TF 505.H15E27"''"'*""'"-*"'T llSlw.i,'iS,^»'»Pefation CORNELL UNIVERSITY LIBRARY ENGINEERING Cornell University Library The original of tiiis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924004853028 EPFICI|tT RAILWAY OPERATION WORKS BY THIS AlHOR American Railway Management John Wiley 4 SonS. f? Published by The Macmillan mpany Restrictive Railway Legislation. 905 Railway Corporations as Public Sivants. 1907 Problems in Railway Regulation. .911 EFFICIENT RAILWAY OPERATION BY HENRY S. HAINES MEMBER AMERICAN SOCIETY CIVit ENGINEERS; MEMBER AMERICAN SOCIETY MECHANICAL ENGINEERS FORMERLY VICE-PRESIDENT AND GENERAL MANAGER " PLANT SYSTEM " OF RAILROAD AND STEAMSHIP' LINES ; ALSO COMMISSIONER SOUTHERN STATES FREIGHT ASSOCIATION ; BX-PRES- IDENT AMERICAN RAILWAY ASSOCIATION THE MACMILLAN COMPANY 1919 All rights reserved 5 i (^ COPTEIOHT, 1919, bt the macmillan company. Set up and electrotyped. Published April, 1919. Nottaooti 5|reB8 J. S. Gushing Co. — Berwick & Smith Co. Norwood, Mass., U.S.A. PREFACE After many years of experience in railway construction, operation and administration, the author has pubHshed a series of works on those subjects, principally in relation to the successive stages of regulation of the railway system of the United States by means of legislation, both State and national. These works are largely the outcome of the author's own personal knowledge, experience and research, and reference is herein made to them; in some cases, where the same matters have been discussed at greater length in the former volumes. It is the aim of the present work to describe the progressive develop- ment of efficiency in the operation of the railway system of the United States, as compared with similar progress in other countries. It has been prepared at the suggestion of many persons who have inquired for author- itative works on the subject, that would be of value especially to students in technical schools and to junior railroad employees, as well as of in- terest to the general reader. With this two-fold object in view, there have been added, in appendices, very complete tables of statistics and much strictly technical information, drawn from official sources. Ref- erence to authorities and explanations of terms have been given in foot- notes. The work is devoted to operation, as distinguished from administra- tion. It deals with facts, and not with opinions; therefore, it does not discuss matters of finance, or rates or labor questions. The author desires to acknowledge the valued assistance of railroad officials and friends in bringing the information down to the present year, and for much that is included in the chapter on the use of raihoads in war-time. In the revision of technical matter, preparation of index and addenda, correction of proofs and general supervision of the work through the press, he has had the aid of Charles A. Hammond, C. E. GWYSANEY, Lenox, Massachusetts, June, 1918 TABLE OF CONTENTS CHAPTER I Evolution of the Railway Primitive Ways of Coramumeation by Land . Development of Water Ways Early Origin of Railways ..... Advent of the Locomotive and tlie Railway Epoch Railway Aid to National Development . Growth of Railway Building in the United States . Transcontinental Systems in the United States and Canada Railway and Water Competition in the United States . 1 3 5 6 8 11 15 16 CHAPTER II Railway Efficiency Increase of Efilciency with Increa,se of Traflflo V(, •, • • • ^^ Volume of Traffic by Territorial Districts in the United States , . .20 Mileage and Tonnage . ' . ' 23 Electrification of Railways 24 Departments of Railway Service 32 CHAPTER III Motive Power 1 Locomotive Efficiency • • • .35 , Coal Handling and Mechanical Stoking .43 , Steam Economy '..,., .i 45 The Compoiind Locomotive and Superheating . . . . . 47 . Wheel Arrangement and Design . . . . . ... .50 I,, Recent Improvement in Locomotive Design — The Articulated Locomotive . . . . . . . . . . . 53 ' Additional Features and Adjuncts 59 J. Maintenance and Standardization 62 , Locomotive Tractive Power in the United States . . . . .64 ^ Economy of Service 66 1^' "Pooling System" in Locomotive HandUng 68 Electricity as a Source of Motive Power 73 Transmission of Electric Power 75 Electric Tractor Design . 80 Recent Installations of Electric Traction .83 Comparative Economy of Steam ,and Electric Traction . . . . 86 American and Foreign Methods of Electric Traction . . . .88 Different Applications of Electric Traction — Gasoline Motors . . 89 H^ Standard of Motive Power Efficiency 92 vii viii TABLE OF CONTENTS CHAPTER IV Rolling Stock ^^°^ Early Forms of Railway Cars — English and American Types . . 95 Sleeping Cars 99 Development of the Running Gear — The Bogie Truck . . . 100 Draw-gear and Couplers .......-• 102 Further Improvements in Automatic Couplers, End-Platforms and Vestibules 104 Hand Brakes and Power Brakes 103 Air-Brake Control of Long Trains 107 Maximum Brake Energy — Safety Appliances 110 Car Wheels — Chilled Iron; Steel, Forged and Cast . . . .112 Steel Construction for Car Trucks and Frames 116 Freight-Car Capacity in the United States. AU-Steel Freight Cars . 118 Comparative Efficiency of Wooden and Steel Frame Box Cars . . 120 Statistics of Freight Cars in the United States 121 Steel Passenger Cars 123 Steel Girder Frames and Other Improvements 125 Advantages and Disadvantages of Steel Car Construction . . . 127 Transition fro in Wood to Steel in Car Construction — Cost and Economy 129 Rolling-stock Defects 131 Comfort and Luxury in Car Design • — Car Lighting .... 132 Electric Lighting ........... 134 Car-heating Improvements — Ventilation : — Painting .... 136 American Inventions . . . '.'".' 140 CHAPTER V Part T. Substructure Roadway Railway Location and Right-of-Way — Economic Alignment and Grade 141 IncUned Planes and the Rack-rail — Ruling Gradient .... 144 Effect of Curvature . . . ... . . . . 146 Railway Construction 148 Viaducts and Trestles 151 Bridges — Early Types I53 Tubular Bridges I55 Trussed Bridges — The Steel Arch . I57 Suspension Bridges IgQ Cantilever Bridges 161 Draw-bridges Ig3 Recent Examples of Bridge Construction ' I65 General Principles of Bridge Design 167 Bridge Foundations — The Pneumatic Process ..... 170 L Cement and Concrete. Reinforcement I7I 1, Tunneling. Ancient and Modern Examples I73 Tunneling for Railways . I74 , Methods of Tunnel Construction 176 Tunnel Work in the United States and Canada I77 TABLE OF CONTENTS ix PAGE yy Tunnel Timbering and Lining 181 I Cost of Single-traek and Double-track Tunnels 184 ,. Economy in Time and Labor of Modern Tunneling Methods . . 184 Ventilation of Tunnels . 186 Recently Built Timnels in America and their Ventilation . . . 192 Sub-Aqueous Tunnehng Examples of Construction 194 Compressed Air Method 196 The Pennsylvania Railroad Tunnels 199 The Sunken Tube Method 199 The Detroit River Tunnel 200 Railway Underground Approaches and Subway Systems in Large Cities 206 Elevated Railways and Subways in New York City .... 208 Special Featiu-es and Ventilation 210 Proposed Tunnel under the English Channel 211 CHAPTER V — {Continued) Part II. Superstructure Roadway Development of Track — The Edge-rail. Wheel Flanges. Bull-head Rail 214 Early American Superstructure. Strap-rails and T-rails . . . 216 Rail and Joint Fastenings . . 219 BaUast 221 Steel Rails -^ The Bessemer Process 222 The Open Hearth Process 223 Rail Failures. Standard Types and Requirements 225 Track-work in England. Ties or Sleepers ...... 227 Pot Sleepers. Steel and Concrete Ties 228 Tie-timber and Preservative Treatment 230 Track Efficiency. Wheel Loads. Tie-spacing 233 Rail Stresses and Wear 235 Problems of Curvature 239 Switches, Frogs and Crossings 240 Maintenance of Way 243 Gauge of Track >. 245 Change of Gauge on Southern Railroads 247 Different Standards of Gauge 249 Turnouts and Sidings 250 Double-track in the United States and Great Britain .... 252 Electric Traction Requirements — Third Rail and Overhead Systems . 253 Monuments for Curves and Right-of-Way 256 Signals- and Interlocking Plants . - 257 Automatic Signals and Track-circuit 259 Electric Interlocking 260 Position-light Signals 262 Signal Statistics 263 Snow Sheds 264 Water Stations and Tanks. Track Water Trough .... 264 TABLE OF CONTENTS PAQH Locomotive Houses 266 Coaling Stations 267 V Coal Handling and Coaling Plants 268 ') Grain Handling 272 :;5 Ore Handling 273 X. Dock and Harbor Facilities. Miscellaneous Freight .... 274 Freight Sorting Yards. Gravity Switching Yards .... 275 Delivery Yards 281 Passenger Yard Tracks 281 Freight Houses. Freight Handling. Warehouses 282 Station Accessories. Scales ......... 284 Design of Station Buildings 285 tC Union and Terminal Stations 288 Train Sheds 291 >^ Ferry Terminals 292 Special Requirements for Elevated and Depressed Tracks in Large Cities 292 Reconstruction of Grand Central Terminal, New York City . . . 294 Difficulty of Providing for Future Growth and Expansion . . . 295 CHAPTER VI TBAFriC Preliminary Definitions .......... 299 : ■, Safety in Railway Travel. Accident Statistics 300 CoUisions and Derailments ■'. 305 Unexpected Stops at Unusual Places ....... 306 Space Interval, How Best Obtained ,i 308 ResponsibiUty for Railway Accidents ! . . •. • • . . • 309 Grade-crossing Difficulties. Electric-traction Dangers .... 310 Expedition of Passenger Service .,..■; 312 High Speed in England and the United States 314 American Methods. Sleepingr-car and Dining-car Service . . . 315 ■ Sanitation 318 Station Conveniences q,nd Luxuries . . . . ■. . . 318 Ticket Syste.ms , . 319 Baggage Handling 321 Express and Mail, Business , . . . . . . . ; . . 323 Variations in Volume of Traffic , 325 Density of Traffic . . , . . , 327 Efficiency in Freight Traffic. Loss and Damage 328 Safe Transportation of Explosives, etc, 332 Through and Local Freight Handling;' 334 Car Interchange and Car Service , ;, 335 Efforts to Increase Car Efficiency. Car Detention .... 336 Car Shortage , 338 Means for Preventing Terminal Congestion. Seaport Rivalry . . 340 Need of Central Authority to Regulate Car Distribution . . . 342 Action .of Interstate Commerce Commission in Car Service Regulation . 342 Special Facilities jfor Transportation of Mine Products, etc. . . . 343 Refrigerator Cars. Cold Storage . , 344 TABLE OF CONTENTS xi ^ PAOE Refrigerator Transportation for Dairy and Perishable> Products . . 346 Pre-cooling and Heating Arrangements 347 Transportation .of Petroleum, etc 348 Special Freight Equipment. Private Car Lines 349 Jlailway Delivery Service in England . 349 Water and Rail Combination Service — Car-floats and Car Ferries . 350 Railway Control of Ocean Steamer Service 352 CHAPTER VII Transportation XyRelation of Transportation to Other Railway Departments . . . 354 ' XFundamental Principles . 354 ' Sources and Causes of Resistance to Train Motion .... 356 Means for Overcoming or Lessening Friction in Train Movement . . 357 Rating and Tonnage Capacity of Locomotives 358 Use of "Pushers" 359 Tonnage Distribution in Trains. Tonnage Rating .... 360 Speed Requirements. Velocity Resistance 362 Brake Action and Brake Efficiency 365 Reduction in Train-weight. Effidency in Train-loading . . . 366 . Light Loading and "L. C. L." Freight . . . . . . .369 Train Make-up and Average Car-load ...... 370 Train Mileage Statistics . . 373 Large Cars and Fast-freight Trains 375 - i Economy of Full Loads . . 376 c5 Classification Yards, Freight and Passenger , 376 Zones of Suburban Traffic . . 379 , Switching Service ........... 380 Wrecking-train Organization ......... 382 Empty-car Mileage 382 Train Dispatching 383 Time-tables and Train-sheets . . 384 - Standard Time 387 ,. Uniform Train Rules • • • • 389 Train Rights and Telegraphic Dispatching. Train Stafif . . . . 391 The Block System ... . . . . . . • .393 Automatic Block System 396 Electric Signaling 397 . , Automatic Stop 398 Interlocking Improvements 399 CHAPTER VIII War Time , ,, , _ Military Transportation by land in Ancient Times and Pi^evious to the Railway Era 402 Early European Use of the Railway for War Purposes . . Development of Raih-oad Strategy during the American Civil War Military Operation of Raib-oads in the Civil War .... German and French Handling of Raih-oads in War. 1864r-1870 . British Railway Preparation for War, and Subsequent Experience 404 406 410 416 419 xii TABLE OF CONTENTS PAGE British Railway Operation during the Egyptian and South African Campaigns 420 Armored Cars and Hospital Trains .....•• 424 Railway Experience in Russo-Japanese War 426 French and German Railway Preparation Previous to the Present War . 428 French System of Military Railway Organization and Operation . . 430 German and French Railway Operation at the Beginning of the World War 433 Railway Operation in Great Britain and Italy 434 Light Railway Construction and Operation in France .... 435 Special Railway Equipment for Heavy Gun Transport .... 436 Railway Bfilciency in Warfare 437 Conclusions as to the Value of Railways in Warfare .... 439 American Experience in Civil and Spanish Wars and on Mexican Border 440 Cooperation of American Railway Association in Troop MobiUzation . 442 The Vast Field of American Railway War Operations .... 445 General Conclusions and Recommendations ...... 44S CHAPTER IX Operation Qualifications and Needs of Railway Operating Officers .... 451 Value of Engineering and Railway Technical Associations . . . 452 Conservatism in Adopting Improvements 453 Statistical Methods . . . . . . . . . . 455 Use of Averages ........... 456 ■Railway Statistics . . . . ■ 458 ' The Ton-mile Average and Passenger-mile Unit 459 Transportation Units of Practical Value 462 Operative and Economical Efficiency ....... 464 Railway Organization 466 Divisional Organization and Authority 469 Purchasing and Distribution of SuppUes 473 Evidences of Inefficiency 474 General Management . 475 Inspection Department 475 Railway Ehgineeripg — Construction Department — Improvements and Standards 477 Theory, Practice and Experience 478 Advisory Boards 473 Distribution of Railway Personnel 479 Training for Railway Employment 481 Methods of Employment in the United States 4S4 Requirements for Successful Railway Management .... 485 Railway Improvements Originating in the United States — Recog- nition by the International Railway Congress .... 4S6 APPENDICES * APPENDIX I. Mileage Statistics ......... 489. II. Locomotive Statistics ......... 499 III. Rolling Stock Statistics . . . . . . . . 514 TABLE OF CONTENTS xiii APPENDIX PAGE I-y. Altitudes, Tunnels, Bridges 530 ^ V. Rails, Signals, Yards, Stations, Improvements . . . 543 VI. Accidents, Traffic Statistics 584 LVII. Transportation Data 599 VIII. Railway War Service 642 IX. Operation Data 654 Bibliography 659 Personal Mention 663 List of Railroads and Railways 665 Index 669 EFFICIENT RAILWAY OPERATION CHAPTER I EVOLUTION OF THE RAILWAY Primitive Ways op Communication by Land: The essential feature of commerce is the distribution of commodities among persons and places, between producers and consumers. In a small way, this purpose may be effected from hand to hand between the persons themselves in the sanae community. Biit the community itself Since 1882, it has been toll-free and, down to 1906, it has cost the State $56,000,000, or about $163,000 a mile. For about six months in the year, the Erie Canal is closed to navigation. Its commercial value is further limited by its restricted dimensions, which are now being enlarged at an estimated cost of $100,000,000 with, in ad- dition, unsettled damage-claims amounting to $62,000,000. The cost of maintenance of the Erie Canal in 1909 was $672,000, and its commerce amounted to 1,600,000 tons of way-freight and 436,000 tons of through-freight. In 1914 its entire tonnage was but 1,316,000 i;ons, equivalent to 72 car-loads of 50 tons daily or to 36 car-loads each way; less than a train-load per day ! The tonnage of one train of 50 cars of 50 tons' capacity exceeds the tonnage of three barges on the Erie Canal of 800 tons, and the train can make the trip from Buffalo to New York and return to the elevator in Buffalo in five days. A barge requires six weeks for a similar trip. While three barges are making this trip with 2400 tons of freight to New York, one train of 2500 tons' capacity can make eight of such trips with 20,000 tons. The relative social efficiency of the rail- road to the canal is, therefore, as eight to one. The distance between Buffalo and New York by the New York Central Railroad is 440 miles. The export rates, including elevator service at both terminals, are 5^ cents per bushel on wheat and. 4f cents per bushel on corn ; equivalent, respectively, to $1.83 and $1.70 per ton, or a rate per ton-mile of 3.11 mills and of 3.09 mills. ^ 1 See " Manchester Ship Canal Stockholders' Association," Statist, July 3, 1915, for a more reliable comparison of the relative commercial value of railways and canals than can be obtained from reports of canals operated as public enterprises. 18 EFFICIENT RAILWAY OPERATION The Great Lakes bear the greatest burden of commerce of any interior water-system in the world ; yet the season of navigation upon them is much shortened by the rigors of winter and, for over seven months an- nually, the communities upon their shores would be deprived of the facil- ities upon which their prosperity depends, were it not for their railroad connections. As a fact, the commercial service upon the Great Lakes is maintained principally as a feeder to these connections. The coastwise service of the Atlantic Seaboard is similarly associated with railroads from seaports into the interior, that thus maintain competition with the all-rail lines radiating from the principal commercial centers. The transcon- tinental lines have now to cope with the Panania Canal for the traffic be- tween the East and the West. In this competition they are handicapped by restrictions imposed by Federal legislation. It would be interesting to see the result, were they permitted to meet that competition with a free hand. CHAPTER II RAILWAY EFFICIENCY Increase of Efficiency with Increase of Traffic In the period of financial embarrassment and competitive strife that followed upon the commercial crisis of 1873, the maintenance of foadway and of equipment had been neglected on the greater part of the railway system of the United States. The service had so deteriorated as to be in- adequate for the requirements of the increasing traffic. In fact, it may be asserted that the railroads in the United States, as a system, were in a state of disintegration physically as well as corporately. Economic efficiency in their operation, as a whole, really began in the stage of reconstruction that thereafter followed. Bitter experience in the previous period had prepared the way for better methods, but only after the financial crisis had passed was it possible for the means to be obtained for putting them into effect.! The subsequent improvement in the general efficiency of our railway system is made apparent in the statistical reports of the Interstate Com- merce Commission. The railroad mileage, which amounted to 93,000 miles in 1880, had increased by 1890 to 166,000 miles ; 73,000 miles addi- tional or 78 per cent, in ten years. In the succeeding decades, there has been no such increase. By 1910, the mileage was 240,000 miles ; in 1911, 242,000; in 1912, 245,000; in 1913, 247,000; in 1914, 252,000. In 1881, Mr. Edward A. Atkinson estimated that our railway mileage would reach 209,225 miles by 1909, while the actual mileage at that date was 192,556 miles. In 1911, it was estimated that the new construction might average 5000 miles pfer annum,'' but the total increase from 1910 to 1914 averaged but 3000 miles per annum.' The mileage per 100 square miles increased from 7.97 miles in 1909 to 8.48 miles in 1914, while the mileage per 10,000 inhabitants decreased from 26.20 to 25.64 miles. The diminishing incre- ment in railway extension has been attributed to the approaching ade- quacy of mileage in the more populous regions, to the cessation of ' For a description in detail of the growth of the railroad system of the United States, see " Problems in Railway Regulation." 2 " Problems in Railway Regulation," p. 301. ' See Appendix I, Table II. 19 20 EFFICIENT RAILWAY OPERATION competition by rival corporations due to rate-fixing by the Interstate Commerce Commission, and to the increasing difficulty in obtaining capital for this purpose upon remunerative terms. Volume of Teaffic by Territokial Districts in the United States The railway system of the United States may be regarded as divided by differences in physical and social environment into three great territorial districts. One of these, which includes those States that are bounded by the Potomac and Ohio rivers on the south and by the Mississippi on the west, but exclusive of Wisconsin and the railway lines from Chicago west- ward, may be termed the Eastern District. The States south of the East- ern District and east of the Mississippi, may be termed the Southern District ; and the remaining portion of the United States, the Western Dis- trict. As so divided, the Eastern District contained, in 1914, about 61,000 miles of line ; the Southern District, 51,000 miles, and the Western Dis- trict, 140,000 miles.' The ratio of mileage of line to area of territory in each of these districts was approximately as follows : District Mileage Square Miles Per Mile OF Line Eastern Southern Western ...... 61,184 51,098 139,948 375,715 449,696 2,201,379 6.16 8.70 15.90 United States ... 252,230 3,026,790 12.11 The increase in mileage from 1911 to 1914, in these districts, was as follows : Eastern District 303 miles 0.5 per cent. Southern District 2,379 miles 4.9 per cent. Western District 6,369 miles 4.0 per cent. United States 8,051 miles 3.3 per cent. It may be added that the Eastern District of the United States is now better provided with railroad mileage than is Great Britain. The transportation service performed by this mileage in 1890 was measured by 76,207 millions of tons carried one mile, and by 11,847 ihil- lions of passengers carried, one mile ; and in 1913 by 301,398 millions of ton-miles and 34,375 millions of passenger-miles. This increase of 295 per cent, in freight-traffic and of 191 per cent, in passenger-traffic was ac- complished with ait increase of only about 50 per cent, in line mileage. The density of traffic in 1890 was at the rate of 487,000 ton-miles and of 75,000 passenger-miles per mile of line. In 1913, the density of traffic had in- creased to 1,245,000 ton-miles and 143,000 passenger-miles. Per mile of ' See Appendix I, Table III. RAILWAY EFFICIENCY 21 line, the efficiency of our railway system had increased in 23 years by 153 per cent, as to freight-traffic, and by 90 per cent, as to passenger-traffic. i The volume of traffic in each of these districts was as follows : Districts Frbight Traffic Millions of Ton-miles Passenger Traffic Millions of Passenger-miles Eastern ... . . Southern Western 154,173 48,543 98,682 16,397 4,489 13,689 United States . . . 301,398 34,575 From the above statement, it will be seen that one-half of the freight- service of our whole railway system, together with nearly one-half of the passenger-service, was performed by the 61,184 miles of line in the Eastern District, or by less than one-fourth of the entire mileage. The Southern District performed one-sixth of the freight-service and one-eighth of the passenger-service with 51,098 miles of hne, or about one-fifth of the total mileage, and the remaining one-third of the total freight-service, together with three-eighths of the passenger-service, was performed by the lines in the Western District of 139,948 miles, or a little more than half of the entire mileage. The relative service performed in these districts may perhaps be more readily appreciated by a comparison of the density of traffic per mile of line, as follows : DiSTBICTS Freight Traffic Ton-miles Passenger Traffic Passenger-miles 2,473,764 1,063,094 736,959 264,498 Southern . . Western 98,236 102,227 United States .... 1,245,158 143,067 Here it will be seen that the density of freight-traffic in the Western District was less than one-third of that in the Eastern District ; that the density of passenger-traffic in the Southern District was about three-eighths of that in the Eastern District ; and that in the Eastern District the den- sity of both freight and passenger traffic was double the average density 1 Traffic Statistics for 1914 Millions of Ton-miles Millions of Passenger-niiles 22 EFFICIENT RAILWAY OPERATION of the whole system. These simple facts indicate the wide differences between the several districts as to transportation conditions and also the necessity for difference of treatment in the regulation of railway affairs within their respective borders.^ In each of the territorial districts, certain railway systems predominate ; as in the New England States, where the New Haven system virtually controlled the traffic situation until its disintegration under judicial pro- cedure. In the remainder of the area included in the Eastern District, the trunk-Une traffic between the Atlantic Coast and the Mississippi River is mainly carried by the New York Central, the Erie, the Pennsylvania and the Baltimore and Ohio systems, as also the vast traffic with the Great Lakes. The anthracite-coal traffic is participated in by the Philadelphia and Reading, the Lehigh Valley, the Delaware, Lackawanna and Western, and the Delaware and Hudson systems. The currents of traffic in the Southern District are physically separated by the Alleghany range. The service east of that range is principally per- formed by the Southern, the Atlantic Coast Line and the Seaboard Air Line systems, whose freight-traffic has been developed from Norfolk as a base, and their passenger-traffic through Washington. A large part of the traffic that parallels the Mississippi River is conducted by the Louisville and Nashville, the Mobile and Ohio, and the lUinois Central systems. The Chesapeake and Ohio and the Norfolk and Western systems are prin- cipally engaged in the important traffic of the coal-regions tributary to their lines, while other hues are occupied with the business from the South Atlantic ports into the interior. In the vast territory of the Western District, there are important cur- rents of traffic identified with the grain-producing region of the Great Prairies and the mineral-producing region of the Rocky Mountains, as well as the transcontinental traffic. In all of these regions, the pioneer systems across the continent maintain their prestige. The Union Pacific and the Southern Pacific systems, long under a common control, are now separated. The Atchison, Topeka and Santa F6 system is a powerful rival for Cali- ' Traffic Statistics, 1914. Millions of Miles Freight TsApric Passbnqeh TBAFrlC DiSTBIOTS Feeight Passenger Increase Decrease Increase Eastern .... Southern . . . Western .... 144,428 50,131 93,760 16,649 4,698 13,911 1,588 9,745 4,922 252 209 222 United States . . 288,319 35,258 13,079 683 Traffic in the Southern District was apparently less affected than elsewhere by the unfavorable conditions that prevailed in 1914. RAILWAY EFFICIENCY 23 fornia business, and the Missouri Pacific has recently opened an independ- ent route to San Francisco. The traffic to the more northern Pacific ports was in the hands of the financial combination controlling the Great North- ern, the Northern Pacific and the Chicago, Burlington and Quincy lines, until the Chicago, Milwaukee and St. Paul system was extended to Seattle. These several systems compete for the traffic of the Great Prairies and the Rocky Mountain region with the Chicago and Northwestern, the Chicago and Alton, and the Chicago Great Western railroads. With the develop- ment of ocean-traffic at New Orleans and Galveston, stimulated by the opening of the Panama Canal, there is an increasing flow of business from the prairie-region to the Gulf of Mexico, in addition to the local products of the great state of Texas. In each of these territorial districts, there are other important lines which share, to some extent, in the through-traffic between them, but which are principally engaged in business of a more local character. Mileage and Tonnage The total railway mileage of the world in 1914 was approximately 714,000 miles, of which the mileage in this country was about 36 per cent. Of the total world mileage about 30 per cent, was state-owned, lying prin- cipally in colonial possessions and in Continental Europe.' It would be of interest to compare the service which our railway system performs with that performed elsewhere, but statistics adequate for that purpose are not available. It is, however, the consensus of opinion that nowhere else is freight-traffic so efficiently conducted. As was impi-essively stated by Mr. Seth Low in the arbitration proceedings in the Train Employees Case, " at the present time a ton of freight is moved in the Eastern territory more than three miles for the value of a postage stamp." When we think of the daily service thus performed in railway operation in the United States alone, and attempt to estimate, for the ^ntire world mileage, the number of years of labor of men and animals that would be expended in transporting the tonnage of commodities and the number of travelers that are carried for distances running up to thousands of mil- lions of miles in a single day, we feel as unable to arrive at an adequate conception of the vital energy that would be thus required as to measure the cosmic energy displayed in the motion of the stars in their courses. Still, some notion may be formed of the value of railway service to communi- ties and nations from estimating the amount saved in interest alone by speedy transportation. The tonnage of our railway systems in 1912 was about 1,145,000,000 tons, and the average haul was about 263 miles. This distance could not be covered by draft-animals, on an average, in less than eight days. Therefore, it would have taken eight years to accomplish by animal power 1 See Appendix I, Table I. 24 EFFICIENT RAILWAY OPERATION that which is accomphshed in one year by railway transportation. This is a saving of seven years in time, and in interest on the value of this amount of tonnage in transit. At an average value of ten dollars per ton, the total tonnage would represent an investment of $10,000,000,000, on which the annual interest at five per cent, would be $500,000,000. A savir^ of seven years' interest at this rate would amount to $3,500,000,000, which represents approximately the relative efficiency of railway trans- portation to that of carriage by draft-animals in the saving of interest alone on the value of goods in transit, an amount that is virtually an annual increment to the capital so invested. Electrification of Railways The characteristic feature of railway service is the production of power for traction from inorganic forces, and its application to the movement of vehicles by rail. Tractive power for this purpose had been obtained solely from the direct application of the expansive properties of water in the form of steam, until inventive genius devised the means for the appHcation of electricity to the same purpose. The commercial application of electricity to railway transportation began in 1887-88, with the introduction of the trolley system on the street- railways of Richmond, Va., by F. J. Sprague. The electric railway, oper- ated on the third-rail system, at the World's Fair in Chicago, in 1893, spread an appreciation of its efficiency throughout the world. In that year, the first franchise for the operation of a railroad solely by electricity was granted to the Northwestern Elevated Railway Company of Chicago. The development of multiple-unit control ^ by Sprague, in 1895-98, led to the general introduction of electric motor-cars on the elevated lines in Chicago and, after 1902, in New York City also. In 1894, the first inter- urban railroad specially constructed for electric traction was put in opera- tion between Cleveland and Akron, Ohio, a distance of 35 miles. The transmission of high-tension alternate current, inaugurated in 1895 by Tesla, in connection with the generation of electric current by water-power at Niagara Falls, resulted in its application to lines of considerable length. From that time, the electrification of steam-railroads became a commercial proposition. Its earliest application was in 1895, in the conversion of the Nantasket Branch of the New York, New Haven and Hartford Railroad into a trolley line. This branch was seven miles in length with sixteen miles of track, operated only in summer. The first interurban electrifica- tion of a steam-road was -undertaken also by the New York, New Haven & Hartford Railroad Company, in 1902, in the conversion of a branch-line between Providence, Warren and Fall River, with 38 miles of track. The experimental substitution of electric tractors for steam-locomotives was made in 1895, in the operation of the Baltimore Belt Line Tunnel, * The control of the motors in several cars from one place on the train. RAILWAY EFFICIENCY 25 covering 7.4 miles of track on the Baltimore & Ohio Railroad, because of the insufficient ventilation of the tunnel. For the same reason, the New York Central & Hudson River Railroad Company was required by legislation in 1902, to supersede steam on its terminal lines in New York City by some other mode of traction. As the company was preparing at that time to remodel the Grand Central Station, it was determined to carry out a scheme of electric traction in a suburban district extending for about 30 miles out, on both the Hudson River and the Harlem lines. As portions of these lines were used jointly by the New York, New Haven and Hartford Rail- road Company, that company Ukewise decided to extend the use of electric traction to Stamford. This was ,the first example of heavy trunk-line terminal operation for all-passenger-trains, and was initiated solely as a poHce measure at a time when no apparatus suitable for such work had been invented. Up to this time, the only electric roads, other than city or suburban lines, were magnified trolley -roads on the multiple-unit or motor- car system. But, two years after the work of electrification of the Grand Central terminal lines had been undertaken, the management of the New Haven road determined for the first time the efficiency of electric energy, with appropriate equipment, in the operation of the heavy passenger and freight service of a trunk-line steam-railroad. Meanwhile, the Pennsyl- vania Railroad management had entered upon the construction of a system of subterranean and submarine lines from New Jersey into Manhattan Island and to a connection with the Long Island Railroad. Between the Westinghouse and the General Electric companies, apparatus and equip- ment were perfected which were adequate to these vast undertakings, in- cluding the electrification of the Hudson River Line to Harmon, a distance of 34 miles, to White Plains on the Harlem line, 24 miles, and the New Haven Une to the city of New Haven, 72 miles from Grand Central Station. In January, 1907, a suburban service with motor-car trains was inau- gurated on the Harlem line to Woodlawn, 13 miles. By July, electric tractors were in use for through-passenger service on all lines in the ter- minal district of six miles, and on the New Haven trains to Woodlawn, where a change was made from the third-rail to the overhead-wire system which had been adopted on the New Haven line, and which was then ex- tended to New Rochelle, three and a half miles farther. In October, this service was extended to Stamford, 33^ miles from Grand Central Station, and this was then the longest mileage operated by electricity on any trunk- line railway in the world. In March, 1910, through electric service on the Harlem line was extended to White Plains, 12 miles beyond Woodlawn ; but local passenger-trains on the Hudson River line had been operated by electricity to Yonkers since August, 1908. It was only in 1910 that elec- tric service for all-passenger-trains was completed to the limit of the elec- trically operated district on the Hudson River line at Harmon. In March, 1913, the electrification of the New Haven lines had been completed to 26 EFFICIENT RAILWAY OPERATION the city of New Haven. By the summer of 1914, the New York, New Haven and Hartford Railroad Company was operating by electricity a six- track hne to Stamford and a four-track hne to New Haven, including in- termediate branch-Knes and freight-yards, for all classes of traffic. This system is the most extensive substitution of electricity for steam on any railway with heavy traffic. Including joint-trackage and controlled lines, it covers 112 miles of line with 633 miles of track. The electrification of the lines of the Pennsylvania Railroad Company in and around New York City was incidental to its pohcy of interchanging directly the traffic hitherto interrupted by the isolation of its city terminals by the intervening estuaries of the Hudson River and of the East River. Begun June 10, 1903, this great undertaking was opened for through- traffic on November 20, 1910. It then extended from Harrison, N. J., to Jamaica, Long Island, 16|- miles. From the tunnel-portal in New Jersey to that on Long Island, the distance is 5.3 miles. Included in that length of line are parallel river-tunnels of 6.8 miles and land-tunnels of about equal length. It connects with the Long Island system, which is stiU the most extensive conversion of a steam-road for handling suburban traffic by multiple-unit motor-car trains. It has electrified four-track, six-track and eight-track lines with a total trackage of about 250 miles. This system is now connected with the New Haven hues at Port Morris by a four- track road, 4^ miles in length, crossing the three passages from the East River into Long Island Sound. The completion of this project affords electric service from Newark to New Haven, a distance of 85 miles, being the longest continuous and total electric service of such magnitude on any electrified steam-lines. Taken as a whole, the electrification of the steam-lines converging on New York City, together with the other engineering work connected with it, constitutes the greatest enterprise in railway construction that has ever been undertaken. Initiated in 1903, it has been carried virtually to com- pletion in eleven years, at a total cost to the Pennsylvania Railroad Com- pany of about $160,000,000, of about $40,000,000 to the New York, New Haven and Hartford Railroad Company, and of probably not less than $100,000,000 to the New York Central & Hudson River Raikoad Company. These companies have also expended millions in purely experimental re- search, while the rapid and extensive development of the several branches of engineering service has made that immediate region, including the New York Subway system, the source of information upon which rail- way managements throughout the world have based their subsequent undertakings. In 1906, the West Jersey & Seashore line of the Pennsylvania Railroad system was electrified for 65 miles from Camden to Atlantic City; this being the first instance of the substitution of electric traction for main- line passenger-service.. Since Sept.. 11, 1915, the suburban lines of that RAILWAY EFFICIENCY 27 system have been operated electrically for twenty miles from Broad Street Station in Philadelphia westward to Paoli. In 1907, electric service was experimentally introduced upon a division of the West Shore Railroad be- tween Utica and Syracuse and also upon a section of the Rochester Division ■of the Erie Railroad, as tests of its efficiency in fostering local passenger- traffic on steam-railways. In 1910, an important electrification for local passenger-traffic was completed upon the suburban Unes of the Southern Pacific Railway Company centering upon San Francisco. In May, 1908, the Grand Trunk Railway tunnel under the St. Clair River at Sarnia was operated electrically, on account of its poor ventilation. For the same reason, in July, 1909, a section of four miles on the Great Northern Railway was electrified, to include the Cascade tunnel, 2.6 miles in length ; and on March 8, 1912, the Hoosac Tunnel on the Fitchburg Railroad, with an installation costing over a million dollars. The tunnel- line on the Michigan Central Railroad under the Detroit River is operated electrically, as is also the Connaught tunnel on the Canadian Pacific Rail- way. These are but a few instances of the many substitutions of electricity on roads with a heavy freight-traffic. In 1912, the Butte, Anaconda and Pacific Railway, a line of 26 miles in the Rocky Mountain region, was electrified, with branches serving copper mines with a trackage of 36 miles. The Chicago, Milwaukee and St. Paul Railway Company has electrified a portion of its main line to the Pacific Coast for 440 miles, covering the summit over the Rocky Mountains. In both of these instances, the moving cause for electrification has been the difficulty of economic operation by steam in a mountainous region, with poor and expensive coal and with water strongly impregnated with mineral salts, as compared with electric traction in proximity to cheap and ample supphes of electric current generated by water-power. For coal-traffic only, the Norfolk and Western Railway Company has recently electrified 95 miles of track in the Poca- hontas coal-fields, including a three-mile tunnel, which it was difficult to operate by steam.^ Originating in street-railway lines, many electric systems have been developed for suburban and interurban passenger-service and inciden- tally for light freight-traffic. Interurban lines, principally controlled by the New York, New Haven and Hartford Railway Company, cover the greater part of Massachusetts, Rhode Island and Connecticut, with track- mileage of 1016 miles, and one may travel by trolley-lines from Boston to New York. There is a similar connection of interurban lines in the Middle States, with track-mileage of 3478 miles, including extensions to Washington, D. C. On April 5, 1909, an electric road was opened from Pullman, a Chicago suburb, to South Bend, Indiana, 76 miles. This road is the western sec- 1 See Chapter III, page 84. 28 EFFICIENT RAILWAY OPERATION tion of connecting trolley-lines, extending for 360 miles to Cleveland. Since that date, there has been a remarkable development of interurban service, assuming more of the character of trunk-line passenger-service, in the region between the Ohio River, the Great Lakes, and the Mississippi, with 6768 miles of trackage. The Ohio Electric Railway Company operates a line from Toledo virtually into Cincinnati, crossed at Dayton by another division from Zanesville to Indianapohs, where it connects with other lines in Indiana that cover that State. Including branches, this company has 548 miles of line. The Western Ohio Railway Company operates a line from Piqua to Toledo and a branch to Cleveland, also other branch Hues ; it has a total mileage of 289 miles. The IlUnois Traction System, extend- ing from East St. Louis to Peoria with a branch to Danville, where it con- nects with lines in Indiana, has a mileage of 297 miles and operates a sleep- ing-car line between St. Louis, Springfield and Peoria. On the Pacific Coast, there is an interurban track-mileage of 2685 miles, of which 1936 miles is in California. The Pacific Electric Railway System, for both passengers and freight, centers in Los Angeles with line- mileage of 609 miles and 1286 of track, of which 12 miles is four-track and 1057 is double-track, with revenue, in 1914, of $9,500,000. At San Francisco, besides the suburban lines of the Southern Pacific Railway Company, there is the extensive suburban system of the Northwestern Pacific Railway Company; and from Portland, Oregon, there radiates a track-mileage of 770 miles.' Electric traction for cross-country freight-haulage was inaugurated in 1907 on a new line of 164 miles, the Spokane and Inland Empire Rail- way. The Oakland, Antioch and Eastern Railway Company, between Oakland and Sacramento, 91 miles, uses electric tractors for freight-service and interchanges freight-cars with steam-roads. At Sacramento it con- nects with an interurban system in the Sacramento valley with nearly 250 miles of line. The line of this road passes through the Cascade Range by a tunnel through solid rock, 3458 feet in length. An example of the complete operation, electrically, of an interurban road for the interchange of passenger and freight traffic with steam-roads, is that of the Une from Albany to Hudson, 32 miles. There are no official statistics since 1912 of the actual mileage of electric lines in this country. In that year, the track-mileage was given as follows : Urban lines 23,000 miles Interurban lines 15,333 miles Total 38,333 miles Time-tables of electric roads are given in the Official Railway Guide for August, 1915, operating in the following regions : '■ Most of these statistics are based on the United States Commerce Report on Electric Railways, 1912. RAILWAY EFFICIENCY 29 West of the New England States, north of the Potomac and Ohio Rivers and East of the Mississippi 22 lines, 1865 miles West of the Mississippi 22 lines, 1989 miles South of the Potomac and Ohio 1 Une, 126 miles Total 45 Unes, 3980 miles^ The commercial use of electric-railway traction abroad was introduced from this country, and the substitution of electric tractors for steam- locomotives followed upon the extension of the Paris & Orleans Railway into its new terminal in Paris, on the Quai d'Orsay, in 1901. Complete substitution of electricity for steam was subsequently effected on the line from Paris to Versailles. After experimenting with electric traction for five years, it was only in 1906 that it was generally applied on the under- ground systems in London. The most extensive substitution of electricity for steam in Europe is on the suburban lines of the London, Brighton and South Coast Railway which, when completed, will cover 173 miles of track. The London and Southwestern Railway Company has also electrified its suburban lines. Some of the other companies have prepared for electric operation of their suburban traffic, and, within a few years, steam- locomotion will be entirely eliminated from the Metropolitan Zone. With this exception, no important electrification of steam-railways has been undertaken in Great Britain. On the Continent, the great suburban traffic of Paris is now conducted by electric traction on the motor-car plan, in connection with its urban underground-railway system. Since 1902, the same course has been tenta- tively pursued with passenger-traffic in and around Berlin ; also at Ham- burg, since 1907. The substitution of electricity for steam, in main-Une traffic, was undertaken experimentally in Germany in 1909, upon a sixteen- mile section of the Magdeburg-Leipsic line. Other experimental service is in progress in Silesia, but, with the exception of a few short local lines on the motor-car plan, there is as yet no extended application of electricity to general railway operation either in Germany or in Austro-Hungary. The proximity to water-power has led to the use of electric traction in the Alpine region of Italy and on some French lines in the Pyrenees ; and the main line to Milan is operated electrically from San Pier d' Arena, near Genoa, through the five-mile tunnel at Ronco, a distance of twelve miles. The lack of coal-mines and the abundance of water-power has in- duced the Swiss government to enter upon the electrification of the 1875 miles of the Federal railway system, to be completed in 1928. The line from Berne, through the Loetschberg and Simplon tunnels, is now operated electrically to Iselle, some 80 miles ; and the mountain section of the St. Gotthard line from Erstfeld through the great tunnel to BelUnzona, 66 miles, is to be ready for electric service by 1920.'' Outside of Europe and 1 For other statistics, see Appendix I, Table II. ' The electrically operated track-mileage in Europe in 1908 was 8811 miles. 30 EFFICIENT RAILWAY OPERATION North America, electric-railway traction is principally confined to city- service, except in Australia, where it is to be substituted for steam on a system of suburban lines owned by the city of Melbourne. In 1913, con- tracts were made with the General Electric Company for the electrifica- tion of 150 miles of line in this system, covering 286 miles of track. Although the substitution of electricity for steam in railway operation has passed from the stage of theoretical discussion to its practical applica- tion in certain limited fields, there still remain unsolved important prob- lems affecting its relative economy ; while the financial issues involved are far more serious. Much more extensive experience is yet required in the observation of results, and in the perfection of devices and methods, to determine satisfactorily the extent to which this fundamental change in the source of tractive power should be applied in railway operation. Apart from its electrical equipment, a railroad cannot be operated elec- trically as it was previously operated by steam. Tracks, yards, terminals, stations, signaling apparatus and telegraph lines must be changed. Special rolling-stock and additional real-estate must be acquired. There is also to be added the cost of power-stations, cables, substations, conductors and feeders, and alterations, estimated at $50,000 per mile even under such favorable conditions as prevail in Switzerland. So that, taking all these items into account, the existing capitalization of our railway system would be more than doubled, were it totally electrified. The cost of installation of 22^ miles of line west of Stamford has been as great as the original cost of construction of the New Haven line from New Haven to the Bronx, and it was estimated that $30,000,000 would be re- quired for coniplete electric service from Stamford to New Haven, 36 miles. The electrification of the Pennsylvania railroad line from Broad Street Station, Philadelphia, to Paoli, a distance of 20 miles, with 93 miles of track, cost $3,500,000, and that of an adjoining section of 12 miles to Chestnut Hill is to cost $1,000,000 more, or an average cost of $144,000 per mile for the whole installation. This is not a complete substitution for steam but merely a trolley-system in the commuting-zone ; nor does the expen- diture include the cost of power-plant, as the current is obtained from a commercial company. The cost of electrifying the thirty-eight steam rail- roads within eight miles of the City Hall of Chicago, and covering 3,476 miles of track, was estimated at $178,000,000, but the extension of electric traction beyond that zone, which would be required to meet operating con- ditions, would increase the total cost to $274,000,000. With the experience acquired on the New Haven Hues, the opinion has been expressed that, "under general conditions, it is improbable that the direct saving resulting from the substitution of electric traction will justify the additional investment for electrifying steam-roads."^ It is quite a 1 "Electrification, N. Y., N. H. & H. R. R.," E. H. McHenry, Vice President,, Bulletin International Railway Congress, November, 1907, p. 1154. RAILWAY EFFICIENCY 31 different matter from the construction of a new line. A great amount of already invested capital must be sacrificed. The transition-stage is ex- pensive and difl&cult. It affects hghting and heating, telegraph and tel- ephone service, signaling and track maintenance, for which both temporary and permanent provision must be made. The simultaneous maintenance of steam and electric service requires the expenses incident to both, without the full economy of either. To secure the fullest economy the new service must be extended over the whole length of the existing operating district for passenger and freight trains. The concurrent use of steam and of electric traction on the same track has been found to diminish the efficiency of each when used separately. Consideration must be given to the con- tinuing obsolescence of electric equipment due to advance in the art of elec- tric traction; and the "scrapping" of over 63,000 locomotives, represent- ing an investment of not less than $1,500,000,000, will not be seriously entertained under any consideration now conceivable. ^ The substitution of electric traction for steam has been influenced, so far, more by considerations of social efficiency than of economic efficiency. It has occurred where steam and smoke have become a nuisance in the entraiices to large cities and, where so introduced, it has been availed of for passenger traffic between such cities and their suburbs. Gradually the benefits of frequent communication, and of stops at short intervals, have been conferred by electric traction upon thickly populated regions between large cities and incidentally some light freight-traffic has been developed, but, as yet, there has been no electric-railroad construction especially intended for trunk-Une freight and passenger traffic. Indeed, the only important instances of such construction for heavy freight-traffic have been upon steam-roads operating where coal was either dear or inferior, or the water impregnated with mineral salts, and where cheap and abun- dant water-power was available for generating electric current. The extension of electric-railway operation will therefore necessarily be independent of the existing system of steam-railroads. It can no longer intrude upon the public highways, but must occupy a separate right-of- way at a higher cost per mile than that of the steam-railroads already built. While, to some extent, the earthwork may be of proportionately less magnitude, the track-superstructure is even more expensive. The multiple-unit motor-car is only apphcable to light traffic. Trunk-lines can only be operated with electric tractors that are far more costly than locomotives and, their weight being quite as great, there can be no saving in bridges, viaducts and other substructures. Though, under favoring conditions, electric railways have diminished the profits of steam-lines with which they were in competition, their own operating expenses are increasing with the renewals rendered necessary by lapse of time, and 1 For further disoussion of this question see "Problems in Railway Regulation,': pp. 286, 294. 32 EFFICIENT RAILWAY OPERATION their fixed charges are also increasing with additional bonded indebtedness. With the improvement of public highways and the construction of road motor-cars, the "jitney" has appeared as a rival to trolley-lines, and, should its present desultory competition be developed under stronger organization, it may become a serious factor in diminishing the net earn- ings of suburban electric hues. It is finally to be recognized that after more than ten years of experi- mental practice on an extensive scale by a number of important corpora- tions, electric traction has superseded steam on less than 700 miles out of the 254,000 miles in the railway system of the United States. Taking all these matters into consideration, there need be no apprehension of steam being replaced by electric traction to such an extent as to affect the existing capitalization of our steam-railroad system. Departments of Railway Service The plant and the personnel of a railroad organization are more im- mediately devoted to the production, distribution and application of a primary source of power for tractive purposes. On this foundation rests the efficient organization of railway service. The sources of tractive power and its distribution by self-propelled tractors are controlled in the Motive Power Department. The construction and maintenance of the vehicles used in transportation are cared for in the Rolling-Stock Department. It is the office of the Transportation Department to apply these tractors and vehicles to the movement of Traffic. The maintenance of the per- manent way and track has brought into the railway service, the Road- way Department, corresponding to a similar organization developed by the former turnpike-service. The substitution of steam for animal power, as a means of traction, was only made commercially successful by the adaptation of the flanged wheel to the iron rail. This led to the control of the tractors, the vehicles and the roadway under a common authority. The physical environment primarily influences the method of railway operation under every aspect, as is indicated by the similarity of organiza- tion which generally prevails. For the division of railway service into operating departments is due to the fundamental distinction between the instrumentalities by which that service is performed. The service in each department is directly affected by the physical environment, since that environment controls the horizontal and vertical alignment of the highway,, the design of the tractors and vehicles, the character of the traffic and the conditions of the train-service. All of these instrumentalities are co- ordinated through the General Management, which is so intimately in- volved in the entire field of operation that its relation to the several de- partments into which that field is divided, may be Ukened to that of the brain to the human organism. RAILWAY EFFICIENCY 33 The human or social environment may be separated into that which is poUtical and that which is commercial. The political environment affects the administrative organization ; for instance, as to how far that organization may be directly or indirectly controlled either by the State or by capitalists. The poUtical environment may also seriously influence operating efficiency by legislation concerning the scale of wages, the terms and conditions of emplojrment, the conduct of the postal service and other matters relating directly to railway operation. The commercial en- vironment directly affects the compensation for the service rendered in the transportation of persons and things, as also the terms and conditions of that service. It also indirectly affects the cost of operation in the prices bf materials and supphes, and in the cost of living. The more immediate relations of railway service to its commercial environment are maintained through that branch of its organization which, in the United States, is known as the Traffic Department. Its relations with its financial and political environment are maintained through its administrative organiza- tion. These several aspects of its social environment may be provisionally ignored in a discussion of the efficiency of railway operation, pure and simple, and attention confined to its physical environment and to those opera- tions which relate directly to transportation of persons and things by rail. Railway efficiency, or the intelligent adaptation of means to ends as affecting railway operation, may, therefore, be considered under several aspects. I. As Technical Efficiency, or the adequate application of the reaction of Energy upon Matter, as the means to transportation by rail. II. As Economic Efficiency, or the utilization of tractive power in over- coming train-resistance with the least possible misdirection of Energy or waste of Matter. III. As Social Efficiency, or the performance of a public service with an assurance of safety, dispatch and convenience to the communities to which the service is rendered. Technical efficiency relates more directly to the mechanical require- ments of railway service. Economic efficiency represents the relative cost of the service performed, while social efficiency represents the relative value of the service which is rendered to communities and to their in- dividual constituents, without special reference to economic efficiency. Economic efficiency in the performance of a service adds nothing to the gross revenue from rendering that service; nor does it wholly diminish the cost of performance, which is largely dependent upon the physical and social environment in which the service is performed. Indirectly, however, by increasing production with the same or diminished expenditure of energy, economic efficiency does diminish the cost of per- formance by extending the total cost over a larger volume of production, and thereby proportionately lowering the cost-unit with a corresponding 34 EFFICIENT RAILWAY OPERATION increase in net revenue from the same expenditure. Therefore, the benefit derived from saving, through increased economic efficiency, should be con- sidered as a percentage of- the net revenue and not of the gross revenue from the service performed. Economic efficiency directly depends upon technical efficiency as developed in the departments of a railway organization that are devoted to the design, construction and maintenance of the motive power, the rolHng-stock and the roadway, which are the material elements of trans- portation by rail. Efficiency in each of these departments will now be separately considered. CHAPTER III MOTIVE POWER Locomotive Efficiency The operation of a railroad is primarily undertaken as the means for transporting persons and things from one place to another by land-car- riage ; mechanical energy being substituted for vital energy as the motive power. That which is sold is traction, and the basis of traction is the output of mechanical power created for that purpose. Thus we may re- gard railway service as a product in all respects similar to the output of steam-power, water-power or electric power, generated for any other pur- pose. In ordinary railway operation, steam as a motive force is applied through the locomotive as a tractor. It may be applied through many separate tractors though not all of similar design. Even the locomotives in use on any one railroad are diverse in this respect, and, if we are seeking for a basis of efficiency in the output of tractive force on a particular rail- road, we should endeavor first to ascertain what its motive power ought to do as a whole, and then what it really does in fact. In other words, there should be a comparison of Duty and Service ; Duty being that which should result, if the entire motive power were worked up to its theoretical capacity, and Service that which, in practice, results from the application of the power to the purposes for which it is intended. Duty, therefore, is that which ought to be done. Service that which is actually done. Duty is efficiency at one hundred per cent. Service is the percentage of efficiency virtually attained. Technically, efficiency is the ratio which the useful work performed in any operation bears to the energy expended in doing it. This ratio, as theoretically estimated, may be termed Duty ; and the ratio, as actually ascertained in any specific instance, may be termed Service. The efficiency of any particular locomotive, as a tractor, may be estimated by the per- centage of Service to Duty expressed in convenient units of measurement ; say in tractive pounds, as indicated by the dynamometer at the locomotive draw-bar, or the draw-bar pull. From such data, the total tractive power of the entire locomotive-equipment of a railway may be ascertained, and, consequently, the estimated and the actual potency of that railway as a means of transportation. 35 36 EFFICIENT RAILWAY OPERATION ' This proposition is illustrated by a statement recently made by the President of the Pennsylvania Railroad Company to the Interstate Com- merce Commission. In the period of ten years, ending June, 1913, the equipment of that company in locomotives had increased 45 per cent., and their tractive power had increased 80 per cent., or 99,521,170 tractive pounds. Here is an example of the increased potential efficiency of a rail- way as a means of transportation. It is also a practical recognition of the fact that the primary purpose in railway operation is the production of tractive power. The tractive power of the locomotive may be resolved into its several elements, and the merits or demerits of any locomotive, with reference to either of these elements or factors in the production of motive power, may be intelligently investigated and accurately ascertained. The field of research is wide, for the independent plants are numerous. The opportunity for fruitful results would be far greater if more attention were given to ascertaining the output of a railroad's tractive power, obtained from the locomotive-equipment, in general, instead of confining the investigation only to locomotives of recent design. If this course were pursued, many locomotives would be found so wasteful of energy that it would be more profitable to discard them than to continue them in service. The economic value of a locomotive depends upon its efficiency in three respects : first, as a steam-generator ; second, as a mechq,nism for the conversion of energy into mechanical power; and third, as a mobile tractor, through its adhesion to the rails. By the transformation of water from a liquid to a gas, energy is developed from he^t in the generator — the boiler. This energy is then converted into mechanical power in the steam-cylinder — the motor, frosvj which it is transmitted for traction through the adhesion of the driving wheels. The combination of these three functions in a self-contained mechanism — the locomotive — is the basis of railway transportation, which is the dominant factor in the mate- rial prosperity of every country into which it has been introduced. Of the three requisites here described, the efficiency of a locomotive as a steam-generator is ascertained by the relation of the number of heat-units required to evaporate a certain quantity of water in a given time. Its efficiency as a mechanism is virtually independent of its efficiency as a steam-generator, though that there is some relation between them may be gathered from discussions as to the relative merits of superheating steam and of its expansion in simple or in compound engines. The third element of efficiency in a locomotive, considered as a mobile tractor, is determined by the relation of its development of power to its adhesive weight ; that is, by its wheel-arrangement. The economic value of a locomotive depends not only upon its efficiency as to each of these requisites, but also as to their interrelation. That is, whether the steaming-capacity of its boiler be duly proportioned to the expansive capacity of its engines, and likewise MOTIVE POWER 37 whether this capacity be duly proportioned to its available adhesive weight. It is the proper relation of these fundamental elements in a locomotive which constitutes efficient design. Chief among these elements is the agency of combustion in the genera- tion of steam ; and here there are two factors of eflaciency, — the evapora- tive capacity of the boiler, and the calorific value of the fuel. Remark- able economy in fuel-consumption has been attained in stationary plants, but the same percentage of efficiency is not to be expected from the loco- motive, because of the different conditions under which is it operated. The opportunities for economy are far greater when the conversion of heat as motive force is accomplished at a central station, than when it is dis- tributed among a number of independent mobile tractors. For, at a central station, ample space may be provided for the installation of appliances for mechanical stoking, for improved draft, for feed-water heating and for making further available the wastage of heat from accessory apparatus. The essential features of the locomotive as a steam-generator are the multi-tubular boiler and the forced draft by exhaust-steam ; these were, indeed, the characteristic elements of the first successful locomotive, — George Stephenson's "Rocket." ^ From that day forward, ihere has been no marked advance, as to design, until the advent of recent devices for super- heating steam, which are still in somewhat of an experimental stage. The violent combustion in the fire-box, and the velocity with which the prod- ucts of that combustion are transmitted through the tubes, in consequence of the forced draft, result in their deUvery at the smoke-box still with a very high temperature, and with a corresponding waste of heat-energy, under the operation of Carnot's Principle.^ For these reasons, the loco- motive-boiler is necessarily a wasteful consumer of fuel. The evaporative capacity of the boiler is not the same at the fire-box as at the tubes. In the fire-box, the effect is produced by radiant heat, directly from the fuel ; in the tubes, by convection from the heated gases rushing through them to the smoke-box. The ratio of total heating-sur- face to grate-area is also an important matter. It should not exceed eighty times the grate-area, and more economical results are obtained at sixty-five times the grate-area. Even with ample grate-area and fire- box heating-surface, the grates and ash-pan should be carefully designed with reference to the quality of coal and the required rate of fuel-con- sumption. It is assumed that 2^ square feet of heating-area will produce one indicated horse-power when the grate is properly proportioned to the 1 See Appendix II, Table XIX. 2 Camot's Principle. Heat-ef&eiency is fixed solely by the temperature of the bodies between which, in the last resort, the transfer of heat is effected; that is, in a steam-engine, by the difference between the temperature of the ex- haust-steam and that of the atmosphere. The less the absolute temperature of the atmosphere and the less relative difference between it and the temperature of the exhaust-steam, the more ef&oient is the locomotive as a heat-engine. 38 EFFICIENT RAILWAY OPERATION heating-surface, and with regard alfeo to the quaUty of fuel. With a fair quahty of bituminous coal, 120 pounds per square foot of grate-area per hour is considered the maximum rate for economical evaporation, and with anthracite, 55 to 70 pounds per square foot, according to its quality.' With bituminous coal, the front end of the fire-box should be bricked off as a combustion-chamber; for a large percentage of bituminous coal burns above the grate as gas and the short time allowed for discharging and refilling the gases in the fire-box (about six times a second), is insuffi- cient for the perfect combustion of each particle ; therefore, the gases must be mixed, either by an arch or a baffle, which forces them through a re- stricted area not less than the flue-area. The baffle and the combustion- chamber not only aid combustion but also increase the radiating-surface, with corresponding increase in fire-box evaporation and in lowering the temperature of the escaping gases. Combustion-chambers lengthen the flame-travel, but the arch, especially on supporting water-tubes, doubles the average length of flame-travel and also possesses the more important advantage of a mechanical admixture of the gases, while the supporting- tubes expedite the circulation of the water; thus insuring a higher rate of heat-transfer. Thirty years ago, brick arches were rarely used, except •experimentally, and, at one time, virtually discarded. This device has since been so much improved that, in 1914, it was estimated that there were 30,000 locomotives so equipped in the United States and in Canada. The length of the barrel of the boiler, and therefore of the tubes, is determined more by the wheel-arrangement than by thermal conditions. Fifteen years ago, tubes were usually from 12 to 14 feet in length, while lengths from 20 to 24 feet are now in use. The ratio of tube heating-sur- face has accordingly increased from 8 to 12 per cent., and, in larger loco- motives, to 20 per cent. In the longer tubes, the heat is better utilized, because the range of temperature is greater between the furnace and the stack. The evaporative value of the tubes varies also with their diameter and with the manner in which they are spaced. The effect of fierce combustion is unsatisfactory, as much unburnt coal is drawn through the tubes and thrown out of the stack by the vio- lent draft. The chief advantage of the present draft-arrangement is in its simplicity, — free from complicated parts and requiring only minor ad- justments. As all other systems of forced draft have been found imprac- ticable, further improvement is probably to be sought in a less violent ' English locomotives use a good quality of coal and have grate-area of about 23 square feet. German locomotives of the same size have about 32 square feet or about 12i per cent, more, being adapted to a lower grade of fuel. The large French locomotives have from 35 to 40 square feet of grate-area ; or, in proportion to weight, about the same as in British practice. American locomotives have proportionately more grate-area than British or French; 60 to 70 square feet being usual for the Mikado type, and 88 square feet for the Santa F6 type. — Bulle- tin of International Railway Congress. MOTIVE POWER 39 draft, mechanically produced. Centrifugal fans as now designed cannot be constructed of sufficient capacity, within the necessary limitations of space for locomotive-use. Promising results have, however, been obtained from blowers of the turbine-type, but some other design than the ordinary exhaust-nozzle must be adopted, if it is intended to use powdered coal as locomotive-fuel. The necessity for the rapid production of steam from a relatively small volume of water is well provided for in the locomotive's multi-tubular boiler, but its steaming-efficiency has now about reached the limit per- mitted by the restrictions due to the gauge of the track and to its support- ing-power. Within the limits thus placed upon additional grate-area and increased heating-surface, the locomotive-boiler may be considered an efficient steam-generator, though not an economical one. Locomotive-boilers were originally constructed in a dome-shape over the fire-box. This form was superseded by the "wagon-top" over the fire-box, with the entrance to the steam-pipe in a separate dome, or rather a cylindrical chamber, attached to the barrel of the boiler. Drier steam could there be obtained than over the fire-box,\yhere the water was in more violent ebulUtion. The "wagon-top" also affords opportunity for a better method for attaching to the boiler-shell the stays by which the crown- sheet is braced. As boilers were designed of larger dimensions, the " wagon top" was extended farther forward. The "Belpaire" type of boiler, with a square top over the fire-box, has also been introduced from France, with the fire-box attached to the boiler-shell by radial stays. On roads using anthracite coal, a wide and shallow form of fire-box, known as the "Wooten" fire-box, is in use for burning anthracite culm. It is of interest to note some of the results obtained in tests to ascertain the ratio of the evaporative capacity and fuel-consumption of locomotives to the tractive power delivered at the draw-bar. In the Pennsylvania Railroad tests at the St. Louis Exposition, in 1904, there was obtained equiv- alent evaporation, from and at 212 degrees, of 16.4 pounds of water per square foot of heating-surface per hour; this being the evaporative capacity of the boiler when forced. Under lower pressure the evaporation per pound of coal was 10 to 12 pounds of water, which declined to two- thirds of these values when the boiler was forced. These figures show the existence of an economic balance between fuel-consumption and efficient evaporation. The ratio of evaporation in ordinary stationary engines is usually from 4 to 7 pounds per square foot of heating-surface per hour, but with proportionately lower fuel-consumption. This result is, however, obtained by the intervention of heat-saving appUances which are not available on a locomotive. In recent practical tests at Altoona, the best record with dry coal was 1.8 pounds of coal per indicated horse-power per hour, and the best per- formance with dry steam was 14.6 pounds of steam per indicated horse- 40 EFFICIENT RAILWAY OPERATION power per hour. The maximum equivalent evaporation, from and at 212 degrees, per square foot of heating-surface per hour, was 23.3 pounds. The St. Louis tests showed a minimum steam-consumption of 16.6 pounds of steam per i.h.p. per hour. The lowest figure of fuel-consumption was 2.01 pounds per i.h.p. A reduction of 10 per cent, in fuel and of 12 per cent, in water indicates the development of motive power efficiency in ten years in the best contemporary practice. In the Altoona tests, the best fuel-performance was obtained from a locomotive making 320 revolutions per minute and developing 1245.1 i. h. p. The best water-rate was by another locomotive at 320 revolutions per minute, developing 2033.1 i.h.p. On an average,^^ in simple engines with 700 to 1000 feet of piston-speed per minute, one horse-power can be obtained from 27 pounds of saturated steam, or from 23^ pounds in compound engines. With superheated steam, the same unit of power can be obtained from 20.8 pounds in simple engines, and from 19.7 pounds in compound engines, including steam for auxiliary purposes. Attempts to economize in fuel by feed-water heating have not been successful. The experimental saving of about ten per cent, has been ob- tained by appliances which complicate the operation of the locomotive in other respects. Still, an open feed-water heater under the boiler may yet be found practicable, using exhaust-steam from the air-pumps and boiler- feed, and partly from the main exhaust. Feed-water containing a considerable volume of solid matter in sus- pension, or salts of lime or of magnesia in solution, may seriously diminish the steaming-efficiency and endurance of boilers, by causing deposits on the boiler-sheets and tubes'. Under such conditions, the feed-water should be purified before it is used. The solid matter may be removed by filtra- tion. Where the water is very alkaline, say as much as 300 parts to the milHon, it has been found necessary to install expensive apparatus for its purification. This may be accomphshed either with soda-ash, quick-lime or gypsum as a reagent, according to the chemical nature of the impurities.