator is well worthy of study, and it is a question of directly practical interest in what way the blow-pipe operates thus serviceably. Looking into an ordinary fire-box, at rest, under the action of the blow-pipe, smoke may be perceived in the fire-box, wending its way into the tubes, which becomes totally invisible at the top of the chimney. By what process is this visible smoke made invisible? It seems to be by absorption, or precipitation, or a little of each, by the steam from the blow-pipe:the jet of steam'paints the smoke.' Imperfect though such a mode of banishing smoke may appear, it is nevertheless true that every one of the coal-burning contrivances falls back upon the blow-pipe, for the means of consummating the extinction of smoke. The action of the steam-blast in destroying visible smoke, while the engine is at worlk, is apparent to the most ordinary observer, inasmuch as the discharge from the chimney, which may be clear and colorless when the blast is on, may become densely brown or black when the blast is off, and may only be mitigated by a discharge of steam from the blow-pipe. That steam possesses a considerable power of precipitating smoke particles, may readily enough be observed over an ordinary domestic fire, by directing the steam spouting from a kettle into the ascending smoke. It is found to banish the smoke in a greater or less, degree certainly not by any act of combustion or other chemical process, but by simple physical action. And, it is in virtue of the same faculty of absorbing or precipitating smokeparticles, that crttain processes for preventing smoke operate by passing the smoke into intimate contact with:water. It is found, similarly, that smoke from stationary chimneys is subdued or mitigated by discharging the exhaust steam into the chimney. Of course, the action of exhaust steam, so directed, in accelerating the draft, and thus in another way diminishing smoke, is distinguishable from its action as a precipitant or absorbent; but, in many cases, doubtless, the prevention or reduction of smoke by the agency of steam ill the chimney, is due mainly to its operation in the latter capacity. 102 THE ECONOMICAL GENERATION OF STEAM. "It cannot properly, therefore, be assumed that combustion is in every case complete when smoke is prevented by the agency of steam in the chimney: —whether the engine be running under the blast, or standing under the blow-pipe. This is a fair subject for practical investigation; meantime, one conclusion of practical value may be drawn from the general experience of the smoke-preventing agency of steam, and that is, that we have to guard against being misled by appearances, and should not be diverted from the accomplishment of the main object-the thorough combustion of coal by ample admission and intimate mixture of air above the fuel." The steam-blast in the chimney, from whatever source, must be chiefly effective in drawing in and mixing air with the gas; and since continuozous smoke prevention requires air, in any case, the influence of steam in absorption and precipitation cannot be of much practical account. 7. The downward draft.-This method of burning coal was employed long before the invention of the steam engine. It has recently been applied with good results in various kinds of stoves. Unfortunately, there are no complete and well-reported experiments on its results in smoke prevention. The general belief is that it does mix the air and gas and ignite them; there is no apparent reason to the contrary. An abundance of air of course strikes the fuel at the point where the gas is distilled, and cannot fail to mix thoroughly. Ignition follows as a matter of course. What practical objections might arise, it is impossible to anticipate. In 1857, Mr. J. K. Fisher, of New York, published a design for a downward-draft locomotive furnace; the water-grate (which would evidently be necessary on account of the intense heat) consisted of tubing through which the feed was to pass. No steam could be made in the grate, and the latter would there. fore be well preserved, on account of its limited surface, so long as a Constant current was main, tained through it. The indispensable counterpart of such an arrangement would be the independent feed. In 1858, the boiler shown at fig. 4, plate 56, was patented by Mr. Daniel Evans, of England. The grate is constructed like the legs of the fire-box and would be likely to stand as well, if good circulation could be maintained in it; forced circulation would probably be required. It is possible that the downward draft may prove important for some purposes, such as light locomotives or city steam-car boilers, which can more conveniently be fed at the top. It appears, at least, to deserve a thorough trial. CONCLUSION.-There is, practically, but one plan of simply admitting air by the aid of the draft, and of igniting the mixture thus formed, which can be relied on as insuring perfect comr. bustion, under ordinary management, and that is Beattie's combustion-chamber and fire-bricks. But the cost, weight, and renewals of this arrangement, offset to a considerable extent, the advantages of perfect combustion. Passing the air and gas over a long bed of incandescent fuel, as in Head's boiler, is, in case of good management, certain to promote thorough combustion. This boiler, however, is not so simple, and with present materials and construction can be hardly so durable as the plane fire-box. The price of perfect combustion, then, is increased cost of repairs. And neither of these arrangements, however valuable, can be applied to old engines at a less cost than from $1,000 to $2,000, depending on the changes to be made in the framing, etc., to make room for the combustion-chamber. -The simple admission of enoughi air to the fire-box, at the surface of the coals, if the.supply is skilfully regulated, if the fire is kept thin and clean, and if good, clean fuel is supplied frequently, a little at a time, will insure very much better combustion than is common in coal-burning boilers, and will prevent smoke, carbonic oxide, and the escape of carburetted hydrogen to such an extent that the burning of coal on our railways will not be offensive to passengers and to adjacent residents, and at the same time more work will be got out of the fuel. At times, when the fire is stirred or much coal added, combustion will not be perfect, but it will be at least improved. To admit perhaps a tenth part of the air required, however, to a part of the fire-box where it can do no possible work except condensation, will not prevent smoke. Deflectors, or mixers of air and gas, especially if they are igniting points also, will very decidedly aid in the perfection of combustion. Combustion-chambers, by giving more time and room for mixture, are also important auxiliaries; but neither of these features are of any value without a plenty of air; nor do they do so much as modify the conditions of air admission. When MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 103 deaflectors are simple in construction, but favorably arranged, as in Eaton's boiler (Plate 45), the expenses of maintaining them, although quite appreciable in the aggregate, are small in proportion to the advantages derived. With proper air admission and judicious firing, this plan is still more likely than simple air admission, to save smoke and money; but it cannot be relied on forperfect combustion. 2. THE FORCIBLE MIXTURE OF AIR WITH GAS AT THE SURFACE OF THE FIRE.-If we admit an excess of air into the fire-box, even in fine jets, and then allow the draft to carry it and the gases in the same general direction, they require all the time and space to mix, that combustion-chambers and deflectors can possibly give them in a locomotive fire-box. And when they are mixed, the work is but half done, for they must be ignited and provided with room to burn upelse they are lost. If merely ignited at the end of the combustion-chamber, they almost instantly enter the flues, where combustion ceases. But if while the gas is rising in a body from the bed of coals, we intercept and break up that body of gas by streams of air driven down upon it and /througyh it in every direction, there will be no mistake about the intimacy and the locality of the mixture of air and gas. There will be a combustible mixture immediately over the bed of coals the instant the gas is generated, without the aid of deflectors, mid-feathers, or combustionchambers; and the mixture will then and there ignite and burn simply because the heat of the incandescent fuel which distilled the gas cannot be prevented from setting it on fire and keeping it on fire so long as air and gas are mingled at a point within its influence. This ignition will be done, moreover, without the aid of fire-bricks, cast-iron igniting points, or the incandescent fuel in other fire-boxes or any heat except that which distilled the gas-the solid carbon burning on the grate. It would hardly be possible to conceive of a more simple plan of coal burning than this. After men had vied one with another to complicate the coal-burning question-and fire-boxes, it was reserved for Mr. D. K. Clark, of London, to reduce both to the simplest forms, by forcing a mixture of air and gas where there was heat to ignite them, by means of the steam jet arranged as shown by figs. 1, 2 and 3, plate 56. These principles may, indeed, have been contemplated by other practitioners, and were, perhaps, approximately carried out. But Mr. Clark's was the first simple and thorough plan of insuring practically perfect combustion, and is still the best. Before detailing the circumstances and facts relating to it, there are other simple but quite effective devices to be mentioned, which embody the same principle to some extent. As early as 1857, the firemen on English lines had observed that by placing their shovels in the doorway of the fire-box so as to direct the inward current of air downwards, the smoke was more effectually prevented than by allowing the air to move straight towards the flues. This fact led to the introduction of a permanent shovel or hood in the doorway, which has since been variously modified, and is very effective in smoke prevention. The instant objection of practitioners will be-it is best to stop and answer it seriously now —that the hood will burn out. It will. Those on the North London line, substantially as shown at fig. 10, plate 67, last three weeks. When they are cracked and warped so as to be unfit for use, three rivets are cut off at A, a piece of old boiler or other plate is attached, and so on. Practically, the hood costs nothing. The arrows on the figure, show the manner in which the air is thrown down upon the fire. In view of the strong draft caused by the blast-pipe, it would be expected that the air would turn short around the end of the hood, and go into the flues. In fact, its momentum is so great that it cuts away the fuel in the centre of the grate by supplying air to the solid carbon as well as to the gases. The fire-door is sometimes removed altogether; when it is necessary to add fresh coal, the hood is easily lifted by pressing down its handle, to the position shown by the dotted line. Some cold air, of course, enters the flues, but not enough to prevent maintaining as high a steam pressure as was formerly done with coke in the same boilers. With careful firing, there is rarely any smoke. Too much air enters, of course, when no gas is distilling. It is, therefore, best to keep a constant supply of fresh fuel on the fire. Were a door so arranged as to shut off the air, the hood would rapidly burn out; a constant stream of air preserves it. If there were no hood, most of the cold air would directly enter the flues; as it is, the door may be kept constantly open without much loss, provided gas is always distilling, so as to utilize the air. For goods-engines 104 THE ECONOMICAL GENERATION OF STEAM. with heavy trains, where a large amount of fuel is required, there is no objection to almost constant firing-a little at a time. And it will be found that this simple arrangement, with careful manipulation, will keep the box full of white flame, and will perfect combustion to such an extent, at least, that the smoke will not be a nuisance. Should any propose to " try it," it is earnestly requested of them that they try it and not a distortion of it (which is too frequently the way new things are "tried"), before concluding finally as to its merits. In 1858, Mr. Douglass, of the Birkenhead & Cheshire Junction Railway, applied to each of 30 engines, a baffle-plate or hood fixed to the inner back fire-plate, over the door, and extending down towards the middle of the fire-box. The results are quite satisfactory; the value of any such arrangement, however, depends on the proximity of the lower end of the hood to the fire. Mr. ILees, of the East Lancashire Railway, uses a hood attached to the door frame (Fig. 8, Plate 56), together with an arch of brick or iron, with tolerably good results. Messrs. Hawthorn, locomotive builders, employ the hood shown at fig. 3, plate 55. The door is composed of a number of inclined plates which admit separate strata of air. A jet of steam is employed below the door to farther mix the air and gas. There are other modifications of the hood, bonnet, deflector, baffle-plate, or shovel, as it is variously styled on different lines. A method employed on the Caledonian Railway and shown at fig. 4, plate 55, is peculiar-simply a triangular box over the doorway, the air entering at the top, where it is regulated by a slide, and dropping down, it seems, by its momentum, to the surface of the fire. This certainly would appear to be little better than leaving the door open, at first sight; the fact is, however, that such a downward direction is given to the air upon entering, that it visibly strikes the coals, however strong or soft the blast may be. There is nothing to burn out in this arrangement-it is almost ridiculously simple. But a large amount of air must enter and cool the flues-not enough, however, to offset its advantages in aiding combustion. Smoke is not entirely prevented, as far as the author has observed —but there is none of that thick black smoke which characterizes too m-uch of the American coal-burning practice. 2. By Eternal Steam Pressure. —The arrangement patented in 1859, by Mr. Frodsham, foreman in the Eastern Counties shops, is shown by figs. 10 to 15, plate 55. The door is hung at the bottom of the doorway (Figs. 10 and 11), and a flat movable deflector (Figs. 12 and 13), is hung inside the doorway-the two, as shown by fig. 15, tending to throw down a regulated supply of air upon the fuel. To insure the mixture of air and gas, two steam pipes enter the sides of the fire-box at the level of the door (fig. 14), from nozzles in which, diverging jets of steam strike, perforate, and farther mix the masses of air and gas, at the surface of the coals. In September, 1859, the author made several trips upon engines fitted in this manner, over the Eastern Counties line. With coarse coal, when the management was skilful, there was no visible product of combustion at the chimney. With fine slack coal, the distillation of gas was more rapid, and there was a little amber-colored smoke. The air without the steam jet, prevented smoke, except at the instant of firing; when the door was shut, and air excluded, the steam jet made no perceptible difference in the combustion; and after firing, with or without the jet, the air being excluded, the smoke was invariably excessive. D. K. Clark's coal-burning apparatus, patented in 1857, is shown by figs. 1, 2 and 3, plate 56, and is variously modified in order to be conveniently applied to different classes of engines. The plan preferred is shown by fig. 3, and consists of 8 hollow stays or tubes as close as possible to the surface of the fire, in the front, and as many more in the back of the fire-box. These tubes are generally 2 inches in diameter, and fitted with a slide, as shown, to close them altogether, when no gas is distilling. From 2 to 3 inches from the outer end of each tube, is a?,-inch nozzle in a steam pipe. Jets of steam issuing from these nozzles and passing through the open air into the fire-box, induce annular currents of air to enter the fire-box with sufficient velocity to spread themselves over the entire surface of the coals, and to mix with the gas as it distils. The amount of steam is regulated by a cock in the pipe. The nozzle is simply apiece of brass wire tapped into the pipe and then bored. The whole cost of the arrangement can hardly be made to exceed $50. The steam may be condensed, to some extent, before it enters thle fire-box, as it is condensed in'the Giffard injector (Plates 62 and 63), and the air is the principal if not the only agent in MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 105 combustion. In fact, the jet is similar in principle to the injector; the air assumes the velocity of the steam, even if the steam is condensed. The author has made a number of trips on engines fitted with this apparatus, over various lines, and has observed the following results, which are also corroborated by the report of observers generally. 1st. When the jets in the front of the fire-box, only, are applied, and fine slack coal is thrown upon the fire, the jet of air blows the fine coal that had risen by the force of the draft to the back part of the fire-box, where it may again lodge, or may, in the course of its passage, become so heated as to part with its gas; it certainly does not go bodily into the flues, for there is very little combustible matter found in the smoke-box. The motion of the currents is much the same as if there were an arch in the fire-box. It may here be suggested that if this fine coal were blown over a dead plate, it would probably lodge, and gradually part with its- gas and be burned. 2d. When after firing or stirring the fire, or at any time when gas is distilling, air jets enter at the front or at the back of the fire-box, or at both these points, the steam jet being strong enough to force in and the tubes large enough to admit all the air required (the sizes mentioned are large enough), the air and gas mix perfectly, ignite instantly, and are burned so entirely before reaching the flues, that no smoke nor amber-colored vapor nor any visible product can be seen at the chimney-top; that is to say, during the ordinary working and replenishing of the fire the use of the steam-inducted air jets, invariably makes the combustion practically perfect. 3d. Upon opening the fire-door shortly after firing, when there is not so much flame as to hide the surface of the fire, the 8 front jets being turned on, there will appear, close to the surface of the coals, 8 hollow cones of white flame occupying nearly the whole of the bottom of the firebox, and the remainder of the fire-box will be more or less filled with white flame, in proportion to the amount of gas distilling. 4th. When the engine is standing, and at any time when steam is shut off from the cylinders, the air jet will prevent smoke without turning on the steam jet in the chimney, and without raising the steam pressure, because no air is being drawn through the grate. Ordinarily, the steam jet in the chimney, taking the place of the blast, must be applied with full force to draw air enough In above the fire, to prevent smoke. But if the ash-pan and damper are not perfectly tight (they never are), more air is coming in at the grate all this time than above the fire, and the steam pressure is rapidly rising. If the stop is long, one of two nuisances must be tolerated at the very time of all others when no nuisance should be allowed-at the time of standing at stations and in streets. Either the jet in the chimney must be turned off and the smoke allowed, or else the steam, will escape at the safety-valve loudly enough to frighten horses, not to speak of wasting fuel and running risk of explosion with the common kind of safety-valve. And the chimney jet itself makes a great noise. But with Clark's arrangement, the chimney jet is either not used at all or may be turned on just enough to keep a gentle current moving through the flues; the air jets above the fire will prevent smoke without raising steam, and this is the only arrangement in use which will accomplish this most desirable object. It is so highly approved for this reason alone, that it would probably come into extensive use even if it prevented smoke at no other time. 5th. The saving of coal, due to smoke prevention by the jet, as compared writh excluding air from the top of the fire, was found to average 18 per cent. This was the result shown by a careful experiment lasting one month, on the Great North of Scotland Railway. It is found, on the contrary, by Mr. Lees and others, that although combustion is improved, condensation is increased and no saving in fuel is made, when air is admitted inbulk, over the fire. In short, Clark's steam-inducted air jets, absolutely prevent smoke, when properly arranged and carefully adjusted, in all cases where the firing is not especially bad. And there is nothing to offset this economy of fuel and this prevention of nuisance. The first cost is trivial. There are evidently no parts that can burn out, and so far, no depreciation of other parts has been traced to it. It has been suggested that the steam will decompose the iron —two years' use have not demonstrated any such effect, and it has not been observed that the 900 lbs. of water, in the shape of steam, which result from the combustion of each ton of bituminous coal, is particularly destructive of fire-box plates. A small amount of steam is required, but this is so small, that the 14 106 THE ECONOMICAL GENERATION OF STEAM. proportionate'saving of coal well burned by this method over coal imperfectly burned, is quite as great on the average, as the saving due to good combustion in any other form of boiler. Besides, the steam jet is turned on only while gases are distilling rapidly from the coal; when the distillation has nearly stopped, the less amount of air required is drawn in by the draft, and when solid coke is burning, the slide over the air-tubes is closed. Thus with two little levers, the engine man can regulate the air supply exactly; the appearance of the fire and of the chimney-top being his guide. It may always be safely concluded that if the coals on the grate are not above 12 to 14 inches thick, and if the exhaust steam is not tinged with smoke, the combustion is practically perfect. A very thick fire might be invisibly wasting fuel in the shape of carbonic oxide. The objection to placing all the tubes in the front of the fire-box is that they would interfere with good circulation. In new fire-boxes, this might be provided for; in those made for wood-burning, the spaces are already too narrow. The jets can be readily distributed, however, so as not to interrupt free circulation. Four holes in each of the four legs of the fire-box are sufficient if fresh coal is added a little at a time and frequently..One rule must be invariably observed in applying this arrangement: the tubes must be placed close to the highest level of the fire, to insure perfect combustion. Another feature is, perhaps, equally important-the steam nozzle should he removed at least 2 inches from the mouth of the tubes. It has been remarked in the introductory chapter, that Clark's jet has been discarded on several American lines, because another and essentially different and absurd device did not work. When the Giffard injector was first heard of, one or two geniuses here pronounced it a " humbug," because a contrivance of their own, which they called the Giffard injector, but " improved " by them, was abortive. Those whose efforts in this direction failed because they had not correct drawings to work from, are of course, not referred to. But it is an unfortunate fact that there are too many "practical railroad men " who are never content to use any device whatsoever, which they have not "improved." Clark's apparatus is the best apparatus in use, as far as the author is aware, for adapting old boilers to coal burning, and it is perfectly and entirely satisfactory vwhen properly applied and used. It is to be hoped that the failure of other and different devices which are called " Clark's jet," will not prevent the extensive employment of this extremely useful invention. Nor will the morbid determination of some of our " economical " railroad men, never to use a " patented article," stand in the way of its introduction. The invention is not patented here. So the magnanimous scruples of the philosophers referred to, need not stand between them and economy and the comfort of passengers, in this particular case. To the majority of railway managers who want only the facts, the author would repeat the conclusion before mentioned: there is no mistake about the excellence of this plan of coal burning. Care must be taken to adjust it properly and to manage it.skilfully. If it fails in any case, it will fail only because it is imperfectly applied and used. 3. GRATES~ Mater-Grates.~For bituminous coal burning, the water-grate is believed to be inferior to the solid grate, because it conducts a large amount of heat from the fire, which might be better employed, at first, in maintaining an igniting temperature in the fire-box. The watergrate is employed with anthracite coal, because the solid grate will not stand. For bitutninous coal which is liable to clinker, the ordinary shaking or rocking-grate could not, of course, be used. The arrangement shown at plate 67, figs. 11 and 12, however, is intended to "slice" and clean the fire, with either the water-grate or the solid grate. Intensifying Combustion, Small Grates, etc.-The employment of solid grates, and even of firebrick or non-conducting grates for bituminous coal, if they could be made to stand the heat and weight, is a part of the general system of intensifying combustion, the value of which was alluded to in the chapter on the chemistry of combustion. Some practitioners go so far as to say that no heat should be transmitted, especially from the gas to the water, till combustion is completed, or in other words, that the fuel should be burned in one place and utilized in another place. There would be practical objections to this system, especially in the limited space allowed for locomotive boilers. But intensifying combustion produces the same effect as far as it goes. This has already been done, to some extent, bly using small grates; and there are, doubtless, practicable means of farther carrying out the principle. MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 107 The room for the boiler has been so small that, in order to get the desired extent of surface, it was necessary to make the fire-box small, so as to leave as much room for flues as could be allowed; hence, the grate was relatively contracted as the size of engines increased, until it became much smaller than the grates used for other boilers. This diminution of size was found to be advantageous, and was carried to an extent not at first deemed allowable. The large grates of old engines were partially stopped up by dead plates, and were found to work better. At first a mere strip of iron was placed next the sides of the fire-box, for the purpose of preventing cold air from running up its sides. Gradually the plates were increased in size, till they now cover half the area of the grate, in many cases. And a corresponding intensity of combustion and saying in fuel have been noticed, much more of course than could be due to merely keeping cold streams of air from running up the fire-plates. New engines were built with reduced grates, some years since. Crampton's great Liverpool had 1 foot of grate to 106 feet of heating surface; Eddy, of the Western (Mass.) Railway, made grates of about this proportion for wood; several in England, made grates 1 to 90 or 96; and an English designer in Austria, gave the proportion of 1 to 125A to a wood burner. D. K. Clark found that engines with small grates worked best, and advised that they should be not larger than 1 to 85, and decidedly intimated that they might with advantage be smaller —coke being the only fuel which he at that time considered suitable for locomotives. Coal burners formerly had much larger grates. McConnell, Winans, Dimpfel, Boardman, and many others have made them as large as they conveniently could. A. F. Smith, of the Hudson River Railway, on the contrary, though preferring a large fire-box, covered a third of the grate with a dead plate in wood burners, which he adapted to burn coal, and therein dissented from the then common opinion that very large grates are expedient for coal. It seems that the preference for large grates arose before the plan of admitting air in jets above the fire was in vogue. It was found by trial that a thin fire, 4 to 6 inches, through which the air could flow in sufficient quantity to burn the gases, produced the best economy; but so thin a fire would not make steam enough unless the grate was large. After the air-jet system was adopted, smaller grates were i made to work well with coal; and Clark's steam-inducted jets make the old coke burners, with small grates, burn coal with freedom from smoke, and with 7'76 lbs. of water evaporated per lb. of coal, which compares well with the average work done by coke. The evidence leads to the conclusion that'&here need not be a sacrifice made for the sake of a large grate for bituminous coal, such as spreading the fire-box laterally beyond the wheels, or extending it backward behind them. On the other hand, it is on the safe side to make the firebox as large as is consistent with the general plan of the engine; observing, however, that narrow water-spaces are far worse evils than small grates. If the grate is too large, the dead-plate is a cheap remedy; and the large fire-box is useful for combustion room. The influence of the grate on the draft is one of its important points. A small grate and deep fire requires a stronger draft than a large grate, and may require a narrow blast-orifice. But the admission above the fire of ~4 of the air, which is allowable in coal-burners, reduces the resistance by more than one-half. On the Chicago & Rock Island, and other western railways, Mr. Malden Wright introduced in 1857 the first very small grate that was employed for coal burning. It was chiefly intended to intensify the heat sufficiently to melt the cinder in the very impure coal used there, so that it would run directly into the ash-pan. In this it was quite successful; occasionally, however, the cinder was carried upward by the intense draft, and became plastered over the flue-plate, stopping up the flues. But the results as to combustion were equally noticeable. The entire bottom of thle fire-box was covered with a sloping pavement of fire-brick, except an opening 12 X16 inches, in the middle (for 16 X22 inch cylinders) in which rested the grate. There were in fact, two grates for alternate use, one in service and the other hanging down. The blast orifice was reduced to 21 inches diameter. The combustion, air being admitted through the door, was very much improved, simply by the intensity of the heat in the fire-box. More smoke was made with the satne air admission, when the large grate was employed. But the draft was so strong that the soft flue material was soon cut away by the flying particles of coal, and the repairs were considerable. The arrangement was used on the Rock Island line for. 9 months, and was, on the 108 THE ECONOMICAL GENERATION OF STEAM. whole, quite satisfactory, but wood being cheap, no especial effort was made to introduce coal at that time, and the plan was abandoned. On the Illinois Central line, Mr. Hayes has for some time employed a grate 39 inches long, by 18 inches wide, for the same kind of coal, the length of the fire-box being 53 inches, and its width 30 inches. The 14 inches dead-plate in front, is employed as a drop-grate for hauling the fire; the 6-inch dead-plates on each side have sloping sides. In the Rogers engines, plate 46, the solid drop-grate B extends backward from B to i the length of the fire-box. The grate used by Mr. Jones on the New York Central line, is shown by figs. 3 and 4, plate 54. There is an evident advantage in a dead-plate in front; the entering air tends to take the shortest course from the mouth of the ash-pan to the flues, and hence the most of it enters through the front of the grate. A dead-plate carries the current towards the centre of the fire. Mr. Hayes has put transverse partitions in the ash-pan, to produce the same results. Perhaps a better plan would be to taper the slots or perforations in the grate, in the same manner that unequal perforations are made in steam-pipes for the purpose of making an equable draft on the -boiler. Another method of intensifying combustion without greatly reducing the grate area, and the blast orifice, has been proposed. The grate would consist of a box say 18 inches deep, projecting below the legs of the boiler. The sides and bottom of the box would be slotted or perforated so as to admit air to the coals in every direction. Much of the air that entered the upper holes or slots would mingle with the gases and perfect their combustion. None of the gases would be blown against the cool sides of the fire-box, but the whole would tend to assume a central current, which would longer maintain its high temperature, and thus insure combustion. Yates' conical grate, plate 52, has proved valuable in one of the particulars mentioned-in admitting air above and at the surface of the fire. Size of Openings.-The fact has been well-established that a thinner fire can be carried with small and evenly-divided openings, than with wide slots such as are employed for wood-burning. The only objection to narrow openings is, that they are more liable to be clogged with clinker. But movable grates, like that shown on plate 46, will keep any opening clear. The practice is so diverse, that there is probably no correct standard arrived at. Since a thin fire allows some air from beneath to mingle with the gases, and especially since it prevents the formation of carbonic oxide, there would appear to be a decided advantage in making the'openings as small and numerous as is consistent with the stability of the structure. fMovable Gruates for cle aningand stirriny the Fire. —The use of the raw, impure coal, from the newly opened mines which supply many of our railways with fuel, renders such grates very generally important in this country. And with the best fuel there is a great advantage inbreaking up the coke as it forms, and in keeping the fire loose and even. On the contrary, emptying perhaps a quarter of the coals into the ash-pan, to acoomplish this object, is hardly necessary or economical. A good movable grate, therefore, was not perfected at the first trial. The grate used in the Baltimore & Ohio R. R. coal burners is shown by fig. 12, plate 54. Two bars are cast together, and the pairs being connected are rocked together. On the western division of the New York Central, the arrangement shown at figs. 3 and 4, plate 54 is employed. The grate is in two sections, which are rocked together. The fire may also be dumped by this grate. Eaton's grate (Plate 45) simply dumps the fire, and is not used to shake the coals. A pendulous grate, each bar resting on a journal, and tending to vibrate with the motion of the engine, is shown by fig. 8, plate 54. Chains suspended across the fire-box, so as to form a bottom or perforated grate, which shakes with the motion of the engine, was at one time employed on the Long Island, Erie and other lines. The plan illustrated by fig. 11 and 12, plate 67, was suggested by Mr. Davis, foreman at the Jersey City Locomotive Works, in 1855, but has not been applied. It need not be associated. necessarily with the water grate, although the blades A would be better preserved from warping and burning, in this case. The thin plate-blades shown, are fastened together by rods and thimbles, and being raised by arms working in slots, they slice upward through fire, cutting masses of coke into small pieces, and pushing any large masses of clinker to the top of the fire, MEANS OF APPLYING THPt PRINCIPLES OF COMBUSTION. 109 where it can be removed. This is obviously the only plan of thoroughly loosening and stirring the fire. For breaking up and throwing down small formations of clinker, as they occur-which is in fact the great object of a movable grate-the plan of Hudson & Alien, shown with the Rogers engine, plate 46, at A figs. 1 and 3, and at fig. 4, has now been largely used for three years, with very satisfactory results. Indeed, few locomotive improvements have been so generally and highly approved as this. A simple rocking grate merely allows whatever lies loosely upon it, to drop through, and this is generally ashes and coals. Formations of clinker adhere to the bars, and spread over several at a time, and are not removed' without some force. Grates which open widely enough to let the clinker fall through, drop and waste much of the fire also. But with the arrangement under consideration, one of the bars shown at fig. 4, rises bodily, while the next bar-falls, and so on alternately for the whole width of the fire-box, thus leaving wide trenches into which fall pieces of cinder, or at least their corners or edges. When the bar which had been raised is again brought down, these pieces of cinder are cut in pieces by the projections on the bars and are forced down into the ash-pan. This action takes place only at the bottom of the fire, so that very little, if any coal is carried down. At the same time the whole bed of coals is loosened. The grate is moved from time to time, as circumstances may require from the footplate, by the lever O. The shape of the openings is also favorable to thin firing. One of these grates made at the Rogers Locomotive Works, has been in constant use for three years on the Great Western Railway of Illinois. This is a very long service for cast-iron. On the Mine Hill & Schuylkill Haven Railway, Mr. Wilder employs a toothed " raker " for clearing an anthracite coal fire. The instrument lies, when not in use, in the bottom of the ashpan. When the teeth, or fingers are raised between the grate bars (as Davis' blades,Fig. ll,Plate 67 are raised) they become connected to an eccentric on the driving shaft, which moves them back and forth, clearing the fire. It has already been remarked that V shaped slots in the tops of bars will fill with ashes, which prevents cinders from adhering to the metal, and is also a non-' conductor. It is reported that on the Taff Vale Railway, fire bricks broken into cubes and strewn upon the grate, preserve the bars without stopping with clinker. The Delano grate, illustrated by fig. 1, plate 54,was applied to several locomotives on the Pittsburg, Fort.Wayne & Chicago and other railways in 1856 and 1857. The coal to be fed was placed in a box say 12 X 18 inches in plan under the foot-plate; the box was moved forward by a lever, and the coal was pushed upwards into the bottom of the fire by a movable bottom or piston, actuated by a second lever. A leaf or plate attached to the front of the box, covered the opening when the box was drawn back. The object of this arrangement was to " burn the smoke " by passing it through incandescent fuel; it did furnish one of the elements of the combustion of the gases heat-but there was not air enough to perfect combustion. The plan was abandoned. 4. VARmIABLE- DRAFT.~-T he smoke-box and its apparatus, and the blast will be referred to in a succeeding chapter. The consideration of the variable blast, and of other arrangements for changing the draft as required, belongs exclusively to the subject of combustion. 1. The Yacriable Blast.~The utility of an unobjectionable variable exhaust for modifying the draft, is practically acknowledged by the great number of attempts that have been made to attain one, and the extensive use which has been made of a number of these plans. It is proved by theory and practice that they are highly beneficial in all cases where the engines are worked at nearly their maximum power. Since the blast, except when gas is distilling from the coal, is the only agency by which oxygen is supplied to the fuel, that is to say, since it controls the amount of fuel burned-and whereas the efficiency of the blast is not only always in proportion to the amount of steamn generation required, the control of the blast is evidently a matter of considerable importance. The blast is strong of course, when the steam follows through most of the stroke of the piston —or when the most steam generation is required; and it is weak when the steam is cut off short or throttled, or when least is required. Thus there is a general relation of supply and demand. But the slightest practical acquaintance with working locomotives, exhibits the fact 110 THE ECONOMICAL'GENERATION OF STEAM. that the steam pressure constantly tends to fluctuate, from a variety of accidental and irregular causes. When running very fast, and when cutting off early in the stroke, the pressure in boilers that do not make steam freely tends to fall. Strong side-winds, bad fuel, snow on the track, bad state of the permanent way, accidental leakages of flues or smoke-arch, and other abnormal causes, frequently keep the pressure too low for economy. On the contrary, some boilers, burning good coal or wood, and well fired, are blowing off steam at the safety-valves most of the time, and are thus wasting an enormous amount of fuel. Now, when the engine-man varies the size of the exhaust opening to meet all these irregular causes, thereby preserving an equable pressure, the following decided advantages result:1st. If the pressure tends to fall, closing the exhaust orifice a little invariably brings it up to a point from which it can be economically expanded in the cylinders. The practice of running with very little expansion and low steam, is of course a waste of fuel-yet it is the invariable method of keeping the pressure up to a working point, when, by reason of the causes mentioned, the draft is insufficient. 2d. If the pressure tends to rise above the point at which the safety valves blow off, opening the exhaust orifice a little decreases the draft, and limits the consumption of fuel and the generation of steam to the requirements of the cylinders. And at the same time, it decreases the back pressure on the pistons. The practice of opening the fire-door to keep the steam low is very bad; the work of transmission of heat, once done, is undone; and the sudden contraction of flues and plates by cold air, is an invariable cause of leakage. Shutting the ash-pan damper is rarely effective to any considerable extent; dampers generally leak enough to admit air for pretty vigorous combustion, especially since all the power of the blast tends to make up in velocity, what the air supply loses in sectional area of entrance. And if the damper is tight, the back pressure in the cylinders is probably increased, since the blast is deprived of the aid of the natural draft of the fire. In coal-burning engines, the variable blast is important for another reason. When slack coal is thrown upon the fire, under a strong blast, a part of it must go up the chimney at once, unconsumed; so-by easing the blast a moment after firing, the coal has a chance to settle —the finest will begin to coke at once. Beattie's fire-bricks, and Clark's jet also tend to prevent fine coal from going bodily through the flues. There are some" old runners " who understand the physiology of the locomotive so well, that they can maintain an equable steam pressure by careful feeding, firing, and throttling, and literally " humor " the iron horse into good behavior; and these men are too often opposed to such innovations as the variable blast and independent feed-pumps. But there are ten times as many tyros, who need all the assistance they can possibly get. Besides, the knowledge of the " old runners," aided by the improvements mentioned, would lead to still better results. The best nozzle is undoubtedly a solid one; the~next. best is a solid end, well fitted and bolted on: but the first can not be varied at all; and the last can not be varied on the road. The reason why the variable blast has not become a common feature, is simply this:-there has been, till quite recently, no good instrument for the purpose the old plan invariably embodied one of two fatal errors. 