^ Water containing organic acids is also objectionable. A road in Florida that obtained its supply from surface-water collected in ponds, was much troubled with leaky tubes. This was attributed to the action of organic acids de- rived from the roots of the dwarf -palmetto, which grows there in abundance. The acid acted upon the crystals of free carbon in the iron tubes, pitting them in minute holes, and brass tubes were therefore substituted. Any set of combustion data lacks completeness in so far as it does not give the calorific value of the fuel consumed, since that value varies from 8000 British thermal units per pound, in slack or culm, to 16,200 B. T. U. • The effect of a water-puriflcation plant was shown on a division of the Missouri Paciflo Raib-oad by a comparative statement of 166 locomotive-failiires from boiler-leaks in January, 1905, reduced to 10 in January, 1906. — "Economics of Railway Operation," Byers, 1908. MOTIVE POWER 41 in cannel coal. For, at last, all computations of the effect of heat appUed as a mode of motion must be measured in heat-units to be of service for purposes of comparison. With such a basis of the comparative values of fuel, the efficiency of different boilers may be measured as to their evapora- tive capacity in pounds of water in units of time, with an equivalent number of heat-units, and their relative economy as steam-generators may thus be accurately ascertained before proceeding to the valuation of the mechanism of the locomotive and of the conversion of steam as a motive force into tractive power.^ Efficiency in heat-units is equivalent to the number of foot-pounds of work per pound Of steam or, conversely, to the number of pounds of steam per horse-power per hour. The heat required to produce one pound of saturated steam from water at zero Centigrade, or 32° Fahrenheit, is about 650 times the amount of heat required to raise the temperature of water one degree Centigrade. The British thermal unit (B. T. U.) represents the energy of heat absorbed in raising the temperature of one pound of water one degree Fahrenheit, and is equivalent to the power required to raise 777 pounds' weight one foot high at latitude 45°, sea level. Similarly, the metrical calorie, or kilocalorie, represents the heat-energy absorbed in raising one kilogramme of water one degree Centigrade, and is equivalent to the power required to raise 426.3 kilogrammes one meter in height, or a kilogramme-calorie. There is a lesser calorie, or gramme-calorie, which represents heat energy of 426.3 grammes.^ Where a reliable supply of fuel-oil can be obtained at suitable prices, it is being substituted for coal, with great advantage as to furnace-deteri- oration and with the reduction to a minimum of the strain upon the phys- ical energy of the fireman. One and a quarter tons of oil are estimated to have the economic value of two tons of good coal, but care must be taken to separate the water that is sometimes associated with the oil. With many oil-burners, the water-supply and oil-supply are carried in a tank-car instead of the ordinary tender, and the cab for the engine-driver and the fireman is placed at the front end of the locomotive. In 1908, out of 56,867 locomotives in service, there were 2,354 oil-burners ; in 1914, out of 64,760 locomotives, 4,140 were oil-burners. In six years, there had been an increase of 1,786 oil-burners, or 76 per cent. The consumption of fuel- oil increased from 30,000,000 barrels, in 1914, to 37,000,000 barrels, in 1915, or 23 per cent. It is now in use upon forty railroads in this country, and the increased consumption is affecting the value of the official statis- tics of coal-constimption in its ratio to ton-mile performance. 1 See Appendix II, Table XX. 2 One B. T. U. = 0.252 kilocalorie. One kilocalorie = 3.968 B. T. U. One B- T. U. per cubic foot = 0.1123 kilocalorie per cubic meter. One kilogrammeter = 7.233 fool^pounds. 777 foot-pounds = 107.424 kilogrammeters. 42 EFFICIENT RAILWAY OPERATION On the Florida East Coast Railway, twenty-five locomotives have been recently converted into oil-burners. The fire-pans are round-bottomed and slope from back to front ; so that any accumulation of oil in the pan may drain out at the forward end, without danger from explosion. A course of brick is set on edge along the sides of the pan, to protect the lower portions of the side-sheets from the intense heat, and to seal -the pan at the point of its attachment to the mud-ring. Air is admitted through a damper at the front wall of the fire-pan, and through a second damper controlling the supply through the flash-hole, which is placed about two-thirds of the distance from the burner back to the rear of the pan. A four-inch length of tubing is inserted in each of the perforations in the front wall of the fire- pan, covered by the first damper. Air-supply is received at this point in firing-up. After steam has been raised, this damper may be partially closed and further air for combustion received through the second damper, which is manipulated by a notched lever in the floor of the cab. The average mileage per ton of coal was 18.11 miles, and, with coal at $3.04 per ton, the cost per mile run was $0.167128. During the same period, with oil at $0,017 per gallon, the mileage averaged .124 mile per gallon at a cost per mile run of $0.137467.1 More than one-fifth of the total coal-production of the United States is consumed by locomotives. In 1906, this consumption amounted to 90,000,000 tons, valued at $170,500,000, and was accounted for as follows : Utilized as motive force 41,000,000 tons Unutilized as motive force .... 31,000,000 tons Consumed in incidental service 18,000,000 tons The incidental consumption of fuel is in starting fires and in keeping up steam while the locomotive is standing, as well as the loss through blowing off at the safety-valve and from coal dropped in the ash-pit at the end of a run: The unutilized expenditure of heat-energy in train-service can be very little further prevented by economizing devices ; but the in- cidental consumption of fuel might be considerably reduced, if proper attention were given to it. t The use of pulverized coal as fuel in the manufacture of cement and in metallurgical processes, has induced experimental tests of its value as locomotive-fuel. A report by a committee of the Railway Fuel Associa- tion summarizes its advantages for this purpose, as follows : 1. Absence of smoke, sparks and cinders. 2. Maintenance of maximum boiler-pressure within an average varia- tion of three pounds. 3. Increase of 1\ to 15 per cent, in boiler-efficiency. 4. Saving of 15 to 30 per cent, in fuel of equivalent value. 5. Enlarged nozzle-area. 1 Journal Am. Soc. Mechanical Engineers, August, 1916, p. 666. MOTIVE POWER 43 6. Elimination of ash-pit delays and expenses, with reduction in time required for firing-up. 7. Maintenance of a relatively high degree of superheated steam. 8. No acciunulation of cinders, soot or ashes. 9. No overheating of fire-box. 10. Elimination of arduous labor in firing, and in building, cleaning and dumping fires. 11. Avoidance of expense in providing various sizes and kinds of fuel. 12. Elimination of front-end and ash-pan inspection, and the use of special tools and appliances for building fires and for stoking and cleaning them.^ If the nuisance caused by smoke and cinders can be prevented by the use of pulverized coal in connection with a hot-air blast, its use within city limits might be a preferable alternative to expensive electrification, Coal Handling and Mechanical Stoking After human efficiency has apparently been exhausted in the develop- ment of mechanical efl&ciency in the locomotive as a steam-generator and as a mechanism, its practical efficiency as a tractor depends upon human efficiency of another kind, — upon the vital energy and the skill of the fireman. The vital energy expended by him in this service is in striking contrast with the demand upon the muscular power of his companion at the throttle-lever ; for the engine-driver expends but Tittle energy in con- trolling the speed of the locomotive. ■ The service that he performs is like that of the marine pilot, except that the course of the train is directed by the flanged wheels and the rails. His attention is given mainly to maintaining the required speed, conforming to the train-schedule and observing the indications of the signaUng apparatus. In fact, the engine- driver is not essentially connected with the Motive Power Department, but rather with the train-service. On the contrary, the vital energy of the fireman is heavily drawn upon in the production of tractive power ; a task that requires skill, as well as muscular strength and physical endurance. Skill in handUng the scoop decreases the demand upon the fireman's vital energy, and training in the disposition of fuel in the fire-box decreases the ratio of fuel-consumption to water-evaporation. Instruction in both of these matters has been undertaken by some railroad managements, but, as a general thing, they haVe not received the attention that their impor- tance requires. The value of such instruction may be measured by a state- ment recently made by Mr. E. H. Coapman, Vice-President of the South- ern Railway in charge of operation, that in six years the consumption of coal on that line had been reduced 31 per cent. He accounted for this i"The Use of Pulverized Coal as a Fuel." Joseph Harrington. Journal Am. Soc. Mechanical Engineers, October, 1916. See also Appendix II, Table XXI. 44 EFFICIENT RAILWAY OPERATION remarkable result by the application of scientific and practical tests as to the quality of coal and by the use of superheaters, but he laid emphasis upon having practical men to teach the fireman, and by recording "the number of scoops of coal that each fireman throws into the fire-box for every hundred miles he runs." It is interesting to note the manner in which this was accomphshed. After a careful adjustment on each loco- motive of the draft-appliances, the valve-motion and the cylinder-packing, and of the admission of air through the grates, the fireman was trained in the reduction of useless consumption of fuel at terminals. The fire was either drawn or banked, according to the time that the engine was standing. When banked, the fire was moved forward on the grates, so that the admis- sion of air would check combustion without cooling the flue-sheet. The furnace-door was operated by compressed air, controlled by a treadle- lever. The opening was set at three seconds, which allowed time for firing a single scoop of coal, averaging 14^ pounds. Each movement of the door was recorded by a device attached to it, and the fuel-consumption was thus readily determined at the end of each trip.^ The demand upon the muscular power and the endurance of the fire- man has been much increased by the introduction of more powerful loco- motives. For every additional horse-power there has been a large increase in fuel-consumption, notwithstanding the improvement in mechanical efficiency. The physical efficiency of the fireman therefore virtually de- termines the practical efficiency of the locomotive. On locomotives of the larger classes, this task has approached the limit of human endurance ; and a great amount of ingenuity is now directed to lighten the fireman's labor by the intervention of mechanical stoking-apparatus. It is stated as a well-known fact that there is a difference of twenty- five to fifty per cent, in the amount of coal burned by different firemen in performing the same work, and that few locomotives of 50,000 pounds' tractive power can be worked to their full capacity by shovel-firing. In a random test of ten classes of heavy freight-locomotives, built in the last three years, it was found that, to deliver their full power, from 4900 to 8000 pounds of good coal were required per hour ; while they were actually getting from 4500 to 5000 pounds only, and hauling trains of corresponding tonnage. As a consequence, all the locomotives of the same class on the 1 On the Chicago, Rock Island & Pacific Railway, during the year ending June 30, 1915, in freight-service there was an average consumption of 16 scoops of coal per engine-mile ; in passenger-service, 7.4 scoops per engine-mile, and 9 scoops per switch-engine mile. A reduction of only one scoop of coal per freight-engine mile and one-half a scoop per passenger- and switch-engine mile would result in the following annual saving : Freight-service, 131,022 tons $294,799.50 Passenger-service, 67,496 tons 151,864,75 Switch-service, 24,045 tons 54,101.25 $500,765.50 Railway Review, April 1, 1916. MOTIVE POWER 45 same runs are now rated according to the ability of a fireman of average efficiency.! , With a locomotive consuming 4900 pounds of coal per hour, the fireman must handle 80 pounds, or five shovelfuls of 16 pounds each, per minute. At this point, his maximum capacity has been reached. To secure the full power of locomotives consuming fuel at this rate, mechanical effi- ciency must, therefore, be substituted for physical eflGiciency.^ There seems to be no question as to the practicability of mechanical stokers on heavy locomotives. They have been tested up to a capacity ■of eight tons per hour, or four times as much as can be expected from shovel-firing. With a superheater, the rating was increased to 5000 tons and, after it was fitted with a stoker, the rating was further increased to 6000 tons. A rivalry followed with the superheated locomotives that were shovel-fired which had the effect of increasing their rating to 5500 tons; so that a single mechanical stoker increased the commercial value, in ton- nage-service, of every locomotive on that division. On a division with ten tonnage-trains a day, mechanical stokers effected an increase of 11 per cent, in the tonnage, although the return-movement was largely in empty cars. The saving in wages and in train-supplies amounted to about $100 per month per locomotive. Wherever mechanical firing has been intro- duced, it has resulted in increased tonnage of from 10 to 20 per cent. ; it also assists in burning a cheaper grade of coal with a more uniform rate of consumption. There were in use, in 1915, 301 stokers of the under-feed type and 531 of the over-feed or "scatter" type.^ With increasing e;cperience, the stoking apparatus is becoming greatly simplified. It is now constructed of fewer, stronger and heavier parts, with very few moving surfaces. A well-designed stoker should handle coal from the tender in any condition, and with a minimum degree of at- tention. It should not require alteration after having been properly ad- justed, and should be controlled from the seat-box, where the fireman can keep a better lookout for signals ; and it should make no noise that could be heard while the locomotive is in motion. For the mechanical stoker to be a success, it should be developed as an integral element of the loco- motive, and not as an adjunct or afterthought. Steam Economy The economic use of the expansive properties of steam is the measure of the efficiency of the locomotive. WTiile economy in fuel-consumption depends upon the design of boilers and of their accessories, the efficiency 1 Experiments conducted in the Pennsylvania Railroad laboratory have shown the thermal efficiency of a locomotive-boiler to be 73.2 per cent, with experienced firemen, and but 59.7 per cent, with inexperienced men . Railway Club of Pittsburgh. Jan. 28, 1916. ' Proceedings Am. Master Mechanics Association, 1915, p. 35. 46 EFFICIENT RAILWAY OPERATION of the locomotive, as a mechanism, depends more directly upon the design of its engines. The locomotive, unlike most steam-engines, is operated under widely varying conditions as to power and speed. Its maximum power must often be developed in starting a train. The consumption of steam may then be diminished by utilizing its expansive properties, with- out impairing the efficiency of the locomotive as a tractor. Its full power, in this respect, is attained when operated at from 700 to 1000 feet of piston- speed per minute. The design of the locomotive, as a mechanism for the conversion of heat- energy into tractive power, is controlled by the necessity of conforming to these conditions. This result has been attained with high-pressure engines of a simple type, with rapid piston-motion and short stroke proportion- ately to cyHnder-diameter. The rapidity with which the reversal of this, motion is accomplished, renders it impracticable to regulate the admission and release of steam by the poppet-valves in use on stationary engines. This difficulty was overcome by the invention of the slide-valve controlled from a driving-axle. A practical difficulty appeared in the sudden reversal at high speed of the parts of the engine in reciprocal motion. The consequent shock to the pin-connections was obviated by a delicate adjustment of the valve, rel- atively to the angular position of the crank-pin, that permitted the ad- mission of steam into each end of the cylinder momentarily before the arrival of the piston there, resulting in a compression of the steam. A reaction then followed against the momentum of the moving parts by which the shock to the pin-connections was neutralized. This process, called "cushioning," reduced the liability of the piston's going through the' cylinder-head by the breaking of a crank-pin. The conversion of rotary motion from the driving-axle into reciprocal motion of the slide-valve by means of the eccentric, and the adjustment of that eccentric to coordinate the valve-motion in proper relation with the piston-motion, is a striking example of mechanical ingenuity. A further development of the eccentric-motion was devised in the link- motion, attributed to Robert Stephenson. This ingenious appKance for controlling the admission of steam into the cylinder at different points in the stroke of the piston, made the expansive action of steam available through a wider range than with the half-stroke cut-off, which was long: afterward in general use in the United States. The link-motion has. therefore been a valuable aid to steam-efficiency, though, for mechanical reasons, it has been largely superseded, on European roads, as well as in this country, by the outside valve-gear, known as the Walschaert type. The demand for greater tractive power in the locomotive could only be met by increased steaming capacity at higher pressure and by larger cylinders. Longer ports were then required to facilitate the admission and^ MOTIVE POWER 47 release of steam, and the necessarily wider slide-valves exposed a greater area to steam-pressure. The consequently increased friction between the faces of the valve and of the cylinder-ports brought such an excessive strain upon the valve-gear that, in the larger locomotives, this difficulty had to be obviated by balancing the slide-valve by means of internal steam- pressure, or by the substitution of the piston-valve. The efficiency of the valve-motion of a locomotive-engine may be illustrated by the example of a locomotive with driving-wheels six feet in diameter, running at a speed of sixty miles an hour. At this speed, the wheels revolve about 280 times a minute, while the motion of each piston is reversed over nine times a second. The valve-motion, which regulates the admission of steam to the cylinder with this frequency, is under a pressure of perhaps 200 pounds to the square inch. From this example, a conception may be formed of the efficiency of a device that combines such delicacy of adjustment under such a strain, and that, too, on a motor whose great momentimi subjects the whole mechanism to sudden shocks by the incessant reaction of the flanged wheels and the rails. In the early British locomotives, the steam-cylinders were inclosed in the smoke-box and connected with cranked driving-axles. This arrange- ment was retained on European roads long after American buUders had transferred the cylinders to the outside of the frames, to make way for the center-bearing truck which characterizes our railway practice. With this change of position, there has been a considerable loss of heat by radia- tion to the atmosphere, obviated to some extent by jacketing the cylinders ; but this loss has been more than compensated by doing away with the cumbrous and unreliable cranked-axle, and by readier access to the moving parts of the machinery. This change has been generally adopted, and the inside-connected simple engine is now a thing of the past. The Compound Locomotive and Superheating The successful application, in marine and stationary engines, of the principle of steam-expansion through a series of cylinders, led to its ex- perimental apphcation to the locomotive by the eminent French engineer, Anatole Mallet, in 1876. In its original form, the exhaust-steam from a high-pressure cylinder on one side was expanded into a low-pressure cyhn- der on the other side, of about double its capacity, through the intervention of a receiver in the smoke-box. In this design (known as the cross-com- pound engine), in order to start a train, high-pressure steam had to be introduced into the low-pressure cyhnder at a reduced pressure. This objectionable feature was obviated, in 1878, by Mr. Webb, of the London & Northwestern Railway, in the three-cyUnder compound, composed of two outside high-pressure cylinders expanding into one low-pressure cyhnder, placed in the smoke-box and connected with a cranked driving- 48 EFFICIENT RAILWAY OPERATION axle. In 1889, one of these locomotives was introduced on the Pennsyl- vania Railroad for experimental purposes. The Baldwin Locomotive Works built cross-compound engines in 1898, and subsequently originated the "Vauclain" four-cyhnder compound, with one high-pressure and one low-pressure cylinder, arranged vertically on each side. The mechanical difficulties involved in doubling the connec- tions with the driving-wheels of engines so placed, were ingeniously over- come and, in this design, compound engines met with more favor in the United States. It was further developed in the "balanced" compound, with the two high-pressure cylinders under the smoke-arch and the low- pressure cyUnders outside. The inside pistons were connected to crank- axles at 90° with the corresponding crank-pins on the outside engines, thus eliminating counterbalancing. Another variety of the four-cylinder compound was the "tandem" compound, built also by the Baldwin Loco- motive Works in 1902, and intended for heavy freight-service, keeping the low-pressure cylinders of larger diameter within the clearance-limits. In this design, the high-pressure cylinders were placed forward of the low- pressure cyhnders, with both pistons on the same rod. A steam-chest common to both served also as the intermediate receiver. The increase of tractive power at one step from 40,000 pounds in the simple engine up to 72,000 pounds, working compound, and even to 86,000 pounds in emergency, was at first severely criticized. Yet to-day there are compound engines working up to 115,000 pounds, and to 138,000 pounds in emergency. Engines have even been designed to give 140,000 pounds' tractive power, compound, and 168,000 pounds in emergency. The compound locomotive has undoubtedly a superiority over the simple locomotive in handling heavier trains with ^ relative reduction in fuel and water consumption, and, it is also claimed, with less boiler-repairs, an improvement in riding qualities and the practical elimination of jerks in starting. Yet the compound engine seems to be losing ground in Amer- ican practice, except as applied in the articulated or Mallet type. The statistical reports of the Interstate Commerce Commission show that, in 1910, there were in use in this country 862 two-cylinder and 1511 four- cylinder compound locomotives. In 1914, there were only 659 two-cylin- der and 1333 four-cylinder compound locomotives.' This decrease of 203 two-cylinder and of 178 four-cylinder compounds is exclusive of the articulated type, of which there were 775 in use in 1914, all with four-cylin- der engines.^ The diminution of interest in compound engines is attributed princi- pally to the growing belief in the superior efficiency of working dry steam 1 See Appendix II, Table VI. " In 1889, there were 689 compound locomotives in use in Europe ; in 1892, there were 1858. In 1900, the French companies had adopted the four-cylinder compound engines. ' MOTIVE POWER 49 in the simple engine over using saturated steam expansively in the com- pound engine. Superheating-apparatus was introduced on the German railways, in 1898, by Dr. Wilhelm Schmidt, and was introduced into the United States in 1906. In 1916, there were 16,000 superheated locomotives in use in the United States and Canada and, in 1917, over 21,000 were in service or under construction. The superheater is estimated to secure an economy of 25 per cent, in fuel, as a direct result of 33 per cent, reduction in the total water-evapora- tion, per unit of power. Its effect upon fuel-consumption has become recognized, though there are no official statistics as to the extent to which it has been applied. Both in freight and passenger service, there are locomotives developing at least one-third more power than would be pos- sible with simple engines using saturated steam and consuming an equal quantity of fuel. They are also operated at less boiler-pressure than with saturated steam. The economy in superheating seems to be confined principally to locomotives on high-speed trains with few stops, and there- fore operated in passenger-service. In the Schmidt apparatus, superheating is accomplished by circulating saturated steam through looped tubes introduced into certain boiler-tubes, enlarged in diameter, before the steam passes into the steam-chest. But the usual type of fire-tube superheater produces its maximum effect only when it is forced to the limit of boiler-capacity. The material in valves, cyhnders and packing, as well as the lubrication, will not withstand super- heating above a certain temperature. In other devices, the steam is super- heated in the smoke-box.' A serious drawback to the eflSciency of the locomotive is due to the back-pressure that follows upon the use of exhaust-steam in the forced draft. Tests upon eighteen different types of locomotives, working under various conditions, showed that for every hundred horse-power used in traction, sixty-six horse-power was wasted through the exhaust. Over ' The tendency is to increase the gas-area available for superheat at the ex- pense of the boiler-tubes. Steam is used at a temperature in excess of 750° F. The hmit is only fixed by the ability of the exposed machine-parts to~ withstand the high temperature. Although superheaters were originally appUed to heavy locomotives using steam for long periods of run, there is an increasing use of them for switchers. There are now over 1300 locomotives equipped with superheaters. — Geo. L. Bourne, Journal Am. Soc. Mechanical Engineers, September, 1917. The benefit of the superheater finds its limit when an increase of cut-off halts a further reduction of specific steam-consumption. The speed and pressure at which this takes place depend upon the proportions of the boiler as compared with the cylinders and wheels. The locomotive should be so designed as to pro- vide the boiler with its proper evaporating and superheating surfaces, so that the largest amount of sustained horse-power can be had. at the speed at which the locomotive is required to operate under normal conditions. The large smoke-tubes now used in this country are 5| and 5} inches outside diameter, with IJ inches outside diameter superheater-unit tubes. — R. M. Oster- mann, ibid. 50 EFFICIENT RAILWAY OPERATION 70 per cent, of this waste was due to the excessive back-pressure neces- sary to produce .the forced draft. Other tests were made in 1914, on a locomotive of the Prairie type, by changes in the front-end arrangement. Assuming the efficiency of the original arrangement at 100 per cent., a draft of six inches' water-pressure was produced at a saving of 34.5 per cent, in back-pressure. Yet this saving of wasted power only increased the total power of the locomotive by five per cent. It reached its maximum effi- ciency at a speed of 35 miles an hour, developing 1350 indicated horse- power with 190 horse-power of back-pressure. Until some kind of blower can be substituted for exhaust-steam, there seems to be no practical way of eliminating the back-pressure. Apart from the improvements already noted, the locomotive is still a simple reciprocating steam-engine, varying but little in its essential features wherever operated. But in the application of steam to railroad traction there is a wide variation in the coordination of motive power to tractive effort. That effort is exerted by the leverage of the driving-wheels upon the rails, and it is here that the necessary relation of weight to adhesion is mani- fested. The differences in practice are due to differences in environment, physically and socially. The ruling-gradients and the character of either passenger or freight traffic control the relative power and weight of the motors, and also the arrangement of the driving-wheels by which that power and weight are made available in the locomotive as a tractor. Wheel Arrangement and Design The early English locomotive, as used in colliery service, was little more than a stout wagon. The power from the cylinders, which were placed in the smoke-box, was more conveniently applied only to the back pair of wheels on a cranked axle. As the value of steam-traction for rapid motion became apparent and was applied to passenger-traffic, this pair of wheels was increased in diameter to as much as seven or eight feet ; the adhesion being still sufficient to overcome the inertia of the few and light carriages of which a passenger-train was at that time composed.^ The diameter of the driving-wheels was limited by the necessity for passing the axle under the barrel of the boiler, with clearance for crank-action, at a height, not so excessive as to raise the center of gravity above the point of safety in operation ; also, by the limit in the length of piston-stroke, which reduced the leverage of the crank-action as the diameter of the wheels was increased, and further by the mechanical difficulties in the construction of such large wheels. The development of passenger-traffic led to increase in the weight of passenger-trains, consequently to increase in size of boilers and engines and in the whole weight of the locomotive. But with in- ' On the Bristol & Exeter Railway, seven-foot gauge, in 1853, a tank-looomo- tive with single drivers nine feet in diameter made 81 miles an hour. MOTIVE POWER 51 creased steam-efficiency, it became possible to maintain a satisfactory rate of speed with smaller driving-wheels ; so that six feet, or rarely six feet and a half, is now their maximum diameter. For economic efficiency in freight-traffic, it is more essential that the entire weight of the locomotive should be made available for adhesion. For this purpose, the forward pair of carrying-wheels was coupled by side- jods to the rear wheels, which were directly connected with the piston- rods. But with, the demand for increased tractive power, the barrel of the boiler was lengthened in front. The forward pair of wheels was then moved farther! back and, in place of the rear wheels, were connected directly with the piston-rods. As speed was of less importance in freight-traffic, the driving-wheels remained of comparatively small diameter, in order to preserve the leverage of the crank-action. The greater dimensions of the engines increased the weight of the smoke-box end, and the accompanying length of overhang caused a pitching motion that was obviated by placing a third pair of wheels farther forward, not connected with the engines and therefore with a relative reduction of the weight available for adhesion. The differentiation of a type of locomotives with a single pair of driving- wheels for passenger-service and of another type with coupled driving- wheels for freight-service existed on European roads, long after the single pair of driving-wheels had disappeared from service on railroads in the United States. A far greater departure from European practice originated in the United States with the use of the center-bearing truck or, in British railway par- lance, the bogie. It served the useful purpose of reducing the rigid wheel- base from the length between the front and rear pairs of wheels, firmly attached to the engine-frame, to the distance between the centers of the coupled driving-wheels. This device was well suited to the frequent curves of short radius and the lightly built track which characterized the Ameri- can roads of that period, but was not so important upon the European roads, that were more substantially built and on an easier alignment. As this truck was so placed as to support the center of the smoke-box, it in- terfered with the position of the engines, which were then moved to the outside of the frame and connected to crank-pins on the driving-wheels. The inside-connected engines, with the objectionable cranked axle, have been superseded on American roads by the outside-connected locomotive, and the obvious advantage of this design has induced its general adoption elsewhere. The desire for a closer adaptation of type to environment, has led to other differences in wheel-arrangement; the most important being the introduction of a pair of low wheels back of the fire-box, to reheve the weight of the overhang at that end. All of these designs, however, are but variations of the original American outside-connected locomotive, which has gained general recognition as a successful example of the economic 52 EFFICIENT RAILWAY OPERATION use of the multi-tubular boiler with forced draft, in connection with the simple reciprocating engine, to develop power in a mobile tractor. There is a much greater variation in wheel-arrangement on locomotives of the ordinary types in the United States than either in the types of boilers or of the engines proper. The nomenclature, as to wheel-arrangement in use in this country, as given in Appendix II, Table XIII, has been generally accepted elsewhere. The locomotives in service in the United States, in the years 1911 to 1914, are grouped in Table XII as to wheel- arrangement in ten classes, of which five may be disregarded, as simply experimental designs. The locomotives in the remaining classes, in 1914, are tabulated as follows : A. All driving-wheels 8,496 B. One pair of carrying- wheels, front 25,268 C. Front truck 18,702 E. One pair of carrying-wheels, front and rear 4,911 F. Front truck, one pair rear 5,980 Total 63,357 Class A includes mainly switching- and tank-locomotives. Class E includes mainly heavy freight-locomotives. Class F includes mainly heavy passenger-locomotives. By far the larger number, about 70 per cent., is included in Classes B and C. Class B is the type in ordinary freight-service, and Class C in passenger- service. There were but 153 locomotives in the remaining five classes. Of the total number in Class B, 20,227 out of 25,268 were "eight-wheel connected," and 5013 were "six-wheel connected." In Class C, out of 18,702, there were 10,812 "six-wheel connected," and 7157 "four-wheel connected." So it may be inferred that, for general use in freight-service, the six- wheel and the eight-wheel connected arrangements are preferred, and for passenger-service the four-wheel and six-wheel connected arrange- ments ; all being provided with a front truck. The number of "Mikado" locomotives (2-8-2), used in heavy freight-service, increased from 671, in 1911, to 1159, in 1913, and to 3287, in 1914. As far as experience has gone with the Mallet type, 594 out of 775, in 1914, had one pair of front wheels and one pair of rear wheels. Of these, 464 had driving-wheels in two groups of six wheels each, and 118 had two groups of eight wheels each. The features of design for the transmission of steam-pressure in the cyhnders to the driving-wheels of a locomotive, bring into prominence the ingenious application of the mechanical powers in changing reciprocal into rotary motion, especially in the distribution among the driving-wheels of the adhesive weight of the locomotive. Here the relation of the poten- tial steam-pressure to the adhesive weight is an important consideration. For this is the final test of the combination of all the factors in the produc- MOTIVE POWER 53 tion of tractive power, as shown by the dynamometer and stated in pounds of draw-bar puU.^ Recent Improvements in Locomotive Design. The Abticulated Locomotive Thirty-five years ago, the boiler and engines of the ordinary American locomotive were attached to a bar-frame between the driving-wheels, which were spaced at a maximum distance of nine feet between centers. As the fire-box was between the frames, the grate-area was about three feet wide and six feet long and the maximum heating-surface was about 1300 square feet, which at the usual steam-pressure established a maximum of eighteen inches for the diameter of the cylinders. In 1881, Mr. T. N. Ely, of the Pennsylvania Railroad, designed a locomotive with the foundation-ring of the fire-box on top of the frames ; thereby adding eight inches to the width of the fire-box and permitting the grates to be as long as could be fired by hand, while the heating-surface was increased to 2500 square feet. In 1895 the Baldwin Locomotive Works introduced another design, known as the Atlantic type, in which the fire-box was extended farther back of the driving wheels and supported by a pair of tr ailing-wheels. This permitted the fire-box to be made as wide ^s clearance-limitations would allow. In 1901 this design was further developed for passenger-service in the "Pacific" type, with three pairs of driving-wheels.^ Twenty-five years ago, the largest locomotive in service weighed about 154,000 pounds, with 34,000 pounds tractive power, and this represented the improvement effected in sixty years. In this period, the purpose had been to attain increased tractive capacity. The details of construction had been improved and the number of wheels increased, but a pound of draw-bar pull required the same fuel-consumption as it had a quarter of century before. From 1889 to 1899, the total weight of a locomotive in- creased from 154,000 pounds to 232,000 pounds, with a proportionate in- crease of tractive power. Now, there are locomotives of the same two- cylinder type with 50,000 pounds greater tractive power than had been attained in 1899. The first systematic plan to secure the utmost power of locomotives, within given limitations as to weight and clearance, was made in 1895 with a passenger-locomotive of the usual American eight- wheel type. This locomotive weighed 116,000 pounds, of which 74,500 pounds was on the driving-wheels, and-with 21,290 pounds' tractive power. Up to 1902, there were few locomotives with higher tractive power than 40,000 pounds. 1 See Appendix II, Table XVIII, for formulas for determining the tractive power of locomotives. 2 A locomotive of this design was built by the Baldwm Locomotive Works in 1913, with cylinders 26 inches diameter and 26-inch stroke, driving-wheels 80 inches diameter, 4,525 square feet heating-surface, 38,300 pounds' tractive power, and total weight of 189,500 pounds. 54 EFFICIENT RAILWAY OPERATION The most recent design for passenger-service, knows as the "Mountain" type, has a four-wheel truck, eight driving-wheels and a pair of trailing- wheels. It has 240,000 pounds' weight on the driving-wheels, with 58,000 pounds' tractive power, or nearly three times as much power as the standard passenger-locomotive of twenty years before. In a type of freight-loco- motive built in this country for the Japanese railways, and therefore called the "Mikado " type, the adhesive weight has been proportionately increased by substituting a pair of leading-wheels for the truck, carrying the rear end of the fire-box on a pair of trailihg-wheels, and concentrating the greater part of the weight upon a group of eight driving-wheels. In the "Santa Fe" type, the weight is distributed among ten driving-wheels. With a given weight per pair of driving-wheels, a locomotive of this type can develop tractive power 25 per cent, greater than one of the "Mikado" type, and have equally high steaming capacity in proportion to adhesion.^ For freight-service on easy grades, exceptionally heavy locomotives of the American type are preferred, with six, eight, or even ten coupled driving- wheels. Simple cylinders, operating at 200 pounds' pressure, have reached a diameter of thirty inches, with main axles thirteen inches in diameter. The main crank-pins, rods and other moving parts are of proportionate size, and their weight has reached the point where proper counter-balanc- ing becomes difficult. Locomotives with four-cylinder simple engines have been tried, but the space between the frames has so limited their capacity that it is impracticable to provide them with the power given by the two-cylinder siniple engines. The effect upon the track of the vertical unbalanced forces in a two- cylinder simple engine has yet to be obviated. The more powerful loco- motives, with cyhnders from 27 to 29 inches in diameter, give maximum piston-thrusts of about 117,000 pounds, with wheel-loads higher than ever before and with reciprocating parts of much greater weight. The four- cylinder balanced compound was designed as a possible solution of this problem. A three-cylinder simple engine is said to have been successfully tested in experimental service, with a large cylinder between the frames connected to a cranked axle, and the outside cylinders to the back driving- ' Principal dimensions of locomotives built in 1915 by the Baldwin Locomotive Works for the Brie Railroad Company : "Mikado" type. Cylinders, 28 inches by 32 inches. ..■ Working pressure, 170 pounds. Driving-wheels, 63 inches diameter. Rigid wheel-base, 16 feet 6 inches. Weight on driving-wheels, 236,950 pounds. Front wheels, 30,200 pounds. Back wheels, 54,910 pounds. Total weight, 322,060 pounds. "Santa Pe" type. Cylinders, 31 inches by 32 inches. Working pressure, 200 pounds. Driving-wheels, 63 inches diameter. Rigid wheel-base, 22 -feet. Total, 41 feet 3 inches. Weight on driving-wheels, 327,250 pounds. Front wheels, 24,450 pounds. Back wheels, 56,000 pounds. Total weight, 407,700 pounds. Tractive power 83,000 pounds. MOTIVE POWER 55 axle. It offers a more even turning movement than is the case with the two-cylinder engine, with better counter-balancing and less destructive effect upon the track. The power obtained from a two-cyUnder engine, with cyhnders 27 inches in diameter and with maximum piston-thrust of 117,000 pounds, can be obtained from a three-cylinder engine with cylin- ders 22 inches in diameter and with a maximum piston-thrust of 78,000 pounds. This decrease of 33 per cent, in thrust means a corresponding reduction in the weight of the machinery, particularly of the reciprocating parts. It is thought that such a three-cylinder engine would be especially efficient in high-speed passenger-service.^ Both the simple and the compound engine were gradually increased in cylinder-capacity and tractive power until the further enlargement of the low-pressure cylinders in the four-cylinder compound engines had reached the clearance-limit ; the permissive diameter being 30 inches, equivalent to 21 inches in the high-pressure cylinder. In the eighties, the American Master Mechanics Association had fixed the maximum adhesive weight at 12,000 pounds on each driving-wheel, yet wheel-loads have now reached 35,000 pounds. A further important departure from conventional design was originated by Mallet in his compound articulated locomotive, in which it was sought to increase the tractive power within the clearance-restric- tions, without lengthening the rigid wheel-base or increasing the axle- weights. This purpose is accompKshed by a modification of the principle of the American long car-body supported by swiveling trucks.^ In the articulated locomotive, the rear end of the boiler is rigidly at- tached to a truck-frame that carries the high-pressure engines ; the low- pressure engines being placed upon a forward truck-ftame, well ahead of the front end. This end of the boiler rests upon a saddle, center-bearing on the truck-frame, which permits of a restricted lateral motion. In each truck-frame is a group of three or four driving-axles. The forward truck swivels on an extension that is pivoted to the rear frame ; the play in the articulation being taken up by strong lateral springs. The exhaust-steam from the high-pressure engines is conveyed to the low-pressure engines by a pipe with an intervening ball-and-socket joint and shp-joint. By the same device, flexibility is given to the superheater connections and, by metalUc hose, to the injector and feed-water connections. The Mallet locomotive, with two groups of three axles each, was intro- duced into the United States in 1904^n the Baltimore & Ohio Railroad for "pusher" service in coal-traffic on heavy mountain-grades. The Erie Railroad followed with heavier locomotives, having two groups of four ' Statistics as to performance and principal dimensions of certain high-speed passenger-locomotives are given in Appendix II, Table XIV. ^ In this respect, the articulated locomotive is a development of the Fairlie locomotive of about 1870, in which two boilers were placed back to back on a single frame supported upon driving-wheel trucks, each carrying its own engines with steam-pipes connected by swiveUng-joints. 56 EFFICIENT RAILWAY OPERATION axles each ; and still more powerful examples of the same wheel-arrange- ment were bailt for the Delaware & Hudson Railroad Company, each of which did the work of two "Consolidation" locomotives.* The Mallet locomotive was then adapted to road-service. The pas- sage around sharp curves was facihtated by the addition of leading and trailing wheels in pony-trucks, radially attached to the engine truck-frame, and by devices for lubricating the flanges of these wheels. Locomotives of this type, with three axles in each group, were built in 1906 for the Great Northern Railway Company.^ In 1909 the Atchison, Topeka & Santa F4 Railway Company was operating locomotives in passenger-service with 4-4-6-2 wheel-arrangement and 73-inch driving-wheels, and in freight-service with 2-8-8-2 wheel- arrangement and 63-inch driving-wheels. The long forward overhang of these locomotives resulted in such lateral play as to compel consider- able enlargement of clearance-limits around sharp curves. To obviate this, resort was had experimentally to articulating the boiler also, at the front of the combustion chamber and just ahead of the articulation of the truck-frames. Two expedients of this character were devised. One was a double ball-and-socket joint in connection with a slip-joint in the outer shell. The other was an " accordion " arrangement of rings, ten inches broad, in V- shaped joints, riveted on the inside and bolted on the outside edges, which effected the desired flexibility. In both designs, the combustion-chamber was prolonged in a drimi, flanged as a sHp-joint, into the forward section of the boiler, which was rigidly attached to the forward truck-frame.' ' Baltimore & Ohio Railroad locomotives : CyUnders, 20 inelies and 32 inches by 32 inches stroke. Steam-pressure, 235 pounds. Weight, 334,000 pounds. Erie Railroad locomotives : Cyhnders, 25 inches and 39 inches by 28 inches stroke. Steam-pressure, 215 pounds. Weight, 409,000 pounds. Drivers, 4 feet 3 inches diameter. Rigid wheel-base, 14 feet 3 inches. Tractive power, compound, 94,800 pounds. Delaware & Hudson Railroad locomotives : M Cylinders, 26 inches and 41 inches diameter by 28 inches stroke. Steam-pres- sure, 228 pounds. Weight, 402,000 pounds. Drivers, 4 feet 4 inches diameter. Rigid wheel-base, 14 feet 9 inches. Tractive power, compound, 1^5,000 pounds. Tractive power, simple, 126,000 pounds. Total length, with tender, 90 feet 6 inches. Height, 16 feet. Width, 11 feet 4 inches. ^ Great Northern Railway locomotives : Cylinders, 21.5 inches and 33 inches by 32 inches stroke. Steam-pressure, 200 pounds. Weight on drivers, 316,000 pounds ; on pony-trucks, about 20,000 pounds each. Drivers, 4 feet 7 inches. Rigid wheel-base, 10 feet. Total wheel-base, 30 feet. Length over all, 44 feet 10 inches ; with tender, 73 feet. Tractive power, working compound, 71,600 pounds. Reversing gear operated by compressed air. ' These locomotives were of the 2-6-6-2 wheel-arrangement, with super- heating and reheaters. Cylinders, 24 inches and 38 inches by 28 inches stroke. Steam-pressure, 220 pounds. Drivers, 69 inches in diameter. Tractive power, com- pound, 61,500 pounds. MOTIVE POWER 57 In working locomotives "compound," it has been customary to carry a steam-pressure of from 200 to 230 pounds, in order that the diameter of the low-pressure cylinders should be kept within the clearance-limits. An experimental Mallet locomotive, built by the Pennsylvania Railroad Company, was intended to work "high-pressure" only, at a pressure of but 160 pounds, by enlarging the diameter of the boiler and using a superheater. This locomotive had an adhesive weight of 437,500 pounds and total weight of 482,500 pounds, with theoretical tractive power of 99,200 pounds. The Pennsylvania system has made but little use of the articulated locomotive. The most powerful locomotive in service in the world, up to 1916, was built by the Baldwin Locomotive Works for the Erie Railroad Company. It is of the triple articulated type, with the tender as the third articulated section ; the wheel-arrangement being 2-8-8-8-2. It has six cylinders ; the middle pair working high-pressure. One cylinder of this pair exhausts into the forward low-pressure pair and the other into the rear pair, which is attached to the front tender-truck frame. The front cyhnders exhaust into the stack, and the exhaust from the rear pair, after passing through a feed-water heater, escapes through a pipe in the back end of the tank. The variation in the adhesion, due to the varying load in the tender, is not important in pusher-service. The locomotive has a superheater and a mechanical stoker, which handles coal from the tender to the furnace without the intervention of the fireman.^ This locomotive has exerted normally 160,000 pounds tractive power, increased to 200,000 pounds by mechanical draft. On a tonnage-test, it hauled a train of 250 Joaded coal-cars for 23 miles over ascending-grades, reaching at one point 0.9 per cent, on a five-degree curve. The total length of the train was 8547 feet or 1.6 miles, and its weight, exclusive of the locomotive, was 17,912 tons. The maximum speed was at the rate of 14 miles an hour, and the highest registered draw-bar pull was 130,000 pounds. With but one engine-crew, it has the capacity of three large eight-wheel coupled locomotives. Such a tractor may be considered a triumph of engineering skill in design and construction, but from a practical standpoint, it seems to have ex- ceeded the limits of economic efficiency. Yet the hmits of magnitude in the construction of articulated locomotives have not been reached, at least jn design ; for there has been designed a "quadruplex" locomotive, mounted on four groups of four driving-axles each, including the tender, with one ' Principal dimensions : Cylinders, 36 inches by 32 inches stroke. Working pressure, 210 pounds. Driving-wheels, 63 inches. Wheel-base, rigid, 16 feet 6 inches'; driving, 71 feet 6 inches; total, 90 feet. Weight (on driving-wheels), 761,600 pounds. Total weight, loaded, 853,050 pounds. The weight of the tender has been more recently utihzed for traction by plac- ing it upon "Mogul" running gear, actuated by superheated steam from the locomotive-boiler. ' ' Southern Railway Duplex Locomotive. ' ' — Railway Mechani- cal Engineer, March, 1917, p. 121. 58 EFFICIENT RAILWAY OPERATION pair of lead-wheels and one of trail-wheels, and a total length between extreme centers of 118 feet. Its estimated weight, in running order, is 885,000 pounds ; tractive power, 200,000 pounds ; working-pressure, 215 pounds. With smoke-box and combustion-chamber, the total length of boiler is 80 feet; with an "accordion" joint in the combustion-chamber, intermediate between two sections of tubular boiler. The engine-cab is in front. In four years, from 1910 to 1914, the number of locomotives in service of the Mallet or articulated type had increased from 200 to 775. It has apparently met with favor where local conditions as to gradients and tonnage had exhausted the economic capacity of the ordinary types of locomotive. On the Pennsylvania Division of the New York Central Lines, where the traffic is principally in coal, 60 Consolidated locomotives, costing $17,000 each, have been displaced by 26 Mallet compounds. They handle the whole tonnage at average speeds and with 35 per cent, less fuel, over long grades of one in 200, and others even steeper, and around eight- degree curves, without assistance from helpers. The reduction in the number of trains has served to postpone building second track that would cost millions. An articulated locomotive, with one pair of lead-wheels, one of traihng- wheels, and two groups of eight driving-wheels in each group, developed its maximum power at seventeen miles an hour. At twenty-five miles per hour it exerted 950 horse-power at the draw-bar. Locomotives of this type, designed for "pushing" service on grades of over two per cent., and rated at 115,000 pounds' tractive power, when wording compound, have handled a train of 7180 tons gross, on the Virginian Railway, on a six- tenths per cent, grade, requiring a draw-bar pull of 110,000 pounds. On lighter grades and at higher speeds, over 3000 indicated horse-power has been obtained. Other tests of this type have shown an increase of 13.5- per cent, in tonnage and increased speed of 4.3 per cent., with a saving of water of 11.1 per cent, on practically the same fuel-consumption per ton- mile. In 1912, experience with the Mallet locomotives in road-service on different lines was summed up in the American Engineer and Railroad Journal as follows : With 103 in service, two did the work of three of the Consolidation type ; with 25, the number in service was reduced one-half, the number of crews one-third and the fuel-consumption also. With 10 Mallet locomotives, the fuel-consumption per 100 ton-miles was 39.2 against 69.8 pounds with Consolidation locomotives, and, on another line, with the same number of Mallet locomotives -having 0-6-6-0 wheel-arrange- ment, three took the place of five of the Consolidation type, with saving of 22 per cent, in coal-consumption, one fireman furnishing steam for their full rated capacity. In general, the Mallet locomotives have been found equally reUable with other large locomotives ; the cost per ton-mile being MOTIVE POWER 59 reduced nearly 30 per cent, with increase in ton-mile movement of about 67 per cent. The cost of repairs is probably double that of the Consolida- tion type. When we think of the average locomotive in the United States as weigh- ing 77 tons and see it speeding along, self-contained as to the production of energy and its conversion into tractive power, we may well view it as a marvel of mechanical efficiency. Still more may we wonder at the per- formance of the monstrous articulated locomotive, 88 feet in length and weighing 225 tons, with 224 tons on the driving-wheels. Additional Features and Adjuncts Accessory appliances to a locomotive have greatly increased since the time when they included only the feed-pump, the bell and whistle, the spring-balance by which the single safety-valve was controlled, the gauge- cocks and the hand-brake. It would be difficult for the modern engine- driver to estimate the varying boiler-pressure by lifting the spring-balance with his thumb or to teU the height of water in the boiler at night by the difference in the hissing sound of water or of steam from the gauge-cocks. But the steam-gauge and the glass water-gauge, as well as the cab-lamp, were innovations not so many years ago, as were also double safety- valves. The steam-whistle, in the form of a trumpet, is said to have been in- vented by George Stephenson in 1833. Up to that time, signals had been given by lifting up the spring-balance attached to the safety-valve. In the United States, the first mention of the whistle is in connection with a locomotive built by Rogers, Ketchum & Grosvenor in 1837. Because of its whistle, this locomotive was bought for the first railroad in Ohio, which was then under construction. The track was laid to suit the gauge of the locomotive, which was 4 feet 10 inches. In consequence, this gauge was made compulsory by statute in the State of Ohio. On the early locomotives, the foot-plate, or foot-board, was merely inclosed by a hand-rail. Cabs were unknown until 1848, when an engine- driver on the Western Railroad of Massachusetts devised one of canvas as a protection from the weather.^ Before the introduction of the sand-box, it was the duty of the fireman and the wood-passer to prevent sHpping by throwing sand, from buckets, on the track, while walking beside the locomotive or standing on the pilot or "cow-catcher," which is a distinctively American device. ^ The reflector headlight is said to have been first introduced on the Boston & Worcester Railroad in 1840. It originated as a safety-device for » "The Eastern Railroad," F. B. C. Bradlee. - Wood-burners were in general use on roads in the United States for many years, and in the Southern States until about 1880. The use of coal was originally confined to the immediate vicinity of the mines. Coal-burners were introduced on the Eastern Raihoad of Massachusetts in 1860. 60 EFFICIENT RAILWAY OPERATION night-runs, when trains out of time ran against each other for the half- way post between sidings. The headlight lost its value for this purpose as train-dispatchers relieved engine-drivers and conductors of the re- sponsibility for making meeting-points for delayed trains. On double- track, the glare of the headUght becomes rather a nuisance than a benefit to the engine-drivers of opposing trains. This is particularly true of the use of electric searchUghts which, in gome of our States, has been made compulsory. In all other countries, simple signal-Hghts, similar to our tail-lights, are still sufficient for the needs of train-service. Far more important than the headlight, has been the substitution for the feed-pimip of that mechanical paradox, the injector. In the injector, patented in 1858 by H. T. Giffard, a French engineer, a jet of steam im- parts to a surrounding column of water, in a pipe, sufficient velocity to force the water into the same boiler from which the steam was taken. In addition to raising the temperature of the feed-water, it supplies the boiler while the locomotive is at rest. Many a collision was formerly caused by locomotives having to leave trains in sidings and then run out on to the main line to "pump up." A still greater improvement was the introduction of the air-brake; although this accessory pertains more directly to train-equipment, as well as the appliances for steam-heating passenger-trains and power for electric lighting. All these train-appliances, however, in so far as they cause steam-consimiption, are a draft upon the tractive power of the loco- motive. Among other modern accessories are the jet-blower, to create a draft when the locomotive is at rest; valves to reheve the back-pressure in the cylinders when the locomotive is drifting, with steam shut off ; spark- arresters that prevent a shower of sparks from illuminating the onward progress of the train and scattering fire along its path. Mention may also be made of sight-feed lubricators in the cab, of oil-pumps and of other auxiliary means of lubrication while the locomotive is in motion. It would be difficult to make long runs at high speed without stopping, but for the use of hard-grease cups wherever possible, instead of oil. All of these standard improvements in the mechanical efficiency of the locomotive are in general use in this country. The design of tenders has also kept pace with the development of the locomotive. As the load of fuel and water was increased, two four-wheel trucks of the type in use ,under freight-cars, though of larger dimensions and more accurately fitted, have been substituted for the original running- gear on three axles, working in pedestals fixed to the tender-frame. Ten- ders carrying 6000 to 8500 gallons of water and 10 to 15 tons of coal weigh empty 23 to 33 tons and, when loaded, from 60 to 82 tons. The type in general use is fitted with a horse-shoe tank surrounding the coal-space. Tenders attached to locomotives which are provided with grates above the MOTIVE POWER 61 trailing-wheels have water-bottoms, with the coal-space above them at the level of the higher furnace-doors. The "Vanderbilt" tender, with circular tank, holds 12,000 gallons of water and, with 20 tons of coal, weighs 206,400 pounds. For high-speed trains, running over 45 miles an hour, water is scooped up from troughs, or track-tanks, 1000 to 2000 feet in length, located on level tangents.^ This wonderful growth in motive-power efficiency, the creation of inventive genius in locomotive design, could not have been achieved, had there not been an equal advance in the metallurgical arts, in mechanical appliances and in machine-tool production. Improvements in methods of construction have accompanied improvements in design. In the earlier locomotives, the tubes were of copper or of brass and the fire-box of copper, as is still usual in European practice. But in this country, charcoal-iron and special grades of steel have been substituted with advantage, for the higher pressure to which our boilers are subjected. With the experience gained in the treatment of basic steel, that material is now in general use. Mechanical details are more thoroughly considered in the connection of boiler-plates by welt-straps, in a more judicious arrangement of riveting, and in the attachment of braces, crown-sheet bars and tube-fastenings, as also in the more careful use of calking tools and in the use of electric welding, pneumatic riveting and other appUances using compressed air. The introduction of water-tubes and superheaters has also required a higher grade of skill in the boiler-making art. This attention to me- chanical details has greatly contributed to the strengthening and endurance of boilers, enforced as it is by more thorough inspection; so that the explosion of a locomotive-boiler is now of rare occurrence, notwith- standing the increase of from 80 to 200 pounds to the square inch in the working steam-pressure. Similar excellence in character and workmanship has been displayed in machinery-details. The use of high-grade steel alloys has permitted a reduction of as much as 2,500 pounds in the weight of moving parts, with a corresponding relief in their effect upon track and tractor. On large locomo- tives the weight of the two cylinders has been decreased some 3500 pounds by the use of cast-steel. Cylinder-castings are made reversible, so that the same pattern may serve for either a right-hand or a left-hand engine. The wear on the lower surface in the cylinders by the weight of the piston is obviated by lengthening the front end of the piston-rod and extending it through a bearing in the front cylinder-head. Metallic piston-rings are kept tight simply by steam-pressure, which is also made available for relieving the wear of the valve-seats in the steam-chest by balancing the flat shde-valve; while the objectionable feature of long steam-ports in cylinders of great diameter has led to the substitution of piston-valves. Here, as in the matter of boilers, mechanical excellence and thorough in- » See Chapter V, Part II, p. 265. 62 EFFICIENT RAILWAY OPERATION spection are shown by the decreasing number of accidents attributed tO' locomotive-defects.' The strain upon the engine-driver's strength in reversing the valve- gear of heavy locomotives has been reheved by the aid of compressed air. Asbestos cloth has taken the place, as a boiler covering, of wooden lagging, which was liable to become charred and to take fire from a lodging spark. The labor of cleaning locomotives has been lessened by reducing the area of brightened surfaces and polished brass fittings with which they were once resplendent. Maintenance and Standardization Shop-efficiency is closely associated with motive-power efficiency as to mechanical excellence and operating economy. In this connection, it is well to dissociate construction from maintenance. The former, in its economic relations, has so little to do with railway operation that, in a discussion of shop-efficiency, it may be but briefly considered. In loco- motive-building on a large scale, as in other manufacturing industries, professional experts may be serviceable, and especially on piece-work, in diminishing useless physical effort and idle time of machine-tools; also in speeding up machinery. The greater the output of such a shof), the wider is the field for the profitable application of such methods. Con- versely, they are less valuable with a small output. It is doubtful whether it is advisable for a railroad company to build its own equipment, unless the annual replacement is on a scale commensurate with the provision of a suitable plant that can be fully employed in such work alone. The loss in obsolete patterns and in duplicate parts is pro- portionately heaAder in a railroad shop than when it can be distributed over a larger output ; while the temptation to find work to keep a full force on the payrolls, as construction becomes slack, does not exist in a commercial establishment. Nor should the interest upon the value of the land-site and the cost of the necessary buildings be ignored in determining the relative economy of buying or of building equipment, as well as the prob- able outlay for depreciation and betterment of buildings and appliances. The total number of locomotives built in the United States in 1914, includ- ing a number sent to Canada, was 2235. The total number in service that year increased by -1232 ; so that not over 1003 were required for replace- ment. From these statistics it would seem that there are few railroad com- panies that could profitably engage in locomotive construction.* Railroad repair-shops are in a class by themselves as to shop-efficiency. It is more of an object to have repairs done expeditiously than cheaply, ' The accidents from locomotive-defects, noted in tlie Report of the Inter- state Commerce Commission for 1915, were 856 in 1912, with 91 persons killed and 1055 injured ; and in 1915, 424 accidents with 13 persons killed and 467 injured. 2 Ry. Statistics of the U. S. for 1914. — Slason Thompson, p. 40. MOTIVE POWER 63 for the locomotive earns nothing while it is still. Much of the work is of a character known as roundhouse repairs, which is largely a matter of manual labor, appUed under conditions which necessarily involve loss of time in going to and from the shop, in taking down and putting up parts of machin- ery while locomotives are under steam, and in other ways familiar to oper- ating officials. The extent to which labor-saving appliances and tools can be economi- cally employed in an ordinary repair-shop depends upon the number of locomotives that need such a shop for maintenance in fair running order. As a general thing, about one-tenth of the locomotives on any road or division are out of service at one time ; and this is the measure of the kind of shop which should be provided for their maintenance. If an expensive machine-tool is to stand idle for half of the time, its efficiency is reduced to that extent, while the interest on its cost and the annual depreciation in its value are running on continuously. Unless it can be shown that the use of such a tool will make it possible to effect a proportionate reduction in the cost of labor, its purchase does not contribute to economic efficiency. Rather than provide each small repair-shop with such costly tools, it would be better to establish central shops for important repairs «.nd general over- hauling, where expensive appliances might be kept constantly employed, and leave the ordinary running-repairs to the division shop, intrusting the character of the work as well as the efficiency of the shop-hands to the supervision of intelligent and experienced foremen. The success which has attended the apphcation of standardization to ■ machinery-output in factories, has brought out enthusiastic advocates for its extension to operations in which handwork is a more important ele- ment. The profitable employment of such methods in railroad repair- shops varies with local conditions. As a general rule, standards are useful for all parts that are simple and numerous. For instance, standard screw- threads and other standards are adopted for bolts and nuts, and this requires that careful attention should be given to dressing drills and other tools. But the more varied the parts are in any mechanism, the more difficult it becomes to estabhsh satisfactory standards for them all. This is especially the case as to parts with large wearing surfaces. With these it may be cheaper to take up lost motion by individual fitting than to condemn both driving-axle and journal-box, for example, in order to preserve standard dimensions. In comphcated machines, it is far easier to estabhsh standards than to maintain them. The value of stand- ardization consists in rigorous conformity to dimensions. Variations that may be negligible if the units are few in number, become a serious hindrance to the economic value of interchangeable parts, where many of such units are in service. These remarks do not apply to the duphcation, ^s an emergency measure, of important parts roughly machined to ap- proximate dimensions. 64 EFFICIENT RAILWAY OPERATION The advantages to be derived from standards in details are not so apparent in general design. Standardization means rigorous conformity, and uniformity is a barrier to further improvement. Each progressive step then implies the scrapping of existing tools or apphances. This statement applies with force to designing machinery for purposes that are continually becoming more extended or diversified. The description given in this chapter of the development of the locomotive, furnishes a case in point. If the locomotive of the ordinary American type had be- come definitely estabUshed as the ne plus ultra of locomotive design, and classified types of this design had been made standard as to all parts and dimensions, the hmits of motive power efficiency would have been irrevocably fixed, and the hope of any further reduction in the cost of motive power would have vanished. It has, therefore, been fortunate that inventive genius has not been thus paralyzed, but that it has still the opportunity to increase the efiiciency of our locomotives, untram- meled by the restrictions that would have been imposed upon it by standardization in design. Locomotive Tractive Power in the United States A comparative statement of the total theoretical tractive power of the entire locomotive-equipment of our railway system appeared in Bulletin No. 31 of the Bureau of Railway Economics. This comparison was made for each year from 1902 to 1910, and the results for the years beginning and ending this period are as follows : Locomotives Tractive Powek Average Power 1902 1910 Number 41,225 58,9 i7 Pounds 768,502,779 1,588,894,480 Pounds per Tractor 18,641 26,955 Increase • .... Per cent 17,722 43 820,391,701 106 8,314 44 From this statement, it would appear that the theoretical efficiency of the total tractive power of the railway system of the United States had been more than doubled in eight years ; and that the num- ber of tractors and their average individual efficiency had increased nearly one-half. This statement may be compared with one compiled from the statistical reports of the Interstate Commerce Commission,"^ as follows : ' See Appendix II . MOTIVE POWER 65 Locomotives Tractive Poweb Average Power 1910 1914 Number 58,947 63,510 Poun(k 1,588,894,480 1,931,953,982 Pounds per Tractor 27,282 30,420 Increase 4,663 343,059,502 3,138 Average increase per annum 1902-1910 1910-1914 2,215 1,141 102,548,962 85,764,875 849 784 Apparently, the increase per annum from 1902 to 1910, in number of locomotives and in average tractive power, was not maintained in the next four years. From this comparison, however, there are excluded unclassi- fied locomotives ; also locomotives of the articulated type, as follows : Number Tractive Power Power per Tractor 1910 1914 : 200 775 14,407,261 61,241,128 72,036 79,021 Increase . . . 575 46,833,867 6,985 Using the conventional "standard locomotive" of 25,000 pounds trac- tive power as a unit of comparison, the average articulated locomotive of I9I4 exceeds three of the simple type in tractive power .^ It is further to be noted that, while there was an average annual in- crease of 752 locomotives from 1890 to 1900, of 2128 from 1900 to 1910, and of 2380 in 1911 the increase in 1912 was only 875, and 1116 in 1913. This indicates either that the locomotive-equipment had gained on the traffic-requirement or that there was a diminution in the purchasing-power of the railroad companies. In 1913, there was an increase of 1382, of which 828 were for freight-service and 216 for passenger-service, 247 for switch- ing and 91 unclassified. Eighty-two per cent, of the total number were in regular-train service, and of these nearly one-fourth were in passenger- service. The remainder were either classed as switching-locomotives or were unclassified ; the latter being for the most part in use on work-trains. Over eleven thousand, or 17 per cent., were thus employed in service un- productive of revenue.^ From the statistics given in Appendix II, Table V, it appears that, while the number of locomotives, exclusive of the articulated type, increased 12.4 per cent, from 1908 to 1914, their total tractive power increased 28.5 per cent. A higher ratio of increase was maintained with respect to the 1 See Appendix II, Table XVIII. 2 See Appendix II, Table I. 66 EFFICIENT RAILWAY OPERATION component factors of evaporative and adhesive capacity ; so that a due relation was preserved in their efficiency as steam-generators, as power- mechanisms and -as tractors. The general increase in motive-power efficiency is made apparent in a statement of train-tonnage on forty-five of the principal roads in this country in 1913, which showed that on sixteen of these the average freight-train load had increased thirty per cent, in the previous five years. As in the previous chapter,' a comparison was made of the relative volume of traSic in the three territorial districts into which our railway system has been divided in the statistical reports of the Interstate Com- merce Commission since 1911, it is of some interest to note the relative distribution of motive power in 1914.^ The ratio of niunber of loco- motives to length of line operated, indicates the relative density of traffic in those districts. The Eastern District, with 26 per cent, of the total mileage, had 45.6 per cent, of the total tractive power; the Southern District, with 19 per cent, of that>mileage, had 16.2 per cent, and the West- ern District, with 55 per cent, of the mileage, had 37.5 per cent, of the total tractive power. Of the total tractive power in the Eastern District, 74 per cent, was applied on sixteen roads ; in the Southern District, 73 per cent, on seven roads ; in the Western District, 74 per cent, on fifteen roads. None of these thirty-eight Hnes were operated with less than five hundred locomotives, and together they controlled 74 per cent, of the motive power of the entire system.^ The average tractive power per locomotive in the Eastern District considerably exceeds that in the others, including those of the articulated type. The extent to which this type has been recently introduced into our railway system is shown separately.^ While the use of the Mallet or artic- ulated locomotive is not general in the Eastern and the Southern districts, it is becoming a substantial element of the motive power in the mountainous regions of the Western District. Economy of Service Although the locomotive as a tractor has attained an efficiency of approximately 90 per cent., its economic efficiency is to be determined by the results of the application of its tractive power in train-service. There are two elements, mileage and tonnage, in this service, passengers being substituted for tons in passenger-service. For transportation purposes, these two elements are employed as factors for ascertaining the sum total of the service rendered ; that is, in passenger-miles for passenger-service and in ton-miles for freight-service. ' Chap. II, p. 20. 2 See Appendix II, Table VII. ^ See Appendix II, Table X. * Appendix II, Tables VI, VIII, IX, and XI. MOTIVE POWER 67 In 1913, there were 63,378 locomotives on our railway system. Of this number, 52,320 were in regular service ; 37,924 being on freight-trains and 14,396 on passenger-trains. The service rendered in that year was at the rate of 7,843,663 ton-miles per freight-locomotive and of 2,341,629 pas- senger-miles per passenger-locomotive. On the basis of mileage, the average -was 19,531 miles for freight-locomotives and 42,629 miles for passenger-locomotives. In 1914, with 38,752 freight-locomotives and 14,612 passenger-loco- motives, a total of 53,364, the respective averages, as to service rendered, were 7,368,713 ton-miles and 2,364,644 passenger-miles. With an in- crease of 828 freight-locomotives, there was a decrease in the average performance of 474,950 ton-miles. With an increase of 216 passenger- locomotives, there was an increase of 23,015 passenger-miles per loco- motive. From the Interstate Commerce Commission reports and from the tables in Appendix II, it will be seen that the average performance per locomotive in the past five years is reported as follows : YEAR FREIGHT PASSENGER MiLEB Ton-miles Miles Passenger-miles 1910 . . 1911 1912 . . . 1913 ... 1914 20,656 19,603 18,907 19,531 17,941 7,237,569 6,998,740 7,134,323 7,843,663 7,368,713 41,510 41,267 42,489 42,620 42,642 2,171,106 2,289,278 2,295,902 2,341,629 2,364,644 Average .... Per day .... 19,327 52.9 7,368,542 20,045 42,105 115.4 2,292,512 6,281 Average per mile run . 379 tons ; 55 passengers According to these statistics, the daily performance of a freight-loco- motive in these five years averaged 20,045 ton-miles, and that of a pas- senger-locomotive was 6281 passenger-miles ; equal to 379 tons of freight or to 55 passengers for each mile run. With forty-ton freight-cars and sixty-seated passenger-cars, this amounts to ten car-loads of freight and one car-load of passengers for each mile run. An additional factor, however, should be supphed in estabUshing a stand- ard of motive-power eflSiciency from these figures ; that is, the time occu- pied in rendering the service. What was the average nimiber of hours per day, throughout the year, in which each and every locomotive in regular service was out on the road between terminals? For, evidently, they were earning nothing except when they were on the road. Assuming an average speed of 13 miles an hour for freight-service and of 29 miles an hour for passenger-service, including all stops and delays between terminals, 68 EFFICIENT RAILWAY OPERATION the freight-locomotives were profitably employed, on an average through the year, for 4.07 hours a day and the passenger-locomotives for 3.97 hours. On this basis, the average daily work of a locomotive was respectively 4900 ton-miles or 1058 passenger-miles. The average rates were 0.733 cent per ton-mile and 1.982 cents per passenger-mile. At these rates, the freight-locomotives earned, on a daily average, $36.91, and the passenger- locomotives, $20.96. The 53,364 locomotives in regular service in 1914 represented an in- vestment of perhaps $1,500,000,000, and an annual interest at five per cent, of $75,000,000. If the motive power represented by this vast sum were employed in manufacturing industries, would its average use for only four hours a day be considered satisfactory? Yet, while a manufacturing plant is usually operated from eight to twelve hours a day, the train-service is conducted day and night. Why, then, should a locomotive be in use only four hours a day ? "Pooling System" in Locomotive Handling By keeping more continuously in use the enormous investment in motive power, the economic efficiency of our railway system would be largely increased. This end was sought in the introduction of the "pooling system"; that is, by assigning no particular crew to any particular loco- motive, so far as applicable under prevailing traffic-conditions. As has been the case with many other advances in railroad efficiency, this effort to increase the profitable use of motive power originated in the United States, and apparently with Colonel T. M. R. Talcott, when General Manager of the Richmond & Danville Railroad. Prior to 1875, the Richmond & Danville Railroad Company, whose line was of five-foot gauge, acquired control of the North Carolina Rail- road, a fine of standard gauge. The fine from Richmond connected at Greensboro, 189 miles from Richmond, with the North Carolina Railroad which, 93 miles farther, connected at Charlotte with the Atlanta & Char- lotte Air Line Railroad, a fine of five-foot gauge. This intervening section of 93 miles was, in March, 1875, changed to the same gauge to permit of continuous-train service from Richmond to Charlotte and Atlanta. The need for this change was so imperative that, until additional equipment of the wider gauge could be obtained, the run of 282 miles from Richmond to Charlotte was performed by the same locomotives which had previously covered the line from Richmond to Greensboro, the crews of the North Carohna Railroad taking charge of the trains at Greensboro. Therefore, thirty-six locomotives did the work before performed by fifty-seven, and with a monthly mileage increased from 1667 to 2683 miles. The result of this emergency measure induced the Pennsylvania Rail- road management to experiment in the same way, in 1876, with 27 engines and 42 sets of men. In the same year, on the IlUnois Central Railroad, the MOTIVE POWER 69 runs of certain freight-locomotives were lengthened from 117 miles to 210, 228 and 252 miles, and, on one division of 210 miles, a complete pooling- system was established, with 18 or 20 freight-locomotives. The experi- ment on the IlUnois Central Railroad was discontinued after a trial of from four to six months ; but on the Middle Division of the Pennsylvania Rail- road it proved so satisfactory to the Superintendent, Mr. James McCrea, afterward President of the company, that he subsequently introduced it on the New York Division. The train-crews were kept together, both locomotive-men and train-men, "first in, first out," and no cleaning was done by the firemen. These experiments in pooling locomotives attracted such attention that they were made a topic for discussion at the annual convention of the American Master Mechanics Association, held at St. Louis in May, 1877. The purpose then uppermost was to obtain increased mileage by long continuous runs of the same locomotive, with relay-crews at the terminals of connecting divisions. Further trials of this plan proved unsatisfactory. The infrequent opportunities for cleaning fires and for inspection led to a divided responsibility between the relay-crews as to fuel-consumption and as to careful handling, and it was this that brought the plan into dis- repute. Attention was then directed to gaining average mileage from the locomotives assigned to a single division, and it was this plan of pooling which was principally considered at the sessions of the International Rail- way Congress at Washington in 1905. The question, as there presented, was the advantages and disadvantages of the practice, with respect to the efficiency and care of the locomotive. The several ways of handling motive power were classified as follows : I. Assignment of a specified crew to a specified locomotive. II. Assignment of two crews in series to a specified locomotive. III. Assignment of three crews in series. IV. Assignment of relief crews to single-crew assignments. V. Assignment of more than three crews to all the locomotives as- signed to a specific service. VI. Complete pooling of all locomotives and crews on any fine or divi- sion, with temporary assignment of either to meet current requirements. A distinction was made as to the objects of poohng, whether as an emergency measure in an unexpected increase of traffic, or to provide for an ordinary seasonal increase, or as a general plan of operation. As might have been anticipated, opinions as to all of these matters differed with the varying conditions of environment in different regions or countries. These regions or countries were classified in three divisions, with a separate report from each division. The report upon the practice in the division including all countries, excepting the United States, Belgivim, England and its colonies, Holland, Denmark, Russia, Sweden and Norway, covered 26 managements, 62,810 70 EFFICIENT RAILWAY OPERATION miles of line and 21,900 locomotives.. The conclusions reached in this report were to the following effect : Complete pooling leads to an appreciable increase in cost per mile and should only be resorted to in case of absolute necessity. It is preferable to use "interpolated auxiliary crews" (that is, relief crews as in No. IV), or else multiple crews (No. V). The double-crew system (No. I) is recommended for switching, subur- ban and shuttle Service. The single crew (No. I) is advisable on fast express-trains. The complete pooling plan (No. VI) had been exclusively in use on the St. Gotthard Railway since 1886. In three years, the fuel-expenditure had increased 5.5 per cent., cost of lubricants, 42 per cent., and cost of maintenance, 11.6 per cent. ; but in the meantime the speed of trains and the weight of locomotives had also been increased. The report upon the practice in Belgium, England and colonies, Hol- land, Denmark, Russia, and Sweden and Norway covered 90 managements, of which 60 had replied, with 60,200 miles of line. Twenty-four man- agements, with 45 per cent, of the total mileage, preferred single crews (No. I). The other 36 managements made partial use of one or more of the several methods of pooling. Of these, there were 20 managements : which reported : 8,249 locomotives on single system (No. I) 3,341 locomotives on double system (No. II) 100 locomotives on triple system (No. Ill) 28 locomotives on mixed system (No. IV) 659 locomotives on multiple system (No. V) 149 locomotives on complete system (No. VI) 12,526 locomotives, of which number 34 per cent, were on one or more pooling plans. The only instance of triple crews (No. Ill) was in switching-service at busy stations on three lines in Great Britain. The report for the United States was prepared by G. W. Rhodes, Assistant General Superintendent, Burlington & Missouri River Railroad in Nebraska. He quoted the experience of Mr. M. E. Wells, a locomotive- engineer, who had been employed in a complete pooling-system on that line from August, 1893, to November, 1899 ; with eight-wheel, ten-wheel, mogul and consolidated locomotives, on fast and slow trains, work and local trains; all with crews ."first in, first out" on different trains to the best advantage of the service. There were 52 crews, 41 in freight-service and 11 in passenger-service; with an average of 37 locomotives, 30 on freight-trains and 7 on passenger-trains ; though, when necessary, the loco- motives in one pool were used in the other. There was a saving of 15 locomotives over the old system. All cleaning was done by roundhouse men, and the filling and cleaning of all lights. No headlight-oil or signal- oil was carried on the locomotive. The large tools, for emergencies, were MOTIVE POWER 71 carried in sealed boxes, the small tools in individual boxes which were taken off with the engineer. Lanterns were checked out of the oil-room and individual oil-cans, with engineer's number, were taken to the oil- room, refilled and returned. After the engineer had made his inspection and turned in his report, the roundhouse-inspection began with the hr)stler ; the boiler-maker inspected the boiler, grates, ash-pan and front-end, and the packer inspected the journal-boxes. In August, 1899, with engine- mileage of 299,205 miles, there were but two locomotive-delays from hot boxes ; one of twenty-three minutes and the other of thirty. The princi- pal trouble was in "a failure to get work done that is found and reported." This excellent summary of practical experience with the complete- pooling plan, enforces the conclusion expressed in one of the reports pre- sented at the Washington session of the International Railway Congress, that, "For the complete pooling-system to combine its maximum ad- vantages with its minimum disadvantages requires a special organiza- tion, which can not be developed for special cases or for temporary purposes." A lack of interest in this question was displayed when opinions were sought from the 217 members of the American Railway Association in the United States, Canada and Mexico; from whom only 84 replies were elicited, of which 48 were unfavorable to pooling. The most of these, however, were either from short lines or from lines that were operated in sparsely settled regions. One point alone seemed to be settled. The attempts on long runs with relay-crews had been abandoned. It was found that a relay-locomotive saved more time and gave better- service than did a relay-crew. The report also called attention to the saving in roundhouse accom- modation which was afforded by pooling. Roundhouses are considered adequate when they furnish accommodation for one-fourth of the loco- motives that are handled at a terminal. At Altoona, on the Pennsylvania Railroad, in 1905, 686 locomotives centered from two divisions. With 52 stalls, 203 locomotives were handled daily with double, triple or mul- tiple crews. Many of them had to stand out of doors in all weather and the need for more accommodation was emphasized as "appalhng." Sim- ilar conditions prevailed at other terminals. . Experience in North America was summed up as follows : Why is the question of pooling still an open one ? It is not contradicted that pooUng increases some items of expense. The main advantage which offsets such increase is the decrease in interest on the capital invested in locomotives. The volume of business and the available supply of motive power are the points to be considered. Under certain conditions of traffic, nearly all the lines pool their locomotives. Notwithstanding the at- tendant disadvantages, the increased mileage or tonnage shows a decrease 72 EFFICIENT RAILWAY OPERATION in cost per mile and per ton-mile. To keep constantly employed the roundhouse force required for complete pooling, there must be sufficient business to have locomotives constantly arriving that require this care. The additional mileage and tonnage, and the relief from extension of ter- minal . accommodaitions, make pooling profitable where such conditions prevail. After a discussion of all the reports presented at the Washington session of the International Railway Congress, the conclusions reached, and unanimously adopted, were expressed as follows : "In Europe and countries other than North America, the general sentiment is in favor of single crews and unfavorable to complete pooling ; except when necessitated by a sudden increase in traffic. For certain serv- ices, various combinations of double or of multiple crews are used ac- cording to circumstances. In North America, pooling is very general, though seldom used for passenger-service; and the tendency to single crews is manifest. The organization, however, depends largely upon local conditions." In the extended discussion that followed the submission of the reports on this question, it was evident that the plan of complete pooHng had not commended itself to many motive-power officials. As one of them said, "Pooling is unpopular because it requires more supervision and because the enginemen are opposed to changing engines." Perhaps it is for these reasons that the complete pooling-system is contemptuoueiy termed "the chain-gang." Yet a practical man, like Mr. Wells (p. 70), found it to result in a saving of 15 locomotives out of 52, or of about 30 per cent., and with no important disadvantages; the principal trouble being "a failure to get work done that is found and reported." Here seems to be the crux of the complete pooling-system, which was touched upon at the International Railway Congress; that is, it "re- quires a special organization which can not be developed for special cases or for temporary purposes," and further, that for such aif organization to be economically effective, "there must be sufficient business to have loco- motives continually arriving that require this care." The economic efficiency of the several ways of manning a locomotive should be determined by the practical test of ascertaining the relative number of hours on the road between terminals of all locomotives in regu- lar service. On this point there are no official statistics. Many opinions are offered with but few facts. These are from roads on which the pool- ing system, in some form, is in general use ; and they indicate an average of six hours a day for locomotives in freight-service. This is over 50 per cent', better than the general average for the whole country. From an analysis of the report of the Interstate Commerce Commission for 1914, the average hours on the road of locomotives in regular service were as follows : MOTIVE POWER 73 Freioht Passengeb Eastern District 3.84 4.64 3.30 3.83 3 63 Southern District .... .... Western District TTnited States 4.95 4.21 4 07 This statement applies to roads with annual operating revenues above $1,000,000, and excludes an average of 18 per cent, of locomotives not in regular service.!; It should be a comparatively easy matter to obtain more definite informa- tion of this character from the daily train-sheets and from a classification of all locomotives, as between freight-service and passenger-service. Such reports should show the exact time that each locomotive entered and left the roundhouse, and its time of terminal departures and arrivals. From these facts, the responsibility can be fixed, as between the Motive Power and the Transportation departments, for the disposition made of the idle time of locomotives. There is at present very little information avail- able on this point. If the motive power of our entire railway system is not profitably employed, upon an average, more than four hours out of the twenty-four, surely it is of importance for general managers, and even for presidents, to know how much of the twenty hours of idle time was spent in shops or roundhouse, cold and unmanned, and how much, in simply standing in freight-yards or at terminal stations, wasting fuel with an ex- pensive crew under pay. Electricity as a Source of Motive Power The progress made in the substitution of electricity for steam in rail- way traction gives importance to a discussion of the relative efficiency for this purpose of such a source of power. When the electric current was generated only chemically, as voltaic or galvanic electricity, its energy was insufficient for any purpose requiring the application of power to over- coming resistance of considerable magnitude, and its usefulness in this respect was restricted to the transmission of signals by the electric tele- graph. About 1863, electric current was generated mechanically in the dynamo invented by Pacenotti of Pisa, and the way was opened for its development in greater volume and intensity; whereby the appUcation of electricity to useful piirposes was immensely extended. The attention of electricians was first directed to illumination by the arc-light, which was introdufced into the streets of New York City in 1882. But illumination by electricity was only made generally avaUable by Edison's invention of the incandescent lamp. Another and widely ex- tended field of electric usefulness was opened up by the invention of the 1 See Appendix II, Table XV. 74 EFFICIENT RAILWAY OPERATION telephone. It is worthy of remembrance how the welfare of mankind in the world at large has been advanced by the researches in the field of electricity and by the inventive genius, in this country, of Morse, Edison, Bell and Tesla. The development of electricity as a source of motive power followed naturally upon its generation by mechanical means, and it soort became extensively used for this purpose in shops and factories as a substitute for steam-driven machinery. The rotary motion imparted directly to the motor from the generator invited its application to the axial rotation of vehicles, which attained commercial success in its substitution for animal power on street-railways in the trolley-system invented by F. J. Sprague.' In a discussion of the further substitution of electricity for steam in railway traction, the first point to be considered is that, save under ex- ceptional conditions, steam and electricity are alike conversions of the energy of heat into motive force. Whether one or the other form of con- version should be chosen, is a matter either of economic or of social effi- ciency. If electricity be chosen, the power is delivered to the tractors from a central source; if it be steam, the sources of power are in the tractors themselves. The economy in coal-consumption that can be attained at a central power-station may be illustrated by experience on the lines of the Inter- borough Rapid Transit Company of New York City. There, it was stated that for every hundred heat-units produced by its original plant, only 7^ units were applied to train-service. Originally, 176,000 tons of coal were consumed in transporting 116,000,000 passengers. In 1913, 332,000,000 passengers were carried on a consumption of 236,000 tons. With the original power-plant, this service would have required 400,000 tons ; and as the price of the coal is rated on the base of thermal units, the actual economic value, as well as the mechanical efficiency of the plant, is de- termined by statistical data of practical importance. Furthermore, the power-plant of 1901 delivered a kilowatt of power per hour with the use of 17 pounds of steam, but the plant installed in 1911 delivers the same unit of power with 11| pounds of steam, which is con- sidered to be the highest rate ever attained for commercial purposes. This improvement is expected to result in an increase of efficiency in train- service of about 42 per cent. Under previous conditions, 300,000,000 gallons of water were used per annum. The water consumption of the new plant is estimated at 200,000,000 gallons, with corresponding economy in coal-consumption. As already stated, there are exceptional conditions which give economic value to the use of electricity in railroad operation. This is obviously the case where it is practicable to substitute gravitation for combustion ' The further invasion of electric traction in the field of railway transporta- tion has been described in Chapter II, pp.- 24-32. MOTIVE POWER 75 in the generation of motive force; as by impounding streams at places where the momentum of a large body of falling water can be applied to the generation of electricity. This is still, however, the conversion of heat into motive force, for it is through evaporation by solar heat that the water has been delivered at the higher level. The topography of Switzerland has greatly favored the utilization in railway operation of "white coal," as the power so generated has been aptly named . ^ The most important examples of hydro-electric plants in this country are those which obtain their power from Niagara Falls. Water- power has also become of economic value on some of the roads that cross the Rocky Mountains ; although on these roads electric tractors are principally employed where 'good steam-coal and water free from mineral salts are difficult to obtain. But wherever there is an available supply of coal, the saving in the generation of electricity by water-power depends very much upon the necessary investment in engineering works, which varies greatly with topographical conditions. The practical economy resulting from the development of water-power for railway traction is not as great as might be expected theoretically ; because of the fluctuation in the amount of power required and because it is not ordinarily possible to locate the hydro-electric power-station at traffic-centers, or even on the . line of railway. The place whence water-power is obtained is fixed by the physical environment, and the electric current must often be transmitted for long distances to the points at which it is to be applied to traction. Great expense is involved in estabhshing and utilizing a hydro-electric plant, and then it may become unreliable because of drought, floods or , extreme cold, or from unforeseen catastrophes. Transmission of Electric Power In order to follow the development of electric traction, it is well to have some notion of the manner in which electro-magnetic force is imparted to the tractor. In the earlier dynamo, the rotor, revolving continuously in one direction, produced a tension or pressure in the flow of current through the conductor designated as a " continuous " or " direct" current. In the alternating dynamo, (for which we are largely indebted to Tesla) there is an alternating pressure or reversal of potential in the current, increasing with the rapidity of . the alternations, and designated as the "single-phase alternating current." An impulse may be imparted from another electro- motive source, so timed that its effects will alternate with the primary pulsations, like the crest of one wave coinciding with the trough of another, and the tension or pressure is thus correspondingly augmented. Similar impulses may be derived from other independent sources, and so are pro- 1 In 1908 the total available water-power in Switzerland was estimated at 750,000 horse-power, of which one-fourth had been utilized. 76 EFFICIENT EAILWAY OPERATION duced two-phase, three-phase or polyphase currents. Without going at present into further particulars, it may be stated that by certain ingenious arrangements in the apparatus, means have been devised for transforming these currents into continuous or direct currents and, in doing so, to greatly reduce the voltage.' '■The use of electricity as a substitute for steam in railway operation, is stiU in an experimental stage. The preferable forms of current and of conductors have not yet been definitely established. In the early de- velopment of electric traction, the continuous or direct current was trans- mitted at the voltage generated at the power-plant, and was applied directly to the motors. Experience, however, proved that continuous current could not be economically generated for long-distance distribution and for heavy traffic. To meet these requirements, this system has been modified by the generation of alternating current, transmitted at high voltage and trans- formed to lower voltage at substations and into continuous current.- Up to 1907, this was the system employed in heavy traction.^, The Erie Railroad was the first steam-road to use power from Niagara Falls. It was applied on an electrically operated line of 33^ miles branch- ing off from Rochester, 77 miles distant from the Falls. The power is generated as single-phase current at 66,000 volts, transformed to three- phase at 11,000 volts at substations and, at this voltage, supplied to feeders for use on a troUey-system. In 1907, the straight alternating-current system came into use on the New York, New Haven & Hartford Railroad. In this system, the current is transmitted as single-phase at high voltage and then transformed to lower voltage in alternating-current motors on the electric tractors. The New York, Westchester & Boston Railroad is operated on the New Haven system in its latest development, with 11,000 volts on multiple- unit motor-cars.' The polyphase alternating-current system, or three-phase system, was inaugurated in America in 1909 at the Cascade Tunnel on the Great Northern Railway. This is a three-phase alternating current fed into line-conductors at high voltage and utilized in three-phase motors. A modification of this system has been applied in the recent electrification of the "Elkhorn Grade" on the Norfolk & Western Railway. Here, single-phase current at 11,000 volts is delivered to the tractors and there ' The Volt is a measure of electric pressure analogous to the measure of steam- pressure by the number of pounds per square inch. * High-voltage current, exceeding 100,000 volts, is transmitted over 200 miles. For transmission up to 20 mUes, 11,000 volts is the usual standard. For longer distances, 22,000, 33,000, 66,000 and even higher voltage is applied. Such trans- mission was made possible by means of devices invented about 1890 by William Stanley, an American engineer. ' In the multiple-unit system, each car is suppUed with its own motors, which can be so connected that all the motors in a train may be operated ooUeotiveiy from a single control. MOTIVE POWER 77 transformed by a "phase-converter" into three-phase current at 750 volts f for the motors.' For long trunk-line work, and for the average density of traffic on American railways, the first cost of the alternating current, transformed into direct or continuous current, known as the "A. C.-D. C." system, would be prohibitory. The only possible alternative would be the " single- phase system" which, as yet, is also too expensive. Still, in the opinion of Mr. W. S. Murray, Chief Electrical Engineer of the New York, New Haven & Hartford Railroad, the single-phase system is the most economical for trunk-hne purposes. He states that the economic trans- mission of power is conserved by the agency of high voltage. It is con- sidered by practical men that 1500 volts is the limit of direct-current trolley-potential, or power of doing work, and there is great difficulty in collecting current from an overhead wire, at this voltage, for operating an 800-ton train. On the contrary, in a paper read before the Canadian Society of Civil Engineers in December, 1913, by A. H. Armstrong, a well-known elec- trical engineer, it was stated that the interurban railroad companies which had adopted the single-phase commutating motors were changing to con- tinuous current as fast as financial conditions would permit.^ The single- phase system had been introduced to obtain a higher voltage than the com- mercial 600 volts. Attention was then given to applying higher voltage to the continuous-current motors with such success that a voltage of 1200 and 1500 volts is being quite generally applied to such motors.' In the V electrification of steam-roads, there is no important installation of the single-phase system except on the New Haven lines. Under exceptional conditions, the three-phase system has been adopted on the Chicago, Mil- waukee & St. Paul Railway, and the modification of that system, known as the "split-phase system," on the suburban lines of the Pennsylvania Rail- road at Philadelphia, and on the Norfolk & Western Railway. In the early days of single-phase traction, trouble was experienced from disturbance of telegraph, telephone and signaHng apparatus, which was for the most part remedied by the use of the metalUc return, and the trouble has nearly ceased with improvements in the motors. Yet there is still danger that adjacent continuous-current conductors may, by electro- static induction, acquire a very high potential. This may be remedied by short-circuiting the two wires of each circuit at frequent intervals with a discharge-coil of very high inductance, which carries the static charge to earth. Placing the continuous-current conductors underground is an » See page 84. , « , 2 In the commutating motor, an altematmg current can be transformed mto a continuous current or vice-versa. ' On the Butte, Anaconda & Pacific Railway, a voltage of 2400 volts is so employed. 78 EFFICIENT RAILWAY OPERATION effective method of protection, but the cost of this is over $5000 per mile.' The difference of opinion as to the relative merits of the several forms of current for electric traction prevails also with reference to the means for conducting the current to the motors, whether overhead or by "third- rail." Several varieties of these two general forms are still under costly experiment on some of our princi{)al lines. ^The third-rail conductor for continuous current had its origin in the conduit that replaced the forest of poles and the maze of telegraph, tele- phone and trolley wires that disfigured the streets in our larger cities. This conduit, it may be mentioned, had itself been borrowed from the cable-railways that were first put in operation in San Francisco. As the early street-trackage was extended and the flow of continuous current diminished with the increasing distance from the power-plant, additional current was fed into the trolley-wire from a separate conductor, at inter- vals, through "boosters."^ The greater volume of direct current absorbed in traction for heavy rail- way traffic requires a conductor of far greater cross-section than a trolley- wire. Resort was therefore had to the conduit-rail, placed on the outside of the track-rail, as the "third-rail system." In this system, the contact- shoe completes the connection with the motor in place of the trolley in the overhead system. Here, again, the difficulty in maintaining the volume of direct current necessary for heavy traffic over long distances led to the introduction of the alternating current, distributed along the line in a sep- arate conductor and transformed at intermediate substations, where it is fed into the third-rail as the current is fed into the trolley-wire through boosters. The position and height of the third-rail must be rigidly maintained and the margin of permissible variation is small. Its continuity is broken at switches and crossings by frequent transference of the conductor to the opposite side of the track or overhead. The position and height of an over- head wire may vary within vertical and horizontal limits of eight and four feet respectively, without losing contact with the pantograph-bows, which have replaced the trolley-wheel on the metallic overhead circuit, q In the New York Subway, current is fed into the third-rail every mile, and the conductor has to be of large surface to carry the great volume of low-pressure current. The Subway runs more trains in an hour than all the through trains in a day between New York and Chicago, and has its 1 Induction is the influence exerted by an electric current, or by an electrically charged substance, upon a neighboring conductor or electrically sympathetic substance, without any direct connection between them. A disoharge-Qoil, or "choking coil," absorbs such induction and discharges it directly to earth. 2 A booster is an auxiliary dynamo placed in the main feeding-conductor, by which additional power is supplied to the trolley-wire, as it is drawn upon by the passing motor-cars. MOTIVE POWER 79 whole equipment virtually in continual service. On no railroad line are there more than a dozen through trains between New York and Chicago, and the electric generating capacity on such a line would be required only occasionally at its maximum. As the capacity in an electric conductor is in proportion to its surface, the third-rail provides the means for the sparkless collection of current at low voltage in large quantities, but it requires a heavy investment in copper conductors, in substations and in frequent power-plants.^ The same energy can be supplied through a small wire by a high-tension current. On the New York, New Haven & Hartford Railroad, the alternating cur- rent is fed into the conductors at only one point between Woodlawn and New Haven. From Woodlawn to Grand Central Station, the New York Central continuous current is used. The improvements effected on the Pennsylvania Railroad made it technically possible to transmit large quantities of low-tension continuous current through third-rail conductors. With 600 volts in the line, a speed of 59 miles an hour was attained on a level, with trains weighing 550 tons and the tractor developing 2000 horse-power. To prevent the voltage dropping with the large volume of current, a third-rail is required of in- creased section, which on the Pennsylvania Railroad weighs 150 pounds to the yard. The first success with high-voltage motors was obtained with the three- phase current, but better results have been secured with the single-phase alternating-current motor. Yet objections have become so apparent that, on the Washington, Baltimore & Annapolis line, return has been made to continuous current with voltage increased to 1200 and 2000 volts. It was also found that a motor-car equipped for the alternating current weighed 60 tons against 40 tons for the continuous current. JThe relative advantages and disadvantages of the continuous and of the alternating current may be summed up as follows : The continuous current has the advantages of good regulation of motors, simple construc- tion, cheaper car-equipment, elimination of transformers and no disturb- ance of telegraph, telephone or signaling apparatus. It has the disad- vantages due to limitation of voltage and to restriction of the distance in which it is effective from a single power-plant, even with frequent sub- stations. The alternating current has the advantage of a greater radius of action at high-tension from a single power-plant, which may be trans- formed at substations into continuous current or used directly for traction by intermediate reduction of voltage on the tractors. But the car-equip- ment is much heavier, with greater initial outlay for overhead conductors and greater working-expense. , It also necessitates special protection for telegraph and signahng apparatus. ^ The continuous-current system, up 1 Substations can not be more than eight miles apart, as this is the maximum distance for economy in copper feeding-lines. 80 EFFICIENT RAILWAY OPERATION to 750 volts, has long been standardized and is generally adopted for urban and for suburban lines. The three-phase system has been fairly standard- ized, but the single-phase system is still in process of development. Electric traction with continuous current was originally restricted to multiple-unit motors. The maximum size of motors was practically 200 horse-power for each truck of a motor-car at a one-hour rating which, with direct current of 600 volts and forced ventilation gave a continuous capacity of 140 horse-power. ^ This is the usual practice with single-reduction geared motors. With the multiple-unit system, there is virtually no limit to the motive power which can be applied to a train, other than the number of axles in it. A New York Subway train of eight cars is usually made up with five motor- cars each having two 200 horse-power motors with which a speed of 100 miles an hour can be attained. With 2000 horse-power, the train is under full headway in three times its length, and the prescribed speed is then maintained with 300 horse-power. With continuous current, motors may be coupled in series, working up to 750 volts, or from 900 volts upward with reversing-pole motors in parallel connection.^ Efforts to advance the efficiency of electric traction for railway purposes have been directed mainly to operating at higher voltage, either by in- creasing the heating-capacity of continuous-current motors or by designing motors to be actuated by alternating current, in order to avoid trans- forming the current at substations. On the suburban lines in Melbourne, Austraha, it has been decided to use three-phase current at 20,000 volts, transformed at substations into 1500-volt continuous current, using the running-rails for return. It is proposed to electrify 87 miles of line, with 288 miles of track, from a centrally situated plant. Electric Tractor Design The electric motor can exert its full power at starting. But as this effort is continued, it is accompanied by an increase of internal heat in the motor which tends to destroy the insulation, if it exceeds a critical tempera- ture. The electric energy must therefore be gradually diminished to pre- vent overheating. Electric motors are, for this reason, rated at a starting- effort, at a one-hour rating and at a continuous rating. The liability to overheating may be diminished by fireproof insulation and by artificial ventilation, with accompanying capacity for operation at increased voltage. ' The capacity of a motor for high voltage is limited by the resistance which the material in its construction offers to the pressure of the current. By this resistance, the flow of the current is checked and becomes transformed into molec- ular heat, which may reach an intensity destructive of the motor-apparatus. This defect is somewhat minimized by artificial ventilation of the motor, ^ In series connection, the current passes successively from first to i|ast through a series of motors. In parallel connection, the current is divided between the motors. MOTIVE POWER 81 In the substitution of electricity as a motive force on steam-driven roads, under ordinary operating conditions, it became necessary to sepa- rate the motor from the carriage as an independent tractor. In taking this step, it was assumed that it was a simple matter to apply the rotary motion and uniform torque, or twisting effect, of the motor directly to the axles of a tractor. This would obviate the objectionable features of un- balanced reciprocal weights and the dead points in the conversion of reciprocal motion in the steam-cylinder into the circular motion of the driving-wheels. The tractors originally designed for the Grand Central Terminal system were magnified motor-cars, operated by continuous current at low voltage, fed through a third-rail to motors acting directly on each axle, and intended to deUver a moderate tractive effort at high speed. In 1909, one of these tractors, weighing 102 tons, exerted a tractive effort of 7100 pounds continuously at 56 miles an hour and 20,600 pounds at a one-hour rating. The four motors could, therefore, give an output of 2200 horse- power for one hour without overheating. In 1913, a tractor designed for regular passenger-service weighed 100 tons, all on motor-driven axles. It developed normally 1400 horse-power, with capacity of 5000 horse-power for short periods, and with tractive power sufficient to take a train weigh- ing 1000 tons at a speed of 60 miles an hour. This design was subsequently modified to carry part of the weight on leading and traihng trucks. In a tractor placed in service in December, 1914, weighing 132 tons, motors were also placed on the axles of these trucks. The tractors of this type can exert a tractive effort of 14,000 pounds at 53| miles an hour, develop- ing 2600 horse-power at a one-hour rating and 2000 horse-power at a con- tinuous rating. Current is collected by eight under-running third-rail shoes, or by two overhead trolleys when at gaps in the third-rail. The trolleys are of the pantograph-type and are held in a raised position by the pressure of the motorman's foot upon a pneumatically operated valve. With experience, it was found that when the machinery, in an electric tractor for heavy traffic, is centered around the driving-axles and the weight concentrated below the top of the wheels, the effect is more destructive to the track than with a steam-locomotive. The motors were then placed upon the frame of the tractor, but still operated directly on each driving- axle, as in the New Haven tractors. The first of these tractors was placed on the line in March, 19Q7, weighing 78 tons. On February 4, 1910, a tractor was tested in freight-service on that line which weighed 175 tons, with estimated handling-capacity of 2000 tons at 45 miles an hour. It handled 30 loaded cars and a heavy locomotive at a speed of 50 miles an hour, making 18 miles in 27 minutes on a track slippery from freezing drizzle. The design of the New Haven tractor is complicated and its weight is increased by the necessity for third-rail operation on the New York 82 EFFICIENT RAILWAY OPERATION Central tracks, and also for transforming the alternating current on the tractor. It is mounted on four 62-inch driving-wheels and two trucks. On each end of the driving-axle, in the hub of the wheel, are seven circular pockets containing hehcal springs for assisting in carrying the weight of the motors and for transmitting the torque from the armature. Into each of these pockets projects one of the hollow pins on the end of the armature shaft. The springs fit into tubes, and these, with the springs, may be taken out from the pockets, which are lubricated. Though capable of sustaining the whole weight of the motors, these springs are normally used solely for transmitting the torque. The weight of the motors is carried on a steel frame, independent of the trucks and pivoted from the journal-boxes of the driving-axles. In this frame, the motors are suspended by springs, so that the swaying of the tractor has no effect upon them. For continuous-current work, the two motors are connected in series in starting, and in parallel at full speed. For alternating-current work, they are joined in parallel at all times. The tractor is controlled from either end by a lever like the throttle-lever, and any number of tractors may be operated from one controller. Two pantograph-bows collect current from the overhead wire. These collectors are pressed against the wires by springs and are lowered by compressed air. For use on the New York Central tracks, a lower overhead pantograph-trolley is required and also third-rail contact-shoes, two on each side of each truck, which are designed for both over-running and for under-running the third-rails. The shoes are lifted out of the way by an automatically operating device when alter- nating current is used. The Pennsylvania Railroad management, with its usual thoroughness, instituted an extensive series of experiments, in which it was found that, at speeds above 40 miles an hour, the steadiest riding was attained with a high center of gravity ; that the nearer the steam-locomotive design was approached in wheel-arrangement, distribution of weight, height of center of gravity and ratio of spring-borne to under-spring weight, the less was the side-pressure registered on the rail-head. In addition to excessive side-pressure, due to the oscillation of a low center of gravity, abnormal track effects were caused by the vertical pounding due to a large non- spring-borne motor-weight with imperfect spring-cushion. The remedy for these defects was found in a radical departure from the original design for an electric tractor as to running-gear and motor-drive. It was in 1909 that the Pennsylvania Railroad Company adopted a double tractor, the halves being permanently connected with the wheel-arrangement of two locomotives of the "American" type (4-4-0), set back to back. Each tractor has a gearless motor on top of the frames, driving through connect- ing-rods to a jack-shaft and thence by main and parallel rods to the wheels. These tractors are virtually on the multiple-unit system, as the complete tractor consists of two separate sections of 1000 horse-power each, and can MOTIVE POWER 83 be controlled in couples. The motors make 264 revolutions per minute at 60 miles an hour, and weigh 42,000 pounds each. A double-header of 4000 total horse-power, weighing 166 tons, can start a freight-train a mile and a half in length, and can make 75 miles an hour with a train of ten Pulhnan cars. These tractors were compared with the standard passenger- locomotives developing 1600 horse-power and weighing 90 tons, and cost- ing $22,000, while the electric tractors were estimated at $50,000 each.i Recent Installations of Electric Traction In the more recent examples of the electrification of steam-roads with heavy traffic, preference has been shown for single-phase current in an overhead conductor, transformed for delivery as a three-phase current into synchronous motors.^ This "polyphase system" has, however, been adopted in each case under exceptional conditions, and is open to the ob- jection that the train-service is restricted normally to predetermined rates of speed. The suburban lines of the Pennsylvania Railroad out of Philadelphia have been electrified since September, 1915, for 20 miles to Paoli, with 93 miles of track. The current is obtained from a power-company; single- phase at 13,200 volts from one phase of its three-phase generating system. It is then stepped up to 44,000 volts for transmission, and reduced again at four substations to 11,000 volts for delivery to the trolley-circuits. From these it is fed to two single-phase, air-cooled repulsion-motors, or "split-phase" motors, of 225 horse-power on each car, connected in series at 2200 volts. In these motors, invented by Professor Elihu Thomson, the current in the generator actuates the motor through induction only, by the repulsive effect of electro-magnetism, as evinced in the action of the mag- netic needle. By an ingenious arrangement of the induction-apparatus, a single-phase, high-tension current, as thus induced, is transformed into several continuous currents at lower tension. There are no rheostats, contactors or similar gear. The regulation is effected by mechanical ' In anticipation of the appKcation of electric traction on that portion of the main hne of the Pennsylvania Railroad traversing the summit of the Alleghany Mountains between Altoona and Conemaugh, 35 miles, an electric tractor has been designed to operate on 11, 000- volt, single-phase, 25-cycle current from an overhead trolley. The single-phase is converted into three-phase current for use in motors, four in number, rated at 1200 horse-power each. Two motors are mounted on each truck-frame, geared to a jack-shaft impelling driving-wheels through connecting-rods. Springs in the gears of the jack-shafts are adjusted to give the effect of solid gears up to 25 per cent, of the weight on the driving-wheels. The draft-strains are carried through the trucks and articulation in a direct plane 34i inches above the rail, to relieve the cab from such strains. Journal Am. Soc. Mechanical Engineers, August, 1917. ^ In a. synchronous motor, the speed of the motor is regulated by that of the dynamo or generator at a predetermined rate which can not be exceeded, but may be reduced by the intervention of resistance-coils, or rheostats, though with ac- companying waste of current. 84 EFFICIENT RAILWAY OPERATION transmission to the brushes of the motor, controlled by the movement of a hand-wheel in the driver's cab. An interesting example of the three-phase system for heavy traffic has recently been installed on the Norfolk & Western Railway, upon one of its most difficult operating divisions with heavy grades, and on which there is originated a large volume of coal-traffic. This division of 29 miles is double-track, except for 3100 feet through Elkhorn Tunnel, with 4 miles of passing-sidings, 7 miles of branch-lines, 5^ miles of mine-sidings and 21 miles of yard-track in five yards ; making a total of about 95 miles of track. The principal grades against east-bound traffic are in two stretches of 1.5 per cent, for about five miles each, one of 2 per cent, for four miles, one for nine and a half miles of 0.4 per cent., and one of three miles of 1.25 per cent. The Elkhorn Tunnel is on a 1.5 per cent, grade, ascending eastward, and is ventilated by the method of forced ventilation devised by Mr. Churchill, Chief Engineer of the Hne. Sixty per cent, of this divi- sion is on curves, with maximum curvature of 12° on the main line and of 16° on sidings. The normal service consists in collecting loaded cars and trains bound eastward, and in delivering return-empties. The coal-traffic had doubled in volume in four years and the average daily requirement had reached 65,000 tons. The maximum trains of 3250 tons each were handled by two Mallet locomotives, with a Mallet pusher up the heavier grades. These locomotives, with adhesive weight of 370,000 pounds and total weight of 540,000 pounds, had a tractive power of 85,000 pounds. The normal speed was seven to eight miles an hour on the grades, and six miles through the Elkhorn Tunnel, with frequent congestion and delay. In the substitution of electric traction, it was decided to use repulsion induction-motors, or synchronous "split-phase" motors, with two rates of speed, respectively at 14 and 28 miles an hour, and the tractors were specially designed to meet this requirement. The current was generated as single-phase of 1 1,000 volts at a central station and was stepped up. to 44,000 volts for transmission to four substations, where static transformers lowered the voltage again to 11,000 volts for delivery to the overhead con- ductor. On the tractors, this single-phase current was converted into a three-phase current at 750 volts for delivery to the motors. The trains when loaded are half a mile in length, and three-quarters of a mile when empty. To start a train, it is necessary that the head and rear tractors should exert their full power simultaneously. The motors there- fore permit full-load current to be appUed for five minutes before starting. This type of motor has the valuable characteristic of "dynamic" or re- generative braking, without the use of air, and with consequent saving of wear to the brake-equipment. When "drifting" on long grades, at a slightly higher speed than the predetermined rating, the static energy in the descending train may be utUized as a source of power in energizing a MOTIVE POWER 85 reverse action in the motor. The motor then acts as a generator of elec- tricity, which is transformed in the apparatus into single-phase current and returned into the overhead circuit. At a rate of speed four per cent, above the predetermined synchronous rate, the motor will deliver its full- load rating back into the Une. The continuous reaction of the motor- torque resists the accelerating effect of gravitation and preserves a uniform rate of speed during the descent. These tractors are technically termed "articulated geared side-rod," being operated in this respect like the Mallet locomotive, as a coupled pair, and with similar wheel-arrangement (2— i-4-2, 2-4-A-2)} The most ambitious project yet undertaken in the substitution of electric traction for steam has been installed on the extension of the Chicago, Mil- waukee & St. Paul Railway to the Pacific Coast, mentioned in the previous chapter. In passing over the Great Continental Divide, from Harlowton, Montana, to Avery, 440 miles, the line crosses the Belt Mountains at an elevation of 5768 feet, the main Rocky Mountain Range at 6350 feet, the Bitter Root Mountains at 4200 feet and the Cascade Moimtains at 3010 feet, with several tunnels, including one of 9000 feet. The maximum grade westward is 2 per cent, for 20.8 miles ; and east-bound, 1.7 per cent, for 24 miles. The line ascends the eastern slope of the Belt Mountains on a con- tinuous one per cent, grade for 44 miles. The difficulty in procuring suit- able feed-water and cheap coal, and the proximity of an ample supply of electric current from a hydro-electric company, induced the railway com- pany to operate this entire district of 440 miles, covering 650 miles of track, by electric traction only ; an enterprise of greater magnitude in this respect than had before been undertaken, and at an estimated cost of $12,000,000. A division of 115 miles was put in operation on December 5, 1915, and another division of 112 miles was to be so operated in May, 1916. Power is delivered from the hydro-electric plant through three lines of copper-tubing, which is an important improvement over wire conductors, since the ratio of weight to transmitting surface is thereby greatly reduced. Three-phase current at 100,000 volts is distributed to fourteen substations, where it is transformed to 3000 volts — the highest continuous current used anywhere for electric traction. This high potential for continuous current was adopted after experience for a year and a half on the Butte, Anaconda & Pacific Railway with 2400-volt continuous current, fed into 1200-volt motors in series. In this installation, two 60-cycle, 1500-volt synchronous motors are connected permanently in series.^ The motors are geared to the driving-axles of the tractor, a coupled pair; each half with eight drivers and a four-wheel truck. Each tractor weighs 260 tons, is 112 feet in length, costs $112,000, and has a tractive ' For details of electric tractors and traffic performance, see Appendix II, Tables XVI and XVII. * In a 60-cycle motor, the current changes direction 60 times a second. 86 EFFICIENT RAILWAY OPERATION power of 85,000 pounds, against 75,000 pounds in the Mallet locomotive which it replaces. The tractors for passenger-service are designed to haul an 800-ton train at 60 miles an hour on a level, or at 30 miles an hour up a one per cent, grade. The freight-tractors can take a 25€0-ton trailing-load up one per cent, grades at 16 miles an hour, or up a two per cent, grade assisted by a pushing tractor. At the summit, the pusher is taken to the head of the train and the motors in both tractors are utilized in retarding speed on the descent by regenerative braking, and without using the air- brakes. By this means, from twenty-five to fifty per cent, of power is recovered and turned back into the trolley-wire.^ ( ' -^ The most obvious objections to three-phase tractors are the compli- cation of a double overhead wire, the danger that the motors do not share the load fairly, and the inabiUty to run at intermediate speeds without rheostatic waste, or to run at higher than synchronous speed to make up lost time. The objections that the motors will not share the load equally is theoretically sound. In order that they should do so, their speeds must be the same ; in practice, this difficulty, arising from slight differences in the size of the driving-wheels, is counteracted by the sUp-adjustment of the motors. But a train pulled by a series-motor, whether with continuous or alternating current, runs slower up-grade or if abnormally heavy, while, with the three-phase motor, the speed up-grade may be practically the same as on a level. Also, the feature of regenerative braking, peculiar to the synchronous motor, is a valuable adjunct to train-service on long and heavy grades. COMPAEATIVE ECONOMT OF StEAM AND ElECTBIC TrACTION Much remains to be done as to design and performance before the electric tractor reaches the approximate perfection that has been attained with the steam-locomotive. The steam-locomotive is a primary source of power ; the electric tractor is a secondary source, whose efficiency depends upon its continuous connection with the power-station. By the temporary separation of the generator from the motor, by means of storage-batteries, a quasi-independent electric tractor has been devised by Beach and Edison for street-railway service. It was experimentally tested, in 1910, between New York and Boston, via Albany, but it does not meet the requirements for steam-railway operation.^ As a mobile tractor, the theoretical advantage of the continuous appli- cation of power by rotary motion has been discounted in the electric ' For additional data see Appendix II, Table XVII. 2 On Sept. 26, 1912, it was announced that a train operated by storage-bat- teries under multiple-unit control, built for the United Railways of Cuba, had been tested on the Long Island Raiboad. The train was composed of three ears, each with four 200-volt motors and 216-cell Edison batteries, recharged from third rail, with mileage-capacity of 60 to 100 miles from seven hours' charging. MOTIVE POWER 87 tractor operating through connecting-rods and cranks. The capital in- vested in electric traction becomes proportionately greater with fewer motors, and that consideration alone points' to its restriction to frequent train-service. Nor does the opportunity for economy in fuel-consumption at central power-stations outweigh the merits of the independent steam- tractor. The saving from this cause must cover interest on the investment in the entire electric plant and for its maintenance, as well as for covering the losses involved in converting mechanical power into electric energy at the power-station, conducting the current to the tractor and there re-con- verting it into mechanical power. The cost of protecting telegraph and telephone hnes and signahng apparatus from induced currents or from electrolysis is also to be considered. While steam-locomotives may be idle for much of the time, it is not possible to interrupt the running of the power-station, though the loads may be reduced, for the whole conduct- ing system must be kept supplied with current. In a papter read in 1914 before the British Association for the Advance- ment of Science, Professor W. E. Dalley asserted that as yet there was no appreciable difference in the efficiency of the electric tractor, as compared with the locomotive, for the expenditure of an equal number of heat-units. After allowing for all losses in both cases, only about four per cent, of the total energy of the fuel consumed appears as mechanical power at the driving-wheels of either tractor. The electric tractor can, for a brief period, exert over double its normal power — a quality that the locomotive does not possess, and one which is advantageous on fines with frequent short and heavy grades, and it can also attain a higher speed more quickly from a- state of rest under similar conditions. The electric motor-car can develop a speed of 30 miles an hour in 30 sec- onds, using the whole weight of the train for adhesion, and giving it superi- ority for suburban traffic with frequent stops ; but this advantage is min- imized when trailers are used, and, with the adoption of the locomotive-wheel arrangement, it entirely disappears. The motor-car wastes no current when idle, while the locomotive must be kept under steam to be ready when required. Electric tractors do not have to be turned, and their superiority for use in long tunnels is manifest, if only as a matter of social efficiency. The multiple-unit motgr-car train can be economically modified to suit fluctuations in traffic. There are but few locomotives that can generate sufficient steam to furnish full-cyhnder tractive power at speeds in excess of twelve miles an hour. Increase of speed must be obtained by sacrifice of tonnage, and to this fact is due the high cost of fast-freight service. With the electric tractor, however, high speed may be attained without corresponding loss of tonnage, since, with a relatively unlimited source of power at command, the maximum drawbar-puU may be maintained at all speeds. But, at present, for general railway service at speeds not exceed- ing 60 miles an hour, there is, after all, no more economical tractor than the 88 EFFICIENT R^ULWAY OPERATION steam-locomotive. In a comparison of the efficiency of these two forms of motive power, it is also to be considered that, where the power is derived solely from a central source, any interruption in the output affects every tractor dependent upon it, and a general disturbance to the service ensues ; while any interruption to the operation of an independent tractor affects its immediate environment only. The advantage of the steam-locomotive in this respect can only be neutralized by the successful operation of an electric tractor which is similarly self-contained, and that is as yet a desider- atum. American and Foreign Methods of Electric Traction In this country, almost all trunk-line electrification has been on the continuous or "direct current " system ; while in Continental Europe, it has been principally on the "three-phase alternating current" system. Apart from other considerations, this difference of opinion upon the fundamental problem of railway electrification has arisen from differences in the re- spective social environments. The extensive area covered by our railway system under similar conditions as to government, language and social habits and customs, has enforced uniformity in railway operation in mat- ters in which an interchange of equipment or an observation of one and the same set of traffic-regulations is necessary for a cheap and convenient public service. The consequence is a 'continuing tendency to recognize the advantage and necessity of uniform methods and appliances in railway operation, and the establishment and maintenance of standards. It is the existence of this tendency that has led to the manufacture of railway equipment and appliances, as commercial undertakings, on a larger scale than elsewhere. This has been particularly the case with electric equipment of all kinds. With the advent of steam-railway electrification, the great industrial corporations entered energetically into this field, while it was yet in its experimental stages. The initial work was taken out of the hands of individuals and concentrated under their own direction and in accordance with their own standards. As these were derived from street-railway experience, it was to be expected that street-railway electric practice should at first predominate, as was the case with the adoption of the "direct-current" system. Electric traction was developed in Great Britain under the direction of these same great manufacturing concerns but, in Continental Europe, the social environment gave prominence to the views of technical experts, who were not so much affected by commercial considerations as by theoretical training. Their respective fields of operation being restricted by national boundaries, there were many centers of development and much diversity in methods and appliances, by which the progress of railway electrification has been hindered rather than advanced. State-control has enabled state officials to exercise a benumbing authority in every stage of its develop- MOTIVE POWER 89 ment. In this country, the advocates of the alternating-current systems had to contend with the influence of the corporations interested in the development of electric traction on commercial lines. In Continental Europe, and particularly where German influence was authoritative, the opinions of official electricians have prevailed. The " three-phase alternating current " system was established in Ger- many by agreement between the several state governments. In this country, the "direct-current" system was introduced by the extension of street-railway experience under the influence of corporations engaged in the manufacture of direct-current equipment. Then the great railway corporations intervened and, after extended and expensive experiments, took in hand the designing of electric tractors for trunk-line traffic which were far more efficient than any that have been originated by the oflBicial electricians of the European states who, in fact, have had to look to this country for a lead in steam-railway electrification.^ DiPFEKENT Applications of Electric Traction. Gasoline Motors At present, electrification of railways should be considered as it affects I. City rapid-transit lines. II. Quick-transit suburban lines. III. Lines directly connecting large cities. IV. Trunk-fines with long-distance traffic. The principal examples in the first category are the municipal systems of New York, London and Paris, in which the "direct current" system has proved its efl[iciency. The extension of the city rapid-transit lines into quick-transit suburban lines has induced the use in common of motor- car trains on the multiple^unit system, and here experience seems to indi- cate a combination of the "three-phase" system with its transformation into continuous current, as fed into the motors. The electrification of fines extending beyond the suburban ' zone has been accomplished hitherto with the direct-current system at low voltage, because it is also a further extension of the city street-car fines with which it connects ; and here, too, the three-phase current, transformed at sub- stations into direct current, seems to be preferable. But the electrification of interurban steam-railroads, without street-car connections, is another matter, since it involves the handfing of freight as well as passengers, and the operation of freight-terininals and intermediate sidings to industrial enterprises. With the growth of traffic on such lines, there is a tendency to differentiate the methods of electric traction for passenger-service and for ' Much information about electric traction has been obtained from the Bulletins of the International Railway Congress, and especially from the Report on American practice by George Gibbs, Chief Engineer of Electric Traction on the Pennsylvania Railroad lines, presented at the meeting in Berne, in 1910. Valuable assistance has also been kindly given by raih-oad ofiaoials engaged in this branch of engineering. 