1st. Cones or any manner of stoppers or short angles in the exhaust pipe, not only interrupt the force of the current, but scatter and deflect it, oftentimes wholly neutralizing its vacuum-creating effect. 2d. A variable exhaust which works hard-upon which the engine man has to serge and jerk at each change, as if he were reversing an engine under steam, is in fact, no variable exhaust at all, for the engine-man will never adjust it. Such is the case with Winans' " flap " exhaust, and with all others whose nozzles or gearing or' joints of any kind become gummed and sticky by the action of smoke and steam, Plates 57, 58 and 68, show four varieties of the variable exhaust which are getting into use in this country. Patrick's (Plate 57) and Addison'a (Plate 68) are alike in their main features, being modifications of Espenshade's arrangement —a horizontal disc perforated by different sized nozzles, revolving upon the top of the' exhaust pipe. The improvements consist in arranging this disc in such a manner that it can be operated. Patrick's is designed for two nozzles; Addisor's for one; but either may be modified to answer both purposes. The disc in plate 57, is MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 11l revolved by the arm of the ratchet e, catching in the notches i of the disc; its other arm k, merely sliding on the smooth periphery of plate b, which lies under the disc, where the ratchet is moved back for a new hold. Fig. 1 is a front elevation of the apparatus, in the smoke-box; fig. 2, a section of the pipes and disc; and fig. 3, a plan, and fig. 4, a reverse plan or bottom view of the disc and ratchet. More than a year's practice shows that these joints do not become so sticky or foul as to be troublesome. The apparatus is certainly simple and comparatively inexpensive. Mr. Patrick has, by personal effort, largely introduced his plan on the New York Central, N. Y. & Erie, Michigan Central, and various western lines, and he has contributed not a little towards economical wood and coal burning, by demonstrating the advantages of employing a variable blast. Addison's plan (Plate 68) is used chiefly on the Baltimore & Ohio line. A single orifice is certainly preferable to a double orifice, because the whole discharge is then in the centre of the chimney. Patrick places two pipes still farther apart than usual. The engravings of Addison's plan explain themselves. The disc and the gearing which operates it, are entirely inclosed and protected from the action of smoke, sulphurous gas, water, and other causes which soon render exposed machinery useless. Both Patrick's and Addison's revolving discs take up some room in the smoke-box. Were the nozzles high, this would largely interfere with the draft; but when the nozzles are near the bottom of the smoke-box, as they should be,* and are in the best American engines, the bulk of the apparatus is not an objection. The shape of all the orifices in these discs, is not impaired for the best delivery of steam; the latest practice shows the propriety of contracting the orifice at the end of the pipe instead of tapering the whole of the pipe. In both these plans, the discs are revolved from the foot-plate. Baker's variable exhaust (Plate 57) is employed chiefly on the Illinois Central line. The ring 1, 2, 3, 4 (Fig. 1), revolves around the pipe and within an external cylinder put on to protect it, and in the ring are the different sized nozzles. The sudden turn in the pipe is an objectionable feature. Parrott & Head's variable exhaust is illustrated in detail by figs. 5 to 8, plate 58. It has been but recently patented, and is on trial on various New England lines. Fig. 6 is an elevation of the apparatus, as it stands in the smoke-box. Fig. 8, is a section of the nozzle and casing. The staves e e of thin metal overlap each other as shown in the plan (Fig. 5), and are compressed, reducing the diameter of the opening A, by thrusting-pieces c, which are driven toward the centre by the curved wedges a on the cap (Fig. 7). The cap, here shown bottom upwards, is held down by set-screws working in a slot, and is revolved by the endless screw, as shown by fig. 6. The elasticity of the staves opens the orifice, when the outside pressure is removed. The shape of the orifice is practically perfect, whatever its diameter may be. And the apparatus can be set as low down in the smoke-box as diaphragms and-petticoat-pipes may require. Should the apparatus be found to keep clean, so as to be adjusted easily, it can hardly fail in any other particular, to contribute largely to economical steam generation. In making experiments in which it was thought that a variable exhaust would be advantageous, the inventors adopted an apparatus then much approved, but found that it acted as a brake upon the engine when the orifice was contracted; and when open, it diffused the jet to such a degree as to render the exhaust nearly useless for the purpose of stimulating the fire. In reasoning upon these defects, they were led to consider the experiments on the discharge of elastic fluids made by Dulong and Arago, and D'Aubuisson. From the results of these experiments they concluded that a conical adjutage of the form represented in fig. 4, the inclination of the sides varying between 7~ and 18", would produce the best effects; that is, would offer the least resistance to the piston, and give the strongest draft. In' constructing an orifice that might be varied within these limits without sensibly deviating from the internal shape required, at its different degrees of opening, and after trying various devices which resulted in making an orifice more or less elliptical, they succeeded ill obtaining one which satisfies all the conditions of the problem, giving a conical ad* See the chapter onr the Smoke-box, Blast-pipe, etc. 112 THE ECONOMICAL GENERATION OF STEAM. jutage free from obstruction, and capable of being contracted to any extent required between the limits of 7" and 18~. The coefficient of discharge with cylindrical orifices, by the experiments referred to, was.926, the theoretical discharge being 1.000. The coefficient of -discharge from the conical adjutage adopted by the inventors, varies from.938 to.917. It therefore appears that the principles they work upon are correct. The custom of many roads has been to buy heavy engines for light work —perhaps in expectation of increase of business; but there has also been a view to get machines that could never fail to make their time. There is prudence in this; but it would be better far to adopt. an appa. ratus that would in a moment fit an engine of economical size and weight, to the temporary work that may fall upon it. And that this msay be done by the blast-pipe is evident from the practice of the early engines; which, by contracted orifices, drove their fires so as to burn 167 Ibs. of coke per foot of grate per hour. This was, of course, wasteful of fuel, and involved excessive backpressure; but it made the small engines do the work imposed upon them: and the practice indicates the course that ought to be adopted in view of variable work, of snow-storms, extra loads, poor lots of fuel, and other accidents. If light engines are to be used, then, as they should be, they ought to have the means of doing more than their ordinary work, in cases of such emergency. At the same time it is of paramount importance that the blast-orifice should, during ordinary work, be as large as possible; and when doing very light work it should be so wide as to offer no sensible resistance to the piston. There is no question but that a great number of engines work with orifices so small as to cause a great loss by back-pressure, merely for the sake of avoiding delay in cases that occupy but a small portion of their working time. If they had the desired apparatus, instead of being permanently fitted to overcome occasional difficulties, they would work with less fuel, and less wear of boiler. There are other variable exhausts in use, which, with all their imperfections, are of considerable value; but their description here would obviously be of no special advantage. 2. Movable Cihimneys, etc. —There are other methods of varying the draft, which may be referred to. In the smoke-arch of the Rogers engines (Plate 46) there is a door, N (Figs. 1 and 2), through which, when opened from the foot-plate, the blast draws a greater part of the air from the external atmosphere, to fill the vacuum. Such a door, well fitted, so as not to be always leaky, would be of advantage even with the variable exhaust, since the latter, when widest open, draws much air through the ash-pan. Fig. 3 of plate 58, shows the movable inside chimney employed on the Hudson River line. It is enlarged at the ends to the shape of what is known in hydraulics as a vena contractca, and is kept in its central position by means of the 4 brackets a a, which are riveted to its sides. The sizes shown in the engraving are employed with a 16g-inch cylinder. The top of the blast-pipe is from 8 to 12 inches above the bottom row of flues. By moving the inner chimney upward, the draft is decreased, and vice versa. The movement is effected from the foot-plate, by means of a cross shaft and arms. On the Mine Hill & Schuylkill. Haven line, an apparently similar arrangement is employed. It consists of a short straight pipe, permanently fastened to the top of the heating chamber (Fig. 10, Plate 60), and another of equal length, which slides over it. A light iron stand as high as it is desirable to raise the outside pipe is placed on one side, with grooved pulleys at the top and bottom; a small wire cord is made fast to the bottom of the outside pipe, and passed over the upper pulley and under the lower one, and thence back to a small winch placed on the end of the boiler, convenient to the fireman. On the side of the pipe opposite the cord attachment, is a friction roller, to prevent the pipe from jamming while being raised and lowered. When the blast needs increasing, the fireman raises the pipe by turning the winch, and fastens it by a pawl to prevent it from sliding down again, till the coal is thoroughly ignited: When sufficient steam is raised and the fire burns clearly, the winch is loosened, and the stack lowered by its weight to its former position. It will be seen that the stack may be secured at any point, between the lowest and highest elevation, so as to secure a regular draught adapted to the requirements of the engine. This arrangement possesses some advantages over the usual method of increasing and diminishing the orifice of the exhaust-pipe to regulate the blast, because it creates less back-pressure. MEANS OF APPLYING THE PRINCIPLES OF COMBUSTION. 113 It will be observed, however, that this pipe is raised to increase the blast, while that on the Hudson River engines is raised to decrease it. The effects are obviously produced by different causes. The Mine Hill pipe is merely lengthened or shortened without disturbing the internal arrangements of the smoke-box. The total height of the Hudson River pipe, from the top of the flues, is not increased, but its depth is varied, and it evidently equalizes the draft through the whole of the flues when it is moved downward, showing that the draft is unequal when it is up; and it is a practical demonstration of the advantage of the diaphragm and petticoat pipes which are now generally employed in American engines. 5. THE STEAM-JET.-The steam-jet in the chimney, for creating draft when no steam is passing through the cylinders, is obviously an important feature, especially in coal-burning engines. It enables the engine man to start a fire quickly and without wasting much fuel in smoke, and to take advantage of stops and of " shutting off" while running down grades, to get a high pressure in boilers which do not make steam freely, and to prevent smoke by drawing in air above the fire. With Clark's jet above the fire it is not indispensable, but it is useful, for reasons already explained. The jet has been found to produce 2 horse power from a 100-horse boiler, in mine-ventilation; at what pressure is not stated. The power expended in producing draft by the blast-pipe is enormous, and were it not waste steam that is used, would be too extravagant to be thought of for an instant. An engine of moderate size vaporises 100 cubic feet per hour; and, when doing heavy work, releases the steam at 70 lbs. pressure-an ordinary pressure in stationary boilers. By the old rule this. exhausted steam contains 100 horse power, and in recent practice, it gives more than 200 horse power, which is expended in blowing the fire. These considerations are unfavorable to the use of the jet as a substitute for the variable exhaust, as has been proposed, even if it will work well with the exhaust. The author has not been able to learn that the steam-pressure is increased in ordinary practice, by the combined jet and blast; the reverse is the case with some engines; yet there is nothing theoretically opposed to maintaining a high pressure by the aid of both, in emergencies when the blast alone is insufficient. The jet-pipe should have a nozzle so small that the stop-cock can be fully open when a strong draft is required, so as to keep a high pressure at the end. Several nozzles are found to be better than one. Mr. Hudson makes the saddle of the chimney hollow, and perforates or slots its inner edge, as shown at M, fig. 1, plate 46. The steam thus goes out in an annular jet as large as the chimney, and is a better air-pump piston wherewith to exhaust the smoke-box, than the solid jet. Mr. D. K. Clark has employed a jet similar in principle: the end of the steam-pipe is bent to a ring a foot in diameter, which encircles the blast-pipe. The top of the ring is perforated. Mr. Hudson had before employed a similar apparatus-a hollow cast-iron ring. The steam-jet in the chimney was, used on the Liverpool & Manchester Railway, by Gray & Chanter, in 1837. 6. CONCLUSIONS.-Before reviewing the locomotive practice, it may be well to mention what has been done in stationary practice,* for the purpose of having all the necessary data at hand. Mr. Prideaux uses a self-acting apparatus like a Venetian blind for admitting air above the fuel, which is opened when fresh coal is supplied, and which gradually closes as the gas of the fresh fuel is exhausted. Mr. Gorman opens and closes the front of the ash-pit and the air-holes in the front of the furnaces, alternately, so that the combustion of the gas from the fresh fuel and of the coke left after its expulsion, takes place alternately. Juckes supplies fuel at a uniform rate by mechanism. In the apparatus known as the Systeme Beaufumb, a partial combustion of the fuel is effected in a furnace surrounded by a water-chamber, and supplied by a fan with just enough of air to form carbonic oxide with the whole of the free carbon, and to volatilize the whole of the hydro-carbons, so that the whole of the fuel is gasified except the ash. The mixture of carbonic acid and hydro-carbon gases thus produced is conducted by a pipe into a combustion-chamber, where, by the introduction of jets of air of sufficient volume, it is completely burned. If smoke is mixed with carbonic acid gas at a red heat, the solid carbonaceous particles are dissolved in the * The following examples are taken from Professor Rankine's Steam Engine and other prime movers. 15 114 THE ECONOMICAL GENERATION OF STEAM. gas, and carbonic oxide is produced. This is the mode of operation of contrivances for destroying smoke by keeping it at a high temperature, without providing a sufficient supply of air; and the result is a waste instead of a saving of fuel. Dr. Marsh supplies the whole of the air for burning the coke as well as the gas, by jets directed downward from above. Such being the stationary boiler practice, in cases where it differs from the locomotive practice, we shall have observed nothing in the former that can be simply and effectively adapted to the latter. The most simple, effective, and thorough means of mixing the air and the gas at any point, are Clark's steam-inducted jets of air. And these not only insure a mixture but they force it to occur at the point where it cannot fail to be perpetually ignited. The first cost of the entire apparatus will not exceed $50; and it is sometimes applied for half this sum: its repairs are literally nothing. And this is all that is required, except judicious management of the fire, to adapt wood or coke-burning engines to the burning of bituminous coal. In new engines, the combustionchamber would doubtless be worth its extra cost in insuring room and time for perfect combustion in case the management of the fire and the jets should not be skilful. It is also of advantage in shortening the flues, thus increasing their durability and promoting better transmission of heat. Clark's steam-inducted air-jets, the chamber in new engines, Hudson & Allen's or some similar movable grate, with sloping dead plates around it, and Parrott & Head's, Addison's, or Patrick's variable exhaust, appear, then, to be the best, and at the same time a very simple combination of parts, for perfecting the combustion of bituminous coal in locomotive boilers. PART III. HEAT, WATER, AND STEAM. CHAPTER I. THE TRANSMISSION OF HEAT. 1. EXPERIMENTS AND PRACTICE. —Tredgold and others have made experiments to ascertain the rate at which heat is transferred from metal to gases and from gases to metal. Other things being equal, it has been found that the rate of transference is as the difference of temperature. But in practice the conditions are different from those in the experiment; generally in the experiments the air has been still, and the gases moving under natural draft; but in locomotive practice the velocity of the gases is so great as to render the results of most experiments inapplicable. The effect of a fan is a sensible proof that the motion of air increases its power of taking up heat. Craddock found that a tube filled with hot water when whirled around at 40 miles per hour, cooled in a twelfth of the time required when it stood still. But in these cases more air is brought into contact; they, therefore, do not prove any thing respecting the transference of heat, to or from a given amount of air or gas. Tredgold states that he vaporized at the rate of 42 Ibs. per foot per hour with a Killingworth locomotive, using a blower. These cases indicate that there may be something in the effect of hard blowing, whether of cold air to take off heat, or of hot air to impart heat. Overman is of opinion that when hot gas blows against a surface at an angle of about 45~ it imparts heat much faster than when it sweeps along the surface of a flue; and that, therefore, in vertical water-tube boilers, in which the smoke current is at right angles with the tubes, the latter will take up heat faster than in the flue-tube boilers. Others, who prefer the multi-flue boiler, on account of its accessibility for repairs, use deflectors, which will cause the smoke to impinge against the flues, to some extent, and also mix it so that all the particles shall come in contact with the surface, and part with their heat. These will be noticed farther on. 2. MOVEMENT OF AIR AND GASE;S.-~In a common locomotive furnace the draft is stronger through the front than it is through the back part of the grate. This has been alluded to in a preceding chapter. It has been found that in a fire-box divided by a mid-feather, about thlreefourths of the fuel were burned in the front compartment. Whether this difference is owing to a shallow ash-pan, or to the greater proximity of the front compartment to the flues, or in some measure to each of these causes, is a matter of opinion; it is in accordance with the opinion of many able engineers to ascribe it in a great measure to the proximity of the flues. The partial vacuum produced by the blast is most decided at or near the flue entrances; and this is ordinarily just over the front part of the fire. Shortly after its entrance into the flues the gas* has a straightforward motion; and it is gene* The term gas is used in this connection to represent carbonic acid gas and steam the products of good combustion, and the hot air and nitrogen which pass through unchanged. Smoke, referring to bad combustion, is not the proper term. 116 THE ECONOMICAL GENERATION OF STEAM. rally believed that it flows so that only its outer surface touches the flues, and its centre does not touch; and so that its centre remains hot, and only its outer surface parts with heat. In the smoke-box, when the blast-pipe is high up, it is generally supposed that the greatest vacuum is at the middle of the upper row of flues, and that it diminishes from this point, so that the draft through the lower and outer rows is less than it is through the upper and middle rows. Petticoat pipes and diaphragms are used to equalize the draft. In the chimney the gas is driven by the steam, either by a continuous jet, when standing, or by intermittent jets when running. If the chimney is too large, the effect of the jet is impaired; if too small, the power necessary to expel the gas is considerable. There is, therefore, a certain size which produces the greatest effect for a given engine. The brick arches and water-bridges before considered, which tend to make a combustionchamber of the top and back of the fire-box, tend also to equalize the draft through the grate; and with a combustion-chamlber they should produce a nearly equal draft. They also tend to throw"' the gases against the back part of the fire-box, and give them. a longer run, and an additional reverberation, which is favorable not only to mixture and combustion, but to the delivery of heat. Jets of air in the front of the fire-box, if numerous, and forced in by steam-jets or a blower, would in part have the effect of the arch or water-shelf-they would turn the gas current backward, as well as jet air through it. And if inclined so as to blow upon the back part of the coals, and burn the solid carbon there, they would in part compensate for the weakness of draft in this region. In the fire-box we have to think chiefly of combustion, as it is effected by mixture of the air and gas; there will be no lack of heat in all parts of it, even though we neglect the reverberation of the currents; but when the gases are thoroughly inflamed, and have rubbed their outside surfaces against the walls of the fire-box, it is well to consider how we may turn the hot parts against the plates of the combustion-chamber. We find baffle-plates of various forms, for this purpose; some place them alternately at top and bottom, so that the hot gas flows downward, then upward, and so on. Others put alternate rings and discs-the discs force the gases toward the water-surfaces in all directions, and cause eddies. The effect of such arrangement, in large flues and combustion-chambers, has been good. 3. SMOKE-BOX ARRANGIEMENTS FOR EQUALIZING THE DRAFT. 1. Petticoat Pipes. —These have been employed under such a variety of circumstances, and with such various results, that no general rule can be laid down as to their proportions. The size and shape of the smoke-box, the size, openings, and position of the pipe itself, the size and height of the chimney, the position of the blast orifice, and other considerations, must greatly affect the result. And when we consider that these devices have hardly been applied similarly in two cases, and that the fuel consumed, the load drawn, and the skill and care of, the engine-man and fireman, before and after their application, can only be guessed at for each particular engine, it seems almost hopeless, in default of well-conducted and special experiments, to arrive at a mathematical conclusion. The great fact, however, that these appliances invariably save fuel, is sufficient for our present purpose, which, as was remarked in the introduction, is chiefly to call attention to the possible economy to be instituted after combustion is perfected. Several varieties are illustrated. The one used in Baldwin's engines, plate 51, gives very good results; a more common form is say 3 or 4 sections of pipe like that shown in Eaton's engines, plates 45 and 45a; the whole of the pipe thus formed being within the smoke-box. Marks' spark arrester and draft apparatus, shown with Eaton's engine, is a continuous petticoat pipe, extending from the exhaust nozzle to the top of the chimney. The lower sections of the pipe equalize the draft through the lower flues, and the upper sections equalize the draft through the wire netting which surrounds the whole of the pipe. The chief object is to give area of escape for the gas, through the meshes of the netting, and at the same time to prevent the exit of sparks. This arrangement, compared with the common American bonnet pipe, has effected a wonderful reduction in the consumption of fuel —not less than 20 per cent.; the amount cannot be exactly stated. 2. D)iap~raymss. Mr. Hudson and others have found that a plate (D Figs. 1 and 2, Plate 46), THE TRANSMISSION OF HEAT. 117 with adjustable slides E, standing behind the blast-pipe, in the smoke-arch, may be made to equalize the draft better than the petticoat pipe. The plan shown (Plate 46), has been patented by Mr. Hudson, and is generally placed in the Rogers engines. The current through the lower flues must pass under the plate, and that through side flues may be equalized by properly adjusting the slides E, while that through the upper middle flues, which would naturally be strongest, has the modified influence of the exhaust. It is evident that this simple plate may be so adjusted in any engine, as to absolutely equalize the draft through all the flues. Another plan has been patented by Wood & Winans, and is employed on the New Jersey and various railways. It is a single plate, and resembles in shape the superheater shown at fig. 13, plate 58. 3. Flues of diferent Sizes.-No experiments of much importance, as far as the author is aware, have been tried in this direction, although flues through which the draft was most intense, have been thimbled with good results. But the solution of the problem appears to lie in suiting the sizes of the flues to the vacuum made in each by the blast, for three reasons: 1st. All practice shows that the movement of fluids through large tubes is faster than that through small tubes, with a given power-hence varying the sizes of flues wouldproduce the result desired. 2d. The petticoat pipe, and especially the diaphragm, cause eddies and considerable friction in the current of gases, and therefore impede their progress and require a smaller blast-orifice. But were flues suited to the draft, the gases would enter the chimney with a gentle curve which would not much impede them. And again, the total flue area would be increased, thus decreasing the total friction. 3d. This would better facilitate the circulation of water among the flues, than any other means. It is a general belief among enlightened practitioners that more water-space mnst be left between the upper flues where the steam from all the rest of the flues and that generated at this point, and also the supply of water to make this steam, must pass. This subject will be farther considered under its appropriate head. Now, if the lower and outer flues are placed quite as near together as at present, and are of large diameter, the draft through them will be increased, and their proximity will not interfere with circulation, since very little steam is generated below them. And if the middle and upper flues are made smaller and placed farther apart, the draft through them will be retarded and a very large water-space will be left for the passage of steam and water. That is to say, the draft through all the flues will be equalized, and the same total amount of heating surface may be preserved. It should be remarked, that a wide water-space should be left in any case, between the flues and the shell for the down-flow of water, and between the middle vertical rows of flues for the up-flow. 4. CURRENT-DISTURBERS IN FLUES.-~Williams, McConnell, Duncan & Gwynne, and others, have introduced spirals, like screw augurs, some with short turns, others with long turns, also rings open on one side like horse-shoes, to disturb the gas currents, and mix the hot and cool parts. (Figs. 6 to 9 and 14, 15, Plate 66.) Their- idea is that currents touch only at the surfaces, and that the portion which starts in the middle continues in the middle, and does not give off its heat; and it is to break up this rectilinear movement that the disturbers are introduced. Mr. Dougherty, of the Camden & Amboy Railway, has tried spirals 15 inches long, with 11 turns, set 1 foot from the fire-box ends: a very palpable and obvious saving in fuel is reported, but the results have not yet been put in the proper authenticated shape for publication. The current-disturbers produce, on a small scale, the effects of the larger arrangements in the fire-box and combustion-chamber. Some who have tried them have not continued their use. Probably they do better in large than in small flues, as they tend to weaken the draft. Their effect is usually attributed to the motion they give to the smoke. But it is evident that they must receive and radiate heat; and that this-radiated heat must be received by the heating surface. A trial of flat strips, of considerable length, even the whole length of the flues, would, probably, be found to have a sensible influence; and, perhaps, would be useful in cases where the flue area is so ample that the friction of the smoke against the extra surface would not countervail their effect. This is the view entertained by H. Conybeare, who, in 1858, patented "heat receivers," the principal 118 THE ECONOMICAL GENERATION OF STEAM. mass of which was in the centre of the tube, and held in place by three or more feathers like the feathers of an arrow. McConnell's disturbers have a similar shaft or stem; but have the addition of obstructors so shaped as to produce eddies in the current, and thus to make the gas impinge against the.tube-surface, and to stir it so that all its particles, and not merely its outer ones, shall touch the heating surface. In marine boilers conducting-pins, 3 inches long and 1 diameter, set about 4 inches apart, have been used, and were deemed advantageous. They were not intended as current-disturbers; but they served that purpose to some extent: they were intended to receive heat, and conduct it into the boiler; in some cases they projected 2 or 3 inches into the water. They have gone out of use, chiefly because they collected scale rapidly; they also interfered with cleaning, and were expensive, and it is probable that some leakage was caused by them. The new disturbers, which are cheap strips of hoop-iron loosely put in, are more likely to prove economical; and, although the evidence is not wholly in favor, we believe them worthy of attention. The credit of introducing these spirals is accorded to the English, and no doubt they are entitled to it, although Mr. D. B. Martin, Chief Engineer in the U. S. Navy, in 1853, proposed to introduce twisted plates into the tubes of marine boilers, to stir up and mix the gases so that all their particles should touch the surfaces. His proposal was rejected, on the ground that it would obstruct the draft. The friction of the gas in flues makes part of the resistance to the draft. The contraction of their aggregate area makes a'greater part of it. Ferules in the ends are still worse. In some cases two-thirds of the force of the blast has been required to get the gas through the flues. The total area of the flues of engines that work with large orifices is not less than T- of the heating surface; but the ferrules sometimes reduce it to 4-5: with iron and steel flues these obstructions are seldom used. Very short flues are made smaller and set closer; but it is questionable whether the gain in surface will compensate for the contraction entailed on the blast-orifice. 5. SMOKE-BOX, CHIMNEY, AND BLAST-PIPE.~This part of the engine is important in relation to the draft, and consequently to the transmission of heat as well as to combustion. Builders do not yet agree as to the best proportions; but there is a tendency in one direction. The smoke-box should be of moderate capacity, so that not too much of the force of the blast may be expended causing pulsation of the gases in it. Mr. Peacock found that partly filling up a large smoke-box enabled him to work with a wider blast-orifice. And it is the opinion of several builders that round smoke-boxes work better than those which are square-bottomed, and have more room in them. The chimney is now generally made smaller than formerly. A general rule has been, in this country, to make the chimney of the diameter of the cylinder, irrespective of all other considerations. Mr. A. F. Smith, of the Hudson River Railway, has tried experiments on an engine with cylinders 16 X 22, whose chimney was 16 inches. Inner chimneys of different sizes were tried; and 9.- inches was found to work best. The boiler has a large combustion-chamber, and about 800 feet of heating surface. The inner chimney was made to slide up and down; a low position was found best for it. (Fig. 3, Plate 58.) The cone at the bottom is on the principle of the vena contracta, which, in water orifices, delivers more than a straight pipe. Chimneys curved outward at the bottom are much in vogue, probably because they look well. A somewhat similar chimney is employed in the Rogers engines with good results, and is shown at fig. 2, plate 66. It is customary, in some engines wherein combustion is bad, to continue the old wood burning-practice of placing inverted cones, or rather pans, in the tops of chimneys to turn over the currents of smoke and solid coal, so that the latter shall not form a volcano, and set the adjacent towns, forests, and bridges on fire. The obstruction to draft caused by these dams, is very great. A better practice has remedied the defect to some extent. The chimney first employed by Mr. Griggs, of the Boston & Providence Railway, and now in considerable use, gives a very free passage for gas and smoke, and prevents throwing out sparks. It is shown by fig. 13, plate 66. The blast-pipe, in English engines, is single. In our coal burners it is often single. Double aind single pipes have been tried on the same engines; sometimes two, and sometimes one had the THE TRANSMISSION OF HEAT. 119 best effect. It seems that other parts affect the case, so that an engine fitted for a single orifice will not work so well with two, and vice versa. Opinions here are divided on this subject, although it is hardly possible that one jet in the middle is inferior to two, both out of centre, other things being in a normal condition. In fact, so much importance is attached to making a true piston of the jet, that the nozzle is set by a T square, on the engines of the Great Western Railway, of Canada; and with obviously good results. When two nozzles are used, it is deemed necessary that they should point to the centre of the chimney-top, and be as close together as practicable. The question of double and single orifices has also a relation to the shape of the chimney. Some builders make the chimney an inch or an inch and a half larger at the bottom than at the top. This shape seems suited to two pipes, both jetting to the centre of the top. We think it worthy to be considered whether the chimney for two pipes might not be round at the top and elliptical at the bottom, or rather formed of two half cylinders, whose axes coincide with the axes of the jets. The size of the exhaust-pipes has an influence on the blast. They are usually made very large, and contracted at the end. McConnell, on the contrary, has made them of moderate size, and united them in a nozzle wider than the pipes. They should be straight when practicable, and curve gently when they cannot be straight; because every abrupt turn diminishes the velocity of the exhaust, increases back-pressure, and weakens the blast. Experiments on water-pipes nave shown that when a given quantity is discharged in 90 seconds from a straight pipe, it will require 100 seconds from a well-curved pipe, and 140 seconds from a pipe with a right-angled elbow. The elasticity of steam probably makes a difference, but it may be a difference for the worse. Abrupt curves near the orifice are worse than when near the cylinder; because the steam has more velocity the farther it is from the cylinder. If a pipe is gently curved, smooth inside, of moderate size, free from strictures and enlargements, and not contracted at the orifice, the exhausted steam is at a high pressure at the beginning, and grows less as it flows, the velocity Increases, and the density diminishes. The momentum increases as the square of the velocity; and it is the momentum that gives efficiency to the blast. The high pressure at the beginning gives a certain velocity, and the expansion during the flow increases the velocity. If, on the other hand, it is unnecessarily large and crooked, and contracted at the end, the expansive force of the steam does not so much increase the velocity; it is little more than is due to the pressure at the orifice. From these considerations it follows that exhaust-pipes not much larger than the ports, smooth and truly curved, will admit of larger orifices than very large pipes, especially when they are ill-curved. To this it obviously may be added that, for single orifices, the small pipe with an uncontracted orifice is best, in that it does not cause back-pressure in the other cylinder, whereas it is shown by the indicator that large pipes, abruptly curved near their ends, and uniting in a narrow orifice, cause considerable back-pressure in each other's cylinders at midstroke. The distance of the orifice below the chimney influences the draft. Mr. Peacock found that a 41 orifice 18 inches below the chimney, made steam better than a 4- orifice 9 inches below the chimney. American builders generally place the orifice low; but bring down a petticoat pipe or diaphragm, before referred to, close to it. Malformations in this region sometimes arise from a union of old patterns with new devices; for example, cylinder patterns made when the blast-orifice was near the chimney have been used with a petticoat pipe, when the blast-orifice was near the bottom of the smoke-box. To accommodate this popition the exhaust-pipe has two zigzags, one of them close to the end. It would have required an expensive alteration of patterns to make a good job; and the petticoat pipe might soon go " out of fashion; " so the matter has been accommodated to save the patterns; but the effect must be considerably impaired by this economy. The finish of the blast-orifice is, by some, considered of importance, while others neglect it. Clark says it should be tapered as nearly to an edge as is consistent with durability, and gathered in gradually at the end; but we find instances of plates bolted on with little regard to thickness; and in some variable exhausts we see the end formed into a large mass. There is theoretically, an objection to this; the smoke cannot pass it easily; it is analogous to a ship with a full stern, which never sails well. But whether the practical effect is of much account when the orifice is placed low, is to be determined by trial. Engines that are good in other respects, may work well in spite of minor defects like this; but if a little can be saved by attention to them, it should not be neglected. CHAPTER II. SUPPLYING AND HEATING FEED-WATER. 1. IMPORTANCE OF AN INDEPENDENT FEED. —This should be evident to every one who is familiar with working locomotives. Independent pumps are opposed only by enginemen who have had trouble with inferior contrivances, or by " railroad men " who, having " got along " somehow, repel innovations. Up to the time of the introduction of Giffard's injector, donkey-pumps were an indispensable feature of most continental locomotives, and were quite common on English engines. Now, Giffard's injector is superseding both donkey-pumps and connected or cross-head pumps; and it is getting into use here, the Reading Railway, among others, having adopted it. However satisfactory may have been its performance for a couple of years or more, on hundreds of engines, there is a plenty of managers who will, no doubt, deprive themselves of the benefits of an independent feed as much longer, while they " try " one of them.* The duplex pump recently perfected by Mr. Henry R. Worthington, of New York, after 25 years' experience in successful hydraulic engineering, is a competitor of the Giffard injector. There could hardly be better means of feeding locomotive boilers, independently of the working of the engines, than at present exist. Some of the advantages of employing these means are as follows: The evaporation of water is not in proportion to the speed of piston, since the cut-off varies the former, in locomotives, to a very great extent. In case of trains whose weight is constantly changing by the addition or detachment of cars, and in case of all engines which are not worked by the most careful and experienced drivers, the feed is frequently varied; and since the delivery of cross-head pumps is not in proportion to the speed of piston, being proportionally less and very irregular at high speeds, the variation of water supply to a cross-head pump is not always certain to produce the same variation in the delivery of the pump. The result is as follows:-No definite dependence can be placed on the delivery of pumps whose speed of piston is varying from say 100 to 600 or more feet per minute; frequent accidents occur from their failure to deliver water enough, and hence the common practice is to carry an excess of water,which causes priming, and is very wasteful. But this is perhaps the least momentous argument in favor of the independent feed. Its great importance is the supply of water when the engine is not running. In case of a sudden detention of any sort, with a fire-box full of fuel, a good boiler will blow off steam till the water is so low that the fire must be drawn. And while engines are standing at stations, they may be generally observed to be vigorously blowing off and wasting fuel, the ash-pan damper being shut, especially if the chimney-jet or blow-pipe is in operation. A damper is neither tight enough nor quick enough in its results, to prevent rapid steam generation and waste while standing; but the copious injection of comparatively cool water (say 180 to 200 in case of the Giffard injector, or when heaters are employed) very soon checks the formation of steam. When an engine encounters a snow bank, the only reason that it ever abandons the encounter, is want of water in the boiler. Short runs are not sufficient to get a pump working, much less to keep up a supply in the boiler.:More time is spent under such circumstances in keeping the pumps clear of ice, and the water over the crown-sheet, than would be necessary without these annoyances to force a passage for * There are examples enough of such incredulity to justify the above remark; for instance: a NlSew England "railway man," in 1858, flatly contradicted the author's statement to him that bituminous coal was regularly burned in England without smoke, and did not wish to be imposed upon by any reports to that effect. SUPPLYING AND HEATING FEED-WATER. 