90 EFFICIENT RAILWAY OPERATION" freight-service, preserving the multiple-unit system, with trailers, for the passenger-service, and gradually introducing tractors of a different de- sign for the freight-service. At the time that it was proposed to substitute electric traction for steam on the Grand Central Division of the New York Central Knes, there had been no experience in its application to high-speed trunk-hne passenger- trafHc. The tractors designed on the lines of the motor-car, developed the unsuitability for such traffic of motors geared directly to the driving-axles. Next, the application of electric traction to the entire long-distance traffic of a trunk-line was undertaken by the New York, New Haven & Hartford Eailroad Company. Here, the tractor was differentiated from the motor- car, with single-phase current transformed on the tractor and fed into the motor at higher voltage than had been before attempted in electric trac- tion. This led to a departure from the third-rail conductor, and a return to the overhead wire in a very expensive form, which compelled an entire rearrangement of adjacent telegraph and telephone lines and of the rail- way signahng-apparatus.' The design of these tractors was complicated by the necessity for providing direct-current operation in the Grand Cen- tral Zone. Although this installation has proved that the entire long- distance traffic of a trunk-line can be successfully conducted on the single- phase, high-tension system, yet the cost of its application between New York and New Haven was so enormous that other railway managements have been deterred from adopting it in its entirety. The character of the electrification of the Pennsylvania lines within its Manhattan Terminal zone was determined chiefly by the operation of the extensive suburban traffic of Long Island, on the direct-current, multiple- unit system. For its high-speed trunk-hne passenger-traffic, however, a notable departure was made from previous practice as to independent trac- tors, by transforming the rotary motion into reciprocal motion in its trans- mission to the driving-wheels. This change affects its through passenger- traffic for only a short distance out of New York, and has not been imitated elsewhere, except on the Norfolk & Western Railway for coal-traffic. ^' The principal substitution of electric traction for steam, has been com- pelled by considerations of social efficiency, — the necessity for providing frequent and speedy passenger-service for short distances. Incidentally, it has been enforced as a pohce measure, to free a populous region from the nuisances of steam and smoke. Its claim to economic efficiency in this field is limited, at present, to the rehef which it affords in the operation of passenger-terminals already crowded to their capacity. Outside of such terminal zones, and such suburban traffic, electric traction had been sub- stituted for steam only in badly-ventilated tunnels. But, with the introduction of the induction-motor, a new field was 1 The peculiar features of electric track and overhead construction are described in Chapter V, Part II, pp. 263-256. MOTIVE POWER 91 opened for electric traction in the conduct of heavy freight-traffic. The importance of this recent development is apparent in the description al- ready given of its installation on the Norfolk & Western Railway and on the Chicago, Milwaukee & St. Paul Railway. In the latter instance, an example is furnished of complete electric traction on a trunk-line of 440 miles, operated under peculiarly difficult conditions. If its promise of economic efficiency should be there fulfilled, the induction-motor, with high-tension alternating current, would seem to be the system best suited for heavy traffic, where high speed is not required. On lines not exceeding a hundred miles in length, through a densely populated region and between large cities, as between New York and Philadelphia, it might be found advisable to conduct the freight-traffic by steam and the passenger-traffic by electricity ; but on separate tracks. As to long-distance freight-traffic and passenger-traffic, on trunk-Unes, there is a growing conviction that the third-rail conductor and continuous current are not preferable to steam for cheap operation; that the New Haven installation is too expensive as a capital investment, and that the three- phase system is efficient only under exceptional conditions. It would seem from the standpoint of motive-power efficiency, as well " as from that of railway efficiency in general, that steam will continue to be the cheapest motive force for ordinary railway purposes until some hitherto undeveloped form of electric traction has been discovered, requir- ing less capital investment and cheaper cost of maintenance. Until then, steam-locomotives will continue to haul heavy trains for long distances over ordinary grades, in regions lacking abundant and available water- power but well provided with good fuel. Eventually, American ingenuity may solve the problems that must precede the electrification of our railway system so as to combine theoretical efficiency with commercial economy, and establish uniform standards of method and appliances. Only when this has been accomplished can there be a just comparison of the respective n fields of steam-traction and electric traction. As the use of electricity in street-railway operation led to its intro- duction on steam-railroads, so the remarkable development in the appli- cation of gasoline in explosive engines for highway-traffic has induced some experiments in its use in "rail motor-cars." In 1904, the Union Pacific Railroad Company placed such motors on branch-fines, where their apparently low operating cost might warrant more frequent passenger- service in thinly populated districts over roads with heavy freight-traffic. In 1911, that company had 113 of these cars in service. One built in 1915 is described as developing 300 horse-power in a seventy-foot car carrying mail, baggage and express, and making 55 miles an hour with a standard steel passenger-car as a trailer.' A gasofine motor for switching-service 1 Weight on driving-wheels, 33,800 pounds ; tractive power, 8200 pounds. Average monthly mileage of 6000 miles, making round trip of, 206 miles daily. 92 EFFICIENT RAILWAY OPERATION was tried on the Pennsylvania Railroad in 1909. It weighed 42,000 pounds, with speeds of eight and of twenty-two miles an hour, and hauled twenty loaded cars on a level track. Some further data have been pub- lished as to the use of such cars on the Ann Arbor Railroad, upon a line of 300 miles. In June, 1911, the three cars there in use were operated at a cost of 13.73 cents per car-mile against a train-mile cost of 60 cents. Under the two-cent-a-mile rate of fare, the revenue was 43 cents per train-mile. The cars, all-steel, cost $22,500 each, and the interest-charge was esti- mated at 13 cents per car-mile. Standard of Motive-power Efficiency A discussion of motive-power efficiency should include a reference to some standard of such efficiency for purposes of comparison. A distinc- tion is to be observed between data as to economic efficiency and as to cost ; for the efficiency of a steam-generator is measured by its absorption of heat in thermal units, while the cost of combustion varies with the calorific value of the coal consumed and also with its price per ton. There have been instances in railway operation in which a very considerable de- crease in the annual fuel-consumption has been neutralized by an advance in the price of coal. On the New York Central Railroad, in 1913, coal-consumption con- stituted 58 per cent, of the total locomotive-cost per mile. Though the average performance in that year increased 896 miles per locomotive, with decrease in total consumption of 60,194 tons, the average price of coal increased six cents per ton. As a consequence, the total expenditure for coal increased by $278,814 ; an increase tantamount to a dividend of five per cent, on a capital investment of $5,576,280. If there were included in this statement the average caloric value of the coal consumed, the impor- tance of ascertaining the relation of cost to efficiency would be made more apparent. The comparative efficiency of locomotives as mechanisms is largely a matter of design which, as to steam distribution, is measured by the in- closed area on tjie indicator-card. As the drawba^-puU exerted is pro- portionate to this area, it is therefore a measure of comparative efficiency. In other respects, a comparison of annual tonnage-miles may furnish a more adequate standard as to the merits of different designs for the same freight- service than an average of the annual cost of repairs. The cost-data as to repairs are not in fact efficiency-data, even as to mechanical excellence of construction. The character of workmanship would be more definitely established by the ratio of hours of labor on repairs to the annual mileage. The main element of efficiency in a locomotive is its tractive power and, for purposes of comparison in this respect, a third group of physical data is required. Such data are obtained from the measurement, by a dynamom- MOTIVE POWER 93 eter, of the power developed at the locomotive-drawbar, or the drawbar- pull. In this whole estimate of comparative efficiency there is an interrela- tion of data as to design, construction, cost and operation, which it is im- practicable to coordinate in a single group, for purposes of comparison, or to measure by a single unit. But it may be considered that, as fuel-con- sumption is at the beginning of such a series of data and drawbar-pull at the end, a standard of motive-power efficiency for most practical purposes, from a technical point of view, may be found in a comparison of the amount of drawbar-pull that is made available by the consumption of a definite weight of coal. Where locomotives are operated in the same service on the same runs and with the same quality of coal, this standard should be sufficient. If a comparison is to be made between locomotives using dif- ferent qualities of coal, the ratio may be determined by the relative calo- rific value, which may vary from 8000 British thermal units for slack to 14,000 units for anthracite or for semi-bituminous coal.^ Tests of this character develop the theoretical efficiency of locomotives. To compare their value for continual service, another factor must be in- cluded in the comparison, in addition to fuel-consumption and tractive power; that is, the factor of time. How many hours per day has the locomotive been in profitable service ? In that period, how many pounds of tractive power have been developed as drawbar-pull, and with what fuel-consumption? Any useful standard of efficiency must take into account whatever information is of practical value, as to the relative merits of locomotives as tractors, for this, at the last, is the true measure of motive-power efficiency. It would not be necessary to obtain such data continuously as to all of the equipment, but only for each locomotive at such intervals as would give confidence in the reliability of the resulting averages as a just basis for comparison. The results which would follow the application of statistics of this char- acter to railway operation may be briefly stated. From an annual compari- son of all the locomotives on any road on the basis as above indicated, a conception might be formed of the varying potential efficiency of that road as a measure of transportation, that could not have been afforded by any other statistical data. The advantages that would result from the general application of the pound of tractive power developed per hour of service as a basic unit in railway operation would be made apparent by such com- parisons as are here suggested. The total theoretical tractive power of 1 Tests of coal in use on the Atchison, Topeka & Santa ¥6 Railway, in 1907, from thirty different sources, gave a range of from 10,261 units for slack to 13,847 units for the best quality ; so that the best coal was more efflcient as a heat-pro- ducer by 31 per cent, over the poorest. Of the different qualities tested, three ranged between 10,000 and 11,000 units, eleven between 11,000 and 12,000, ten between 12,000 and 13,000, and six were over 13,000. William Garstang, International Railway Congress, 1910. 94 EFFICIENT RAILWAY OPERATION the entire locomotive-equipment of the railway system of this country at a given date having been ascertained, the potency of that system as an instrumentality of transportation might be definitely established ; as also the ratio of the service which it actually performed at that date to that of which it was theoretically capable. If the valuation of our railway system, which is now in progress, were to include statistics, not only as to the number and commercial value of the tractor-units belonging to that system, but also as to their theoretical and actual tractive power in pounds, such statistics would have great practical importajice in a discussion of the relative efficiency of the rail- ways of which that system is composed. The efficiency of the locomotive as a tractor is the real basis of efficiency in railway operation.^ 1 Much of the technical information in this chapter has been obtained from the " Report of sub-committee on Railroads, and Locomotives of To-day," and the contributed discussion on " Steam-Locomotives of To-day," pubhshed in the Journal of the American Society of Mechanical Engineers, January, 1915 ; also from the publications of the International Railway Congress ; from " Economics of Railway Operation," M. L. Byers, 1908; " The Utilization of Fuel in Loco- motive Practice," W. F. M. Goss, in Bulletin 402, U. S. Geological Survey, 1909; and personally from E. P. Ripley, President, Atchison, Topeka & Santa Fe Rail- way System ; from E. M. Coapman, Vice President, Southern Railway Company, and from C. A. Goodnow, Assistant to President, Chicago, Milwaukee & St. Paul Railway Company. CHAPTER IV ROLLING-STOCK EabIiY Forms of Railway Cars. English and American Types The rolling-stock on the British railways was developed from the colliery-wagon and the stage-coach. As the colliery-railways became utilized for general trafi&c, the wagons were used for carrying goods, pro- tected from the weather by tarpaulin-coverings. Live stock was trans- ported in wagons provided with railings, and fourth-class passengers as well, without seating them. Afterward, these wagons were roofed over and closed in for general goods-traffic. The passenger-carriages were similarly developed from the stage-coach by placing three coach-bodies back to back as compartments. The gentry preferred to remain in their own carriages, loaded on an open car, rather than to be associated with the commonalty, until a compromise was ef- fected by assigning them, as first-class passengers, to special compartments more luxuriously upholstered. This class-distinction was further extended by separating the working people, as third-class, in plainly furnished com- partments. Luggage was placed on the roofs, railed in, or in the spaces between the curved bodies, as in the boot of a stage-coach. As the early railways on the Continent were, for the most part, planned by British engineers, the wagons and carriages introduced by them have persisted as types to the present time. In fact, it was not until 1892 that passage from one car to another, otherwise than along the running-boards, was made practicable by the introduction of "corridor-trains." In the United States, different physical and social conditions have con- trolled the character of the rolUng-stock, as of other factors in railway operation. The incentive to railway construction was not to secure cheaper and more expeditious means of carrying coal to market. The governing condition was the want of highways for commercial and social intercourse. It was not a question of putting flanged wheels under existing coal-wagons and stage-coaches, but of providing vehicles for goods and travelers in the cheapest way. There were a few early imitations of the Enghsh railway- carriage, and the name of "coach" for a passenger-car still lingers in the railroad man's vocabulary, though the name of "wagon" for a freight car has not gained recognition. The freight-car for goods was from the be- 95 96 EFFICIENT RAILWAY OPERATION ginning closed in and roofed over, and, as a safety measure, sliding doors instead of hinged doors were used on all cars with side-openings. The flat car, or platform car, was more common than the open coal-car, some- times known as the "gondola," as liunber was a more general article of railway trafHc, and for this reason the flat car was longer than the English coal-wagon. But the one feature which profoundly differentiated American rolling- stock from the English types was the center-bearing four-wheel truck. Its superiority to the* rigid wheel-base is obvious in its facility for contin- uous adjustment to the alignment of the track, whether that be due to curvature or to defective maintenance. Though this is a matter of greater moment upon hastily and cheaply constructed roads than upon those which are really "permanent ways," it has also the merit of neutralizing the horizontal and vertical oscillations of a vehicle in motion. For these reasons, the center-bearing truck was in general use in the United States long before it had gained recognition elsewhere, and with far-reaching con- sequences in other respects. Rolling-stock design may be separated into three elements : the body, the underframe, and the running-gear. The body of the car, being merely a receptacle or a shelter, has been variously modified to meet the differing requirements of commerce, of travelers and of railroad operation. The underframe receives directly the shocks and strains from starting and stopping a train and from variations in speed while in motion. These strains are transmitted to the car-body by the bracing in the side and end frames, and to the center-bearing truck through the main cross-sUls or bolsters. The general structure of roUing-stock is designed to resist the strains occurring in train-service. The lighter the rolling-stock, the less is the shock upon its structure in starting and braking a train. The greater the speed, the more important it becomes to reduce the weight of the car relatively to the acquired momentum of the train. The relation of empty weight to tonnage-capacity is also of prime importance in connection with the profitable application of tractive power in train-service. This matter is kept carefully in view in designing freight-equipment, though but little attention is given to the seating-capacity of passenger-cars in proportion to their weight. Here, the comfort and convenience of travelers has been foremost, and considerations of social efficiency have overborne regard for economic efficiency. Social efficiency has also been influential in the numerous designs of rolling-stock for special classes of traflJc. Hence the varieties of freight- cars, as coal-cars, lumber-cars, live-stock cars, furniture-cars, etc. ; and in passenger-service of mail-cars, express-cars, baggage-cars, smoking-cars, parlor-cars, sleeping-cars, dining-cars and observation-cars ; and in railroad working equipment, snow-plows and wrecking-cars. ROLLING-STOCK 97 Many of these various designs originated in the United States and were influenced by peculiar traffic-conditions, as the grain-car, the tank-car, and the refrigerator-car. The ordinary box car was built of enlarged dimensions for carriage of light and bulky commodities until the size of the "furni- ture" car reached the roadway clearance-limits. The open live-stock car was closed in and furnished with facilities for feeding and water- ing in transit. For the transportation of valuable animals, padded stalls were provided, heated and ventilated, with accommodations for attendants, and such cars were well characterized as "palace stock-cars." One of the most valuable of the special types of freight-cars that have originated in this country was due to the physical and social environment in which it was developed. The extensive mileage of our railway system, with a uniform gauge of track, covering half a continent from ocean to ocean and from the Great Lakes to the Gulf of Mexico, has invited an in- terchange of commodities produced under different conditions of soil and climate that was facihtated by the intense competition for such long-haul traffic between rival transportation-lines by land and by sea. The radius of profitable distribution of perishable commodities was limited by vicis- situdes of weather and by their tendency to decay. These conditions acted adversely to each other. The inherent heat of vegetable products was hastened if the car containing them were tightly closed ; if the car were ventilated, they were liable to damage by dust or by a sudden changp of temperature. These difficulties were overcome by the adaptation of uie common household-refrigerator, first to the fresh-meat traffic. The car- body was lined with material that is a poor conductor of heat, and its con- tents were thus protected from external variations of temperature. The inherent vegetable or animal heat was absorbed by ice, introduced through the roof into boxes at the end of the car, and the temperature within was accordingly controlled by the indications of a thermometer exposed to view from the outside.' The framework of the English passenger-carriage was so strengthened by the compartment bulkheads that no reliance was placed upon the side- frames to stiffen the underframe, and side-doors were therefore not objec- tionable as a structural feature. The American car was not only longer, but, as it was not divided into compartments, dependence was placed upon the side-frames to give rigidity to the underframe, and side-doors were only provided for in baggage-cars. Entrance was therefore made over platforms at the ends; a construction that admitted of communication throughout the train. This feature, which was a matter of course in a democratic country, was an objection where class-distinctions were ob- served. The advantage of such intercommunication in railway operation has, however, forced its adoption abroad, though with reluctance. The seclusion so highly prized was made less desirable by occasional acts of 1 See Chapter VI, pp. 344-348. 98 EFFICIENT RAILWAY OPERATION robbery, and even of murder, in isolated compartments. A compromise was accepted in the introduction of a passage-way or corridor along one side of the compartments. In Germany and in Switzerland, and to some extent in Italy, the open car is coming into use because of its greater seat- ing capacity ; so great has been the effect of the substitution of a center- bearing-truck for a rigid wheel-base ! From a social as well as from an economic standpoint it has dominated the whole passenger-train service. For many years, the underframe was not sheathed underneath. The occupants were subjected to noise from the running-gear, to dust and draft through the cracks of the loosely jointed and uncarpeted floors and to possible injury from the floor-joists being torn out in case of derailment. Further protection against this contingency was even sought in construct- ing the whole body of the car with staves, hooped like a barrel, with peep- holes in the sides and entrances in the heads ; the idea being that, in the event of a derailment, the car-body would leave the trucks and roll over instead of being smashed to pieces. A car of this construction was long preserved as a curiosity on the South Carolina Railroad. In the days of the strap-rail, a broad and stout plank was suspended between the trucks, and just above each rail, by braces from the underframe, as a pro- tection against "snake-heads" penetrating the floor. Though the interior fittings of the early American equipment were less luxurious than in the European compartment-carriages of the same period, necessary provision was made for the comfort of their occupants which was painfully lacking in the same class of equipment abroad until a much later date. The arrangement of the seats in parallel rows was due to the end entrances, and the turn-over seat-backs obviated the necessity for turning the cars to avoid riding backward, as well as the propinquity of strangers facing each other. By degrees, the car-bodies were lengthened from forty to fifty and sixty feet, to seat a larger number of passengers, with relatively less dead weight. The increase in length was accompanied by improvement in trussing the side-frames, by a greater use of iron rods as tension-members and of iron plating to strengthen the joints. The space in the underframe between the flooring and the sheathing was filled with sawdust or other material to deaden the noise from beneath. The deck-roof with upper light and ven- tilation was substituted for the low continuous roof. While no class-dis- tinction was observed, smoking was only allowed in specially designated cars. These cars often included a "conductor's office," which served as a private lounging-room for favored friends and, in the Southern States, one end of such cars was usually reserved for negro passengers. There was a gradual improvement in the treatment of the exterior of the car-bodies with respect to paint and varnish, approaching more nearly to the superior finish of English railway-carriages which had been inherited from coach-builders. The interior finish depended upon the price of the ROLLING-STOCK 99 car and the taste of the purchaser. Hardwood veneer was generally used, varying in material and design with changing fashion in furniture. The upholstery varied, under similar conditions, from enameled cloth to carpet- stuffs and furniture-plush. Sleeping Cars A substantial improvement in passenger-equipment followed upon the introduction of the sleeping-car, which was developed in the characteristic American environment, beginning with an experimental car on the Cum- berland Valley Railroad in 1836.^ By 1859, sleeping-cars were in use between Albany and Buffalo. The ordinary passenger-car was fitted up with three tiers of berths, as in the open steamboat-saloons. No bedding was provided other than mattresses and pillows, until the service was undertaken by companies organized for the especial purpose. One of these sleeping-car companies came under the control of George M. Pullman in 1864, and to him the traveling public owes the specially designed and luxuriously fitted cars with which his name will always be associated. To his mechanical judgment and power of organization have also been due many of the improvements in design and construction of passenger-equipment and in provision for the comfort of travelers, now in general use in other countries as well as in the United States. Mr. Pull- man introduced the six-wheel truck to sustain the heavy weight of his cars, also steel-tired wheels. But his most valuable improvement was the vestibuled platform, patented in 1887, which provides a protected passage- way from car to car, ingeniously contrived to utilize the entire end-surfaces of adjacent cars as buffers. The adaptation of cars to night-service in this country was followed by the construction of cars specially designed for long-distance journeys, as parlor-cars, dining-cars and observation-cars. In the provision of such facilities for comfort and convenience of travelers, the railroad managements of the United States have been pioneers.^ For the comparatively short railway trips in Great Britain, sufficient rest at night could be obtained in the ordinary compartment-carriages, with the aid of traveling-rugs. For the longer continuous journeys between London and Scotland, sleeping-carriages were placed on the West Coast Route in 1873, and Pullman cars on the Midland Railway in 1875. "Corridor sleeping-carriages," with eight wheels and forty-two feet in length, were introduced on the London & Northwestern Railway in 1883. Before that time, no vehicle over thirty-three feet in length had been in use, because of the diameter of the turntables. Upon the Continent, no other pro- vision was made for comfort in night-travel than a hired pillow, until long after luxurious nocturnal accommodations had become general in the United States. 1 "When Railroads were New," p. 171. 2 See Chapter VI, p. 316. 100 EFFICIENT RAILWAY OPERATION Development of the Running Gear. The Bogie Truck The mnning-gear of English rolling-stock was developed from that of the four-wheel coUiery-wagon, with inside journals and without springs. Subsequently, pedestals were attached to the side-sills of the underframe and the axles rotated in boxes fitted for vertical motion in the jaws of the pedestals. Bearings of brass or other anti-friction metal were interposed between the journals and the top of the boxes, which also served as recep- tacles for lubricating material. Before the introduction of springs, the shocks from inequalities in the track-surface were transmitted directly to the underframe. The length of the original EngUsh car-bodies was hmited by the load which could be carried upon an underframe supported by a pair of axles, and by the length of rigid wheel-base that could safely trav- erse the maximum curvature in the track. Provision was afterward made for longer car-bodies by introducing an intermediate axle, but the rigid wheel-base was still restricted by the track-curvature. While different traffic-requirements induced differences of design in car-bodies, the design of the underframes in American rolling-stock has varied from English prototypes in the manner in which resistance is offered to the stresses of tension and compression in train-service. It was from the long-coupled wagon swiveling on a king-bolt, that the American center- bearing truck was developed. The underframe of the rigid wheel-base was separated, by this device, from the underframe of the car-body and became an independent unit in rolling-stock design. The rigid wheel-base could be shortened nearly to the diameter of a car-wheel, and the under- frame became a truck-frame. One of these truck-frames was connected with the underframe of the car-body, near each end, by a center-bearing, pivoting freely upon a center-pin or king-bolt through the cross-sill or body- bolster. The rigid wheel-base, so shortened, traversed curves which were unsafe for the older type of running-gear. The shocks from the track, that were before transmitted, directly to the car-body at four points, were now divided at eight points, and the resulting angular motion was further lessened by its reception at two points in the median line of the car-body. The center-bearing truck was a decided advance in efficiency, since it permitted the construction of roIHng-stock of greater length and loading- capacity than was before admissible. It is attributed to the inventive genius of Mr. Jervis, the engineer who built the Mohawk Valley Railroad, between Albany and Schenectady, about 1831, and was introduced under the eight-wheel passenger-car built by Ross Winaiis, in 1833, for the Balti- more & Ohio Railroad, together with his invention of outside journal- bearings. Steel springs are said to have been first used under the loco- motive "York," on the same road, in September of that year. Prior to that time, in England, as in this country, passenger-car bodies were carried on four wheels, slung in leather braces like the body of the stage-coach ROLLING-STOCK 101 from which they had been developed. The first steel springs were prob ably of the semi-elliptic type. As late as 1870, there were ten-ton freight- cars in use on Southern roads which were carried on trucks without side- sills or cross-sills. A long arid heavy semi-elliptic spring resting in seats on the journal-boxes served as the side-sill, and the ends of the truck- bolster were attached to the middle of these springs. The center-bearing truck was subsequently modified in important de- tails. Among these was the floating bolster, resting upon steel springs, either coiled, elliptic or semi-elliptic ; and equalizing-springs that received the shock from the journal-boxes through a bent bar upon which the springs rested, that in turn supported the side-sills of the truck-frame. The purpose of this complex spring-system is to divide and minimize the shocks transmitted through it from the track to the car-body. The efficiency of the center-bearing truck for this purpose, and also in shortening the rigid wheel-base, has led to its general use throughout the world. Its effi- ciency has been further increased by the introduction of the six-wheel truck, which not only conduces to easier riding qualities but also permits the construction of rolUng-stock of increased capacity and weight.' The use of steel as a construction-material has, however, brought about a return to the use of the four-wheel truck under heavy passenger-equip- ment. On the Pennsylvania Railroad, the four-wheel truck is used under cars 70 feet in length, seating 88 passengers, weighing (light) from 118,000 ' to 120,000 pounds and 140,000 pounds maximum loaded weight; the bodies weighing from 93,000 to 96,000 pounds. The load-limit is pre- served of the standard 5J by 11-inch journal. With coil-springs over the journals, elliptical bolster-springs and space for lateral play of the bolsters, the equalizing-springs were discarded without affecting the easy riding of the cars ; though perhaps this might not be the case on a poorer track. Investigation proved that much of the jolting was due to the action of the unbalanced forces in the truck-frame when brakes are appUed, and this was remedied by anchoring the dead-lever to the body-underframe. A pair of four-wheel trucks weighs from 10,000 to 15,000 pounds less than a pair of six-wheel trucks of the same carrying capacity, which is a saving of eight to eleven per cent, in a car weighing 120,000 pounds. The saving in cost of maintenance has been found to be approximately in pro- portion to the reduction in the number of wheels and axles. In an ex- perimental test on the Pennsylvania Railroad, thirteen cars with six-wheel trucks offered as much resistance as fourteen with four-wheel trucks.^ • Mention should also be made of the Lightner journal-box, an American invention, in which the interposition of a plate under the top of the journal box and over the bearing, made it much easier to remove a worn-out bearing by slightly reheving the weight upon it. ^ ^ . , ^ , , ^"Fourrwheel Trucks for Passenger Cars," Roy V. Wnght, Journal Am. Soc. Mechanical Engineers, March, 1916., Cast-steel truck-frames are also used to some extent under freight-cars exceeding 60,000 pounds in capacity. 102 EFFICIENT RAILWAY OPERATION The introduction of the center-bearing truck in the construction of roUing-stock in the United States brought about a transfer of the axle- bearings from the underframe of the car-body to the truck-frame. In this change of structure, the stresses from the superincumbent weight were transmitted through the side-sills to the center-bearing by the interposi- tion of the body-bolster, which therefore became an important member of the underframe. The body-bolster should be sufficiently rigid to transmit these stresses without sharing any part of the superincumbent weight with the side-bearings. The side-bearings are intended simply to diminish the lateral swaying of the car-body in rounding a curve, or in passing over low joints, frogs or crossings. There is no tendency in the truck itself to re- sume its normal position after leaving a curve, except from the reaction of the flanges against the rail. If this reaction be neutralized by friction of the side-bearings, the flanges may grind against the rail long after the curve has been passed. Reduction in flange-wear reduces rail-wear, as well as train-resistance. It is therefore important that, when a car is at rest, the body-bolster should sustain the entire superincumbent weight, without assistance from the side-bearings. Drawgear and Couplers A description of the characteristics of American rolling-stock would be incomplete without some notice of two important accessories — the drawgear and the brakes. The drawgear has beeii developed from the links and hooks by which the tramway coal-wagons were connected in trains. Injury from shocks in train-service was obviated by the inter- vention of a pair of spring-buffers in compression at each end of the wagon. This simple arrangement has continued in British and Continental equip- ment to the present time. Neither the spring-buffers nor the coupling hooks were ever in use in the United States. A more complex drawgear was devised that replaced them both. The buffer-shocks were taken by a central cast-iron drawhead connected by stout rods with an interior cast- ing, containing a set of steel springs through which the strains of train- movement were transmitted to the underframe 'of the car. The cars were connected in trains by links held by pins through the drawheads. This loose coupling connection became increasingly objectionable as cars were built of greater capacity and weight. The free play of three or four inches in the links, rendered necessary for facility in coupling, produced successive shocks to each car in a train, as it was started, or else its motion was checked. As locomotives increased in tractive power, the trains were of greater length. In a' train of twenty cars, the locomotive would move some six or seven feet before the last car would be in motion, and the mo- mentum thus suddenly imparted to it resulted in severe shocks to the drawgear and the underframe. When the train was in motion and its speed was slackened, the forward car in the train received the full shock ROLLING-STOCK 103 as the slack was taken up. These alternating effects took place at each change of grade and were so severe on the couplings, pins and drawgear as well as^on the car bodies, that twenty freight-cars made about the practicable limit of safety with a fully loaded train. The link-and-pin drawgear was even more objectionable from the hazard in coupling cars so equipped. As early as 1869, railway managements in this country were seeking some device for diminishing this hazard. Dif- ference in the height of drawgear was so common a cause of injury in coup- ling, that a uniform height of drawheads was a condition necessarily jfre- cedent to the adoption of any couphng appliance operated by impact, and the yardmen did not care so much for self-couplers as for deadwoods over drawheads. By 1875, the managements had determined to experi- ment with couplers without links and pins, though no such appliance had as yet been devised that was applicable to freight-cars. Two devices had appeared for coupling passenger-cars which involved a novel principle, that of coupling by impact with hooks in a vertical plane; the Miller coupler and the Janney coupler.^ In 1885, the Master Car Builders Association tested forty-two self- couplers. Twelve of these were further tested in the next two years and, in these tests, the fact became established that link-and-pin couplers could not be efficiently used on freight-trains with power-brakes. In 1888, it was determined to experiment only with devices that coupled automatically in a vertical plane, and to establish a uniform contour-line for the coupling surfaces of all such devices. This conclusion was made of practical effect by the patentees of the Janney coupler, who generously waived their exclusive right to the use of such contour-lines j and all couplers of this character were thereafter designated as the "Master Car Builders' type." This appliance was further improved by an attachment that enabled the coupler to be operated by hand from the side of the car and which could be so set as to prevent the hooks from catching each other when it was not intended that a car should be coupled by backing against it. In 1893, this type was made compulsory in. the United States by the passage of the Railway Safety Appliance Act, to take effect January 1, 1898. The action of this coupler was rendered more eflBcient by the adoption of a standard height of drawbars for freight cars by the American Railway Association, fixed at 34|- inches with a maximum variation of three inches between empty and loaded cars.^ In 1893, forty-four per cent, of the casualties to train- men occurred in coupling cars; in 1909, they were reduced to seven. per cent. 1 Charles F. Hatch introduced Miller platforms and the air-brake on the Eastern Railroad of Massachusetts in 1872. "The Eastern Raiboad," F. B. C. Bradlee, p. 78. The Miller platform and hook-coupling are still in use on the Boston, Revere Beach & Lynn Railroad (narrow gauge). ^For further information on this subject, see "American Railway Manage- ment," p. 38, and "Problems in Railway Regulation," p. 318. 104 EFFICIENT RAILWAY OPERATION Further Improvements in Automatic . Couplers, End-Platforms AND Vestibules r The increased length and tonnage of freight-trains has correspondingly- increased the severity of the shocks and strains to which drawgear is ex- posed. Its component parts have been gradually improved in design and strength to meet these conditions, until it has become an expensive and comphcated accessory to freight-equipment. With closely coupled cars, it iS necessary that the resilience of the buffer-springs shall allow sufficient compression to enable the locomotive to make at least one revolution of its driving-wheels before it has to overcome the inertia of the entire train. Yet if these springs are too weak, the resulting impetus causes destructive shocks to the drawgear and unexpected train-partings that add coilsider- ably to the repair-account. The manner in which this troublesome problem was solved in this country and the general adoption of the solution through- out our railway system, are in striking contrast to the course pursued on European railways with reference to the same subject. It is an example of the value of government authorities' cooperating with railway manage- ments in furthering railway efficiency. From 1871, the same matter had been under consideration by British railway managements, in connection with the Board of Trade. By 1886, they had decided to use a "coupling-pole" with a hook at its end, to render it unnecessary to go between the wagons to hook up the couplings. Even after American managements had generally adopted couplers of the Master Car Builders type, their use was deemed impracticable on European rail- ways because of the universal use of separate spring-buffers. In 1899, a British Royal Commission reported that an automatic coupling-system was desirable and recommended experimental research. In 1908, at a com- petitive test conducted by the Italian Society of Railway Engineers, the American type was found to be too heavy and costly, and too difficult to adapt to European rolling-stock, and the first prize was awarded to a coupler of Italian origin. In a competition under the auspices of the French government, the first prize was awarded to a device of French origin. Up to June, 1914, the subject of an automatic coupler adaptable to European rolling-stock was still in an experimental stage. In couplers of the Janney type, the vertical hook opens and closes by rotation around a pivot within the drawhead ; a feature that distinguished it from the previous Miller type, in which the horizontal movement of the coupling-hooks was controlled by springs. The Janney type was also unique in that the slack between the opposing hooks was taken up by springs in compression. These couplers were first appUed to passenger- cars, and their additional weight caused the end-platforms to sag until by improved construction they were more firmly attached to the under- frame. , ROLLING-STOCK 105 A further improvement was effected in the vestibuled platform, first introduced on PuUman cars. The deck-roof was extended over the plat- forms, which were also inclosed, including the steps. By this means, per- sons passing from car to car were protected from exposure to the weather and from falling from the platform while the train was in motion. The interior of the cars was also protected from drafts and from dust through the opened doors. The open end of the vestibule was faced with broad steel plates that surrounded the passage-way and were connected with the end-frame by stout half-elliptic springs with a flexible covering, by which the faces of opposing vestibules were firmly compressed together. This device in itself constitutes a powerful buffer that distributes a collid- ing shock over a large surface, and the shock is thereby transmitted to the body-frame. In case of collision or derailment, it is now no longer pos- sible for the end-platforms to override and for the ends of the cars to telescope. The whole train reacts as an unit to such shocks, and the in- jurious effect of such accidents is accordingly lessened. This feature has been imitated on European carriages so far as to provide a covered way through the train, but, unassociated with couplers of the American type, it has not the same potential value in case of accident. Hand-brakes and Power-brakes The improvement of drawgear, as accessory to the speedier and safer connection or disconnection of cars, as well as in securing them more firmly while in train-service, has been accompanied by an even more remarkable development in brake-gear,' for arresting the motion of the cars, either singly or in trains.^ The brake to the horse-drawn vehicle was adapted to the four-wheeled coal-wagon by a lever at the side of the frame, acting on wooden brake-blocks applied to a single pair of wheels. Increased power was obtained by a wheel, mounted on a staff at the end of the wagon, which operated the brake-lever by winding a chain on the staff ; the pres- sure thus obtained being held by a ratchet. The application of this form of brake to the four-wheel truck required a double leverage in the inter- mediate rigging. Then this double leverage was duplicated to operate the brakes on both trucks from either end of the car and, at this stage, the development of the hand-brake virtually ceased ; except that the wooden 1 It is difficult for railroad men of the present generation to believe that train- service could be safely conducted without any braking apparatus; trains being checked or stopped solely by reversing the action of steam in the cylinders. Yet it is stated that, on the Newcastle & Frenchtown Railroad, in Delaware, "when the signal of an approaching train was heard, the slaves around the station would rush up to it and seizing hold of the cars, drag back on them with might and main, while the ticket agent stuck a stout stick through a wheel, and the whole train was thus gracefully brought to a standstiU." M. R. Pugh, Trans. Am. Society of Civil Engineers, December, 1911. 106 EFFICIENT RAILWAY OPERATION brake-blocks were first plated with wrought-iron for which cast-iron brake- shoes were afterward substituted. The hand-brake was not a train-brake; it was a car-brake. The speed of the train was arrested by the application of the brake to each car separately ; an operation that was repeated successively on the several cars by the brakemen. Its efficiency depended, therefore, upon the num- ber of brakemen on a train, upon their promptness in responding to the whistle-call for brakes and upon their muscular strength ; as well as upon the leverage power of the brake-rigging. When freight-trains got up to twenty cars, the drawgear then in use reached the limit of its inefficiency, and the same was true of the hand-brake. The greater momentum of the heavier trains could not be diminished with sufficient rapidity for efficient train-service. Economy in train-service was accordingly restricted by the normal efficiency of the hand-brake, and the necessity became ap- parent for the application of brakes by mechanical power. Several means were devised for accomj^lishing this purpose. In one device, known as the Creamer brake, the motive force was derived from a spring coiled on the brake-staff, like a watch-spring, which being released by connection with a rope stretched through the train, set all the brakes at once. But, with this device, the brakes could only be released by wind- ing up each spring separately ; therefore, it was of use only in bringing the train to a stop, and not in modifying its speed. Attention was meantime directed to the employment of atmospheric pressure in the vacuum-brake that originated in England. By a steam-jet on the locomotive, the air was exhausted from a line of piping through the train connected under each car with a bellows-hke device, the diaphragm of which was attached to the brake-rigging. This apparatus was in quite general use in Europe before it was introduced in this country under the name of the Eames brake. It was adopted on the New York Elevated Railroad, where it was not altogether satisfactory, as sufficient vacuum- pressure could not be maintained with stops at short intervals, nor did it act efficiently on trains exceeding a limited number of cars. A more efficient method of braking a train was devised by reversing this process and using compressed air. The power obtained from an air- pump, operated from the locomotive-boiler, was applied to the brake- rigging by the motion of a piston in a cyhnder under each car. This plan of the direct or "straight" air-brake was patented by George Westing- house April 13, 1869, and was tested ori the Pennsylvania Railroad in November of that year. Its merits were so apparent that it soon supplanted the vacuum-brake in this country. It had the advantage of maintaining a higher initial pressure than that of the atmosphere and of retaining it during an application of the brakes, without a continuous draft upon the boiler for steam. It arrested the motion of the train more rapidly and was more efficient in gradually modifying its speed. But when ROLLING-STOCK 107 the direct air-brake had been brought to this stage of efficiency, its useful- ness was terminated by a further development of the same principle, again due to the inventive genius of Westinghouse and patented by him January ' 29, 1873. Instead of operating the brakes directly by air-pressure, it was now used indirectly in the "indirect action" apparatus. The piston that actuated the brake-rigging was restrained by the constant presence of air in the reverse-end of the cylinder at a predetermined pressure. Whenever from any cause this pressure was slightly reduced, a connection was opened between the other end of the cylinder and an auxiliary storage reservoir under the car, which had been filled directly from the train-pipe with air at a higher normal pressure. The piston then acted on the brake-rigging ; provision being made for the escape of the air from the reverse-end. By this device, the braking-power was constantly energized and acted inde- pendently on every brake in the train, whenever the pressure in the train- pipe was reduced, either by the release of air in the locomotive-apparatus, or by a valve connected with a cord in each car, or by the train breaking apart. In this latter event, the rear portion of the train will be stopped automatically before it can collide with the forward part. Subsequently, an attachment was added in each passenger-car and in the freight caboose- car, which, by a similar release of air-pressure, operated a whistle-signal in the locomotive-cab, in place of the signal by a bell-cord stretched through the train. Air-brake Control of Long Trains The improvement in drawgear and in brakes was contemporaneous. Without the principle of coupling in compression, the application of mechan- ical energy to braking was impracticable. Until both of these principles had been brought into general use, the length of a freight-train was neces- sarily limited by the strength of the weakest link or pin that connected its constituent units together. The conditions which limited the number of cars in a train, relatively restricted the tractive power of the locomotive by which it was drawn. Indeed, the economies effected by the introduc- tion of locomotives of the articulated type would have been impracticable without the vertical-hook coupler and the quick-action air-brake. In providing for the greater strain brought upon the drawgear by power- brakes, the automatic coupler became more compUcated in construction and much more costly than the link-and-pin drawgear. By reason of the increased capital investment so required, the experi- mental stages of this experience were principally Confined to passenger- trains, which had been brought into general conformity as to equipment of couplers and power-brakes, and largely through the use of Pulhnan cars, before the resulting advantages had been made sufficiently manifest to warrant their application to freight-trains. This next step was hastened 108 EFFICIENT RAILWAY OPERATION by the pressure of diminishing freight-rates upon the cost of service. It was the narrowing margin between the ton-mile rate and the ton-mile cost that forced the construction of cars of greater capacity, their operation in longer trains and the introduction of loqpmotives of increased tractive power, which consequently placed the American managements in the van of railway progress. Yet neither of these elements of greater efficiency could have been so successfully utilized but for the contemporaneous intro- duction of the vertical-hook coupler and the air-brake, which have made it practicable to operate freight-trains of four times their former length, and indirectly have contributed to profitable increase in the tractive power of locomotives. Beyond the limit of twenty-car trains, the necessary control of speed was not to be obtained by hand-brakes ; but the application of power- brakes to freight-cars could only come into experimental use on refriger- ator-cars and stock-cars which were already equipped with vertical-hook couplers, and which, therefore, could be kept together in the train. Then a few more cars were so equipped and placed at the head of fast freight- trains. But this was found impracticable while the other cars in the train were still equipped with link-and-pin couplers. For as the power- brakes were applied in the forward part of the train, and the slack between the following cars was taken up by the arrested motion, the resulting shock told too severely upon the structure of the rolling-stock. Though the number of cars equipped with power-brakes gradually increased in number, they were of no more use than hand-brakes when merely dis- tributed discontinuously through a train. Only when it became com- pulsory to equip all freight-cars, could the number of cars in a train be increased to correspond with the maximum tractive power of the larger locomotives. The momentum acquired by a train of fifty cars, each loaded to fifty tons' capacity and headed by a heavy locomotive, is so much greater, proportionately, than that of a train of twenty cars moving at equal speed, that greatly increased efficiency was demanded of the air- brake apparatus in arresting its motion. The chief direction for improve- ment lay in securing quicker action through the train. Though it was a matter only of seconds, still, in reducing even that brief lapse of time, there was a field for obtaining better results.^ These facts were established in a series of tests between competing devices, including buffer-brakes as well as vacuum and compressed air- brakes, conducted, in 1886, at Burlington, Iowa, by a committee appointed 1 Tie Department of Safety of the Interstate Commerce Commission recom- mended the installation of air-gauges and emergency-valves in the caboose cars of long freight-trams. By means of the gauge, variations in the pipe-line pres- sure might be /observed and measures taken to insure the safety of the train if the pressure should fall too low, or if the train had been uncoupled for any purpose and then re-coupled ; or to indicate the effect of leaks in the pipe-line or of inefficient inspection at the terminals. The valve would also give control of the braking power of the train in an emergency. ROLLING-STOCK 109 by the Master Car Builders Association. In these trials, made with trains of twenty-five and of fifty empty freight-cars, it was found that the shocks caused by the application of the brakes to the forward car in a fifty-car train, some ten or twelve seconds before the brake-pressure was effective on the rear car, were too severe upon the drawgear and under- frames. In 1887, a similar committee conducted another series of tests at the same place on fifty-car trains with an improved Westinghouse ap- paratus, electrically controlled. In these trials, the lapse of time between the effective application of brakes on the first car and on the last car in the train was reduced from twelve to five seconds, and stops were made in two- thirds of the distance, as compared with the trials in the previous year.^ Notwithstanding these results, the committee still doubted the practica- bility of applying electrically-controlled brakes in general freight-service. Mr. Westinghouse then quickened the action of the mechanism without an electrical attachment and, by the end of the same year, fifty-car freight- trains were being satisfactorily operated with his improved apparatus. The efficiency of the "quick-acting" brakes on passenger-trains was being gradually affected adversely by the increase in weight and number of cars in a train and by the higher rate of speed. In 1894, the Westing- house Company introduced a "high-speed" attachment, by which the degree of pressure in the train-pipe, and in the auxiliary reservoirs under the cars, could be varied above the normal pressure at the discretion of the engineman. By this means, the effective pressure on the brake-shoes could be increased or diminished with the varying speed of the train and the condition of the rail-surface. The importance of this improvement had been made apparent in experiments conducted by Mr. Westinghouse and Sir Douglas Galton in England in 1878. The purpose of these ex- periments was to determine coefficients of friction between the brake-shoe, the wheel and the rail at different speeds. From these tests, a definite ratio was established for the application of varying degrees of brake- pressure under different conditions and without skidding. It has taken many years of experience to establish a satisfactory rela- tion in braking-operation between the locomotive, the individual cars and the assembled train. Brakes are now applied to every wheel in the train, except the locomotive truck-wheels.'' There is a more accurate adjust- ment of brake-leverage and greater thoroughness in design and construc- tion of the brake-rigging. Increased efficiency has been attained 'by making the apparatus more positive and responsive in application and release by electric control, by maintaining the brake-rigging in uniform condition, 1 In 1886, a flfty-ear train, at a speed of ^0.3 miles an hour, was stopped in a distance of 354 feet in 16 seconds ; and at 40 miles an hour, in 922 feet m 22^ seconas. In 1887, similar trains at 211 miles an hour were stopped m 160 feet m 7 seconas, and at 36 miles an hour, in 460 feet in 14 seconds. ,. , . j j ^ On French trains, the brakes on the driving-wheels are apphed mdepenaenuy by steam-pressure. no EFFICIENT RAILWAY OPERATION by a better method of applying a heavier brake-shoe, and by the use of the clasp-brake with two flanged shoes to each wheel.^ The grip of the brake- shoe has been improved by bits of cast-iron inserted in soft steel, and by layers of steel in diamond-formed meshes placed in the molds in which the cast-iron is poured. The ordinary style of electro-pneumatic brake has been in use on electric lines for over eight years under several thousand cars, and, on the New York Subway, has given good results under the most severe service in the world. Its efficiency was exhibited in tests made on the Pennsyl- vania Railroad in 1913. The experiments were made with a train of twelve steel passenger-cars with four-wheel trucks, weighing nearly 1000 tons. At a speed of sixty miles an hour, its kinetic energy was estimated at 224,000,000 foot-pounds, equivalent to 70,000 nominal horse-power. At this speed, with the ordinary "high-speed" brake and under usual working conditions, such a train could be stopped within 2000 to 2200 feet. On this trial, with the same apparatus and at the same speed, the train was stopped within 1600 to 1800 feet and, with improved appliances, in a distance of 1000 feet, or within its own length ! It is of interest to note how this was brought about. Rating the usual working pressure at 100 per cent., or the weight of the empty car, the emergency pressure is available to 125, 150 or 180 per cent., as may be desired. With 150 per cent, of working pressure, the stop from 60 miles an hour was shortened from 1600 to 1400 feet. With the electro- pneumatic brakes, the distance was reduced to less than 1200 feet and, in one trial with flanged brake-shoes, to 1049 feet. The use of flanged brake-shoes was estimated to shorten the stops approximately 12 per cent., and the relative proportion of area bearing upon the wheel to the total area of the face of the brake-shoe was found to exert an important influence upon brake-performance. Maximttm Brake Energy. Safety Appliances The demand for increased efficiency in brake-apparatus was illustrated by Mr. W. B. Turner, Chief Engineer of the Westinghouse Brake Com- pany, in the discussion upon the paper by Mr. Dudley." In 1890, with a 1 Clasp-brakes, Master Car Builders Association Rules, 1915. All passenger ears with four-wheel trucks, weighing 96,000 lb., and all with six-wheel trucks, weighing 196,000 lb., should be equipped with clasp-brakes. New York, Ontario & Western Railway Hot Box Delays. Comparison of clasp-brakes and single-brakes on four-wheel trucks. In thirty months, cars with clasp-brakes made 1,873,500 miles with 10 hot boxes, an average of 187,350 miles per box-trouble. The remaining equipment, with single-brakes, made 11,407,200 miles with 88 hot boxes, an average of 130,760 miles. 2 See "Brake Performance on Modern Steam Railroad Passenger Trains," S. W. Dudley, Journal Am. Soc. Mechanical Engineers, November, 1914, p. 373. ROLLING-STOCK 111 train-weight of 280 tons at 60 miles an hour, the energy to be dissipated amounted to about 33,000 foot-tons and the stoppage-distance was 1000 feet. In 1913, with a train-weight of 920 tons at the same speed, the energy was 111,000 foot-tons, or ahnost four times greater. With the old- style air-brakes, the collision energy of such a train, at the point where the first train would have been stopped, would still be 48,000 foot-tons, and there would yet be about 460 feet to run. But the modern train used in the Pennsylvania Railroad tests, equipped with the new apparatus, can be stopped in 860 feet, at which point with the old equipment it would still have a speed of 43 miles an hour and a collision energy of 57,000 foot-tons, or about twice the enei^y of the train of 1890 at the beginning of the stop. The stoppage of a heavy train at high speed within a brief space of time has been successfully accomplished by combining its component vehicles into a consolidated mass, thus ehminating their relative motions. The problem now to be reckoned with is the continuous adjustment of brake- pressure to varying conditions of car-weights, train-speed and rail-surface. As the matter was put by Mr. N. A. Campbell, of the New York Air Brake Company, in the discussion on Mr. Dudley's paper, it is the adjustment of resistance to rotation and rail-adhesion. Further improvement in brake- apparatus is being directed to this problem, but when comparative tests are reduced to distances of less than a hundred feet and to periods of less than a second, theoretical efficiency would seem to have been approached within 90 per cent. The energy dissipated in the conversion of motion into heat, in the action of the air-brake, has been illustrated by the state- ment that, while the acceleration of a heavy train from a state of rest to a speed of 60 miles an hour cannot be accomplished in less than six minutes, the same train at that speed can be brought back to a state of rest by the improved brake-system in one-eighteenth of that time, or in twenty sec- onds ! The air-brake, in connection with the automatic coupler, has greatly facilitated train-service; yet the importance of these inventions in this respect, however highly estimated, is not to be compared with their value as an element of social efficiency. As applied to cars of steel construction, they have minimized to a great extent the effect of collisions or derailments upon the occupants of trains so equipped. They have also contributed largely to the welfare of trainmen by reheving them from liability to in- jury occurring in the use of handbrakes and liak-and-pin couplers, espe- cially in handUng freight-cars. Protection from such injuries has been further secured by attention to minor details in car-construction, such as the rigging for manipulating couplers without going between the cars, and also running-boards, ladders and grab-irons (or handholds) for access to car-roofs. The character and position oif these appliances has been made the subject of regulation in the Safety Appliances Act of 1893, as amended April 1, 1896, and March 112 EFFICIENT RAILWAY OPERATION 2, 1903.1 Improvement in other details might tend still further to miti- gate the consequences of railroad accidents as, for instance, by more thoroughly interlocking the couplers, to prevent the separation of cars in a train, and by firmly keying the center-pin to the body-bolster instead of relying solely upon safety-chains, to prevent the dragging of a car-body off the trucks. Car Wheels. Chilled Iron; Steel, Forged and Cast The mechanical and economic advantages derived from the introduction of converted steel into the useful arts, affected roadway and locomotive construction and design before its influence was apparent in car-construc- tion. The front pair of wheels on the colliery tram-wagons was of cast- iron as early as 1750, though the back pair was still of wood to give a better grip to the wooden pole or "convoy" that was used as a brake until the wooden rails were plated with iron.^ The type of wrought-iron wheels that subsequently prevailed was derived from the ordinary spoke-wheel with iron tire, and it long continued in general use on European railways until replaced by steel-tired wheels. The independent character of the evolution of a railroad system in the United States is seen in this as well as in other details of rolling-stock construction. The production of a fine quality of soft gray iron in the Eastern States and the lack of mechanics skilled in making wrought-iron wheels, resulted in the use of cast-iron wheels with chilled treads, as a cheaper substitute. These wheels were at first of single plate but were afterward cast in double plate, .with brackets to support the flanged rim. Such skill was attained in the manufacture of these wheels, in the mixture of iron of different qualities, in the heat treatment, in chilling and in the molding and casting, that no serious attempt was made to use wrought-iron for this purpose except for driving-wheel tires.* Whether for lack of the proper grade of iron or of workmen skilled in their manufacture, cast-iron wheels have not been favorably con- sidered by European railway managers. Yet in this country, they are in successful use under heavier loads and at speeds equally as high as are customary elsewhere. In 1886, cars of 50,000 pounds' capacity, weighing 31,000 pounds, were replacing cars of 20,000 pounds' weight and capacity, and the weight of the 33-inch wheel was gradually increased from 450 to 500, 600, 700 and 750 pounds, but without any theoretical basis for calculating the stress in ' By order of the Interstate Commerce Commission, March 13, J.917, these requirements were made effective July 1, 1917. Up to May 1/3, 1916, out of a total freight-equipment of 2,510,214 cars, 1,821,086 or 73 per cent, had been so equipped. * " Internal Transport and Communication in England," Edwin A Pratt 1912. ' The depth of the chiUed or white metal should not exceed one inch or be less than a half-inch at the middle of the tread. ROLLING-STOCK 113 cooling or the shocks from use. About 1909, the Master Car Builders standard for the 33-inch wheel was modified to conform to changing con- ditions, but the introduction of cars of 60,000, 80,000 and 100,000 pounds' capacity was putting them to a still severer test. To meet these require- ments, the wheel-weight was increased to 840 pounds under freight-cars and to 950 pounds under tenders. Long-continued brake-application occasions abnormal strains in the wheels from heating. On a car weighing 40,000 pounds and loaded to 100,000 pounds, a brake-pressure of 3500 pounds per wheel, at ten miles an hour on a long grade, generates heat that must be dissipated through the brake-shoe and the wheel-rim. There is also a lateral thrust that may reach 40,000 pounds to which the wheel-flange may offer a shearing resist- ance ranging from 40,000 to 125,000 pounds. Such a lateral thrust when the wheel-rim is heated by long-continued brake-action, may result in a broken flange. The effect of a heavily loaded car upon track out of order, or when striking guard-rails and crossings at high speed, likewise causes shocks that defy calculation. Taking all .these couditions into consider- ation, it is surprising that a brittle metal like cast-iron can be adapted, by skill in metallurgical treatment and in mechanical design, to offer successful resistance to the strains and shocks to which a wheel is now submitted in railroad operation. Yet chilled wheels are made that are capable of satisfactory service under heavy freight-equipment ; they are still liable to broken flanges, although the flanges have been thickened as much as the frog and switch clearances will permit. Sharp, flanges cannot be remedied, although flat spots may be ground out of the tread, as may be done with steel wheels by turning. Cracked plates are principally due to the heating action of brake-shoes, which also causes broken flanges from circumferential cracks in the tread.' The first use of steel in wheel-making was in the substitution of crucible steel for Lowmoor iron in driving-wheel tires. The purpose was to obtain greater resistance to wear and to tensile strains, but it was too expensive a material for car-wheels. Converted steel, being a cheaper material, replaced crucible steel for driving-wheel tires, and steel tires on cast-iron centers were used in locomotive-trucks and tender-trucks. They were gradually introduced under Pullman cars and other high-class passenger- equipment. The solid pressed-steel wheel then followed, doing away with 1 Derailments from wheel failures in eleven years, from 1902 to 1913. Inter- state Commerce Commission Reports. Broken or burst wheels a'rok Broken flanges 6,625 Loose wheels \'\ia Miscellaneous defects ^'^^" Total 11.753 Out of 33 270 derailments in this period, the ratio caused from defective wheels was 3.85 to 1, as compared with broken raUs, and 2.39 from broken flanges alone. 114 EFFICIENT RAILWAY OPERATION retaining-rings and bolts, and the operations for removing and applying the tires, by which the first cost and upkeep were so much reduced that for many years solid steel wheels have been used in place of steel-tired wheels ; though their cost in proportion to their wearing qualities prevented their general use under freight-equipment. In 1903, Mr. C. T. Schoen established the manufacture of sohd forged and rolled steel wheels on a commercial basis. Up to 1908, this wheel was in use on 9000 locomotive-trucks, on 32,000 passenger-cars and on 165,- 000 freight-cars and tenders ; a total of 206,000. In 1907, Mr. J. M. Hansen developed a process for forging steel wheels by hydraulic pressure alone and, in April, 1908, the United States Steel Corporation began the manufacture of wheels of this character. The steel wheel is a safeguard against broken flanges. The flange of a chilled wheel can be broken with a pressure of 60,000 to 100,000 pounds, while a pressure of 526,000 pounds was required to break off the flange of a steel wheel. The Master Car Builders' standard axle has an ample factor of safety with four axles loaded to 150,000 pounds, which is the limit for chilled wheels. With 10 per cent, excess loading, which is permitted, the hopper-car weighing 39,000 pounds with steel wheels, can be loaded to 121,000 pounds and still be within the permissible load on the standard axle. The 33-inch wheel of rolled steel weighs about 800 pounds and may make 240,000 miles with three turnings against a life of about 80,000 miles for chilled wheels. The American Steel Foundries corporation has undertaken to meet the requirements of heavy freight-service at a reasonable cost with the "one-wear" cast-steel wheel. This wheel is cast in a mold on a rapidly revolving table. With the first flow of metal, an alloy of manganese is introduced and thrown into the surface of the tread to harden it. The wheel is next submitted sto an annealing process to relieve internal stresses from cooling, and is then ground to the standard contour of tread and flange and made perfectly round. The finished wheels are each submitted to a drop-test of 500 pounds from a height of 6|- feet to detect imperfectio^is, and are then paired, as to circumference, by tape measurement, in order to diminish the tendency to sharp flanges and ' ' slid-spots, ' ' caused from pairing wheels of unequal sizes upon the same axle.' 1 Weight of wheels and length of guarantee against defects. } Weight, Pounds Guaranteed ChUled wheels 30-ton car Chilled wheels 40-ton car Chilled wheels 50-ton car Rolled or forged steel . . "One-wear" cast-steel . . 625-690 675-700 725-800 750-800 600 Guarantee for steel wheels same as for chilled wheels. 6 years 5 years 4 years ROLLING-STOCK 115 Wheels are removed when slid-spots exceed 2^ inches. These flat spots in rolled or forged steel wheels are removed by turning ; those in wheels with chilled or hardened tread, by grinding. A process has been de- vised for filling these spots in cast-steel wheels by electric welding and re- storing the contour of the tread by a portable grinding wheel. Another defect exhibited in chilled wheels is known as "shell-outs," also caused by the heating action of the brakes, by which the surface of the tread crumbles in spots. The hammering effect at high speed of such defects in the tread has been experimentally determined at from twenty to sixty per cent, of the static load.* When a steel wheel is removed on a foreign road and replaced by a chilled wheel, the Master Car Builders' Rules .require that the owner of the car shall be paid the difference in value, if the responsibility for the removal rests with the foreign road. The charge for removing, turning and re- placing steel wheels is $3.25 per pair, and, on an average, every wheel is turned twice. Consideration should be given to the saving effected by doing away with the turning ; for the investment in shop-room and machine- tools for this purpose and the accompanying labor-charge and cost of maintenance, is evidently very large. The use of steel wheels under freight-equipment is confined principally to steel hopper-cars of heavy capacity. It was vaguely estimated that in 1914 about 3,000,000 steel wheels were in use out of about 20,000,000 wheels under aU classes of train-equipment. On the Louis viUe & Nashville Railroad there are 83,000 "one-wear" steel wheels in use. Steel-tired wheels are still used under passenger-equipment and heavy tenders. On the Delaware, Lackawanna & Western Railroad, 15,000 solid steel wheels, 33 inches in diameter and weighing 815 pounds, had been used under steel hopper-cars up to September, 1915. A 36-inch solid steel wheel, weighing from 875 to 900 pounds, is used under new passenger-equipment and under all tenders. In a comparative test of rolled-steel wheels and steel-tired wheels, 36 of each kind, made on the Atchison, Topeka & Santa F6 Rail- road System in passenger-train service, there was an average mileage be- tween turnings of 105,000 miles for the rolled wheels, and of 111,000 miles for the steel-tired wheels. The rolled wheels add about 4000 pounds more weight to a seventy-foot passenger-car as compared with steel-tired wheels.^ The cost per mile run of the solid steel wheel and of the chilled wheel is 1 A flat wheel with a spot three inches long, under a car weighing, loaded, 50,000 pounds, strikes a hammer-blow of 104,000 pounds at a speed of sixteen miles an hour. Railway Mechanical Engineer, February 16, 1916, p. 60. ^ Much of this information has been obtained from the managements of the railroads mentioned; also from an article on "Car Wheels" by George L. Fowler, in Cassier's Magazine for March, 1910; from a report on the "Use of Steel in the Construction of Locomotives and Rolling-stock" by D. F. Crawford, General Superintendent of Motive Power, Pennsylvania Lines West of Pittsburgh ; from Proceedings of International Railway Congress, Berne, 1910; and from Amer- ican Steel Foundries, April, 1916. 116 EFFICIENT RAILWAY OPERATION about equal, allowance being made for interest on the increased invest- ment in the steel wheel; but the general use of the latter would add about $315,000,000 to the investment in car-wheels on our entire system. Steel Constetjction fob Car Trucks and Frames The substitution of iron for wood in truck-construction was first ap- plied to the diamond-truss truck-sides for freight-service ; but under cars of very large capacity, that design is now being superseded by the solid steel truck-side, in which the side-rail and pedestals are cast in a single solid piece. Wooden side-rails are still continued in service under passenger-cars, being strengthened for the six-wheel trucks by iron plates. From 1895, when, converted steel came into general use, it replaced iron for all car-construction, including axles, excepting journal-boxes and other details, for which cast-iron is equally serviceable and cheaper.' Much ingenuity has been displayed in the construction of car-springs. Beside the varieties of elliptic and semi-elliptic and coil springs made of steel, rubber has also been largely used to diminish the shock between the journal-boxes and the side-rail of the truck. Attempts have also been made to utilize the elasticity of air in compression, but steel still pre- dominates as the most suitable material. The use of steel in freight-car construction has greatly contributed to economic efiiciency in railroad operation by increasing the possible pro- portion of tonnage to empty weight. This matter was made a subject of general interest by promoters of narrow-gauge projects. They empha- sized the fact that the standard-gauge freight-cars weighed twice as much empty as their possible tonnage-capacity, while, in the narrow-gauge equipment, the two were equalized. As a consequence, advocates of the standard-gauge strove to neutralize this argument by improving their equipment in this respect. The ordinary freight-car, at first, like, other features of railroad construction, was a development of the colliery-tramway. The under- frame was similar to that of the old coal-wagon ; a rectangular framework, with some perfunctory bracing, that carried the pedestals in which the axle journal-boxes had a little vertical play. In the United States, this underframe was adapted to the center-bearing truck by the intervention of a cross-timber or body-bolster, to which the truck was pivoted by a center-pin or king-bolt. The strains from traction were transmitted from the drawgear to the trucks indirectly through this underframe, and in a manner that racked it at every joint in its structure. As the demand increased for heavier loads and longer trains, this tendency to distortion 1 An important step toward simplicity in construction has been introduced by the Commonwealth Steel Company in making the whole freight-ear truck- frame of a single casting, thereby eUminating bolts, nuts, etc. ROLLING-STOCK 117 of the underframe was resisted empirically by increasing the dimensions of its members and by strengthening its joints and bracing. The first really efficient improvement in the underframe was the intro- duction of central longitudinal sills in pairs, reinforced against tensile strains by iron rods connecting the two drawgears at the ends of the car, by which the shocks and strains of train-service were transmitted directly through the train, without otherwise affecting the structure of the under- frame. As automatic couplers came into general use, their increased weight was transferred from the end-sills to these center-sills. The body- bolsters were also framed into these sills, which thus became the backbone of the car and the spinal column of the train. To this foundation, strength- ened by the floor-joists and the intermediate braces, the car-body was framed and, though many differences in detail were developed, the under- frame remained unchanged in general design until the substitution of steel for timber in car-construction. This change was initiated when the loading-weight had reached 60,000 pounds. This load, concentrated at the center-bearings of the trucks, caused the body-bolsters to droop, and they were accordingly stiffened by flitch-plates between layers of plank in the composite bolster. The sub- stitution of an all-steel bolster permitted an increase of loading to 80,000 pounds. Even in cars 36 feet in length, the side-sills were not sufficiently rigid to sustain this weight without undue strains upon the underframe, which was strengthened by six truss-rods extending from bolster to bolster. As it was not practicable to further increase the load, which was de- sirable in mineral traffic, attention was then directed to the design of an all-steel underframe, which was introduced in 1897 on the Bessemer & Lake Erie Railroad for cars of 100,000 pounds' capacity and 46 feet in length; though it did not come into general use until 1901. This radical change in construction was not important in cars for general traffic, which could not be uniformly loaded to such a capacity as to warrant the addi- tional weight. Still, by degrees, steel was used for underframes, upper- frames and for roofing; and an all-steel hopper-car of 80,000 pounds' capacity, weighing about 42,000 pounds, was introduced on the Pitts- burgh & Lake Erie Railroad in June, 1897. The tendency at present is for all freight-cars to have steel underframes; hopper-cars and hopper- bottom gondolas of 110,000 pounds' capacity all-steel, and lighter gondolas with steel hoppers. It has also been found necessary to build caboose-cars with steel under- frames, to prevent them from being crushed where heavy "pushers" are used. Similar damage to the ends of wooden box cars from the severe strains to which they are subjected has induced the introduction of ends of corrugated steel for cars of large capacity. The use in Europe of goods-wagons, built wholly of iron, dates back to 1866, but, to any noticeable extent, only to 1880. Even in 1910, they were 118 EFFICIENT RAILWAY OPERATION only in experimental use in France, Great Britain and Russia. These open goods-wagons are usually mounted on four wheels for a load of six tons, with six wheels for twenty tons and with eight wheels for greater loads. The coal-wagons have eight wheels and carry from 27 to 40 tons. In fact, only a small percentage of freight-equipment in Europe is either entirely of iron or of iron with wooden floors. Cars of this construction are objected to because the repairs are more difficult, take more time, and are more costly. They have to be frequently coated with red-lead and tar as protection against rust, and, as covered wagons, they do not suffi- ciently protect the contents from temperature variations. On account of the scarcity of timber, the climatic conditions and other features of "a tropical environment, freight-equipment of steel construction is more gen- erally in use in India, South Africa and AustraHa. At the end of 1908, out of 94,850 goods-wagons reported as of "iron or with iron underframes," 29,505 were in use in Europe, principally in the German and Austrian em- pires ; and 65,345 in India, South Africa and Australia. Up to 1910, no passenger-cars, except for underground-railways, had been constructed wholly of steel, except in America. Fkeight-car Capacity in the United States. All-steel Freight Cars Increase in capacity of freight-cars in the United States dates from 1855 to 1860, when eight-wheel cars of 10,000 pounds' capacity were substituted for four-wheel cars of 7000 pounds. By 1865, there were box cars of 15,000 pounds and coal-cars of 20,000 pounds. In 1873 the standard capacity of the several classes of freight-cars was as follows : Light Capacity Capacity to Total Weight Box .... Stock . . . Gondola . . Coal . . . Coal (4-wheel) Coal (4-wheel) Coal Hopper . Pounds 18,700 20,295 17,280 17,350 7,635 7,635 18,750 Pounds 28,000 20,000 28,000 30,000 10,000 12,000 37,000 Per Cent. 59.94 49.62 61.81 63.36 56.82 61.22 66.37 The notoriety given by promoters to the relative advantage of narrow- gauge equipment as to the ratio of paying load to lighter weightj gave an impetus to the further development of this subject on standard-gauge roads, so that, by 1876, 40,000 pounds was the capacity-standard, 50,000 pounds in 1883, and 60,000 pounds in 1885, for all classes. In 1895, the Pennsylvania Lines West of Pittsburgh built 70,000-pound cars for coal ROLLING-STOCK 119 and ore traffic.' From 1890 to 1908 the average increase in weight and capacity of the entire freight-equipment of Pennsylvania Lines West, increased as follows : Average weight 45 per cent. Average loading, loaded cars 70 per cent. Average loading, all cars 50 per cent. At present the wooden box car of 40,000 pounds' capacity weighs about 22,000 pounds ; that of 60,000 pounds weighs 30,000 pounds ; that of 80,000 pounds, 34,000 pounds ; and that of 100,000 pounds, 38,000 pounds. The proportion of empty weight to loading capacity is, therefore, respectively 55, 50, 42.5, and 38 per cent.^ Cars have also been introduced of 110,000 pounds' capacity, and, experi- mentally, of 120,000, 130,000, 140,000 and 150,000 pounds. In 1911, the Pennsylvania Railroad Company built hoppers and gondolas 52 feet in length to carry 160,000 pounds ; and, in 1913, a gondola car was built for the Norfolk & Western Railway Company to carry 200,000 pounds on six- wheel trucks, with length over all of 46 feet, 10 inches, outside width of 10 feet, 4 inches, and total weight of 65,200 pounds, or 32 per cent, of its capac- ity.^ 'Increase in Cab and Load Weights, 1876-1898 Percentage of Total Weight Light Weight Paying Load Total Car Load 1876 20,500 20,000 40,500 53.62 42.38 1882 24,000 40,000 64,000 37.50 62.50 1889 27,700 60,000 87,700 31.59 68.41 1895 36,000 80,000 116,000 31.04 68.96 1898 38,500 100,000 138,500 27.80 72.20 10 per cent, excess 38,500 110,000 148,500 25.93 74.07 " Economical Size and Capacity of Freight Cars." L. F. Loree. Inter- national Railway Congress. Paris Meeting, 1900, XVIII, p. 63. ^ Estimated Cost and Weight op Steel-underframb Box Car. American Railway Association Standard Capacity Weight Cost Ratio of Allowable Loading to Weight Pounds Pounda 60,000 80,000 100,000 38,200 41,400 43,000 $1,025 1,100 1,145 1.7 2.1 2.6 Ten per cent, overloading allowed. Byers, " Economics of Railway Opera- tion," 1907, p. 647. ' In 1915, the Norfolk & Western Railway Company placed in service 750 steel gondolas of 90 tons' capacity and has ordered 1000 more. With permissible excess-loading of 10 per cent., the paying load is 75 per cent, of total weight. Rail- way Age Gazette, March 31, 1916. 120 EFFICIENT RAILWAY OPERATION In 1910, out of 243,809 freight-cars, the Pennsylvania Railroad Com- pany had 74,466 all-steel, 53,662 with steel underframes, and 115,681 of all-wood construction. The all-steel cars were gondola, hopper, coke and flat cars. The steel underframes were box cars. In twelve years, with an average of 11,000 of such cars in service, 18 all-steel and 126 steel underframes had been destroyed, against about 20,000 of wooden con- struction. During this period, the larger car-building companies had built about 311,000 all-steel cars, of which 300,000 were gondolas, hopper and dump cars for coal, coke and ore;. and 241,000 with steel underframes, of which 75,000 were also gondolas and hopper cars, and 125,000 box cars. Out of 280,000 freight-cars built in the United States and Canada in 1907, 72 per cent, were of some form of steel construction. Out of 133, 11 7 freight- cars ordered in 1911, 28,418 were all-wood, 54,605 had steel underframes and 50,094 were all-steel. In 1914, one-half of the total freight-equipment was either all-steel or with steel underframes. Eighty per cent, of the all- steel cars were open coal and ore cars.^ Comparative Efficiency of Wooden and Steel Frame Box Cars A comparison of the economic efficiency of standard wooden box cars of 60,000 pounds' capacity with steel-underframe cars of 100,000 pounds' capacity appeared in the Railroad Gazette in November, 1897. A train of thirty steel-underframe cars with empty weight of 510 tons would carry a paying load of 1500 tons. A train of wooden cars of equal gross weight 1 At the end of 1915, the Pennsylvania Railroad Company had 6500 steel box cars, representing an investment of $9,000,000. These cars are of 100,000 pounds' capacity and weigh 50,000 pounds with complete equipment. All-steel Cars Ordered up to JANtrARY 1, 1916 Description Number Per Cent, op Total Hopper and gondolas Box and house Tank Flat Miscellaneous Total Also 672,171 with steel underframes. 431,132 30,562 26,306 24,047 17,953 81.4 6.7 5.0 4.5 3.4 530,000 Addition to Steel Freight Equipment, N. Y. C. & H. R. R.R. Co. 1911 1912 1913 Total All-steel . . . Steel underframes 4,600 11,836 7,329 18,754 8,445 29,627 20,374 60,217 Total . . . 16,436 26,083 38,072 80,691 ROLLING-STOCK 121 (2010 tons) would weigh, empty, 703 tons with paying load of 1307 tons, or 193 tons less of paying load. The saving in length of train would be more than 500 feet, with proportionately less resistance in rolling-friction and journal-friction, and a saving in repairs on account of fewer wearing sur- faces and breakable parts. The proportion of paying load to empty weight could be much increased in cars with steel underframes, were it practicable to increase their length ; but that cannot be done without a correlative change of spacing in the arrangement of doors in freight warehouses, coal- pockets, grain-elevators and other terminal facilities, which could only be made at an enormous cost. In another comparison of cars of similar capacity and construction, the life of a wooden car costing $525 was estimated at fifteen years and that of a steel car costing $810 at thirty years ; the cost of repairs to a wooden car at $40 per annum and of a steel car at $20. With interest at five per cent, the saving in the steel car, per ton-capacity per annum, was stated at 42 per cent., or $1.04 per ton. On a fifty-ton car, this amounts to $82 per annum, and on 500 cars in thirty years to $1,230,000. This estimate does not include the saving in cost of train-service due to difference in empty weight and carrying-capacity.^ ■ From more recent experience, it appears that acids in coal and coke corrode the floor-sheets and sides of steel cars quite rapidly so that, during sixteen years, the repairs on the bodies prac- tically amount to rebuilding them; and that it would perhaps be more economical to scrap the body after twelve years' service. The total cost of freight-car repairs increased from $61 in 1908 to $80 in 1914.'' Statistics of Freight Cars in the United States In 1905, the total number of freight-cars on our railway system was 1,727,620. In 1910, there were 2,133,531. Increase, 405,911. In 1914, there were 2,325,647. Increase from 1910, 192,116.' The increase from 1905 to 1910 was at the annual rate of 81,182 cars. From 1910 to 1914, the annual rate was 48,029 cars. In 1905, the total tonnage-capacity of our freight-equipment was 53,036,495 tons; and in 1910, 76,450,660 tons. Increase, 23,414,165 ' American Engineer, Car Builder and Railroad Journal, March, 1898. * On the Bessemer & Lake Erie Raikoad, where steel cars had been in use since 1897, the average cost of maintenance as compared with wooden cars was as follows : Yeab Wooden Cars All-steel Cars 1905 1906 1907 1908 $71.14 94.11 118.50 53.06 $27.83 30.74 39.21 32.11 ' See Appendix III, Table II. 122 EFFICIENT RAILWAY OPERATION tons. In 1914, the tonnage-capacity was 90,848,630 tons. Increase from 1910, 14,397,970 tons. The annual rate of increase from 1905 to 1910 was 4,682,833 tons, and from 1910 to 1914, 3,599,492 tons. From the diminution in the annual rate of increase in both the number of cars and in the tonnage-capacity, it would seem that the freight-equip- ment of our railway system was gaining upon traffic-requirements.^ In 1905, box cars constituted 47 per cent, of the freight equipment, and coal-cars, 37 per cent. In 1914, about 44 per cent, was in box cars and 39 per cent, in coal-cars, which is an indication of the growing importance of the mineral traffic. The number of flat cars remained unchanged in nine years ; the number of stock-cars increased one-half, and the number of tank and refrigerator cars has about doubled, as has also the number of cars classed as "other cars." In 1905, 54 per cent, of the total equipment was in cars of 25 to 30 tons' capacity and 26 per cent, of greater capacity. In 1914, 60 per cent, was above 30 tons and 27 per cent, was of 50 tons. In that year, 59 per cent, of the coal-cars and 8 per cent, of the box cars were of 50-ton capacity and over. The stock and refrigerator cars were principally of 30 tons' capacity and the tank-cars of 40 to 50 tons. Thirty- six per cent, of the flat cars were of 30 tons, 39 per cent, of 40 tons and 16 per cent, of 50 tons. This statement excludes cars in railroad use only (124,719 in 1914), and cars belonging to private concerns. By far the largest proportion of the rolling-stock of our entire system is owned by comparatively few companies. In 1914, the proportion as to freight-equipment was as follows : DiSTKICTS COBPOBATIONB No. OP Cars Capacity Eastern Southern Western Total 16 7 15 38 33.9 per cent. 12.6 per cent. 25.6 per cent. 72.1 per cent. 35.5 per cent. 13.2 per cent. 24.2 per cent. 72.9 per cent.^ The relative ownership of passenger-equipment of the same companies was as follows : DiSTBIOTS Corporations No. OF Cabs Eastern Southern 16 7 15 38 37.7 per cent. 9.0 per cent. 27.6 per cent. 74.3 per cent.' Western Total 1 The progressive increase in tonnage-capacity from 1905 to 1914 of the several classes of freight-equipment is given in detail in Appendix III, Tables III and IV. 2 Appendix III, Table V. 3 Appendix III, Table VII. ROLLING-STOCK 123 Steel Passenger-cabs Steel was not introduced into passenger-car construction to provide for greater capacity in proportion to empty weight, as in freight-car con- struction ; but as a protection against fire, collisions and derailments. It was not so much a measure of economic, as of social, efl&ciency. It is true that iron was being gradually employed as tension-rods, in reinforcing wooden underframes and, in the form of plates, for sheathing the side- frames; but steel, as an important element in construction, was first brought into use for underframes in 1903, on the Chicago suburban lines of the Illinois Central Railroad Company, and at the suggestion of George Westinghouse. In 1904, twenty all-steel cars, designed by George Gibbs, were put in service on the New York Subway lines. In 1905, the Long Island Railroad Company equipped its electric lines with 134 all-steel passenger-cars, and all-steel motor-cars in considerable numbers were in service on the electrically operated lines in New York, Boston and Phil- adelphia. These cars were intended only for urban and suburban traflBc, though occasionally a baggage-car or a postal car of all-steel had been ex- perimentally placed in ordinary railway service. The first substantial adaptation of all-steel construction for regular passenger-train equipment, was in connection with the operation of the Hudson River terminals of the Pennsylvania Railroad Company. Up to February, 1908, that company had built 686 all-steel cars, of which 392 were day-cars and 75 combination passenger-and-baggage cars. By 1910, there were 3117 all-steel cars in regular railway service.^ In 1911, of 4075 cars ordered, 62 per cent, were all-steel and 14 per cent, were built with steel underframes. Of cars in use in December, 1910 and 1911, respectively, all-wood cars were 98.2 per cent, and 87.2 per cent. ; steel-underframe cars, 1.0 per cent, and 3.5 per cent. ; and all-steel cars, 0.8 per cent, and 9.3 per cent. There was an increase of 6642 all-steel cars from January, 1909, to January, 1913. Between January 1 and July 1, 1913, orders were placed for 1140 passenger-train cars, of which 1064 were to be all-steel and the remaining 76 were to have steel underframes. The construction of all- wood passenger-cars has now practically ceased.^ ' See Appendix III, Table VIII. * Additions to Steel Passenger Equipment, N. Y. C. & H. R. R.R. 1911 1912 1913 TOTAI, All steel Steel underframes . . Total . . 297 135 432 288 173 461 521 300 821 1106 608 1714 124 EFFICIENT RAILWAY OPERATION The transition from wooden to all-steel construction of passenger-train equipment was too great an undertaking to be generally attempted. The first experimental all-steel cars were so heavy and costly that efforts were made to combine the respective strength and Hghtness of steel and wood in a composite car, in which the frames and the side-sheathijig below the window-sills were of steel; wood being still used for the upper siding, flooring, roofing, inside-Hning, doors and window-frames. ' In some of these composite designs, the two materials were so judiciously combined that the resulting weight was only 2.5 per cent, greater than in the standard car of wooden construction of equal seating-capacity. Nevertheless, the growing opposition of th.e Public Railway Commissions, encouraged by popular dread of wooden cars of any kind, became a serious obstacle to the use of composite cars. As experience was gained in all-steel con- struction, it was found practicable to build an all-steel car of equal weight per car-seat by increasing the seating-capacity.' The usual tendency to conform to tradition and to experience in any change of a typical character, was too strong to admit of any marked de- parture from the principles hitherto followed in wooden-car construction. The decrease in the proportion of weight to length and seating-capacity, was gained only as structural design conformed to engineering practice in the use of steel beams. The purposes to be kept in view were to con- centrate the structural strains at the center-bearings and to resist the shocks in service that tended to distort the body-frame. In wooden con- struction, these ends were met by using the side-sills as foundation mem- bers, strengthened by tension-rods and by trusses in the side-frames; the end-sills being kept from sagging by truss-rods carried over king-posts placed where the side-sills were attached to the body-bolsters or transoms. This type of construction was analogous to that of a trussed bridge-span, but was limited, as to length between bolsters, by the available depth be- neath the window-sills for suitably bracing the truss-panels. Such designs could be strengthened by the substitution of steel as a construction mate- rial, but with increase in weight. Although steel cars built in the experimental period were much heavier 'Passenger Train Equipment. January 1, 1916 All-steel . . . Steel underframes AU-wood . . . Total . . Additions in 1915 1250 340 906 2496 Under Construction 1075 16 3 1094 In Service 14,286 6,060 41,382 61,728 6744 wooden cars were retired in three years. Steel cars in service increased from 629, in 1909, to 14,286 on January 1, 1916. See Appendix III, Table XI. ROLLING-STOCK 125 than wooden cars, their weight with further experience was considerably- reduced. Pullman sleepers built in 1907 weighed 72.5 tons ; those built in 1910 weighed 63.6 tons. There has been hkewise a reduction in the weight of day cars, in proportion to their length and seating-capacity. The first day cars built by the Pennsylvania Railroad Company were 58-^ feet long, weighed 103,620 pounds and seated 72 passengers. Cars built at a later date, 70f feet long and 78 feet between couplers, weighed 116,290 pounds and seated 88 passengers, with a reduction from 1433 to 1323 pounds per passenger, and from 1774 to 1646 pounds per hnear foot. Wooden day cars built at the same time, 70 feet in length and seating 80 passengers, weighed 106,260 pounds, or 1327 pounds per passenger and 1519 pounds per linear foot. These weights may be compared with those of day cars having composite bodies and steel underframes, as follows : Length, Feet 66 70 72 72i Weight, Pounds 110,230 117,290 118,170 119,050 Weight, per Linear Foot 1670 pounds 1675 pounds 1641 pounds 1642 pounds Variations in the weights here noted are to some extent due to the inclusion of accumulators and of axle-driven dynamos and appliances for lighting. Steel Gikdeb Frames and Other Improvements Greater advantages from the change of material were to be gained by the use of steel in beams, either simple or compound. Here, there is an analogy to a vertebrate type of structure, instead of a trussed type ; and it is on this principle that the general design of steel passenger-equipment is now being developed. It had its origin in the longitudinal central girders that were adopted in wooden freight-car construction, to transmit the shocks in service through the train, without affecting the body-structure of each car.^ In steel construction, these girders are combined in a box girder, which takes the place of the side-sills as the foundation member of the whole car-body, as well as of the underframe. The evolution of this design may be Ukened to that of the structure of a ship's hull, in which the keel is the principal member and the kelsons are the subordinate members of the whole frame, into which are built the body-frames or ribs of the hull. In the further development of this type on the Pennsylvania Railroad, / 1 Freight-cars should withstand shocks from compression of 300,000 pounds, whenever they are subjected to switching, and such shocks are often equivalent to 500,000 pounds. The center-siU construction is based on 400,000 pounds' com- pression. A dynamometer-car, weighing 51,000 pounds, ran down a grade at seven miles an hour and collided with a train of loaded cars with a shock of 607,000 pounds, the limit of the dynamometer record. 126 EFFICIENT RAILWAY OPERATION the central. box girder, the side-sills, the floor-beams and the body-struc- ture may be combined in panels of uniform dimensions, each panel includ- ing two window openings ; so that the car-body may be of any practicable length between the center-bearings and be continued beyond those points in end-sections of different designs. As the rigidity of the structure does not depend upon trussing the side-frames, side-doors may be placed in any panel. Steel castings may be used in the underframes of the end-sections, with simplification of design as compared with underframes built of beams, with greater resistance to injury from collisions and without increase in weight. A provision, not required in wooden construction, has to be made for expansion and contraction in the longitudinal members. The substitution of the central box girder for trussed side-sills as the foundation of the underframe, led to the elimination of the transom or body-bolster. The conseqifent lowering of the car-body brought about considerable changes in the construction of the trucks, which were also entirely of steel, and in the arrangement of the springs. These modifica- tions are experimental, and have not yet fully met expectation as to easy- riding requirements. A considerable decrease has been effected in weight, however. The four-wheel truck under the standard wooden car on the Pennsylvania Railroad weighed 16,000 pounds and the six-wheel truck, 21,700 pounds. The trucks under the steel cars weigh, respectively, 12,500 and 19,500 pounds. The method of panel-units has also been applied in side-truss con- struction in the motor-cars of the New York, Westchester & Boston Rail- road. The panel-units are of pressed-steel sheets, one-quarter inch thick, connected by deep flanges. Each panel consists of the outside plate of a post, the letter-board and the diagonal bracing under the windows. These units are riveted to a plate at the top, to the side-sills at the bottom and, up the center, to pressed channel-shapes which complete the posts. The window-sills are formed by Hght, continuous double-channels (UU) with a light-channel intermediate post to the side-sill. The vertical load is carried mainly on the sides; the center-sills, transmitting only buffer- strains, are held in ahgnment by the panels formed by the cross- and side- sills. The roof of pressed-steel carlines is riveted to deck-sills which are carried on the bent-in ends of the side-posts ; so that the car-body and frame practically form a structural cage. The sides are sheathed with one-sixteenth-inch steel sheets, carrying no load-stress but merely form- ing a curtain. The car has Pullman vestibules, and, without the motor- apparatus, weighs 85,000 pounds and seats 88 persons. The American car-roof is characterized by the clerestory, which gives more height to the interior, and provides light and ventilation from above. This form of roof, however, adds nothing to the longitudinal rigidity of the car-body, nor transversely, except at the ends and by a few inter- mediate straight carlines. In fact, the clerestory is essentially an inde- ROLLING-STOCK 127 pendent structure, superimposed upon the car-body itself. Its mechanical design is structurally weak, as each carHne in it has six joints, and it is so low in proportion to its length that it is insufficiently braced lengthwise. Experimental efforts have been made, in all-steel construction, to sub- stitute a roof of oval cross-section, with carlines in single pieces attached to the side-posts, or even made continuous with them. But this form of roof diminishes the height of the interior vaulting and excludes the features of upper lighting and ventilation ; nor does it appear to be sufficiently stronger than the clerestory-design to offer much greater protection in case of collision or derailment. Mention may here be made of an all-steel car of novel design, built for transferring baggage through the Hudson and Manhattan tunnels in connection with the through-train-service on the Pennsylvania Railroad. To avoid intermediate handling, the baggage is stowed in closed wagon- trucks at either terminal. These trucks are wheeled into side-openings in the car, over bridges formed by lowering the siding, and are held securely during transit by an interlocking arrangement controlled by air-pressure, which likewise controls the manipulation of the side-openings. This design deserves more particular attention, for it contains the germs of an innovation that might be of value in the handhng of freight in less-than- car-load lots. Advantages and Disadvantages of Steel-car Construction The effect upon economic maintenance of the general use of steel in passenger-car construction is as yet uncertain. Ten years of experience has proved that, as with freight-equipment, rust is quite as serious a matter as is decay in wooden construction. Doors and window-frames of pressed steel soon rust out, and serious damage from this cause, especially behind the deck-screens, is only discovered when a car is shopped for general repairs. The steel siding is easily dented and can not be restored to its original surface. The roofing sheets become warped, even with vertical joints to provide for expansion. The joints themselves are rapidly abraded by the friction of flying cinders, and are corroded by the action of coal-gas. These objections affecting the economic construction and maintenance of all-steel equipment, as well as others more directly affecting the comfort and convenience of passengers, are strengthening the tendency toward the adoption of standard passenger-equipment of a composite character, with the underframes of steel beams, the body-frames of structural shapes and the end-sections largely of cast-steel construction. Wooden flooring would then be laid on fireproof material, and the interior finish made of hard-wood veneering, with wooden doors and sash. The sheathing below the windows would also be of wood, and that between and above them of sheet steel, 128 EFFICIENT RAILWAY OPERATION or perhaps of aluminum, or the newly discovered rustless metal, stellite, and the roof of wood, canvas-covered. Steel-car construction is as yet in an experimental stage. The steel underframe of the box car, with wooden body and flooring, and of the plat- form car, also with wooden flooring, can readily be standardized and, in many details be brought into conformity with similar equipment of wooden construction. The same may be said of all-steel cars for coal and ore traffic. But many problems involved in the designs of steel passenger-equipment are still to be determined in the light of experience. These designs may be classified as the composite-type, substantially similar to the wooden car ; the all-steel car, founded on transoms or body-bolsters and side-sills — the truss-type ; and the car founded on a central longitudinal girder — the vertebrate type. The general adoption of this latter type would admit of the standardization in detail of passenger-equipment, of the manufac- ture of the several parts as stock, of ready repairs, and of a reduction in cost that should compare favorably in point of price with the wooden car, and with manifest superiority in economic and in social efficiency. - There are other difficulties to be overcome in all-steel passenger-car construction. The desire to banish fire-risk induces the disuse of wooden roofing, siding and flooring; still, consideration should be given to the discomfort thereby occasioned to travelers. For steel plating radiates heat rapidly and is sensitive to sound-vibrations. It is therefore more readily affected than wooden siding and flooring by change in external tempera- ture and by noises. These objections are measurably removed by the use of a mixture of asbestos and magnesia with other non-inflammable materials, either in sheets or in plastic condition ; though it is possible that the anti- inflammable property of some of these materials may diminish or disappear with chemical reactions induced by moisture. Wood must, however, still be used for sash-frames, for seat arm-rests and for blocking to which the sheet-metal can be attached by screws. This wooden material will weigh about 400 pounds. Other inflammable material may be required for seat- coverings and draperies. The style of decoration must conform to these changes in material, of which there is a premonition in the substitution of baked enamel for paint and varnish in the interior of the cars on the Hud- son and Manhattan tunnel-lines. Steel-car construction has necessitated novel mechanical processes. Experiments in connecting roofing plates by riveting and soldering proved unsatisfactory, and the plates are now welded by the oxy-acetylene process. Even such joints, being inelastic, in time become cracked and permit of leakage and rust. There are also evidences of electrolytic action where materials of different electrical affinities are brought in contact. These are but casual instances of the necessity for attention to details that will be required in anticipation of the complete substitution of steel in passen- ger-car construction. The change is so fundamental that it will be long ROLLING-STOCK - 129 before such details can be reduced to the uniformity which has been at- tained in many features of American rolling-stock, especially as to freight- equipment.' Transition from Wood to Steel in Car Construction. Cost and Economy The great changes in general design, and consequently in mechanical details, which follow upon the change in materials for construction, should be effected with careful attention to the preservation of existing uniformity in these respects. The economies to be expected from further standardi- zation in roUing-stock design will, therefore, necessarily be deferred until experience has justified the opinion that the experimental stage has been passed in steel-car construction. During this transitional period, no im- portant economies can reasonably be hoped for. New methods must be devised for meeting changed conditions, and "speeding-up" will be im- practicable until these methods have been coordinated and applied to standard rolUng-stock on an extensive scale. The shop-equipment re- quired for working metal must be provided, while the appliances for wooden construction will be still in use. In addition to the enormous sums of money that will be needed to finance the general substitution of steel equipment, the shop-conditions connected with it will also call for a very considerable expenditure before economic efficiency can be restored. It is a much simpler and less expensive matter to provide the tools and to obtain the skilled labor for working in wood than in metal, and the reasons which apply to the restriction of locomotive-construction to establishments spe- cially equipped for such work, apply with even greater force to the construc- tion of steel rolUng-stock. The opportunity for economic shop-efficiency is perhaps greater in car-construction shops than in locomotive-shops. The units to be built at one time are more numerous ; and an absolute conformity in design may extend to minute details ; the mechanical pro- cesses are fewer and simpler ; American wood-working machinery has been so perfected that much of the work can be performed by unskilled labor. It is, however, more difficult to organize car-repair work efficiently, for 1 It has taken many years for the Master Car Builders Association to establish standards for the multiplicity of parts involved in the construction of running gear, of brake- and coupler-rigging and of safety-appliances; matters of vital impor- tance in the repairs of cars interchanged throughout our vast mileage. In this general interchange, certain dimensions should also be standardized, in order to comply with the minimum roadway clearance-restrictions existing on our railway systeip. After long consideration, the Master Car Builders Association has rec- ommended standard dimensions for closed freight-cars to meet these require- ments, as also to establish the relation between the length of such cars and the spacing of warehouse-doors. As recommended, the inside dimensions of box cars are 36 feet in length, 8 feet 6 inches in width and 8 feet in height ; special equipment not to exceed 40 feet 6 inches in length, 8 feet 6 inches in width and 9 feet in height. Maximum outside width, 9 feet 2 inches at 13 feet above top of rail. 130 ' EFFICIENT RAILWAY OPERATION much of it is of a minor character which is done by hand-work upon out- door-tracks where it is likely to be isolated from supervision.' The rapidly increasing substitution of steel rolling-stock indicates the necessity for corresponding increase in facilities for such construction in commercial estabUshments, so far as may be warranted by the financial condition of our railway system. In this connection, it may be of interest to arrive at an approximate estimate of the cost of this substitution. It may be assumed that all cars of 60,000 pounds' capacity, or less, are of wooden construction and should be replaced with steel underframes ; say at a cost of $600 per car. In 1914, there were about 850,000 cars of this description." The total cost of replacement would, therefore, be over $500,000,000. Add to these figures the cost of sohd steel. wheels at, say, $315,000,000,' and the increased investment in freight-equipment would total $815,000,000. On December 31, 1915, there were some 41,000 wooden cars in passenger-service and the cost of their replacement was estimated at $528,000,000. Therefore, the general substitution of steel for wood in the rolling-stock of our railway system would require a capital investment of not less than $1,343,000,000. In the substitution of steel for wood in car-construction, consideration must be given to the extent and character of passenger-equipment, as well as of freight-equipment. The total passenger-equipment, in 1914, num- bered 53,466 cars. Of this number, 23,490 were first-class passenger-cars and 13,607 were baggage, mail and express cars ; or about 70 per cent, of the whole. Up to 1910, about 3100 cars were of all-steel construction and principally of these two classes, including 487 sleeping-cars. In 1914, the whole number of sleeping-cars owned by railway companies is given as only 636, and it is probable that these cars are now all-steel. There were 1282 dining-cars and 561 parlor-cars, which will ultimately be replaced by all-steel also. The remaining 13,890 cars are principally combination and second-class cars.* It may be reasonably expected that, out of this total of 53,000 cars, not more than from 15 to 20 per cent, will be of all-steel construction. But steel underframes will ultimately prevail as with freight-equipment and, in addition, the end-frames and platforms will be of cast-steel, as a safety ' In May, 1914, the American Railway Association appointed a Committee on Standard Box-oar Design. A sub-committee of ofBcials of the mechanical departments and of representatives of prominent car-building companies recom- mended three types of box cars : 1. Double-sheathed wooden cars of 60,000 to 80,000 pounds' capacity. • 2. Steel-frame, single-sheathed oars of 80,000 pounds' capacity. 3. All-steel cars of 80,000 to 100,000 pounds' capacity. The reduction of the many different types to these standards would greatly diminish the number of repair-parts required to keep foreign cars in service. All the parts could be standardized and be held in stock by manufacturers 2 See Appendix III, Table III. 3 See page 116. * See Appendix III, Tables VI and VIII. ROLLING-STOCK 131 device. The exterior of these cars with wooden side-frames will probably be partially sheathed with steel, but steel finish inside will not be favored because of corrosion from moisture. As abeady suggested, the body- frames may be of structural shapes instead of wood, and the flooring under- laid with fireproof material. The 'trucks, however, wiU be of steel. How long it may take to complete this replacement is mere conjecture. The average annual replacement may be estimated at 5000 cars of all descrip- tions. On this basis, it would require "ten years to carry out such a substi- tution as is here contemplated. In the nine years, 1906 to 1914, inclusive, the annual increase in passenger-equipment has averaged about 1500 cars! There should be shop-facihties, therefore, for the construction of at least 6500 cars per annum. Some Ught on this subject may be obtained from a statement issued by the Census Bureau, June 15, 1916, as to the con- struction of steam and electric cars in 1909 and in 1914, which is sum- marized in Appendix III, Table IX. From this statement, it appears that m 1914 there were 131,799 freight-cars and 3558 passenger-cars built in the United States, exclusive of 2821 electric cars. It would seem, therefore, that the output should be virtually doubled to replace wooden passenger-cars by steel cars within ten years.' Rolling-stock Defects It is a difficult matter to establish any satisfactory standard of com- parative efficiency in the Rolhng-stock Department from statistical data. Something of this kind is, however, suggested by reference to the inspec- tion reports of the Division of Safety of the Interstate Commerce Com- mission. A statement of the comparative percentage of defects in safety- appliances, compiled from these reports for 1910 to 1914, as to roads with equipment of 10,000 cars or more, is given in Appendix III, Table X. Where this statement shows a variation in five years of only between 1.5 and 2.0 per cent, of the total number of cars inspected belonging to a par- ticular company, it may be inferred that the shop-efficiency there is of a higher character than where, in the same period, the variation was be- tween 7.0 and 14.6 per cent. Other variations, as between 3.2 and 11.1 per cent, were probably due more to financial considerations than to shop-conditions. Taking the twenty-seven companies included in the table, the comparative result is as follows : Defects 1910 1911 1912 1913 1911 Under 5 per cent. Under 10 per cent. Over 10 per cent. 14 12 1 17 10 9 16 2 11 10 6 11 15 1 ' Information regarding steel-car construction not previously acknowledged, Jias been obtained from a valuable paper on "The Construction of Iron 132 EFFICIENT RAILWAY OPERATION From this comparison, it may be assumed that an average under 5 per cent, indicates a relatively high standard of shop-efficiency, and that such a standard has been attained in at least one-half of the railroad shops in this country. An inference as to the weaker points in car-construction may be derived from a further report as to derailments caused by defective equipment. In ten years, from 1906 to 1915, the annual average has been as follows : Broken wheels 275 Side-bearings 94 Broken flanges 575 Arch-bars 179 Loose wheels 109 Rigid trucks 124 Other wheel-defects 104 Power-brake hose 206 Broken axles 379 Couplers 181 Brake-rigging 421 Miscellaneous 375 Draft-gear 226 Total average 3,248 Under each of these heads, there has been a fluctuation from year to year with no apparent relation to the character of equipment or to car- mileage ; and no marked decrease in any item during the period. The high average of derailments from broken wheels and flanges should be materially reduced with the increasing use of steel wheels. Derailments from defective draft-gear and brake-appliances are largely due to inefficient maintenance and inspection. Comfort and Luxury in Car Design. Car-lighting Economic efficiency is a paramount consideration in the design and construction of freight-equipment, but it yields to social efficiency where passenger-traffic is concerned ; and in no other country as in the United States. In designing passenger-cars, the relation of weight to seating- capacity has hitherto been disregarded. Their interior fittings are the best of their kind, and decoration has been carried perhaps to excess, espe- cially in Pullman cars. No crowned head of Europe enjoys more luxurious appHances for railway travel than are provided for the American citizen in his own land. This is conspicuously the case in apphances for fighting and heating passenger-trains. The English stage-coach proprietor made no further provision for the comfort of his passengers than to seat them, more or less inconveniently, within or on his coaches ; nor did the railway companies, in the transition period from horse power to steam. The railway carriage was the conven- tional stage-coach in all respects. The journeys were short and not usually prolonged into the night. The traveler who wished for fight pro- vided his own pocket-candle until the railway companies consented to light each compartment dimly by a candle placed in the roof ; and this Passenger Cars in the United States of America," by F. Gutbrod, Member of Board of Works, Berlin, published in the Bulletin of the International Railway Congress, November, 1912, et seq.; and from operating officials of some of our principal railway systems. ROLLING-STOCK 133 long remained the standard lighting-system on English railways, until it was replaced by oil-lamps in 1850. As the four-wheel truck led to the long, open car on American railroads, so did the open car lead to a different system of train-hghting ; first with a candle-lantern at each end of the car, for which an oil-lamp was after- ward substituted. But neither of these illuminants gave more Ught than was sufficient for passengers to grope their way to or from their seats. An improvement in car-Ughting followed upon the introduction in the early sixties of mineral-oil or kerosene, which gave a brighter light with less labor in cleaning and filling lamps, and less dirt and damage from grease-spots. After the clerestory or deck-roof became a common feature of construction, additional lamps were placed along the center line of the car; and this remained the standard lighting-system for many years. The advent of the air-brake brought about a change in car-lighting, as in other features of train-service. About 1880 Mr. Westinghouse passed compressed air through a "carburetor," suspended under the car-roof, which contained absorbent material charged with a very volatile petroleum product, known as "88° gasoline." The resulting inflammable vapor was conducted through piping to suitable burners. It was found that changes in temperature greatly affected the carburetor, and the system remained in an experimental stage until about 1885, when it was much improved. As this "carburetor system" then equaled the kerosene lamp, as to bril- liancy of illumination, and usually required attention no oftener than^every ten days, it was gradually taking its place when rival methods of lighting with compressed gas supplanted it. The carburetor system is, however, still valuable where there are no facihties for storing either compressed gas or electric current. The use of coal-gas for car-lighting originated in Europe, about 1870. Gas from city-mains was stored in a reservoir in the guards' vau, being shghtly compressed by water-displacement, and was thence conveyed through the train by pipes with hose-connections. This system may be considered as a normal successor to the carburetor system, as the latter pointed the way for its improvement by storing the gas under compres- sion in individual reservoirs under each car, whence it was piped to the burners. The Pennsylvania Railroad Company developed this plan, but the illuminating power of gas proved to be considerably diminished by compression ; a loss which it was unsuccessfully attempted to restore by mixture with acetylene gas. The "Pintsch system" of enrichment with defiant gas has since been substituted, in which the gas has been dis- tilled from crude petroleum and is not decomposed by compression. The Pintsch system originated in Prussia, in 1870, but did not come into general use until the government took over the private Hues in 1880-84. By 1900, this system of gas-lighting had been extensively developed in all countries. In Europe alone, it was in use upon 118,000 carriages. Oil- 134 EFFICIENT RAILWAY OPERATION gas loses but little of its illuminating power under compression and can be stored at 90 pounds' pressure under each car in sufficient volume for twenty- four hours' continuous use. Attempts have been made to increase the bril- liancy of the light by the use of Welsbach mantles, but with doubtful economy because of breakage. The risk from fire in derailments or col- lisions has been greatly lessened by strengthening the car-reservoirs, but there is still a possible risk from gas-leakage. As calcium-carbide became commercially available, its reaction with water furnished an illuminant in acetylene gas, which could be readily produced without the aid of storage gas-holders or a compression-plant. Acetylene gas-Ughting was introduced from Canada into this country in 1899, upon the Great Northern Railway. The generator in each car is charged through the roof with 150 pounds of carbide. Water is also sup- pUed through the roof. As it rises up from the bottom of the container to the grating on which the carbide is placed, the latter is slacked and the precipitate drops to the bottom. The gas forms a slight pressure upon the surface of the water, which automatically regulates its production and consumption. There are also a condenser, storage-reservoir and a regu- latoi-, to insure pure, dry gas at a steady pressure. The generator is pro- vided with a safety-valve, and, to prevent freezing under the floor, heater- pipes are connected to the covered conduits under the base of the generator ; 4.7 cubic feet of acetylene are produced per pound of carbide. Pure acetylene gas can not be compressed without risk of explosion from in- crease in temperature but, in the Pintsch system, it has been found prac- ticable to use oil-gas mixed with 30 per cent, of acetylene without danger and with doubled illuminating power. Electric Lighting As the substitution of the air-brake for the hand-brake prepared the way for the compressed-gas system, so did the substitution of electricity for animal power, on the street-railway, furnish facilities for another system of car-lighting; for electric car-Hghting originated on the street-railway, where the current was taken directly from the power-line. Electric light- ing on steam-railways was tested experimentally, in 1885, on the London, Brighton & South Coast Railway, on the Northern Railway of France and, in this country, on the Pennsylvania and the Boston & Albany roads. The generator was, at first, operated on the locomotive and afterward trans- ferred to the baggage-car, where it was actuated by steam from the loco- motive or by an oil-engine. As the train was deprived of lights when de- tached from the primary source of energy, electric lighting did not become really practicable until about 1897, when accumulators, or storage-bat- teries, were placed in each car. By 1905, it had been introduced on the principal lines in Europe and in America. ROLLING-STOCK 135 Where power-plants are available, accumulators may be charged inde- pendently of the locomotive. Still, this requires that the cars shall be held at such stations for many hours for recharging. The accumulators were therefore made movable, to admit of an exchange of accumulators before those in use had become exhausted. A further improvement was soon introduced on the Caledonian Railway^ in Scotland, in maintaining the lighting-power with current regenerated by mechanism connected by a belt with a revolving axle under the car. Difl&culties have been overcome in the adjustment of the voltage and amperage to variations in the speed of the train, and the life of the accumulator is prolonged for a considerable time after the car is detached from the train. A valuable improvement has been effected by the suspension of the generator from the underframe of the car ; by this change, nearly a ton in weight is removed from the truck to the spring-borne car-body, with accompanying automatic adjust- ment of the driving belt to the varying motion of the truck, and with greater facility for inspection. The operation of this system has been found so satisfactory that it has met with general favor. The "head-end" system of lighting from a generator in the baggage- car was first installed by the Pullman Company, in January, 1887, on the "Florida Special Train," operated from New York to Florida, with accumu- lators in each car. Even at the present time, this system is in successful operation on the transcontinental lines out of Chicago. In the service to Seattle, over the Chicago, Milwaukee & St. Paul Railway, there are eight- een sohd trains of ten cars, each lighted on this system, with an accumu- lator in the baggage-car, one in the middle of the train and one in the rear car. They are fully charged during the later part of the night and the generator is not operated in the day-time or on heavy grades. Electric lighting adds to the comfort of passengers by furnishing a bril- liant hght that may be easily placed for reading, though it can not be so readily graduated as gas or oil can be, when not so required. Its bril- liancy may be reduced by the intervention of a rheostat, but with no saving in current ; and the simpler device of veiling the light is equally efficacious. The electric equipment is estimated to add nine tons to the weight of each car. The batteries are self-consuming to an expensive degree and, where they are recharged, to some extent, by a generator on the train, this re- quires an appreciable draft of power from the locomotive. Electrically- lighted cars. can not be used where electric current is not available nor can cars not so equipped be introduced into an electrically-lighted train, save as appendages. The additional expense hardly seems warranted, even as a concession to social efficiency. For the reasons here given, in the present stage of the art of car-lighting, the compressed-gas system is probably preferable from the standpoint of economic efficiency. 136 EFFICIENT RAILWAY OPERATION Car-heating Imphovements. Ventilation. Painting . Car-heating followed the same line of development as car-lighting, from the stage-coach to the railway carriage. At first, travelers relied upon wraps and furs to retain animal heat and then resorted to heated bricks and portable hot-water receptacles. At length, the railway com- panies themselves supplied such appliances for the first-class carriage com- partments, to be exchanged occasionally, as they became cold ; and this was where the matter stood on European railways until the central system of car-heating was introduced. It was not so with American railroads; for here, again, as with car- lighting, the open car suggested a different plan. A stove was placed at one end of the car, with wood as fuel. The wood was gathered up by the brakeman at the wood-racks from which the tender was loaded and piled around the stove, to be supplied at the fancy of the passenger seated nearest to it. The temperature thus regulated by him and by frequent drafts from opened doors, decreased toward zero at the opposite end of the car. As coal supplanted wood for locomotive-fuel, it was supplied to the car- stove which, for economy's sake, was locked and filled at the will of the brakeman. To save himself trouble, he filled it full and the cast-iron stove was often red-hot, to the inconvenience of those compelled to sit near it and to the general discomfort of all who breathed the vitiated air, thus deprived of oxygen, except as it was regenerated by drafts through opened doors and window-cracks. This primitive method of car-heating was subsequeiltly replaced by the "Baker heater," by which hot water was circulated through pipes along the sides of the car and under the seats from a boiler inclosed in a closet at one end of the car. Under this system, a more equable tempera- ture was maintained and could be rationally controlled. The heater was, however, objectionable on account of fire-risk from occasional overheating or from its destruction in an accident ; though it is still retained, where it is impracticable to utilize a central system. As the principle of heating an entire train from a single source is only practicable when that source is the locomotive-boiler, its application to such a purpose met with considerable opposition at first. Objections were based upon the draft of power thus diverted from traction, upon the dif- ficulty of securing a uniform system of pipe-connections and other neces- sary appliances on trains in through-service over different lines and in different countries ; as also in heating cars when separated from the loco- motive. Profiting by experience gained in estabUshing uniformity in automatic couplers and in braking-apparatus, uniformity in heating-ap- pliances was more readily secured in this country than in Europe, where central heating only gained recognition about 1905. Like the central system of car-lighting, the central system of car-heating ROLLING-STOCK • 137 has originated with the transmission of energy from the locomotive. As compressed air was obtained from the air-pump, steam-heat was drawn from the locomotive-boiler and piped through the train. The defective operation of the earlier appliances was gradually remedied. A circulating system has been evolved in which the return-circulation of steam is main- tained by a vacuum-pump on the tender, with drips and taps in the train to carry off the water of condensation. This system satisfactorily provides for the comfort of passengers, yet, like the approved methods of car-light- ing, central heating is not so satisfactory from an economic point of view ; since either of the two systems in common use makes a heavy draft upon the boiler-capacity and fuel-consumption of the locomotive. Electric heaters have been successfully introduced into trolley-cars and experimentally on some electrified steam-roads, but as yet electric heating has not reached a point at which it affects car-heating in general. The heating of a passenger-train is inseparable from its ventilation. The conditions are so dissimilar from those under which ventilation is elsewhere required, as to make it a much more difficult problem. It is an easy matter to change the air in a train in motion, but not so easy to pre- vent an inrush of drafts accompanied by dust and cinders. Again, the conditions are different in vestibuled cars with closed doors from those in cars whose doors are frequently opened upon uninclosed platforms. Then, too, provision is to be made for cars at rest, as well as in motion. Ventilation was first practically associated with car-heating in the "Spear Stove" system. Screened hoods at diagonal corners of the car were connected " with boxing around coal-stoves. The warmed air was forced by motion of the train through a flue or duct along each side of the car just above the floor, whence it escaped through openings and passed out through the deck-sash. This system of ventilation disappeared. with the abolition of individual car-heaters and the advent of central heating, which was enforced by an order of the Interstate Commerce Commission, dated May 19, 1905. There are now. two recognized systems of central heating. One is the direct system of piping along the sides of the car and under the seats ; air being supplied through the deck-sash. Dust and cinders are excluded by wire netting, and a violent inrush of air is measurably checked by trailing the deck-sash with the direction of the train. In moderate weather, the change of air is for the most part in the deck-roof and does not adequately descend to the breathing line. In winter, the cold air drops into the body of the car with momentary relief from overheat, but in objectionable drafts. In summer, the movement is not sufficient for good ventilation. This system has been somewhat bettered, as to the deck-sash, by directing the air-currents upward and outward through the roof. It is in use in Pull- man cars, and in ordinary service on many roads. The indirect system originated on the Pennsylvania Railroad, about 138 EFFICIENT RAILWAY OPERATION 1890. In this system, air is introduced through hoods at diagonal corners of the car, as in the Spear-stove system, too high for dust to enter, while the smoke from the locomotive rises still higher or is diverted to the side of the train. Fine cinders, passing through the netting at the intake, fall into a hopper. The air is then admitted into horizorital ducts contain- ing the heater pipes. As it rises thence into the body of the car, it escapes through- "Globe" ventilators in the deck-roof; the deck-sash being im- movably closed. It has been experimentally determined that 62,400 cubic feet of air per car per hour can be changed, with all ventilators open, at a speed of 30 miles an hour ; from 27,000 to 30,000 cubic feet with ventilators closed ; and 23,000 cubic feet in a train standfng still and ventilators open. In a standing train, the fresh air may be increased ui day-coaches by opening the doors for a few minutes. With steam at 20 pounds' pressure, about 60,000 cubic feet of air per hour is about the maximum volume that can be properly heated, say. to 70 degrees, or 1000 cubic feet to a passenger in a sixty-seated car. In a run of 237 miles, in an external temperature of two to five degrees below zero, an internal temperature of 70 degrees was main- tained with 30 pounds' steam-pressure, with a radiating surface of 328 square feet in each car. The usual proportion is one square foot for 16.9 cubic feet of volume, and for every 268 cubic feet of fresh air per hour. This proportion must be exceeded in all-steel cars, in which the radiating surface is increased by winding wire around the piping. The air in a car can be changed every four minutes while the train is in motion but, unless the areas of the intakes and exits are so equalized as to balance the air-pressure, one end of the car will be colder than the other. Moisture in the air, which is essential to proper ventilation, quickly dis- appears in a close and heated car, and more complaint is caused by improper heating than by insufficient ventilation. To secure efficiency in both re- spects, a balance of air-pressure must be maintained, with not less than 1000 cubic feet of fresh air per passenger per hour. This result may be attained, some day, by automatic devices, and is only practicable with the indirect system. To heat and waste the required volume of air is a more expensive undertaking than was at first appreciated, and it makes a considerable draft upon the steaming capacity of the locomotive. Ventilation in a Pullman sleeping-car, with double sash and vestibules, is far more difficult than in open coaches. "When the berths are made up and the curtains closed, the movement of the heated currents through the deck-roof tends to draw vitiated air from the smoking-room into the body of the car, where there is an excess of heat and lack of ventila- tion, and especially in the lower berths. While the train is in motion, these conditions are but partially remedied by wire screens in the windows and cold drafts from the deck-sash that disturb the occupants of the upper berths. The window-screens are difficult of adjustment to the direction ROLLING-STOCK 139 of the train and to changes in the external temperature, and it would be an improvement to provide, instead, a ventilating panel in each lower berth that could be manipulated by its occupant. The smoking-room and toilets should be separately ventilated. When sleeping-cars are separated from the train, to stand for some hours fully occupied, the heating system should be fed from a stationary plant, as originated on the Lehigh Valley Rail- road ; but the resulting overheat and lack of ventilation can only be remedied by the addition of an artificial exhaust, until the car is again in motion. Under exceptional conditions, as in dining-cars and smoking- cars, and in sleeping-cars occupied for some hours before they are taken into train, a forced-draift or blower system, acting through the air-ducts, is especially desirable. Although the heating and ventilation of passenger-trains, and par- ticularly of sleeping-cars, can not yet be said to have been adequately ac- complished, still much has been achieved toward the solution of an ad- mittedly difficult problem, which, as affecting the health and comfort of passengers, may be regarded as an element of social efficiency in railroad transportation, justifying the somewhat extended consideration which has been given to it.' The decoration of passenger-equipment is rather a question of taste and, therefore, of social efficiency ; though the use of paint as a preserva- tive is a matter of economic importance. Oil is the preservative element in all paints, and the question of cost must be considered with reference to endurance and lasting qualities, before adopting cheaper substitutes for linseed-oil. The mechs^pical processes employed in external painting must be carefully conducted, to insure adequate mileage before repainting be- comes necessary. Upon the qualities of varnish also depends the frequency with which equipment must be shopped, and the conditions of tempera- ture and dust under which the varnish is applied are matters in which economic shop-efficiency can be displayed. In the selection of painting- materials, attention should .be paid to their chemical composition and probable reaction under weather-exposure. This is of equal importance in painting freight-equipment, as is also the choice of colors which should lend conspicuousness for the easier locating of stray cars. The lettering and other symbols of ownership should also be designed with reference to ready identification. The cost of painting freight-cars has been sensibly reduced by spraying the coloring matter with compressed air. ^ 1 For furtlier information on this subject, see Proceedings of the Master Car Builders Association, 1908 ; and American Engineer and Railroad Journal, 1908, p. 313. 2 " Colorizing, a Protective Treatment for Metals," H. B. C. Alhson, L. A. Hawkins, Electric Railway Review, July 30, 1915. "Metallic Preservative Coatings," E. H. Fish, American Machimst, August 26, 1915. "Painting Iron and Steel," James Scott, Journal of American Society of Mechanical Engineers, August, 1916. 140 EFFICIENT RAILWAY OPERATION Amebican Inventions In closing this chapter, it may be noted that four of the most eflBcient improvements in roUing-stock design are of American origin, — the center- bearing truck, the vertical-hook coupler, the air-brake and the vestibule buffer-platform. The railway passenger-train was originally an assem- blage of independent units loosely associated in series, but, thanks to the inventive genius of Janney and Westinghouse, and to the influence exerted by Pullman, it is now a closely-articulated organism, with reliable means of communication throughout its structure; Compressed air constitutes its pneumatic system; its own energy of motion is subservient to its illumination, with individual reading-lamps, electric fans and bell-calls as luxurious adjuncts ; while heat is imparted from the locomotive by a cir- culation akin to that imparted by the heart to animal life through the arteries and the veins. CHAPTER V ROADWAY PART I. SUBSTRUCTURE Railway Location and Right-of-Way. Economic Alignment and Grade The ways of communication by land between communities are con- ditioned, as to routes, by their physical and social environment ; — physi- cally, as to the topographical features and material resources of the inter- vening region ; socially, as to its density of population and its commercial requirements. In point of economic efficiency, the normal service should be performed with the least expenditure of motive power ; therefore, the route should be as nearly straight and level as may be practicable. In point of social efficiency, it should be so located as to provide for the great- est volume of traffic that may reasonably be expected to be accommodated by its construction. It need scarcely be said that this desideratum of a theoretical minimum of effort combined with an estimated maximum of service is practically im- possible. The just mean is to be found in such a compromise of economic and social requirements as will result in a reasonable return from the estimated investment of capital. If the region to be traversed is greatly diversified by mountains and streams, and its material resources are either scanty or scattered, the location of the route should conform to its configuration, subject to the restrictions imposed by the nature of the motive power to be used. But if communication is to be provided between densely pop- ulated communities, . or for a heavy traffic in commodities produced or concentrated within narrow limits, the cost of construction should be balanced with the cost of operation per unit of transportation, in the estab- lishment of the horizontal and vertical alignment of the route. In modifying the topographical environment of a community, in order to facilitate its means of communication, the engineer does not deal with human passions and emotions nor with metaphysical abstractions, but with concrete facts. These he must mold in accordance with physical principles which he cannot control and to which he must conform, regardless of pre- conceived opinions or of personal predilections. The efficient location of a commercial highway should, therefore, be preceded by such knowledge of 141 142 EFFICIENT RAILWAY OPERATION the topography, chmate, material resources, population and commercial conditions of the region to be traversed, as will enable a reasonable com- promise to be made between its physical features and its social require- ments. Soon after the line has been- provisionally located, options should be secured for the right-of-way. It is better to pay well for a title in fee- simple than to accept a mere easement, even as a gratuity. Wherever practicable, the width should not be less than one hundred feet. Through wild lands, a width of two hundred feet may often be obtained at a nominal cost. In course of time, the wisdom of this course becomes apparent in freedom from local interference and in saving subsequent land-damages, as business development shall require additional elbow-room. Until so required, the outer zones through wooded lands may remain in forest, and the open land may be leased to adjacent farmers. The right-of-way notes should be recorded at once in permanent form, the metes and bounds carefully established and prominently marked, and the titles duly registered. Inattention to these matters, while the details are fresh in the minds of all interested, has often resulted in tedious litigation at heavy cost. On any line of considerable length, its ruling gradient is not determined by the' relative altitude of its termini, but by the rise and fall of the country which is to be traversed. If there be any intervening mountain range, the ruling gradient is to be determined by the relative cost of surmounting that range, either by developing the line in length, by tunneling or by the introduction of exceptionally heavy gradients at critical points. The cost of construction in either case being approximately equal, exceptional gradients may serve to keep the ruling gradient down elsewhere, with subsequent economy in operation, where the traffic is in heavily loaded trains. For if, thereby, the normal train-loads over the rest of the line may be increased by even a small percentage, the cost of assisting such trains over the critical' grades will add but slightly to the general cost of the serv- ice. In the location of a line across an existing line, it is advisable, wherever possible, to bring the new line parallel to the other for a considerable distance on each side of the intersection, with station-platforms between them. This plan- facilitates transfers and track-connections, and also gives a better view of trains approaching the crossing. Of the two departures from a right hne between termini, whether ver- tical or horizontal, the latter is of less importance economically and may be of great advantage socially, where divergence is required to reach centers of considerable traffic. A divergence of ten miles to the right or left of the middle of a line a hundred miles in length, adds but two miles to the total distance and inappreciably to the cost of operation. Nor are long tangents of material advantage. A hne may profitably conform to the topographi- cal features of a region by frequent curves separated by short tangents. ROADWAY 143 There are, indeed, some advantages in operating over long tangents, by decreasing the risk of coUisions or derailment, when emergencies occur in train-service ; but, as a general rule, the saving of distance by the preservation of long, straight reaches of hne does not justify any consid- erable addition to the cost of construction. The cost of operating addi- tional mileage does not add materially to the total cost per transportation unit. This is evident from a comparison of the cost, in this respect, of the several lines between New York and Chicago ; as, for instance, on the New York Central line, which runs due north for 150 miles before trending westward, without affecting its importance as a through line to the West and with its traffic largely increased from the trade-centers that it serves, by reason of this divergence. The vertical alignment of any highway must conform to the mode of transportation to be utilized upon it, whether it be a mountain path for porters only, or a traU for pack-animals, or a macadamized turnpike for stage-coaches. In any of these cases, it is conditioned by vital energy and by muscular power. But with a railway, it is a matter of mechanical power and of adhesion to the rails. The cost of train-service is more un- favorably affected by operating over frequent changes of grade, within the limits of the ruling-gradient, than over a long ruling-gradient continuously to a summit ; since no economic purpose is served in successively surmount- ing the intervening elevations on a longer route. Inconspicuous breaks of grade, though slightly diminishing the cost of construction, may therefore seriously affect the operating value of a railroad location. Changes of grade should be connected by vertical curves of sufficient length to relieve the shock to a long and heavy train as the slackened draw-gear is stretched by the application of steam to the locomotive.