121 the train. A considerable amount of time is employed by freight engines on American single track lines, while waiting at stations for passenger trains to pass, in hammering themselves to pieces by running forwards and backward, over switches, frogs, and bad side-tracks, to " pump up." It is often necessary to preserve the full fires and the ability to make the greatest amount of steam of which the boiler is capable. This requires an enormous waste at the safety-valves, when, for instance, the train stops at the foot of a grade. But with a heavy injection of cool water by the independent feed, the fire may be increased and the steam kept down, and, at starting, the feed may be for a few moments shut off altogether, so that the full power of the boiler may be brought to bear. In managing the draft, the fuel, and the cut-off, with reference to varying trains, loads, grades, curves, temperature, wind, and quality of fuel and water, the supply of water is the chief element of successful engine-driving. There are many times when the fire could be economized if a vigorous feed could at any moment be applied to check the undue formation of steam,vor if, for the purpose of keeping up the boiler pressure, a little tardiness of feed were sure to be compensated for by an increased supply of water at the required moment. As it is, with a heavy train, te pump must be kept at work, whether fuel is wasted or economized. In cold weather, a momentary stoppage in the flow of water allows a pump time to freeze. At every station and at every shutting off of the feed, this is the peril of the ordinary pump. The independent pump, or injector, is not obliged to stop long enough to freeze up at any station. It may be kept constantly at work, and, if exposed to the cold, may be kept in working order. But it need not be so exposed, as it may be located close to the boiler. However the ordinary pump may be connected to the engine, it is generally not only in the way of some other part, but it can rarely be protected from the cold. In northern countries, the failure of the pump is the cause of some of the worst accidents, delays, and breakages in the winter. How to keep it in working order is the grand aim and particular perplexity of the engineer for a third of the year, while 18 inches from its position on the engine, heat enough is wasted to keep it at the best working temperature. When connected to the cross-head, the pump brings a diagonal strain and uneven wear upon the cross-head and piston rod, thus materially increasing friction and repairs. It is a principle in mechanics that a certain fixed speed for machinery will accomplish the most work with the least power, and that an increased or decreased speed will require more power for a given effect. The ordinary locomotive pump, to a greater degree than any other machine, violates this rule. Its action is not regular. When most water is required, its movement is slow. When the greatest regularity of feed is necessary, its movement is so fast as to render its action uncertain, and it wastes a maximum power with the production of a minimum result. Low water is undoubtedly a frequent cause of explosions; at least engine-drivers think so, and every good engine-driver will testify to the fact that the chief cause of anxiety and often the cause of distress of mind in running a locomotive, is the uncertainty of the water level. This distress and anxiety of mind unfit a man for duty, and particularly for the difficult duties of the enginedriver, which demand unusual coolness and presence of mind. The great reason why the common pump is peculiarly and unnecessarily defective, is because its supply is in no way dependent upon, coincident with, or proportionate to the demand made upon it, but only dependent upon a function of the engine to which its action may not be necesary, viz., its locomotion. The independent feed, however, may be adjusted to the simple requirements of the boiler pressure. 2. TIIE GIrtARD INJECTOR.-This invention is illustrated by plate 62; fig. 8 is a full-sized section of the more essential part of the size employed for locomotives, as built by Messrs. Sharp, Stewart & Co., of Manchlester. Plate 63 shows an elevation and a longitudinal section of the injector as improved in detail and constructed by Messrs. William Sellers & Co., of Philadelphia; and the manner of applying it to the locomotive boiler. As the injector has not been introduced to any considerable extent in this country, an explanation of it, somewhat in detail, may not be inappropriate. It is simply this: —A jet of steam flows from the boiler through a conical nozzle, say W inch in diameter, in which nozzle it comes in contact with the feed-water. The momentum of the steam carries the water through the nozzle, across a space open to the 16 122 THE ECONOMICAL GENERATION OF STEAM. atmosphere, and into the end of another nozzle which communicates with the boiler, the steam being condensed and the water being thus warmed by it, before the stream issues from the first nozzle. Fig. 1, plate 62, is a vertical section of an injector of the dimension usually fixed upon locomotives, and fig. 2 is a transverse section; fig. 8, plate 62, is a full-sized vertical section of the conical steam and water-passages. Fig. 3 shows the application of the injector to stationary or locomotive boilers. The steam from the boiler is admitted through the pipe A furnished with a cock B, and passes down through the perforated cylinder or tube C, which is made conical at the bottom, the area of the aperture being regulated by the conical rod D adjusted by the screw and handle E. The jet of steam issuing from the orifice of the tube C encounters the feed-water in the chamber F, which enters from the feed-pipe G; the supply of feed-water is regulated by raising or lowering the tube C by means of the handle H and screw of quick pitch. The stream of feed-water propelled by the steam-jet issues from the upper orifice I, and passes into the mouth of the lower pipe K leading into the boiler, the intervening space L being open to the atmosphere, so that the stream of water can be seen through the sight-holes M at this part of its passage while the injector is at work. A check-valve is inserted at N, to prevent the return of the water from the boiler when the injector is not working. The overflow pipe O carries off any overflow occasioned in starting the injector to work, and the. sight-holes M are covered by a circular slide. In starting the injector to work, the handle H is first turned into the position suited to the pressure of steam in the boiler; this permits the access of water to the instrument, and regulates its admission. The steam-cock B is then opened, and the handle E turned slightly so as to elevate the screw-rod D, which admits a small quantity of steam to the conical opening I. A partial vacuum is thus produced in the chamber F by the rush of steam through the opening I, and the water flows into it. As soon as this happens, which can be observed at the overflow 0, the screwed rod D is gradually raised until the overflow ceases, thus giving full liberty to the steam to act upon the water at I and drive-it into the boiler through the pipe K and the valve N. The supposed relative movements of the steam and water when the injector is in perfect action are shown in figs. 9 and 10, plate 62. Fig. 9 shows the result of an excess of steam, and fig. 10 of an excess of water. In case of an excess of steam, the opening I is entirely filled by it and the water forced back through the pipe G, the steam escaping through the overflow 0. In case of an excess of water, the quantity of steam rushing from the orifice of the tube a is so small in proportion as not to permit of its reaching the orifice I before it becomes entirely condensed; its velocity is thus lost, and there is not power enough left to overcome the resistance of the pressure in the boiler, and the water escapes at the overflow 0. Some months since, Mr. Robinson, partner of Messrs. Sharp, Stewart & Co., and Mr. Cross, the engineer of the St. Helen's Railway, made experiments with injectors of various sizes, manufactured at Manchester, to ascertain what effect the temperature of the feed-water, the vibrations and concussions caused by the action of the break passing over switches and switching cars, would have upon the regularity of the water passing through the instrument. The general result ascertained by these experiments and by general practice since, is that the injector works at all steam pressures up to the maximum working pressure of the boiler, 110 Ilbs. per square inch; and draws water from the tender of any temperature up to 110~ Fahr.; and that neither the sudden application of the break, nor any shock produced in passing bad switches or in switching cars, interferes in any way with its efficient working. The only difficulty experienced is when the water in the tender has become hot and at the same time very low in level, under which circumstances conjoined the degree of vacuum capable of being produced in the water-chamber of the injector is not sufficient to lift the water to the height of 29 inches above the foot-plate: this inconvenience, however, is readily obviated by lowering, the injector so as to bring the water-entrance within a few inches of the level of the bottom of the tender. In using the injector no difficulty has been experienced in so regulating the openings for steam and water as to produce a constant and regular supply of any required quantity of water to the boiler without waste from the overflow pipe. Several sizes of injectors are made to suit different purposes: that generally adopted for loco SUPPLYING AND HEATING FEED-WATER. 123 motives measures 8 millimetres or 0'32 inch in the smallest diameter of the throat K. (Fig. 1, Plate 62.) This diameter is taken as the standard point of measure for the power of the instrument, and the designation of the dimensions of the French injectors in millimetres has been adhered to in England, from the convenience of the decimal system of measurement. The size of 8 millimetres' throat is called No. 8 injector; another size, No. 6, of 6 millimetres' throat being made for stationary boilers of ordinary size and pressure, and corresponding smaller ones, No. 4 and No. 2, for smaller boilers and for agricultural and portable engines: a larger size No. 10 of 10 millimetres' throat is designed for large locomotive and other boilers of great evaporating power. In the case of a goods engine, drawing a load of 24 wagons up a gradient of 1 in 96 about two miles long, on the Lancashire Railway, Mr. Lees has found that all the water given off by evaporation during the ascent could be easily replaced by a No. 8 injector, without reducing the steam-pressure 1 lb.; the initial pressure of 100 lbs. having been maintained during the whole time, and rising immediately the summit was attained. For the purpose of ascertaining the limits of the circumstances under which the injector can be worked, a series of experiments was made by Mr. Robinson, with instruments of different sizes fixed to stationary boilers working at 60 Ibs. pressure; one of which being connected with an adjoining boiler in which the pressure could be reduced to any desired amount, gave great facility for measuring the power of the injector when feeding by lower into higher pressures. The temperature of the feed water also could be varied at pleasure, by introducing into it either hot or cold water as required. The general reslts obtained from these experiments were, that water could be forced into a boiler by the injector when the steam-pressure was not below 5 lbs. per square inch; that the temperature of the feed-water might be raised up to 148~ Fahr., requiring to be varied in the inverse proportion to the pressure of steam; and that surplus power was developed by the instrument, available for forcing water into a boiler at a higher pressure than the one from which the steam was obtained, the injector having been effective with steam of 245 lbs. pressure above the atmosphere in forcing water into a boiler at 481 lbs. pressure. The particulars of these experiments are given in the following tables: in all cases the surface of the water in the supply tank was at least -2 feet below the level of the water-chamber of the injector, the vacuum in that chamber being from 1 lb. to 1 lb. below the atmosphere during the operation. TABLE I.-Maximum Temperature of Feed admissible at different Pressures of Steam. -Pressure of Steam, lbs. per square inch,.. 10 20 30 40 50 100 Temperature of Feed, Fahrenheit,... 148~ 138~ 130~ 124~ 120~ 110~ TABLE II. —Quantity of Water Delivered at different Pressures of Steam, wifth No. 4 Injector, the Temperature of the Feed-water being 95~ Fahr. Pressure of Steam, lbs. per square inch, 5 10 20 30 40 50 60 Water Delivered, gallons per hour,. 93 |124 150 186 210 244 259 As an addition to this experiment it may be stated that with a No. 8 injector, working at a pressure of 150 lbs. per square inch, as much water was passed through the throat of 8 millimetres or 0 32 inch diameter as was supplied to the tank by a pipe of 1l inch bore from'a cistern at least 20 feet higher than the tank. TABLE III. —Quantity of Water Delivered at different Temperatures of Feed, by a No. 8 Injector, the Steampressure being constant at 60 lbs. 2er square inch. Temperature of Feed, Fahrenheit,..... 60~ 90~ 105~ 120~ 130~ Water Delivered, gallons per hour,..... 972 786 698 486 382 124 THE ECONOMICAL GENERATION OF STEAM. TABLE IV. —Low-pressure Boiler supplying High-pressure Boiler: Maximum Difference of Pressure admissible. PRESSURE OF STEAM. I Temperature of Feed. Difference of Pressure. Result. Low-pressure. High-pressure. Lbs. Lbs. Lbs. i 74 50 59 9 40 55, 15- Slight overflow. 38 56, 18- Overflow increased. 86~ 36 57 21 I. Delivery almost ceased. 74~ 50 511 137 49 12 Slight overflow. 33 50- 172 Overflow increased. 96~ 30 521 221 Delivery almost ceased. 74~ 45 47 2 34 45, 11- Slight overflow. 31 451 141 Overflow increased. 106~ 24 481- 24L Delivery ceased. The theory according to which the injector is supposed to act, is thus stated by Mr. Robinson. The pressure on all parts of the interior of steam boilers being equal, some reason mustbe sought why steam taken from one part is able to overcome the resistance opposed to its entrance into another part of the same boiler. Looking at the construction of the instrument itself, it is evident that when it is in operation, and the valve N, fig. 1, open or removed, the pressure in the boiler acting through the pipe K appears free to resist the entrance of the water into it. It may be assumed that steam or water escaping from a boiler into the atmosphere, does so at a velocity proportioned to its pressure or density, as in diagram 6, plate 62. If a pipe conveying steam were turned directly back into the water of the same boiler, it is evident that equilibrium would ensue, and no effect be produced. If, on the other hand, a break were made in the continuity of the pipe so as to leave an interval open to the atmosphere, the steam would rush from one. pipe and the water from the other at velocities proportioned to their different densities. But in the construction of the injector the feed-water chamber F is placed at this break in the pipe, as shown in diagranms 5 and 7, and fig. 1; and this arrangement accounts for the power of the steam to overcome the resistance to its entrance into the receiving-pipe below: for the jet of steam being concentrated on the water at I forces its way through the its way thsurrounroded by the surrounded by thefeedwater, by contact with which it is gradually condensed and reduced in volume and velocity until it is entirely converted into water at the throat K; while, by this contact with the steam from I to K, the feed-water has a velocity imparted to it proportioned to the steam-pressure in the boiler and its own temperature, and being nearly non-elastic it thus acquires momentum sufficient to overcome the resistance of the water in the boiler. This explanation will, perhaps, also serve to account for steam of lower pressures being better adapted to act upon water at higher temperatures, since the lower the pressure of the steam the less must be its velocity in passing from the orifice I to the throat K, and consequently the more time will be given for its condensation, a condition necessary when the higher temperatures of feed-water are used. This explanation also agrees with the fact that the higher the steam-pressure the more rapid is the stream of feed-water into the boiler, since in passing from the orifice I to the throat K, at a higher velocity and at a higher temperature, a larger volume and a lower temperature of water are required to condense the steamjet; and its capacity for this water being in proportion to its temperature, the quantity of feed carried into the boiler is proportionately increased.The temperature of the water rises about 60" in passing the injector, by the condensation of the steam, as measured at the overflow. But a part of the heat of the steam is lost by expansion, * The above facts are taken substantially from Mr. Robinson's paper on the subject before the Institution of Mechanical Engineers. SUPPLYING AND HEATING FEED-WATER. 125 although the steam is condensed; and thus the instrument will not produce perpetual motion without a constant supply of power, as some seem to suppose. Practically, the injector and a good pump are believed to work with about equal economy, although an able French mathematician makes out the former to be rather the more economical. Its repairs, of course, will be hardly measurable. But, for locomotives, the cost of feeding the boiler by any means bears no important proportion to the value of an independent feed. Feed-water heaters will also be efficient, although the water is heated by fresh steam in the injector. The best feed-heaters-those which are most simple and durable, for complicated arrangements have invariably been abandoned~ raise the temperature to about the point at which the injector may take the' water with advantage. Thus far, there has been found to be no incrustation in the nozzles or other parts of the injector; and on the contrary, no wear in the nozzles has been observed. Either of these defects could be easily remedied, should they occur, by having separate nozzles to be applied from time to time, as occasion might require. The construction of movable cones sliding one within the other, adopted in the injector, effectually prevents any difficulty arising from small orifices, since the actual openings are large, though admitting of regulation down to the smallest size. Fears are expressed that the injector may not prove trustworthy; that the feed may sometimes stop; that sometimes it cannot be started; and that it cannot be adjusted to meet the various requirements of the locomotive boiler. The author has had some opportunities, here and abroad, of experimenting with the injector on this point; and the similar results of various reliable experiments have been published. If there is an excess of steam or of water, there will be no delivery into the boiler; nothing, however, can be more simple than to adjust the relative steam and water supply, and to watch the result, not by the precarious indications of pet-cocks, but by looking at the stream itself, as it flows from one orifice to the other, through the open air. A few seconds is ample time to adjust the supply of steam and water, and then the feed starts and continues for any length of time —while there is water in the tender; and so long as this adjustment is not altered, the delivery cannot be stopped except by placing an obstruction between the nozzles; when this is removed, the delivery instantly begins again. When the injector is quite cold, it will not start at once; but it may be warmed by blowing steam through it for a few moments. And when it is too hot, it will not start; but the water soon cools it. In normal practice, the injector begins to deliver within at least half a minute-generally less time —from the moment of adjusting the steam and water supply, To what extent the amount of delivery can be changed, it is impossible, as far as the author is aware, to state with exactness. The practice shows, however, that the feed may be varied sufficiently to meet the varying demand. When once started, the injector should probably be kept feeding to the end of the trip, except in case of very long stops. The injector has been used on locomotive and other boilers in France, for nearly two years, and in England, for nearly one year. Messrs. William Sellers & Co. are the sole agents here, and have been making special tools for the manufacture; and injectors are now furnished and guaranteed by them for use in this country. Nearly a year ago, injectors were ordered by the Northern Railway of France for 457 locomotives —two for each. Injectors are being applied,.probably as fast as Messrs. Sharp, Stewart & Co. can make them, to English locomotives. The Eastern Counties Railway Co. have recently ordered 30 new engines, which are to have one injector each, and no pump of any kind. In view of these facts, there can be no doubt of the reliability and usefulness of the injector for locomotives. 3. THE DoNKEY PUMP. — Probably the best steam pump now made, is the duplex pump by Worthington, before mentioned. It is shown on plate 68. Its internal arrangements are similar to those of the single pump by the same maker, which is probably better known than any other steam pump in this country, and is far superior to those ordinarily used abroad. The advantage of a double pump is obvious; the strain and wear are distributed over double the surface, the dead centres of, the single piston are passed without jarring, and the stemn-valve of one engine being moved by the direct action of the other while the latter is at half stroke, the motion is not only right as to time, but regular, and without the jerk and strain that often attend the movement of a tappet valve-motion. The advantages of the Cornish pumping engine over the fly-wheel engine, are too decided and too widely acknowledged to require discussion here. A fly-wheel pump requires the water to move at certain determinate but irregular velocities, in order that the reciprocating motion of the 126 THE ECONOMICAL GENERATION OF STEAM. piston may be converted into the regular rotary motion of the fly-wheel; the water, on the contrary, to be moved without great strain upon the machinery and waste of power, requires not only a uniform motion, but, so to speak, an elastic pressure, which may be modified by the water itself. The Worthington pump produces, to a great extent, the results of the Cornish engine. The direct elastic force of the steam, instead of the momentum of the fly-wheel, is the force which moves the water. The duplex pump will invariably start by the simple application of steam to the steam-cylinder, at whatever point the piston may have stopped. And it will invariably deliver when moving at any normal velocities. The boiler may be flooded when required, or the proper feed for the least possible steam generation may be supplied. Some of the advantages of an independent feed which have been mentioned, must obviously be due to an extreme variation in the amount of feed. The pump shown in detail on plate 61, is from the working drawings of Messrs. Beyer, Peacock & Co., of Manchester. This is the ordinary form of locomotive donkey pump employed in England and on the continent. It is used only as an auxiliary. As a job of mechanical construction, it is excellent. The auxiliary pump, by Worthington, shown on plate 67, is very simple and cheap; and as an auxiliary, is perhaps quite complete enough for all practical purposes. The steam-valve —a slide on a circular seat-is moved by hand, so that the fireman, without what can be called labor, performs the duties of an expensive valve motion while the steam-piston does all the work. This little machine costs about $50, applied to the engine. It is large enough if worked to its full power, to supply the boiler, in case the other means of feeding should fail. Those who purpose to continue the use of cross-head pumps, should not fail to employ it. This is said advisedly. The amount of steam wasted at the safety-valves of locomotives, especially while -standing, is enormous; an injection of cold or comparatively cool water, will stop this excessive steam-generation when dampers cannot stop it, for reasons already stated. The means of preserving the proper water level are of the highest importance. In snow-banks-which engines could work through well enough if they could only have water —the short runs back and forth are not long enough to get a cross-head pump in working order; so sometimes half a dozen engines at a time are rendered powerless; the boilers must often be emptied to prevent their freezing, and they are then filled at the safety-valve, when the storm is over, with the greatest difficulty. Or the fire is drawn, and the complement of water is put in at the safety-valve in cases where the boiler is not emptied. Half the troubles of winter running in northern climates would be avoided by the use of this simple auxiliary pump. The other advantages of an independent feed have been alluded to. There is a general and quite correct impression that a single pump or injector, however excellent, should not be relied on, when the water-supply is so important as it is in case of the locomotive. An accident may occur, which will disable the regular feeder; and although an auxiliary might rarely be put into service, it avoids all risks, and for this reason it is highly desirable. Now, instead of paying for an extra complete steam pump or injector, less than half the money will supply an auxiliary, and insure safety and economy, if operatives will perform their duty. 4. CONCLUSIONS AS TO INDEPENDENT FEED. —There was, perhaps, a tolerable reason for not employing an independent feed on locomotives, a year or two since. Very few pumps were adapted to the purposes required, and many experiments resulted in failure. Now the case is very different. The Giffard injector, tle Worthington duplex pump, the Worthington auxiliary pump, and other improved steam pumps —all of them well adapted to locomotive service, are in the market. It has been advisedly remarked, that the use of single-riveted longitudinal boilerseams, which have but half the strength of the plate, when the double welt joint possesses 80 per cent. of the strength of the plate, is absurd. Yet a very plausible argument can be constructed, in favor of the single-riveted joint. But what possible excuse can there be for wasting fuel in the wholesale fashion which invariably attends a dependent feed; for stalling engines in snowbanks; for allowing pumps to freeze; for hammering out frogs and side tracks, and cutting loose from trains, in " running to pump;" for carrying low steam in the fear that temporarily reducing SUPPLYING AND HEATING FEED-WATER. 127 the feed will necessitate drawing the fire; for burning flues and fire-boxes; and for suffering the never-ceasing troubles and expenses due to a water-supply which cannot be governed by the requirements of the boiler? 5. FEED HEATERS.-Feed-water heaters are much used for stationary boilers, and have been tried on many locomotives, with a saving stated at 20 per cent. of the fuel. A seventh of the heat is about all that can be saved directly by heating the feed to 212; but when we consider that this saving, or half of it, will allow of less violent firing, and a wider blast orifice, we may credit the accounts, if well attested. Most of the early heaters tried on locomotives were smoke-box heaters. They were generally unsuccessful, chiefly because they interrupted the draft by occupying room and creating eddies; and whether situated in the smoke-box, or forming an inner chimney, they soon became so coated by the action of alternate moisture and soot that they were practically non-conductors. In popular phrase, they would " sweat " whenever a fire was lighted in the boiler-that is to say, being full of cold water, they would condense the watery products of combustion, after which soot would adhere to them. Exhaust steam was employed to heat feed water by Sharp, Stewart & Co., and others in England, and quite extensively by Mr. Ebbert, then Loc. Supt. of the Galena & Chicago Railway. But instead of placing the heater in a convenient position, since the steam could be brought to it wherever situated, some have adhered to the old place and form-the smoke-box and the small water-tubes-which have many disadvantages. Feed-heating is accomplished like condensation, in two essentially different ways-surface heating, and heating by injection. Whenever heating by injection-that is, letting the exhaust steam into the feed-water-is mentioned, in this country, we are immediately informed that the tallow passing from the cylinders with the steam, back into the boiler, will cause priming. If an excess of tallow is used, this may be true;-it is interesting to know, however, that several varieties of injection-heaters have been employed for years with great success in England. They occupy less room, weigh less, and requiring very little surface, since they are merely a box where steam and water can meet, they cost less than the surface heaters, in which the steam can only impart its heat to the water through an extended surface of sheet metal. And it is probable that, in any case, the judicious use of tallow will not cause priming. Steam is a good lubricator, and enough to fill the cylinder when running down inclines, is less expensive than tallow. The most simple and compact heater that has appeared, and, obviously, the most effective, for its cost and dimensions, is that by D. K. Clark, of -which three forms are illustrated on plate 60. The principle is similar to that of the steam-jet for coal burning-the forcible and immediate intermixture of currents or jets of water and steam brought into direct contact, and travelling together. One or more jets of steam are discharged freely and directly through a pipe or other passage, or chamber of suitable form, into which also the water to be heated is delivered, and through which it is passed in conjunction with the steam. In this confined passage, the steam, in virtue of its initial velocity, forcibly impinges upon, disperses and mixes with the water, and is quickly condensed, and the water is raised in temperature by the heat of the condensed steam. The jets of steam should be so adjusted, as by suction to draw and conduct the water into and through the heating-chamber, after the manner of the blast-pipe. When more than one jet of steam is used, the jets may be placed side by side, circularly or otherwise, or they may be placed in succession, heating the water by stages. The water may be delivered round or within the jet of steam, and may be introduced in successive stages in one heating-chamber. The steam-nozzle' may be round or it may be annular, square, oblong, or otherwise formed, exposing more extended surfaces of contact than when circular, for the mixture of the steam and the water to be heated. A free vent may be provided for uncondensed steam. Syphon or drip-pipes may be fitted to the heater, to discharge surplus water. In another form of heater, the steam and water are discharged upon a perforated surface, and the steam mixes with, and drives the cold water through the apertures. In surface water-heaters, the steam is caused to advance helically or by alternate deflection, or 128 THE ECONOMICAL GENERATION OF STEAM. otherwise to impinge upon the heating surfaces. The heater may be cylindrical, rectangular, or otherwise formed. The water surrounding it may be caused to circulate against the heating surface by ribs, or other contrivances, with or without the aid of pumps. The steam space may be annular, with internal and external heating surfaces. The heater may be subdivided into a number of pipes or compartments. One of Mr. Clark's plans has been employed on a stationary boiler in the North London shops, with such good results that it is about to be applied to locomotives. There can hardly be a doubt of its success. The whole apparatus costs about $50, so that any one can afford to risk a failure by adopting it. Beattie's feed-heater, as formerly employed, is shown in detail on plate 38. It did not prove entirely satisfactory; and the one shown on plate 59 has been substituted-con a considerable number of the engines of the London & South Western Railway. This apparatus consists of two iron annular tubes, the smaller one passing inside the larger. There are two cast-iron ends, A, A, brazed on the ends of the large pipes, C, C. Care must be taken to have the ends of the large pipes turned bright, and the casting cleanly bored, to insure the brazing metal taking hold. A stuffing-box is formed in the end of the casting, and a brass nut or gland screwed into it: the packing consists of one or more india-rubber rings made the same size as the pipe; and an iron ring is put next the india-rubber for the nut to turn upon, otherwise it would drag the india-rubber. The small pipe at the smoke-box end passes into a stuffing-box in the elbow B; a wrought-iron ring is brazed on to the end for the packing to press against; and a wrought or cast-iron flange is screwed on to the end next the fire-box, to form the connection with pipe D, which passes on down through the foot plate, and communicates with the condenser E. The condenser is of cast-iron, as shown; F is a copper sieve, about three inches deep, perforated with holes 13 diameter, to allow the water to fall in a shower. G is a throttle valve, to vlhich a float H is attached; this regulates the quantity of water in the condenser, so that only as much as the pumps use is admitted, thereby keeping a clear space for the exhaust steam to be admitted. The water is admitted into the condenser by means of the cock K; should any thing cause the disarrangement of the throttle valves, so as to shut out the water, an arrangement is made at the bottom, so that by opening the cock, L, the water may be admitted both into the condenser and feed-pipes. The course of the water is from the tender through the centre pipe into the sieve, where it falls into the condenser; the feed-pipes, M M, then convey it to the pumps; the left-lhand pump forces the water through a two-way cock, 0, in the left-hand side of the boiler, where it passes upwards into the side apparatus, and crosses over the boiler through the pipe, T, and along the apparatus on the right-hand side into the check-valve, W. The exhaust steam from the blastpipe passes through the internal pipe of the side apparatus on the left-hand side of engine, and down the copper pipe, D, into the condenser under the foot-plate, and is regulated by a cock, P P, and through the internal pipe of the side apparatus on the right-hand side into the tender by means of the cast-iron pipe, N, and is regulated in like manner by the cock, P P. Should the water in the condenser become too hot for pumping, the cock L should be immediately opened to admit cold water from the tender. This apparatus is unnecessarily complicated. The same results are more simply produced by Clark's system. Beattie's heater, however, is found to be valuable, and, as was shown in referring to his coal-burning boiler, quite indispensable. Both the water in the condenser for immediate use, and that in the tender, are heated. Many impurities settle in the tender if the water is heated there. In the injection heater, the water is not under boiler pressure, as it usually is in case of the surface heater, so that the apparatus lasts longer, and a little leakage does less damage. But the water cannot be so highly heated as it may be in the surface-heater, since if its temperature is above 212~, the formation of steam prevents the pump from operating successfuilly. It is not probable, however, that steam can be generally spared to heat the water above this point. Messrs. Sharp, Stewart & Co. have used Beattie's heater with success. Messrs. Beyer & Peacock have found injection-heaters advantageous; and generally, it is observed, that;o hurtful impurities SUPPLYING AND HEATING FEED-WATER. 129 from the cylinders are conveyed to the feed-water. Where much tallow is employed, it is customary, on some lines; to stop the entrance of exhaust steam to the condenser, for ten minutes after the tallow has been applied to the cylinders; after which, no tallow passes into the heater. In any case, this temporary remedy would render the plan of injection entirely feasible. Surface-heaters are of two kinds. The water is forced through pipes in contact with steam, or it passes slowly through a large vessel while the steam flows through small internal tubes or flues. In most cases, the water is heated after it passes the pump. The Giffard injector is therefore better applied in connection with the surface-heater, since the injector does not work well with very warm water, as was shown. The first plan of surface-heater, mentioned, is exemplified by Ebbert's, plate 60. The water passes from the pump-pipe, G, through the saddle of the chimney, into one of the nest of tubes D (shown in section at fig. 3), and out at E F into the boiler. The concentric space between the inner and outer chimneys is filled with exhaust steam by means of the pipe d, the steam being regulated by the valve e covering the pipe h, which is cast with the saddle. Fresh steam from the boiler may be admitted at n, which is better than wasting it at the safety-valve. The condensed water is drawn off by the pipe n. This apparatus certainly heats the water to 212~, and effects a considerable saving of fuel; but its repairs have been so excessive that it has been abandoned on several lines. Its chief fault is, that it does not allow the impurities of the water to settle. Purifying the water should be a considerable advantage in feedheating, and the friction of the water through the tubes brings an excessive strain on the pump. Wilder's heater, fig. 10, plate 60, is employed on the Mine Hill & Schuylkill Haven Railway. It is prevented from overheating by passing water through it continually, the waste going back into the tender. The second form of surface-heater is illustrated by Hudson's, plate 46, and by Eaton's, plates 60 and 451. Hudson's is an improvement on a plan which has been more or less used for many years. It consists of a cylinder filled with small flues, F, the feed-water being pumped slowly between them, and passing from one end to the other; during its passage, many of its impurities are precipitated, and are readily withdrawn. The exhaust steam, regulated by a cock in the pipe, H, passes directly from the exhaust ports of the cylinders, through the flue-tubes, and if condensed, the water from it may run out at the pipe, L, or it may be conveyed into the ash-pan. If not condensed, the steam may pass through the pipe, G, to the chimney top, or it may be confined under pressure by the cock in the pipe G. The results are satisfactory —the exact saving of fuel and cost of repairs have not been ascertained by experiment. The position of this heater is favorable; its weight is low, and it is not in the way of other parts. It cleanses the feed-water more thoroughly than any other plan which the author has observed. Several plans, similar in general arrangement, have recently been patented in England. Eaton's heater, shown in detail by figs. 4, 5 and 6, plate 60, and as applied by plate 45, is an extremely simple apparatus. It is formed of an outer and inner pipe, J and K-the inner pipe K being connected to the exhaust pipe by tile pipes, L and M. The waste water and steam from the inner pipe, K, are carried off by pipe N. From motives of economy the outer tube, J, is cast in one piece together with the side pipe, 0, 0, for conveying the cold water from pump P to the outer extremity of pipe J. The cold feed-water is forced by the action of the pump through the annular space between the inner and outer pipes, and during its progress to the check-valve, Q, is heated by the, exhaust steam contained in the pipe K. The inner tube, K, is a plain wrought-iron tube, 5 inches diameter, and the joints are made water-tight at each end, by stuffing-boxes packed with india-rubber rings. The delivery pipe, R, from the pump, is also connected to the feed-heater by the stuffing-box joint, fig. 6. It is preferred to carry the waste steam fromn pipe K, into a tubular apparatus fixed as low as possible inside the tender, but this has not yet been applied to Mir. Eaton's engines. The paramount object has been to reduce the apparatus to the cheapest form and fewest parts. The feed-water is heated to 180~. The cost of repairs is trifling. The first cost of the apparatus does not exceed $50. A feed-heater using exhaust steam, is sometimes employed to vary the blast; but it cannot perform its own functions properly while taking the place of a variable blast. Supposing the boiler pressure to be falling, the admission of cold water, by allowing all the exhaust steam to go into the 17 130 THE ECONOMICAL GENERATION OF STEAM. smoke-arch, will not so speedily remedy the difficulty as to heat the feed with a part of the exhaust, and to increase the velocity of the rest of it, by contracting the blast-orifice. Velocity, not quantity, is the obvious cause of an efficient blast. It has generally been observed that feed-heaters have been most successfully employed on lines managed by their inventors. That exaggerated statements, especially about the repairs of heaters, are sometimes made by interested parties, is very probable; it is equally probable, however, that the care with which these improvements are watched and followed by their best friends, has much to do with their success. Similar care on other lines, would bring forth the same desirable results. It is impossible to state the exact economical results of feed-heating —either the saving of fuel or the cost of repairs; because no experiments which fairly estimate all the conditions, have been made. It is quite sufficient, for present purposes, however, to know that there is a saving worth making, and it is very obvious that the cost of maintaining such heaters as Clark's and Eaton's, cannot materially detract from this economy. It would, therefore, be unreasonable to neglect this improvement any longer, on lines, at least, where fuel is expensive. CHAPTER III. V A POR I ZA T ION. Having treated of combustion, and of the transmission of heat and the supply of water to the boiler plates and flues, we next have to consider the imparting of the heat to the water. It has sometimes been supposed that little attention is required on this point; but the general belief of engineers now is, that the efficiency and durability of boilers depends very much on the proportions and arrangements which affect this part of the process of generating steam. 1. MAXIMUM RATE OF VAPORIZATION.-To find how fast an overheated boiler could make steam, when the water foamed over on starting the engine, or opening the safety-valve, Craddock heated a vessed to a dark-red heat, and observed the rate of vaporization while it was cooling down to 500". He found the rate for this range to be 270 Ibs. per hour per square foot of surface1 8 times greater than the average rate per foot in the best long-tube locomotives. He then observed the rate from 500~ down to 212~, and found it 101~ lbs. per hour. Richard Prosser reports that Church's boiler, with a fan blast, on a trial, vaporized 13 gallons —104 lbs.~per hour per foot of surface. Jacob Perkins. heated a very thick iron cup to a white heat, removed it from the fire, and poured in a measure of cold water, which evaporated in 90" —not touching the metal, but remaining in a spheroidal state. A second measure evaporated in less time, and so on, until the seventh evaporated in 6"0 and the eighth measure did not boil. There was, therefore, a temperature of maximum effect, which he believed to be about 80~ above the temperature of the steam for low press-. ure but as the pressure increased he thought the difference of temperature greater, that is, the most effective temperature of the surface in contact with water would be more than 80~ above that of the water. Perkins' experiment seems to disagree with Craddock's; the latter insisting that the temperature of greatest effect is much higher than Perkins supposed. C. W. Williams reports that the plates of the 4-inch water-spaces of the steamship Liverpool's boilers, between the furnaces, were blistered and cracked by the heat; and a tube was put in, with a try-cock on the end, to ascertain the conditiolln of the water in those spaces. It was found that when the fires were strong nothing but steam issued. Similar evidences have appeared in locomotives; the fire-box plates have bulged and burnt out in a few months, when the waterspaces were narrow, but not when they were wide. Also, when flues have been set close together there has been rapid deterioration, indicating overheating, and a spheroidal condition of the water. Water-spaces, which facilitate the access of solid water, prevent such deterioration of the metal, and probably also prevent the repulsion of water in the most rapid firing in locomotives. But it is said that in steamboat boilers, where the fan-blast is used, and the bottom of the boiler has been near the grate, and the fire has been foul except in spots, the firing at the spots has been so intense as to burn the boilers; and, it was presumed, to repel the water from the plates. From all the evidences it is concluded that the temperature of maximum effect is that which barely allows the contact of water; and that the excess of temperature of the metal over thewater is greater as the pressure is higher; and never less than 80~ above that of the water, and probably much higher than that under high-pressure. It also appears probable that this temperature is often and excessively increased in boilers with narrow water-spaces, especially if the pressure is 132 THE ECONOMICAL GENERATION OF STEAM. not kept up. In fact, there are cases which indicate that nothing but steam is present to take up heat in a great part of the spaces contiguous to the fire. From the little evidence we have on this point, it appears that steam takes up heat only.6 as fast as solid water of equal temperature. But as the steam is more or less superheated in such cases, it is likely that it is less than half as effective as water, even at the degree of solidity found in well-proportioned boilers. 2. WATER SPACES AND CIRCULATION.-The general result of all practice, as far as definitely reported, and of all experiments on the subject of narrow water-spaces, is, that they are disadvantageous-that they practically destroy the efficiency of more heating surface than they add to the boiler. Without repeating these results, it may be said, generally, that narrow water-legs burn out faster than wide water-legs. In anthracite burners, this is especially noticeable; the plates of 3-inch spaces often fail in a month; those of 4-inch spaces last at least a year, without perceptible deterioration. When there is burning of the metal there is also deficient generation of steam; burning is occasioned by dryness of surface, or the presence of steam instead of water. In some cases froth, and in others mere steam have issued from try-cocks in the lower part of boilers, in which the firing was strong, and steam has been found at the lower try-cock while there was water above. Experiments in open vessels have shown that the'water is driven out from tubes of three inches in diameter, that are open only at the top; it returns for a moment, steam is made; and it is again driven out, with violence and noise, and something like explosive action. In a considerable number of boilers in which the flues were from to -1-inch apart, new flue-plates have been put in, and from 15 to 25 per cent. less number of flues have been set from 1 to — inch apart, with a very evident, but, we regret to say, not definitely measured increase of steam generation, with a given amount of fuel. 3. MOVEMENT OF WATER AND STEAMI.-The usual movement of water and steam in a boiler with 22-inch legs, and tubes set L-inch apart, and as close to the barrel as they sometimes are, is probably much as follows:-When running, there is so much steam mixed with the water as to raise the level about 5 inches, Next the water-shell of the fire-box and barrel the water is not solid, but is less foamy than it is near the heating surfaces. Under the flues it is nearly solid, over them it is foam. In the middle and up to the third row from the top there is probably superheated steam, which rises through the upper flues and foam, and gets saturated. In the legs there is foam, and sometimes only steam. There the foam is most dense, it goes down and drives up the thin foam, and it is doubtful whether there is any thing so solid as half-and-half anywhere in contact with the heating surface. The down currents mix more or less with the up currents, and get a good deal lightened with foam, which renders the circulation slow. Circuzlatin Passages. —If outside pipes or passages are provided, free from heat, so that there can be a free down-flow, forced by the weight of a column of water a foot high or more, in the legs, and five or six inches in the barrel; there will be a quiet circulation, more steam will be made, and the metal will not be overheated, unless the pipes are small. Or if inside pipes are provided for down-flow, the same benefit is attained, if there be room enough. Jacob Perkins, about 1830, made experiments to test the utility of circulation, and took out a patent for apparatus to promote it. This plan was tried on the Liverpool & Manchester Railway, and found to increase the production of steam and the durability of flues. The arrangement tried there consisted inputting curved plates between the outer shell and the flues, to separate the downcurrent of water from the up-current of mixed water and steam. Mr. A. F. Smith of the Hudson River Railway, has made another application of this principle. He has applied circulating partitions in the water-legs of the fire-boxes; the partitions being inclined, so that the water-current widens as it ascends. The plates were made adjustable by screws, so that they could be inclined more or less, and put more or less distant from the outer shell. The results have been manifestly good, but are not definitely reported. Preliminary to the introduction of these circulating plates, Mr. Smith made experiments on a small scale. A tin vessel with a flue up through its middle, was filled with water, and subjected VAPORIZATION. 133 to the heat of gas-burners, and the action of water and rate of vaporization were noted. A circulating cylinder, intermediate between the flue and the outer shell, reaching nearly to the bottom and nearly to the top, was then put in; and caused a freer and smoother circulation, and more rapid generation of steam. The action of the circulating plates not only increases vaporization, but diminishes the disturbance which sometimes causes priming, and probably, at all times more or less moistens the steam by throwing spray into it. Mr. Joseph Nason, who in early life was with Jacob Perkins, and familiar with his experiments, has applied this principle in a small boiler, and with such satisfactory results, that he was induced to patent an improvement on this method of promoting the circulation. He first tried a vessel with a single 2-inch tube 2 ft. long, with a'-inch circulating tube, in a fire urged by a blower to as high a temperature as the tube was likely to sustain in practice. When the circulator was withdrawn, the water was driven out with great commotion and noise; but when it was replaced, there was a free and quiet circulation, and separation of steam. The circulating tubes had formerly been straiglht, so that the foamn ascended all around them, and interfered with the access of solid water to the circulator. To remedy this, Mr. Nason prolonged the vaporizing tube upward, and turned the circulating tube through its side, so that it received solid water below the level at which the foam issued from the annular space. A casting was made for this junction, into which both pipes were screwed, and the casting was screwed into the bottom of the boiler. The tubes were set as close together as was consistent with the strength and durability of the tube-sheet. The boiler is an improvement on that which was originally invented by Rumford, and improved by Perkins. It has worked satisfactorily, although the tubes are small; and is a good illustration of the utility of circulating partitions. Mr. W. G. Hamilton, engineer of the Jersey City Locomotive Works, proposes to place outside down-flow pipes upon the fire-boxes of wood-burning locomotives, with narrow water-spaces, for the purpose of adapting them to the more intense heat of coal. Although the plan has not been thoroughly tried, as far as the author is aware, its failure would manifestly be contrary to all the well-known principles which have been mentioned. To appropriate the whole of the original water-space to the up-current of steam and water, the down-flow pipes for solid water should be, in the opinion of engineers who have examined the subject, equal in capacity to at least four 3-inch pipes. It is to be hoped that the plan will not be condemned, as Clark's jet has been, because it did not work on a small scale. It is not probable that a couple of 1-inch pipes would "warrant a further trial," any more than did four'-inch steam inducted air-jets for coalburning, as previously mentioned. Two 3-inch copper pipes on each side of the fire-box, could be conveniently applied to almost every variety of modern locomotives. The pipes would be bent on a large radius at either - end, so as to form, in these bends, an expansion joint similar to that of Dimpfel's water-tubes. They would be secured to the outer plate by flanges, and would enter the boiler at the upper ends, as far as possible above the level of the crown plate, but low enough to insure their being filled with solid water. It might be found expedient to extend them to a central pipe, which should take solid water at the extreme forward end of the boiler. At their lower ends, down-flow pipes would deliver water to the extreme bottom of the water-space. They would not only widen the actual capacity of the water-space, by the amount of their area, but they would greatly facilitate the velocity of the current, and increase the amount of solid water brought into contact with the heating surface. It would be absurd to conclude in advance that this plan is not feasible, since the whole of the evidence, such as it is, sustains it. In anthracite-burning engines, particularly, the destruction of fire-plates is so great, that such simple and probable preventives as this, should have been tried, as they were suggested long ago. On the whole, there is sufficient evidence to show that water may be kept in very narrow spaces, by means of plates or pipes, to separate the down-currents from the up-currents. Whether this practice is better than that of ampler water-spaces-greater simplicity of construction, and facility of cleaning, is a question for designers and builders, and for those who use - and repair locomotives, and have opportunities to balance all the advantages and disadvantages attendant on the weight or lightness incident to either practice. Lightness has been sought, without sufficient 134 THE ECONOMICAL GENERATION OF STEAM. examination of the consequences, and has in many cases not been attained; that is, a ton of boiler with deficient water-spaces has not made so much steam, as a ton of boiler with less metal and more water; and the cost of repairs has been much greater. In the barrel of the boiler, beyond the combustion chamber, there has been no great change of late; flues are, with few exceptions, set as close as formerly: but forasmuch as there is less heat in the flues of a combustion-chamber boiler, than in the fire-box ends of flues in a common boiler, the narrow spaces are less objectionable. The most enlightened opinion, however, is in a considerable degree opposed to present practice: it is in favor of more space between the flues, and more room for down-flow between the flues and shell. And the use of circulating plates to make a down-channel next the shell, would be in accordance with the views of several who have thought and experimented on the subject. Arrangement of Flues, — The arrangement of the flues in vertical rows, is rather gaining favor, on account of its aiding circulation. Mr. Harvey Rice two years ago arranged the tubes of an engine in diverging rows, to give more room as the steam ascended. Mr. Swinburne is in favor of 3-inch tubes set wide apart. Some years ago he built a Cornish boiler, to assist a tubular boiler that did not make steam enough. The Cornish boiler, with much less surface, made all the steam required; and his observations on these boilers convinced him that the crowding system is erroneous. Mr. Beattie's short flue boilers differ little from Cornish's boilers, except in the shape of the flues, and the quantity of water they carry. A moderate extent of well-arranged surface receives heat from the smoke, and imparts it to the water as effectively as a much greater surface arranged in the old way, even when the combustion is equally perfect. The multitubular plan, at first with coke, worked much better than the old plan; probably because the blast-pipe came with it, and the load of water was lessened, and great talent was concentrated upon it. But since these advantages have been applied in aid of the old plan, by the introducers of the combustion-chambers, the extreme confidence at first yielded to the multitubular plan has been moderated; and the subject has been reviewed. The conviction is, that not only the combustion-room, but likewise the water-room, has been too much obstructed and choked; that froth and steam linger where solid water should be; and that the metal does much less than its full duty in imparting heat. Rate of Draft through Flues.-In a marine flue-tube boiler, Mr. Isherwood states, zinc has been melted in a second flue from the top, while tin was not melted in a lower flue, proving that the heat was more than 773~ in the upper, and less than 440~ in the lower, with a natural draft. In this boiler the smoke-entrance and smoke-box are 3 feet deep; the difference of gravity between the smoke in the furnace and in the smoke-box, for a column 3 feet high, is too slight to make much difference in the quantity that passes through the upper and lower flues; it is, therefore, probable that there is another cause for this difference of temperature. Williams found that foam received heat slower that solid water, in the proportion of 8 to 10. The upper flues are immersed in foam, the lower in water. The rise of water-level is 5 inches, in some locomotives, that is, when the steam is shut off, the water-level immediately falls 5 inches, showing that there is, say 20 cubic feet of steam mixed with the water. From this it is evident that there is a great volume of thin foam in contact with the surfaces; and it may be suspected that where the flues are - or, inch apart there is little more than spray, and sometimes dry steam, in the upper region of the mass of flues, while there is solid water at the bottom. Hence, it appears that so long as flues are equally spaced and of equal sizes, an equal flow of heated gases through them is not wanted; there should be less through the upper row, so that the flues which are not cooled by solid water shall be allowed more time to cool their gas. The petticoats and aprons are useful in this regard, so far as they unequalize the flow; but it is questionable whether they are as efficient as the current-disturbers may be made. Duncan & Gwynne, as we have stated, use plates twisted into short turns like an auger, others twisted less, and some with one twist in a yard. They may extend the whole length of the flue, or only some 15 inches, as Mr. Dougherty has used them. This kind of mixer allows the required facility of retarding the gases more or less. Let us now suppose that we leave the lower flues open; in the next row we put short plates VAPORIZATION. 13 5 slightly twisted; in the next, larger plates; and so on until near to the top row, where there is least water, in which the plates may run through, and have short twists. Thus we may vary the resistance, and help the reception of heat to equalize the temperature of the gas at the front ends. To ascertain whether it is equalized, we may use alloys that melt at different temperatures. Place in each trial flue several bits; and at the end of a trip see which are melted; and thus we may find how to adjust the disturbers to any required nicety. It is highly probable, however, that the employment of flues of different sizes, in the manner referred to in a preceding chapter, would better meet the various and complicated requirements of transmission of heat and vaporization. If we simply equalize the flow of gas through flues equally spaced, by petticoat pipes and diaphragms, we evidently prevent the highest possible amount of vaporization, for reasons just stated; if we compel it to flow wholly though the outer flues, its velocity is so great that it cannot transmit all its heat to the metal. But if we decrease the size of the upper and middle flues, by increasing the water-space around them, and at the same time preserve the total flue area by increasing the diameter of the outer flues, the results must be substantially as follows: 1st. The current of gas through the outer flues can be slow enough to part with all its available heat; 2d. The current through the middle and upper rows will have a large body of comparatively solid water to take up its heat; and 3d. There will be a plenty of room for the steam formed below to rise to the surface, in addition to an ample water-space. There is hardly any feature of locomotive economy so little understood as this, and there are not many features of greater importance; it should be thoroughly tested by an experiment which, by embracing all the working conditions, would establish a useful and final conclusion. No individuals, however, who are capable of conducting such an experiment in a manner which would be accepted, have either means or leisure for it; and it is vexatious to experiment for the public goo(^ when one-half the railway managers neglect many improvements which long practice even has established. Were there but one railway, such experiments would be tried-witness the early trials on the Liverpool and Manchester Railway. But with so many railways, and 9,000 engines, there is little prospect of speedy and definite improvement. CHAPTER IV. SEPARATION AND SUPERHEATING. WHERN, in virtue of good combustion, transmission and. vaporization, steam is at least economically formed in a boiler, the practical business of " the economical generation of steam "-its delivery to a point where it can go to work-may be but half accomplished. Its tendencies are so great to carry water with it into the cylinder, and to condense, if its temperature at formation is not maintained at every point of its passage, that, in boilers badly adapted to these conditions, abote one-half the fuel is known to have been wasted after it had been once utilized. 1. SEPARATION.-It is to be regretted here, as at many other points in this checkered history of steam generation, not only that practitioners have not decided as to the best means of separating the steam from the water, but that they most positively disagree as to the operation of the same means. While one believes that the steam dome should be over the coolest part of the boiler, where there is least vaporization and commotion, another believes that steam is hotter and therefore dryer when taken from the hottest part of the boiler. While one insists that the perforated pipe, where its slot or perforations decrease in size nearest the throttle-valve, maintains an equable draft upon the whole of the surface of the water, and therefore does not take steam fast enough at any one point to induce foam to rise with the steam, another insists that there is always spray in the steam at the point near the water-level, where the perforated pipe must be placed, and that steam is only drained by rising into a high dome. In practice, well-made pipes and large domes generally appear to give similar results, and sometimes, the ultimate working of boilers with comparatively narrow water-spaces and small domes, is not materially different from that of boilers which are designed to carry dryer steam. To conclude from these facts, however, that the question of separation is of minor importance, would be to forget that, perhaps, fifty other features of boiler and engine-each with a small percentage of economy or loss, within its power, and the functions of each, not definitely known-may have compensated for a very great loss by priming. Some experiments have been made and some results of practice have been authenticated, which decidedly show that the loss of fuel due to priming, is from 10 to 30 per cent., and that the general remedyis, 1st and chiefly, ample steam room; 2d. Taking steam as far as practicable from the water, and 3d. Allowing it to develop all the distance possible, without reducing its temperature and pressure, around deflecting plates, before it enters the delivery pipe. Nothing more definite than this has been determined; and it is probable that less attention has been paid to the subject of separation, since the subject of superheating has been revived, in the hope that the latter process would better meet the required conditions. The construction of steam domes and steam room, witlh, reference to the delivery of dry steam as well as strong and cheap boilers, has been referred to under the head of boiler-making. Superheating, or heating the steam a second time, will not only insure the dryness of the steam, but it will allow the use of much simpler and stronger boilers; and herein, therefore, appears to lie the proper solution of the problem. 2. SUPERLHEATED STEAM. —In ordinary practice, steam exists in three conditions. When the steam-room is ample, the water pure, the pressure high, the water-spaces wide, and the firing moderate, it is pure steam. When the steam-room is contracted, the water foul, the pressure low, SEPARATION AND SUPERHEATING. 137 the water-spaces contracted, and the firing intense, the steam is water-logged, or filled with spray.* And when there is heating surface in the stean-room to vaporize the spray, and give a temperature above what is due to the pressure; the steam is superheated.f Pure steanm, of atmospheric pressure, when heated out of contact with water, is expanded ^ of its volume for each degree of temperature. But water-logged steam is expanded to a much greater extent, not merely by superheating, but also by the conversion into steam of the watery particles suspended'in it. The advantage in the latter case is greater than in the former; because, in addition to the advantage of superheating, there is the advantage of preventing the loss of hot water, i. e., of the fuel which heated it. If the pipes and cylinders be not so protected as not to lose heat, either by external radiation, or by imparting heat to the exhaust, there will be condensation; and power will be wasted. Superheating to 50~ is considered sufficient to prevent condensation in cylinders and pipes ordinarily protected. We have, therefore, to add to the amount saved by superheating, whatever may be saved by avoiding condensation, which is estimated at from 10 to 30 per cent., and is a very variable quantity, since the extreme conditions affecting it, are to be met in daily practice. The advantages of superheating have not been accurately determined either in practice or in theory. But it is interesting, and may be useful, to know the results that have been arrived at, in trials believed to have an approximation to truth. The economy of superheating the steam, instead of using the fuel to make fresh steam, depends on three things: 1st, How much heat is required to give an equal increase of power; 2d, How much surface is required to transmit the heat; and, 3d, How much less condensation does the heat suffer in the cylinders and pipes. The first question depends on the specific heat of steam, that is, how much heat is required to raise its temperature; and this is not certainly known. Crawford, one of the early experimenters, states it 1'55; Delaroche & Berard, at *847; and Regnault, at'475. Delaroche & Berard's statement has been generally adopted; but Regnault has acquired the reputation of being the most accurate experimenter, and Rankine and other authorities adopt his results. The unit of heat adopted in treating of these subjects is that amount of heat which will raise the temperature of lb. of water 1~. Now, according to Delaroche & Berard, to raise the temperature of 1 lb. of steam 1~, will take'847 of a unit of heat. To vaporize a pound of water at 120 lbs. pressure, requires as much heat as would raise it to 1200~. If the feed-water is at 60~, it leaves 1140~, or 1140 units of heat as the quantity to make l lb. of steam. Dividing this by the number expressing the specific heat,.1,- == 1346, gives 1346~ as the increase of temperature that will be given to a pound of steam by the heat necessary to make a pound of saturated steam from water at 60~.. If this be divided by the number of degrees that are required to double the volume of steam,:346= 2'8, it gives an increase of volume of 2 8 by the fuel that would give only 1 if employed, to make fresh steam. The fuel expended is 2 in each case: but the volume is 2 for the moist steam, and 3-8 for the superheated; showing that a gain of 90 per cent. might be made if we could superheat to that extent. But moist steam at this pressure has a temperature of about 350~; and as shown repeatedly in practice it would injure the cylinders and superheating apparatus to heat it much above 500~. It is, therefore, unsafe to take more than an eighth of this amount of superheating, which will raise the volume of steam to 1'35, and the consumption of fuel to 1'25. This gives ~i'3 5-' 833, for the consumption of fuel, or a saving of one-sixth as the theoretical saving by superheating from 350~ to 518~. It is, however, considered by the old experimenters that the specific heat decreases as the vol* The terms " intense" and " moderate" firing are here used to express extremes which have been considered under the head of Combustion. t There are some who suppose that steam is not perfected instantaneously; that there is an intermediate state between steam and water, in which the volume is much less than that of perfect steam. They adduce the experiments of the late James Frost, which they have repeated; and claim that 4~ will double the volume, at atmospheric pressure; and 12~ more will increase it another volume that is, 16~ will expand it to thrice its volume at 212~. The usual explanation of such a phenomenon is, that the steam is water-logged, and the increase of volume due to vaporization. 18 138 THE ECONOMICAL GENERATION OF STEAM. ume is diminished by compression-that to superheat a pound of steam compressed into a ninth of its volume under atmospheric pressure would require less than.847, how much less is not known; but some experimenters regard it as very much less. So far as this is true the gain would be greater than we have shown. And this is in favor of superheating high-pressure steam. According to Regnault's statement,.475, the heat applied to superheating will be 5 times as effective as that applied to vaporization; and the fuel used to make and superheat a pound of steam, at 518~ will be 1.07; hence the economy will be I. - -=.792; or over 20 per cent. saving. On the second point, the extent of surface required to superheat, compared with the surface required to make fresh steam, we have not much light. C. W. Williams says, that in an experiment-under what pressure is not stated-6 feet of surface in solid water took up as much heat as 10 feet in steam. We have no authority for it; but there is an appearance that dense steam takes up heat faster than rare steam, how much faster, and what relation the density has to the power of taking up heat, does not appear to be known. But if Williams' statement that steam will receive heat.6 as fast as water, is correct, it must follow that superheating surface is really more efficient than generating surface; for since the heat is 2.8 times as effective, and transmitted.6 as fast, it follows that.6 X 2.8_1.68 is the rate at which the superheating surface would give power; that is, it is 68 per cent. better than an equal surface exposed to water, provided the difference of temperature be equal. But as the temperature of the steam is to be 160~ higher than that of the water, there must be a deduction on that account, the amount of which will depend on the temperature of the gases escaping from the furnace. Assuming 350~ for the temperature of the water, and 600~ for the gases, leaves 250~ for the difference; assuming the mean difference of temperature while the steam is superheating, we have 2.8 X.6 X.67=- 1.ll; or the surface used for superheating is a ninth more effective than equal surface used for vaporization, exposed to the same smoke. If Regnault's estimate of the specific heat be true, the case will stand 5 X.6 X.67 =2.01; or the superheating surface will be twice as effective as the same surface used for making steam. The weight and bulk of a superheater would be less than that of an equal extent of flue surface, because the flues would be set as close as is consistent with strength of the tube-sheets, and there would be no water around them requiring a large circulating space; but it would have to be inclosed in a long smoke-box, and its attachments would be of considerable weight; still it might, on the whole, not exceed the weight of an equal extent of water surface. It has been suggested that in Beattie's boiler (Plate 39a) the region of the tubes might be occupied by a superheater formed by tube-sheets riveted to the barrel, except for a sufficient space at the top to allow the steam, to pass over; the combustion-chamber terminating six inches from the tube-sheet of the superheater. This would be light; but whether the heat of the gases would not be too great for it is a matter to be tested before it can be called feasible. Pieces of metal should be tried for a long time in the front end of a long combustion-chamber, to determine the temperature and probable durability of metal in that region when exposed to steam. If Williams' statement be correct, and steam will receive heat.6 as fast as water, such an arrangement might last tolerably well. And there have been cases in which superheating apparatus has been exposed to the direct heat of the fire, and has lasted a considerable time. But in other cases the apparatus has been soon disabled; in the steamship Arctic it was exposed to the heat inside the furnaces; and in a few hours a cast-iron elbow was cracked, so as to let out more steam than could find its way up the chimney, The adoption of smaller and thicker superheating pipes of wrought iron in some of the recent English practice has been more satisfactory. The third point, the condensation in the cylinders, is quite unsettled; it has been held in England that from 18 to 40 per cent. of the water is carried off in the liquid form, by priming or by condensation; but recently Prof. Rankine has made experiments which indicate that the tables of pressure and density of steam hitherto used, are erroneous; that steam of 130 lbs. pressure is but 190 times the volume of water, instead of 210 times, as has been believed. This at once takes II per cent. from the supposed waste of water; and reduces it to a small amount in the best engines. It may be added that little or no allowance has been made for leakage by experimenters on locomotives, while those lwho experiment on other engines sometimes attribute much loss to this cause. There is probably a loss by condensation that is of sufficient consequence to SEPARATION AND'SUPERHEATING. 139 add something to the reasons for superheating; but those who attribute the whole advantage of it to the avoidance of water, are probably in error, theoretically, however correct they may be as to the fact that the practical advantage ceases at about the point where the cylinders work dry. D. K. CLARK, in " Recent Practice," says: " To convert steam into a perfect gas, it must be heated above the temperature due to the pressure; at 130 Ibs. above the atmosphere it requires 50~ additional temperature. It is established that the full measure of advantage by expansive working is only to be had by the adoption of a process of superheating; and recent experience in ordinary engines would showv something like an economy of 20 to 30 per cent. of steam, when the process is thoroughly applied." This estimate agrees nearly with the results of trials of stationary engines in England and this country, within three years. The following notes of experiments and practice in these matters may suggest useful ideas: Dr. Haycraft, of Greenwich, England, about thirty years ago, constructed a small engine, which hlie worked in the ordinary way, noting the workl done, and the consumption of water and fuel. He then applied heat around the cylinder, and also passed the steam through a superheater, and found an enormous increase of work from a given weight of fuel. Encouraged by the result, he -adopted means to heat the steam to a temperature that soon spoiled his cylinder and packing. But every increase of temperature gave some increase of power from the fuel. Warned by the injury of his engine, he proceeded cautiously to ascertain the point which would give the best effect, having due regard to saving of fuel and repairs. He constructed a faggot of tubes, which he heated by a separate boiler, in which he could so regulate the pressure as to superheat the working steam any number of degrees. Through this faggot he passed the steam on its way to the engine. A thermometer was so placed as to show the temperature to which it was superheated. One degree showed an improvement; two degrees about double; and each additional degree showed a sensible improvement, up to about twenty degrees; and at this point the increase of effect ceased. Dr. Haycraft concluded that what he at first took for a new and important principle was in reality but a remedy for a disorder incident to small and unclothed engines and boilers and steampipes; and he saw that Watt had fully understood the disorder, and, by the steam-jacket, had endeavored to remedy it. The clothing of the Cornish boilers and steam-pipes, not generally noticed at that time, had also produced the advantages which Dr. Haycraft thought he was the first to produce. Four years later, Mr. A. S. Woolcott, of New, York, in experimenting on a very small engine, to test a new cut-off, accidentally made the same discovery. He was troubled with the " spitting of water from the exhaust-pipe;" and to avoid it he enveloped his engine in a case, which he filled with steam of a higher pressure, from another boiler. With this help the engine ran ten to twelve times faster, with the same number of gas-burners to supply steam. He drew the same conclusion that Dr. Haycraft had drawn; but had not at that time published. In 1850, Mr. James Frost, of Brooklyn, before the publication of Dr. Haycraft's experiments, made a course that led him to the samne results that Haycraft at first attained; but did not apply such excessive heat as to injure his engine, and turn him to such a course as might have revealed to him the real state of the case. His experiments attracted the notice of Mr. E. KI. Collins, who caused them to be repeated on a larger scale. There appeared to be a great saving of fuel: in consequence of which it was decided to put a superheating apparatus on board the steamship Arctic. A series of pipes were put into the furnace, exposed to the full heat of the fire. On the trial trip a cast-iron elbow cracked, and leaked so badly that it became necessary to shut off the steam from the whole series. The pipes were taken out; and the Arctic, a few days afterwards, departed on her last trip. Messrs. Penn & Son, of Greenwich, have lately fitted a steamship with a superheater, exposed to the smoke; and have stated that it saves 30 per cent. of the fuel. It appears, however, that the temperature of the uptake is too high for the superheating pipes. The Allaire Works, of New York, state that their steam-chimneys save about 20 per cent., and that they are as durable as the furnaces of the boiler. 140 THE ECONOMICAL GENERATION OF STEAM. On locomotives there have been some trials: McConnell has used a tubular apparatus in the smoke-box; and is still engaged in modifying his apparatus: Sharp, Stewart & Co. have tried the steam-chimney, and a superheating chamber in the smoke-box; but the results have not yet given satisfaction. Mr. James Millholland has, for some time, employed the faggot of steam pipes shown in the smoke-box of his engine (Plate 53), for the purpose of exposing a larger surface of steam to the heat of the gases. Mr. Richard Eaton has still farther carried out this plan, as shown by plate 451. Mr. A. F. Smith has applied the arrangement shown by fig. 13, plate 58, to some of the engines on the Hudson River Railway. In these three instances, the results have been obviously good — exactly what they have been, it is impossible to say-they have been measured by a standardthe consumption of fuel- which measured every other result, and every other varying external and internal condition at the samne tine. The results of partially drying the steam, for it could not have been superheated in these small vessels, cannot be said to have been measured either as to quantity or quality, but what is known, is in the right direction. Mr. D. K. Clark proposes a superheater (Figs. 3, 4 and 5, Plate 66), which will give a very large surface contact of steam with the gases, in a small compass, and evidently possesses considerable advantages. It has not yet been applied. Mr. Thomas Prosser has for some years used a boiler invented and built by himself; and has recently submitted it to the Navy Department, by whose order it has been experimented upon by Chief Engineers Isherwood, Everett and King, of the Navy. There is a full and minute account of these experiments in Isherwood's " Engineerizng Precedents," from which we gather the following particulars: The boiler consists, first, of a vertical cylinder 30 inches diameter and 261 inches high; second, of a square chamber formed of 2-inch tubes 50 inches long, set close together, and screwed by necks into the bottom of the cylinder; third, of a water-slab, 11 inches thick, which forms the bottom of the chamber, and into which the long tubes are also screwed by necks, and which is perforated by 100 thimbles 1 inch diameter, to make it serve as a grate; fourth, of smaller water-slab, that is placed 20 inches above the grate-slab, and forms the crown of the fire-box; fifth, of 25 2-inch tubes, which connect the crown-sheet slab to the bottom of the cylinder; sixth, of 24 l~-inch flues, which pass through the crown-sheet slab, and through the 2:inch tubes and the cylinder; seventh, of 76 i~-inch flues through the cylinder. The water fills the outer tubes which form the square chamber, and also the annular space between the 24 2-inch tubes and the -11-inch flues that pass through them. One of the 2-inch tubes is left without a flue, so that the water- may descend through it from the cylinder to the crown-sheet slab; and thus a circulation is established by down flow through this tube, and up flow through the annular spaces. The boiler contains 113 square feet of surface below the water-line, and 54 square feet above, in the steam-room. When at work it is probable that the foam rises so as to add very much to the vaporizing surface, and lessen the superheating surface. During the trials the pressure varied from 100 to 140 lbs., and the steam was from 26~ to 39~ hotter than the water, averaging 30~ of superheating. The steam-room was 20 inches deep, when the water was still. The temperature of the gases in the smoke-box was 440~ —about 60~ degrees hotter than the steam. Thus we see that when a third of the heating surface is above the water-line, and the boiler worked so slowly as to reduce the escaping gases to 440~, there is less superheating than is desirable. Were the temperature 620~ or more, as in locomotives, the difference of temperature would be four times as much, and the superheating considerably more. But as it is improbable that any thing like solid water was in the lower part of the boiler, and that as much steam was made as the surface should have made, we cannot consider this experiment as a guide. It merely shows, in a general way, that an upright multiflue boiler will superheat the steam to a useful extent when the escaping gases are at a temperature lower than is usual in locomotive practice. If, instead of this form of boiler, there had been an upright boiler of the usual form, of equal area, the surface exposed to steam would have been:one-half greater in the same depth of steamroom. SEPARATION AND SUPERHEATING. 141 Besides the practice before mentioned, there has been considerable desultory practice on a small scale, with upright multiflue boilers, in which the flues pass through the steam-room, and for about a foot at the upper ends, act as superheating surface. These boilers are much liked, although it is theoretically objected that the flues are liable to burn out at the upper ends. We have not learned that this objection has any foundation in fact; on the contrary, it is generally said that the lower ends fail first; and the fire-box crown, when impure water is used, is liable to incrustations, and is the really perishable part. But the working of these boilers, when wellmade, and clean, is said to be economical and efficient: and this reputation of them tends to confirm the view, that a moderate extent of superheating is advantageous. Penn & Sons consider that the condition of economy in superheating is, that the escaping gases shall be used for that purpose; thu, saving heat that would otherwise be wasted. Haycraft considers that a separate fire may produce economy by superheating. His trials were with very small engines, that are most liable to waste by condensation; and Penn's practice has been with large engines. Their opinions are probably due to their different experience. If Penn be right, a very long locomotive boiler would'be, as economical of fuel as a shorter boiler with a superheater; but if Haycraft be right, it would be well to shorten the boiler, and use the ends of the tubes for superheating. And as a precaution against rapid wear of the superheater, it is of course prudent to put it beyond violent heat. The neglect of this means of mitigating the well-known evil of condensation in the cylincers, has been owing to the trouble of a complex apparatus attached to an engine, but not a natural part of it, and in the way when examination or cleaning is required. There must be a strong reason for departure from simplicity, in locomotives; which, some complain, have already too much complexity. But this is not so much a mere matter of convenience as to warrant the neglect of an apparatus that can save 20 or 30 per cent. of the fuel, and relieve the boiler from overwork. If it be true that any thing near this percentage has been saved; and if the higher pressure, or other conditions of the locomotive, do not exempt it from the need of such an adjunct; it is worthy of careful consideration whether this apparatus may not be improved, and applied so that its saving may not be countervailed by inconveniences. 3. CONCLUSION.