^ On a straight Une, under 1 The effect of intermediate changes of grade, or of " undulating " grades, varies with the acquired momentum of the train as it approaches an ascending grade. In the case of an ordinary passenger-train approaching upon a level a series of undulating grades of one per cent., or 62.8 feet per mile, at a speed of 50 miles an hour, its acquired momentum may be assumed as equivalent to a "potential lift" of 88.75 feet before coming to a state of rest. At a summit, one mile distant, there would still remain a potential lift of 88.75 - 52.8 = 35.95 feet, correspond- ing to a speed of about 33 miles an hour on a level, without additional power from the locomotive. If the next descent be 0.6 mile, and 31.7 feet fall, the additional momentum due to the acceleration on the descent would result in a potential lift of 35.95 + 31.7 = 67.65 feet, corresponding to about 44 miles an hour. On the next ascent of 0.4 mile, a rise of 21.12 feet, the train would arrive at the summit with a remaining potential lift of 67.65 - 21.12 = 46.53 feet, corresponding to a speed of 36 miles an hour. The effect of this series of undulating grades upon the performance of the train, for this distance, would be to reduce its speed from 50 miles to 36 miles an hour with an absolute rise of 52.8 feet, without additional power from the locomotive. Assuming that the changes of grade were connected by suitable vertical curves, there would have been a uniform pull upon the drawbar of every ear in the train,. and consequently with no jerking effect upon the pas- sengers. See "'Economic Theory of Location of Railways," A. M. Wellington, page 348. 144 EFFICIENT RAILWAY OPERATION favorable conditions, a locomotive may ascend a gradient of 1 in 22.5 by- adhesion alone, though the theoretical limit is about 1 in 16, or 330 feet to the mile. On such a gradient, its economic efficiency would be nil, as it could exercise no tractive power at the tender-drawbar. Inclined Planes and the Rack-kail. Ruling Gradient Inclined planes were frequent features of EngHsh tramway construc- tion, following upon the use of steam hoisting-engines in the collieries. The maximum gradient for animal power was between 25 and 30 feet to the mile, and this limit seems to have been as much determined by the difficulty in controlhng the speed of the loaded cars on the descent as in maintaining the average efficiency of animal traction in the opposite direc- tion. Where this controlling gradient could not be preserved, resort was had to inclined planes. There were several double inclined planes on the Stockton & Darhngton Railway. Although locomotive traction had by that time attracted much attention, the practical value of the principle of adhesion had still to be tried out, and it was seriously proposed to operate the Liverpool & Manchester Railway by a continuous series of double inclined planes, in twenty-one sections, in the total distance of thirty-two miles. It was to decide this matter, that the celebrated Rainhill com- petitive tests were made, in which the victory of the "Rocket" opened up a new era in railway traction. Railway construction in the United States underwent a similar de- velopment. There was an inclined plane on the Quincy tramway. The tramway of the Delaware & Hudson Coal Company ascended the Lacka- wanna Mountain with a rise of 800 feet in 3^ miles by inclined planes of 1 in 12 and 1 in 20. The Mauch Chunk road was also an early example of tramway operation by animal traction, in connection with a series of inclined planes.^ Inclined planes were retained in steam railway construc- tion. On the Baltimore & Ohio Railroad, as originally located, there was an inclined plane, 41 miles from Baltimore, ascending 80 feet in 2150. A second plane ascended 100 feet in 3000. From the summit, 813 feet above sea-level, the Une descended by one plane of 160 feet in 3200 and by another 81 feet in 1900. On the Mohawk & Hudson Railroad, opened August 9, 1831, the ascent of 185 feet from the Hudson River to the upper level in Albany, was oper- ated by a twelve-horse-power engine. On the Portage Railroad in Penn- sylvania, built over the Alleghanies in 1834, there was for thirty-six miles a succession of ten inchned planes, operated until 1853 by engines of thirty- five horse-power at a rate of four miles an hour. The summit was 1398 1 See Appendix IV, Table IX. ROADWAY 145 feet above the eastern canal-basin, 1171 feet above the western basin and 2311 feet above sea-level. The longest plane was 3116 feet with a rise of 307 feet. At the Columbia end of the Philadelphia & Columbia Railroad, opened October 7, 1834, there was an inchned plane rising 90 feet in 1800, and at the Schuylkill terminus, one of 196 feet in 2800, or 369 feet to the mile ; the intervening maximum gradient being 44 feet to the mile. This inclined plane was the scene of a notable performance on July 9, 1836, when a locomotive, built in Philadelphia by William Norris, ascended it in two minutes. This exploit was so unprecedented that the announce- ment was received with incredulity, until it was repeated ten days after- ward in the presence of an officially appointed commission, and was sub- sequently commented upon by European engineers.^ A locomotive of the "Camel" type (0-4-0), built by Ross Winans, was operated on a temporary track over the Kingwood Tunnel on the Baltimore & Ohio Railroad, on a grade of 530 feet to the mile, hauling one car-load of material at a time. Practically, the ruling gradient rarely exceeds 52.8 feet to the mile. In exceptional cases, this may be exceeded and the normal train assisted at such points by "pushers" or "bank-engines." Gradients beyond the limit of adhesive traction may also be ascended by means of the rack-railroad. A cog-wheel or pinion engaged in a toothed rack on the track was patented in England in 1811. In 1812, a locomotive on this plan hauled coal from Middleton to Leeds, a distance of 3^ miles. A half-century later, Sylvester Marsh adopted this system for the Mount Washington Railroad in New Hampshire, on a gradient of 1 in 2^. The rack had pin-teeth cut in angle-bars. In Switzerland, there has been a considerable development of the rack-railroad on the Abt system. This system consists of a multiple-rack, its sections laid side by side, with the teeth breaking joints or "staggered," so that the driving-wheels are con- stantly engaged with the rack ; the rack combined with adhesion working on steep grades, and adhesion only on easier onest The system has been specially adapted to mountain-roads engaged in excursion-traffic. The Mount Pilatus Railway, near Lucerne, is operated on a gradient of nearly 1 to 2 by a double rack with vertical teeth on each face. The Jungfrau line has teeth cut in the head of a T-rail. In England, the ruling-gradient is generally much easier than in the "United States. On the Liverpool & Manchester Railway, it was 1 in 900, except on the inclines in Liverpool and at Rainhill. On the Great Western Railway, it is 1 in 1320 for a long way out of London. One of the steepest ' The locomotive employed on these occasions was of the following dimen- sions : cylinders, lOJ inches diameter by 17f inches stroke ; driving-wheels, 4 feet in diameter ; truck-wheels, 30 inches ; 78 tubes, 2-inch diameter by 7 feet long ; weight on driving-wheels, 8700 pounds; total weight, 14,930 pounds; trailing load, 31,270 pounds. With 80 pounds' steam-pressure, the ascent was made in 2 minutes 24 seconds. "When Railroads were New," C. F. Carter. 146 EFFICIENT RAILWAY OPERATION gradients, on the Midland Railway between Birmingham and Gloucester, is the Lickey incline, which is 1 in 37 for two miles.^ The restriction of the vertical alignment by the ruling-gradient affects also the horizontal aUgnment. To keep within this limit, advantage is taken of the work which has already been performed by water-courses in excavating valleys and ravines in mountain-slopes and hill-sides. Here the value of the practiced eye of the experienced locating engineer is seen in fitting the hne to the face of the country.'' The summit is generally to be sought in a mountain-pass or gap, and the line can usually be developed in sufficient length to attain this elevation without exceeding the ruhng- gradient. In some cases, this purpose can only be accomplished by resort to unusual expedients, such as the spiral tunnels on the St. Gotthard Rail- way, or on the lines through the Rocky Mountains ; or else by zig-zags or "switch-backs," as on the line from Bombay up the Ghauts to the in- terior plateau of Hindustan.' An early example of this method of overcoming excessive elevation in railroad construction is the Mauch Chunk Switch-back.* Other instances may be cited, though of a temporary character. The Baltimore & Ohio Railroad was operated for some time, in 1852, over the Broad Tree Tunnel, near Wheeling, by a switch-back, 2^ miles in length, with grades of 293 to 340 feet to the mile. In 1878, the Atchison, Topeka & Santa F^ Rail- way crossed the Raton Summit by a switch-back, 3^ miles in length, with six switches, on a maximum grade of 316 feet to the mile, with Mogul locomotives weighing 110,000 pounds. Effect of Curvature In balancing the relative effect of gradient and curvature upon the assumed normal train-load, the maximum permissible degree of curvature should not coincide with the ruling-gradient. The resistance due to curves alone, under ordinary ionditions, is based on the assumption that each 1 Classification op Usual Gradients Heavy . Moderate Easy English lin 100 1 in 200 1 in 400 American 1 per cent, or 52.8 feet to the mile 0.6 per cent, or 26.4 feet to the mile 0.25 per cent, or 13.2 feet to the mile ''-Between Philadelphia and Harrisburg, on the Pennsylvania Railroad, the old line constructed by the State is crossed every half-mile, for long stretches, by the new line which, though never more than a few hundred feet away, has hardly one-tenth of its curvature. " Economic Theory of Location of Railways," A. M. Wellington. ' An extended discussion of the relative ef&cieney of inclined planes, spirals and switch-backs will be found in "Railway Location," by A. M. Wellington, 1915, Chapter XX. * See Appendix IV, Table IX. ROADWAY 147 degree of curvature equals a straight gradient of 1.5 feet to the mile. Curves of six to ten degrees do not Umit the speed of fast trains nor mate- rially lessen train-loads, but a tangent of not less than four hundred feet should be secured between such curves for easing the entrance and de- parture of trains by the intervention of transition-curves. On a level line, consolidation locomotives take 80 or 90 cars around 8 and 10 degree curves at 15 miles an hour. They are operated with ease around 14 to 16 degree curves and are in general use on roads with even heavier curves. Transi- tion or easement curves are introduced between heavier curves by extending the curvature farther back on the tangent at a gradually decreasing rate.^ The average horizontal alignment of roads in the Mississippi Valley and in the Southern States is comparatively straighter than in the East. The percentage of level hne is, however, somewhat greater in the East. In the Rocky Mountain region, the Central Pacific line of 872 miles averages 52 per cent, curvature per mile, of which 105 miles average 151 degrees. On the Colorado Central Railroad of 34 miles, the average curvature is 420 degrees per mile.^ On the early English lines, there were no curves of more than one degree, but subsequently curves of two and three degrees were used. In Holland, of a total mileage of 945 miles, 62 per cent, is level and but 27 miles are on gradients from 0.5 to 1.5 per cent. In Germany, 25 per cent, of the mileage is between these gradients and a little over 25 per cent, is curved ; while in Norway, with 50 per cent, of curved hne, 37.5 per cent, is between the same gradients. ' Simple curves are described with, a single radius ; compound curves, with two or more radii. Reverse curves are curves of contrary flexure, usually separated by a short tangent. In England, the radius is expressed in chains of 66 feet; in the United States, by the angle subtended by a chord of 100 feet, which is the length of the chain used in this country. Curves of over 8° are usually run in with 50-foot chords and those over 16° with 25-foot chords. The radius of a 1° curve is 5730 feet or nearly 87 English chains. The radius of any sharper curve may be obtained by dividing 5730 by the degree of curvature. Tliis is correct up to an 8° curve. The equation of gradients per degree of curvature varies with the character of the traffic; 0.02 per cent, being commonly used for light curves and for freight-tracks where the speed is slow, and 0.05 per cent, for very sharp curves. For table of curve-equations, see "Railway Location," Wellington, page 652. ^ AvEBAGE Alignment op Railroads in the United States in 1880 No. OF Lines Miles op Line CcEVATUnE Per Cent. Level Line Region Curves per Mile Per Cent, of Curvature Degrees per Mile Eastern . . Western . . Southern . . 99 49 17 5372 8558. 3511 1.88 0.78 1.10 35.5 16.9 27.6 55.9 16.9 31.5 22.8 21.4 22.0 'Railway Location," A. M. Wellington. 148 EFFICIENT RAILWAY OPERATION An illustration of original location in conformity with the principles of economic operation is afforded in the extension of the Lehigh Valley Rail- road through New Jersey from Phillipsburg to Perth Amboy, under Robert H. Sayre, Chief Engineer. Although this hne incidentally furnishes an entrance into Jersey City, the chief purpose in its construction was to pro- vide for coal-traffic to tidewater. This purpose is accomplished by virtu- ally concentrating all intermediate differences of elevation at the summit in Musconetcong Tunnel, 12.5 miles from Phillipsburg, on an ascending grade of 22.0 feet per mile, thence descending for 6.5 miles on a 47.5 feet per mile ruling-gradient to a point 150 feet above sea-level. The remain- ing 40 miles to tidewater is virtually a gentle descent, broken by slight changes of grade at railroad crossings. As a consequence, unbroken trains of 50 cars, weighing gross 3300 tons, are handled to tidewater over an intervening elevation of 255 feet, assisted by only one pusher for 12 miles from PhiUipsburg to the summit. Trains of 70 empty cars, weighing 1260 tons, are assisted for 7 miles to the summit by a single pusher. The coal- traffic over this line for the years 1910-1914, has averaged annually2,131,525 tons.i Railway Construction Though the railway, or railroad, has given its name to the system of transportation that now dominates all traffic by land, its inception ante- dates the advent of that system by two hundred years, originating merely as an improved road-surface, just as macadamizing superseded stone-paved highways ; and it is only of late years that it is really becoming a more integral part of the roadway itself. In fact, the railway proper is still superimposed upon the highway, and so the two may be separated in a discussion of railway efficiency. Thus disassociated, as superstructure and substructure, the latter is really the "permanent way," for there is little permanency in the superstructure. Ballast, timber, rails and fast- enings are all transient elements of the superstructure, and, from this point of view, attention may first be given to the substructure or "per- manent way." In England, the permanent way was modeled upon the construction methods developed by Telford and by Macadam in the early part of the » Lehigh Valley Railroad. — Phillipsburg to Perth Amboy Elevation above Chanoes of Miles Total Sea-level. Feet FiT.EVATION Phillipsburg . . . 217.0 Summit .... 472.0 +255.0 12.38 Lansdowne . . . 181.3 -290.7 6.52 18.90 Bound Brook . . 31.4 -149.9 24.60 43.50 Raritan Junction . 98.0 + 66.6 13.30 56.80 Perth Amboy . . 31.5 - 66.5 2.80 59.60 ROADWAY 149 nineteenth century. Between 1818 and 1829, over one thousand miles of turnpike road of this character had been built in England. These roads rivaled and surpassed the ancient Roman roads in sohdity, and in the design and execution of viaducts and auxiliary structures. Their align- ment and gradients were skillfully adapted to the topographical contour of the environment, and, in aU these respects, the experience of the highway engineers of that period was equally valuable when directed to railway construction. With ample capital at their disposal and with growing experience, the English railway engineers undertook works of increasing magnitude, in which they developed improved and novel methods of construction and design. The railway alignment was adapted to mechanical traction by easier curvature and by Hghter gradients, sometimes at great expense for heavy earthwork, viaducts and tunnels. Because of the attention given to these matters in the original construction of EngUsh' railways, but Httle alteration in their permanent way has since been necessary to meet the increased requirements of traffic, and the same thoroughness of execution marked aU of the accessory structures. From time immemorial, the excavation and embankment required to bring the surface of a highway to the desired grade have been performed by manual labor, long assisted only by the pick, shovel and hand-barrow. In the fifteenth century, the labor of this character was somewhat lightened by the invention of the wheelbarrow, which continued to be the main reliance for short hauls until it was supplanted, under favorable conditions, by the horse-shovel or scoop. For longer hauls, the contractor's track-equipment has replaced the dump-cart ; but the chief appliance in expediting earth- work, and in reducing its cost, is the steam-shovel. By its superior capac- ity and speed of operation, it is now practicable to lengthen the average haul and the consequent balancing of earthwork between cuts and fills, before resorting to borrowing or wasting material. A man shoveling gravel or light soil ought to handle about 10 cubic yards in a day of ten hours. An ordinary railroad steam-shovel of 70 to 90 tons, under similar condi- tions, should handle at least 1000 cubic yards, at a cost of about $50.00. Compare this result of mechanical energy with that of vital energy.^ Gunpowder, invented for destructive purposes in warfare, has been equally powerful for similar purposes in highway construction. Gradually, it has been superseded by blasting powder and by other explosives better suited for work of this character. The slow process of drilling blast-holes by hand has been displaced by appliances operated by steam, by electricity or by compressed air, which have served greatly to diminish the amount ' Capacity OP Excavating Appliances. Barrow, 2 cu. ft. Drag scraper : No. 1 5^ cu. ft. No. 2 — 4J cu. ft. No. 3 — 3| eu. ft. Wheeled scraper: No. — 7 eu. ft. No. 1—9 cu. ft. No. 2—12 cu. ft. No. 2J — 14 ou. ft. Steam-shovel bucket, 1 cu. yd. (minimum). 150 EFFICIENT RAILWAY OPERATION ' of manual labor and to shorten the period of execution where work is carried on in material too hard to be broken up either by the pick or by the steam-shovel. Earthwork construction often involves operations of great magnitude, as to quantities of material to be moved, yet, save in the particulars just mentioned; it affords but little opportunity for advance in engineering effi- ciency. A road-bed built with due regard to the natural slope of the mate- rials of which it is composed is virtually imperishable, so long as its slopes are protected by careful sodding or by other suitable precautions, and its foundation is kept secure by adequate drainage. Much difficulty is experienced at times in obtaining a stable founda- tion for an embankment on treacherous soil. A notable instance occurred in the construction of the earliest raUway intended for general traffic, the Liverpool & Manchester Railway. The line, as located, crossed a boggy tract, known as the Chat Moss, for four-and-a-half miles. The foundation proved to be of saturated, peaty matter from ten to thirty feet in depth, resting on a clay and sand subsoil. An embankment of 277,000 cubic yards in content consumed 670,000 yards of material and was only com- pleted by virtually floating it on the bog. Such treacherous foundations are sometimes encountered unexpectedly. In one instance, the line ran rather diagonally for some distance across an open marsh, through which there trickled an insignificant stream. As the embankment was extended, it settled so much that a temporary trestle- work was built to carry on the work. At one point, a pile went out of sight at the first blow, and the hammer with it. Two pilings, each sixty feet in length and doweled together were required to reach a solid foundation. As the work went on, it appeared that the meadow occupied the site of what had been a deep water-cpurse, whose meanderings crossed the line in several places. After losing a large quantity of filling, recourse was had to a floating foundation of long saw-mill slabs, crossed in alternating layers, upon which the material had to be carefully distributed to prevent it from breaking through the natural surface and capsizing. Intercepted water-courses must be passed through the road-bed, making the openings of ample cross-section, to provide for abnormal floods. Across small streams and in low embankments, the culverts may be left open, or, in high banks, arched over with masonry, constructed in advance of the earthwork. In some cases, tubes of heavy earthenware may be ad- vantageously employed, in sections alternately inserted in each other, so that the water may flow over and not against the inner joints. At open culverts, the ends of the bank should be carefully protected against wash-outs, the upper ends of covered culverts protected against seep- age through the embankment, and their lower ends against undermining by the outfall of flowing water. Such culverts may also be constructed of iron tubes, plain or corrugated. ROADWAY 151 Viaducts and Trestles If the grade-line be projected above the natural surface beyond the height within which earthwork may be economically employed, resort is had to viaducts. At an early period in ancient history, such structures served a useful purpose in providing water for fortified places. The several aqueducts which supplied Rome at the height of its power, still fulfill that object to some extent, or stretch in ruined arches for miles across the Campagna. Elsewhere throughout the Empire, similar monuments bear witness to the skill of Roman engineers.^ Viaducts were unnecessary while traffic was borne by porters or by pack-animals ; it was only after wheeled vehicles came into general use that excessive gradients were found objec- tionable in highway building, and more especially when engineering prac- tice was directed to railway construction. In viaduct-work, as distinguished from bridge-building, the length of the spans is governed by an adjustment of the balance between the relative cost of the foundations, piers and connecting-superstructure, according to the character of the available building-material and the nature of the soil. Any advantage that may be derived from increasing the height of the grade-line, is but little diminished on account of the accompanying cost for increasing the length of the supporting-piers. The cost of the abut- ments and of the floor-system is practically independent of the length of span. The cost of the piers depends mainly on the character of the soil and on the height of the grade-Une. In a viaduct with many arches of long span, large quantities of material are required to fill in the spandrels in order to resist the tendency to deformation at the haunches. As masonry-work of this kind fulfills no structural purpose, it is more economical to sub- stitute trussed girders on piers. For any given load and type of super- structure, the cost of the girders for one span varies nearly as the square of the span-length, and the total cost is least where the cost of one pier equals the cost (erected) of the main girders over one span. The simplest type of a railway viaduct, the wooden trestlework, was generally adopted in early construction in the United States, wherever ' timber was abundant. The bents, on a low grade-Une, might be merely two piles, capped and placed ten feet apart, carrying simple track-stringers extending aver two bents and jointed on alternate bents, the whole struc- ture connected by mortises and tenons, pinned with draw-bore. The stringers were jogged vertically and pinned to the caps. On a grade-line up to ten or fifteen feet above the natural surface, the bent was often temporarily of three piles, braced by planks spiked diagonally across ' Pont du Gard. — Aqueduct carried for 820 feet on three tiers of arches, the two lower tiers from 60 to 75 feet span. Segovia. — Aqueduct 2410 feet lo^, 109 arches in two tiers, 102 feet high. Mamz. — 2100 feet long, on between 500 and 600 piers. Antioch. — 700 feet long and 200 feet high. 152 EFFICIENT RAILWAY OPERATION them. When the piles began to decay, they were cut off below the line of permanent moisture and a framed bent was then erected upon them. On a dry soil, the framed bent often rested on mud-sills. On a still higher grade-line, the bents assumed the character of piers, spaced farther apart and carrying trussed stringers. Very lofty viaducts so constructed became pyramidal structures of framed timber, and others of similar design were subsequently built of iron beams and columns.' In any system of trestling over ten feet in height, the bents should be braced transversely and diagonally in each additional height of ten feet, as well as transversely in the floor-system and by longitudinal braces or walings. The floor-system should be substantially constructed, with heavy guard-timbers lined with iron plates on the track-side. Safety in train- service would be further insured by placing a re-railing device at each end of every viaduct and bridge with an open floor. Iron viaducts are usually spaced in 30-foot spans, with double bents braced in pairs ; though they are also built in spans of 45 to 60 feet, to suit local conditions. The cost per linear foot is thereby but little affected, as a heavier floor-system is required with increasing length of span. One of the earlier examples in Europe is at Freiburg, Switzerland, built in 1862 and still serviceable. It is double-track, 225 feet in height and 1100 feet in length. The early American viaducts of iron were of slighter construc- tion. Several have proved insufficient for the increasing service to which they were subjected, and have been replaced by more substantial struc- tures. The Kinzua Viaduct, on the Bradford division of the Erie Railroad, was single-track, built in 1880-1881, 2053 feet in length ; and was composed of twenty towers, varying in height from 30 to 285 feet, from top of masonry to top of ties, with a batter of one to six. On account of its height and manner of construction, it vibrated under the passage of trains to such an extent that speed over it was restricted to five miles an hour. It was replaced, in 1901, by a structure of somewhat novel design, buUt upon the original masonry. The details and dimensions are the same in all the towers from the top downward. The height of the stories is 62 feet, sub- divided longitudinally by bracing, but without diagonal bracing within the tower. The latticed columns, composed of plates and angles, measure 37 inches, transversely. The intermediate transverse struts are box lat- tice-girders, respectively 4, 6, 7 and 8 feet in depth. The longitudinal diagonals are built of two lattice-channels, connected by diaphragms at their intersections and with the columns. The only longitudinal struts 1 The Portage Viaduct carried the Erie Raihoad over the Genesee River at a height of 250 feet, across a chasm 900 feet wide, in spans of 50 feet ; and was built in two years of 16,000,000 feet of timber at a cost of $175,000. It was opened on August 9, 1852, destroyed by fire in 1875, and replaced in forty-seven days by a bridge of steel. "When Raihoads were New," C. F. Carter. ROADWAY 153 are at the bottom of the tower. The tower-legs are bolted to the masonry by two 1^-inch bolts. The floor is carried by plate-girder deck-spans, 9 feet between centers ; upon each tower they are 38 feet 6 inches in length and 4 feet 6 inches in depth. The intermediate spans of 61 feet are 6 feet 6 inches in depth. The total weight of the deck-spans is 638 tons ; and of the towers, 2715 tons. Expansion is provided for by twelve freely-moving girder-ends on friction-rollers. The change in length of the track-stringers, due to a range of temperature of 75° F., is about ten inches in the length of the structure. The work of erection was carried on from each end by a traveler, spanning 160 feet, having an old tower in the middle. The work consumed four months with 120 men. Viaducts of trestlework, up to twelve and fifteen feet in height and several miles in length, were common where railroads were built through the cypress-swamps that border the streams along the South Atlantic coast. Since the introduction of reinforced concrete, these timber-struc- tures are being in many instances replaced by an iron superstructure on concrete piers, or by a series of low arches carrying an embankment. Via- ducts of imposing dimensions are also constructed of reinforced concrete.' The viaducts on the Florida East Coast Railroad are of an even more am- bitious character. The extension of this line to Key West is built for some sixty miles along the low-lying range of islets forming the Florida Keys. The numerous intervening channels are crossed by viaducts of reinforced concrete piers and steel superstructure. One of these, the Long Key Via- duct, is 2^ miles in length, with arches of reinforced concrete in 50-foot spans and 30 feet above low tide. In this region, where there is neither building stone nor brick-clay and where the sea is infested with teredo and limnoria, this extension would have been impracticable but for the use of reinforced concrete. Bridges. Early Types Deeper water-courses must be spanned by bridges, and in designing such structures engineering efficiency of the highest order has been displayed. Bridge-buUding was a development of house-building. Where building- stone was abundant, or clay suitable for brick-making, the mason became the bridge-builder, and the arch which supported the wall over door or window openings was a fundamental principle in bridge-construction in 1 The " cut-off " opened for traflftc in November, 1915, between Clark's Sum- mit and Hallstead, Pa., on the Delaware, Lackawanna & Western Railroad, in- cludes two remarkable viaducts of reinforced concrete. Tunkhannock Viaduct, 2.375 feet in length and 242 feet above the bed of the stream, includes ten spans oif 180 feet each and two of 100 feet. The double-track roadway is carried by a series of stilted arches of shorter span, superimposed upon the longer spans. This viaduct contains 167,000 cubic yards of concrete and 1140 tons of steel. The Martin's Creek Viaduct is 1600 feet in length and 150 feet above the bed of the creek. 154 EFFICIENT RAILWAY OPERATION the time of the Roman RepubUc. The economic limit for masonry-arches was about a hundred feet span ; where timber was plentiful, the car- penter replaced the mason, the principle of the rafter or beam superseding that of the arch. In its simplest form, the span of the timber-bridge was limited by the dimensions of the available timber and by the weight of the normal passing load. If the beam was of sufficient length to span the opening, but was of insufficient cross-section to carry the load, engineering genius provided the remedy. The strain at the middle of the girders that supported the floor- beams was partially relieved by erecting a king-post at that point and bracing it against the ends of the girder, thus transferring the strain by compression to the abutments. Where the span was so long as to require a king-post of excessive height to give the bracing the proper angle, an alternative plan was to divide the length of the girder by two queen-posts connected by a tie-beam or strut, thus carrying the strain through the braces to the abutments at the required angle. Still another method was to place the king-post or queen-posts beneath the girder, supported by an iron rod strained over their ends, by which the stress from the passing load was transferred to the abutments by tension instead of by compression. Both of these methods, it will be seen, were borrowed from roof-carpentry. The next advance in timber-construction was to cover a span exceeding the length of a single beam by bolting several together with lapped joints, and then combining the methods of compression and of tension in a single trussed girder. The many designs of this type may be classed either as paneled trusses, in which the stresses are concentrated at widely separated points in the upper beam, or top-chord, and in the lower beam, or bottom- chord ; or else as lattice-girders, composed of many braces and counter- braces bolted or pinned together at every intersection, forming a web which is connected with the chord-members at many points. The trussed girders, with their top-chords and bottom-chords respectively connected by lateral bracing, virtually form a box girder, which may carry the floor- beams either on the bottom-chords, as an overgrade or "through" bridge ; or on the top-chords, as an undergrade or "deck" bridge. In the latter case, the interior of the bridge is stiffened by cross-bracing against swaying under a passing load. Where the grade-line is at a sufficient distance above high-water level, a considerable saving in the height of abutments or piers may be made by carrying the roadway on the top-chord. By connecting the trusses continuously in a bridge of three or more spans, a theoretical advantage may be obtained of 49 per cent, in dead weight and of 16 per cent, in the live load. Timber-arches were used as early as 104 a.d. in the bridge built over the Danube by order of Trajan and, in the longer spans of later design, the increased stresses were resisted by a combination of panel trusses with arches built of beams. By the middle of the eighteenth century, the art of ROADWAY 155 building truss-bridges had made great advances in Europe. A bridge was then built over the Rhine at Schaffhausen with spans of 172 and 193 feet. It was constructed in panels with vertical posts, the stresses at the middle of the top-chords being transferred by long braces and connecting tie- beams through the end-panels to the piers and abutments. The same engineers built the Wettingen Bridge of 390 feet span, which was the longest timber-span ever constructed. As suitable timber became scarcer and the supply of iron ampler, the principles of construction, developed by working in timber and stone, were applied to the use of metal also. Wrought-iron was used in combination with timber, or with cast-iron posts or struts, and arched bridges were built, principally of cast-iron.* The superior efficiency of wrought-iron over timber, when used in ten- sion, was more fully exhibited in connection with an entirely different prin- ciple of construction, — that of suspension. This principle had been put in practice anterior to historical dates in the rope-ferry and for foot-bridges over mountain-gorges. In the latter part of the eighteenth century, it was applied to highway-bridges ; the roadway being suspended from parallel chains stretched over the span, strained at the abutments over lofty piers and anchored in masses of masonry. Wire cables were subsequently sub- stituted for chains. In 1810, Telford built a bridge over the Menai Strait of 570 feet span, which was suspended by iron bars linked together. This was followed by a suspension bridge of 870 feet span at Freiburg, Switzer- land. Such was the state of the art of bridge-building when it was applied to railway construction in England. Arched bridges continued to be used for moderate spans under a high grade-line, but, where a lower grade-line left no room beneath it for an arched bridge, the opening was spanned by girders on piers. Long spans in highway bridges were practicable because the normal loads bore so light a proportion to the weight of the bridge itself, that their added strains were a negligible element in a bridge-design. But, with the development of railway transportation and the demand for heavier loads at high speeds, this was no longer the case, and engineering skill was called upon to meet this exigency. Ttjbulak Bridges In 1845, a plan was required for carrying the Chester & Holyhead Rail- way over the Menai Strait, already crossed by Telford's highway bridge of 570 feet span. The work was intrusted to Robert Stephenson and WiUiam Fairbairn. The coefficients of the strength of materials com- • Early oast-iron bridges : Coalbrookdale, over the Severn, 1773-1779, 100-foot spans of five cast-iron webs ; still in use. Wearmouth, 1793-1796, arches of cast- iron voussoirs. Southwark, over the Thames, 1814-1819, center span of 241 feet and rise of 24 feet, of cast-iron ribs. Paris, over the Seine, 1800-1806, and 1834- 1836. 156 EFFICIENT RAILWAY OPERATION monly used at that time were largely empirical, as were also the formulas in which were expressed the inherent stresses in structures and the occasional ones to which they are subjected from moving loads. The principle of construction that was adopted was that of the girder. Girder-bridges of short spans had already been constructed of iron in the form of a box, and the box-girder type was made the object of ex- periments in which the proper functions of the top and the bottom- chords of a girder and of the web connecting them were for the first time definitely determined and reduced to formulas by the eminent physicist, Eaton Hodgkinson. As a result of these experiments, the Conway Bridge was designed as a pair of box girders of rectangular section, each large enough for the pas- sage of a train through it. The top and bottom were of cellular construc- tion, connected by sides of heavy plates stiffened by ribs and gussets. This bridge gave such satisfaction that the design was repeated in the Britannia Bridge, over the Menai Strait, as a pair of continuous tubular girders.^ The material was assembled on the adjacent shore ; the girders were floated to the bridge-site and raised into position by hydraulic appli- ances. Soon after the completion of the Britannia Bridge, Isambard K. Brunei built the Saltash Bridge, near Plymouth. In the principal spans, the top- chord was a single " lenticular " arched tube, and each bottom-chord was a pair of chains composed of pin-connected links, suspended by hangers from the outer edges of the wide top-chord. The trusses from which the road- way was suspended were therefore parabolic in form, the idea being so to dispose the material as to offer the greatest resistance to the strains to which it was to be subjected, with the least practicable dead weight.^ In designing this bridge, Brunei displayed that independence of precedent which distinguished his genius also in the adoption of a seven-foot gauge for the Great Western Railway, and in the construction of the steamship Great Eastern. The Britannia Bridge marked the advance in engineering eflSiciency, due to the demand for heavier loads at higher speeds in railway trans- portation, which raised bridge-building from an art to a science. Yet, notwithstanding the recognized ability of the engineers who planned it, , the tubular girder has only been repeated, by Stephenson himself, in the 1 The Britannia Bridge, completed in 1850, was composed of two clear spans of 460 feet each, and two of 230 feet each, 104 feet above high water. The tubular girders were 1511 feet in length, 15 feet wide, 23 feet deep at the ends and 30 feet at the center ; each girder weighed 4680 tons. Forty per cent, of the total weight of the girders was in their stiffened sides. Proportion of depth to span, 1 in 16. Encyclopsedia Britannica. ^ Saltash Bridge : Two spans of 455 feet each, and seventeen shorter spans. Top-chord, a lenticular tube, 17 feet wide, and 12 feet deep. Bottom-chord, of 14 links in each chain of a section 1 by 17 inches. Eno. Brit. ROADWAY 157 Victoria Bridge at Montreal ; ' while the Saltash Bridge remains the unique example of its type. With further experience, the trussed girder was found to afford better opportunity for economical disposition of materials, for faciUty in construction and erection and for efficient distribution of stresses among its members. This conclusion was greatly strengthened by results attained in the United States. 'Trussed Bridges. The Steel Arch At an early period, bridge-builders in the United States were skillful in the construction of trussed girders. The favorite type was the lattice- truss. It was simple in design ; it required no timbers of unusual dimen- sions, and could be put together without iron-work and framed by ordinary house-carpenters. By increasing the height of the trusses and by combin- ing arches with them, these lattice-bridges were built of considerable spans, but for railroad purposes, the span rarely exceeded 125 feet.'' It was impracticable to equalize the stresses between the truss and the arch and, with the increasing loads, the pin-connections were compressed and bent ; so that, under the consequent deflection from passing trains, the integrity of the structure was imperilled. Resort was then had to empirical designs of other types. In 1830, the Howe truss was patented; a panel-truss with vertical rods as tension members and inclined struts or braces as compression members and for counterbracing to resist the wave of deflection from passing trains. In 1840, this type was further improved. At the panel- points, the braces rested against iron angle-blocks, with tubes gained in between the chord-pieces. Through these tubes, the tension-rods passed and were screwed up against wrought-iron gibs on the exterior faces of the chords, thereby furnishing a ready means for restoring any loss of camber. The chords were given greater transverse strength by interposing iron plates to separate the chord-pieces which, at the staggered joints, butted against castings with separating webs provided with ribs that were gained into the continuous pieces at the sides. In 1844, the Pratt truss was introduced. This truss was a development . of queen-post bracing, being composed of two or more upright queen-post systems, combined within one trussed girder as primary, secondary and suc- ceeding structures, supporting each other in transferring stresses from the middle of the girder to its ends. In 1847, Squire Whipple developed the theory of stresses in the members of a truss which was apphed in the 1 Victoria Bridge over St. Lawrence River : 25 spans of 244 feet each, replaced about 1900 by a bridge of modern construction. 2 The Amoskeag Bridge at Manchester, N. H., was built in 1792, with six spang of 92 feet. The Bellows FaUs Bridge over the Connecticut River, 1785- 1792, had two spans of 184 feet each. The Colossus Bridge, over the Schuylkill River, was a flat-arched truss of 340 feet. Enc. Brit. 158 EFFICIENT RAILWAY OPERATION Whipple-Murphy type. This was a queen-post design, somewhat similar to an inverted Pratt truss. In the Warren truss, the struts and ties form alternate equilateral triangles between the chords. A Warren-truss bridge of 390-foot span was built over the Ohio River at Louisville in 1869. An iron bridge of this type was built on the Great Northern Railway in Eng- land.i Up to 1850, bridges built of white pine were limited to about 150- foot spans ; but, as railroads were extended down the Atlantic coast, the yellow-pine forests furnished timber of larger dimensions and capable of resisting greater strains, and the length of spans frequently exceeded this limit. These longer spans were also combined with timber-arches, either attached to the sides of the trusses or cut into the bracing and connected with it by turnbuckles, by which means the strains were adjusted between the trusses and the arches. The conditions that turned the attention of European engineers to iron-bridge construction were likewise operative in the United States, and, between 1847 and 1857, there was a similar tendency to empirical designs, with the compression members of cast-iron and the tension members of wrought-iron.^ The Fink bridge was a suspended girder trussed with an inverted king-post system, beginning with a principal post in the middle of the span and the half-spans successively subdivided at intermediate points of support. The Bollman bridge was of a somewhat similar design. But in the end, the panel-truss with pin-connections superseded them both. As was the case in Great Britain about the same period, there had been little accurate information accumulated as to the variations in the quah- ties of iron employed in bridge-construction, or of the effects of static stresses and of dynamic shocks upon the accepted types of railroad bridges. Even in 1876, a bridge of the Howe-truss type, built over the Ashtabula River with iron rails for compressive members, collapsed beneath a train, because of faulty connections and the crudity of the design as to details. With the advance in theoretical knowledge, there was an accompanying improvement in structural specifications that was furthered by the en- trance into iron-bridge building, as general contractors, of iron-works possessing ample capital. These concerns began to furnish their own designs and specifications, which were prepared by proficient engineers. A novel feature in these specifications was their conformity to assumed conditions and lipiitations as to the stresses which they could sustain with safety.' The first specification for concentrated axle-loads appeared in ' Norwich Dyke Bridge, 1851-1853 : 250-foot span. Top-chord, hollow cast- ings. Bottom-chord of •wrought-iron links , pin-connected. 2 In 1852 James I. Shipman, Chief Engineer of the Alton & Sangamon Rail- road, in Illinois, estimated for iron bridges in its construction, and based his esti- mates on the cost of an iron bridge built by Squire Whipple across the Canal Basin at Albany. George T. Hammond, "Discussion on Railway Development," Trans Am. Soc. C. E., Dec, 1911. ' A specification for short spans, prepared in 1871, limited the load to two tons ROADWAY 159 1875. The principal requirement, as to wrought-iron, was that it should have a tensile strength of 66,000 pounds per square inch. Experimental tests afterward reduced this maximum to 52,000 pounds with an elastic limit. of 26,000 pounds. The introduction of rails of Bessemer steel into the United States was followed by the use of that material in the Eads Bridge, over the Mississippi River, at St. Louis ; completed in 1874. This bridge has three arched spans of unique design. Each span consists of four double ribs, in which the voussoirs are composed of hexagonal steel bars, clamped together in tubes by wrought-iron coupUngs.^ For fifteen years afterward, steel was only used for large members, as for chord-bars. The first large all-steel truss-bridge in the world was built by General William Sooy Smith over the Mississippi River at Glasgow, for the Chicago & Alton Railroad Company. At that time, there was an adverse opinion among engineers as to such use of steel on account of its erratic behavior when subjected to shock and vibration.^ It was not untiL1890 that the use of steel for all bridge-work had become general. Two grades of steel were recognized ; one of soft steel with ultimate tensile strength of 54,000 to 62,000 pounds per square inch, and a medium grade of 62,000 to 70,000 pounds. After the sub- stitution of open-hearth or basic steel for Bessemer or acid steel, but one grade was employed, with a tensile strength of 60,000 pounds per square inch. The adoption of the independent-panel truss for long spans was ad- vanced by the introduction of Bessemer or converted steel as a building material, which made it possible to obtain beams of increased dimensions and of more reliable chemical constitution. The Covington Bridge, built over the Ohio River in 1888, was for some time the longest span of this type.' The construction of arched bridges was similarly affected by the use of steel beams in the trussed ribs. The Roebling bridge over the Niagara having become insufficient for the requirements of increasing traf- fic, it was replaced in 1896-1897 by a steel structure with a center span of 840 feet, the longest span of this type yet constructed.* Recently, the per linear foot, or for 30 tons on driving-wheels within a space of 12 feet, with a loaded train of 20 tons in each 22 feet, proceeding at a speed of 30 miles an hour. M. L. Byers, Proceedings International Railway Congress, Berne, 1910. Vol. I, pp. 68-80. See the same, also, for other information about bridge-build- ing in the United States. ' Eads Bridge : Center-span, 520 feet and two side-spans, each 502 feet. Rise of center-arch, 47J feet. Side-arches, 46 feet. Upper and lower members of each rib trussed 12 feet apart. Double-track, and above the tracks a roadway, 54 feet wide. Bnc. Brit. 2 Engineering News, March 30, 1916. ' Covington Bridge : Center-span, 550 feet. Side-spans, 490 feet, 67 feet wide between centers of trusses, 84 feet deep, carrying two railway tracks, two carriage-ways and two footways. Total weight in three spans, 5000 tons. ■•Roebling's Niagara River Bridge, 1852-1855: Four wire-cables, 10-inch diameter, in a span of 821 feet, 245 feet above the river. Box gu-der, 18 feet 160 EFFICIENT RAILWAY OPERATION Chesapeake & Ohio Northern Eailway has been carried over the Ohio River, near Portsmouth, by two riveted-truss spans of 775 feet each, ap- proached on the Kentucky side by a viaduct of 1063 feet, and by one of 823 feet on the Ohio side. Suspension Bridges The suspension principle afforded the opportunity for economical construction of long spans on a high grade-line, without the use of false works ; but their extreme flexibility made suspension bridges sensitive to deflection caused by moving 'loads. For this reason, this principle was not adopted in railway construction until about 1850, when a plan was devised for applying stiffening girders to the chains that supported the roadway, with an estimated saving of three-sevenths of the weight of the girders in an ordinary truss-bridge of equal span and carrying an equal live load. Such a railway bridge was built at Vienna, in 1860, of 264-foot span. Two chains, vertically parallel and four feet apart, were stiffened by bracing ; the truss then forming an inverted arch from which the floor was suspended. American engineering afforded an early example of the suspension principle in the wire-cable bridge over Niagara River, which was at the time the longest span of a railroad bridge in existence. It was stiffened by trusses, chiefly of timber, which formed a box girder, carrying the railroad track on top and a highway below. Iron girders were sub- stituted for timber in 1880. This design of a suspended box girder was repeated by Roebling in the Brooklyn Bridge, built in 1872.' The catenary curve assumed under the effect of gravitation by a wire- cable becomes more parabolic in form in a chain of pin-connected bars, which is also employed in suspended railway bridges. The trussed sus- pension bridge is also used in railway work for very long spans. A bridge on this principle has been designed for crossing the Hudson River at New York City, with a span of 3200 feet.^ But under these conditions, the suspension principle loses much in the way of economy, on account of the loftier piers and more massive anchorages that are required, as well as the increased weight and cost of the suspended trusses. The failure of the Tay Bridge in 1879 exercised an unfavorable influence upon its use for such spans, and led to the abandonment of a similar design for the Forth Bridge. deep and 25 feet wide. Niagara FaUs Bridge, 1896-1897: Center-span, 840 feet. Side-spans, 190 and 210 feet. Weight of center-span, 1629 tons. Two- hinged parabolic arch with trussed ribs, carrying two electric tracks, two roadways and two footways. 1 For bridges in New York City, see Appendix IV, Table VIII. 2 A suspension bridge was planned for carrying the Pennsylvania Railroad into New York City with a span between towers of 3000 feet. The towers were to rise 600 feet above water level, and the bridge was to carry sixteen tracks on two decks. ROADWAY 161 Cantilever Bridges The impetus given to building long spans by the advance in the metal- lurgical and mechanical arts, as well as by the better comprehension of the mathematical theory involved in structural stresses, is evidenced in the apphcation of yet another principle in bridge-construction — that of the cantilever. This principle, which is fundamentally the same as the mechanical power applied in the cant-hook, appears in its primitive form in the strengthening of the simple beam by the projection of a support from the abutment, like the bolster or corbel under the roof-rafter. This device added strength to the beam to resist the stresses to which it was subjected, as they increased in magnitude from the middle of the span. In so doing, there was an inversion of the strains of compression and tension from their original relation to the neutral axis in the beam, con- stituting the contrary curve of flexion which, as in the arch, modifies the thrust from a horizontal to a vertical direction. The trussed girder, continuous over three spans, is an example of this change of thrust, and the point at which it occurs was practically exhib- ited in a test of the bridge built over the Boyne, in Ireland, in 1852-1855.1 By this division of the length of the girder, each of the Side-spans slightly exceeded in weight one-half of the central span. Upon the completion of the bridge, a test was made of the accuracy with which the point had been determined at which the strains of compression and of tension were neu- tralized by their inversion in the curve of contrary flexure. The rivets in the top-chords of the central span were cut at that point, toward one of the intermediate piers, and the shore end of that side-span was lowered an inch with the effect of opening the separated points of the chords by only •j^ of an inch. The same process was repeated in the opposite direction with a similar result, thus proving the correctness of the design. The same principle is also employed in the so-called "hinged-arch," hinged at the points at which the opposing strains of compression and of tension are neutralized ; and with economy in the distribution of mate- rials in the structure. Such a bridge was built over the Viaur River in southern France. The ribs are arched Whipple-trusses, with a center- span of 722 feet, and the roadway is 380 feet above the water-level. The shore-ends are two half-spans, each of which was so combined with a half of the center-span as to slightly overbalance it, and each half of the struc- ture was connected by hinges with its supporting pier. The bridge was erected without false-works, by balancing the two half-spans on each side as the work of erection was extended in each direction toward the middle of the central arch and toward the abutments. From the arms of the cantilever, counterbalancing weights may be 1 This bridge of trussed girders, 22i feet deep by 541 feet in length, is supported by two intermediate piers with a center-span of 264 feet. 162 EFFICIENT RAILWAY OPERATION suspended, and the additional strains resisted by increasing the depth of the compression-member, and by straining truss-rods over a post or tower erected on the pier; or by substituting a trussed girder as the tension- member. In either case, the strains from the suspended load act as they do upon the jib of a crane. By increasing the dimensions of the canti- lever, vertically and horizontally, it may cover a half-span to any length commensurate with the. factors of safety of the materials of which it is composed. The cantilever design has been adopted for many railroad bridgels of long span on a high grade-line. Among these are the Niagara Bridge built in 1883, and the Poughkeepsie Bridge over the Hudson River.^ In- dependent truss-spans may be erected without falseworks by temporarily connecting them over the piers as cantilevers and then building them out- ward until they meet in the middle of the span. Two cantilevers may be opposed to each other with an intervening space covered by a trussed girder span suspended from their opposite arms, forming virtually a con- tinuous girder, hinged at the points of contrary flexure. In many places, this is the most economical plan for a bridge with a long span and on a high grade-line; as the cantilevers can be built out from each pier with- out expensive falseworks or obstruction to navigation. The suspended truss may be assembled on the cantilevers and rolled into place or be raised into place from the water-level. The Forth Bridge was designed on this principle. The shore-arms of the side-cantilevers are loaded to balance half of the weight borne by the opposing arms and two hundred tons besides.^ The Queensboro Bridge, over East River in New York City, has a span of 1182 feet of cantilever type with suspended pin-connected trusses. The bridge over the St. Lawrence River at Quebec^ which collapsed on August 29, 1907, during erection, was the longest span yet attempted of this type. This bridge had been designed by an engineer who was recognized as an authority on the subject, and it had been constructed by one of the leading bridge-contractors in the United States. After an investigation by a commission of experts, another design of the same type was accepted. On September 11, 1916, the suspended truss fell into the river as it was being ' Niagara Cantilever Bridge : Double track, 910 feet in length. Center- span, 495 feet ; subsequently strengthened. Poughkeepsie Bridge, 1886-1887 : 200 feet above water-level. Five river- spans of 547 feet each and three shore-spans of 208 feet each. The trusses over the second and fourth piers are extended as cantilevers over the adjoining spans. The shore-piers carry cantilevers projecting one way over the river-openings and the other way over a shore-span, where they are secured to anchorages. 2 Forth Bridge, 1882-1889, opened 1891: Total length, 5330 feet. Two center-spans, each 1710 feet. Suspended trusses, 350 feet long. Cantilever towers, 260 feet high. Piers, 145 feet high. Six cantilevers, 680 feet each. Double- tracks carried on an internal viaduct of lattice girders. Total weight, 38,000 tons, exclusive of approaches. ROADWAY 163 lifted by hydraulic jacks from floats below into its position between the cantilever arms, apparently from a failure in some part of the lifting ap- pliances. The following year a new suspended truss was successfully lifted to place, completing the bridge.' A fixed span may be extended over the supporting piers in cantilever- arms. There is an example of this design in the Harahan Bridge over the Mississippi River at Memphis. To one of these arms is suspended a truss which is supported at the other end on a pier.^ Draw-bkidges Where a bridge crosses a navigable stream on a low grade-line, an open- ing must be provided sufl&cient for the passage of watercraft. This opening is spanned by a draw-bridge, which may be operated vertically as a lift- bridge, or horizontally as a swing or turning bridge. The early highway draws were lift-bridges of the bascule type. This contrivance was bor- rowed from the bridges over the moats of feudal fortresses, which were raised by levers resting on the entrance walls and counterweighted within. Modern examples of this type have usually two leaves, meeting when closed in the middle of the opening. The girders revolve on a horizontal axis near the face of each pier, with the counterweighted arms describing an arc in chambers within the piers. The Tower Bridge over the Thames in London is of this type.' Railroad draw-bridges are usually of the swing or horizontally turning ' Original Quebec Bridge : Length between abutments, 3240 feet. Cbannel span, 1800 feet. Suspended truss, 675 feet. Shore cantilever-arms, each 562i feet. Total weight of metal, 32,000 tons. Designed to carry two steam-rail- road tracks, with a highway and an electric railway on each side ; all between the main trusses. New Quebec Bridge : Two piers, each supporting a cantilever with a river- arm of 580 feet, carrying a suspended truss of 640 feet, making a channel-span of 1800 feet. Shore-arms, each 515 feet, outer ends resting on piers connected with one shore by a truss of 140-foot span and, with the other, by a truss of 269 feet. Total length, 3239 feert. Width, 88 feet between trusses. Double-track, with two sidewalks. Contracted to be completed by December 31, 1915, for $8,560,000. Opened for traffic, December 4, 1917. 2 Harahan Bridge, opened July 15, 1916: Total length, 2549 feet, connected with a steel viaduct, 2363 feet in length. Double-track, 14 feet between centers, with highways 14 feet wide on cantilever-brackets outside of the trusses. Clearance inside, 8 feet on each side and 24 feet above base of rail. Channel-span of 621 feet, extended over piers in cantilever-arms of. 186 feet each. Each arm supports a suspended truss of 418 feet. One of these is supported at the other end by a pier which also supports a deck-span of 347 feet. The other suspended truss is also suspended to a cantilever with arms of 186 feet, making a span of 790 feet ; the other cantilever arm feeing anchored. As so constructed, there are succes- sively a deck-span of 347 feet, and three through spans of 607 feet, 621 feet and 790 feet. ,, , , . , ' Tower Bridge : 200 feet clear opemng. Double-bascule, rotatmg through an angle of 82 degrees, on steel shafts 21 inches in diameter and 48 feet long. Lifted in one minute by hydraulic power. Short-arm girder, 62 feet 6 inches in length, carrying 365 tons of counterweights. 164 EFFICIENT RAILWAY OPERATION type. Over narrow openings, as over canals, they may be operated from one of the piers, as was the case with the early "jack-knife" draws. In this style of draw-bridge, each rail is laid without ties on a separate stringer or girder hinged to the pier. The outer ends of the girders, while swinging, are suspended by links from a gallows-frame on the pier and are connected by a parallel-ruler hinge which holds them at gauge when the draw is closed. The suspension-chains or rods are attached to the ends of a cross-arm, which is pivoted at the middle of the parallel-ruler hinge and are lifted by levers operated on the pier. The draw is opened by a rope attached to the outer end of the draw which is manipulated from the end of the pier- guard. In later designs, the draw is opened and closed by winch-operated gear, meshing in an arc fixed to one of the girders. When the draw is open, the girders fold close together at an angle of about 80 degrees to the line of road. The usual type of railroad draw-bridges is the pivot-draw, resting upon a turntable supported by a pier midway in the opening, and leaving a passage on each side. A trussed girder supported in the middle may be so balanced that the strains on each side can be equalized. This is the prin- ciple of the cantilever as operative in the pivoted-turning draw-bridge, of which the bridge over St. Louis River at Duluth is a notable example."^ The weight of such a draw-bridge, including the turning apparatus, is about the same as that of a fixed span of eqi^al length carrying the same load. The first four-track draw-span in America was built about 1897 over the Harlem River, for the New York Central & Hudson River Railroad Company. It has three trusses, 389 feet between end-pins and 26 feet apart, carrying the four tracks on a ballasted floor and weighing 2500 tons. The draw-bridge over Conneaut Harbor, Ohio, on the Bessemer & Lake Erie Railroad, carries four tracks with only two main trusses ; two tracks are between the trusses and one each on an overhang. The floor-system is supported by deep overhead transverse trusses with cantilever extension to the outer ends of the overhangs. There is no center-tower ; the stresses in each truss are brought to one point on the center pier. The, entire dead weight of 1400 tons is carried, when the bridge is open, by a compound box girder on a pintle-bearing without the support of a turntable. This construction was made necessary by the limited height of the tracks above high-water level.^ A reliable locking-apparatus, connected electrically 1 Duluth Bridge : Opening of 500 feet, spanned by a pivot-draw, 58 feet wide, carrying double-tracks between trusses and carrying on each side, on canti- levers, an electric street-raihoad and a footway. Opened in two minutes by electric motors. Ends lifted four inches by electric motors, which also release the latches and raise the rails. 2 Conneaut Draw-bridge, built in 1912, is of the Warren triangular-truss type, 235 feet long, 32 feet 7 inches between centers of trusses and 67 feet in width over all. Height from base of rail to masonry, 5 feet 8 inches, and floor-system 3 feet in depth. Center-bearing of phosphate bronze, 34 inches in diameter. ROADWAY 165 with the track-signals, is indispensable to the uninterrupted passage of trains over a draw-bridge. Otherwise, all trains should be brought to a full stop at least 600 feet from the opening. Recent Examples of Bridge Construction The effects of increase in traffic and of the corresponding increase in the tractive power of locomotives upon the requirements in bridge-con- struction, are shown in the history of the Pittsburgh & Lake Erie Railroad's bridge over the Ohio River at Beaver, Pa. The original bridge, built in 1878, was renewed in 1890 upon the single-track masonry with a structure thought sufficient for many years to come. But, by 1897, the increased tractive power of the locomotives on the line had permitted an increase in train-tonnage of from 1600 to 2500 tons. About 1902, there was a further increase in tractive power with a corresponding increase in train" tonnage to 3500 tons, which was subsequently increased to 4200 tons- With these changes, there was laid a double-track approach to the bridge ; then a single-track gauntlet across it, then four tracks to a point near the bridge, and, finally, it became necessary to reconstruct the bridge itself. The piers were not large enough for double-track, and it was desirable to relieve navigation by removing one of the channel-piers. For these reasons, in 1906, it was decided to build another bridge on a new location, 90 feet in the clear above low water, with a cantilever-span of 769 feet between centers of piers. The location of this span compelled one arm of the south cantilever to be entirely over land, with an approach-span at the north end of the other cantilever. The bridge was designed for two gauntlets. The tracks in each gauntlet have their adjacent rails a foot apart between centers. Both freight-tracks lie toward the center-line of the bridge, 13 feet on centers, and the passenger tracks are 15 feet on centers. The side-clearance is 7^ feet from centers of passenger tracks to the trusses and 7 feet to the skid-girders ; and the overhead clearance is 21^ feet above the rails. The length of the anchor-arms is 320 feet, of the cantilever-arms 252 feet, of the suspended truss 285 feet, and of the approach-span 370 feet. The trusses are 34^ feet between centers and 30 feet in the clear. The river-clearance and grade-conditions made it necessary to have a shallow floor-system with only 6 feet between the clearance-line and the top of the rail. A special feature of this construction is the protection against derailment. There are nine lines of stringers, four under each gauntlet and one half-way between. The ties are of white oak, chemically treated. Maximum load on overhangs is 6100 pounds per lineal foot of track. The bridge is turned by two pairs of pinions, connected by equalizers, on a rack with pitch diameter of 41 feet. "A Four-track, Center-bearing Railroad Draw-Span." L. H. Shoemaker, Trans. Am. Soc. C. E., December, 1912. 166 EFFICIENT RAILWAY OPERATION 7 by 9 inches, 12 feet long, placed on edge and sized to 8^ inches. They are three inches apart, with three lines of spacing-blocks on each half of the floor. The surface of the ties is 1^ inches above the tops of the floor- beams and ^inch away from them, in order to allow cinders to be blown off the floor-beams by passing trains. Wooden guards could not be used on account of the steel-car construction with drop-doors ; so, beside the gauntlet tracks, there are four inner guard-rails, making twelvB lines of 100-pound rails. The rails are 33 feet long, supported on double-shoulder tie-plates secured to the ties by screw-spikes. The splices allow f-inch expansion at each joint. Skid-girders prevent the car-bodies from striking the trusses in case of derailment. There are two girders for each truss, respectively four and six feet above the ties, with chords on each side of the truss made up of 6" X 6" angles and web-plates — 12|" X i" plates on the inside and lOi" X i" on the outside. These chords are 5 feet 4-|- inches apart and are connected by 3^" X 3^" angles. The track-faces are smooth, with counter- sunk riveting and are provided with expansion-joints. The floor-system is designed to carry a train on each track, weighing 6000 pounds per linear foot, each preceded by two locomotives weighing 426,000 pounds each. The dead load averages about nine tons per linear foot. The wind-loads were computed at 300 pounds pressure per linear foot on a train and 30 pounds per square foot on exposed surfaces of trusses and floors ; 10 pounds per square foot being treated as uniformly dis- tributed, and 20 pounds as a moving load, wherever it produced the maxi- mum effect on the members. The total weight of materials is 6184 tons. The bridge wag tested with eight locomotives and twenty loaded cars in two trains, one for each track, weighing on each track 2,842,200 pounds, placed side by side ; the trains being entirely in the cantilever arm and the suspended truss. There was a maximum deflection of 0.467 inch on the cantilever and suspended truss and of 0.069 inch on the anchor-span. The bridge was completed in May, 1910.i A bridge recently constructed over the Ohio River between Metropolis, 111., and Paducah, Ky., has a channel-span of 723 feet between centers of piers, which is probably the longest single span in the world by fifty feet.2 1 "The Pittsburgh & Lake Erie Raih:oadv Cantilever Bridge over the Ohio River at Beaver, Pa." A. R. Raymer, Trans. Am. See. C. E., September, 1911. 2 Bridge over Ohio River at MetropoHs, Illinois : Double-track. Channel-span on south side of the river of 723. feet, four spans of 655 feet each and a span of 304 feet on the north side. South of channel-span, a deck- truss of 250 feet. A via- duct-approach on each end, composed of tower-spans of 30 feet, with intervening clear spans of 65 to 90 feet. Total length, 5442 feet. Pin-connected trusses. Track 113 feet above low-water mark and 52 feet clear head-room above high water. Railway Age Gazette, July 23, 1915. ROADWAY 167 General Principles of Bridge Design The preparation of a bridge-plan involves primarily a consideration of the weight and speed of the normal moving load or live load. The axle-loading and spacing of the heaviest locomotive which would probably be operated upon the line in question are used in the computation of stresses in the various members.^ The height of the grade-line above the natural surface of the opening to be crossed, the width and depth of the water beneath and the character of the underlying soil, are important factors in determining the division of the superstructure into spans, within the physical limits of safety in the materials to be used, and also in deciding upon the type of superstructure. As the span increases, the effect of the live load becomes more negligible, until the point is reached at which the dead weight of the superstructure itself restricts any further length of span. The superstructure must be designed with_ reference to wind- pressure, and the effect upon the fioor-system of the wave of deflection and the shock from impact of passing trains, and must provide for expan- sion in all its parts under changes of temperature. The length of span is proportional to the depth of the truss, which may vary from one-eighth to one-twelfth of the span.'' Because of improvements in the state of the metallurgical arts, in the control of the chemical elements, and of the molecular constitution of iron and its heat treatment, as also in the tools and appliances for fitting, handling and erecting heavy bridge-members, structures once imprac- ticable are now designed with closer approach to theoretical requirements ; there is also a general agreement about the chemical and physical qualities of the materials required, and as to the tests for their determination. Each member has a definite function and is adjusted to a calculated stress. It has therefore become possible to reduce the empirical factors of safety in bridge-construction with material advantage as to cost and with greater economical efficiency. American practice is characterized by types easy of construction and 1 The live load in early practice was estimated at one ton per foot of line, but the weight of a locomotive of the Consolidation type corresponds to a load of 2§ tons per linear foot. The total weight of a bridge when loaded was at first limited to five tons per square inch of section of truss-members. This limit is now raised for steel structures to nine tons per square inch ; the live load being estimated at twice its actual weight. ^ For parallel steel chords, one-tenth of the span, the theoretical limit is 1070 feet. For parabolic or bowstring chords, one-eighth of span, 1280 feet. For flexible suspension bridges of linked bars, with depth one-twentieth of span, 2800 feet. For stiffened wire suspension bridges, depth one-tenth of span, 2700 to 3600 feet, and for one-eighth of span, 3250 to 4250 feet, aocordiag to the assumed factor of ss|.fety. In practice, the limits as to length of span are less than here given. 