-Whether the locomotive boiler really needs a separate superheater to keep its steam dry is not yet settled beyond dispute. There are engineers of repute who think it does not; and that the true way to avoid water in the cylinders, is to give ample steam room, and to keep a high pressure, and to be careful about the cleanliness of the water. Recent boilers, with combustion chaimbers, have long raised crowns, with domes set farther forward than they used to be: thus more room is allowed for steam; and the dome is over a part where the ebullition is probably less violent than it is immediately under the domes of boilers that have no combustion room in the barrel. It is claimed that the steam from these boilers is free from the mnisty particles that accompany the steam in boilers that are deficient in steam room; and that therefore they do not need a separate vessel to vaporize the misty particles of water. That part of the water in cylinders which comes by condensation, has been much lessened by lagging, and bright metal casings; and in some instances by jackets in which the exhausted steam is allowed to expand. Beattie's Cylinder, plate 66, fig. 1, is an instance of the practice on the London & South Western Railway. It may be said that exhaust steam is not so hot as that entering the cylinders, and therefore is a condenser rather than a heater. It is hotter than the atmosphere, however, and is good lagging at least. The engraving is given principally to show a manner of steam jacketing, by using fresh steam around the cylinder, instead of exhaust steanm, in a manner not only consistent with the best construction of cylinder, but important to it. The lining is a sheet of steel, welded, turned and bored, and shrunk into the outside cylinder. The casting is set on end, and the bottom of the steel cylinder is temporarily filled by a disk of wood, so that it forms a bucket, which being filled with water, prevents the heating and expansion of the steel. The casting is then warmed by a slow fire till it expands sufficiently to allow the steel cylinder to enter. Watt's jackets had both steam from the boiler; and sometimes from a small boiler of higher pressure. If a jacket be paid for, and carried over the road, it is not easy to see 142 THE ECONOMICAL GENERATION OF STEAM. why strong steam from the boiler should not be in it, rather than the cooler steam from the exhaust, which is not better than so much confined air. If strong steam be in the jacket, there must be a simple way of removing the water of condensation. It may be blown into the tank, from time to time; or it may be pumped into the boiler. In some recent small engines it runs into the boiler by its own gravity, the cylinder being situated above the water line. And in a few cases of portable engines, the cylinder is inside the steam dome; a renewal of old practice;and as it is the practice of Watt and his immediate successors, and based on careful experiments, it is worth considering. If a self-acting means of forcing this water up into the boiler could be devised, steam-jackets would be more likely to find favor. It has been suggested that the momentum of the steam flowing into the steam-chest, might be made to do this, on the principle of the hydraulic ram, or that the Giffard injector might be adapted to this purpose. Pressure is an important element in relation to dryness. Regnault's experiments have shown that high-pressure steam contains more heat than low-pressure steam, and, when expanded, is slightly superheated. This, in addition to the restrainft it imposes on ebullition, tends to dryness. And if the cylinder could be kept at the temperature of the boiler, there is reason to believe that high-pressure steamn, free from mist, would need no superheating; but these reasons, though sufficient to warrant careful observation are not sufficient to refute the., arguments, based on successful trials, of those who maintain that a superheating apparatus is necessary, even at 150 lbs. pressure. Excepting the early experiments, many of the later developments as to superheating or increasing the temperature of dry steam, in distinction from simply drying saturated or waterlogged steam, have not been referred to, chiefly because it is believed that simply drying steam is at present all that can be economically undertaken in locomotive practice. Wethered's method of mixing saturated and superheated steam is undoubtedly an excellent way of regulating the temperature, so that it shall not injure the cylinder. Any other feature of economy in employing mixed steam, has not been fully established, although such results are possible. The plan has not been applied to the locomotive. On the whole, it is highly probable, that drying the steam by heating it a second time-that is to say, making dry steam out of the steam-and-water now delivered by locomotive steampipes-is entirely feasible, and that in addition to the direct result aimed at, boilers may be simplified and strengthened; and withal so cheapened as to allow the drying apparatus to be applied without increasing the total cost of the locomotive. But the first step in attaining to any considerable economy, without random and costly experiments, is to definitely measure what has already been done in this direction. The general principles and features of the system have been mentioned. The result is now entirely in the hands of the practitioners themselves. CO-ALBURJNING PASSENGER ENGNE _.o,-0,1o;10.11_, Soet/h,~gk'TI~,sm-nJ _/?a.t.-frer.".,5"the rail on the wood, but more especially by the want of sufficient firmness in the bearing of the sleeper on the ballast and the want of strength in the fastenings. On badly ballasted roads, or where the surface-bearing on the ballast is too small, it is very common to see some of the sleepers loose, or hanging by the rail. Thus, whenever a train passes, the sleeper is forced down with a crushing blow. Sound timber is often worn out; it is, therefore, of little advantage to undertake the cost of preserving wood from rotting, until a better system of bearings is employed. On some of the Austrian roads, where timber is valuable, the sleepers are planed smoothly all over, both as a means of preventing decay and of improving the bearing surface. On several roads where the flat-footed rail is used, the sleepers are grooved by machinery to receive the foot of the rail. This aids the spikes in keeping the road in gauge, although it weakens the sleeper and hastens its decay, by giving a lodgment for the water. Comparatively, then, the cross-sleeper system requires a larger amount of timber and a very much greater width of ballast, than the longitudinal system, for a given permanence of way. WFooden Blocs. —In France, on one railway, the experiment was tried of cutting most of the sleepers in two at their centres, leaving one whole at every 10 or 12 feet to bind the track together. The object was to destroy the spring of the sleeper, which, taking a bearing on its own centre, will deflect at the ends by the weight of a train, and thus chur tthe ballast in such nma2nner as to lessen its own bearing and hasten the general disturbance of the track. The experiment is understood to have succeeded well, and W. Bridges Adams of London, has recommended the general observance of the same practice. The forms of cast-iron sleepers, thus far introduced, are generally detached rectangular blocks, either lengthwise or across the track, and these have been found to bear better and to require less frequent packing than the transverse sleepers. Tlhe practice of sawing the sleepers in two would not answer excepting the separated ends were well packed in ballast, as there would be a tendency to strain them from their fastenings-but, if these were secure, there can be no doubt that the bearings would be better than with the present whole timber. On the Boston & Providence line, some of the joint sleepers were at one time turned half round, forming longitudinal bearings. The results are said to have been good. The cast-iron sleepers, used abroad, which will be further mentioned, are quite as advantageous as tofor2z, as in durability of material. They are found to pack better and to lie more quietly. They do not spring out of the ballast and hammer back into their beds, on the passage of a train, after the manner of long elastic beams with a partial bearing. In short, they are bearers and not levers, and they utilize all,their surface, by transferring the load directly to the road-bed. Wooden blocks dressed to similar forms and sizes, and say 4 inches thick, would give all these advantages of form, besides furnishing ample elasticity, and dispensing with an amount of ballast between the rails, which would be indispensable to an equally firm cross-sleeper road. The rail-joint proposed by Mr. Zerah Colburn, and shown on plate 26, figs. 15 to 17, is a longitudinal or block joint-sleeper, being in fact, a section of Diimpfel's longitudinal system. (Fig. 13, plate 36.) It is very evident, from the principles and experience already mentioned, that this jointsleeper would give all tavanag the advantages of the block or cast-iron system as to form, besides making a much better joint than the chairs or wood splices commonly used. The rail has an ample bearing on so much of the timber that it could not rapidly crush into it, and it has more vertical stiffness than a dozen ordinary wood-splices. The longitudinals would of course have the same bearing on the ballast as that of a common joint sleeper, and this would be effective bearing, under the load and not at the end of a flexible lever. Cast-iron Sleepers. —The high price of timber, and the desire to provide a permanent way without the many disadvantages of the transverse sleeper system, have led several permanent-way engineers to design varieties of cast-iron sleepers, some of which have been tested for a long time, and are now in use on some of the English linles. The use of cast-iron, for this purpose, is so much a matter of cost, that the subject can claim no such attention here as it receives in England; the advantages of cast iron as to form, are of special importance. There, an ordinary road requires $2,100 per mile of single track, for timber 164 PERMANENT WAY. sleepers-these being preserved to last twice as long as those used on American roads-probably twice and a half. The plans proposed, of cast iron, weighing from 255,000 to 358,000 pounds per mile of single track, would cost from $3,575 to $5,000 a mile, would probably last as long as wood, and would be worth a considerable sum as old iron when ultimately renewed. In the United States, the sleepers would cost say $550 a mile, the cast-iron from $5,000 to $9,000, making the consideration of cast-iron quite out of the question without some extraordinary advantages were to be derived from its use. It will be interesting, however, to observe the more prominent plans of cast-iron roadway brought-forward in England. Samuel's cast-iron, timber-cushioned sleeper, the property of the Permanent Way Company, is clearly shown by figs. 10 to 12, plate 30. The casting is 42 inches long, 16 inches wide, weighs 132 pounds, and gives 4-66 square feet of bearing surface, each sleeper. So laid as to make the total bearing surface equal to 12 square feet per foot of track forward, would require 100 tons (of 2,240 lbs.) per mile. For 2 s quare feet bearing per foot of track, 133 1 tons of 2,240 lbs. This sleeper was applied, for an experimental length, on the Eastern Counties line, some time about 1850, or say 10 years ago. It has been repeatedly reported as doing well, but it has now been taken up because it was too light, and no more has been laid down. On the South-eastern line, Samuel's sleeper, somewhat improved in design, has now lain above two years, and is perfectly satisfactory in all respects. The bearing is on the surface of the ballast, the base is more concentrated than with the sleeper-tie, and this plan has therefore proved to be easy to pack, to maintain a good position, and to admit of fair drainage. In 1850 and 1851, 200 miles of P. W. Barlow's cast-iron sleepers (Fig. 12, Plate 19) without the wood cushion, however, were laid down on different lines-100 miles on the South-eastern, with old rails. Nine miles of the sleepers required to be renewed in the first five years after laying down. The present pattern is made in two sizes, one 38 by 14 inches, weighing 137 pounds; the other 53 by 15 inches, weighing 182 pounds. The first would give 3-7 square feet of bearing per sleeper, the second 5-52 square feet. Laid to give 2 square feet of bearing per foot of track forward, would require of the first plan 1741 tons, of 2,240 pounds, per mile of single track. Of the large sleepers, 155 tons, of 2,240 pounds, would give the same bearing. These sleepers would be spaced, where laid down so as to give but 1, square feet of bearing per foot of track forward, requiring 131 tons of small sleepers or 1161 tons of the larger pattern. It will be seen that the bearing of the rail in the wood is very short, rendering the wood very liable to crush. The bearing surface on the ballast is further below the rail than in Samuel's plan, requiring more bearing per foot of track to give equal steadiness. This plan also introduces bolts and nuts, which must be, to some extent, an objection. M. De Bergue's cast-iron sleeeper (Figs. 1 to 9, Plate 30) is also in some use. This is more concentrated in its bearing than either of the others, being nearly equilateral, or 20 by 14 inches, weighing 46 to 56 pounds and giving nearly two square feet of bearing each, requiring but 100 tonls (of 2,240 pounds) per mile, to give -a bearing of 13 square feet per foot of track forward. With 20-feet rails, and blocks spaced 30 inches, centre to centre, this plan would require 22,124 pieces per mile, being 4,224 blocks alone. The sleeper is designed, as yet, only for the flatfooted rail. It was applied on the Great Northern line, in a siding, where the traffic was so heavy as to wear the rails out in 18 months. The sleepers were of but i inch thick castings, were 18 by 14 inches, and spaced 2 feet apart centres. Whlen the rails were renewed, one-fifth of all the sleepers were broken. The thickness has been successively increased to g inch, and is now down of that thickness on the London & South-western and the Lancashire & Yorkshire railways — for small lengths only, amounting altogether to a few miles. It requires to be packed with fine sand, but the engineer of the South-western railway states that it is easier packed, and preserves its bearing better than any other plan of permanent way in use on that line. The castings require some care in cooling them at the foundry, but so far, on the South-western line, scarcely one per cent. of the whole number laid down have broken. Figs. 13 to 2B6, inclusive, of plate 30, show modifications of the cast-iron system. These involve the use of more or less wood, as cushions or as keys. Fig. 5, plate 31, shows Greaves' Cast-iron Spheroidal Sleeper. This is in use on a few SLEEPERS. 165 English lines, and has been laid down to a considerable extent on foreign railways. It is difficult to get precise and reliable information, sufficient to decide its character. It has been disapproved of by some engineers who have tried it, from not having sufficient stability, from being difficult to pack, and from beng somewhat rigid. It is not necessary to trace further the application of cast-iron for sleepers, but it does deserve remark that these plans are found to take a better bearing, and require less labor in packing than the cross-tie system. Upon this point there is no division of opinion or experience. Stone blocks had disadvantages in their rigidity, in their own weight, by which they sunk in the ballast, and in the depth of bedding which they required-altogether sufficient to countervail their advantages of form, and to throw them out of use. With the plans of cast-iron sleepers, it is found that 1- square feet of bearing, per foot of track forward, answers as well as two and a half square feet of nominal bearing of cross sleepers. If in consideration of the high price of timber, and the advantages heretofore indicated, the English engineers should use blocks of preserved timber, the bulk of sleepers would be reduced from 5,500 cubic feet per mile to 3,300 cubic feet, a reduction of 40 per cent. If the mere disposition of a smaller quantity of wood into a better form can afford advantages in the permanence and good condition of track, and in labor and expense of maintenance, all railway companies are interested in the fact. Zonyituudinal System of the Great WFeslern Railway. —The longitudinal system of the Oreat Western railway, shown by plate 7, has been variously modified from time to time. The stringers have been tried experimentally 10 X 10. inches, and in deep strong ballast they make a good road; but the standard of main-line timber has long been 1 5 x 7 inches, with transoms, 7 X 5 inches, at 11 ft. apart, framed with strap-bolts. The longitudinals are always laid to break joints, and the rails do so too, not only with the timber by an overlap of at least 3 ft. 6 inches, but also with each other. The longitudinals when well bedded and settled in the ballast do not roll; they have 21 feet of bearing surface per foot forward, which is the same as in cross-sleeper roads. The rail is about 10 inches above the bottom of timber. The great and excellent principle of the longitudinal system is that it is packed continuously, and gives a bearing immediately under the load. There are no cross sills or other pieces under the main timbers, the transoms or ties being on the same plane. But as carried out on the Great Western line, the longitudinal system has several disadvantages in respect of its own maintenance, although it is admitted to be easy on the rolling stock. As the rail peculiar to this system (the bridge rail) has been made deficient in vertical stiffness, and as the timber cannot compensate for the want, it follows, as is proved in practice, that the line springs on the passage of a train. And, in springing, the rail must crush into the timber. A practically stiff rail might bear up on irregular and remote supports, and might, if sufficiently stiff, lie smoothly on rough ground without ballast, but such stiffness can only come from the rail and not from the timber. If the rail " gives," the timber must give too, and crush at the same time. The bridge-rail used on the Great Western is estimated to have but fivesevenths of the stiffness of the same amount of iron rolled into the usual form. Hence, from this want of strength, and from the yielding nature of the wood when pressed upon by the flat foot of the rail, the timber has been considerably "' mashed " in, so that near the Paddington station, at London, cross-pieces of hard board have been interposed between the rail and the sill, so as to make a continuous floor for the former. The Great Western longitudinal system, for this reason especially, is rather more expensive as to maintenance than the cross-sleeper system. It presents a diffculty also in the way of drainage, for which it does not lie in the right direction. Being deep in the ballast and deflecting more or less under heavy loads, it forms longitudinal channels in which the water is likely to collect and remain, unless the amplest precautions are taken against the evil. There is also the difficulty of removing the sills when they require to be replaced —especially the delay, which is sometimes very inconvenient. For the stability of all plans of superstructure and for the efficiency of drainage, the height from the underside of the sleeper, or from the bearing on the ballast to the top of the rail, is important. Plate 31, fig. 1, shows the old stone-block system, fig. 2, the present, and fig. 6 the 166 PERMANENT WAY. former Great Western system, while fig. 4 shows the ordinary plan of cross sleepers. The other figures will be referred to in another connection. On examining plate 31, however, the difference in stability and ease of drainage appears in striking contrast to those first referred to. The Great Western railway is one of the most expensive in England as regards actual cost of maintenance. This may be partly due to its wide gauge, its heavy equipment, and the preponderance of its passenger business over that of freight. The cost has been as follows for three years: 1854. 1855. 1856. Cost per mile of Road,.... $1,420 96 $1,799 55 $1,311 94 "'L" " Run,... 15-24 cts. 15-24 cts. 12-96 cts. This is considerably in excess of the cost of the same item for other English roads, excepting that for the south-eastern, which has some 100 miles of light cast-iron sleepers, and which has proved, on the whole, quite as expensive. RShe Sandwich System.~M. Mr. W. Bridges Adams put down, seven years ago, on the Eastern Counties line, a short distance of track wherein the rails were supported throughout their length by side timbers 7 inches wide, bolted through, the timbers forming a bed of 15 inches, testing on the ballast. The common 5-inch rail was used. There was no deficiency of vertical strength, and hence the plane of the rail was preserved, and the timber was not found to crush, there being no mortising or tenoning, and the timbers being on the sides of the rail there was no difficulty in removing them when necessary for renewal. And, being comparatively shallow in the ballast and not deflecting sensibly, the difficulty of drainage was diminished. This plan, called the Sandwich system, and now more largely used, overcomes the disadvantages of the ordinary longitudinal system. More recently, Mr. Adams has modified this plan so that it might better embody the peculiar advantages of the system; and he and others have adapted it to the various circumstances of railway working. Some of these modifications are shown by plates 5, 21, 22, and 33. It is, substantially, a deep but light rail, suspended by the head between continuous longitudinal sleepers on either side, the sleepers being bolted together through the rail, every 3 feet, and the two rails with their sleepers, being tied together by iron rods or wooden ties, every 6 to 8 feet, to preserve the gauge. The side timbers may be a little shorter than the rail, as at fig. 4, plate 5, thus allowing room for a separate fish-joint and preventing the necessity of separating the rail and the timbers when the rail is to be reversed. Or the side timbers may break joints with the rail and with each other, the fish-piece being enclosed by the same bolts that bind the timbers together, as at fig. 5, plate 5. This is undoubtedly the better plan, as it saves extra bolts and preserves the continuity of the entire track. And long sections may be reversed at a time, which will prevent the necessity of disturbing many joints. The side timbers constitute the sleeper and rest directly on the ballast. This system embodies several distinct and important principles; viz.: The rail is supported immediately under the head, the load being taken off through the side timbers, whereby the central web of the rail is relieved from all lateral strain which would cause it to buckle, and (serving only as a stiffening keel), can be much deeper and thinner than the web of a common rail, which must sustaian the entire action of the passing load. The rail in plate 32 weighs 70 pounds per yard, is 8 inches deep and $ inch only in thickness. This thin, deep rail has double the vertical stiffness of one 3.- inches deep with a 3-incl stem. The rail in plate 5 is 5- inches deep, with a -inch web, and weighs but 60 lbs. per yard. Thus these deep rails will not "bend in detail" or successively as the wheel rolls along, and will not, therefore, crush the timber. That this bending in detail actually goes on with low rails, the experience with the ordinary longitudinal system as well as the universal experience with the cross-sleeper system, fully proves. The bending crushes the timber, where a practically stiff rail would distribute the pressure over an extended surface. It is only by suspending the rail by the head and clamping it closely at the sides that it is practicable to employ a rail so deep as to be absolutely stiff under all loads. The side timbers also increase the stiffness of the rail about one-half. That the rail does not crush into the timber-even a common 5-inch rail (Plate 28), with a slight bearing surface-is abundantly proved by the fact that the wood and the iron absolutely rust SLEEPERS. 167 together, and must be pried apart in order to be separated. The author has witnessed an experiment of this kind on the North London line; similar results have been published by other observers. The timber suffers no mechanical depreciation whatever, after seven years' service under a heavy and constant traffic on the Eastern Counties, North London, and other lines. One set of timbers has been known to wear out two sets of heavy double-headed rails, equal to four sets of singleheaded rails, and it is now attached to its third set. Therefore the timber may be preserved from decay by any process which does not decidedly weaken it, with peculiar advantage. Cross sleepers, on most American lines, are mashed and mechanically destroyed long before they begin to decay. The whole depth of this superstructure would be but 5 to 8 inches-the depth of the rail itself; whereas,, parts of the Great Western track are 13- inches deep from the top of the rail to the bottom of the sills. The stability of the superstructure is in proportion to the coincidence of the plane of support with the surface of the rail. The longitudinal system, as usually laid, rocks laterally, owing to its height above the support. Adams' plan is but little more than one-half as high. (See Plate 31.) The motion of the trains over a sandwich track is extremely easy and-uniform. There is a feeling of smoothness but not of hardness and rigidity. The rail has a uniformz and elastic bearing. There is, substantially, the same amount of metal at all points. Rigid joint sleepers by the side of flexible intermediates and anvil chairs and joint fastenings are out of the question. There can be no injurious rigidity. Since the stiffness of the structure insures smoothness in the entire tread of the rail, the mere elasticity of the wood is sufficient to relieve all the jarring common to a less smooth but more rigid system. The timbers being bolted on the sides, can be easily examined, and, when necessary, easily renewed; a great advantage over the ordinary longitudinal system. While sufficiently deep in the ballast to be well supported in place, this system is not so deep as to give trouble in drainage. The ordinary longitudinal system, 13- inches deep, lying in channels of that depth, is difficult to drain, and the cross sleeper, if it has ballast enough' at its ends to afford any adequate support, forms quite as deep a channel as the sandwich sleeper, which may, indeed, lie almost on the top of the ballast. The uniform depth and quality of ballast required between the rails, and the extreme width of ballast required for the cross-sleeper system are unnecessary for this. The whole of the bearing is immediately under the rail. In proportion as the superstructure is deep, so must the quantity of ballast be increased. This plan would require the same depth of ballast, beneath the sill, as any other system, but above the bottom, three to five inches could be used with Adams', where the ordinary cross-sleeper road, as made in England, requires a foot. This is obviously the cheapest system known. The leading fact about it is that by paying for one rail, two are secured. It gives two rails for he _price of one rail of the ordinary weight. When the top is worn out, the track is simply reversed, and a new head is ready for service, having been preserved in the ballast, and not worn and notched by resting in chairs, as is the case with double-headed rails employed in the usual English fashion. (Plate 28.) Less timber is required than in the cross-sleeper system, because every part of what is employed, is utilized. The sleeper is a bearer under the load and not at the end of a lever. And less ballast is required, as has been already shown. A modification of the wrought-iron 7-inch beams now made, could be applied with this system. These beams now weigh 60 pounds per yard, and even this iron, suitably disposed, would fulfil all the conditions of strength. Plate 5 shows a full-sized section of a 60-lb. rail with 4 by 5 inch longitudinals. The same timber is employed for cross-ties to preserve the gauge. Plate 32, fig. 1, shows a section of the rail and supports, taken at the joint. The side timbers average 71 by 71 inches, and give 23 square feet of bearing per foot of track forward, and take a little more timber than cross sleepers giving the same bearing. The side timbers are shown as fastened with wooden treenails and keys. At the joints, a pair of dishing plates, shown in section, and also by dotted lines in the plan (Fig. 2), preserves the stiffness at that point. At all other points in the length of the rail, as in fig. 3, the wooden balks clamp the sides of the rail. 168 PERMANENT WAY. Fig. 8 of plate 21, shows another of Mr. Adams's plans. An 8-inch rail -- web, 7 5 pounds per yard, and supported by timber 61 by 7?, dressed to shape. A cast plate, recessed in the timber, is used at each joint. Size of plate, 18 by 6 inches, by - inch. Weight 22 pounds. The principal object accomplished by this plan is the elevation of the tread of the rail so far above the timber that room enough will be left for the flange of the wheels without grooving the timber as shown on plate 5. Mr. Adams's estimate for this plan is: (the English plan having double-head, with chance to reverse; the American rail having single head.) Bearing Surface on Ballast,. 13,200 square feet per mile, single.'" " of Rail on Wood,.. 2,200 " " " Cubic feet Timber, per mile,... 5,867 Vertical stiffness of Rail,... 64 Weight of Iron,...... 132 tons 9 cwt. One mile of American cross-sleeper system. Ties 2- feet centre to centre. Bearing Surface on Ballast,.... 13,362 square feet per mile, single.'': of Rail on Wood,... 704 " " " " Cubic feet of Timber, per mile,... 6,336 Vertical Stiffness of Rail,.... 18 Weight of Iron,. 108 tons 3 cwt. Fig. 7, plate 21, shows another of Mr. Adams's modifications, with a 6-inch rail, ~ inch web, 40 pounds per yard, supported in two balks, averaging 6 by 41- inches. There are a pair of wrought-iron joint-plates at each joint, I foot long, and weighing 3 pounds each. This plan gives 10,560 feet of bearing on the ballast per mile of single track, or 2 feet per foot of track forward. The rails have 2,338 square feet of bearing surface on the sleepers per mile. The rails per mile, are 62 tons, 17 cwt., and the timber 4,400 cubic feet per mile. Plate 33 shows one of Mr. Adams's plans for forming a stiff foundation with plank. Crosspieces of plank, 16 inches long and 2- inches thick, are placed under the whole length of the rail. Every 6 or 9 feet, one of these planks is carried across to the other side, to tie the track together. Beneath the planks is a keel, formed by lengths of planks, 7 by 2- inches. The rail is rebated into the upper planks, and the whole structure is held together by staple bolts, as shown: castbrackets are used at the joints of the rails. The plan shows a 40-pound rail, or 62 tons 17 cwt. per mile. There are 4,470 cubic feet of timber, giving 14,080 square feet bearing on ballast, and 2,640 square feet of bearing of rails on timbers. The upper timber, when decayed, could be easily slipped out at the sides. The whole would afford much vertical stiffness, the bearing surface would be near the rail, and the structure would be elastic. Another plan is shown on the same plate, in which T-iron ribs, 25 pounds per yard, take the place of the vertical planking. This makes equal to a 65-pound rail, with 3,150 cubic feet of timber noer mile, with the bearings as above. Either variety of the Sandwich system (Plates 5, 21 and 32) may be considerably simplified, if a superior quality of iron is insisted upon, by making the rail as nearly right angular under its head as it can be rolled, so as to prevent the necessity of dressing the timber to fit the shape shown at plate 5, for instance. The groove for the flange may be very cheaply made by fitting a circular cutter in an ordinary wood planing machine and passing the timber through in the ordinary manner. It has been objected, that this groove will fill with ice in winter, upon Northern roads because ice occasionally forms entirely across the road-bedbetween the rails, to a height which sometimes carries the wheels on their flanges and causes them to run off the line. Now, since the timbers would not necessarily touch each other within an inch or more, at their ends, any rain or snow water would run out of the groove. It is possible, however, that the groove might fill with sleet which would gradually turn to ice, and thus lift the flanges. This can be only settled by experiment. The timbers may be trimmed to the line A B, plate 5, or the plan shown by fig. 8, plate 21, might be adopted, should this prove a serious objection. The last-named plant raising the tread of the rail high enough to clear the flange-requires a little more weight of SLEEPERS. 16 9 metal, but it most effectually prevents trouble from ice, and it obviates the necessity of trimming or grooving the timbers at all. The only other objection to the sandwich system which has been mentioned, is that it will not last well without ballast. We never heard of any system that would. It is indeed true that a springy, mud-and-water roadbed absolutely requires a superstructure that will yield and accommodate itself either to churning the earth in wet weather, or to following the undulations it may have assumed when caught by the frost. On such a foundation the superstructure must either be light enough to wring and twist without permanent set and strain, or else it must be provided with hoggingr ames like a North River steamer, so that it may transfer its bearing a hundred feet each way, till it can find enough bottom to carry its load without deflection. And as road-beds approximate to this condition, they are likely to strain a stiff sandwich rail more than they would strain the common American deflecting chain track. Still, it is not probable that any such disadvantages, on an ordinarily good American line, would offset to any appreciable extent, the decided and signal advantages of easy traction and the double life of the rail. A little ballast, however-half what would be required for a first class cross-sleeper track,-would add to those already named, the crowning benefit of cheap maintenance. The practice of tampering with standard improvements, and modifying and mutilating them till their valuable features were entirely lost, and then pronouncing the original a "humbug," has been illustrated in the first part of this work. Here is another chance for similar results; and we expect to hear the sandwich system condemned by some "shrewd railroad men," who will make an utter failure of some essentially different and absurd sandwich system of their own. To those who really wish to avail themselves of the long and successful practice in England, in this matter, it is hardly necessary for us to repeat the fact that a deep rail, though not necessarily a heavy one, is absolutely necessary to the full success of this system. A shallow rail, supported under tlie head only, will deflect under each wheel, and rapidlymash the timber, loosen its fastenings, and go to pieces. A deep rail is literally the backbone of the whole system. And it will, in all'cases, be found highly economical to insure at least a foot of good clean ballast under the sleepers. There is no mistake about the excellence of this system. It is peculiarly adapted to this country, as it furnishes an excellent track without the excessive use of ironllwithout requiring any more iron than is now employed in a second class cross-sleeper road. And, in this country, it is patent to the public. Its advantages may be briefly reviewed as follows: 1st. Its first cost is less than that of a medium cross-sleeper road, that is to say, a good American road. 2d. Its cost for maintenance can hardly be half that of maintaining a good cross-sleeper line. 3d. Two rails are furnished at the price of one. 4th. The firm, non-deflecting and uniform tread of the rail, decreases the cost of traction in like proportion. 5th. The maintenance of rolling stock is decreased in the same degree, for similar reasons. It should be further remarked, that the joint-fastening of a deep girder rail, like either of those contemplated or used in the sandwich system —that necessarily vexed question which can never be satisfactorily settled in case of shallow rails —is fully provided for. A fish-joint, deep enough to preserve the stiffness as well as the continuity of the rail, and situated between the nearly parallel tables of such a rail as that on plate 5, so that it shall require but little pressure to keep it in place, certainly fulfils every attainable condition of a thorough joint as to lightness, cheapness, strength, durability, and elasticity. Thus circumstanced, the fish-joint requires but slight fastenings, since the strain is almost solely a vertical strain on the fish-pieces themselves, as girders resting on firm supports, and a tensile strain on the web of the rail. Andl when inclosed, as at fig. 5, plate 5, two extra bolts, at most, are all that are required to hold it in place. Again, the only difficulty ever experienced with a deep fish-joint, is the loosening of the nuts; in this case, the elasticity of the timber is found to keep the nuts tight, by constantly straining the boltheads and nuts apart, and creating a degree of friction against which they cannot easily turn. Dimpfel's Longitudinal System. —One form of this system is illustrated by plate 36, fig. 13. It is a double or single-headed rail, supported both under the head and under the bottom flange by 22 170 PERMANENT WAY. continuous longitudinal sleepers. With a deep double-headed rail, the extra depth of timber is quite unnecessary for both stiffness and for bearing of the rail on the timber. Whatever may be the theoretical opinion of American practitioners who have not observed the working of Adams' sandwich system, the fact is that the small bearing under the head, in case of a deep rail, is quite sufficient. The rail and timber would hardly rust together if there was play between them. Therefore, any extra depth of timber not only requires more material, but embodies the serious defect before mentioned, of top-heaviness. And if top-heaviness is prevented by deeply imbedding the sleeper, good drainage will be impossible. But while Dimpfel's system is not likely to be a successful competitor of Adams', it'may still be a great improvement over the cross-sleeper system with the common rail. The modification or adaptation of it, by Mr. Colburn, to joint sleepers (Plate 26, Figs. 15 to 17), has already been mentioned. The deficient stiffness of a low rail is compensated for by an increase of bearing. A common American rail, then, which makes a moderately smooth and uniform track, on cross sleepers, would be certain to answer much better on at least four times the bearing. The only serious objection to the plan for the ordinary pattern of rail, is the height of the'rail above the base of the sleeper; but this is not likely to offset the decided advantages mentioned. It is still an improvement upon the Great Western longitudinal system, and its modification on some of the Southern lines in this country. Seaton's Longitudinal System. — This is illustrated on plate 23, figs. 24 to 26, also, for street purposes, on plate 6, fig. 11; and it has been long enough in use to prove its advantages over the cross-sleeper system. It has some obvious features of simplicity over the sandwich system, but these are not necessarily advantages, because Seaton's system requires a heavier rail, and therefore a larger first cost than Adamns', although when laid as it should be, it is unrivalled in surface, gauge, and economy of maintenance by any permanent way the author has ever observed. A 90-lb. rail has been employed with this system on parts of the London and North-western and the Great Western lines since 1857. The saving in first cost, over a first-class English crosssleeper system, with an equally heavy rail, is said to be about $500 per mile. The cost of maintenance is estimated from data which are believed to be authentic, as one-third less than that of the best cross-sleeper system. Upon examination, in 1859, after the passage of about 100 trains daily, for two years, the rails showed no abrasion at the joints, while the fastenings of the rails to the joints and of the longitudinal timbers to the ties appeared absolutely tight, and did not appear to have been repaired or touched since the rails were first laid. The ballast was not superior to that on several of the best American roads. Upon removing one of the rails from the timbers, the creosote on the latter did not appear to have been worn at all; the rail had taken an even bearing on the whole of the timber which it covered, and had neither moved upon nor crushed into the wood. The screws were all tight and the joint fastenings firm. In addition to the general excellence of the longitudinal system, this modification of it owes its peculiar advantages to the following features: The sleepers have the greatest possible bearing on the ballast, 17 inches width being standard; the whole structure is so low and has so broad a base that it has no tendency to rock; for this reason it need not be deeply imbedded in ballast, and thus interfere with drainage; not only vertical, but lateral support is given to the rail by the sleeper itself in virtue of their respective shape, and independently of the other fastenings, which are merely to keep the rail from jarring out of place. The great feature on which these systems depend for their excellence, is evidently, so much stiffness in the rail itself, that it will not bend "in detail," or under each w:heel, and thus mash into the timber. The unworn creosote on Seaton's timbers, and Adams' sleepers and iron rusted together, are absolute proof that no movement of the rail occurs —thlat it does not bend enough to mash into the timber, but only springs zwit the timber, just enough to prevent jarring. The rail transfers its load. to so much surface oftirber that there is no crushing pressure at any point.. But if the rail is so light as to yield at any one point more than the elasticity of the timber will allow, the timber must be crushed, and the chief advantages of the system are sacrificed. While Adams' rail may be very light and still sufficiently stiff, on account of the thin deep web admissible, Seaton's entire rail must be proportionately heavy. It does not make the most economical use of iron, however simple may be its fastenings. And for reasons already SLEEPERS. 171 mentioned, it requires a fair amount of ballast, and would hardly do so well on a mud bottom as flexible straps of iron floated on cross sleepers. These points have been especially referred to id view of the fact that several American managers, still bent on the suicidal policy of saving first cost, are proposing to lay a mere angle-iron after Seaton's method, affirming that so long as the sleeper has bearing enough, 50-lb. iron is sufficient. It would reasonably appear, in consideration of the principles and results mentioned, that the longitudinal system of permanent way, probably in the form of Adams' sandwich plan, or, if sufficient first cost will be undertaken, of Seaton's plan, will eventually supersede the cross-sleeper system; and that it will not only decrease the cost of traction and repairs, but happily settle the now vexed question of rail joints and fastenings. It is to be hoped that discredit will not be thrown upon these plans by the failure of " improvements " and modifications designed to save first cost. We know of no cheaper system, fit to be used, than Adams' sandwich system. To construct cheaper permanent way than we now have, with the expectation of saving in the long run, is perfectly unreasonable. CHAPTER IV. RAILS. IRON AND MANUFACTURE.-Iron made during the past ten years has generally proved quite inferior to that made previously. In 1835, the rails for the Stockton and Darlington Railway were made to the following specification. Best No. 1 cold-blast mine iron was, first, run out in a finery fire; second, puddled, and the balls shingled under tilt hammers; third, rolled into bars; fourth, these bars were cut, piled, heated, and hammered into blooms; and, fifth, these were reheated and rolled into rails. For the Leeds and Selby Railway, the iron was made nearly as above, viz.: Best No. 1 cold-blast mine iron was puddled and shingled into a bloom, which was cut and rolled into a best bar. This was again cut up and rolled into a rail, without any intermixture of inferior iron. Such rails sold from $2 50 to $5 00 a ton above the prices of " best " bar-iron. Most of the rails made during the past ten years have been made of hot-blast cinder iron, the tops and bottoms not always reheated, but rolled directly from puddle bars into rails; recently, however, both English and American engineers are turning their attention to a better quality of iron.* The rails for the Royal Swedish, and for the Bombay and Baroda lines, are made much in the same manner. The tops and bottoms, of reheated iron, are each made from a pile of puddled bars-the central bars being merely of puddled iron. The pile for the Baroda rails is -, by 1 2l inches. That for the Royal Swedish rail is 9 inches square. In both cases, the central bars of the pile are to go through, that is, to be of full length. The tops and bottoms are sometimes, however, specified to be 2 inches longer than the others. The size of the piles is to be particularly noticed. The following table will show the relative reduction in rolling. Area of Reduction RAIL. Wreight of Rail. Size of Pile. Area of Pile. Finished Rail of Section. Pile in Rolling. Lbs. Inches. Square Inches. Square Inches. Baroda Rail,.... 66 71 X 121 93'75 6-9 13-6 to 1 Eastern Counties,.... 65 9 X 9 81 6-8 11-9 to 1 Royal Swedish,... 62 9 X 9 81 6'5 12-'5 to 1 American Rail,.... 69 6 X 6 36 7'2 5 to 1 * The following is extracted from a specification made by the Eastern Counties Railway Co., some two years since, for 4,000 tons of new rails:-" Each rail is to be made from a pile 9 inches wide and 9 inches deep, consisting of one bar of iron 9 by 1- at the top and bottom. The intermediate bars not to exceed I inch in thickness, and to be alternately 6 inches and 3 inches wide, so as to break joint. The pile is to be rolled at a welding heat into a bloom 5 inches wide and 6 inches deep, which being again raised to a welding heat, is to be rolled into a rail. The bars, 9 inches by 1i inches, which form the top and bottom of the pile, are to be manufactured from such a mixture of ores (being all mine iron), as shall produce the closest and hardest wrought-iron, and shall be drawn from the puddle ball under a hammer, which shall be equal to a five-ton tilt hammer, into a slab 9 inches wide by 2 inches thick, which slab shall be heated sufficiently for its reduction to tlhe required thickness of 1 inch. The bars, of thickness not exceeding 1 inch, for forming the central part of the pile, are to be manufactured from such a mixture of ores (being all mine iron), as shall produce the closest and toughest wrought-iron, and shall be drawn from the puddle ball under a hammer, which shall be equal to a five-ton tilt hammer, into a slab or bloom of convenient form, the sectional area of which shall not be less than 20 square inches, which slab or bloom shall be reheated sufficiently for its reduction into bars of the required thickness, not exceeding I inch. The use of cinder or cinder pig will not be permitted under any circumstances. The rails are to be dipped while hot, in hot linseed oil, and are to be perfectly protected from the weather until this is done. Should any of the rails laminate, break, or otherwise fail, within a period of three years from completion of order, the company will, at their own expense, take such rails out of the line, and the contractor shall be bound to exchange them for an eqlual quantity of sound rails, to be delivered when required, firee of all cost, at the cormpany's wharf, Blackwail, (rl the Petcl:crboroonli station." RAILS. 173 Thus the improved English rails receive from double to nearly three times the working given for the standard make of rail now in use in this country. The American rail is rolled direct from a pile 6 inches square-the improved English rail is rolled from the pile into a bloom of say 6 inches square, which is subsequently rolled into a rail. Within the last year, the Eastern Counties Railway Company has ordered rails to be guaranteed to last seven years, and to be manufactured in accordance with the general principles just mentioned. The price of these rails was a little over $56 (~11 10s.) per ton; meanwhile rails were selling in England for a trifle less than $27 (~5 10s.) per ton. Only a mill with heavy rolls and ample power, can work piles from 9 to 10 inches square, but with piles of that size, the rails would have more working on them and the iron would be made more dense. The same result would be approximated, however, by working a smaller pile into merchant bar, which might be cut and piled to make the rail. In the latter case, however, there would be a risk that the iron would not have enough cinder to weld well. The puddle-balls have been usually squeezed, in English rolling mills, by means of " alligator squeezers." These have comparatively little power, and hence do not shingle the iron thoroughly. The rotary squeezer not being in use in England (to any considerable extent) the tilt or steamhanmmer is much resorted to, for shingling the puddled iron. The hammer is, undoubtedly, much better than the " alligator," but is not so efficient as the rotary squeezer. The hammer requires considerable time to work the puddle-ball, during all of which the latter is cooling. The ball, when first taken from the puddling furnace, is below a fair welding heat, and thus the hammer does not have its due effect. The rotary squeezer does its work in a few seconds, and acts equally upon all parts of the ball, thus clearing it of much cinder. During observations made at a large American rolling mill, the puddled bars shingled by a hammer, showed a waste of 12 per cent. from the pig metal from which they were made. The bars, made from squeezed blooms, showed a waste of 14 per cent., showing that the squeezer had expelled 2 per cent. more cinder than the hammer. In re-eating this iron, for covers for rail-piles, the bars from hammered blooms showed a further waste of 7 per cent., while those from squeezed blooms wasted only 5 per cent. Charcoal blooms are too hard, however, for a squeezer, and must be shingled by a hammer. It will be observed that the wearing out of rails consists largely in the peeling of the top stratum of iron from the rest of the mass. Upon examination it will be found that there appears not to have been a sound and thorough weld between the strata. Since we know that a bar of superior iron is generally used for the top of the pile from which such rails are made, while the rest of the pile was of an essentially different quality, we only observe a result which might have been predicted from the beginning —that metals which are not homogeneous cannot be rolled into a solid mass. There appears to be quite as much necessity for good iron in the interior as in the exterior strata of a rail-pile.' Much attention is being paid to the selection of the pig iron for conversion into rails. Castiron varies in strength from 9,000 to 45,000 pounds per square inch —a wide range, indicating the necessity for a careful choice of material. Good pig, honestly and thoroughly worked, will make a good rail and an additional expense of $18 a ton in the manufacture of rails will give a quality that will wear twice as long, and hence be worth twice as much ($60 X 2=$120) as the ordinary iron of the day. WEIGHT OF RAILS.~The lighter the rail, the more thorough will be the working in it, for the reason shown in speaking of the sizes of rail-piles. The preference is now almost invariably in favor of lighter iron, as the heavy rails wear out soonest. It cannot be, however, because the rails are heavy, per se, that they go to pieces so fast —for the resistance of iron to wear and blows should be as its weight —for a similar quality in each case. It must be chiefly because lighter rails are better made, that they endure longer than heavy iron. The elasticity of a light rail, however, has something to do with its durability. The 45-pound rails, made by the Ebbw Vale Company in 1837, for the Philadelphia & Reading railroad, have stood a wonderful wear. They were made similar to those already specified for the Stockton & Darlington line their manufacture being superintended by Solomon W. Roberts, C. E., now of Philadelphia. The 64 and 68-pound rails, since laid down in their place, have gone 174 PERMANENT WAY. to pieces with less than one-third of the same wear. The old rails-perhaps partly crystallized by their long use-and certainly worked dry of cinder in re-rolling, do not show their original superiority when reworked into heavy bars. In 1854, rails of 85 to 100 pounds per yard were considered by English engineers to be the best. Since that time, it is found on the Eastern Counties line, that the 95-pound rails made the worst road, were less durable, and in course of time became the most dangerous-as compared with 75-pound rails. The London and North-western officers report that there are many more failures by breakage and in other ways, with the 82-pound rails than with the former 56-pound rails. None of the 62-pound rails, laid in the Grand Junction line, in 1837, were taken up until after 1.849. The 80-pound rails then put down in place of the few removed, showed more wear after from twelve to eighteen months' use, than the old rails after twelve years' wear. The North London line, perhaps the heaviest worked in England, is having laid down 72-pound rails in place of those of 85 pounds, formerly used. The last rail-a n pattern-for the Great Western line, is 62 pounds per yard, in place of 72 and 92-pound iron, formerly used. The Southeastern new rails are 65 pounds per yard-the old rails, 75. The Eastern Counties last pattern is of but 65 pounds, where one of 95 pounds was formerly used. The New York & Erie road has had rails of 58, 60, 63, 65, 68, 72 and 75 pounds laid down in place of 56-pound railsthe heavier iron almost invariably proving inferior. The most favorite patterns of rails on Continental roads, are of 62 pounds per yard. In France, the heavy English patterns formerly used are being replaced by lighter iron. Mr. Bidder has remarked, that the only way to get good rails was to contract for 85 pounds per yard, and to then arrange with the manufacturer to supply a better quality, of 75 pounds per yard for the same gross sum. Mr. Strapp, Engineer of the London and South-western line, was offered, in 1857, new rails delivered in London, at $29 28 a ton, while within twenty-four hours afterwards he was offered $26 20 a ton for the old rails he had taken up after 19 years' wear. An American manufacturer, whose experience has peculiarly fitted him to judge of the proportion of rails, when lately asked into whatform he would dispose 70 pounds of iron per yard, replied, " I would use no such amount in any form." The true course is in the use of the best iron, in moderately light rails. The costper ton is increased but by a comparatively small amount, while the durability may be doubled. One advantage of a light rail is, that its very size insures more thorough working being expended upon it. So, in all the lighter rails ordered for English roads, the quality is provided to be greatly improved. RBut recently, since American railway managers are getting into the fashion of using light rails there is great danger of going to extremes. By decreasing the height and hence the vertical stiffness of rails to save weight, the bars bend in detail under each wheel, which greatly increases both the lamination and wear of the iron, and the resistance t tthe rolling of the wheels. To decrease the stiffness of the rail would be fatal to such systems as Adams' and Seaton's, for reasons already mentioned. FORM OF RAILS.~The double-headed rail is adopted in England for convenience of fastening as well as for changing the wearing surface when one becomes worn. It entails the use of a chair on every sleeper, by which the cost of fastenings is much increased, and so, too, the bottom of the rail, bearing in cast chairs, becomes more or less indented, whereby it is often made unfit for reversing. Either table or head of the Engish rail contains but little more iron than is placed in the foot of an American rail, and this iron is not so favorably disposed for vertical strength in the thin, broad, as in the thick, narrow base. The resistance of wrought-iron to extension and compression, is as 90 of the first to 66 of the second. Hence, the head of a rail should be to its base (where it is not to be reversed) as 90 to 66. And the iron requires to be kept closely to the web, to afford its entire strength. At 3 inches from the web, it would do very little good umless, as in Mr. Brunel's 6-inch wide-base, to prevent crushing into the sleepers. With a rail of 2 —inch base, Mr. Barlow cut away E inch on each side, reducing the breadth to 1- inch, when the strength was apparently increasecd. This, of course, was due to the slight inaccuracy of the hydraulic press used to make the experiment, RAILS. 175 but it showed that the strength could have been but inconsiderably diminished. The flanges are brought into use only by their lateral adhesion to the vertical web. Thus, while the base must be wide enough to prevent crushing into the sleepers, say 4 inches at least —this width is only for lateral stability-not for vertical strength. But as the English rail obtains ample lateral support from the chairs, it can be and is made much deeper than American rails. A rail is merely an elastic beam, and the stiffness of rectangular beams is as the cube of the depth-doubling the depth increasing the stiffness eight times. The flanges on the rail somewhat modify this result. If a rail deflects, the wheels must roll up a grade. Professor Barlow found, in experimenting with the rails of the London & Birmingham railway, that the descent obtained by the deflection of a rail between any two fixed supports, gave no advantage of gravity, and that as much extra power was required as would be due to the ascent of a grade, equal to' the slope of the deflecting surface, for one-half of each bearing of rail, or, in other words, half of the deflecting slope for the whole length of line. That rails do deflect is proved by the oozing out of mud from under the sleepers in wet weather, and the continual product of dust in the same place, in dry weather. The experiments made some years ago, on the Camden & Amboy railway, showed the deflection of an ordinary rail, under the weight of a 20-ton engine, at passenger-train speed, to be equal to a 45-feet grade against the wheel. Professor Barlow found the deflection of a 50-pound rail, 38 in. deep, 3ft. bearing, under a load of 5 tons at rest, to be equal to a 25-feet grade (-084 in a span of 36 inches). This would be increased from 50 to 100 per cent. with the same load at rapid speed. The experiments of the Parlimentary Commission on the application of iron to railway structures, showed in one case, where the load traversed at 30 miles an hour, an increase of deflection of 150 per cent. over that observed with the load at rest. These were with the rail firmly supported by the sleepers. But the deflection of rails is as the cube of the distance between supports, and hence, if one sleeper should fail to give support, the span of deflectionswould be doubled, and the deflection itself increased eight-fold. If two contiguous sleepers should yield so as to give no support, the deflection would be increased 27-fold. With our shallow rails, in case of trains at rapid speed,* and the sleepers not always giving full support, a heavy grade is continually opposed to the passage of our trains. This must absorb much power, for, as has been mentioned, there is no advantage derived from the descent of the deflecting slope, above the loss by the irregularities, so that the ascent of the opposing slope represents a total loss of power. All that part of the deflection caused by the failure of the support of the sleepers, is chargeable to a bad construction or bad maintenance of road-bed. The remainder is due to the rail itself. The excess of power required to work the traffic of American as compared with English roads, corroborates this view. Our roads oppose the most resistance. Depth being so important an element of stiffness, the lighter T-rails now being put down in England, are of the full de ptU-5 inches-of the former heavy rails. The reduction in weight is in the head and stem. But if stiff rails are used, and they are not perfectly straight, or curved correctly in curves, and if the road-bed is not smoothly maintained, a fair amount of ballast being a condition of good maintenance, they will make a hard and rigid line, and will be rapidly split, besides doing much injury to the machinery. A light rail, once bent, will ease itself by its elasticity. A stiff rail, once out of line, will give and receive a blow. The Camrden & Amboy line was laid partly with 7-inch rails, 92 pounds per yard. They proved too rigid, while from their great weight and the difficulty in rolling them, it is probable that they were inferior in quality. Rails of 4-} inches depth are being now put down on the same line with much success. It is to be remembered that all the iron which has stood such heavy traffic on English roads, for from 12 to 20 years, was from 4- to 5- inches deep —mostly 5 inches. * It is perfectly established that the deflection of beams, when the load is in motion, is greater than with the same load at rest. Professor Barlow found that 6,500 pounds' weight on a single driving wheel of an engine deflected a 45-pound rail, 3 feet bearings,'12 of an inch when at rest, and'l177 inch, equal to a grade of 52 feet per mile, at 20 miles an hour. The Parliamentary Commission, on the application of iron to railway structures, demonstrated the same fact. The effect of passing rapidly over ice, so often instanced in this connection, is not to the point. Ice does not represent an elastic beam with remote supports. It is supported uniformly by the water, and time is required to displace thlis. 176 PERMANENT WAY. Depth is, among the considerations of form, very generally overlooked. On the other hand, there are certain popular opinions as to the form of rails which examination fails to support. One is that the steep " pear-head " is essential to strength, and another, that a very thin head and flange necessarily compel the use of the best iron. The only possible use of a heavy pear-head is to prevent the edge of the rail from breaking down. Any more iron than is necessary to prevent such failure, is thrown away, as the iron is not in a place to assist the strength of the rail. The steep pear-head of the Buffalo, Corning & New York road (Plate 16), and others, are rarely seen on an English road. If the iron in the- head is sound, a very thin head is enough to prevent breaking down on the edge. This is proved as follows: In plate 12 is shown the old Reading 45-pound rail, which stood anll enormous traffic for 20 years. The head, although very light did not break down. It was made of the best iron. The edge of the head, after use, permanently deflected downward, the section being taken from the worn rail-the dotted line showing the original form. This shows that there was an advantageous elasticity in the iron of the head of the rail. In plate 12 is, also, shown the Camden & Amboy old rail, taken after many year's use. The head, although very thin, did not break down as the iron was good. Plate 12 shows, also, the rail of the Boston & Lowell road, after having been run nearly two years on the bottom flange. It has since run nearly four years more, the trial having commenced in the autumn of 1854. The dotted line shows the original form when laid -and it will be seen that the edge has taken a set of nearly -j inch, without, however, breaking down. By referring to plates 11 and 28, it will be seen that, for very heavy English rails, the head is comparatively light-rarely averaging more than 11 inch in depth, the inner corner, where the head unites with the stem, being a curve of short radius. As a,general principle, the better the iron, the lighter may be the head of the rail, without danger of breaking off; and reciprocally, the lighter the head, the greater the probability that the iron will be well-worked and, consequently, good. It may be mentioned that many American roads are already availing of this principle. The Boston & Worcester rail of the last pattern, is light under the head, similar to the old Harlem rail, plate 25. The Michigan Central, new rail, is also very light under the head. The best rail ever laid on the Nlew York & Harlem road, as shown by plate 25, was put down 12 years ago, and has since sustained a very great wear. It will be observed that it was quite light under the head. Plate 8 exhibits several improved American Rails. That of the Cleveland, Columbus, and Cincinnati road was, together with Adams's bracket joint, taken substantially from the design made by the author, as shown on plate 15. This rail and joint, now in use on a fairly ballasted and drained road-bed, constitute one of the best sections of railway track in the United States. But, on the other hand, a light flange does coJmpel the use of good iron. Rails have been often laid out with very thin head and foot, with the express purpose of compelling the use of the best material. The designers did not know the character of iron, however. While good iron is likely to be improved by working it clown to small dimensions, a very deep thin flange can be easier rolled from cold-short iron than from tough red-short. And such a flange offers a temptation to use brittle iron. The security from cinder, sought by means of a thin flange, is quite balanced by the exposure to the risk of cold-short iron. They must kntow that the iron is origiginally good, and then put it in such shape as sh1all insure thorough working. The width of head of English rails is generally 2- inches — some are shown 2i inches, some 21 inches. Considering that the top is generally described with a radius of 5 to 54 inches (or just the height of the rail itself), it does not make so muclh difference what the width of the head is, as the edge receives a bearing only when the sleepers are considerably worn. With newly turned tires, the bearing is theoretically but a point, although there is practically such deflection that the weight of the driving wheels of an engine, rolled upon a piece of gold-leaf, incorporated a portion of it as large as a dime, into the surface of the head of the rail. Although rail-iron crushes under a pressure of about 8 tons to the square incl, and although a wheel can bear, tleoreticallly, on but a point on the rail, yet the elasticity of the metals in con RAILS. 177 tact may, without crushing, afford a bearing of measurable base. Suppose a 5-feet drivingwheel to bear on a geometrical point merely on a plane rail at a distance of half an inch on each side of this point, the wheel would be only'0041 of an inch clear of the rail. This distance is about the same as the thickness of a sheet of the paper on which this book is printed. Were the tire and the rail to yield, each, one-half of this-which both could probably do without crushing the fibre of the iron-a full bearing of one inch would be had; while an equal amount also would, probably, be obtained in the width across the rail, thus making a full square inch. Almost inappreciable as this compression appears to be, it would, nevertheless, represent a grade before the wheel, equal to 43 feet per mile (-5 inch. 0041=122); or 1 in 122. If the actual contact was for ~ inch on the rail, the resulting grade would be 22 feet to the mile. As the mere rolling-friction of railway engines and carriages is nothing like what would be due to the gravity on either of these ascents, it thus probable that the surface yields (to be restored by elasticity), under a very moderate pressure, and to a certain depth.'The tendency to further yielding decreasing rapidly with great comparative extension of the bearing acquired, the compression is not such as to oppose a decided resistance to the rolling of the wheel. This conmpression is quite independent of the deflection of the rail as an elastic beam-the distinction can be understood by hanging a weight by a thin wire to a wooden beam: while the whole beam will be deflected, the fibres under the wire will be compressed or indented in addition. Before concluding these remarks on form of rails, it should be observed that a high rail, moderately square under the head, is the best adapted for splicing at the joints. The ordinary American rail can never be well-fished. It is so shallow that when the splices are punched for the bolts, the greater portion of their strength is gone. The under side of the head of the rail, also, is so steep tha tthe splice cannot obtain a bearing sufficient to afford any considerable stiffness except by a strain on the bolts beyond their strength. Zentyg of Sails.-English rails are usually 18 feet in length. Many have been made as short as 15 feet and 16 feet. More lately, the length is being made 21 feet. On the Manchester, Sheffield & Linconshire PRailway, rails of 30 feet have been considerably used. The Rhymney Works once made a Barlow rail, 52 feet 6 inches long, for the Paris Exhibition, but it is not probable that such masses of iron can be either economically or soundly rolled. Several long lengths of rolled iron have been made in the United States-iron was rolled at Troy, for the Collins steamers, of over 60 feet length, from piles of over 700 pounds. Wrought-iron rafters have been rolled at Phoenixville, Pa., for the U. S. Capitol, 51 feet 2 inches long. Rails 30 feet long are now quite extensively rolled at several American mills. With the present appliances for making rails, piles of much over 500 pounds' weight are not likely to come out perfectly sound. Therefore, however desirable longer rails and fewer joints'may be, these results seem to conflict to some extent, with the soundness of the bar. RE-ROLLING RAILS.-Whether re-rolling old rails will result in a new product of good quality is. entirely circumnstantial. It can be told by working a sample of the iron and in no other way. This subject is of very great importance to railway companies, and yet the most contradictory opinions exist in regard to it. Some say, without any qualification, that old rails should work up into the best quality of new iron —others, that they are all utterly worthless for re-working into rails. The officers of many roads go so far as to accuse iron-masters of retaining old stock, sent them for re-rolling, and putting off their customers with raw iron. Now, the 7principles uponl which old iron is re-worked are perfectly simple, and if understood will often save nmuch money as well as hard feeling. The iron of commerce is never pure. It is an alloy of iron, carbon, silicon, sulphur, phosphorus, manganese, aluminium, etc. These foreign elements, in properi proportion and mixture, form the cinder of the iron —without much cinder the iron is "burnt," and can never weld. In pig iron, these elements are in excess —the carbon so much so, that the iron may be melted, as in ordinary foundry operations-whereas, when the iron is dlecarbonized by puddling, it only becomes naclleable under a high heat-ordinary " wrought" iron. lNot only is wrought-iron chemically the same as cast-iron, except in the proyortion merely of the foreign matters, butwrought23 17 PERMANENT WAY. iron, at every successive " working " or refining, while otherwise chemically unchanged, loses more and more of these foreign matters. Just in proportion as the cinder is worked out-or the purer the iron becomes —so will the iron be made softer and more fibrous, and at the same time become more difficult of welding. Iron with excess of cinder, although it is " raw," welds without trouble -provided there is-neither copper nor zinc in it, nor an excess of sulphur. On the other hand, while a single old tire will drawa down into the most excellent bolt iron, a pile of old tires, laid up in a heating furnace, would weld with some difficulty and irregularity, and if worked into a new engine-tire, would be quickly shattered to pieces. Rails rarely wear out —they laminate or crush in the majority of cases. Where they laminate, it proves that they are not thoroughly welded, although crushing may occur from too mzuch cinder, worked irregularly into the ironsand it often occurs from decay of cross-ties or bad ballasting. Sound welding of the rail-pile being so necessary to form a good rail, it follows that only such iron as has sufficient cinder, uniformly distributed, can be depended on to weld well. In re-rolling old rails, it is thus entirely circumstantial whether the product will be a sound bar —although the iron itself will be improved in any case, and as a bolt, or a tension rod, would be stronger than before. The whole distinction is that between goodfibre and good weld, and this distinction may be very wide indeed. Cast. iron has the least fibre and the best weld. The first thing to determine is, whether the old rail will weld readily. If the old rail was raw and cindery, it will be improved by re-working, and if worked enough, will infallibly make a good rail. But if, on the other hand, the old rail is of highly refilled iron, such as the original Ebbw Vazle, or " E. V." rails, it will not weld ill re-working, and the product may go to pieces in three months under ordinary traffic. This, although at first sight paradoxical, is really very simple. For while it amounts to the fact that an originally poor rail will, if not cold-short, make a good rail by sufficient re-working, and that an originally very good rail will make an inferior rail by re-working-the reason is clear -the first was not worked enough to secure a good fibre —the second becomes worked too much to. secure a good weld. Most of the very superior rails, laid down in this country above twenty years ago, when re-worked into new rails, without liberal intermixture with new iron, have failed for the reason assigned. On the other hand, it might be expedient, in the case of very cindery old rails, to heat and hammer them into blooms, to be re-worked into bars for the rail-pile. THE PHCENIX IRON COMPANY'S PROCESS.~A11 important improvement in the manufacture of deep rails and girders with wide flanges, is illustrated by plate 16~. A tough red-short iron is likely to be unsound when squeezed' from a solid square pile into the thin edges of a girder; for this reason, the pile is made to assume the external shape and the proper arrangement of fibre of the finished product. The top and bottom parts of the pile are separately formed; the bars intended for the web are then keyed into place, so that the pile shall hold together under handling, and the whole of the iron is then equally compressed between the rolls. Not only is the product made sounder, but its manufacture is cheapened by this process. And since the excessive squeezing required to change the general shape of a square pile is unnecessary, a heavier pile may be managed in the rolls, and longer as well as sounder rails may be produced. THE TUBULAR RAIL.~The full-sized tubular rail of Mr. E. W. Stephens, as made at the Crescent Iron Works, Wheeling, and the various stages of its manufacture are illustrated by plate 15-. It will be observed that, while substantially the ordinary pattern of; rail is preserved, it is lightened by the amount of iron left out of the head. By reference to the respective stages of the rolling process, it will be observed that the metal is more thoroughly compressed than the solid pear-head can be; more cinder is consequently worked out of it, and the metal is left comparatively dense, sound, and pure. The shrinkage of the hollow rail is observed to be 1} inches less, after coming from the rolls, than that of the solid rail of the same weight, for 30-feet lengths, which indicates the purity of the metal, to some extent, although the temperature of the tubular bar is less than that of the solid bar, upon leaving the rolls. Experiments farther show that the RAILS. 179 tubular rail has more vertical stiffness and more elasticity than a solid rail of the same exterior dimensions. The greater density of, the hollow bar would tend to promote its durability. This rail is now on trial; while opinions based upon some two years' use of it are generally favorable, it would be impossible to pronounce finally upon its merits. TIIE CONTINUOUS RAIL.-Various patterns of the continuous or compound rail are shown on plate 26. Its history on the New York Central Railway, where it has been thoroughly tested, and on other American lines, has been as follows:-The compound head of the.rail shown at fig. 3, plate 26, although presenting, for a few months, the best surface ever known on an American track, soon failed from the following causes; and when failure commenced, it was rapid and total. The wheels, worn concave on other rails with a narrower tread,* ran on the two sections of the compound rail alternately and did not always take a fair bearing on both; hence the wear was excessive. The lamination,of the inside edges tended to sunder the two parts; and the action of frost was to strain or break the rivets and to separate the sections of the rail. The next plan of continuous rail, shown by fig. 2, plate 26, remedied the first difficulty mentioned, by providing a continuous base, but the abrasion of the head and foot at these points of contact, in the absense of a rigid connection between them, soon destroyed the rail. But the wear and tear of machinery, according to the testimony of the engineers of the New York Central, was considerably reduced by the superior smoothness of this rail. How much, is not definitely stated. It was, therefore, deemed advisable to make farther use of some form of continuous rail. Mr. Winslow, of the Albany Iron Works, then introduced the plan shown by fig. 1, plate 26, which seemed to embody all the practicable features of the: continuous system, and which has been quite extensively employed, with results, which, though obviously superior to those of the common American system, are not reported with accuracy. There are, however, certain principles bearing on the case, which establish the advantages of the continuous rail for certain purposes. These have already been referred to in speaking of elasticity of permanent way and of the longitudinal system. The continuous rail embodies some of the advantages of the latter system. A deep rail call obviously be fished in a more durable and economical manner than by making the whole rail in two parts; but such a rail requires a well ballasted road-bed, for if supported by an uneven or yielding bed, its points of ultimate support will be so far apart as to cause the rail to permanently bend; thus it will be both rigid and rough-the worst possible condition. A low and more yielding rail, however, will adapt itself to what bearing it can get; being flexible to a greater degree, it will follow the undulations of an uneven road-bed, and it will at the same time be slightly elastic, becaues it will bend in detail or under each wheel, and it will not be both rigid and rough, like a deep rail on a poor road-bed. Either a flexible rail or a tolerably good road-bed, appear to be necessary to any degree of economical working. Of course there can be no comparison between the intrinsic value of these two plans-a deep rail on good ballast, and, a flexible rail on sand or soil, or whatever road-bed is conveniently made. But if railway managers either cannot or will not perfect the beds of their roads (and we are sure. that many of them would do so, if directors and shareholders did not insist on dividing all the money as fast as it is earned), the most reasonable thing they can do is to secure the best form of light flexible rail. But the first defect of the common low rail is quite as serious as that sought to be avoided by the use of a low rail, viz.: the want of continuity and stiffness in the joints. For reasons already mentioned, continuous and uniform elasticity is especially important on an indifferent road-bed; therefore, anvil-chairs, or large masses of metal or even timber at the joint, are certain to cause more rapid deterioration at the ends of the rails than at any other ~points. It is simply impossible to fish the joints of low pear-headed rails; any cheap method of preserving their stiffness, necessitates a large mass of metal and consequent rigidity. The mrost that can be done at small cost, is to preserve their continuity —to bring both ends dclown together. Therefore the common low rail cannot be economically jointed in any manner which will not compromise the result aimed at, viz.: uniform elasticity. The continuous rail is a compromise between these almost irreconcilable elements —continu- All wheels, in fact, will wear concave, without the tread of tile rail is as wide as that of the wheel. 180 PERMANENT WAY. ous stiffness and continuous elasticity. It does not preserve the entire strength at the joints, but it prevents much deflection, and it preserves the continuity of the surface, that is to say, both contiguous ends of the parts of the rail yield together, so that severe hammering by the wheels is prevented. And the weight is the same at all parts, so that the elasticity is uniform. The price of the continuous rail is $5 per ton above that of a solid rail of equal weight. This makes the joints cost about $1 each; and the history of joints costing no more than this, has not, we believe, proved them to be superior to those furnished by the continuous rail. HARD RAILS.- The first requisite in a rail is soundness. If this is secured, the head may be made hard and the durability increased. But there is a risk, in attempting to unite hard and soft irons, that the weld will not be secure. Different irons bear different degrees of heat. The hard iron may be at a welding heat in the pile, while the soft or fibrous iron is yet below its own welding heat,-or, while the latter is ready to weld, the former may be burning. In either case, the union of the iron is imperfect. Hard-headed rails have been found to split or peel off, on English railways-Barlow's saddle-back rail was especially troublesome in this way. Uniformity of iron is important to secure soundness. The Ebbw Vale works has turned out specimens of rails, made from steel produced by the Uchatius process. The quality of the steel is pronounced to be equal to that used for razors. The heads of rails are being artificially converted into steel by Dodd's process of case-harden-ng. This is being practised on rails for crossings and about stations. Switches are being very generally steeled in the same manner. The cost of converting rails by this process is about $8 per ton. It is not the.practice in England, to roll the steel into the rail-pile, as has been clone, to some extent, in this country, without much success, since the hard material peels from the soft, for reasons already stated. It is probable, however, that puddled, or semi-steel, possessing so much of the nature of iron; together with its steel qualities, can be rolled so firmly upon iron that it will not peel. By this process, the cost of steel-headed rails may not be greater than that of iron rails, since less material will have equal strength and stiffness. CONCLUSION.-The present movement as to rails'turns chiefly on a better proportion of bar and a better quality of iron. It has been customary to waste twelve tons of iron per mile, worth $700, under the heads of our rails, for no other reason than that such inferior iron was used as to crumble down on the edge. All the iron put in rails should be worked- to double the amount generally practised, and while the whole cost might be increased one-third, the wear of the iron would be fully doubled. Experience has proved this, again and again. English roads are taking up this reform in earnest, and some are paying above twice as much for their iron as the current prices of ordinary bars. CHAPTER V. R AIL JO IN TS. TEiis subject has already been referred to, in sonme detail, ill discussing the different systems of permanent way.* It is proposed to classify the various plans of joints and fastenings, and to refer to the principles and general results of each class. Since it has already appeared that the fish-joint ~a simple splice between the upper and lower tables of the rail-is at once the most effective and economical method of jointing deep rails, and that deep rails are a necessary feature of the best permanent way; and since most of the other varieties of rail joints, besides the fish-joint, have not been long enough in service to warrant definite conclusions as to their respective ierits, it is not proposed to go into a detailed discussion of their principles. The plates appended, illustrate quite fully the best practice as well as the recent invention and adaptation in this cde partment, and in the absence of definite and authentic results, the reader will prefer to draw his own conclusions from them. All attempts to make satisfactory and thorough work in jointing our low, pear-headed rails, must result in eking out an intrinsically bad system of permanent way. CLASSIFICATION OF JOINT FIXTUREs.-All fixtures at the adjacent ends of rails may be classified as simple chairs, splice or fish-chairs, and simple splices. The chairs merely give increased bearing on the sleepers (but rarely enough to prevent the mashing of the latter), and hold the rail laterally, but they neither transfer the stiffness of one rail to the next, nor preserve the continuity of the surface. They simply transfer the bearing of a projecting and deflecting rail-end to a single sleeper, which itself has not bearing enough on the ballast to support the load. The splice-chairs give the rails a seat, preserve their continuity, and may, but do not necessarily preserve their vertical stiffness. They also hold them laterally. Splices simply preserve the continuity of rails, and may or may not preserve their stiffness. 1. The simple chair, lilke that shown by fig. 6, plate 26, makes a very poor road and is veryextensively used because it is cheap. No chair at all would be cheaper and not much worse. This pattern of chair (Fig. 6) is sometimes made so heavy as to bring down both adjacent rail ends at the same time, or to preserve the continuity of the surfalce, when it fulfills the condition of a splice to a certain extent. 2. Splice-chairfs are of three kinds: 1st, Sleeve or longlipped chairs (Figs. 5 and 7, Plate 26), solid pieces not capable of taking up the lost motion caused by wear, and not stiffl without deep vertical webs or ribs. W7irhen long or resting up~on two sleepers, they hlave proved a better fixture for low pear-headed rails than most of the adjustable chairs, since they require no attention, remain tighter than ill-fitted bolts and keys, give so large a bearing as to prevent much deflection, and in any case allow only a slight vertical play between the rail ends. The device shown by plate 13, adds the advantages of the sleeve-chair to those of the splice; it preserves the continuity and may preserve the entire strength of the rail, without causing excessive rigidity. It will be observed that the brackets forming the splice are welded to the bottom piece. Three sizes of this joint are furnished; of the largest size, the side and bottom pieces are respectively 18 and 15 inches long; of the second size, 16 and 13 inches long, and of the third, 14 and 11 inches long. * See also the section on " Elasticity of Permanent Way," page 157. 