168 EFFICIENT RAILWAY OPERATION erection. There is a preference for trusses of great depth, with long mem- bers of rather small cross-section, giving an appearance of lightness, and for the use of eye-bars for tensile members. For spans up to fifty feet, riveted plate-girders are favored and riveted trusses up to seventy-feet span.i For longer spans, pin-connected trusses are more generally pre- ferred here than by European engineers. It is customary to prepare a general plan, designed upon one of the accepted types, with specifications, strain-sheets and drawings sufficiently in detail to secure to bidders an understanding of the requirements. Much of the structural material is designed of standard dimensions and is carried in stock by the manufac- turers. Wide discretion in the details of construction and manner of erec- tion is given to the contractors, who are in general provided with competent engineers and experienced workmen, and are well equipped with tools and appliances suitable for their work. Under these circumstances, economy, accuracy and expedition in bridge-construction has been greatly furthered. European engineers usually individualize in bridge-planning, even to details of manufacture. They favor riveted construction and shallower depth of truss, involving an increase in the sections of various members and a general increase in the weight of a bridge. The two schools are, however, over-lapping. Eye-bars are becoming more common in Euro- pean design, and riveted bridgework is increasing in America. With increasing axle-loads and higher speeds, the effect of impact from dynamic shocks has assumed greater importance. In earlier prac- tice, impact-effect was rather a matter of empiric assumption. Recent experimental tests have tended to enlarge our knowledge in this respect. It is now considered that the more serious effect of impact is principally due to imperfect counterbalancing of the driving-wheels. The maximum effect appears to be about 50 per cent, in spans not exceeding fifty feet, decreasing with the length of span until, for spans over a hundred feet, the maximum effect may be assumed at 15 per cent. It is especially cumula- tive where the rate of revolution of the driving-wheels synchronizes with the normal rate of vibration in the structure, but it is negligible at speeds under fifteen miles an hour.^ The increase in axle-loads and in train-loads, as well as in speed, has encroached upon the factors of safety in many of the earUer bridges to a point at which it becomes necessary to replace them, unless they can be strengthened where they are critically weak. To renovate such bridges 1 The longest single-span girder in use is on the New York, Chicago & St. Louis Railroad ; 130 feet by 9 feet 8J inches in depth. To shield the floor against blast-action and smoke-gases from locomotives passing beneath it, the under side of the floor is protected by a cement-gun coating and by cast-iron blast-plates, three feet wide, hung directly over the center-lines of the tracks. Engineering News, December 9, 1915. ^^M. L. Byers, Proceedings International Railway Congress, Berne, 1910. Vol. I, Sec. II, p. 72. ROADWAY 169 with assurance of safety, required a knowledge of their weak points which could only be obtained by computation of the additional stresses, and by experimental tests as to the points at which these stresses had gained upon the original factors of safety. There was further required an approximately accurate knowledge of each bridge in detail, from a careful examination of the condition of its several members. Where the stresses are concen- trated at pin-connections, the normal factor of safety may have been diminished by wear to an extent difficult of detection from casual obser- vation. The riveted truss, in which the stress is distributed among the rivets, has fared better under increased loads. But in cases of short spans, it has been found preferable to replace these with riveted plate-girders. The floor-system usually exhibits the first symptoms of distress from overloading. Other serious defects are' made apparent by loose rivets, by streaks of rust and by cracked paint and generally by excessive de- flection and vibration under passing trains. There is ordinarily an igno- rance of the history of the earlier bridges, as to the quahties of the materials of construction and as to the manner of erection that overshadows with so much doubt any scheme for strengthening them, as to render it advisable to replace such bridges with structures of known conformity with the changed requirements. Even with bridges of more recent construction and of a magnitude requiring the best knowledge attainable as, to their design and materials, the demands as to speed and loading have so far exceeded what has been recognized as normal, as to call for extensive alterations to meet the changed conditions. The bridges over the East River in New York City are cases in point. Within twenty-six years after the opening of the Brooklyn Bridge, it was estimated that an expenditure of not less than $15,000,000 would be required for their partial reconstruction. Be- sides strengthening the details of the Brooklyn Bridge, the number of Rapid Transit trains permitted upon it at the same time has been mate- rially reduced, and another deck is to be added to the middle span. The Williamsburg Bridge, opened in 1903, has been strengthened in the main trusses, and supporting piers placed under each land-span and each end of the main span. The end-pins, 10 inches in diameter, have been replaced by others, 13 inches in diameter. As to the Queensboro Bridge, opened in 1909, a board of engineers reported in 1908, that all Elevated Railroad and Subway trains should be excluded, the live load reduced one-half on the roadway and one-third on the sidewalks, and the dead weight reduced by 2000 pounds per linear foot. In 1914, it was proposed to reconstruct the floor-plan by placing the trolley-tracks outside of the main trusses on the lower deck, and the sidewalks outside on the upper deck ; one-half of the width between the main trusses on the lower deck to be utilized for Subway trains, and half of the upper deck for Elevated Railroad trains. On the Manhattan Bridge, completed in 1910, it has only been found 170 EFFICIENT RAILWAY OPERATION necessary to shift the anchorages about three inches to provide for the wider car-bodies of the Subway trains. Bridge Foundations. The Pneumatic Peocess Progressive improvement in the arts connected with bridge-building , has been as conspicuous in foundation-work as in superstructure. The empirical experience of the well-digger in sinking kerbs through loose soil or water-bearing strata, had resulted in such skill in the use of sheet-piling, coffer-dams and caissons that, for a long time, railway engineering de- veloped no novelty of importance in foundation-work. The steam pile- driver had supplanted manual labor at the winch-handles; light-houses and piers were founded on iron piles screwed into the sea-sands or sunk by water-jets, but engineering work'of this character, continued on the same general lines until the emplojonent of compressed air in subaqueous con- struction. Diving-bells had been much used for subaqueous work until the intro- duction of compressed air for this purpose. It is stated that in 1778 Smeaton sank the foundations for a bridge at Hexham, in Northumber- land, by the use of compressed air. In 1830, Sir Alexander Cochrane (afterward Lord Dundonald) patented the use of compressed air in working in water-bearing strata for bridge-foundations and pier-construction, though it seems not to have been practically applied. In 1849, Treger used pneumatic caissons for driving through river-sand. In 1851, Sir Charles Fox applied an expedient known as "Potts' Pneumatic Process" in the construction of a bridge over the Medway, at Rochester, in connec- tion with the first iron caissons used for subaqueous foundations. Potts' process was introduced into the United States about 1852, and was first employed in founding the piers of a bridge over the Great Peedee River in South Carohna during the construction of the Wilmington & Man- chester Railroad, now a part of the Atlantic Coast Line. The river-bed, to a considerable depth, consisted of coarse sand, easily scoured by freshets ; conditions very unfavorable for founding masonry-piers upon a timber-plat- form, either by a coffer-dam or by a caisson. Under these circumstances, the pneumatic process was adopted by the Chief Engineer, L. J. Fleming. The piers were composed of cast-iron columns in pairs, the draw-pier being a group of five. These columns were in sections of nine feet in length and six feet in diameter, bolted together by internal flanges in lengths suflBicient for the top of the column to be above water when its bottom rested on the river-bed, supported by a surrounding stage of piling. A cast- iron air-lock was attached to the upper end of the column and connected with a steam air-compressor on the staging. The sand .from the interior was sent up through the air-lock in canvas-bags and, as it was excavated beneath the chamfered edges of the bottom-section, the column descended by its own weight, being kept plumb by guide-piling. When a depth was ROADWAY 171 reached below the scouring action of freshets, the interior was made water- tight with a layer of cement and then it was filled up in the open with con- crete. This process was repeated at the bridge over the Santee River, on the Northeastern Railroad, and at the Savannah River Bridge on the Charleston & Savannah Railroad ; the latter being left incomplete at the beginning of the Civil War. Both of these roads have since been merged with the Atlantic Coast Line. The pneumatic process, as thus practiced, had objectionable features. The air-lock had to be hoisted off to attach each additional section, and the passage of men and of materials through it was slow, and inconven- ienced by the circumscribed space. The air-content of the whole column had to be maintained at a pressure just sufficient to counterbalance the varying depth of the river. If the pressure were insufficient, the water rose inside ; if it were in excess, the air blew out underneath. As it rose in bubbles along the surface of the column, there was a consequent reduc- tion of skin-friction against the sand thus set in motion, which took place unequally and tended to throw the column out of plumb. Occasionally its descent was arrested by excessive friction, and it became necessary to order the men out and to overload the column with rails to force it down. Snags buried deep in the river-bed caused serious delay in cutting them free of the descending colimin. If they protruded above the surface of the river-bed, they were liable to shoot out when freed, followed by a blow- out that imperiled the men inside. These difficulties were obviated to a great extent, and the value of the pneumatic process largely augmented, by transferring the air-lock from the top to the bottom of the pier, thus converting it into a caisson of any desired area, connected with the outer air by open shafts. The caisson served as a foundation for a masonry pier, whose increasing weight main- tained its descent as the excavation progressed^ and the number and dimensions of the shaft could be suited to the rapid handling of materials both ways and at the same time. After a sufficient depth had been reached, the caisson could be filled in and remain as a foundation. Electric illumina- tion was substituted for the oil-lantern and the telephone replaced sig- nalling by hammer-blows upon the shell of the column.^ The pneumatic caisson underwent a far more important development in its application to subaqueous tunnel-work in connection with the excavating shield. Cement and Concrete. Reinfokcement As the advance in the metallurgical arts, following upon Bessemer's discoveries, inaugurated a new era in iron-bridge building, so the contem- ' The foundations for the Eads Bridge at St. Louis, completed in 1874, were sunk by the compressed-air process to rock-bottom, 136 feet below high water. The piers supporting the Forth Bridge, 1882-1889, were sunk forty feet beneath the river-bed upon pneumatic caissons, 70 feet in diameter. 172 EFFICIENT RAILWAY OPERATION porary development in concrete work gave a similar impetus to masonry- construction. From the days of the Later Republic, concrete had been generally used in Rome for building important structures, with the wall- openings formed in brick ; of which the Basilica of Constantine, or rather of Maxentius, is a notable example. The lime burnt from the materials in that vicinity proved particularly suitable for work under water, but, as these natural earths were not widely distributed, hydraulic lime was not generally used. In 1791, a method was introduced for preparing it arti- ficially from marl ; the product being known as Roman cement. Regions abundant in marl profited greatly by this discovery until it was found that the elementary materials, lime, silicates and clay, when mixed in suitable proportions, would produce hydraulic lime superior to, Roman cement, known in the United States as Rosendale cement. The manufacture of this artificial or "Portland" cement has now become a leading industry in every country in which its component elements can be obtained, with a greatly increasing substitution of concrete for stone and brick work. In the United States, the production has increased as follows : 1880, 42,000 barrels ; 1890, 335,000 barrels ; 1900, 8,482,000 barrels ; 1910, 68,205,000 barrels.^ Concrete is usually composed of about one part of cement to two of clean sand and three to five of broken stone or bricks, or gravel. It was formerly used only as rubble or molded in blocks, known as b^ton in France, where blocks weighing 350 tons each have entered into the con- struction of break-waters and other marine works. In 1868, Morier, a French gardener, invented a method of making water-basins from cement strengthened with wire-netting. From this invention there has been a remarkable development of the art of masonry in the use of concrete molded in forms or frames in connection with iron wire or rods, so com- bined that their respective resistance to stresses of compression or of ten- sion are advantageously utiUzed. By 1880, reinforced concrete had be- come a material of recognized value in railway engineering. Because of its ready adaptability in many forms for walls, floors, beams or columns ; its fire-resisting qualities; its cheapness as compared with masonry of brick or stone, and the expedition with which concrete work can be carried on to completion, it is extensively used in superstructures, piers and foundations of important bridges, as well as in architectural work. 1 Portland cement is a compound of calcium, silica and alumina, being a mixture of 1.7 parts,. by weight, of lime to 1 part of silicious matter containing soluble silicates with an additional component of alumina and iron, burnt to in- cipient fusion. The resulting clinker contains from 60 to 64 per cent, of lime, 19 to 25 per cent, of silica and 5 to 9 per cent, of alumina. After it has been ground' there is added about 3 per cent, of gypsum or calcium sulphate, "to regulate its set- ting time. Portland cement attains its maximum strength after hardening from six to twelve months. George P. Diekmann. Journal Am. Soc. Meeh Ene May 1916. G.A.Rankin. /6id., Sept. 1915. ROADWAY 173 Recent and conspicuous examples of structures in reinforced concrete are the Hudson Memorial Bridge over Harlem River and the viaducts on the extension of the Florida East Coast Railway; a project which would have been impracticable with any other material.^ Cement containing a con- siderable proportion of gypsum should not be used in concrete work exposed to sea-water, as its consequent reaction tends to its gradual disintegration. The usefulness of reinforced concrete will be much enhanced if the suggested photographic examination with Roentgen rays of the steel reinforcement can be made practically available. Tunneling. Ancient and Modern Examples Where the established grade-line must penetrate the natural surface beyond the limits of economical excavation by an open cutting, engineering provides the efficient means for prosecuting the work by tunneling. Tun- neling is an ancient art, originating in quarrying and mining operations. In Ancient Egypt, subterranean tombs were excavated in solid rock, and tunneling was applied to the drainage of inundated lands. Lake Copias was drained by tunneling 150 feet beneath the adjacent surface. The art of underground excavation was further developed in connection with the water-supply of cities. On the island of §amos, an aqueduct built about 625 B.C., included a tunnel 4200 feet long of a section eight feet square. In the construction of navigable canals in Europe, further experience was gained in subterranean excavation and, with the general advance in en- gineering, tunnel-work of considerable magnitude was undertaken, such as the Steinedge Tunnel, three miles in length, on the Huddersfield Canal in Yorkshire. Resort was also had to tunnel-work in the construction of important highways through mountainous districts, of which the Simplon Road is a notable example. Accuracy of alignment is imperative in driving tunnels. Before the invention of telescopes, the alignment was established by the water-level in connection with an instrument for measuring angles of deflection. The general direction of a tunnel of considerable length could only be preserved by prosecuting the work from intermediate shafts at frequent intervals ; and the penetration of the timnel beneath the natural surface depended on the depth to which it was practicable to sink these working-shafts. In 82 A.D., Lake Fucino in Italy was drained by a tunnel 3^ miles long and 10 feet by 6 feet in section, aHgned by forty shafts, some of them 400 feet in depth. The skill now acquired in the adjustment of the instruments used in alignment, is exemplified in the spiral tunnels on the St. Gotthard Rail- ^ On the electric railway recently constructed between Chur and Arosa in Switzerland, there is a bridge of reinforced concrete with, a span of 315 feet in the clear and a rise of 140 feet, intended for a normal load of 44-ton motors and 33- ton trailers. The fan-shaped centering rested on three reinforced concrete t '•=!. Engineering News, March 9, 1914. 174 EFFICIENT RAILWAY OPERATION way and also in the Simplon Tunnel, where it was impracticable to test the instrumental work during construction, from the absence of shafts through- out its length of twelve miles.^ Tunneling fob Railways The introduction of railways gave a further impetus to tunnel-con- struction. The necessity for keeping the ruling-gradient within the prac- ticable hmits of the adhesive weight of the locomotive, often left no alter- native to tunneling, if the volume of traffic justified the additional expense. To meet such conditions, railways were projected to cross mountain ranges at great elevations, with the increasing confidence due to progressive experience and the appUcation of engineering genius and skill to devising improvements in methods and appliances. The Alpine region of Central Europe affords so many examples of this character as to warrant some de- tailed account of the latest tunnel-work in that region.^ The penetration of mountain-ranges by railway tunnels calls for thorough preliminary investigation, with especial attention to geological conditions. These can be predetermined with approximate accuracy where the stratification is at a considerable dip and at right angles to the pro- jected line. Such predetermination increases in uncertainty as the sub- tended angle decreases between that Une and the strike of the stratifica- tion, and with its lesser dip. It becomes virtually impracticable where the stratification is contorted or folded. It is also important to determine whether the rock will be found soUd, broken or decomposed. In lime-stone regions, the conditions to be expected are variable ground, with frequently great pressure, strong inflow of water and occasionally fire-damp. From present experience, the zone of rock-pressure does not increase with the depth below the surface, and has not been especially dangerous in long tunnels. In a part of the Simplon Tunnel where there was great pressure for a distance of 46 yards, it was due to an unstable mass, not more than from 600 to 1000 feet in depth, and did not correspond to the pressure which should have been produced at 4000 feet below the surface. The movement of sections of the superincumbent mass, the percolation of water, excessive temperature and foul air, often expose the working forces to imminent peril and greatly delay the prosecution of the work, with material increase in its cost. At the Tenda Tunnel, in Italy, a fault 1 In the oonstruction of the tunnel under the Mersey, between Liverpool and Birkenhead, 1881-1886, shafts were sunk on each shore for a drainage-drift- way, 180 feet below quay-level and outside of the tunnel-section. From these shafts, offsets were made to the center line of the tunnel, which was then pro- jected from a base line of twelve feet, upon a rising gradient of 2 feet in 1000 to the middle of the river, with a total length of 1350 yards. The main tunnel, driven down-hill in advance of the drift-way into which it was drained, was con- nected with a total error of 2i inches. '^ For elevations attained in railway location, see Appendix IV, Table I. , ROADWAY 175 of 47 yards led to the abandonment of the work by the contractor. After great delay, the passage was effected at a cost of about $60,000. At the Loetschberg Tunnel on the line from Spiess to Brieg, in Switzerland, the geologists predicted soKd rock under the bed of a mountain-stream which was 558 feet above the tunnel-section and, as supposed, with 328 feet of solid rock overhead. Yet, at a point nearly two miles from the tunnel- mouth, the stream burst through a fissure, filled the tunnel with sand and gravel to within a thousand feet of the entrance and buried twenty-five men. The Une was in consequence diverted and lengthened 805 yards. In the Karawanen Tunnel, on the new line from Vienna to Trieste, three miles in length and completed in 1906, the rock-pressure set a section in motion extending for 1| miles. Nearly two years' delay was thereby occasioned. In double-track tunnels, these difficulties are more readily surmounted, and the lengthening hauls of materials and supplies within the tunnel are less troublesome and expensive. For these reasons, even where the second track is not demanded by present traffic considerations, in the end the double-track tunnel may be found to have been built at less cost than for a single track. The Simplon Tunnel was planned for a single track with passing places. As the work progressed, unexpected conditions as to tem- perature called for fresh air in volume requiring a conduit of dimensions for which there was not sufficient space in the working-section. A ventila- tion-heading was therefore driven beside the tunnel, to be utilized as the bottom-heading, whenever a second trmnel should be required.^ Two tunnels so close together may cause the masonry to give way from excessive pressure, producing serious obstruction in both tunnels. For these rea- sons, the plan of the Loetschberg Tunnel was changed to double-track at an increase in the estimated cost of $1,250,000.^ The inside dimensions of single-track tunnels in Europe seem sufficient to insure safety in operation, but it is deemed advisable in future con- struction of double-track tunnels somewhat to increase the width, as the use of larger and longer rolling-stock is gradually decreasing the clearance between the loading gauge and the tunnel walls. In curves, the clearance may be slightly increased by shifting the tracks more to the inner side.' ' The enlargement of this heading for a second track was commenced on the north end December 20, 1912, and on the south end March 30, 1913. The tunnel was completed May 29, 1914, at a cost of $5,307,500 ; the estimated cost having been $7,000,000. The cost of the double tunnel, exclusive of ballast, truck and electric installation, was $16,598,000, or $255.25 per Hnear foot. ^ In France, single-track tunnels of standard gauge rarely exceed 1 \ miles in length, though, on electrically operated hues, there are such tunnels from 3 to 4| miles in length. In Italy, there are steam-operated single-track tunnels from IJ to 3i miles in length, and a single-track tunnel constructed in Switzerland, in 1910, is 5\ miles long. ' The usual cross-section for double-track tunnels in France is from 26 feet 3 inches to 26 feet 11 inches at imposts, 24 feet 7 inches at rail level and 19 feet 8 inches above rails ; though some recent tunnels are 28 feet 6| inches in width. Recent tunnels in Austria are 26 feet 11 inches at imposts and 21 feet above 176 EFFICIENT RAILWAY OPERATION Methods of Tunnel Construction There are two methods for carrying on tunnel-work, known as top- , heading and bottom-heading, which are adapted by modifications to meet varying conditions. In the Alpine tunnels and in long tunnels elsewhere in Europe, except in France, the bottom-heading has been found prefer- able. With top-heading, the time for completing the tunnel after the headings have met has been longer, and it is more difficult to maintain the working-tracks and the temporary drainage. The piping for ventilation and air-supply has to be shifted oftener, and taking out the excavated material at the level of the imposts interferes with the masonry-work. With bottom-heading, the temporary track may usually be kept close up to the working-face and is not subsequently disturbed. The drainage gives less trouble, as a drainage ditch or even a completed culvert may also follow closely upon the advance in the heading. In some of the Alpine tunnels, instead of top-heading, the bottom-heading has been extended upward in successive stages with inclined blast-holes. The rock thus of- fers less resistance, and difficulties in ventilating a top-heading are avoided.' Bottom-heading is undoubtedly preferable when the work is not less than three miles from the entrance, especially when there is reason to expect rock-movement or a considerable inflow of water. A third method has been introduced in tunnel-work, known as under- heading, to which attention was directed in Switzerland, in 1897. By this method, a heading is first driven upon the center line, and beneath the tunnel-section, for an independent drainage-tunnel; it is about nine feet wide and seven feet deep, as to the interior dimensions of its masonry-lining. This underheading is kept in advance, followed by the usual bottom-head- ing. As the underheading receives its lining, air is blown through it into the tunnel proper, and the working-faces above are ventilated by a tempo- rary installation, where compressed air is not used for drilling. In solid rock, the lining may be kept at a distance of 600 to 700 yards behind the working-face of the underheading. This distance allows the use of a larger number of trucks to hold the excavated material before a train is shifted out, and with corresponding expedition in the progress of the work. The underheading-track is used for receiving material from the bottom-heading as well, with relief to the track in the main tunnel and with better oppor- tunity for adapting its timbering to pressure from rock-movement. In case of an inburst of water, it is carried off through the underheading, without interruption to the work on the main faces. Where there is a movement of the rock, the lining in the underheading is carried as clpsely rails. Single-track tunnels in France are 16 feet 5 inches at imposts ; in Aus- tria, 18 feet ; in Italy, on older lines 15 feet 1 inch at imposts and 16 feet 5 inches headway, and on newer lines 18 feet at imposts with 19 feet headway. 1 In the Albula Tunnel, there was a saving of from $18 to $27 per linear yard by this method, as the blasting was easier and the ventilation was better. ROADWAY 177 as possible to the working-face, and access is afforded to the bottom-heading through successive openings in this lining as the bottom-heading is ex- tended, and with greater facility for timbering. The completed under- heading serves not only for drainage but also as a conduit for permanent ventilation and for the cables required for telegraphs, telephones, electric lighting and signaling apparatus, as well as for electric traction. It has been proposed to lay the permanent way directly upon the vaulting of this secondary tunnel, supporting the rails in cast-iron chairs with under-ribs imbedded in concrete, which is rammed in after the track has been lined up. This plan renders it unnecessary to bring other track material into the tun- nel for the standard-gauge track, which can thereafter be used with greater advantage for the carriage of construction-material. Mountain tunnel-work has been but rarely carried on under air-pressure. In a tunnel in Italy, at the beginning of the work, mud was encountered in a semi-liquid state, containing large bowlders, in which it was impracticable to use a compressed-air shield. Sections constructed on the surface, in lengths of about 18 yards, were sunk to position upon compressed-air caissons for 205 yards. The work was then carried 175 yards farther with an air-shield, and at a cost of $1060 per yard. The time consumed in subterranean excavation by manual labor alone under primitive methods discouraged the prosecution of tunnel-work of considerable magnitude. On the Lake Fucino Tunnel,^ 30,000 men were employed for eleven years ; a work which could now be accomplished in as many months with a tithe of that labor. Even with the use of blasting powder, the Harecastle Tunnel, 1.6 miles in length and 9 feet by 6 feet in section, on the Trent & Mersey Canal, begun in 1766, was only completed in 1777. TUNNEL-WOEK IN THE UnITED StATES AND CANADA There was but little tunnel-work in early railroad construction in the United States. The lack of capital led to the adoption of heavier gradients in crossing the mountains between the sea-coast and the Mississippi Valley than had been used in Europe. The first railroad tunnel was on the Alle- ghany Portage Railroad at Staple Bend, four miles east of Johnstown, 900 feet in length, and completed in 1833. The abolition of inclined planes on this route was accomplished by heavier rock-excavation, of which the most important was the tunnel at Galitzin, a mile in length. In the construction of the Baltimore & Ohio Railroad over the Alleghany Range from Cumberland to "Wheeling, 201 miles, there were eleven tunnels, having a total length of 11,150 feet. Farther south, in Virginia, North and South Carolina and Georgia, State aid was freely given toward tunnel-construc- tion through the Blue Ridge. Scarcely had such assistance been extended to the Georgia State Railroad, between Atlanta and Chattanooga, where the work was least expensive, when the Civil War interrupted further 1 See p. 173. 178 EFFICIENT RAILWAY OPERATION progress. Subsequently, the extension was accomplished in Virginia, and in North Carolina, but, to this day, work on the Rabun Gap Tunnel in South Carolina has not been resumed. The longest of these tunnels is the Big Bend, 1.23 miles, on the Chesapeake & Ohio Railway. In the Easterri States, the Bergen Hill Tunnel afforded the Erie Railroad access to New York Harbor at the beginning of the Civil War. The most important tunnel-work, however, was the Hoosac Tunnel, 4.7 miles, begun in 1855 with the methods then in vogue and completed in 1875. This was the first tunnel-work in America in which compressed-air drills and nitro- glycerin were employed. In the Mississippi Valley, a vast area was opened to railroad construc- tion in which no heavy mountain-work was encountered until the under- taking of a transcontinental railroad made it necessary to cross the Rocky Mountains. Here the Union Pacific Une was located through a favorable pass, and likewise the Southern Pacific line. It was only with the develop- ment of mining operations in this region that railroad tunnel-work in the United States became comparable in interest with that in the Alpine re- gion of Central Europe. The ingenuity displayed in the location of the Rio Grande Southern, of the Denver & Salt Lake and of the Denver & Rio Grande roads averted the necessity for very long tunnels, even at alti- tudes between 10,000 and 12,000 feet. The Atchison, Topeka & Santa F6 Railway attained the Pacific Slope through the Raton Pass Tunnel, 2000 feet in length, at an altitude of 7622 feet, and on a maximum grade of 185 feet to the mile.^ In later construction, both in the United States and in Canada, the Continental Divide has been pierced by tunnels, from two to five miles in length, in which American engineers have fully kept pace with European practice. The increasing abundance of available capital and the desire for greater tonnage-capacity by reducing gradients has given an impetus to tunnel- work. On a railroad with present length of two thousand miles, there have been added since 1904, on branch-lines twelve tunnels, with aggre- gate length of 7863 feet, the longest being 4770 feet ; while on main-line betterments, there have been added thirty-two tunnels, with aggregate length of 32,251 feet and a maximum length of 3291 feet.' 1 Appendix IV, Table III. Tunnels op Recent Construction in United States and Canada Railroad Name Length, ft. Completed Great Northern . . . Cascade 14,400 1905 Terminal Seattle 5,141 1905 Ca. Clinchfield & Ohio . Sandy Ridge 7,760 1914 Canadian Northern . . Mount Royal 17,000 1914 Del. Lack. & Western Nicholson 3,630 1915 Ch. Mil. & St. Paul . . Snoqualmie 11,890 1915 Canadian Pacific . . . Connaught 26,400 1916 ROADWAY 179 Single-track tunnels of recent construction are 16 feet wide, with clear height of 22|^ feet from base of rail. Double-track tunnels have the same clear width outside and same height over center of each track. Where electric traction is used, a smaller cross-section area may be used. The single-track section of the Pennsylvania Railroad tunnels in New York City has 225 square feet of area above the track. The Snoqualmie Tunnel through the Cascade Mountains, 60 miles east of Seattle, was opened for traffic in January, 1915. At the west end for 436 feet the work was advanced by top-heading through yielding and saturated material. The remaining distance, through harder material, was advanced by bottom-heading, 8 by 13 feet, driven at subgrade along the north side of the tunnel at 1000 to 2000 feet ahead of the bench. From fourteen to thirty 9-foot holes were drilled for each shot. Following the heading, the wings were driven to full-section width, after which the trap or stoping timbers were placed. Bench-openings were driven to full- section width at intervals of 150 feet. Although the rock was hard, it was so stratified and filled with soft talc-seams that the tunnel was lined throughout with concrete, averaging 6.1 cubic yards per linear foot.^ The tunnel was ventilated at each end by the exhaust method through a two-foot pipe that opened at the end of the enlarged section. An auxiliary plant at each end forced air into the heading through a ten-inch pipe. The longest railroad tunnel in North America has been recently con- structed in connection with the relocation of the Canadian Pacific Railway in British Columbia. The new line, of about 10 miles in length, replaces the line over Rogers Pass, with a reduction of 4.3 miles in distance and of 552 feet in summit-elevation, and the elimination of 2500 degrees of curva- ture, 4-J miles of snow-sheds and some very heavy bridging. This im- provement was effected by the construction of a double-track tunnel, 26,400 feet in length, known as the Connaught Tunnel. Work on this tunnel was commenced in September, 1913, and the new line was opened for traffic on December 9, 1916. TunneUwork in America has usually been prosecuted on the top- center-heading method. The Snoqualmie Tunnel, on the Chicago, Mil- waukee & St. Paul Railway, was excavated on the bottom-heading plan, 'Snoqualmie Tunnel Average daily progress 9.5 feet Maximum daily progress 25.0 feet Average time between shots, divided as follows : 15.5 hours Breaking down roof and mucking back 2|-3 hours Setting up cross-bars, drilling and mucking out 7 hours Taking down, clearing and shooting 1 hour Waiting for heading to be clear of gases 1 hour In the Grand Trunk Pacific single-track tunnel- work in sliding material, the concrete lining, 21 inches thick, averaged 5.57 cubic yards per foot of tunnel with 680 pounds of steel reinforcing rods. 180 EFFICIENT RAILWAY OPERATION driven on one side. In the Connaught Tunnel, the advance was made by a parallel drift or heading, driven fifty feet to one side of the tunnel-center, followed by the main heading, with which it was connected at intervals by cross-cuts. This pioneer-heading was 7 feet wide and 8 feet high in rock, and furnished the means during construction for air-and-water pipes and other appliances, and for carrying supplies. It was also used for,con- veying material from the main heading to a point in the main tunnel back of the shovels. For these purposes, it fulfills the functions of the "under- heading" method, recommended by engineers in the Alpine region. The pioneer-heading was started, at the east end, 700 feet west of the portal and 60 feet above the main-tunnel level. By this plan, 700 feet of pioneer-tunnel was saved and the soft-ground heading reduced. The heading ran nearly level until the grade of the main heading was reached. Solid rock was struck 600 feet from the entrance and, at this point, an inclined cross-cut was started into the main heading. The west pioneer- heading was started on an incline 300 feet long, from the outcrop, 700 feet east of the west portal and 150 feet above the main-heading level. This location provided dumping ground and avoided soft-ground tunneling. The headings were driven with light hammer-drills of hollow steel, with water-attachment, mounted on a light horizontal bar. The cross-cuts into the main heading were from 1500 to 2000 feet apart. There were in all 19,610 linear feet of pioneer-tunnels and twelve cross-cuts of about 40 feet each. The pioneer-tunnels were discontinued for one mile in the center of the tunnel ; connection being made by the main heading only. The main heading was located midway of the full section, eleven feet wide and nine feet high, with its bottom six feet above the subgrade. The rock consisted largely of schists, except for 1200 feet of glacial drift at the east end and of 400 feet of soft ground at the west end. No trouble was experienced from water-infiltration in the rock, and timbering was re- quired only in the rock near the center of the mountain. Work was started January 15, 1914, and was prosecuted from both ends. The headings met December 19, 1915. The material from the enlarged section of the tunnel was excavated by shovels operated by compressed air, and was , loaded directly into twelve-yard cars. The cars were hauled to the en- trance of the tunnel by standard-ganige compressed-air locomotives, and thence to the dump by steam. On account of delay in driving the tunnel through the soft ground at the ends, when the main headings met, the shovels employed in the enlargement were two miles apart and did not meet until July, 1916. They advanced 27,749 linear feet in 540 days; an average of 46.1 feet per day. In the enlargement, radial shooting from the central heading was employed, instead of drilling the holes parallel with the axis of the tunnel ; the holes being at an inclination of about one in four from the direction in which the tunnel was driven. About one and a half miles of the tunnel, including the soft ground at the ends, ROADWAY 181 requiring concrete-lining. The tunnel was finished eleven months ahead of the contract-time at a cost of about $6,500,000, including the- ventila- tion-system. The work was prosecuted under unusually favorable con- ditions as to water-infiltration and rock-pressure.' Recent Tunnel Construction Snoqualmib CONNAUGHT Material Length, feet , Track , Finished section . . . , Excavated area . . . , Yards per lin. ft Heading , Heading per lin. ft. Excess of excavated area, per lin. ft Black Slate Quartzite Conglomerate 11,890 Single 352 sq. ft. 517 sq. ft. 19.2 sq. ft. Bottom 4.0 cu. yds. 3.5 cu. yds. Slates Quartzite Schists 26,400 Double 526 sq. ft. 831 sq. ft. 30.76 sq. ft. (Solid rock, 25.0) Center-heading, Pioneer Drift 4.0-2.1 cu. yds. 3.0 cu. yds. Tunnel Timbering and Lining The methods of timbering in tunnel-work vary so much under differ- ent conditions that they can not here be considered within reasonable limitations. The difficulties encountered, often unexpectedly, are over- come by ingenious devices that require graphic illustration to be under- stood. There have been cases in which the passage through a fault has cost over $1200 a yard for timbering, the materials alone costing $200 a yard. In the recent Austrian tunnels, from the bottom-heading in advance, shafts were opened to the top of the arch and from these the top-heading was driven. From the top-heading, for a certain distance back, the full section was excavated in separate rings, each nine to eleven yards in length, in which the hning was built. The intermediate sections were then excavated and lined, the inverts put in, ring by ring, the bottom-channel constructed, and the sole of the completed section covered with concrete. In this method of excavating by individual rings, crown-timbering is used in the full section ; side and intermediate sills are put in at the height of ' See "Methods Adopted in Construction of Connaught Tunnel," J. G. SuUi- van, Chief Engineer. The Cornell Civil Engineer, January, 1917. Also "Con- struction Methods for Rogers Pass Tunnel," A. C. Dennis, Mem. Am. Soc. C. E.,. Trans. Am. Soc. C. E., December, 1917, p. 448. For other information as to recent tunnel-work in the United States and in Canada, see paper on "Tunnels" by Charles S. Churchill, Assistant to President, Norfolk & Western Railway Com- pany, presented at a meeting of the International Engineering Congress, San Francisco, September 20-25, 1915. 182 EFFICIENT RAILWAY OPERATION the imposts on which the crown-timbering is supported, with bearers, under the cross-sills where the pressure allows this to be done, and shoring up the sills and bearers by props on the tunnel-sole. This self-supporting timber- ing has been found to satisfy all requirements, may easily be put up by experienced miners, and requires no unnecessary excavation. A minimum amoimt of timber is cut to waste, much of it may be used a second time, and the timbering may be changed with safety while the masonry is carried on. In tunnels with two inchnes, where the ends are nearly level, the ruling- gradient is usually 2.5 to, 3 feet per 1000, to facilitate drainage. On ac- count of less adhesion in tunnels, the maximum gradient should be less than in the open in the proportion of 2.6 per 1000 when the outside gradient is 10 per 1000, and of 5.6 per 1000 when outside it is 25 per 1000. In tunnels with an interior summit, it should be nearer the lower entrance, if possible, so that, if the work there gets behindhand, it will not be neces- sary to stop the work on the other side. Although there is great variety in the character and dimensions of masonry-lining, certain points are common to most of such work. Ex- perience has shown that, under excessive pressure from the superincum- bent mass, the cross-section should be as nearly circular as possible, with extensive use of inverts. While it is not difficult to adapt the cross-sec- tion of a double-track tunnel to the probable pressure, inverts often pro- duce fractures in a single-track tunnel with an egg-shaped cross-section. The minimum thickness of the lining of a double-track tunnel is usually about two feet. In exceptional cases, arches have been built four feet thick, of large blockp, squared on all faces. The thicker the lining, the greater is the cross-section of excavation and the proportionate pressure. The thickness of the lining should, therefore, be diminished as far as pos- sible by excellent material and workmanship. As rock-pressure generally begins to show only after some time has elapsed, the hning should be put in as soon as practicable. In double-track tunnels, the drainage-channel is in the middle of the floor ; in single-track tunnels, it is frequently near a side-wall, where there is no invert. In passing through sections of rock under pressure, or through saturated ground, the materials, dimensions and shape of the lining must conform to the conditions so encountered. In moderately stable ground, any building material may be used, though cement is preferable for mortar. Brick may be used to advantage, as it hastens the setting of the cement ; but not in very thick arches, as the extrado joints become too wide. The brick should be well-burnt and porous, but where good stone is available, 'it is preferable, especially if cut on all surfaces. Rubble-work should not be used in arches, nor even in side-walls, if stone of suitable strength and size can be conveniently obtained. Where currents of air reduce the temperature below the freezing point, materials that are acted on by frost ROADWAY 183 should be excluded. Reinforced concrete has become valuable for tunnel- lining where the masonry is exposed to great pressure. As failure in the vaulting generally takes place near the concave surface, the reinforcement should be so arranged as to distribute the pressure on the inner third of the block, over its whole cross-section. It is difficult to make a tunnel-lining absolutely impermeable by water. Usually the extrados is covered with dry stone laid on cement. The water then collects at the imposts and is led to the main drain by channels in the walls or outside of them. Even under slight water-pressure, this method is not satisfactory, but may be improved by covering the layer of cement with tarred sheets carefully bricked in. The best method is the injection of cement behind the masonry with compressed air. Where this is done, the layer of dry stone should not be used, as the voids require a large quantity of cement. The injection should be made upward from the lower part of the lining. Catch-water drains must be provided to collect any spring-water, as it will waste away much of the cement. They should be watertight, to prevent their obstruction by extraneous matter. Water containing sulphates should be carefully excluded, to prevent chem- ical reaction with the mortar, which has been found completely decom- posed by water containing a hundred grains to the gallon of calcium sul- phate. In this case, a layer of concrete was covered with one of asphalt, which is a slow and costly measure. The use of mechanical apphances for the drainage of a completed tunnel should be reduced, as far as prac- ticable, by resort to gravity. Where such drainage can not be provided directly, the water may be collected into pits and pumped into the drain- age-ways. Bronze-studs should be placed in the hning and their position carefully determined in order to test its stability under pressure. Holes should be drilled through the masonry wherever there has been an inflow of water. Any sudden variation in the normal infiltration should be immediately investigated. It is of doubtful advantage to finish the arch before the side-walls, as in top-heading. Where the supporting material is apparently stable, accidents frequently happen from the imposts giving way. Even in very hard rock, the removal of the underlying material exposes the arch-masonry to serious disturbance and, however much care may be taken in blasting, the lining of the arch is often damaged. The restoration of a lining that has failed under pressure may not cause serious obstruction in a double-track tunnel, as a single track can usually be maintained through it under a temporary roof. Where this can not be done, resort must be had to extraordinary expedients. In some cases, the bottom of the tunnel has been lowered to make room for the work overhead. Iron beams serve a useful purpose, and reinforced con- crete may facilitate the prosecution of such work. As a general rule, the work should be separated from passing trains by a partition-wall, or it 184 EFFICIENT RAILWAY OPERATION may be carried on outside of the lining. In view of the difficulties encoun- tered in the restoration of a defective lining and the loss from obstructed traffic, it is of the first importance that the original masonry should be carefully planned and thoroughly well executed. Cost of Single-track and Double-track Tunnels From the statistics given in Appendix IV, Table IV, it appears that the completed cost of single-track tunnels in Europe, over a mile in length and under six miles, has been from $200 to $300 per linear yard and, in exceptional cases, $432 and $560 per yard. The cost of the Simplon Tunnel, 12.2 miles, including the ventilating passage, was about $677 per yard. Double-track tunnels between one and five miles in length have usually cost between $300 and $400 per yard and, in exceptional cases, from $540 to $849 per yard. The St. Gotthard Tunnel, 9.2 miles, cost $749 per yard, and the Arlberg Tunnel, 6.4 miles, $823 per yard. The dimensions of some of these tunnels are given in Appendix IV, Table II. It is usually assumed that two single-track tunnels will cost about 30 per cent, more than one double-track tunnel in the same location. It is of doubtful economy to construct a very long tunnel unless the traffic war- rants a double-track. Though the St. Gotthard Tunnel was built for two tracks, at first only one was used. But, within the first year of operation, the second track was needed. In the Simplon Tunnel, where the increase in freight-traffic was slower, the want of a second track was felt after three years of operation. The saving in interest can only balance the extra cost of a second single-track tunnel after ten years have elapsed. Under ordinary operating conditions, there must be intermediate passing places in long tunnels, which are expensive. Economy in Time and Labor of Modern Tunneling Methods The cost of tunnel-work and the time consumed in its prosecution have become greatly diminished with the introduction of machine-drills and high explosives. In the Alpine tunnels, the advance with machines has been from three to ten times greater where the hand-work averaged less than a yard a day, and, on an average, seven times as much. The dif- ference diminishes as the hand-work exceeds a yard a day. At the rate of only a three times greater advance, it was found that a cubic meter, or 1.308 cubic yards, was excavated by hand at about three-fourths the cost of machine-work. This rate of cost, however, must vary with the relative cost of labor, and the time saved is an important matter. In very hard rock, the average advance by hand-drilling does not exceed a rate of one meter (3 feet 3f inches) per day. In the Simplon Tunnel, with machine- drilling, the average advance in each period of three months was about 656 yards, with a daily maximum of 26 feet 3 inches. With hand-drilling ROADWAY 185 the daily advance in nine instances averaged less than a yard ; in fourteen instances, less than two yards anti, only under exceptional conditions, as much as 2^, 3^ and 4 yards. With machine-drilhng, the daily average in four instances was less than three yards ; in five, less than four yards ; in six, less than five yards and in one case, 6.08 yards. In bottom-headings, machine-drills are generally used only in the main heading; their use in different working places is mainly a question of power. In a long tunnel where water-power is available, it is advisable to do all the breaking-out work by machines ; especially with the increasing cost of manual labor.i In machine-drilling, percussion-machines are generally used. The rotary cutters employed in coal-mining have not given satisfaction in rock- work. Compressed-air machines have been found preferable to those operated by hydrauHc power on account of their simpler construction, with less loss of time for repairs and with better working results. Per- cussion-hammers, operated by compressed air through flexible hose, may be used to advantage where it is inconvenient to use machines. The com- pressed-air machines are of two types ; either the drill is advanced auto- matically or by hand. Each has its advantages and its disadvantages. Electrically operated machines are also in use. There are three kinds, the solenoid, the crank-and-spring and the rotary type. The solenoid apparatus is the simplest, but it quickly becomes hot and it is not possible to withdraw the drill if it becomes jammed. It was therefore abandoned in the construction of the Jungfrau Tunnel, where the crank-and-spring machines gave more satisfactory results, but were eventually given up because of the frequent repairs. With an advance of 1.2 meters (about 4 feet) per set of fifteen holes, these machines averaged about three hours, and the compressed-air drills from two to two-and-a-half hours. The electric rotary machines have been but little used in tunnel-work, as it has not been practicable to obtain the pressure required for very hard rock without excessive friction. The electro-pneumatic percussion drill, com- bining the portable compressor and the air-drill in one apparatus, has proved the most efficient of the drilling machines. Tunnel-working has been facilitated by the substitution of petroleum 1 Comparative Progress OF TUNNEI/-W0RK Tunnel Date Length in Miles Section Average Daily Phogbess Mont Cenis St. Gotthard . . Arlberg .... Simplon .... 1857-1871 1872-1881 1180-1891 1898-1905 7.98 9.14 6.36 12.26 26'3" X 24'7" 26'3" X 24'7" 25'3" 16'5" X 19'6" 2.57 lin. yds. 6.01 lin. yds. 9.07 lin. yds. 11.63 lin. yds. Machine-drilling introduced at Mont Cenis in 1861. With hand-labor 9 inches per day ; with compressed-air drills, 45 inches per day. St. Gotthard. Compressed-air drills. 186 EFFICIENT RAILWAY OPERATION and acetylene lamps for the ordinary miner's lamps, and by electric light- ing in the completed archway. It is adtisable to use safety-lamps if there be apprehension of fire-damp ; accidents from this cause have occurred, even where its presence had not been previously indicated by geological conditions. With machine-driUing, intermediate shafts are not of much advantage for working purposes within 1500 to 2000 yards of the nearest tunnel-ends. In sohd rock, as much time is required to sink a shaft 300 feet with machines as to drive a heading 1500 to 2000 yards. Headings are driven more slowly from a shaft, and material is handled through it at much greater cost. Sinldng a shaft in water-bearing strata may lead to disturbance of the superincumbent mass, and add to the difficulty in starting the head- ings from it. IncUned shafts are preferable to vertical shafts, as the former can be sunk four times as rapidly. It has been found that four times the quantity of material can be removed through an inclined shaft as through a vertical one, and the rate of progress at the working faces is materially better. Much time is lost, in drilling, by clearing away the excavated mate- rial, amounting in different cases from an equal number of hours to nearly double as many, including the time required for putting in the blasts and for firing them. Attempts to reduce this loss of time by the aid of mechan- ism have not so far been satisfactory. In the removal of excavated mate- rial, mechanical traction has superseded animal power in the completed portion of the tunnel. In a top-heading it is stopped at the first ramp, where the material is dumped into the cars from hand-trucks, or the cars are run down into the train. Where the rock-pressure has made it diffi- cult to provide passing-places of sufficient width, the empty cars have been hoisted into excavations above and the loaded cars passed beneath them. In tunnels with a bottom-heading, compressed-air locomotives clear away the broken rock from the working-faces, but this requires powerful com- pressors. Tunnel-traction has also been performed with locomotives actuated by superheated water. Steain should not be used in the working- section of an artificially ventilated tunnel. It is a waste to pump air into a tunnel at great expense and then to vitiate it unnecessarily. As electric traction has now come into use in the operation of long tunnels, it should also be used in the completed portion of such a tunnel during construc- tion, but the trolley wire can not be taken beyond it. In any tunnel- installation, ample power should be provided for all requirements and sufficient water-power is usually available for hydro-electric plants. Ventilation of Tunnels Tunnel-ventilation gives occasion for serious problems during con- struction, as well as in after-operation. They become more difficult of solution at the great elevations at which long mountain-tunnels are driven. ROADWAY 187 At such altitudes, the rarity of the atmosphere and the external tempera- ture are conditions to be considered as affecting the working forces and the requirement for an artificial supply of air. Rock-temperature is also an important consideration. The temperature of the earth is constant at a depth of 65 to 175 feet and about 2° F. higher than the mean annual sur- face-temperature. From experience in driving deep tunnels, it is in- ferred that, below that depth, the temperature increases one degree for each 80 feet under a valley, for each 100 feet under a plain and for each 150 to 200 feet under a mountain. In the Jungfrau Tunnel, at an elevation of 10,500 feet, and where the rock was 1640 feet thick, its temperature at the working face was 39° F. Blasting raised the air-temperature to 55°, but, in the intervals, the air through the ventilators and from the drilling machines reduced it to 35°. Under these conditions, the men felt no ill effect froiji the altitude. In the St. Gotthard Tunnel, the highest temperature was 87° F. and accordingly it was calculated that the highest temperature in the Simplon Tunnel would be about 107°. In fact, on the north side the temperature reached 137°, while on the south side it was from 18° to 36° below the pre- dicted temperature. In the St. Gotthard Tunnel, the temperature rose a degree for each 160 to 180 feet in depth while, in the Simplon Tunnel, it rose at the same rate for each 120 to 150 feet. The erroneous prediction was due to insufficient data as to the mean annual surface-temperature, and to the cooUng effect of the numerous springs encountered in the St. Gotthard Tunnel. To this cause was also attributed the lower tempera- ture on the south side of the Simplon Tunnel. It was subsequently ascer- tained that the heat of the earth passes to the surface quicker through strata which dip considerably than through horizontal strata. Artificial ventilation becomes necessary for efficient work when the temperature in the tunnel exceeds 77° -F. Within certain limits of dis- tance, ventilation is much improved when compressed air is used as a motive force, though the exhaust from the drills alone has been found in- sufficient for this purpose. If a ventilating shaft communicates with the heading, it may be equipped with an exhaust-fan. As the length of the tunnel approaches three miles, an independent source of ventilation be- comes indispensable for an ample supply of fresh air, and resort must be had to blowers and conduits. A ventilation-plant should be planned for operation after the tunnel has been opened for traffic. Its economic efficiency for this purpose de- pends not only upon the capacity of the blowing apparatus, but also upon the relation of the cross-section of the conduit to its length. The resistance offered to the passage of air through a pipe varies inversely as the fifth power of its diameter. In the Simplon Tunnel, the minimum quantity of fresh air required for a working-section of 1000 yards was fixed at 1060 cubic feet per second. 188 EFFICIENT RAILWAY OPERATION There was not room in the heading for pipes over 20 inches in diameter, and to furnish such a volume of air through such a pipe would require about 50,000 horse-power; whereas the same service could be rendered with 18 horse-power through a conduit of 8 feet 3 inches in diameter. For this reason, the ventilation was effected through a drift-way of 65 to 75 square feet in cross-section, driven at a distance of 56 feet from the main tunnel. Sufficient ventilation was provided at the working-faces with 883 cubic feet of air per second, but the current could not exceed a velocity of from 12 to 15 feet per second without inconvenience from draughts and dust. For this purpose, a secondary apparatus was installed of small turbines and blowers, with water-jet blowers supplying 53 cubic feet of air per second. The installation was operated with 10 horse-power from the high-pressure water-supply to the drilHng-machines. By these means, the temperature of the air was kept below 77° F., so long as the rock-temperature was below 95°. When this temperature was exceeded, the air-temperature was re- duced by spraying the rock from a pipe about 10 inches in diameter, sup- plying 22 gallons per second at an outside temperature of 34° F. in winter, and of 46° in summer. This pipe was insulated by charcoal in another pipe, and delivered the supply at 46° to 59°. With the highest rock- temperature experienced (133° F.) and an air-supply of 883 cubic feet per second, the spray was delivered at 50°. Under these conditions, the tem- perature within the tunnel at five miles from the entrance was maintained below 77°. In the more recently constructed Austrian tunnels, 206 cubic feet of air per second was supplied through piping of 2 feet 3^ inches diam- eter in the finished portion and of 1 foot 7f inches diameter ia the headings, with 360 horse-power. The natural ventilation, which may have been sufficient during con- struction, may be inadequate to meet the requirements in after-operation. This depends in some measure on the difference in height between the tunnel-entrances, their situation as to environment, and on the direction and strength of the prevailing winds. Natural ventilation is usually suf- ficient in double-track tunnels and in single-track tunnels electrically oper- ated ; except in the Simplon Tunnel, which is on steep inclines where the heat becomes oppressive at the summit. In the Cochem Tunnel in Prussia, 2.6 miles in length and double-track, artificial ventilation was required after the trains exceeded seventy per day. In an investigation of 84 single- track tunnels in Italy, twelve had deficient ventilation normally and six- teen occasionally. In the tunnels with normally defective ventilation, double-heading was used on gradients exceeding 10 per 1000, and the num- ber of trains in both directions in twenty-four hours was not less than twenty-six. In France, the natural ventilation was insufficient in four single-track tunnels and in one double-track tunnel. In Grand-Brion Tunnel (1285 yards, single-track) on the Grenoble line, where double-head- ing was in use, it was found impracticable to use triple-heading. ROADWAY 189 The movement of air in a single-track tunnel becomes reduced with the length of the tunnel. As a consequence, the air may be less vitiated in the longer tunnels, yet it causes greater suffering because of the longer expos- ure.' In a single-track tunnel with a continuous grade and not more than half a mile in length, the atmosphere became oppressive when the num- ber of trains was over ten in twenty-four hours, heavily loaded up-hill. In some short up-hill tunnels, in the most unfavorable months, a car- bonic-acid content was found in the pusher-cab of 4.6 parts per 1000, and of carbonic oxide up to 8.1 parts, also sulphurous acid. The tempera- ture in the cab often ran up to 99° F. and even as high as 140° in the sum- mer months, when there was no wind. Where this is the case, pushers should be dispensed with, as far as possible, by the use of more powerful locomotives, and even by reducing the gross weight of the trains. Slip- ping is thus diniinished and the time shortened in the tunnel. Coal con- taining sulphur should not be used, nor coke, as it produces large quanti- ties of carbonic oxide and of hydro-carbons. Oil-firing produces no sul- phurous gases and but little smoke and, if the train be stopped, the com- bustion may be cut off. But the crew suffers from the higher temperature of the gases and, if there be an excess of oil, the smoke is worse than with coal. Rather successful results have been obtained by drawing air from the bottom of the tunnel and blowing it into the cab with a small blower, operated by a steam turbine on the locomotive ; the air is pumped through a strainer and the asphyxiating gases absorbed. This expedient, however, affords no relief to the track-hands, and, under such conditions, adequate improvement in a foul tunnel can only be effected by electric traction or by a modern ventilation-plant. If the volume of carbonic gases in the air in proportion to the quantity of coal consumed be taken as a measure of vitiation, the relative values corresponding to good and bad ventilation may be expressed numeri- cally as 4 and 11 parts per 1000, respectively. It has been proposed to ■establish 6 parts per 1000 as the limit of vitiation with freight-trains, and 3 per 1000 in the case of passenger-trains.* Where there is an interior summit, the vitiated air is withdrawn at that point through a duct connected with exhaust-fans at the ends of the tunnel ; 1 The movement of a train in a tunnel 274 yards in length produced a current of 24 feet per second, but only at- half that rate in a tunnel over three miles long. * Volume op Air Requiked fob Ventilation. Assuming that 29 cubic feet of poisonous gases are produced for each pound of coal consumed per mile, then in order to maintain the atmosphere in a tunnel at 2 parts of carbon dioxide per 1000, there will be required 29 times the number of pounds of coal consumed multiplied by 500 and divided by the intervals in minutes between passing trains. For example, in a tunnel one mile in length with fuel consumed at the rate of 32 pounds per mile and a train passing in each direction every five minutes, 32 X 29 X 500 ^^35500 cubic feet per minute, the volume of fresh air required. 2§ Enc. Brit. XXVII, 409. 190 EFFICIENT RAILWAY OPERATION fresh air being supplied through the portals. This method, known as the "Guibal" system, is also in use at the Severn and Mersey tunnels, where the air is withdrawn at the point at which descending grades meet and the heavier gases become concentrated. At the Mersey tunnel", the top of the exhaust-opening was parallel with the fan-shaft. As each blade of the fan passed the opening, there was a momentary stoppage of the air- current which imparted a vibratory motion to the fan that was injurious to the apparatus and gave a trenaor to the atmosphere, which affected the neighborhood. This was obviated by cutting a V-shaped opening in the shutter at the chimney-entrance ; thus gradually decreasing the aperture and allowing the air to pass into the chimney in a continuous current in- stead of intermittently. The exhaust-system is unsuited to a tunnel on an ascending grade. Here the process should be reversed, and air forced into the tunnel under pressure in a closed chamber at the upper portal, thus causing an induced current in the open tunnel against an ascending train and driving the prod- ucts of combustion to the rear of the train instead of following it in its as- cent.' In providing for artificial ventilation on this principle, considera- tion must be given to the resistance offered by the current engendered by the ascending train which, in a single-track tunnel, is from two to three times that produced in the open. In a double-track tunnel, it is of little effect. In tunnels with gentle gradients and heavy traffic, an artificial current of ten feet per second has been found sufficient, regardless of the direction of the train .^ Leaky conduits materially reduce the mechanical efficiency of a ven- tilating plant. In a conduit with well-made joints, its mechanical effi-. ciency was 82 per cent, of its volumetric efficiency in a length of H miles, and but 50 per cent, in a length of 3 miles. The loss by leakage was 0.5 per cent, per 100 yards of conduit. The conduits should be installed during the prosecution of the tunnel-work. In a double-track tunnel, it should be under the tunnel. Where electric traction is used, the tunnel can be ventilated by putting in a shutter midway in the underheading. When the shutter is closed, air can be blown in at each entrance and escape 1 Enc. Brit. XXVII, 407. ''Air Resistance in Simplon Tunnel Clear Cross-section, 250 square feet Speed in Miles per Houk (Resiatance in pounds per ton of 2240 lb.) 31 37 434 Running in direction of current . . . In opposite direction .... Tn onpn air 11.20 15.68 7.39 14.34 21.73 9.18 17.92 28.00 11.20 ROADWAY 191 through adjustable openings. Where the traffic is ahnost incessant, no plan of ventilation yet devised has proved satisfactory under all condi- tions. If a tunnel exceed 1^ miles in length, it may be advisable to build it double-track ; though with an interior summit it may even then require artificial ventilation. But if artificial ventilation be required in a single- track tunnel exceeding 2 miles in length, the installation of the necessary plant, with the working expenses capitalized, may approximate the original cost of a double-track tunnel. The "Saccardo" system of ventilation, invented about 1896, is in .gen- eral use in Europe, except in the Simplon Tunnel. In this system, an annu- lar conduit occupies the space between the masonry-lining and the loading- gauge at the portals. In long and steep tunnels, a current of air is produced in the opposite direction to that of the passing train, of sufficient volume and velocity to overcome the air-current which accompanies the train, and, in addition, to neutralize any natural current in the same direction. This sys- tem has two defects ; the great quantity of power required, and its low efficiency, which could be increased from ten to fifteen per cent, if the ap- paratus at the entrance to a tunnel were arranged both to blow in the air and to exhaust it. In the Ronco Tunnel (double-track and four miles in length) between Genoa and Milan, exhaust-fans, 18 feet 10 inches in diameter and operated with 700 horse-power, produced a descending current at 10 feet per second and an ascending current at the same rate, which do not, however, free the tunnel from smoke sufficiently to render visible the block-signals in the middle of the tunnel. Two additional fans were therefore installed on either side of the signals ; one of 125 horse-'power driving air toward the upper end, and the other of 50 horse-power blowing in the opposite direc- tion. The total cost of these installations was about $150,000. In the Pracchia Tunnel (single-track and 1.7 miles in length on a steep grade) between Bologna and Pistoia, it was proposed to neutralize the cur- rent of air produced by an ascending train, and then to sweep out the tunnel by a current at 10 feet per second. With a double-header, at a speed of 15^ feet per second, the power required for this purpose was as follows : Nattjbal Cttrrent in Same Dibection WATER-PHBSSnRE Fan-bbvghttions PER Minute In Tunnel At Outlet of Injection Horse-power None 10 feet per sec. . . 9.84 12.28 14.17 18.82 86 99 loa 147 With a train moving at a speed of 33 feet per second, 122 horse-power was required to neutralize the accompanying current, and 233 horse-power to produce a current of 6^ feet per second in the opposite direction. The 192 EFFICIENT RAILWAY OPERATION original installation of 320 horse-power was increased to 440 horse-power, in order to obtain satisfactory results. At the Giovi Tunnel (double-track, 2 miles in length) on a gradient of 30 per 1000, near the Ronco Tunnel, fans were installed at the upper end, 20 feet in diameter and producing a current of 20 cubic feet per second, when running at 90 revolutions per minute. With 29 trains ascending daily, 23 triple-headers and 6 double-headers, on an average interval of 19 minutes, the air is breathable, but it requires from 20 to 25 minutes to clear the tunnel of smoke. At the St. Gotthard Tunnel (double-track, 9.2 miles in length) 4415 cubic feet of air are forced in per second. When assisted by a natural current of 6^ feet per second, 750 horse-power is required to reduce the vitiation to 6 per 1000 on the standard scale. The ratio of the work done in moving the hir in the tunnel to" the sum of the power developed at the motor and the resistance opposed to the train by the assisting current, is from 0.51 to 0.57, as against 0.40 at Pracchia. In the Simplon Tunnel, ventilation is provided by closing with a movable screen or curtain the entrance near which the ventilators are situated, and then changing the air, either by exhaustion or by blowing in fresh air. The curtains are of heavy canvas on iron frames, balanced by counter- weights and moving on rollers in grooved channels. They are raised by electric motors operated from a signal-cabin where approaching trains are indicated by automatic rail-contacts. Other signals notify the motorman whether the curtain is up or down. In case of misunderstanding, the tractor could tear through the curtain without causing serious damage. The installation at each entrance consists of two fans for blowing or for exhaust, each 12 feet 3 inches in diameter and actuated by a turbine of 200 horse-power. They are operated either in series for pressure, or in parallel for volume. The air is blown in at the north entrance and ex- hausted at the south entrance, with a volume of 3178 cubic feet per second at 275 revolutions per minute, which is sufficient for electric traction. The highest temperature is 80° to 82° F. in the south end, and 77° to 79° in the middle of the tunnel.' Recently Built Tunnels in America and their Ventilation Artificial ventilation of tunnels in America has followed about the same course as in Europe.