182 PERMANENT WAY. 2d, Bolted chairs, of which many varieties are shown in the plates. Some of those patented here were long before patented in Europe. Their leading trouble is shaking loose. Without expensive fitting, nuts Till jar off. They therefore embody many devices for securing the nuts, but the longer English experience has proved larger and better bolts and nuts cheapest in the end. They are of several classes, those which gripe the foot of the rail only, and preserve its continuity but not its level, since the iron used is mostly in the neutral axis as to strength. Some have deep bottom-ribs and are very strong, but they bring the strain primarily on the bolts and nuts. Another class is a modification of the fish-piece, between the tables of the rail; it does not bring all the strain on the bolts, since these only hold it laterally in place, the web of the rail taking the strain of supporting the wheels. Both the fish and chair are sometimes in one piece, bolted laterally to the rail. The best modification has proved to be the Adams's bracket (Plate 15, Fig. 2), which gives a fish and a bearing at one or both sides of the rail, on the sleeper. A single bracket-joint, applied with a Z shaped rail (Plate 36, Figs. 11 and 12), has some obvious advantages. The bracket gives ample steadiness and lateral support to a deep, double-headed rail. The peculiarity of the spike head and the nut fastening will be observed. In several cases, the sleeve-chair and the fish-joint are used together (Figs. 2 and 3, Plate 21.). The Camden Amboy ring-joint, on a chair, and some of its modifications (Fig. 12, Plate 36); also raised sides of a chair to form a fish, and to carry the wheel over the break (Figs. 14 and 18, Plate 86), are in exceptional use. The wooden splice alone (Fig. 4, Plate 36), is valuable for lateral support of the joint, and better than nothing for vertical stiffness. One great advantage of wood, in connection with bottom or side plates or both, is, that nuts resting against it will not easily loosen, by reason of its elasticity. The joint shown at figs. 6 and 7, plate 21, is comparatively smooth, elastic, and durable, and is the best wooden joint in use. As ordinarily made, the bottom timber is 2 by 8 inches, and 21 feet long; the side timber 3g by 4. inches, and 2- feet long. The iron chair weighs 6 lbs., the bolt and nut, 13 lbs.; 2 bridge spikes, 9inches long, take the place of 6 spikes. The whole costs about 65 cents per joint. The joint, fig. 1, plate 21', consists of a 33-inch bottom-piece, resting on two sleepers, a 36-inch side-stick, and a 17-inch iron fish-piece secured by two bolts. The cost of the joint is $1. 3d, Keyed joint-chairs are of nearly the same varieties as bolted chairs. Their grand defect is, that iron enough to prevent their splitting, as the key is driven, will give them an anvillike rigidity. Wedges or keys under the rail, between it and the chair, if of wrought-iron, may be light, and are very effective as to continuity, though they give very little stiffness. Wedge fishes, or fishes fastened by keys, or held in place by longitudinal bolts, all tend to pry open the chair, and involve very great weight and rigidity. The stiffness of all joints which gripe the foot of the rail only, has been found to depend on their length and their individual stiffness. Sufficient cost will make a good job, but no cheap, light structure will answer, although a device which will bring both adjacent ends of the rail down, when one is depressed, is found to prevent much lamination. The value of the fish principle depends mainly on the shape of the rail. If this is high, the high splice or fish it allows will be stiff, and if square under the corners, the splice will bring little strain on the bolts or keys, and will hold fast in its place. Mr. Adams has proposed to employ pieces of old, laminated rail, for bottom splices. His method of fastening them (Fig. 2, Plate 21), is obviously impracticable. The author has suggested the use of long gibs or troughs of rolled iron to hold the two rail flanges together, these gibs to be prevented from sliding off the flanges of the rails by pins put thlrough them and the flanges. The plan is on trial. A form of splice-chair (Plate 29), neither keyed nor bolted, has been but lately put down. The Adams bracket hooks at the bottom of the rail, upon a sleeve-chair or bottom-piece, so that when the joint tends to settle, the least downward motion hugs the fish part into its place, and draws the opposite lip of the bottom-piece close upon the foot of the rail. There are no nuts nor keys to jar loose, and the weight of the train tightens the joint. 3. Simple fish-splices are used upon and between sleepers. The suspended joint is too elastic with a low rail and bad ballast, while the joint on a sleeper, without a uniformly smooth road-bed, is too rigid. The general principles of the bolted splice chair, as mentioned, apply also RAIL JOINTS. 183 to the simple splice. A wedge-splice, held laterally in place by projections, formed upon the bottom of the head and the top of the foot of the rail (Plate 21-, Fig. 10), has proved valuable as far as used. The excellence of the fish-joint in all cases where rails are deep and nearly square under the head, is obvious. The tensile strength of the web of rail itself resists the downward pressure of the wheels, that pressure being transferred through simple plates of iron standing on edge. All, that is required of the fastenings of the splice is to hold it laterally in place, and if the rail is nearly square under the head, only slight fastenings are required-in fact, the splice would remain in place without fastening, while the weight was upon it. If the splice is at any other point than between tlhe upper and lower tables of the rail, itsfastenings have to resist the weight of the train. It was originally intended, when the fish-joint was proposed in England, to employ two sleepers 6 inches apart, at the joints, and to drive a plate, a little wedging, between the jaws of the two chairs and the side of the rail, covering the joint equally on each side. This was never adopted, but, instead, the sleepers were removed say 12 inches each way from the joint, and a pair of plates, say 18 inches long, 3 by I inches, bolted together througli the rail by four - or ^ inch bolts, an allowance being made by oval bolt-holes, for the expansion and contraction. The fish-joint, upon this general arrangement, was designed by W. Bridges Adams, in 1847, and has been applied throughout the London & North-western, the Eastern counties and some other lines. It is being extended to the London & Soutlh-western, and others. The Frenclh are beginning to adopt it. The Indian and the royal Swedish railways have it down or contracted for. The fishljoint, with key-bolts, was first used by Robert H. Barr, of' Newcastle, Del., in 1843, but with the low American rail, it had soon to be discontinued. It las been stated, that the underside of the liead and upper surface of the foot of the rail require to be of a shape to afford a good bearing for the plates. In this respect, English rails are always superior to the common American pattern. The rail shown at fig. 1, plate 8, is particularly designed upon these considerations. The want of perfect fit, and of further provision for taking up the wear on the edges of the plates, has caused them, in some cases, to get loose. The nuts often tend to workl loose-on some lines apparently to a great extent. The outward strain is such tliat, as soon as the nuts are loose, all support is gone. Much of whatever difficulty is experienced, undoubtedly arises from inferior workmanship. The fish-plates, bolts, etc., are contracted for in large quantities at low prices, and the mechanical arrangement is not responsible for consequences due to such causes. In punching fish-plates, unless care is used, they are apt to be swelled on the edges, opposite the holes. This may reduce, as it often does, the actual bearing to- one-eighth its intended amount. The nuts are quite apt to be convex-faced, by which their holding friction is greatly reduced. The bolts are apt to be cut with a rough thread, by which their hold is greatly lessened. So, if the holes are not accurately punched in coincidence with those in the rails, the strain especially when the rails alter their length by heat, would naturally work the joint loose. Thus, the quality of the workmanship exercises so important an influence upon the stability of the arrangement, that tlhoroughly good workl must be understood, from the start, as quite essential to any success. Yet there is an undeniable tendency of the nuts to worlk loose —there are, probably, twenty devices, patented and unpatented, for retaining the nuts of fish-joints in place. Mr. Adams proposes to chamfer thle inner edges of the nuts, and to drive a flat, dovetail key between them, as shown by fig 9, plate 19. This reduces the bearing surface of the nuts. The fish-plates of tlhe Royal Swedish railways, have concave-faced nuts, by which the corners get a hold in the side of the plate. The fish shown at fig. 1, plate 8, is a very thiorough job. A split key outside the nut, cannot fail to keep the latter tight. The fish-plates must not, of course, come in contact with the sides of the middle web of the rail, as room is required to draw them to fit when worn, and as there is an elasticity, when they bear only on the edges, which relieves the jar on the nuts. The nuts of some of the fish-joints used in France, are made with a conical end, screwed into a countersink in the fishlplate. Some of the English fisli-plates are now grooved on thleir sides (see Fig. 12, Plate 17), 184 PERMANENT WAY. to hold the head of the bolt, and to catch the nut when it turns square with the groove. This weakens the fish. The fish-plates of several of the French railways are similarly grooved for the bolt head only, which is made of T form. These plans appear to regard it as necessary to prevent the bolt, rather than the nut, from turning, but of the two it would appear that the nut was much more in need of this protectopn. Fig. 5, plate 20, shows a long nut for two bolts, the bolts being screwed into it. This same plate also shows a plan of bolting the fish-plates through the joints of two chairs. This is one of the " fighting patents," so called, of a company interested in railway inventions, and is not, as we understand, adopted in practice. Saamuel's double-fish, figs. 1, 2, 3, 4, and 17, plate 20, show Pole's tapped joint, where all the bolts are screwed into one of the fish-plates, tapped for the purpose. This avoids the nut, which is apt to be in the way of the flange of the wheel, and, doubtless, secures a good hold of the bolts. But it requires an accuracy of work hardly to be expected in such materials. If the holes are not tapped in one plate exactly in coincidence with those punched in the other, the bolts would be greatly strained, there being a tendency to wring them off in the neck in turning them up, and to prevent also the head from, taking a good bearing. If a bolt should thus break off in the fish-plate, it would prove very inconvenient to remove it. Fig. 16, plate 20, showing right and left screws for this purpose, are open to the same objections, and to a greater extent. The joint shown by figs. 1, 2, and 3, plate 36, may have any amount of stiffness by deepening the fish-plates. The method of fastening it is very simple. This joint has been severely tested, and has been in successful use for a year on the New York Central road. The fish-plates on English lines are usually 18 inches long. For the PRoyal Swedish railway, they are 15 inches. For the North London line, they are 27 inches long. They are sometimes rolled as a trough, but are more commonly made flat and chamfered on the outer or inner edges-where the stem of the rail is vertical, generally on the outer edges. The bolts are quite commonly RF inch-more lately they are made i inch, in diameter. defective workmanship and badly-shaped rails have in some cases rendered the fish-joint abortive. This has been the case especially with the suspended joint. On the Eastern Counties line, after being used between sleepers, for several years, it is being gradually removed. The fishplates, after use, show an abrasion of their upper and lower edges, in such way that they appea)r to have been bent. But between the points indicated as the original positions of the ends of the rails, a burr is left on the upper edges and on the inner sides-from which may be seen how the iron has been worn away from these points. That good fish-joints are largely a question of workmanship may be readily observed by comparing those on the South-eastern and on the Brighton lines, where they run side by side; on the former the joints are well made, and they are rarely loose. On the Midland Railway, the deflection of the fish-joint was found so troublesome, that a sleeper was driven beneath the joint. This sleeper was merely left loose, and could not, therefore, have its proper effect. The deflection of the fish-joint may be readily observed during the slow passage of trains. Its stiffness may be easily increased in the manner shown by figs. 1, 2, and 3, plate 36. On the northern Railwayr of France, some 350 miles of 5-inch flat-footed rail is laid with fish-joints resting on sleepers without chairs. The two adjacent sleepers are 30 inches from the joint sleeper; the rest are 36 inches apart. This appears to be all that is required for the 34-ton Engerth engines. On the Eastern Counties line, all the bolts and nuts of the fishjoints were renewed in 1856, andsthe cost of this was equal to that of all other repairs of the line. It is the intention to remove this joint altogether fromw this line, on which it is now laid down for 1,000 miles of single track. The rails, however, to which this joint had been originally applied, were old and of a form unfavorable for obtaining a proper bearing of the fish-plates. On the Great Northern line, an examination of 250 consecutive joints, showed that 261 bolts (26 per cent of the whole) were loose, and six came entirely out within 48 hours after they had been carefully tightened up. The nut, when firmly screwed up, takes a certain bearing in the thread, and by the strain upon it must insensibly produce a change in the form of the thread —so much, that being afterwards screwed up to a new position, it will tend to work back to its old bearing — and with this, the structure is again loose. This tendency can only be restrained by maklting:First, fish-plates of good strength, with a bearing of but moderate angle with the horizontal RAIL JOINTS. 185 line of the bolt-say 30~. Second, making the bearing of the edge of the fish-plate continuous, instead of at the swellings caused in punching the holes. Third, making the bearing a surface instead of a line-fitting to at least half an inch of the width of the under side of the head of the rail. Fozrth, making good fitting bolts, straight and square under the head. Fifth, employing all the available friction surface of the nut. Sixth, making a full, clean thread in the nut. To this must be added a certain lateral elasticity in the plates themselves, by making them deep and thin (where the rail is deep enough), and in all cases fitting the plates so as to clear the sides of the stem of the rail. In short it may be laid down as a rule that with a good form of rail, and with careful workmanship, the fish makes a good job with the ordinary cross-sleeper system. Its peculiar excellence for the Sandwich system has been already referred to. Almost if not all English railways are laid with joints opposite; that is to say, on the same sleeper. In this country, as is well known, opinion is divided upon this practice. It has been usual to make both joints on the same sleeper. The arguments against alternate or broken joints' are those which apply in the case of a very bad track. One of them is that the alternate yielding of the joints causes the train to roll, and that the blow taken at each joint tends to wear out the middle of the opposite length of rail. These arguments assume the very state of things, however, which it is the purpose of broken joints to avoid. For the yielding and blows at the joints must be much less where each has the full bearing of an entire sleeper to support it. Thus, while there are twice as many shocks with alternate joints, they are each reduced to less than one half the force of the shocks with opposite joints-twice as many sleepers and a portion of the strength of the rails being interposed to lessen their effects. The great objection to the rolling motion given to the train, is that its effect upon the running gear is more destructive than the simple vertical motion caused by opposite joints. With properly fastened joints, however, the rail will be practically continuous, and it will then make no difference where the joints are placed. The allowance for the expansion of joints is, of course, the same in all countries, with equal variations of climate. The necessary allowances for a 20-feet rail are as follow: At 100~ place the rails in contact. At 30~ at a distance of -014 inch. 90' at a distance of -016 inch. 20~ " 131 " 80~ " " -032 " 10~ " " -147 " 700 " "'049 c 0~ "'" 163 " 60~ " " -065 " -10~ " " -179 " 50~ " " -082 " -20~ " " -296 " 40~ " "'098 " — 30~ " -212 " A variation of 10" or 12~ is said to produce a force of expansion in iron, equal to one ton to the square inch. A variation of temperature of 150" would, thus, in its produced expansion or contraction, cause a strain equal to the strength of the iron itself. Engine tires, however, are often'E stretched one inca in a length (of circumference) of sixteen feet; and although the interval from. red heat to being cooled with water is not two minutes, the iron is not weakened. Iron is indlee strengthened by being subjected to tension while at a high temperature. Professor Jollnson found a gain of some 20 per cent. to be due to what was called " Thlermo-extension." Track, in this country, where laid with close joints in cold weather, has been raised vertically one foot, and thrown latterly two or three feet by expansion. In England, the same thing has occasionally happened. In June, 1856, a train running at 40 miles an hour was thrown off the inside of a curve, on the North-eastern railway, in consequence of the rails being bent by restricted expansion. This was with 82-pound rails, fastened every three feet by heavy chairs, and fished. at the joints. On the other hand, several interesting examples show that the fastenings may be sufficiently strong, or that the heat absorbed by the rail may be so carried off by the ground, as to keep the rails in place, even without any allowance for expansion. I. K1. Brunel has stated that he had welded together 100 feet of bridge rail, and riveted together another length of one quarter of a mile These were securely fixed to the longitudinal timbers, and accurate gauges set at each end. The result was conclusive, that there was no accumulated motion, either at the extremities or at any part of the two lengths. Barlow's saddle-back rail, which is bedded into the ballast, is now 24 186 PERMANENT WAY riveted closely and firmly together for five or six miles, without any practical inconvenience from expansion. Mr. John Hawkshaw once erected a parapet, 750 feet long, formed of cast-iron plates, nine feet long and one inch thick. This was firmly riveted together and held down'tightly, and was proved to retain its position perfectly under all actual variations of temperature where put down. Practically, however, for a climate like that of this country, there is no authority for disregarding the expansion and contraction of rails. Hence, the longer the rails and the fewer, consequently, the joints, the greater the expansion at those joints. A rail laid down, as has been suggested before the Franklin Institute, 300 feet in one continuous length, would at 40~ require 1 inch for expansion; and when the ends had worn out from such a dangerous allowance, the whole rail vw-ould have to be taken up. CONCLUSION.~No track is complete, nor can it economically carry a heavy traffic, excepting with thoroughly spliced joints, making the rails practically continuous. No mere chair or seat can be sufficient, since the foundation upon which the whole rests-the earthwork-cannot be depended upon for its permanent support. Thoroughly jointing low, pear-headed rails is impracticable; any cheap fixture which will pre. serve the stiffness of the joint will be destructively rigid. For a deep rail, the fish-joint is probably the most cheap, simple, and durable that can be devised. And a deep rail, especially if it forms a part of the Sandwich system, is certain to be a feature of the best permanent-way. The Sandwich system finally settles the question of rail-joints, by embodying the fish-joint in the most simple manner; indeed, nothing but the fish-joint would be appropriate to it. No excellence of joint fastenings can compensate for defective earthwork or drainage; bad ballast; small, irregular, and decayed sleepers, or for weak, unsound rails. Before we can have good, smooth roads, admitting of easy traction at high speeds with a low rate of maintenance, each detail must bs perfected so far as the means have been indicated conjointly by science and experience. APPEND IX. STREET-RAILS. THE standard forms of American street rails, together with several improvements not yet in extensive use, are shown on Plate 6. The history of street railways in this country has been long enough to establish certain important facts and conclusions which will be briefly mentioned, regarding the adaptation, permanence, and economy of these worlks. While the chief object of using rails or tramins in streets is to decrease the cost of transportation by decreasing the rolling resistance to the wheels of vehicles, the conditions of traffic in crowded city streets and on suburban roads, are by no means identical. The traffic of crowded streets is, to a great extent, cross and diagonal, instead of parallel to the streets, because each door is a stopping-place for one or more of the throng of goods and passenger vehicles, while each of these vehicles is constantly turning out to pass or overhaul the others. Any considerable obstacles,.therefore, to cross and diagonal traffic, such as deeply grooved rails (Figs. 7, 8, and 9), which catch the wheels of miscellaneous vehicles and tend to keep them in straight lines, always straining and frequently breaking them, are pronounced a nuisance by the public, after a fair trial. But in long suburban stretches of wide road, and in all city streets which have little local business, but are employed as through routes for passengers, the traffic is mostly straightforward, and little, if any, public inconvenience is caused by any reasonable kind of tram or rail. An essential feature of a city rail, then, is its smoothness or general level with the paving. A certain depth is required for the flanges of wheels;: —inch is found sufficient, as in the case of the Philadelphia rails, Figs. 4, 5, and 6; and 1-inch is believed to be ample, except on short curves, where there should be a cast-iron trough, holding both sides of the flange of the inner wheel, as commonly used; the flanges of the outer wheel run on a flat plate. A projection of half an inch from the body of a flat rail or from the surface of the pavement, for the tread of the oar-wheels, opposes no objectionable resistance to cross and diagonal traffic or to the wheels of miscellaneous vehicles which run on and off the rail and use it. Good adhesion, however, as well as easy traction, is necessary. Horses slip upon a flat surface of iron when it is not quite clean; and the wider the rail or tram the greater the trouble in this direction. For this reason, and because there is so little straightforward traffic in crowded streets that miscellaneous vehicles could not advantageously use a wide tram-rail to run on if they had. it, such a rail is not best suited to the most busy comimercial streets. A narrow rail, furnishing merely a tread for the car-wheel, projecting but half an inch above the level of the pavement on the inside, and flush with it on the outside, as in Fig. I, would interfere in the smallest degree with the adhesion of horses. Whlile so small a rise of the tread or depression in the paving, as at E Fig. I, would lot catch the wheels of miscellaneous vehicles and hold them in a straight line, to any considerable extent, there would be some more abrasion of the stone at E than at other parts, and it would ultimately wear into a deep gutter. At G, however, the rail and the paving are so nearly level that ordclinary wvheels would not tend to run more on one part of the paving than on another. To prevent any extra wear of the stones due to this cause-the catching of wheels in the groove left in the paving for the flange of the wheel-it is probable that a rail like that shown at Fig. 6 or Fig. 10, except much narrower, with a deep web or base like those shown by Figs. 1 and 11, would accomplish every desired result. The bottom of the groove for the flange of the car-wheel would then be of iron, which would last as long as.the tread of the rail; while all parts of the rail touching the pavement would be practically level with the pavement as at A, G, Fig. 1. The excessive wear of stone, due to the cause just mentioned, is a very marked feature of all street railways where the iron projects much above the stone. But the most marked and unfavorable result of this same cause is the settling-the hammering lown of the stones adjacent to the rail. When rails are laid upon balks of timber, as at Fig. 3, the paving stones on one side of the timber do not break joints with or have connection with those on the other side; and they, therefore, have less permanence and solidity than the remainder of the pavement, where the stones, by breaking joints, wedge and adhere together, and where the weight upon one stone is distributed over a much larger surface of ballast, than its own area. -The pavement next the sleepers, then, is not only weaker than the rest, but in case of such rails as 1Ss APPENDIX. those shown in Figs. 7, 8, and 9, it carries, perhaps, ten times as much traffic, because the high sides of the rail catch and hold all wheels which attempt to cross them at an acute angle. The trough in the rail itself (Figs. 7 and 8) is enough to break a great many wheels and axles, but the irregular gutters, soon worn and depressed at the sides of the rail in the pavement, are much worse. This is a notorious fact as to the New York City railways in busy streets which have a light paving. Not only next the sleepers, but at other points, the paving is beaten down; omnibus-wheels, which are set to a wider gauge than the rail-track, sometimes make two or more gutters-the one next the sleeper and thne others outside the track, where their opposite wheels run in the same continuous straight line. Not only must the surface of city rails, then, be nearly flush with the paving, but the paving adjacent to them must be as permanent and firm as the rails themselves. Seaton's rail, Fig. Il, provides for this necessity better than any other, while its head, especially with the modification mentioned-a side tongue to form the bottom of a groove for the flange of the car-wheel-is well adapted to the convenience of miscellaneous traffic in crowded streets. The paving stones at the side of the rail are beveled so as to rest on the sides of the A, just as the stones in Fig I rest on the bottom flange of the rail. Thus (Fig. II), the adjacent paving has not only the bearing due to its own area, but that of the entire width and a large portion of the length lof the sleeper. With such rails the paving would not settle at all; it could only wear, and it would not wear faster than the rest of the paving if a narrow side flange or tongue were rolled on the side of the head, under the flange of the wheel. The timber has a very wide bearing on the ballast, and the system has advantages which are mentioned at length in the foregoing Chapter on Sleepers. For principal streets, however, like Broadway or the Strand, it is important to have an essentially permanent way. Wood is perishable, and the extra first cost of iron in such a situation is more than compensated for by its durability. In fact, there is evidence enough to show that the entire pavement of such streets should be of iron; in that case some system of coping, like that shown by Figs. 13 and 14, would make a cheap and durable railway. With stone paving, the rail shown at Fig. 1 has advantages over any combination of wood and iron. The vertical stiffness of street rails has not been referred to; all street rails in use have an excess of stiffness over steam railway bars, with reference to their respective loads. Yet there is an evident advantage in stiffness, and this rail (Fig. 1) provides it in excess. Nor could the iron be much lighter, for it must have depth enough to embrace the paving stones, as shown, while the web could not be as light as that of the SandAwich rail, which is supported under the head and is prevented from buckling by the side timbers. If cast iron trams (corrugated to give horses a firm foothold) were laid on both sides of this rail, much less depth would be required, and the rail would still be stiffer than any form of flat rail (Figs. 4, 5, and 6). The bottom flange is as wide as the wooden sleeper would be, in addition to which the rail may hang from the head upon the tops of the adjacent paving; and, in turn, the paving may be supported not only by the bottom flange of the rail but by the paving on the other side of it. In short, the rail and the paving, being fished together, form a continuous support. Fig. 15 is a double-headed rail, which fulfils the same conditions and furnishes two rails at the cost of one. The flanges may be narrower than here shown, which would better adapt them to crowded streets. And the paving would be laid flush with the top surfaces of the rail, thus preventing their excessive wear. There is no advantage to railway companies in laying a rail which obstructs miscellaneous traffic, for the very process which racks and breaks miscellaneous wheels and axles, loosens and dislocates rails and sleepers. For through-traffic streets which have little local goods traffic, and for suburban roads, the objection to flat or tram-rails (Figs. 4 and 16) which has been mentioned the slipping of horses-does not hold. The traffic is in straight lines, and horses have little occasion to step on the rails at all. But the use of a wide flat rail is of especial and decided importance to the public convenience, since it gives the advantage of easy traction-the advantage of a railway over a stone pavement —to every vehicle in the street. This is the case in Philadelphia where the wide tram is used; the carts and wagons generally, not only save in time and cost of maintenance, in virtue of good permanent way, but they keep ozt of the way of the cars, which is an advantage to the railway company. All vehicles need not be of the same gauge to enable them to run on the trams; a variation of ten inches is allowable on such a rail as that shown by Fig. 4. Since the convenience of the public should be first consulted in granting the monopoly of passenger traffic to private companies, it is outrageoug to allow the use of such rails (for horse cars or light steam cars) as those shown at Figs. 7, 8, and 9-rails which are a positive nuisance, as every driver of a vehicle in New York will testify. And it is unnecessary to allow the use of the rail shown at Fig. 12, or of any other rail that the public cannot use, when one of the same cost (Fig. 4) will give every teamster and citizen a railway to drive upon, without inconveniencing the railway company in any important degree. City railway charters are not any too dearly paid for, in this country at least. The rail shown at Fig. 16 possesses the advantages of the compound and the double-headed system, besides dispensing with wood and locking the paving together, as in Fig. 1. lWhether or not its advantages are sufficient to warrant its greater first cost is yet to be tested. For strictly suburban railways, where the wide highway afords ample room for miscellaneous traffic, no reasonable railway system will interfere with the public convenience. The simple question is, which is the best railway irrespective of outside traffic? The conditions are exactly those of the steam railway, except that cross sleepers would not allow a good path for horses and are inadmissible; in fact, the longitudinal system is better for any railway. The choice, then, lies between the Sandwich system (Plates 5; 21, 32, and 33) and Seaton's system (Plate 23, Figs. 24, 25, and 26, and Plate 6, Fig. 11). These systems are discussed in detail in the foregoing Chapter on Sleepers. APPENDIX. 189 ADAPTATION OF MACHINERY. Inasmuch as the extent and cost of permanent way far exceeds that of rolling machinery, it is evidently cheaper to put the required amount of uniform flexibility and elasticity into the train than into the track. There must be flexibility and elasticity to relieve the jarring and consequent destruction of track and machinery, consequent upon and in proportion to the deviation of the surface of the rails from a true plane, and the tread of the wheels from a true circle. Sensitive springs between the rolling machinery and the engines and cars resting upon them, are certainly of the greatest importance. A considerable portion of the economical working of European railways may be traced to the use of long, elastic springs. The springs of American engines and cars, on the contrary, have a short, hard action, which is only rendered tolerable by the superior equalization of strain and weight peculiar to American machinery. In any case, however, the entire weight of the wheels, axles, axle-boxes, springs, etc., falls directly upon the rail, entirely unrelieved by the springs. The unsuspended weight of some American engines is about six tons. The only relief for the rapid destruction of rolling stock and way, which now visibly results from this cause, is to place an elastic material between the tire and the wheel, as well as to preserve a uniform elasticity in the track. Then the simple rail will be the anvil, and the simple tire will be the hammer. In fact, it is quite evident that if sufficient elasticity were put into the wheels and train generally, none whatever would be required in the track — the rails might be laid on rigid iron or stone piers. Elastic Wheels.-In the " horse-foot " wheel-tire of Mr. W. B. Adams (Plate 26, Fig. 12), the tire is formed with a deep internal rib abutting against the face of the wheel, and a continuous hoop-spring of tempered steel overlaying a continuous hollow in the internal periphery of the tire on which the wheel rests. The wheel is forced on by a gentle pressure without heating the tire, and is retained by a flat ring sprung into a groove at the back, thus preventing the wheel from getting out, and no other fastening is needed. The comparative simplicity of this wheel is apparent; and, thus completed, it is analogous to the foot of a horse, standing on a continuous spring. This system has been successfully practised on the North London Railway, where a set of four solid wrought-iron discwheels, 31 feet in diameter, was fitted with Staffordshire tires so constructed, and placed under a carriage weighing 5! tons. At the same time, a set of wrought-iron spoke wheels, of the same diameter, were furnished with Lowmoor tires, shrunk and riveted on in the usual manner, and were placed under a similar carriage. The following figures (14 and 13, Plate 26), show in section the respective wear of the tires after running 45,000 miles in the same trains; and they indicate that the Staffordshire tires on springs, had worn on the tread one-half less than the Lowmoor ordinary tires. Wheels on this system have worked satisfactorily also, on the Eastern Counties Railway, for upwards of two years; under carriages, break-vans, and coupled engines. Mr. George H. Griggs, locomotive superintendent of the Boston and Providence Railway, has for some years employed cushions of hard wood, about one inch thick, between the rims and tires of driving wheels. A system of slots, say i- inch deep and 4 inches wide, is formed in the rim of the wheel. The pieces of wood are driven into these slots, and form an almost continuous wooden rim around the wheel. The tires are shrunk upon the wood, and are instantly cooled by dropping the whole into a tank of water, so that the wood is not charred. The first cost of the apparatus, in case of cast-iron wheels, is trifing, and the wear of tires is increased at least one-half. Mr. Griggs' plan is beginning to come into general use. The Bissell Truck. —But the trains and track require lateral as well as vertical relief. The European train, severely rigid laterally, without trucks and swing beams, shows the first effects of wear on the flanges of tires and on the insides of the rails. The Bissell truck for locomotives, illustrated by Plate 69, is specially designed to relieve the lateral rigidity, and in so doing, it accomplishes other important functions. Since the truck does not turn on its own centre, but on a centre between its own and that of the forward driving wheels, all the axles of the engine become practically radial to any curve the engine may run upon. At the same time, they are always rigidly parallel to each other, when the engine is on a straight line, while the common swivel-truck sometimes vibrates on straight lines. When the engine strikes a curve, its front end rises upon the inclined planes shown in Plate 69, and its gravity tends to bring it to the bottom of the planes again, and to keep it there, on straight lines. IBesides saving at least one-half the ordinary wear of flanges, on the English plan of engine, it enables the leading wheels to be placed much farther forward, and more weight to be thrown upon the driving wheels. It also allows the engine to run upon shorter curves than the 6-wheeled engine or the common American engine can pass at all. It is a safety truck, preventing the wheels from leaving the track in case of bad joints or slight obstructions on the rail. The Bissell truck is already in use upon above 100 American locomotives, and a year's experience with it on the Eastern Counties line has warranted its farther introduction there. SWITCHES AND FROGS OR CROSSINGS. Wood's Safety Switch. —This invention is so fully illustrated by Plate 35-1, as to require no further explanation. Its design is to prevent trains from running off the track, in whatever position the switch may be set, and in whichever direction the trains may be running. It is in extensive and successful use on many of the railways in New York and in adjacent States. Dick's Frog.,This improvement is illustrated by Fig. 1, Plate 352, as lying in the track, ready for service. That the ordinary frog, or crossing, affords too small a bearing for wheels, is a notorious fact. A simple castiron frog so rapidly fails, as to render its use inexpedient under any circumstances. The best common frogs are steel plated, the point or tongue being of solid steel. Upon these points comes nearly or quite the whole weight of every wheel, while passing it. In the arrangement here shown, the rail of the main track which crosses the side 190 APPENDIX. track is continuous, and practically without any break whatever, since it touches the point of the frog, leaving no channel for the flanges of wheels when they run on the side track. When wheels are switched upon the side track, however, they open a channel for themselves, which closes up again, since the rail referred to is a spring, being fastened as shown, so that it normally rests against the frog, but may be moved aside, to allow flange room for wheels taking the side track. Thus, the shock of concussion, and the consecluent wear and strain of track and equipment —always severe in case of ordinary frogs —-are almost wholly avoided. The frog is arranged upon separate and independent chairs, with open flanges, so that the several parts can be separately and readily removed for repairs, without disturbing the other parts. And it is arranged with large open spaces under the rail, so that small stones, snow and ice will fall through, and not interfere with the working of the spring rail. The cost of the common steeled frog is from $30 to $60; the average price is about $40. Dick's frog and a full set of chairs, is sold for $16. The two extension rails, bent and fitted to the chairs, are furnished at three cents per pound for any length desired; the iron is not materially injured by use, and its value when worn out being deducted from the first cost of the frog, the latter is found to have cost only one-quarter as much as the common steeled frog. Burleigh's S'witches and Crossings. —To prevent the crushing action of the outer edge of the wheels, a projecting piece (see Plates 34 and 35), inclined at each end, is rolled upon the Tongue-rail of the switch, to give support to the flange of the wheel, while passing over, which relieves the outer rail from the blow that would otherwise be felt. This projecting piece, or fiange-bearer, also gives additional lateral stiffness to the Tongue-rail, and prevents it from springing and opening at the point. A filling piece, or flange-bearer, steeled on the top and bottom surfaces, is introduced for a similar purpose in the crossings; being bolted between the point and wing rails, it protects them from injury. This filling piece has also another use — it effectually braces the wing and point rails, and renders the whole crossing as rigid as a beam. The importance of this is evident, as the movement which ordinarily takes place between the various points of a crossing when trains are passing over, and which causes a series of severe blows both on the Permanent Way and the Rolling Stock, is entirely avoided. These flange-bearers are also made reversible, so that when one side is worn out, the whole crossing (with double-head rails), or any part thereof, may be turned over. TABLE OF THE'WORKING PRESSUEE OF LOCOMOTIVE BOILERS MADE OF PLATES WHOSE ULTIMATE TENSILE STRENGTH IS'30 TONS OR 67,200 lbs. PER SQUARE INCH OF SECTION. —Fromr Clark's "Recent Practice." WORKING PRESSURES, FOR DIFFERENT KINDS OF LONGITUDINAL JOINTS. WORKING PRESSURES, FOR DIFFERENT KINDS OF LONGITUDINAL JOINTS. ~Welded Joints. Riveted Joints. Welded Joints.. Riveted Joints. ~ I:E^,fwn TDDouble-rivet'd Double-rivet'd' Single-rivet'd Dnuble-ivot'dDouble-iivet'd Single-rivet'd Scarf-Weld. Lap-eld. Double-Welt. Lp. Lp. eld. Lp-d. Double-Welt. Lap. Lap. Lap~~~~~~~~~~~~~~~Is Lap. I. La-,x ed inches., l bs. i e bs.p. PP,-. lbs. per I[. p lbs. per Ib. le bs. per inchs. pe. p bs. per Ibs. per sque iich. sque in ch. squar e inc h. square inch. are e. square i.. squ inch. square inch. square inch. square inch. 183 123 150 133 112 122 82 100 89 T5 229 155 188 167 139 153 103 125 Ill 94 30 275 18T 225 200 16T 45 -1 183 123 150 133 112 r- 321 26. 21. 17..... 21 5 I 867.. 300... 244 l. 200 16 112 136 121 102 115 77 94 83 70 yV 208 140 170 151 127 144 96 117 104 88 33 i 550 168 204 182 152 48 -j- 173 115 141 125 105 7~~~~~~~~~~~~127 T - 292 238 203 164 8 333.. 23...29 I7... 29188 153 103 125 Ill 93 108 72 88 78 66;~-~~~~~~~~~~~~~~~~~~~~~~ 153 72 8 ^ 191 128 156 139 116 5 135 91 110 98 82 36 - 229 154 187 167 139 51 162 109 132 117 98 9 267... 217... 189 154 1 305 250 216 176 i 141 95 116 102 86 i 102 68 83 74 62, Il 176 11 145 128 108 i5 12 SG 10 ^ 3 -1 212 143 174 154 129 54 i 153 103 125 III 932 17 247... 203... -.1718... 146 i' 282... 231.. 0204... 167 1Gi4 110 131 88 107 95 0'79 59 Ir~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ^6- 64 10 34 19 10 ^ 120 81 99 88'74 42 197 132 161.143 120 5 -1 145 97 119 106 89 -^ 230... 189.... 139 262 214 1_... ]... ^ 193 158 NoUe 1. The blanks respecting the y^-inch andl 1-i-nh pnlfats de~note that. t~he pressures~~ are1 u~ncerta~3in, bu7t may be a.s-~~ ~ ~ ~ ~~~~~~~~ IVIr~V)ULIU ILUVI JVU~V ~, rv~~- u -~- -~ —— J ~-/- o,- / w / 98-T- o sumed at leazst equal to those for i-inch plates. 2. The workringr pressures for 1-inch and ^,-inch plates, are estimated in the ratio of the thickness from those for -1inch plates. 3. The BUESTING PnEESSTEE is five times the wrorking pressure. 4. To find the working pressures of boilers of other metals undernamed, multiply the pressures giv-en in the Table by the annexedl multip~liers: Best Stalffordshire plate,..'2 (or, roundly, deduct one-fifth.) Best Amnerican plate,.. 1'27 (or, roundly, add one-fourth.) Ordlinar~y American plate,...'I (or~, roundly, add one-tenth.) Cast-steCin Pate,.... -6 (or, roundly, add two,-th-irdg.) RESULTS OF EXPERIMENTS ON THIE COMPARAiTIVE TENSILE STRENGTII, &o., OF VARIOUS KINDS OF STEEL AND IR1ON PLA`TES, BY M~ESSRS. ROBERT NAPlER & SONS. _N'te.-Z-All the plates were taken promiscuously from Engineers' or Merchants' stores except those marked S, which were received from the Makers. Breaking weighit per 1^,. *. Breaking weight per square inch of square inah of.:adured.riginal sectional area. Contratetion. of I- ^ I ^^/sectional area. ^ ^^ ^ Elongation. Characteristics of Fracture.~~~Eongtion Chraceriticsof racure - S ^ Names of the Makers or Works,.___ titi 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Fracture. Highest. Lowest. Mean. Mean. Mean. AMean. ~ ~1~~ 2 3 4 5 6 7 8 9 10 STEEL PLATES. Granular. lbs. lbs. lbs. lbs. lbs. lbs. per dent. per cent. per cent. 4 r T. Turton and Son. Cast. S. 5..' 9 I 7'1 lWholly granular, fine, and semi-lustrous. 95,299 10 05,9063 T 76 O~Wolygau I I Naylor, Vickers, and Co. S...... L: 87,972 7672 81,719 0 84,435 11'4 -108,125 21ranular and rous; very fine. 6 - I) S. C 95,196 ~~~~~~~~~~~~82,428 87,150f8,5 112,018 f'108,125 2201T224Gauranfi Moss and Gambles.,~...S. L 81,588 67,977 75,594I 728 105'554 2821}8 77 I2,338 Mossa9da050es.","'.'.'.'.'. j34^ ^ 1^ i~t^^^ l ~ ll'l21 19 8 0 Wholly fibrous; very soft and fine. Snrriga 1twl,&Co.. S. C 71',796 67,538 69,0S2 112,546 886199,450 Homo:geneous Metal....'. C 105,732 84,211 97,070 1 1 860 11,9661o I o4 110'2 141'6 73 G ranularandfibrous;eryfine. 2e s y G. ude'te. S. C $4,724 62,485 78'580 t7,9 78,245 5$ 0, 0 214 4 Mersey do~~., pud''te.'. 108,906 92,676 101,450} 93,209 109'552 0,4 - -4 Q7 I2 1 Lmntd ir t ~ ^^" " ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~'" 1' S'7-: 1 Laminate, fibrous, andaslightly~anular. (Shi p pla tes)......'.' 99,468,1 100,649 9.474 5'4 125 2 10 3 Ihar "Ld" ~'. S., 10,20 0 95,946 1 9,11 98,472 \4'' ti f 4-08 Granularand laminated;fibrous. C)....... C 88,40 25298,0 2,,...... S. 86,908 67,184 7,046 2.1 12'5 6'96 17l 80~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~7,6."' 8,3 ^ 05 57 2 IRO..... S. Cl68,848 66,524 67686 73,634 I' 2 S........ plates),02065,045 1 51 582-77,520 63,073 32-3}^ -857 ~lomnaer fifous, f slightl y granu la r. I I ^ -.'~~.'.'.~.~~~.~.^ ^ ^~~~~~~~~~~~~~~~~~0689 9382 51,278..,6079^ ^ 102'..2. 4 49,3169 5'3 8}60 SI; Blocadiern pddo.-S.C.'....S~. ^ L^ 1 1 5,9 4- 6130 ^114 bo ") " LSF.... C'9,44L 81,047 845,98 87,877 97,9 4 -2 2 4 9,, (Boiler), 7413 95,227 6,2 72~~~~~~~~~~~~~~~~~~~~~500 92,130 7161 fi~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~' 597 9610 5,320 51 f" 0 5T6,07 01, 0 -4 8-6 o2 6,8 2 rnua n a _l^,,,,.... ^ ^ ^^75, 73,699 76,6461 5.2 IRON PAlTEgoS.Best~oiler ~. ~.'. ~.~... ^ ^ ^ f^ ^ ^1 -Crystalline. per cent. 4 Grlasow Ship L 62,544..' 561,172 8143 50643 fU 1 - 2980 24- I- 14"-1 4 Ga;fbe rte ore rstl eeal o 4 ISMakers-tampuncetain.~.'.'.~.~.. C ^ 1 ^ 4456,546 5 5 8,95 61,038 1:,9 21. 20 7 L ~~~~~~~~~~~~~~~~61,184 51,541 57,246I5, 0 69'650t 18'9 15'6 t2, 9'65 1 1~~~~~~~~~~~~~~~~~~~0,4 50,1703 56,495 63,078 12'3 6'8 f I I ^sendBest~~~est..~.~.~.'.~.~. ^ ^ 1^ ^.^ ^ ^ 1 ^30 U \ ^ U 6 40,51?0,160eCoorlgenerall 0 C ~~~~~~~~~~~~~~~~~~55,368 47,426 50,515 5T,883 61,065 1-. 1 owfi^ - 5B2st.^ 6;~. -. ~71;6.~.' ^^ 55 2^ }^ t11.6 19 * C 50 e 4 t,074 46,441 50,009 55,w62 6'9 5'9 tJ [ Bradley and Co. S. C. L ~~~~~58,534 58,889 55, 831 53,191 67,406 6136 17'2 13'1 125 9'00 3 11 5' C 55,414 ~~~~~~~~~~47,532 50,550 55,206 t 186 9'0 18' 0 8 19,,, L.F.. 60,985 54,687 56 996 66,858 Is.'0 121' 94 Dullish gray of vaiou hds ir ls n eealie {1........ 55,89T747,410 51:251 ~ 418 56,070 61,464 8'6 5 Co ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~0675,9 55,T08 52,566 65,652 1-'.o 1 49.425 1 j.54 ~ ~ ~1'0099S2 -:.107 7.Wls etBetL 5,0 41,002 47;410 47,020 51,521 6'7 5'8 40.0 0 51~~~~~~~~~~~~~~~~~4,935 49 8'7 10' K'~'B.... 49'44142'581 ~~~~~~~ ~~~~~~~~~~~~~~ ~~~~46'630 1 48'848i 4=993 3 C 49,441 42,581 10-2 6,1~~~~~~~~~~~~~~~~~~~~~~~~4' 1: ]Dull gray; fibre rte ore K. M". B. L 51,285 ~~~~~~~43,232 46,404 45,584 1,896 8'4 5} 2 C 54244,195 44764 47,891 49,893 6'443 3 Consett:L 54,403 4T,613'51,245 48,979 -9131. 17 89 7-8 4 ry ir os 4 11 C 4T,613 46,062 ~~~~~~~ ~~~~~~~46,712 52,050 55,616 10'2 611' 2'7Gay irel~o 6 Glasgow,'B Shoip e'.337.747 47,773 49,016 2 48 Glsgw Si.. 58808,7 778 46,064 4'1 2'$S 48Gray; fibres rathe ore rsal eeal o Makers' stamp uncertain L ~~~~~49,838 48, 290 47,598 4410 53,182 9'18 4 59.1 44,140~~~~~ ~~~~' 4'20 16Dull gray; fibre rte ore 6 44,512 ~~~~~~~~~~~~~ ~ ~~~~~~~~~~~~~~~~~~~ ~~87,0 40,682 48,426 48,804 5'0 27'41 ]Mal Nrs'ed, Bstapnertai.L 4,9982 43,401 434,835} 4445'84608, 47 580 8'4} 2 3~~~~~~~~~~~~~~~~~~~~,4 C,6 44,83~0 49t5~~9I8 "4 C 43'- 3S0 41,56,43622 4'9 2' 42'9 16 Gray; fibr generlycore 6 Govan, Best.... 49,6TT788,398 43,942 45,78 8. 3 34 240 12 0, 6248 43,255 2'6 [ 1'4 10 1_~_06 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~43'069836'460389'544 40.62 ~L denotes that the strain was,applied len~qthwiay's of the plate; C, crossweays. TABLE OF THE PRINCIPAL DIMENSIONS.OF VARIOUS STANDARD AMERIOAN AND ENGLISH' LOCOMOTIVES. Dimensions in inches of Flues. Areas in sq. ft. Proportion of heating surface Kind to NAME OF BUILDER. of Firebox. Fuel. Diam- Length.. Com- lREMARKS eter. D'' bustion lotal 5 ~ -. No. tance Grate. Firebox. chm Flues. heating. inches. inches. o 5 g ft...d apart, bpr surface.. I o Bogers L. & M. Works. wood 16 X22 66...... 21-165 2 II 12.5 78 00 90T 985 \ - 61 to 8 sq. ft. of grate covered with dead-plate. " "..~....b. coal 16 X22 66.... 3 190 11 8-6 - 14 78 35 670 783 6 - Iron Flues. ~ ".....a. coal 13 X22 66... 13811 9..21 80 00 595 695 _ Jersey City... coal 14 X22160 57 59 351 3~ 160 11 8 - 14.4 68 34 563 665;; Locomotive Works.. wood 13 X22 551 51 421 36' 21 102 1} 10-3~ 10.8 66 00 513 579 F. P. Dimpfel... coal 16 x22166 60 67 39.146 11 13. 18 99 80 745 824 1 1 Water-tubes, turned up to iron-sheet,-average length. ReadingR...... coal 15 X20671..... 170 1 11-5 24.5 64 4 634 702- 21~ 37 Combustion chamber 5 inches deep. Boston and Providence R.R.. coal 15 X2066 60 42 38 2. 120 2 2 17.5 82 00 720 802 Brick arch over front half of grate. 2112 New York Central R.... coal 14 X24 56........176 6 10 90 42 375507 Chimney tapered from inches at throat to 16 at top and bottom. cc cc...coal 17 X20 48........ 15011 12-4: 3.76 85 00 657 742 ry 1 Daniel Gooch... coke 18 X24 96........305 2.. 21 142 00 16271769 1 ^ GreatWesternrlway, England, 1847. "Gt.Britain" variety. Sharp, Stewart & Co.. coke 18 X24 60........142..... 10.56 86.7 00 865 951.7 1 -a Freight, six-wheeler. 1850. Daniel Gooch..coke 18 X24 96........305 2... 23.62 149.7 00 1590 1740 ^ ^ Gt. Western, "Courier" variety, 1850. I pr. drivers. J. V. Gooch.. coke 1414X2178...... 2 &S 181 1 10 12.4 75 00 823 898 1 4.6 London &South-Western railway, lpr drivers. Crampton.. coke 18 X24 96........ 300 2-^ 12-6.. 21.5 154 00 2136 2290 -^ Great Liverpool, 1849. See D. 96. Cl on 00il 2 y 126ac1nery, for particulars of these engines. D. K. Cla-!Lk.. coal 15 X20 66 48 36 42..1501 11-41 10.5 58 00 750 808 _'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'' 6 . T - -—, —--—. —--------—.-.. —--—. — -. I, —).I,I; 11. I'VIIN _'I' -1 ,,,\JPQ,,hPvVI. -:. :11 1,,;; 1 I ,, r, k" , ",`_.. I I,,i _ _ _ I I.1 '.'I 11, I'.,I.1"; 1-1 1 .11 II .,,,) .,,,.,,,,,, 2; I..!77--7_ ---- -- - ,.1 Iir".-'_: - - __ - - __ -t -,.,-'r -, - - - 7' -7- - - - — =-, — i.,,,,I - -::77___ia;_ -If_ 7:I ,.,.," -- _5 1 1 'l,"11._. -- ___ -..- - — I ——.,- __,; .,. 1. _ -__-__:- _ -F. 11.. 1, ,- I. --- - - - -,I .: I,; !,;. I I1,,,, __ _ ,I. a__7_14 I,,:' ----------, - —-.- _ i.'I- __ --.1 -,, 'i- Z- - - -_____ -I1 f, t, "''. _ I --. - __ - - - _.___II 11I; `__K, - ;, 1,, ,j,1,;I-, --'-_-Z7Q-_- -i 1, -",-,,_ _7 1 ;,,.1 i, , - al -. ',.. [ ), :-, , ).4-11 i` -11;.! I- ,;',,; '., i "i -_i i ",,;,;, j, I - I J "k, ",,.i;: .r.'.1":, -, -:.i -I,,: i. I",,_I I,,:,.: I,, 1, I" , I;:,. - - I I,. - - - I I. 1.'t P-._ - I! I I !; 1 1 I,.. IIi 7_.,'. ".,, .,,,, , tj .!,:1 A I Ii 1,I.. —.. . I - _ _ ISm. Ty, ,,, I ",,, I E l,, 111iiIiIiIiiiiiiiii,!.-E X I:! 1 I',' "I /, —--—,- -_ -, _ _ _H I v/'' /,/,,,,,.",''A. I., i,"I 1'._ V, ""I"", __, I 1/1,. /, I 1, 4'. I,,, I I I., %",, I I -,'' I,,',x l,\, 1, \I,. - S''\ " ".,,,."-,,, ,,,,)!,,,\- 11I " 1,\ -I., I -, I I 11,,,,.:,O,, ", 1;.,!,v,,"",x, , ,\\ ""', ,. , - ",",' , 1,, \"\ \,,! I,;. I'I, I \ I IN _ , I 1A.11.1 \ ", II /,, I,.I I\'I'll'' I, 1,,: "'X " i I,4, 1, I I I, ",,,.,;,, \\,.", -- " ".; ".,\, \_,\: ", 1,'',,,,,,'' / /, ,, I, //,I,, "I; " I" I t, 1 "I,/,/,, _'I, "',,, "/.,, /I1, ,\',,,II I1, 1,,I III II "W ".,, I, I _;_ I, \`, " . II,. ', I,I I'l., " / -,,, I [ 1 ", "/;" I lil, I,- -, I \ I % - 1,! I?./I. llz:,, /' V,/. "",,'.t,!-, ", II -',_ I 1.,/ , ,,',", 1. t I 1 I, I, ,,, \ I W! 1'1,. I,,,,, I I,\,I 1I I, 1, I " I',''; el I, I 11 I/ " /,, [ , /.,.I II,, I kI \\, , ,,,.,I.,II,%I I i 21. I. I ", I. J\ 1Ip I I I I, I 1 Ii.", I I.,,, " I I,.. ,.. I I", 1'1 1, -, ,,,,I -, \,,, I , 1, I i I I I i I,," , II;,I,.I,,, "` " III11 11':, I 11.1'.,.111..,";i,,"/,X`' \1111,, \\,,. II1,/,; /1,//,.,",,I.,.,, /',I.".\ 1, I,,i., I 1'11",\,,, \II 11 0, , I I,," I I, I ,.I %,I,, I I,,/, i,/ ., -',,, -',, ,".,/,, ,,,, //",,,:, ,,,'',,, I, " 1! /"",,, ", ". I I I f, 1) 1 " I., ., I " i,, II I I"I N,11 I I m ,, 1 -, 11\,., I 1 I I,, ,, " I i, i 1/`,,I//, 1,.,I.1,,,,I,,':/,.`,\,",, I I-".,,, 1,I,/, I.,:,. I. ". " i I q I q'i,, . 1, ,.,,, I, i': ! " -/11 '', ',,, /,,'. ':," ". I q 11,I': I I,,,;.,;".11,I ,. ) 1.I I'l l, - - N- ,,,,,, I. " 1 1j I / "i /;".. "". I r' I 11 I:.,,, I 1, ": ", k, I, ",'' _ I:, I, -;, lil 1,, ; I I;., I I,I I, "", "' " I , \ A,. 1i J,,N ,, 11`,,,,,,., ,'''' I I,. 1 /,.,,,,,,,,,, ",, i /,/,; .I."- (,/,, ",, X#/ ly''IIII 1,, I I; I 1. , j ! " I 1,;, I.,.. 1 iIIII,:,,,,,I ) - ,. 4:_, ", I\\ ",i I,,,, J 1,,;,,;,;II I, I, _,: I"I', - II I. i, I,\ I.. i t,,,I,,I I,. V,-,,, , 2,I1\ I I;.-,I I'll -I\.i!, ,,,,,,, i,, ",,'- \.V 1,1,:, -I,(.'',,,-.., I..''. 1. 1. ,-, _;,'/.I1,''.Ili k \ -11.1'',',.,, X111 I:t! I I, 1 1. i'i'_ _,.',`i', ,1. "I ,,.: I". b'' ",'P ,,`,,- 1''...; 1\,. I 11'111" "', 11 :-,.'' .. 1.1 1 11- O 1-!. _' )', I q. f,.,, "! iF, I k,: 1,,",; 11 \ I,!I, 11 1 1 III I I . I ,,,I - Z 1111 ill:I I ).;I,". 1, 1, I',,,,, i! _ -1 I!.. " 1/4, ,, I ; I I;!, ,, ,1 I I,,\ I, %:.% 11"; ,,,,I l I E R,, J, "'i!,:J,...."',", .I' ", "i "I 1 . _7": " A j, /V i ":,,\,t ! "!11 I,, ,,r,!,, I1 I 11I -1, I 1I \1 ,,,,," I,,!.),., I,, II,/,I1, I I ",,, I, I,, i! I, I I" II II\,, I iI,,,. j,,'t.,-.,.,,. I' ),,.,\ 1;, II.,, A " 11;:,:, /4, ,,,,,..,.I I I, II ,,,IIII`, ,;..1I, I,.It..iI iI;I 11,iIII., ";- ",,,,, ", z!z-4-., , ,, /,"."1.,I'' _ ___ __ I.,JII; !:. /,,,';;,Jl I ..,,,,, 1 -/II,(!,, , ", ",,), ,",\','i 1,.,tI,.":, II',,I., 1; /Z 1Ii, I I 41, I I. I (, 1';";, I t, I,\, I F);".. I !, I I i I,, , I_,, 1', ".. . ',`, / ".. I:, "I II I,, 1,'I 1'1;,,,,,, \, i I, -, $,", J "t 1, d t,,; —,,I \) i,;',i I - 1 4:,; /, "', ",. V,1'/"_,,,,! I 1` I,\, , ,;,.'.1, = _.1,,,. ,,,/.." I/I,: L" ill 1"'',"II '',,",,, I,,;,.. I 1\ t.', .;,kI,,\ \, lz N \, /.., -,,,," (,, "/, I,,,,1',`,,.,1. I W) I;j,,Iii 1.11, ,'7,'. -! 11',: i 1 I;=/,',il,,,,,.,,,t,'//;,,,,,'/., I;, \,,,1.yI I.,,, kf ) I,il / i,11 I;, ,,i' I ",- ,, II- ill -, 1!:,/I,I,, fI,j,,, "I,,,,,,,,,j, I,"/,I","" ",,, ",; NY L;j-. , I ;",,I, -11,r/Ij I.1,(/ 1) 1I,,,,,,.I i,-_//,)/: L _L _" 4 - -. - -- - -I,-". 1 "I i!"I,I,,!. 1,i I,,,,,/`,!((,,,/,,,,'//,,I..)I",,,,/,,,. I =' // 1' 11! " I i",,! I I,:!,/'11', 7, -- - - - -7 -- - I, I, ____ I _ L , -_ - —, _, _,_ - _,!.., J - II; __ __,"I (' i "/,,-, - 7/j, ,,, —',.;IZ, I.1_ -__ _!', i, ___ = J4_-'_ -.1, I, ,,;44 ,_.._J,,, I;-,,"),, I\'It,:,. I., :,,Ii , .. I ). i - -I!; i,. I I I I, \, I "I ,,, 1,i;', .:, Im 11; ',l, 1 " I,i I , : 11) ""\,,II I I 1, i ,.,' i I, 11,,/ I.,,, li/,.,I'V ,, i'b,, "k, i, I I11,,L " "' ,,, I,Ii` " . I !'.'), ,I - 1,I!..I,,..I; 1 1,, ", /' ,, /' ,!..,,,,.,,,,:, j,,"",,,,,;,! ,',',I_)'. j,:') ,i-,I "I'I I 11 IIi I 11I I -11. I.1, . ,,,I," —,I ;I, I.,11 ),.\I. I.,,,_:, ",1. -": N :, I , I, "! I",, t, ,, I I i( I,. II I II, t,, 1_ ',, .,,, \,\, ,: i i i II., I I ,III I N I,, -.II.,,,,.,II,,I I! )!II,,,,, I1i I1 1 1; I i,,.;)',) /I ):,:I,,,, "\, i, I,; \, " ",k, I,'/11,, _l,, "/, /I,, - I, I;, /// -,,,. '!, i I'! 1", . 1 I I I,,,, 'Ij,),1,, ;; I,,.,I " xI,,) I/!',,,,,/,I. 1 "/,,;I,;,,,/,,,,,,. I A ,,.,. //,,,,,, /",, ",., I/ I /I,,f,, I,;"7, )) I1.) ,"/., i!", (J," ,I I /,,, 5 1, 1&,, .pI 11', /11,/,`! . I,,,,, li I'll),/.,J r, "I " 1,, i,, W,\ -'_ I,. — - /, j, ",I. " I i,,.,,,, I I I-11, "". \.' .,"' ! ")"","" 14 i.,")',',,,',";,jl,l,.t ";,,, I I/ ': l ,,., I..'',\\,\., -," ";,, i I [ 1)11.!V";- /fl), I,, I( Ii , "J " j;\",,,,,\.;. -!, ,)-.,,, I i.;, - - -'Ili II I i,,'//, N,1 1.1,:,!,i'!, i " ,, /(,,, ",/:,,I,\.11 - , ",., j', !Vj,,,,\: 1,I.(, "Y'" .,,,, j fIii , 11 ) //'111,.I,j,,,,, - " " L! 'l _)_ )11 ,,,, I.I - __ - _,, ,,-'. I - -,,-,,I, 1 _ _, _14 —_- "_, - - _ L-; I - _EH41 _ 41" II' I 1',) (f,;, - L,,, -,,.,;7',--'-,`q i' il lI, — --- II,, -. -,I,I', ,I,..,;. L 1..I " i,; I'Y , I I);I.",, 1\ \I "\,,,.,,,,,,., I I,,1__ _77___7, =. I! 1. .,(,, ., ,,',,7 .il I -('I; /f 1.1,\,,,, -OVO)'"O'.M, __ _ "; k,.'. I, ). I\k,, ,!," " I WI I I - I,,,I 1, i, ,,, "'I I,,,,, dl,,:, -':!,'/ 1, ) II , " /,, / y 11, -"(,,!) 1,.-,U; I)) i k i': 1( 1,,,,,,,,,"I /;",''', II,11/11',.I.I I ___!__ -If,:, ,-, -;I,,/ I';, ",,, /I I, If, ,,11;,i \, \, I", a ", ,,,,,//11 "I /!,! /j " ,,.I "II .., I i I \! I "-,I,\:6 (' I, - , "I,1 ", I. I " ". \( f Iv I I I /,,r " J W, i, . I , i ", 1,I.../I.I,; II k; ",, ) (. I; ):,1 "i i/,,, 111,( "'(Ii,,' ' "!;j i f/'. 1 I I.11-1 '. k I, I, I 1;k' 1. 1'% "I I I. ".\. i i, "'I — 1/1, ", If I,, 1_;,\,,) ) \ \k. /,.! I /,,,,,;,,,,,:,I =',,,j,:,,,(k, / I I t, CI I, i:I1,-,,,,I II ,I,)r -,,I,; —- -, lo I I,;/: / "I,,It, ), i -. I "i \ I I i., \i, .), ) &ik,, I"'', I'l,,,,/ I, )i,,,,,I! I jI (. \, / II ". I 1 I ,, 1; —— ,- - ==. I,1,,.;.,,, ).0 I.I.I,/,/ ,!,Ii,", 1, 1i ,,:) o -.,!.'',;,!(, 111I I I. I /,/,, ", If'',, I!,I1 1)!,j "i j ,) I = — ! Ti,.,, , I" i i , i'l 1,, ".1 I I; 1,; I (I I ",: "., "10 \, .;'. I\ -!,.U ,; " I !'r,,.,--,, l )\`Vklt, :,q.d 4, I ,. I,,,,. d!,,,,,,,I,r", -,.! II.,,/.II "! I I'..!,,,,,I, "' ( ) ". 1, f, ""', e" y I," 1r,. ( )(", I I IIf, i,'i ,,, J ,) - 1, , -.; I I I i') 1; -; : jI,)" "'!,I,,', ,, /11,I I If "I- II %.i"I-. "T I" r,, "..,II,,,. If,I,//i,)! 'i`.,! - 1111'il ),,, 1 I i.,//!I ', , fi ,, ,',",., . I, I, ,/, i,, j'.',, ''I i "k ', i Il'. I1,!, ,,\. I E!.iII,((/':' ,I)/, NV\ I j t i( 1, I 1., I,,,II(Y'., I, Ii! - 1) i ,.'.q '\ 1,-,'';,%'' M f Ili'', " -1 , 4;. - -. —I );iI I J), 11,", \,,`, " ill;!,!, / - 1 I ,! i I f;;., I .,')"., i . \ -j- _ _,_. __.- ___" I, - - I, "I/((, /,,, ).'`11 "., i)!) ;,11/,..) (, I. I 7/1 ) I - J I,, — II, - i'. )/ - W) j ",,, /, /, ! 11 - / 2j,,,f;1 j..I 1. (,; ", I,,V,..,A, \ ,, I "), -;, j, I, i',, ;,,, ,,,/,, i //, /// ,[ i ., jll: ' ,,f),, 1'I ,,",,I III, 41" I , / I,,,,, I:, ., 4 ' I ,'I"' 111` i' I, -)!1 Iic ! 0 ', ",,,,,, ,//',., I:;,( (" l1.I,,,11 ".di! "11 - - -, I: pi,\,, I'',,,,, i I,1, I. 1,I., ,; ",.,li I,,F11,,,I,.,'',II ", '\' I,'' !,,It ".,11f... i l il,,, I w ,,,'P W 1, fillIf,,,"f 1,," I..,',I1,'.,", ",I.i.!i ,/"I // jVII II)l"S' I,1, "'. t 1,0,;,/ / " /; i P,/ ) )!,,Ili, I ,III,, /,iI,,,,1, i (,.!,,"',:.I"II."I,, 11 I I,,,,," /.I,.,, I I II I 1.,, I,,.,II II.h, I I; II! l ,4!;,,,,;,,!"/, HF1, 11 it",-(i,,I`,1 11. I." 1 1,, "I " /,I /if, )'',I;if.11',. ,1 111A I;,",,,,III.",:1II 01I II,;s, I I /II 1.,, i .\ r ,I, "W1;j % \ \ - ". I 1 III"; ",, II,;(iIijj,IIIkI I 41,..I- I I i I1 I,.' ,.1, \" "", I I,,,'Ili,,I II "., i,'1: ili.,,`iI 11 "i, ,, , (Yl ,I, ", f,II11/,,,,.,, ,'i ,,,(, , f ikI,t'.,, ""'; ! ",''; 111 iI II ',,. II,.N I I,/, I,,,,,,i,'; , i 1 1.i)[," ,/i", I1; 1,rl,,,,.j,,,.,,).;,',I"i I i,,1I" ,,,,,, I,,,,I,,.I.I_ij I ",;'! ,! ,!',, JAI,; (",!,ii "i ;; t!I. I !,., I,,tI IIII 1"II, Ii!\ I IT V II I I I,'.I! ) .,,)/'/,(. 1%.,.,,;",: IIIj, 1 r I i II;I, [ I,,\',,I'11,, I,, I I 111;,,, I'T )I,1,,J,,') 1., Ii) I.. 1 I [i ' ' llif I 1/1 I 1/, h, , l i ' I. I II, 1: - , 110:". I " T, )", I, ".II11-"-I, i, i i, I' .,,,.1 : ",, , i,,,,, I, 1111;, 1. 1.11'Il:j "!),I.(),'1'1 " f;, I i, I i 1 ,,' I ,I, i,, . ; I. tY, "I..'I i,,''I 11." 1 N 1., o 1''! (.), " 1, I, i ).1;,,III I''I".I.- 11. ^""~~~~~~~~~~~~~~~~~~~~~~2 I ^ ~~~~~~~~STREET BtAILS I ~~~~~~~ 1i.. Fig. 4^ llL ^ 6 NO ~ ~ ~ ^ ii~-^oooie~ls.~t~lL61s icnt Lcmtv 81s oti.41s "k,~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~x Tibripitt51s ^-^~ ~ ~ ~ ~ ~ ~~~~~Ptsbr",ocutve8~s.. iy(l- 3ls C'en-t 7ocom-s.v 688lTbs. Bosto ^___________________________________________________________ ~- ^.^^^leydeL ^th~~~~~~ofJ.BiertISO~~~~~Foac-3,wFi..T ~~~~~~~~~~~_Z:Z ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~' opXIO'~' it-z ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~ __ 2v ZP ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L~~~~~~~~~:'7~ x~FT%..:,,.'.....'..~ _H_ E~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:~-~T~;E —7::/i~EZ-E:_';?~Z-:!S2:F''~~~ 7-'-::..;~ 7~ L- b~ - -----— _7 _:. —~.:cZ:T;Z577%Z Z.7,;/72~iiPT':IZ-;:'ZJZ~ZZ11c~7;;t'E ~-;;ZL'77'7 TZ.I'_::;~;ZT;Z-_-_L_?'~7~:;"L,.LT, ~i' —Z_:I-"T-5ZZ- ~/?/// qxez'q %q~~~~~~~~~~~~~~~~~'~>/:d~ 16~~~~. A l~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~, 4i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i,~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~_~p~j'j... -'S,'~,,,I~:iiq-a~: I ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~,:',B..~a,>:,d Pla-te 8 ME R IC__ N RAIL'S and JOINT S So. -:',cic",7i R R. 5 7 ts.~CLt~~~relvejancuub~~~~ Coud b ////^//// I C&e Cut R. ////////// S ~~60 ~s.~'/7 I/////' I^PcU ffM i Ou l b' 5 I.:'~,?,,f?:,;,:,,~ /,,;,,?x?//,d!,:,.....,x,,',,,..,,.,,,,,,;/,,,,,., ~,~,~~,,,~.,~~~//~/ / A_. T,.~~~~~ ~~~~ ~ F'oHe clel ][)'RBOITCIIEIRIES PLANOFPPLESERYING TIMBERFROM I V A A, Oz, gwg N. 11% M, A 9.4 Viz- M R5 b'ON n VQ Allok'S;K,.T 14,00 Alls R IF, 15 1, -M WN Zerah Colburnael. P1. 10. ]?]~e~ [o h;~?}~/,(!ql~,s' (V''eoSo/i l'~ (] (~[~'tz7iel: |''I;OJ1' g S 011l:1~.,' O ( It'EO S 0 TING,I_,EEPER S ~E] eva tio| on I111) iH_,,f,< 5: 11''I`TI S IfllE-I~ ~E R I,, | 0/j / 11',! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l l J lil'li v A v~~~f===;; c tf L I, [, h ~~~-~,. 1 l F ~~ -- - 1g --: flalrt7~. 0111: ll~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1ii |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...........R,,;~.1-lS1 i l ~~~~~~~~~~~~~~~~~~~-r-?............1l r. =~~~~~~~~~~~~~~~~~____~ ~__- _ ~,1!?,,:,I -"'"2 i __ __ __ _-___.!__ J ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -~ —-=-'-'-'' — *_.Ia1!bg i f v== —I=-'-=- -- C- ~~~~~~~~~~~~~~~~~~~~~~~~~~.........1 " |...... il ji]|t —7- I~~~~~~~~~~~~~~~~te a, Hea tiq~e. == tJl t~~~~~~~~it~~~~'-~ I i!I Iii, 1l~~~~~~~~~~~~~~~~~~~ii ril~~~~~~i I.E ffi __ L ffif 11~~~~~~~~~~~~~~~~~~~~~~~~~~~~_ 1..'_....,.... i...-. -T — CJ. \\'!" —-? i' 1P - ~'4~., L~~~~~~~~~~~~~t~~j L__.I'' 1~,-__. jljj CY, c/.I~~~~~~~~C14 2~~~ ~~7.72737t7~e (~~~~~~~-7%:!.)(~ ~ ~ ~ ~ ~~I;rt~~~~~ —-— r —-=e_~ ~~t,','. _==~~-,!~ IJ ji' iT-___ __'...... I~~~~~~~~~~~~~~~~~~~~~, I!~ i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i t',t! Ii i!':~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i,,~~~~~~~~~~~~ ~ —- 88,, i Zi-7-, j I r~~~~, L__ iI ~.. /I~ ~ ~ ~ ~ ~ ~ ~''~...-'... i j /1 I i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............................. G-.j W~ MO. :Fiji~~~~~~~~~~~~~i NI~~~~~~~~~~~~~ 1'. I I o H Ciiiii. i, v del.Ii i ~ i il / p~.,;..~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i I~~~~~~~~~~;II r:~~~~~ l^'- -. —^..... I:.~. -.- _'~~~~~~~~~~~~~~~~~~;;.:~', [ l^: 1'-;4>:' ^.'. Ili ^~j' ^ ^ ~. ~ ~.:>-::: -'>>?-" — I' ~:':::/:.:,:'.<>>::.::>'> ^^\ ^ - ^ ^^ ^~~~~~~~~~~~~~~~.>.'' A ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~' ~'.-"' 1':4.I'~ ^ 1 ^~~~~~~~~~~~:.,^ t i~~~~~~~~~ ~~~~~ F~~~~~~~~~~~~~~~~~~~~~~~~~ ____ fc i1iim~ii~ m~i~~ii lll~nll~l~.lnll n ll'~ill..l..~... l..,. ~~. l.l~lll. l~ll~ll~ll~~ i i i i i i.iiiii~ ii i ii iiiiiiii'ii ii ~' i ~ 1 l i l ii ii~i i ^llllll^.h. ll~~~n 1.~ ~ ~. m~ l. J ~M W M^~ ~l. iiliiin iiiiii1im : —.,...,..........., _... P__ ____late 13 TYNG' S BOU-GHT IRON SSPESSIOI rAL JOINT. Fi~g I.'~l ~..~ a(~e ^~~~~:,:'',/Z://:// \^^^ - \ I I:'F ~.?... r / / I / l i' Ii __-.~~.",:- t i.i.....'.. i i: l / f' l "' -....',' i'.,i i I~~~~~~~~~~~~~~~'.T..B~~olle ~ I.. i e I i j l, ~~i /,'l, ] i — -t i,,\~,,,, i, ~ t',',','..................'' T "~, -........, I' 1''.'.!~~~~~~I~ - _!I.~~~~ 1ii:- i' ~:, i -'.' —- ----- —' —'- [i",,:i —.' _ __ I',..,_ _ _______' _._' / / /, ____,. -_, _ __ _ _ _ _ _ _, I! i,'l! / I!' 11:'1',~ ~ ~ ~ ~ ~ ~ ~ II i. //'/ ii /~!!,' i (': j~ i i /i / / i~ i I1 v,., ~ ~ ~ ~ ~ ~ ~ Iii II j/ ~i l. /1!i;" J~' ~'.l,,'" t:'i ( t\. i ~! I i. ~.. -, ~ S~ ld ",,I ~, "V':. q ~~~~~~~~~~iIIIi e / _ \ I _ i1 1!'~~~~j j- II I.~~~~~~~~.i' == i 0 i i _ i ~~I I.,.I I = W 11~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ l~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I XX,,W~f,:: A A.I.Rolley ~ ~ ~ ~ ~~~/ / el)Ij iii ~113:0ERI TLAIFl T F`KE^GCH; BAL c\\\\`~\\;~\\\:'\\\ AIuS TRIUAINBAI,C OMMON FORJM. ~~ _Z~e~L~~7;ic~-B t Al-t d, - cra Trar R.. R.\ J.FlGr7of' Weste krn., C(FranC^~~S~ s.C'ordJOva anQZC F^sishs and S'cJ~=v'icwl~o at.Tmt; I~~~~~~~~~~~~~~~~~~~~~~7 7llbbsy...k r(7l/pl.7d 7 1.0, If noll 2loplkem Iy- ~la "' ~'' A..L.IfTollov dol.;L 'RAILSS. 5 8T b. D E EHAD KAtL Mao-J&nckdJoin an d fa, If Br acklets 071 Cuzrves. Ceeln.^Ei..i'... 581b.RAIL WmTRAD S IN KI T -JOINTI 58 1-1). RML S BRAC ~ ~ ~ ~ ~ ~ ~ ~''''j_________________________ ____________________ _ _______________.____________^ _______' ~'. ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~... ---- A:-L.Hfoll.evN del. 'No~~~~~~~~~~i I *'~~~~~~~~~~~~~~~~~~N 9^2^J No. _u^^. ^ ^ ^^ 1. STEP~~-HE.NS' TUBULAR RI. -.as.iiade at the Cre s cent Ir o i Works, ^~ ~~~~~~~~~~~~~Vt.'i~, ~Va. ~~~~~~~~~~~~~~~~ ~.,.._1' ^ ~. ~ ~ ~~ ^ll^ I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.k'''~.'.,?',%'' "2,,'"~.',,,'x"CX,"'',,',,". I,',,,,,,~,' ~'~',~~,,, x.,'~'1"."'"',,x,. ""~,"','x'.' ~', x., """,O,, ~~ —— _._ —,..,,,.,,~, —,: -__ —. - ~_._. _.__.__............,,,..~_,,.__,,~.?_..,,,,,,,,.._....,,...,/~- ~_.......____._...__ _ _ __ _ _ THE PHOEX IRON. CGo's. PERMANENT WAY For the^Penn27syZ;-ia;Rin^Juwy.:. ilel Fit U siP". p L. Co'15.-Pa ctlnt Process. Fig. i L. \ XI ^^\ \ vv \v ^\ \\o\ \\\n\ e v e\^ ^ ^ ^^ ^ ^^ IIII I I I'~~"~""".C _______________ ~.'~. A.JJ.Ho~~~~~~~~~~~~~~~~~tev~~~~~~~~el... ~ Lith~~~~~~~~~~~~~~~~~~~~~~~~ofJ.Bieril80BroadwayJN.^'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~11Wt. ofpr.;yho fishes 22 ~ o14rs 71 - 771~/~~~~ 3C 7r Barlow7a (x ~W~oodywoses1 JFisJ1;ingre-Jaw chalir. Tr Wtof pair cicuJ'sq 052 lb~s.,pa7. of fishs 28'~^4sJe' ^ls;4 spike's 4 ~s ~li~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L- -— i - ---'Trj. W hs2 Uft-Iz / l ~s., 4 7)01s^, n65s ~. nrlas~mrs 5^12 17 p spfes ls,=l i~' ii 1 1~ " t''. ~~'.Adons CastIron. B achet,' 16I2 0 W lIb" L7. L~~~~~~~~~~~~~~~~~~~~~~~~"' ~, ~ _,, ~ =^~ —---— ^^^' -X Wt ff~~~pcar hncf chairs~ 6^ ~~s~.,i~o'lts (^'nuts ^^lbs.) 4 spikes' VCs. T: Ft. o I cx' i~ a, i r 44 1 s.,pn64b~ ltsi~s X 28 tn~}Vt ofpdr o fises 2 lbs., 4k es &7 4 U ~\.!,??.~ll ~ ~ halld.raS.~~~9 ~* _Plate, 18, I.Allll, BAIL1S JOINTS. I~ ~ ~ 1.~i: SlIclatirs Chair..211oc'ifca-tiornl onf lrson s' Wedge. Chauir~ \ Z-~~ -HfarTingh(XTms Jub \ ^~~~~d cets t.' -lnrute edzc~ietes 5 ~ TIcrsN O.^V.^avae^. Lvlli"i-fJ.Hivn^Oru-on.Sl-N7 -- -------- i" ICTSI^TO^'S SIII _TPI _ _ _ I /''''? —-1 ~ ~ =_' —TU —=-~'~=-'-T-' ~-= —— ~ -=3 2 -; —-= — =' —.',. _["uvp ~ 0'P?__d___d__4d- _ _,~;g i ^^'T./' T iS^ i' l _____ ____ _ 1 ^ l~i~i-^^'.^P O PZp~,.S,/'uii v 9,.T,:,,: -19, fe~~~~~~~~:::;_-:~_:=_I~_' —------- U.,, w ~~_L-___~______I ,St,(tt#u /' r(ohb/Ic /X'/. I/t/r/// o/''//z// //i 1. s/;,'''//S,.,/v, i/o//.s'.' /A,. WE 7.4111(....... h( _ I/ =s'.''(// //-' /. _' 7._._'"^i/~ _a/1 _i'~lt'i' =r _. =l-/ ——. h, ---- -----— /h',,. I;,,\' -,l'-t!~': (S( —— l \ I = L "'~:C'! D i ~f:;'S::::_ "~,I i - _- -S ==' 1~ 1.:1:n~, ivlei'^l.c\' s. i" lh(~h(^l__- ~ c'-d i - i-J-:: —' _/ o i-lil..... ".1.lB i ii =. 1141 ~_l_~~te 2:1 RAI L,lOIrIT S | ft~~ctrsa7TsJ7-iL tft 6.{/i. Adwrsts'I~ G~iD & Ke~Y fiOi/Zt. 5 t d=4 =.s'Bri dge, ~ [ Jo-zn. = ==! == sser r_~~~~~~~~~~ -,i,' " I a a!::?;," Gij1r 5 11lS.j/i~lil. i 1irirII: fi l I I 11 I ~ ~ ~ i'i*~i ~' ~If iii ~ II''~~~~I ~~~~~~~~~~~~~~~~~~~~~ — --—.-=~-~'-~ II,,~ l', ~ ~ ~, 1 41,,~!'1 11.,|..... C e Adoze~r~us'3: Jo-U~..,zcz /.~ t --- r- a\\-= —-_ DJ —--— =-_ ___r-~;,~=-.~~~ ==_ —C= -;-~ —--— 3'` nBc7~h~"7.. Jji~- wn~T ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~=._~ ~:!!i~~~~~~~~~~~~~~~~~:',\\,. = = =q,....'q',,.', ~,''?l-' 1, |~~1 W @= Le Rnetsb(J.e-Wjz Joi~nA t(hcrtizr. 5.... 1'tI 9fiI ii 11L.I —[o]_ de l,,'1 X\XaB@D~1it; de I 11 * r/ -, E\ \ W \ \\T-,A iii~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,-, -; -- E.= _.i L & _=_g ~ AL-.cS.Ho ley d e~l. l.-ith~.c~f' J~l ~icll GO -P'liltc,.l,2;t. N.'r. j RAZ~.AIL JOI-NTS. ClITae l tdZ f C Z.'ZZTurti.l~lol eFi-s7,.Tozt S, Sterl&e', Co]1 (.Inr: nftr,Jolitt. r'- w71v-. walrough.-y/~t' Ch2fau\7 Ri"eardiitr I.nffR:W 2 3. [$X~iJ~Kr-as',-t Ve Joinft N.Y..A. -Ee1. ~R fW.. ^' ^..5. I=4=9g=t 2 =EC ll.......... _\\> ~~~~~~~~~~~~~~~~~~~~~~~~~-~~~~~~~~~~~~~~~~~~~~~~~t — =rI ijl:I-^__6'________ 3-. 0... 5 = _f~~I~~I —-~~ —-;:~~=-'=-~~~-~~_;,- -X..X * — 000=f= — a;-f Dikes JointforMattredRail Ilinoi Ce..e..e is.. ^ 9 I~~~~~~~~,' g 1()2....................! ~.'1ozdi'fi'cation0 ot'^fhe 1Tr'imbie Joint7z y | i. 1 7 7{Ro c k/ i7 6iclti^ W,-. 5./ -( /~ 77 -', _-_ _ _ ^^~^l. _.~^'~~. ^??'~~~'.~__ _. __.~.. _. A..LJ y:r ~".A.L.3ollev~~~ /jjj1111 1104"11.4-09 HO'~l'77'' I~; I I4 Y. ^ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~f'^^p^M~^')'^^ ^MO~^ ^/^'W'^.V'^m~-~~ — isv.) s^)ia'tH__~i_ 'W^Tand JOJJT T S _ P~~~~~~~~e-1-t~~~~~~^ " e' 9.2~~~~~~y. ~ ~ ~,- 3 1 Am -WAYz- H T()1By, 1 @69 ~~~~~~~~~~~~~~~~~~ — S {1 —-- 1- 1 ~~~~~~~~~~~~~~~~~hqi t1 I S LA~lwm6u LI II' B ^^ ~ 1,.,.13 T't~~~~~~~~~~~~~y. /G. 1|~-Tq.19 Tivg.21. ^. \"$- \^if ^gg5.~^{ | ~~~~~~~~~~~~~~S5ea ton's TernuzzZzn^ Waty 34..5 (/t/fiiulz, 26. ^'.~~~' )l~^^ —-^~^i. ^ ^^ c3 fe-:-f~^-~~^^ o —-— ~ ^ ^.. i j ^~~~E~ ^ ~( \ ^ 11 ^^ "'~~~~~~~~~~~~~~~~~~~~~~~~~~C ^~i t.< i ~~~~~~~~~Zi~~~^ r I l l....., I ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1TI I I(A' I. I 0 L..',.... KE.... Ff,, i~i ti,I Ilm~~~~~~~ ~~~~~~~~~~~~~~~~~~L...., I' i ~i "' ii''i' i I I Ijl ii i i 1;1 ~ l/il I I 1!iii i ~l.... I i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i IW /I!i l!' Ji: lE',,ii J iliii/! i Ii iIi i i i~' ~,lllIij~ (I!6'I'''"i iii?:,:; I'' 1i! ~: i?: j i i Fidlia' i ~~~~~~~~~~~~X!,,I,.....,... i ii;; iiiiiill~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~/iiii ~!ii i i I qi /:O i,,,'/' i:~ ~ ~ ~~i?'i~, 1 i I fI I I 1 A —i, Holley del..~,'[/~iii,1in ii'., i II iliiii ~~~~!~ i ] I J i ii;,i ii!..f///!!" k l;',i i i ~' ~ {'~ ~ ~ ~ ~ ~ ~ ".... j'i I;;:!i i ii ii jj~ ~~/,i i!i; i~ i!'iii!~j/ ~, ~ i I iI / I i I 1 iI:': /'i~ ii 1:i i,,~/r~. ti11;q I~ ~ ~~ ~ ~~~~~~:Jhi...' j i i'i t"', I' i( i I j I I i i I Iij; jil/ /jj/ijjj/~jjj/J'i'h~( i 7: i'i' j1it i ~ ~~~~~....' I I I jI I i I/I /I! I i iliiiiiili!F: iiui' l: >~>lil Ii:!::! i:!, i /llilt II';'hHl i I!!II ]?' ~- ~_~.................. i/:'///:,%~-+: —. — --., IIII;~~'~ / /~ ~ - -: ill~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i -:::II:I I Iijjli I I;I I -.,i''i i iii ~iiii i I i ii Iiii ~iii i~i I ll i,ili I i,i,i,i,,,,,,i,,' i.....E.T~. Holley del. ]5-ikh. J.-Bien. 60 Fulton Street % Y.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~';' . TlaAt 26. iMlSCELLANEOlUS y.|~~, ~ig.,].S'Tig.. 4: L'4 60 Ihs. I —, — Ewm~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~a @ }h'?MS-a o' ~, (y 105cyaen.'\,'',^- ^ - ~l ~^-~^ —.-' i ^ j ___ ___. ffy'drrouene.. ~,.._x /.'. —I ~~~C07om77V~i07o of~cr vaycrogez..Effects of. 7L'-iat-1171io^07. \ ~~~~~~~7.71 fl607irie or' 64CaL7?cLn~e. Z ^ r <'.t}1. C ol bi^ i "i.~ L~T ~~ ~~ ^' ~ ""' " ~ ~ ""*~~ -— ~" —~ -- -~~~~ I ^- -- COAL 1URNIN ( LOC( OMOT BOILER,. i \_ ^ _(^~' Lond~on? d,.YOl.lt/ ll',,"e;Un A~la..t/!?[': I~~~~~~~~~~~~~~~~~~~~~~~ ^~~~~~~~~~~~~~'',X i~~~~~~~~~~~~~~~~~~~~~i ^~~~~~~~~~~~~~' 0 (O - - - - -- )'''\','~_- lilll —-i- - -/^ ^fLT " ^''' i I^^X ^ ~ ^^^ _ ^i / I.-I.-i' ^ ^ ^''^' ^'\' i'x " "''"'""- -'~', i....... ^r Z0ra0b('o0buvi' ( —----—. _/.. 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~:~ - ii ",,..-" ~-q-~ ~-k~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~D Zo'rah C'olbutmnT~ d__._. '.1^7^'/'/'..".- l~~~~~~~~~~~~~~l~ ~ ~ ~L. ^ 1^!,?~~~~~'?":~~~ ~i~~~~~~~~~.~^~~ ~~~ ~''2 1,-' - ~.' +' ~., te e. e'. i 3.' 9 9' 9 I'' 111 ~.. I - 1 3.3 1...!~ Q ^\\\\, I' _ T~\ "~\ ~ I ( I ^ 9 t tS -S I i; /.: " -.'' J~..'A 1 ~;,,::;,.. ~ I. 40i i~ ~. ~'"' I0 o-S. o; ~'7^7^ -. e.9 I3:l DP~9,~ "' j -- O O I S e~: /^. "'^~^^lp } S * I'^ X ~ ~' + ~ * I ~' ~ ~. ~. ~' *J //'/ {/,.' L-_~_ ~ 1 ~ ".''.1 1 I f \-~, -.. M...... _ _;' I i. _' ~-'i -': —:z-l-': e "c', o" (I c. I'~,G~l - ~ ~- vrCx"~ "-'"' "'x "- -,-.".,..,x'~ ~ o o oo.:`7 {; t: 1?; C' Co r.',-.~ "~'- C.......: )-:'P,' I. -'"~ ~'','//,',,','/,, ~ ^ 1 l.. } I ~...'0 i' ^~ ~' ^1.'(^' ^. ^ ^-^^',' 0, 72';'//'/___,':~' " 11 t l',ll:l""l'" ~~i.i-',// //t...'', I'1'w i /.Z 7> i~.^ ~1 X, ~ -. t~i!,;// It I,,.,.,,'0~, i ~ _,:,,,,., ~ ~, ~, - - ~ -, ~ J- - - -T. ~ - _ ~~ ~, ~, ~ - ~ ~ ~ - ~ - ~ ~ ~ - ~ ~ ~ -','~,~ ~ -,,, ...........~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'~aP-3TP''\,(~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~...... "..-II.................... c~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1111 MI.................... iei 7j Co I C~~~~~~~~~~~ 1 -1- ~;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I'll Ill-II Hii -I -i - i ta~ ~ ~~~~~ii cp i1r~~~~~~~~~~~~i i i I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~o Oi',II.,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~j ~~~~7, ----------— ~ ~ ~ ~ ~ ~~ ~~~~~~~~~~~~~~~~~~~~~~ —--- -- -- -EL ------ ~-i ii - -- - I 04 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 1-!-~9 —6 411 IVA I~~~~~~~~~~~~I rj~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........ ~,-~' ]J' l ]ili~l! i// jI "% ~~~~~~~~~~~~~~~~/. /~~//I: i;'1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~'~' I'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~IIII f/ j i i; I~~~~~~j: I j jiI~~~'i:I; ii I!!j'rtj -........................ —i I' I -~~~~~~~~~~~~~~~~~~~~~~~~~~I I I: -.....,%....~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I~j/jIIiIi/ Tt~ ~ j! t!~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l[I~i ~~~~~~ii ~') ~,).l'~', T, i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iiil ~ 1"' I "I~~~~~~J.E.McCONNELCS. I I ~ ~~COAL -BlI-,*RNING BOILE RLr~ 1 p^~ ~ ~ ~~~~~~~~~~~~~~~~~~~0. l~~a~-1 a ^' O' 0 "~ ~ — ~-'-^ —--— ~ -^^^^ —------- ------— I- ^^ ^^ ~'~ ^ ~_:)_^ ~"~- ~ ~ ~ ~ -.~~,~~,,~~~r^~-rnT~~r.^~r^^ r^-,rrr'^^TT~^,r^~^-,^rrrrr-rT^^T~~n~^,rrrmT~r^'mrr^^, - "^ ~ ^- ^^11^1^ 1)^~ ^~~~~~~~0 io'ii^~ i ~^ ^ i ~ ~ ~ ll:IIIIIZIZIIIZI'llllllj ^.-^ ~ ~ ~ ^ ^ ^ n n i e) ~ n e ~ i ~~ ~ i 1~' I I 0, / ^~^~~~e~~~~ei~ ~ i~^~~^iie^^1~^ I I~~~~0 l. i i^^^~).. _____^__^_~_^_,._^~~~~L. ~ i - - i a~~~~~~w.,~~~~~ ~ ~ Nf\y dc'].e ^ i^ ^I '.:'o0 dJ:'J ~zo??ooV'~ a..00dd"I 4Jo.'.ldl;)t A:.............o ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~_o <'. "^ n ~ ~ ~ ~ ~ ~~~~0 n a ar] O1C-1... GO oY o( el oi10 0.1 10CI0:3^3^:,,~ ~ ~ ~~~"'..~..............o,:: r'^E a i- FJ~ 1^ ID t~ n an^ o o c ^ oe )o o ^ -~,'i(~~~~~~ Saa ) I r' J n I, c n - o o (^o o': i: C; n,i It O O CC) OZI'H FD GDI IF E0 0 0 0 0:0 0 0 ) G 2 a 2: aM:e la a a r^ ^ o! o -o oooo ~: o ~o l S' Jo(':o,' t7 r Ci Ml ~ a Ir i 5',C 0-_'^n c a a no o o o Q c~. ~~~~~~~~~~~~~~ C ( 7mm I O^) 0 O C 0) ( 0:0? 0 n0 - -. D p: ~ \C-0 [~ 0 Oi: x'. hI %, o:. II........ ]..:. /~ ~~~~~~~~~~~~~~~~~~~~ ~~~ ^i^) (DI 0, C. C.) ^ Cy^^^^^; O ^ 0~0 000 /O 0 ^~~~~~~~~~~~~~~~~~~~~~~~~~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~' ^^' o^O O\QOil o o'''3 o0 oO Q~0 -l:~o oi i oo^ o: (~:, i...o.o.. Jo!' 0 C) 0 O (D C ~i.................: 1 o 0 0 oG OO OOO TY? t^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r I ) ) ( ~'" 1 i (Q (p (^)''~ ^'~ ~ ^'~ ~~~~~~~Q (P i ~~~~~~~~~~~~~~~~~~~~~~~~~~~\?:.:.) ^; /:,:.15/ ~~~~~~~'' {DO' I I-I........__: ~- _.. 6~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~._ L_.':-' ~~~~~~~~~~~~~~~~~~~0.... I'~ C —~~-:~ ~~ ~ —-::-, k.. -z —-Z' —' —-' " 4 -—....... ~.'- -—.:..-~..,,._~._... L~~~~~~i...... it"~~~~~~~~~~~ ~ ~~~~~~~~~~~~~~' -'- -.-x —r —.- --- _: C B-AIs COA tL BURNING- BOILER RJOtIN BEWR CES COALN BRN BOILER. Aa nche fe7\ S/zeie. l d L IncolnMre W. Patented Ocl/tm.u. Longitudinal Section, of-ireboax. Scale 1' 7. /0 07ne Foot. j i ~ —~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-~~~~~~~~~~~~~~ —-------- I -- \ — ~ ~-~j —i BC~~~~~~~~ —----------- Zerah- Colb-urrn cel. 1........_~. _...... I ^ ~~~-~ ~ -~~~"-~-~ ~ —.-'....................... i~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 11^.,~~~~~~~ Id I I ^ ^ ^ ^ ^ ^ ^. ^ ^ ^ ^ ^ ^:.........~ ~............. i~~~~~~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ II:' jI %ll~ll a: I i,,l,l% % ~% % ^ <&'<& ^ "& ~ i %'........i..................... i ^ ~^' ^ @ @ ^ ^ ~ @ @ ^ @ ^ @ @ @ @ @ @ @ ^@ @ @ "^ @ ^ ^ ^ @ "'-lr: —~" _:"- -'-.._-::. ~' ~ j;^''^C"''' ^ t /O~^ ^ ^ ~^~^ ^ ^ ~^.@ ^@@^^ ^^ @ @^ @@@ @@^,^@._'~_ ~ j;'CG'_CX^';lC)J~ \. ] I!^ ~ ~ ~^~^ ~ @ @ ^@ @~^ @ @ ~ ~@ @ @ ^' ^'^^t / ^'^i^J i p^^CC-.).-;^. \k-11-1 to \C.1- ~ "& ~ ~ ~&~'~Q % % ~ % ~~ ~ % @ ~ ~ @"'."& @ ~ ~~@ 1 9 @ ~ ~ @ @ @ ~ ^ ~^f ~. " ^^;! ~~ Q..O..L....^.^,...I .._~~~~~~~~~o[P -,F. 0~~~~~~~~~~~~~~~~: ~"'. i 00~~~~~~~~~~~~~~~'.0 O 0 0 0 WIII~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,' X\ Hill~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~/ll,, W_,i~:::::~~~ 0~~~~~~~~~~~~~~~~~~~~~~~~~~/ 0........-..... _..... 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~/'"D / C. Scl J,/6 to.',w' 0 0 0 (D' 0 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.... ~____~ Tii ~ ~ ~ ~ ~ _Fq Illi~t'!!!ii~i~i i i~l'l iltlllll~EAT ON" S ODO RO ABU N GL 0C(i)_!0 T IVE,.,AN~~~~DMARKS' SMOKE S_ _TACK_,f, illtllili~~~~~~~~~~~~~~~~~6-wl evo'^-~ ~~~~~0 000~ 0,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............ T' ~ ~ ~ ~ ~,' i. I IIIIIIII~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~t1~ ~~~~~~ ~ ~ ~ Hl 11 //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~w III/ I11111/1 II I ~ tL.E...'R<.~. E.....~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ —------- -------------— ~~~~~~~~~~~~~~~~~~~~~~~~~~ ----------- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~?"~0o0oo0850 ----------- ---- --- --- --- --- --- --- -- --- --- -- ------------ -- --- --- --- -- --------- --- ------- O 0 ~ —- --- //~~/ \ x\.' ^ ~' _~~ ~~~~~~~~~U )-. - I J^i __J _li ~ bjt'1~i \ ol:..... \ i ^fj o7'P r -'^-ii' eI _~~~~~;;.. -' o, -- { L -; ^~1. ^ ^-^-^ ( ^=^^-~- i I ^ r' -: i'.__^ ~'. 3o_~~p~~, &:_~3c: /~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~/'/ / O~~~~~~~~~~~~~~~~~~~~~~~~ II ""~'......... (- I 9^ ~~~~~~~~~~~~~~~~~~~~~~~~~~_/,"""""'9: ~'ld... evil, ----------- ~~~~~~ ~ ~ i ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I= I...l ll1... S ~~~~~~~~~~~~~~~'U10 A IaIiiu- "41LL '''P'oa-9 081 LU -0' P jo'l. /0 ~v'. - -. ~... I I~ ~ ~ ~ ~ ~ ~~~~~~~t.-^.... f....... 0' ~..... H...II )I.......... El; I' ^ ^illH.U.V SJI'VII~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~j^S~~~~~~~~!Ii ~.,~' - ~',- I 1' f ~P~ ~i~~ I ~'^^'^ ^^^^^~ 0 0.0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~' 0 2- 0 ";.. C)#,, - -",'. I T H E B O A R D M A N C OA I. B V~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~ lN IN G B O ILE R I~~~~ A1 131 11N U f ^...,....,....^^^-~^h -.^____^^ -.1 l~l 7-, illp —~~~~~~~~~~ I ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-jl A.L~~~~fioiley~~~f (1e i II;~ ~ ~ ~~~~~~~~F C)~ ~ ~ ~ ~ ~ ~ ~~~ ~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~OOd' /, (D~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~, 000 o S i i, CO I i oo r; 13 30 jI j / i,4> —CJ[10 1O e C-1O 00 0O C; ~O (C~7 V: O (3 O 0 0 0 C;3''3 / i~~c~o Ij~i ~~ O O () C) o Cc(!ct "D C'):c o o 0 r 0 00 cZ~ca Ic0 o C)~ 01 0 0 0 0:00 C 00 0 C C)o~ 0 0 C) 0 C) 0 O otC, 0 0 Cj ( Dj C D ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ i F. P. IDIMPFELE'S COA~L BIURNING BOILE R.186' Tzbes _l,'4"ou.tsidl, edCiamzfe.&r... —J "'Y - "~'~-'~'~ —'~:=~": F,-,,,_'~L?..................................................... /'"I, I' I~~~~~~~~~~~~~~ O O 0 O. O D.? 0 0;/ O r:2, O',.f C) o C), o (,. o,' o o i~ ~ ~ ~~;'/ ~' O O ~~~~~0 0O 0 0 O 0 0 0 O iI x-~ " /1 P DI _ e C P. t?. P i mqZe l a el_ M..W..BALDWIN.C~2C c: C o alB urain g B oil er, --------— ^ ^ ^ - ~ ~ --- i o'( A.O(~.hV!.ol.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L~~f'J l(} J) ~I, f,,;\. 0 c._M~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~d. et. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ - ---- ! —— ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ —--- i: f. c II~~~~~~~~~~~~~~~~~~~, i; H~~~~t.~L't~ TES's COA H MOR WO1) B-1;ttN3rITN 1/0011,E~il ~I i)~L-,I3C) 0. > JD h 0 / / C 0 0 O C C) 0 C.)c O -L VJ':i~~~~~__.~.......,.....,.,a: C)I' Di C Pd. I i F~q1l 000OO iao1 I 0 O'O C)1 j C) (D C)~. C. C, -. 00' 0OU)C C~,~o DO) 0 i C? O 0 C) 0 C C _0 0(\~ il 3OC,O L ino~"~ C) C) 0', C) C) 2! C, O O Ci C!,,,,! 7'7,,'!.,', ~.dl " ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~'I -.. 0;~~~~~~~~~~~~~~~~~~~~~; ~:!-'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ f-...... ~,' (~f~~f~~-~i ~ ~ -~ 1~ ~~~ ~7j C~~~~-h i ~ ~ ~! ~~- — Ff —-—:;l —— ~-'-'- ~~~~~~~~`l~~'-j,~~;, —---- ----------- C I b',.:~,,'.5~~~~~~~~~d 1 C~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ C7 ~ ~ ~~~ ~~.3(:'00 ).',. i' 3 / i_,~~~~~~~~~~~c,:''.~.T,.:,~...'''I 3.,''',?Zi?'' -....~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ~~~~~~~~~~~~~~~~~~~~~~~~~~,, -/^/el.: 2.. JAMES MIHL-OHOLLAkNIDTS. ~ if^^o oooo \ nAntlwracite CoalBli rinuiPas sender Loc( o oieO O C) (ECTilll~~!. G0o 0000000 -P'J-delphia.)0(0`0 0^ 000 _Po~d~o2~1o^~: n^~~ //l~: I ~a'7''~. 00000^ DOOOQOOO D' 00^00 -0000000 00000 0000000' 00 (DOC 000000 0000 O0 U~ —'-' ^ x ^ \_/ _17~~~~~~~~~~~~~~~~~~~~~It F~~ i ~ ~~~~~~~~~~~~~~~~~~~ ~ s ~oo,_,o'aoooo'' ~7'L~"z~Z ~~, i~~~~cl~~T alla I We I.' O ~C~) ~ O / v ~~~~ C~~~~~~ C 9X~x_~a rlV~OO3 LI.. Ii? ^~~~~~~~~~~~~~~i Z ___ ^.w~w^ ^ ^^^^^^ ^ ^^'/&!^~ ~ ItI~ ^^lT~ i~]n1111^^''~L^ \\~~~~~~~c It i.i! Wlll^^w ^_ \\ { i ^ ^^ l!^ ^^^^^^^^ ^^_^_^^^ ^^~~~~~~~~-co ^^ - m ~ ~m ^ ^ rr"' ---------— ^ —----------- ---—.-.. —, —.^.^~ —------- "^^^l~~~~~~l^ ^ ~ ~ ^ \' ^~~~~~~~~~~~4 lure ~ ~ ^, —-~ ^ lip~~~~~~~ Fij il'k^~ ^..^ ^ ^^^^ i ^ __^___^ -_ __^ ^ ^ ^ ~ -^.' i __' ^ ^'~A l i ^',' ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ i ~ ~:' c ell~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~;. Oo~~~~~~~~ % O.i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ~~~~~~~~~~~~~~~..~.'........~. 1..~. r. i~ ~~~.'-.............~~~~~~~~~~~~~i ~.,~ Ail.,''\\\~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ----- --------...................... ~~ ~ ~ i BEAT'TIE'S FEED"WATER HEATER. gi I''___'' IT[' -Z~~~~nd6VZ, 4~~~~~~~, SOMA& Ve~~~~~~~~le~~.fif oJ-Be-nlSwaoy-ay-.T Fi~. 2.I AX ~ ~ ~ ~ / 2,.~~~~~~ I[ —~~~~ ~~~-~!. — -----— ~"Jl~~' ~~.................. -13~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~_~:.o'j.~e.oLO.'_O~.~,-~ "T~ F El E iN k ii'1i,..i'i. -,I. ~~~~~~~~~~~~~~~~~~~~~~~Cok, ilp i~~~~~~~~~~~~~~~~~~~til 1~~~~~~~~~~~~~~~~~~~~~~~~~:'3 Ft~~~~~~~~~q~~ 6s! ------- ------ F~~~~::ig..:'e:, t:1m!111,A ~~~~~~~~~~~~~~~~~~~~.............' ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~7- -~ —+ —- --------- ~'................................:~~~~~~~~~~~~~~~~~-t,':,:,~~~~~~~~..... {!~.~ -—:: —:.',...tl q.I._-_! I. -".~ l PI. 61. 1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ j Il Ol T- -< -dfLs 6Sow I~~ ~,1 Irs inPj __________ |^^' af " ~ c\ a 8:' _: LW4-.XVj1 — 11 ~ ~ 0 o.- I01 ~I ~LV~1 ~III~. f ~ NoU f^__g ^ L'c r._ l'< ~''.'l~i'.~~~~~~~~~~~~.~ ~olll-,~ ". P~~~~~~~~ 0t " F~~~~~~~~~ ~~~ ~~. _~, ~ _Fad T41 atew 17: 8 9 10 Jus. T F" Ck-IFFAIR Scale -'116 Yt, _FHJ'Caj SeCtiOM Of -17T INJECTOR j 6(40 T 0 H A for I;pe- dun' ioi s as -ilt by -ha S PStewaTt Co................................ MMWL Xa —acli-ester. IM-H Tt' q Fail _VeP city Feed 141'0'11 7. TrallS -Se S, W7 C fia-,-,tb e7F 8. Tidt Si'_-e Sechon ol, ve- "Srizc, X,? 6, XOeomoti 0 JCFig..8. x N N X it Li X, ----- ------ s 2_11 xx z Xx HIM 11 ZZ X, ",/X ZZ NNI IN -Fuj"l S1,7V _A.L._Y,,)17,ev del. er GIFFARD' S INJECTOR, /~:::1:~! ll:!i!) PATENTED APRIL 24.1860, W BA;ILLAM- -SE LLE RIS & Co SOLE:AKERlS. l~~~~~~ft ~~~~~~IN THE UNITED STATES. Sieav-n, h-071, T4 0 fill!~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:::,.............. ~:.....I'~~ ~ ~ ~~~~~.................': _________________________..________ i 6'5 W.J,~ y-a.el. ~ ~~ ~ ~ - ~~ ~'' ~~ " ~ ~ ~ ~~ ~ ~ ~ ~ ~ ~ ^ ~^. ~~ -~.-..'~'''' ..Plate 64 31z B OILER IMAKING. 1. _________ ^ ___________,i3 r - 39i -~-^ —--- ^'t^T ^~^ —-^~ i_. 9~~~~~~~~~~~~~~9.,~ —~ 1~ —----— I I~ -'\@. ~ ZO~~~~~~ ~ __ _________ Jf Q^ ^^^__ —~_______ ^_ ______. I.. ^____________________________j~~3 f ____________ ~~^'0-^ ~~^^ ^^^7^7- Bifill H ~ -^ ~~~ ~ ^^^^^"" —----- _____________^_______________ S..Z4 111 j________________________________~^." -A.TJ.~~~c n-e HolierHR (iei ' ^, ( I' i ~'.i I ^' * JL.I..i~~~oile^ del. ~ .nc~cc~tt~:e:~~ r~3~~in-c~~. Plate 66. ---------— ~~~~~~~~~~~~~~~~~~~~~~~~~,,-~-I —~~I-~ -- ~ (I _ I: 2 i i ji'' i''! l j (il~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~!il IK~~~~~~~~~~~ d lil iij~ | Fig. 3 m ff 7 tM~~~~~~~~~~~mpe74Aga~~~~~fz7C. l. 4 1: 1 10~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ "' I \ r I i'' iFig. 8 1@ X I 9 I~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. t~)X@ I I ~~jl;~:4~~~~~~~~~~~~~ i XV X: l l t! = = X I I 1\ t>; i S 11~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~11 -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I at~:I -t 1C 1 i iiii I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~: i. = i )' Fig-i,(l~lcnre~e~l.. _.1.67 lISCELLANEOTTS I W9 -' T. 2 -Fig. 6 1-9 13 -Fig, ~ ~ L-._ _^. 1^_________ __ _ —L l _ _ __ Ol~ l ^^' ^7g ~ j -— ^, ^ ^~6 —fl~- l~..^c-__-~f -&M^ 11'-F.:12 II. ~; —-. -\ ^ 1 F'w. WORTHINGTON'S iDP]:_EX STEAM[ PUMP. a~2 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ j-~/v*. f~~~~~~~~~~"~e,^^: ax,,x' %:)=, ~5',y. 1.^ ^ i~-: ~~.-:sl( —-—.,/ ^ THE BISSELL TR-CK. I __FrL~ocomnftvtge. Pwc(-byT.L-Le Bsa[s N.T. 11 I~___ _ _____uii —-IIR_.__........j ~ ~~.. ~~i:[i gi.,. _____l~ T I _ ii V- ~ —.~: M'~\\ i'^~~ ~ ~ ~ ~,'. ~ I j-' ^^ -ig -. — - ~-~~~~~~~~ ~~ ~~~~~~~~~~~~~~~~ ~ ~ ~ ~ ~ ~~~~~~~~~~ ~:'"L','.. ^'^-i.^ /!!.ii!i~iii! r -l.. -:- i^ - all 1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i 7.?/ e.'cthe:lien del'. L it. of.'l.Dim'60.L~ lton,~':