^ An improvement upon the "Saccardo" sys- tem has been devised by Mr. Charles S. Churchill, while Chief Engineer of 'Most of the information as to European practice in tunnel construction and operation has been obtained from the Proceedings of the International Rail- way Congress at Berne, in 1910, in which the subject is considered at great length and with ample illustrations. For a list of the principal tunnels in Europe, see Appendix IV, Tables II and IV. 2 For a Ust of artificially ventilated tunnels in the United States, see Appendix IV, Table V. ROADWAY 193 the Norfolk & Western Railway, and has been generally adopted. It was an objection to the Saccardo system that the air-conduit surrounded all the sectional area of the tunnel at the portal and decreased the clearance limitations at that point. In the Churchill system, the air-conduit or nozzle is constructed outside of the portal in funnel-form, tapering down to its required sectional area about fifty feet from the entrance of the tunnel. This conduit therefore serves as a blower nozzle, through which fresh air may be forced into the tunnel without contracting its sectional area. The Elkhorn Tunnel on the Norfolk & Western Railway, and the first installation of this system, affords a typical example of its successful ap- plication. This tunnel is single-track, 3000 feet in length, on a gradient of 1.4 per cent, or 1 in 71, and with a sectional area of 235 square feet. Fresh air is delivered at the lower end at the rate of 400,000 cubic feet per minute at a velocity of 1700 feet per minute. The coal-traffic through the tunnel in recent yearfe was handled by two Mallet locomotives, one acting as pusher, at a speed of seven to eight miles an hour. The trains were from 50 to 60 cars, with a gross weight of about 3500 tons. These loco- motives consumed 173 pounds of coal per minute, emitting 43,000 cubic feet of gas ; including 496 cubic feet of carbon monoxide, 4253 cubic feet of carbon dioxide and 34,500 cubic feet of nitrogen. The bad and the inert gases in combination amounted to 21,000 cubic feet per minute during the 4^ minutes that the train was passing through the tunnel. As the ventilation was conducted, the smoke of the first locomotive as well as that of the pusher was forced ahead of the train, so that the men on the locomotives experienced no discomfort and the tunnel was clear of smoke as the train passed out. The increase of traffic, accompanied by the use of heavier locomotives, compelled a resort to artificial ventilation also in the Horse Shoe Tunnel on the same line, though on a much easier grade. This tunnel is single-track and 3291 feet in length, with sectional area of 300 square feet. The ventilation is by two electric fans at the upper end, supplying 540,000 cubic feet of air per minute. The Big Bend Tunnel, on the Chesapeake & Ohio Railway, is single- track, 6500 feet in length, with an ascending grade of 21 feet to the mile for two-thirds of its length from the west end and a descending grade of 4 feet per mile thereafter. It is ventilated by two blowing-fans, each of 14 feet diameter and 7-foot face. They are designed to deUver a current of air at 1600 feet per minute, but with 200 horse-power only 1200 feet per minute could be attained. These fans are placed at the east end of the tunnel and blow against the heavily loaded trains instead of blowing with them, as at the Elkhorn Tunnel. Owing to the length of the Big Bend Tunnel, if the fans had been placed at the west end, the speed of the trains would have had to be less than that of the air-current, say 1000 feet per minute in order that the smoke might be blown ahead of the train. Then, only one train could have been passed through in each 15-minute interval. 194 EFFICIENT RAILWAY OPERATION as the block including the tunnel is two miles in length. But with the fans at the east end, it is practicable to maintain a speed of 2000 feet per minute with a train interval of 7 to 8 minutes.' The Connaught Tunnel, on the Canadian Pacific Railway, is ventilated by fans at the higher end, which are driven by Diesel engines, each of 500 rated horse-power at sea-level. They are operated only when a train is ascending the tunnel, driving fresh air against the train and for a sufiicient length of time thereafter to clear the tunnel of gas. As oil is used for fuel on this section, the fire can be shut off in case of an emergency stop and there is no production of carbon monoxide.^ The cost of an installation on this plan has been from $25,000 to $40,000. Higher air-velocities are used than heretofore in European practice, which serves to keep the track in dryer condition. In many installations, the spacing of the trains requires the operation of the fans at full-speed only at short intervals. A single-track tunnel a mile in length is cleared in 4.8 minutes by fresh air supplied at a velocity of 1100 feet per minute and in 3.3 minutes at a velocity of 1600 feet per minute. In a double-track tunnel of the same length and of a sectional area of 450 square feet, with trains at 15 miles an hour, spaced at 5-minute intervals, the total emission from the stack of a heavy locomotive during the four minutes of its passage will amount to 43,000 cubic feet within the total contents of the tunnel, which is about 2,400,000 cubic feet. Under these conditions, the tunnel should be cleared in four minutes with a delivery of 600,000 cubic feet of fresh air per minute at a velocity within the limit of 1300 feet per minute. The tunnel would, therefore, be practically cleared of smoke within the five- minute intervals of passing trains and, even if this interval were somewhat diminished with increasing traffic, the gases remaining in the tunnel would be so greatly diluted as to cause but httle discomfort in operation. The increasing use of electric traction in badly ventilated tunnels will mate- rially lessen the necessity for artificial ventilation.' Subaqueous Tunneling. Examples of Constkuction The progress made in the art of tunneling was in course of time applied to passing beneath broad rivers and estuaries, instead of bridging them. The earliest example was the underground passage of the Thames River ; commenced in 1823 as a footway, but subsequently enlarged as a double- track railway tunnel. This work was made memorable by the employ- ment of a rectangular excavating shield for driving the tunnel through the 1 Railroad Gazette, February 20, 1893. ^ See "Ventilation of the Connaught Tunnel," J. G. Sullivan, Chief Engineer. The Cornell Engineer, February, 1917. ' The information as to artificial tunnel ventilation in America has been ob- tained from a paper by Mr. Charles S. Churchill, M. Inst. C. E., on "Ventilation of Tunnels and Subways in America," published in the Proceedings of the In- stitute of Civil Engineers, Vol. CC, 1914^1915, Part II. ROADWAY 195 silty river-bed, invented in 1818 by Sir Marc Isambard Brunei, father of I. K. Brunei. This shield was of iron, 36 feet wide and 22 feet high, covering the face of the work. It was of cellular construction with twelve cells, each 3 feet wide and 22 feet high, subdivided horizontally into three cells, closed with sliding shutters. The shield was forced against the silt by hydraulic jacks and the oozy material was excavated with safety, as it was removed in such small quantities that the mass could be kept under control; though the workmen's candles were frequently extinguished by the occluded gases. The shield was moved ahead through an iron tube in which the masonry-hning followed. Depressions formed in the bed of the river were filled up with clay, but the water broke through in such quantities in May, 1827, that the work was stopped and not renewed until 1836. It was completed on March 25, 1843, at an average cost of $6500 per Hnear yard. The work could have been executed more cheaply and expeditiously, had it been projected on a level fifteen feet lower, where it would have been driven through the stratum of London clay, which is im- pervious to water. In 1865, P. W. Barlow patented a circular shield in connection with a cast-iron Hning of the completed tunnel back of the shield, and, in 1869, associated with J. H. Greathead, he constructed a subway of this character under the Thames at Tower Hill. A vertical shaft, 10 feet in diameter, was sunk for 60 feet into the clay stratum on each bank of the river, and tliese shafts were connected by a tube 1350 feet in length and of seven feet internal diameter. The shield of ^inch plates was slightly tapering, with the larger diameter forward, to reduce the skin friction. As the shield advanced, the space around the tube was grouted by compressed air. In the same year, A. E. Beach patented in the United States the use of hy- drauUc rams abutting against the lining of the completed tunnel for pushing the shield forward, and utihzed his invention in 1873 in an attempt to build a subway under Broadway in New York City.* The next important subaqueous construction was that of the Mersey River Tunnel between Liverpool and Birkenhead. After preliminary operations in 1879, the main tunnel was begun in August, 1881, and opened for traffic in January, 1886. It is about 4000 feet in length and 26 feet in width, with a brick Hning 2.25 feet in thickness, and was driven at a depth of 180 feet below the level of the quay. It was intended to keep the bore in solid rock, but, as with the Kandersteg Tunnel on the Loetschberg line,^ the geological predictions were at fault. For a distance of 200 feet, the crown of the tunnel was six to seven feet above the rock, in the ancient river-bed filled with glacial drift, bowlders, clay and sand, under a depth of 100 feet of water. With the aid of powerful pimiping machinery, this part of the work was accomplished without the use of an excavating shield or of compressed air. The tunnel was operated by steam for twenty years > See p.- 208. ^Seep. 175. 196 EFFICIENT RAILWAY OPERATION before the introduction of electric traction.' The subway connecting East Boston with the city proper, 1.4 miles in length, of which 3400 feet is under the harbor, is the first important example of a shield-built, mono- lithic concrete arch. The earliest subaqueous tunnel in America was built in 1889-1890, to carry the Grand Trunk Railway under the St. Clair River between Sarnia and Port Huron, and is 1.13 miles in length between portals. In 1904-1905, the New York Rapid Transit System was carried under the Harlem River through tubes, 15 feet in diameter and 400 feet in length. The top of the tunnel is 28 feet below high water and 3 feet below the river- bed. McBean, the subcontractor, who afterward built the Subway Tunnel under the same stream,^ here adopted a novel and ingenious method of con- struction. A trench was dredged to within seven or eight feet of the required depth and a space inclosed of the width of the tunnel from shore to mid- stream with sheet-piling cut off two feet above the outside top height of the tunnel. On this piling was tightly fitted a temporary flat timber- roof, three feet thick, covered with five feet of dredged mud. Water was expelled from this chamber by compressed air, and the lower half of the tunnel was built inside of concrete, surrounded by a cast-iron shell. The sheet-pihng was then cut off at mid-height of the tunnel, and the upper part of the tunnel-shell was lowered in sections through the water to serve as a roof until the upper half was completed. CoMPKESSED Air Method A great advance in the construction of subaqueous tunnels followed upon the application of the experience gained in the use of compressed air for bridge-foundations. The caisson or air-lock was utiHzed horizontally instead of vertically, in connection with the excavating shield, and with far-reaching consequences. It is said to have been so employed in Ant- werp, in 1879, but it was brought into more prominent notice in the same year in the construction of a tunnel under the Hudson River from Hobo- ken to New York City. This enterprise was projected by D. C. Haskins, financed by British capitalists, and was planned with two tubes of brick- masonry, 16 by 18 feet. When the northerly tube had been driven about 1200 feet from the New Jersey shore, the air blew through the silt, twenty men were drowned, the tube filled with water and the work was temporarily abandoned. In 1886, the City & South London Railway was carried beneath the Thames. Because of the exorbitant demands for a site on the river-bank, the headings of the tunnel were begun from a cribwork in the middle of the river, from which shafts of cast-iron rings, 13 feet in. diameter, were sunk 1 The system of drainage in this tunnel is described on page 174 (note), and that of its ventilation on page 190. 2 See p. 205. ROADWAY , 197 into the clay. Here the hoisting machinery was placed and material expeditiously removed and deUvered by barges. The land-damages for the approaches under the shores were also lessened by superimposing one tube over the other. These tubes were 10^ feet in diameter, of cast-iron plates nearly an inch thick, in rings of 20 inches in length connected by three-inch flanges. Where sand and gravel were encountered, the air- lock was used ; but the shield was not used as an aid in excavation until it was found that the work could be greatly expedited by arming its cutting edge with projections that broke up the clay, as it was forced forward by the hydraulic jacks ; and an advance was attained of 13 to 16 feet in twenty- four hours. The doorway in the face of the shield was placed near the bottom, so that, in event of a sudden inflow of water, the air-pressure would still maintain a breathing space in the upper part of the shield. The Baker Street & Waterloo Railway, opened in March, 1906, was also carried under the Thames from a cribwork in the river. This tunnel, of two tubes, 12 feet internal diameter and 23 feet apart, was driven through a bed of sand and gravel in a deep depression in the clay, of which no indication was given in sinking the foundations of the Charing Cross Railway bridge, 250 feet away. As the shield advanced, it was overloaded with clay, and the air-pressure was varied from 21 to 35 pounds per square inch, to suit the height of the tide. Blow-out pipes were carried into the shield, to sweep out the vitiated air, before it became mixed with the tunnel- air between the air-lock and the shield. The Glasgow District Tunnel, 1891-1896, was carried beneath the Clyde under air-pressure. ' Before it had advanced 80 feet, and only 13 feet under the river bed, it had been blown out ten times ; but the work was continued by filling the holes with clay. With the experience thus gained in the use of compressed air, work on the Hudson River Tunnel was resumed in 1890 by British engineers, and extended for 1800 feet, when it was again suspended because of financial difiiculties. In 1901, the company was reorganized by W. G. McAdoo, as the Hudson & Manhattan Railroad Company. An iron fining was substituted for the brickwork, a parallel tube was added, and the tunnel was opened for traffic on February 25, 1908, as an electric railway connect- ing the Delaware, Lackawanna & Western Railway station with stations on Sixth Avenue at Fourteenth and Nineteenth streets. The successful inauguration of this tunnel was an incentive to undertaking similar work, to facifitate the entrance of the Pennsylvania Railroad into New York City, which had been hitherto impracticable. From this beginning there have resulted examples of underground engineering in and around that city, unequaled elsewhere for the ingenuity displayed in design and for celerity of construction under unusually difficult conditions. In sub- aqueous tunnels alone, there are the original twin tubes of the Hudson & Manhattan Railroad, its second pair of tunnels from Cortlandt Street to Jersey City, subsequently continued to a junction with the Pennsylvania 198 EFFICIENT RAILWAY OPERATION Railroad tracks at Marion ; the two tunnels of the Pennsylvania Railroad under the Hudson River and under East River to a connection with the Long Island Railroad; the "Belmont" Tunnel under East River; the two twin tubes of the New York Subway under East River and one pair under Harlem River ; also a projected tunnel under East River to relieve the traffic over the Queensboro Bridge. In connection with these subaqueous works there are three independent systems of subways in New York City ; one from the Cortlandt Street ter- minal of the Hudson & Manhattan Railroad to a connection with the orig- inal Hoboken-Tunnel line under Sixth Avenue and beyond Nineteenth Street to Thirty-third Street ; the Pennsylvania Railroad subway from its Hudson River tunnel to its city terminal and beyond to its East River tunnels; and the municipal system of subways as originally operated, and as now being extended to connection through its East River tunnels to Long Island, and through its Harlem River tunnel into the Bronx and beyond. The Hudson & Manhattan Railroad Company has also built a connection beneath the New Jersey shore between its terminals in Hoboken and in Jersey City, and, jointly with the Pennsylvania Railroad Company, is operating a suburban hne from its Cortlandt-Street terminal to a transfer station in the outskirts of Newark for a connection with the main-line trains of that company, and over an independent line beyond into New- ark.' The tunnels of the Hudson & Manhattan Railroad were driven through the silt simply by displacing it. Instead of permitting the material to enter the openings, the shield was pushed bodily forward ; the semi-fluid silt flowing around the tube as it progressed. Great care was required where the tube passed out of the sand or silt into rock which required drilling or blasting. At such places, the river-bed was covered with a layer of clay to enable the air-pressure to be adequately maintained without caus- ing a blow-out between the soft and the hard material. In one of the subway-tunnels, a man was blown through thirty feet of ooze and water into the open air, yet he escaped unhurt. The ordinary maximum pres- sure was 36 pounds per square inch ; though in joining the tubes, as the ends met, the pressure was temporarily increased to 51 pounds. In the sub- way-tunnels, the air-lock was not used. The shield was driven forward by seventeen hydrauHc jacks, each exerting a thrust of 350,000 pounds, and by their relative pressure the tube was kept in aUgnment under the vary- ing conditions of external resistance. Under 22 pounds' pressure, the men worked seven and a half hours in an eight-hour shift with a half-hour in- terval ; under 35 pounds, they worked in two shifts of two hours each with a four-hour interval ; under 45 pounds, they worked but one hour and twenty-five minutes in eight hours. ' Details as to tunnel construction in and around New York City are given in Appendix IV, Table VII. ROADWAY 199 The Pennsylvania Railboad Tunnels All previous subaqueous tunnel-work was dwarfed in the construction of the Pennsylvania Railroad tunnels under the Hudson and the East rivers. The bed of the Hudson River is of silt, so loosely compacted that it would have been impracticable to drive a tunnel through it in open air. The main level of the tunnels is therefore kept at a great depth, so that the weight of the superincumbent material may be sufficient to resist the neces- sary air-pressure. Indeed, the silt is of such a light consistence that much of the advance was accomplished by pressing against it with such force that it exuded through the openings in the shield. The twin-tubes of the Pennsylvania Railroad tunnels are built of cast- iron rings, 2^ feet long and 23 feet in diameter, each composed of eleven segments and a key, and of a total weight of 11^ tons. They are lined with concrete, leaving 18 feet clear internal diameter. A concrete walk, 6 feet in width, extends along each side of the tube at the height of the car-win- dows, in which are imbedded the conduits for the water-pipes, the power- cables and the signal-wires. As there is but six inches' clearance above the car-roof, ventilation is assured merely by the trains forcing the air ahead of them and the fresh air following. When the faces of the Hudson River tubes, 6000 feet in total length, had approached within 125 feet of each other, the alignment was tested by connecting the parallel tubes on each side with six-inch pipes, through which instrumental observations were taken. It was found that the tubes were out of ahgimaent one-eighth of an inch horizontally and three-fourths of an inch vertically. The Sunken Tube Method Another plan for the construction of subaqueous tunnels has been successfully apphed under suitable conditions. In 1845, De la Haye, of England, suggested making a submarine railway by constructing wrought- iron tubes above water in sections 400 feet long, bulkheading them so that they would float, towing them to the tunnel-location, then admitting water and sinking them to a suitable bed. M. Belgrande, in 1866, built a pair of sewer-tunnels on this plan under the Seine at Paris. Each had a diameter of one meter, was 156 meters long and was made of iron plates. The first masonry-tunnel on this plan was constructed in 1893-1894 by Mr. H. A. Carson, in the outer portion of Boston Harbor, for the Metropolitan Sewer. The sections of 50 feet in length were made of a combination of brick and concrete, with exterior wooden staves, four inches thick. Their external diameter was a fittle more than nine feet, with external flanges for bolting contiguous sections. Temporary bulkheads were inserted and were tested for tightness by exhausting the air and measuring the rare- faction by a vacuum gauge. They were made in cradles above the water, 200 EFFICIENT RAILWAY OPERATION were lowered by vertical screws and towed to their positions, filled with water and lowered to saddles sunk in trenches dredged in the harbor- bottom. The sections were bolted together by divers with rubber gaskets between the flanges. The trenches were filled in, the bulkheads removed and the water pumped out. The spaces which the bulkheads had occu- pied were then closed with masonry. The whole tunnel, 1500 feet in length, was water-tight, true to hne and level, and satisfactory in every way.^ The Detroit River Tunnel The most important subaqueous railway tunnel in America, outside of New York City, is that under the Detroit River, built by a combination of the railroad companies interested in the international traffic between, Detroit and Windsor. The Detroit approach of 3669 feet includes an open cut of 1510 feet ; the subaqueous portion is 2668 feet and the Windsor approach is 6449 feet, including an open cut of 2900 feet. The total length of the line as constructed is, therefore, about 2.42 miles. The tunnel consists of two single-track tubes with a maximum gradient of 1.5 per cent, on the Windsor approach and of 2 per cent: on the Detroit approach and a maximum curvature of two degrees. The entire excavation was in blue clay. On the Detroit side, it was found necessary to use compressed air for 450 feet from the river, and for nearly all the way on the, Windsor side. The approaches are in plain concrete, except for some longitudinal rods in the inverts. In Windsor, the clay came in at the working-face about as fast as it could be removed, and excavation for the dividing-wall between the tubes was carried on for about 2000 feet with a shield above the bottom-drift. As the subaqueous section was approached, the move- ment of the shield became so erratic that it was abandoned, and the re- mainder of the wall was built in headings under compressed air. The work was then completed with "side-shields," semi-elliptical in shape. These shields extended from over the springing of the arch in the center- wall around the bottom of the invert, including a ring of timbering, 10 inches thick, outside of the concrete. The outside shell was of f-inch steel, reinforced with angle-iron. The main body was 7^ feet long, with a heavily reinforced cutting edge. It was divided on the working-face for six pockets, through which the clay was excavated and deposited on a tail-piece of ^-inch steel, until the timbering was placed. Each shield was supplied with twenty-one 5-inch hydraulic jacks, with 20 inches extension, worked up to 8000 pounds per square inch. Later, the two bottom-jacks were replaced by 8-inch jacks, and another 5-inch jack was added next to the central wall on top. The clay was cut with semi-circular knives, pulled by two Tnen and guided by a third. By means of a conveyor, operated by compressed air, the 1 H. A. Carson, "Discussion on the Detroit River Tunnel." Trans. Am Soc. C. E., December 1911. ROADWAY 20 1 excavated material was loaded upon narrow-gauge cars standing on a track carried by a working-platform behind the shield. These cars were cable- hauled to the shaft, elevated to the surface and dumped on fiat cars. The excavation was carried about two feet ahead of the shields. As the shields traveled along the central wall, there was little trouble in keeping them in line, but it was more difficult to keep them to grade, than if they had been of a full section with jacks on all sides. In the drifts,' the usual rate of progress was about twelve feet in twenty-four hours, and with the side- shields about ten feet. Compressed air was used as required. The De- troit approach was built in the same way. The water-proofing in the approach-tunnels was made of alternate layers of felt and coal-tar pitch, varying in thickness from seven to twenty- two layers. The concrete was laid in runs of twelve feet and each run was allowed to set sufficiently for the form to be moved before the succeeding run was made. The concrete for the center-wall was deposited from the surface through chute-holes, except for the work under compressed air, where the top-drift was made high enough for the concrete for the entire wall to be delivered in cars. The subaqueous portion of the tunnel is 2668 feet in length and is level for about 1000 feet, but both ends are on curves. The method of construction was entirely new, as applied to subaqueous tunnels. The plan consisted in dredging a trench of the required width and depth, sink- ing to correct position water-tight steel tubes and surrounding them with concrete deposited in the water. Before the tubes were unwatered, the trench was back-filled with clay to the full height of the exterior concrete deposit. By means of two water-tight openings, the concrete-lining was built in the atmosphere. The tubes are 23 feet 4 inches in diameter, of f-inch plates,, sunk in pairs, 26 feet 4 inches between centers. The sections were 262 feet 6 inches long, except one of 238 feet 6 inches and the closing section of 64 feet 6 inches. The tubes were reinforced by circumferential stiffener angles, 4 inches by 3 inches and f inch thick, riveted to the inside at intervals of 12 feet. They were further stiffened for sinking by 12 steel rods, I inch square, radiating from a cast-steel ring in the center of the tube, and bolted to the stiffener angles. For depositing the exterior concrete, the tubes were sur- rounded with steel diaphragms at intervals of twelve feet, to the ends of which was attached wooden sheathing, six inches thick at the bottom, di- minishing to three inches at the top. Each pocket thus formed could be filled from top to bottom, independently of the adjacent pockets, with con- crete of a minimum thickness of three feet. The bottom of the diaphragms was made straight to give a level bearing. The bottom of the trench was from 60 to 80 feet below the mean water surface. The average depth of the river-bed is 36 feet, with a minimum depth of 18 feet, and a maximum of 48 feet. For a distance of several 202 EFFICIENT RAILWAY OPERATION hundred feet in the middle of the river, the top of the tunnel is from three to seven feet above the river-bed, with a depth of 41 feet in the main channel. It was found that a dipper-dredge could not be used on account of the strain upon the anchoring-spuds, and the dredging was done with a clam-shell bucket of three cubic yards' capacity. The average perform- ance was 600 yards of clay per 10 hours, with a maximum of 1400 yards per day and of 25,000 yards in one month, place-measurement. Care was taken not to waste clay from Canadian waters across the international boundary, in order to prevent the imposition of United States customs' duty. The depth and width of the trench was ascertained by sweeping it with a 24-inch "I "-beam, 48 feet long, suspended from a derrick scow. At the points for joining the tube sections, a grillage of "I "-beams was placed at the proper level upon a deposit of concrete, by lowering it from derricks on the scow which carried the apparatus for sinking the tubes. The grade was determined by a steel mast, 80 feet long, which served as a level rod, and was also suspended by a derrick. The ends of the tubes were fitted with bulkheads, four inches thick. In each was a 14-inch gate-valve just below the line of flotation and, at the top, a 2-inch air-escape valve with hose reaching above the surface, after the section had been sunk to its position. There were two semi- bulkheads in each tube, 9 feet deep and 48 feet from the end-bulkheads. These bulkheads formed an air-chamber at each end, by which the section was kept on an even keel, by controlling the sinking through the air-escape valves. As the section sank, the air from the middle of the tubes escaped through a 4-inch pipe back of each bulkhead. The openings for these pipes were capped by divers afterward. The time required for submer- gence of a section was about two hours. Four air-cylinders were used for floating the sections, two over each tube and connected to the diaphragms. Each cylinder was 10 feet 2 inches in diameter and 60 feet long, divided into three compartments, fitted with water-valves and air-escapes. These compartments were so proportioned that, with the end ones empty, and the tubes filled with water and com- pletely submerged, about 18 inches of the top of the section was above the surface. The middle compartment was then filled and the section was lowered into position. The submergence was more accurately controlled by two 5-ton counterweights suspended from derricks on scows along- side, by which additional load could be imposed as required, and the section could be raised or lowered a fraction of an inch by regulating the water in the air-cylinders. After the section was in position and suffi- ciently anchored by concrete, the air-cylinders were detached by divers and used for the next section. The total weight as submerged was 550 tons of metal, 240 tons of wooden material and 110 tons of air-cylinders : 900 tons in all, weighing 516 tons when entirely submerged. The section was held in position against a maximum current of 3| miles an hour by ROADWAY 203 concrete anchors weighing 22 tons each, under water, and buried 700 feet up-stream. The sinking apparatus was carried on a derrick-scow, 35 by 120 feet. The sections were connected by a joint consisting of a "pilot-pin" 6 feet long and 6 inches in diameter, fitting into a socket in the end of the pre- ceding section. The end of the pin was slotted to receive a key which, when driven home, secured the end of the section in position. An annular pocket, 4 inches deep and 8 inches long, was made water-tight by rubber- gaskets and, after the tubes were unwatered, this pocket was filled from inside with grout. The connection was further strengthened by bolts outside, placed by divers. The position of the section as to line and grade was determined by steel masts, one on each tube at the east end and one on the west-bound tube at the west end. Each face of the masts was marked with a center-line and graduated in feet, referring to the center of the tube. The grillages were set several inches below the bottom-grade of the diaphragms, and the level accurately secured by shims placed by divers. When a section was finally deposited, it was anchored to the grillage by a turnbuckle. The total time consumed in sinking a section and in making ready for the concrete was from three to seven days. The concrete-mixers were carried on a scow, 36 by 155 feet, fitted with spuds 20 by 20 inches and 90 feet long. On one side of the scow there were three tremie-pipes, 26 feet 4 inches between centers and extending 89 feet above the water. The pipes were in 16-foot lengths, 12 inches in diameter. When a section of tubes was in place, the tremie-pipes were lowered into one of the pockets formed by the diaphragms, to within a foot of the bottom of the trench, with the middle pipe between the tubes and the others out- side. The lower ends of the pipes were usually held from three to five feet in the concrete and the flow regulated so that the pipes were kept filled with concrete during the run. The placing of the concrete was in- spected by a diver, who saw that it was carried to about six inches above the top of the diaphragms, and this was verified by soundings. There was a hydrostatic pressure of about 30 pounds per square inch at the bottom of the tubes. The total quantity of concrete, 101,900 cubic yards, was placed in a working-period of 18 months. The maximum quantity placed in a day of 16 hours was 1069 cubic yards, and the maximum for any one month was 10,287 cubic yards. For the back-filling, gravel was used to a height of eleven feet above the foundation concrete. The remainder was filled with clay, as excavated by the dredgers, and then covered with riprap to the depth of two feet. For unwatering the tubes, the west end of the westernmost section and the east end of the other alternate sections had beeri fitted with bulkheads of 10-inch by 12-inch yellow-pine timber, heavily braced to wit"hstand the hydrostatic pressure, with a 10-inch valve for unwatering in each bulk- head. An opening was connected through every two alternate sections by 204 EFFICIENT RAILWAY OPERATION means of the 14-inch inlet-valves used in filling the tubes. An access-pipe, 4 feet in diameter and extending above the surface when the section was in place, had been provided at the west end of each tube in the first section, as also a 12-inch discharge-pipe. Centrifugal pumps were placed under these pipes and the pump-shafts led up through them. For carrying water to the pumps, a 12-inch pipe had been laid through each of the first five sections in advance of the sinking. After the first section had been un- watered, the remaining sections were un watered in pairs. When the sections were all unwatered, they were foiind to be watertight. Tha concrete-lining of the full section, except over the access-pipes, was placed in four runs ; viz. : the invert to 12 inches below the bench- wall ; the two side-walls to 8 feet above the bench, and finally the arch. The concrete was mixed on the surface at the Detroit end and delivered through chutes in the access-pipes upon a platform suspended at the level of the top of the side-walls, and advanced with the progress of the work, the narrow-gauge track being shifted from the platform to the completed invert. The cars were drawn by cable and thirty minutes was the longest time required to deliver the concrete to the Windsor connection ; a dis- tance of more than 2700 feet. Steel forms were used for the lining, 24 feet long for the inverts and 12 feet for the arch. The lining was completed in exactly twelve months. The connection between the subaqueous tunnel and the western ap- proach was made in the dry, in a coffer-dam. At the eastern approach, the end of the approach-tunnel had been covered by a steel plate, which served as a water-tight surface against which the tubes abutted. To prevent an inrush of water into the approach-tunnel, three concrete bulk- heads with steel doors had been built into the work, 50 feet apart. The closing section. No. 11, was 64 feet 6 inches long, and the gap to the next full-section was 65 feet 4 inches, when No. 11 had been placed against the end of the approach tunnel. The interval of ten inches between the last two sections was closed by divers with a joint of special design. The closure was sealed with concrete when the tubes were unwatered, and covered by a plate bolted to 1;he tubes. The difference in level through the completed tunnel was ^ of a foot. The contract for the tunnel was awarded August 1, 1906. The work was commenced October 1, 1906, and completed July 1, 1910. The tunnel was opened for all traffic in October, 1910. It has overhead clearance of 18 feet to the top of the semi-circular arch, which has a radius of 10 feet. The top of the bench-walls is 5 feet 3 inches above the rails and the distance between them is 11 feet 6 inches. Safety-ladders, staggered in 25-foot intervals, are built into the bench-walls. The tunnel is lighted by 16-candle- power 110-volt incandescent lamps, staggered every 20 feet, with two separate circuits. Conduits for telegraph, telephone, power and signal circuits are built into the walls, also splicing-chambers every 400 feet. All ROADWAY 205 water is drained into five sumps ; one at each portal, one in each approach near the river, and one under the middle of the river, with capacity of 30,000 to 40,000 gallons. Each sump is provided with a pump and electric motor. Under the center of each track is a gutter, 6 inches deep in the approaches and 10 inches in the subaqueous tunnel. The tunnel is pro- vided throughout with a 6-inch water-line, and hydrants and alarm-boxes every 100 feet. As the tunnel is operated electrically, there is no special system of ventilation. The track is of special construction. Each rail, of 100-pound section, rests on ties 8 by 11 inches and 3 feet long, except every fifth tie, which is longer to receive the third-rail on the right-hand side in the direction of the traffic. The inner ends of the ties are at the edge of the gutter. The ties are held in place by concrete to a depth of 5 inches at their outer ends, with one-inch slope to the gutter. The line is operated with purchased current, three-phase, 60 cycle, 4400 volts, converted at a substation costing $260,000. The third-rail is of the under-running type, 70-pound section, and of a total length of 18-^ miles ; though the third-rail section is only about 4.4 miles. The signahng is arranged to make one block between the portals and the interlocking plants in each terminal yard operated electrically. The yards are lighted by arc series-lamps. The location of the approaches destroyed two terminal yards in De- troit and one in Windsor. The latter was readily removed to a more convenient location. In Detroit, it was necessary to reach the main switching yards, three miles west of the tunnel. This required the cross- ing of some thirteen streets by separating the grades and, within the limits of 2| miles, there were over thirty industrial tracks. A union passenger- station was also built in connection with the tunnel.^ The construction of the Detroit River tunnel is remarkable for the ingenuity of the method devised for the construction of the subaqueous portion, for the thoroughness with which the details were prepared, for the rapid progress of the work and its successful completion without serious accidents. It was estimated that by this method of construction, $2,000,000 had been saved, with the advantage of building the tunnel on a level fifteen feet higher than would have been practicable with com- pressed air, and a corresponding saving of 1750 feet of length in the ap- proach-tunnels. The sunken-tube method of construction was employed in carrying the Lexington Avenue Subway under the Harlem River through a tunnel on which work was commenced in 1914. This tunnel is in four tubes, each 19 feet internal diameter, incased in quadruple sections 200 feet in length. Here the descent was regulated by the admission of compressed 1 " The Detroit River Tunnel," W. S. Kinnear. Trans. Am! Soc. C. B., D cumber, 1911. 206 EFFICIENT RAILWAY OPERATION air into chambers at the ends. The "buoyancy of the sections was over- come by loading them on top as they sank between guide-piles to their position in 20 to 26 feet of water ; the top of the tunnel being 28 feet below high water. In all these cases of subaqueous tunnels, the depth below the water line is so much less than the necessary height above it for bridge clearance that the resulting gradient is considerably easier. The expensive con- struction of piers is avoided and their obstruction to navigation ; as also the occupation of valuable riparian property for approaches. Railway Undeegbound Approaches and Subway Systems in Large Cities The difficulties encountered in the entrance of railways into populous cities have been overcome by expensive construction and by engineering skill. The preservation of a terminal station in the heart of a great city could only be secured by elevating the approaches to it upon viaducts, through which the streets are passed, as with the London & Blackwall Railway. The exigencies of city traffic were partially relieved in Berlin by the Stadtbahn, opened in 1878. This is a four-track line of twenty , miles ; twelve miles of it costing $18,000,000. It is carried across the city for the most part on embankments and masonry-arches. An elevated railway of the same kind was built in Liverpool in 1894. But the traffic along the streets could not be conducted by railway lines on masonry, and recourse was had to iron viaducts spanning the roadway. . An extensive system of this character was developed in Chicago beginning about 1880, and another in Boston twenty years later. Structures of this character have cost from $300,000 to $400,000 per mile, exclusive of equipment, terminals and land-damages. Though the iron viaduct was preferable as a substitute for the surface-railway as a means of rapid communication, the piers are a serious obstruction to street-traffic, and the nuisances caused by passing trains to the occupants of adjacent buildings were but little diminished by the introduction of electric traction. With the improvement in underground construction, relief was sought by placing the railway lines beneath the streets instead of above them; a change of level that was invited where rivers intervened in the approaches. These "subways" may be said to have originated in Paris in the construc- tion of passage-ways beneath the streets for sewers and for pipe lines. ^ Now, the underground mileage of urban and suburban lines, constructed to meet social requirements, far exceeds the railway-tunnel mileage rendered necessary by topographical conditions. 1 In 1843, a subway was built beneath Atlantic Avenue, Brooklyn, as an ap- proacli by the Long Island Raib-oad to its former terminus at South Ferry, but was afterward filled in. It was a double-track tunnel, 21 feet wide and 17 feet high. George T. Hammond, " Discussion on Railway Development." Trans. Am. Soc. C. E., December, 1911. ROADWAY 207 The introduction of railways into cities by subways was undertaken on a grand scale to expedite the suburban traffic with the metropolis of London by the construction of the Metropolitan Railway, which was opened in successive sections in 1863, 1865, 1868, 1871 and 1876. The circle around the city was completed in October, 1884, by the construction of the District Railway. The total route-mileage of 101.8 miles was built at a cost of $132,000,000, and was operated by steam, until electric trac- tion was introduced in 1905. The internal section was' made with a width of 28^ feet to accommodate trains of the Great Western Railway, which was then of a gauge of 7 feet f inch. The line was kept so close beneath the surface that the platforms were reached by stairways at a minimum of 18 feet from the street. The work was carried on by "cut and cover." The trench was excavated in the open and was filled in after the masonry was built. Where the line was very near the surface, the space was bridged with heavy girders with jack-arches between them to carry the pavement. The tunnel, which is elliptical in cross-section, was built without inverts, and it became necessary to introduce concrete struts between the walls and beneath the tracks to resist the external pressure. The more recent subways in London have been constructed in tubes, with a total mileage of 35 miles and at a total cost of $105,000,000. The tracks are usually in single tubes from 10^ to 16 feet in diameter internally, enlarged at the stations. Though for the most part driven with exca- vating shields, compressed air was but seldom used. They were built on a much lower level than the Metropolitan Railway, to avoid the maze of underground conduits, to diminish danger to the foundations of adjacent buildings, and, in some instances, to pass one tube under another.^ Access from, the street is usually had by "elevators or lifts, at depths of 100 to 190 feet, accommodating from 60 to 75 persons, and limited to a maximum speed of 200 feet per minute. Moving stairways have also been tested experimentally.^ ' In 1893, a subway for electric traction was constructed in Buda-Pest, Htmgary, by "cut arid cover," with the rail-level but nine feet beneath the surface of the street, which was carried on steel beams close together with jack-arches between them, and a total thickness of only twenty inches. The entrances to the stations are through "kiosks" at the street corners. With the exception of the Untergrundbahn, now under con- ' In the construction of the Glasgow Central Railway 7J miles in length and opened in May, 1890, it was necessary to provide underpinning on each side of Argyle Street for two miles. Sheet-piling, 12 by 6 inches, was driven to a depth of 25 to 30 feet from an overhead traveling stage, moving on wheels at the street- level. ^ Most of the information as to tunnel-work in Great Britain has been obtained from a paper by Francis Fox, Mem. Inst. C. E., published in the Proceedings of the International Railway Congress in Berne, in 1910, from which have also been compiled the statistics as to London subways in Appendix IV, Table VI. 208 EFFICIENT RAILWAY OPERATION struction in Berlin, the only other subways of importance in the cities of continental Europe are in Paris. The Metropolitan Railway, with about 50 miles of line, begun in 1898, is for the most part underground and is carried under the arms of the Seine for distances respectively of 306 feet and of 132 feet by tunnels driven with compressed-air caissons. The new terminal station of the Paris & Orleans Railway, at the Quai d'Orsay, is also approached through a subway along the bank of the Seine. The subways for the street-railways in Boston and in Philadelphia have been constructed by "cut and cover," with the exception of the Boston subway to East Boston through a subaqueous tunnel.' The sepa- ration of the railway and street traffic in and around Chicago has been prosecuted since 1892 in a complication of steam and trolley lines that is unparalleled elsewhere, as it includes either the elevation or depression of 838 miles of steam-railroad tracks. Chicago has also a unique system of subways for freight-traffic only, underlying its business-district, and connecting the basements of important business-houses with all the freight- terminals. The tunnels of concrete, with cross-sections of 12f feet by 16 feet and 7^ feet by 6 feet, are driven through firm clay. In 1905, there were 65 miles of this subway, operated by trolley-motors on a two-foot gauge. Elevated Railways and Subways in New York City Subway systems have been developed in and around New York City which, as to cost, extent and engineering difficulties, are of a magnitude that has commanded attention and admiration throughout the world. The character of the work has been diversified by reason of the geographical position of Manhattan Island, isolated by broad rivers from the populous shores of New Jersey and of Long Island, and by the configuration of the island itself, which has restricted the local lines of communication within a long and narrow space. Its area is closely covered with buildings, many of whose foundations are based on a soil saturated with subterranean waters, and others with basements two to three stories beneath the surface, extended as vaults under the sidewalks. Beneath the pavement between these vaults, there is a maze of sewers, of gas and water mains, and of con- duits for electric cables, uncharted and many of them in a precarious state of disrepair. An effort to relieve the streets of their congested traffic was attempted in 1868 by A. E. Beach, editor of the Scientific American and inventor of the plan for pressing tunnel-shields forward by hydraulic rams.* He ob- tained a charter for a conduit under Broadway for the transmission of parcels by pneumatic propulsion. In 1873, the project was amended to undertake the construction of a subterranean railroad from the Battery to Harlem River. The work was commenced near the Astor House, and a section of the tunnel had been completed, when it was interrupted by 1 See p. 199. « See p. 195. ROADWAY 209 litigation that effected its abandonment. In February, 1912, it was found to be still in good condition. In 1870, a street-railway was built in New York City, with a track supported by a row of iron columns along each curb. This was replaced in 1878-1879 by a more substantial structure spanning the carriage-way and capable of carrying, four tracks. From this beginning, have been developed the extensive systems of elevated railways in New York and Brooklyn. The first important underground railway work in New York City carried the New York Central line beneath the street-crossings to Forty- second Street. It has been recently rebuilt on a far more extensive scale in connection with the construction of the new Grand Central Station. In its present form, the approach is so completely underground with structural supports of immense strength that the surface has been made available for building-sites of great value. The Pennsylvania Railroad line has also been brought into its new terminal station in New York City by a subway-approach from the Hud- son River tunnel and extended to the East" River tunnels through Thirty- second and Thirty-third streets in subways 42 feet wide and 21 feet high ; the tracks being separated by a dividing-wall. Beyond the eastern tunnel- portal, the line of the Long Island Railroad is continued for some miles by either elevating or depressing its tracks, with a further extension to the Port Morris terminal of the New Haven line, by bridging the three outlets of East River into Long Island Sound. ^ The subway in connec- tion with the Hudson & Manhattan tunnel has already been mentioned in this chapter. The municipal subways in New York City had their inception in the organization of the Rapid Transit Commission to facilitate passenger- travel within the city-limits. Work on >the original Rapid Transit System was commenced March 25, 1900, and the line was opened to Broadway and 145th Street, October 27, 1904. The 20.47 miles of track, including the Lenox Avenue Branch, of which five miles is on viaducts, cost $40,000,000. The extension under East River into Brooklyn was opened January 9, 1908. The system was completed with the extension to Van Cortlandt Park, when its line of 20 miles from end to end had cost $60,000,000. A subway was next built as a loop between the Brooklyn and the Williamsburg bridges, and then followed the project for the second system. The entire system of the dual subway, when completed, will comprise 629 miles of tracks at an estimated cost of $507,000,000. Taking all these public works together, including the East River bridges, the in- vestment within and immediately around New York City for facihtating personal intercommunication only, by elevated and underground rail- ways, may be estimated at not less than $1,000,000,000. In addition, a ' See Appendix IV, Table VIII. 210 EFFICIENT RAILWAY OPERATIOlSr plan is now about matured for removing the freight-tracks of the New York Central Railroad from the streets on the West Side by a line beneath Riverside Park and a viaduct through the blocks for the remainder of the route to the freight-terminal in St. John's Park.^ The subway system of New York City has a croas-seetion of 26 feet 3 inches wide by 13 feet above the rail. In the original form of construc- tion a metal-lining of steel stanchions was placed every five feet against a concrete wall, filled in with an invert of 20 inches of concrete. The roof of concrete jack-arches between the cross-girders was 31^ inches between the tunnel and the surface, further supported by iron columns between the tracks. Subsequently, this plan was changed to one wholly of rein- forced concrete, in which the thickness of the roof was increased to 5 feet 3 inches between the middle row of columns. The Pennsylvania Jlailroad double-track subway, on the Hudson River side of the terminal station, was built with a full arch, concrete side-walls and brick vaulting. On the East River side, there are two double-track tunnels, each with a brick arch resting on side- walls. Special Features and Ventilation The underground work in New York City has been prosecuted through unstable soil and through solid rock, under every conceivable difficulty from subterranean waters, from obstruction by existing subways, sewers, water-mains and cable-conduits, and in the protection of the foundations of adjacent buildings; All approved methods and appliances have been skillfully utilized ; many of them have been greatly improved, and, with added experience, others have been developed which are remarkable for the ingenuity displayed in devising them. Tube-construction has been but rarely resorted to. On account of the weight and dimensions of -the excavating shield, it is moved and guided with difficulty, and it can not be accommodated to varying cross-sections. If the shield is at all near the surface, the pressure which it exerts may raise the pavement at times as much as two to three feet in height. The usual method has therefore been that of "cut and cover," opened at night and covered by day. In later construction, the streets are supported by timber and planking, permitting continuous work beneath. In Forty- second Street, where the rails are 36 feet below the surface, a trench was opened only to the width of one track, and the wings for the other three were cleared out by cross-drives ; the roadway above being supported by beams and girders as the work progressed. In Europe, wherever arched masonry could be used, it has been found less expensive than a metal roof, though the latter has the advantage of bringing the stations nearer the surface. Reinforced concrete has been more generally used for subway-construction in the United States. In 1 Further information on this subject is given in Appendix IV, Table VII. ROADWAY 211 underpinning the foundations of adjacent buildings, it is unadvisable to place reliance on iron girders as permanent supports. There have been instances where the web of a girder, originally an inch thick, had been foiind reduced by corrosion to the thickness of a sheet of writing-paper. The movement of materials in crowded streets is often such an obstruc- tion to the normal traffic as to warrant considerable expense in devising ways of transport ; as where the bank of a river or canal, or a railroad- track, might be reached within a reasonable distance by cross-drives. Where subway-lines are operated by electric traction, no provision has to be made for the removal of gases from coal-combustion. The Metropolitan Railway in London was notorious for bad ventilation before it was electrically operated. The proportion of noxious gases, amounting to from 60 to 89 parts per 1000, was then reduced to a maximum of 11 to 14 parts. Artificial ventilation has, however, been applied in the elec- trically operated "tubes." Since 1902, the Central London Railway, six miles in length, has been ventilated by an exhaust-fan of 5 feet face and 20 feet in diameter at one end of the line, driven by 300 horse-power. Several times during the night, the doors of the intermediate stations are closed and the line is swept from the other end by a current of fresh air, by which the proportion of noxious gases is reduced to 7 parts per 1000 ; the outside air at the same time containing 4.4 parts. In some of the later tubes, electrically-operated fans are placed at half-mile intervals; each exhausting 18,500 cubic feet of air per minute. In the Boston subway, the air is completely changed by exhaust-fans every fifteen minutes. The air in the Pennsylvania Railroad subways in New York is changed auto- matically by the passage of the trains, but in the subaqueous tunnels there is an emergency provision of exhaust-fans. In the Rapid Transit sub- ways, reliance is placed upon natural ventilation, though not altogether with satisfactory results. The ventilation is effected through openings in the walls of adjacent areas under the sidewalks. These openings are 60 feet in length by 5 feet in width, covered by gratings for foot-support and to prevent the entrance of rubbish into the subway. Pboposed Tunnel under the English Channel A discussion of railway tunnel-work may well conclude with some reference to the projected tunnel under the Channel between England and France. Such a tunnel for a carriage-way was proposed to Napoleon Bonaparte, in 1802, by a French engineer. Thom6 de Gamond, another engineer, prepared five separate plans for a railway tunnel, between 1834 and 1856. One of these was reconsidered by I. K. Brunei, Robert Stephen- son and Joseph Locke, and was the subject of an exhibit at the Paris Universal Exposition in 1867. In 1869, Thom6 de Gamond - obtained the support of an Anglo-French committee for the incorporation of a company 212 EFFICIENT RAILWAY OPERATION to obtain a concession for carrying his plan into effect. Then the diplo- matists intervened and it was not until 1874 that the British government authorized a joint commission to prepare an agreement under which the work might be undertaken. In 1876, the commission presented its report for ratification by the legislatures of the two countries. In 1875, the French government had already sanctioned the formation of a Channel Tunnel Company, granting it a concession for the railway connection in France, for a period of 99 years from the date at which the tunnel should be opened for traffic. In England, three companies, that had separately occupied themselves with, the matter, were merged in 1886 into the Submarine Railway Company, which began work by driving a heading for 1|- miles, of which a mile was under the sea, when public opinion was aroused against the project, through the press and by petitions to the government. Military and naval experts expressed alarm at the possi- biUty of invasion by way of such a tunnel, and the attempt to obtain parliamentary sanction was withdrawn. Similar attempts in 1887 and in 1906 were equally unsuccessful, but a favorable change in public opinion is expected upon the termination of the present European War. Recent improvement in the methods and appliances for subaqueous tunnel- construction have rendered the project more feasible than heretofore. From geological examinations, trial borings and soundings, valuable information has been obtained as to the formation of the bottom of the sea and as to the composition and stratification of the underlying material. In addition to the heading driven from the English shore, one has been driven from the French coast for an equal distance under the sea. From this experience, it has been ascertained that a continuous stratum of cretaceous rock underlies the Channel and extends under both shores, which is sufficiently compact and impermeable to admit of driving a tunnel through it without lining. The plan now recommended is for two parallel tunnels 50 feet apart, circular in cross-section and connected by cross-drives every hundred yards, with an independent drainage-tunnel. As the drainage-tunnel progresses from each shore, the main tunnel would be driven from it at from four to seven different places. The gradient would not exceed 20 feet per 1000, and at the lowest point the tunnel would be 328 feet below sea-level and 164 feet beneath the bed of the Channel. A system of ven- tilation has been devised for supplying fresh air to the work during con- struction and for exhausting vitiated air when in operation by electric traction. On the French side, the line would diverge from the Northern Rail- way between Boulogne and Calais to Wissant on the coast, and thence descend to the sea by a loop-line over a viaduct half a mile in length and 46 feet high, in order to meet military objections raised in England, and to permit of the destruction of the viaduct by a British fleet. On the Eng- ROADWAY 213 lish side, the entrance to the tunnel would be at the back of Shakespeare's Cliff, west of Dover, under direct fire from three forts. It is estimated that the work can be executed within a period of seven years and at a cost of $80,000,000. The whole length of the tunnel would be 33.6 miles and the total time for express-train service between London and Paris would be about five hours, or two hours less than under present normal conditions.' ' For additional information, see Proceedings of International Railway Congress, Berne, 1910. Vol. I, Part IV, p. 61. CHAPTER V {Continued) PART II. SUPERSTRUCTURE Development of Tra.ck. — The Edge-bail. Wheel Flanges. Bull-head Rail The substructure of a railway is in many respects much like that of other highways, except as required by a different mode of traction, in- volving heavier loads and higher speeds. It is only because of its super- structure that the railway becomes no longer a public highway, open to all comers. In its inception in England, the railway track was merely an improved surface of the ordinary road, enabling it to bear up better under heavy traffic from the colUeries. With this purpose in view, in 1633, broad beams were laid longitudinally at bad places in the roads. From this beginning was developed the "stringer track," of pieces of oak or fir, about six feet in length, four to six inches wide and four to five inches thick, pegged down to sleepers placed about two feet apart, so that each stringer was supported by three sleepers. This form of track was next improved by an arrangement known as the "double- way," in which a rail of beech was laid on top of the oaken stringer and took the wear. The space between the rails was filled with cinders or broken stone to protect the feet of the horses from the sleepers. Then at the curves, or on steep grades, there were laid strips of iron, known as "plates," two inches wide and half an inch thick. About 1738, this practice was extended until the "tramway" or "dramway" generally supplanted the double-way.^ About 1767, cast-iron plates were substituted and, in 1776, to strengthen these plates as well as to guide the wheels, they were cast with a flange from two to three inches high. Such plates were six feet in length, three inches broad, half an inch thick and from 47 to 50 pounds in weight. The under-raU was no longer used; the plates were spiked directly to the sleepers. This form of track was known as a "plate-way," and to this day, in England, track-hands are termed "plate-layers." Where these plate-ways crossed a turnpike, the raised flange was very objectionable. In 1788, William Jessop, who was building a plate-way, conceived the idea of transferring the flange to the wheels and turning the plates on edge as "edge-rails" ; and the "railway" was born. 1 The first section of the Great Western Railway, opened in 1838, was stringer- track with a heavy rail. 214 ROADWAY 215 These edge-rails were cast in three-foot lengths, fish-bellied and with top and bottom flanges. The lower flange, at one end, was spread out as a foot and spiked directly to the sleeper, and a socket in this end received the plain end of the adjacent rail. As the spread-end was found to break in service, Jessop, in 1797, invented a cast-iron "chair," in which the ends of the rails were inserted. In 1805, the edge-rail was rolled in wrought-iron in lengths of twelve feet and of a section l^inch square, supported on stone blocks at intervals of three feet ; but the cast-iron rail was not superseded until 1820, when Birkenshaw had patented an edge-rail with a "T" head. This rail was rolled in lengths of 15 feet, of a section If inches wide at top, weighing 25 pounds per yard ; with a channel or groove on one side of the lower part of the stem, by which it was secured to the chairs with wooden keys.' Later, these rails were rolled in a flsh-bellied pattern with a head 2j inches wide, fastened to the chairs with pins or bolts. Such rails were laid on the Stockton & Darlington Railway in 1822-1825. In 1829-1833, George Stephenson used a pattern weighing 35 pounds per yard, 2^ inches deep at the chairs and flsh-bellied between them to 3^ inches. A projection on one side of the end of the stem fitted into an opening in the chair, and an iron key on the opposite side held the rail in position. Difficulties in rolling this rail led to the abandonment of the fish-bellied pattern, and attention was directed to the "parallel rail" rolled in 1834, with a knob at the foot of the stem and weighing 42 pounds per yard. From this section was developed the "bull-head" pattern, rolled in lengths of 15 feet and weighing 68 pounds per yard, secured in the chairs by wooden keys. In the meantime, in 1835, the "double-head" pattern was introduced to be turned bottom upward after the upper surface had become worn. A section of this pattern was rolled in 1838, weighing 78 pounds per yard ; the additional weight being intended to reduce the cost of chairs and sleepers by placing them five feet apart. After an experience of some thirty years, it was found that the bottom surface was so much worn in the chairs that the rail could not be turned to advantage, and the bull-head rail became the standard pattern in Great Britain. In 1875, such rails were rolled in 24-foot lengths weighing 83 pounds per yard ; in 1893, in 30-foot lengths weighing 85 pounds ; in 1896, in 36-foot lengths, weighing 100 pounds, and subsequently in lengths of 45 feet.^ » The Liverpool & Manchester Railway was laid with edge-rails of forged iron, in lengths of 15 feet, weighing 35 pounds per yard and supported every three feet on stone blocks. 2 For further information as to the genesis of the railway track in Great Britain, see " History of Inland Transport and Communication in England," B. A. Pratt, London, 1912 ; and "Modem British Permanent Way," C. J. Allen, London, 1915. 216 EFFICIENT RAILWAY OPERATION Early American Structures. Strap-rails and T-rails Railroad transportation was introduced into the United States at a time when labor was as difficult to obtain as was capital. Early operation was all-important, and cheap construction of more immediate benefit than easy grades and curves. There were but few auxiliary ways for the delivery of materials from any considerable distance ; so that the best use had to be made of those which were at hand. Therefore, the road-bed was ballasted with sand, the cuts and fills were left unsodded and exposed to wastage by weather. The bridges were often built by house-carpenters with timber cut from adjacent forests and wrought into beams with the pitsaw. The "strap-rails" were about 2^ inches wide and f inch thick, laid on 6-inch by 6-inch stringers, to which they were fastened by spikes through their surface and about 12 inches apart. The recurring blows from flange action threw the stringers out of line. This irregularity was corrected horizontally by wedges or keys, driven beside the stringers in jogs in the mud-sills that supported them, and vertically by wooden shims. The track was of such slight construction as to be rightly described as "a hoop tacked to a lath." This description is not applicable to early railroad construction in the Eastern States, where more approved methods had been derived from English experience. The Quincy tramway of 1826 was laid on granite sleepers 8 feet apart, which were obtained from the quarry that it served. The pine stringers were 12 inches deep, covered with a strip of oak, to which the strap-rails were spiked. The Boston & Lowell Railroad, built in 1835, was a remarkable example of track-construction, with a continuous foundation of parallel dry-stone walls, supporting stone blocks into which oaken plugs were inserted, to which the rails were spiked. By 1845, this form of construction had been abandoned for a track laid on chestnut ties, 7 feet long, 6 inches deep and 31 inches apart, on a bed of gravel two feet in depth. The T-rails weighed 56 pounds per yard, in 18-foot lengths ; the joints being secured in "clasp " chairs of 20 pounds' weight. The Bos- ton & Providence Railroad track of 1837 was of a similar construction and, ' after eleven years of operation, it had been necessary to replace but 750 rails, or only 2^ per cent, of the whole. The average life of the white-cedar sleepers had been from seven to eight years. On the Western Railroad of Massachusetts, in 1841, 56^pound rails were laid on sleepers 7 Jeet long and 7 inches deep, 3 feet between centers. The sleepers rested 'on longi- tudinal sills, 8 inches wide and 3 inches thick, supported at the joints by other pieces 3 feet long. Four pieces of this length were also placed under the joint-sleepers.' I In the report of the Boston & Lowell Railroad Company, made to the Legis- lature m February, 1838, it is stated that, "The foundation for the first track of rails IS laid with dry stone walls in trenches from 2J to 4 feet deep and aboii* 18 inches thick. The rails are laid, part on stone blocks, and a small part on stone ROADWAY 217 Sleepers of unhewn timber were apparently in general use, as the stringer-track of the Mohawk & Hudson Railroad, built in 1831, is de- scribed as laid on sleepers 8 feet long by 7 inches in diameter and 3 feet between centers, with a strap-rail -^ inch by 2^ inches, "with the upper curve rounded to 1|- inches in width." The sleepers rested on stone blocks in a bed of broken stone. The tramway of the Delaware & Hudson Canal & Railroad Company in 1829 is described by Mr. Allen, its chief engineer, as "formed of rails of hemlock timber in sections 6 by 12 inches, supported by caps of timber, ten feet from center to center. On the surface of the rail of wood was spiked the railroad iron — a bar of rolled iron 2| inches wide and \ inch thick." It is not surprising that such a track should have been insufficient to support locomotives weighing seven tons on four wheels. The gradual extension of the Baltimore & Ohio Railroad represented the continuing development of track-construction, from longitudinal sills of granite, laid in trenches filled with broken stone, to stone blocks sup- porting wooden stringers, and to "the log-rail, formed of trunks of trees, worked to a surface on one side to receive the iron and supported by wooden sleepers." The original strap-rail was 2^ inches wide by f inch thick, in lengths of 15 feet, beveled on the ends and pierced with 11 oblong holes.^ blocks and sleepers, all of which are supported on the trench-waUs above described. The stone blocks are 3 to 4 cubic feet each. The stone sleepers are 7 feet long and average 8 to 10 inches square. The wooden sleepers of chestnut and white cedar are 7 feet long, 7 to 8 inches in diameter. This track is mostly laid with rails of the fish-beUy pattern, and is set on chairs which are fastened to the blocks and sleepers above described." F. J. Wood, "Discussion on Railway Development." Trans. Am. Soo. C. E., December, 1911. ' In a report dated December 1, 1829, Colonel Long, chief engineer of the Baltimore & Ohio Railroad Company, describes the track-construction as follows : "The tops of fills and bases of cuts were alike made 26 feet wide, with side-slopes of li : 1 for flUs and 1 : 1 for cuts. Of the 26 feet, 20 were macadamized with broken stone of 2 and 2j inch size, laid to a depth of 4 inches. Trenches were dug through this pavement, 4 feet apart, to receive the ties and at each end a pit was dug, 18 inches long, 12 inches wide, and 12 inches deep, which was fiUed with rubble to form a foundation for the tie. The ties, of locust and cedar, were 8 feet long and 7 inches in smaller diameter, and were notched to receive the wooden rails, the outer edge of the notches being placed to a true spacing of 5 feet. The wooden rails were of '6-inch by 6-inoh Southern heart-pine, in lengths of from 15 to 40 feet, and were set in the notches with keys. The iron rails, on which the wheels were to run, were of wrought iron, f inch by 2i inches and 15 feet long, appropriately rounded on their upper sides, and perforated with elliptical holes about 15 inches asunder. At the joints they were scarfed on an angle of 60° with the sides, and laid on a plate rt inch thick. The nail-holes were counter-sunk, allowing the nail-head to be driven below the touch of the wheel, while the ellip- tical shape of the hole took care of expansion and contraction. For a width of 9 inches on the inner side of each wooden rail, coarse broken stone was laid, leav- ing a space of 2i feet in the center which was fiUed with finer broken stone to form a path for the horses. An alternative construction of stone rails was contemplated, in which case the stone rails were to rest on continuous rubble walls built in trenches to a depth below frost. Wooden railS were adopted for first construction, because it was believed that the fills would not sufficiently compact to receive such per- manent work as stone, for 4 or 5 years." F. J. Wood, "Discussion on Railway Development." Trans. Am. Soc. C. E., December, 1911. 218 EFFICIENT RAILWAY OPERATION Stringer-track continued in general use in New York until 1847, and in the Southern States for ten years afterward. Under the rolling effect ■ of passing wheels, the strap-rail tended to break at the spike-holes and to buckle up in "snake-heads" at the ends, where it was mitred to lessen the wear. This led to the substitution of the "chub-rail," which was wider and thicker than the strap-rail and was rolled with a low flange on the inner edge, through which it was spiked to the stringer, instead of through the upper surface. Rails of this pattern were in use in Georgia as late as 1870, weighing from 30 to 40 pounds per yard. There was also a pattern with a double flange of an inverted U-section, spiked to the stringer alternately through the inner and the outer flange. Rails of this pattern, rolled at Mt. Savage, Maryland, in 1844, weighed 42 pounds per yard. The buU-head rail never found favor in the United States. In fact, it was anticipated by the inverted "T" pattern, usually known as the "T-rail" or "flat-base" pattern; as the upright T-rail had been displaced in England by the bull-head pattern. The inverted T-rail or " pear-head " rail was an American idea, having been invented, in 1830, by Robert L. Stevens, who had the first lot rolled in England. In May, 1831, this lot of five hundred rails, in lengths of 15 feet and weighing 31 pounds per yard, was probably laid on the Camden & Amboy Railroad. The next lot was of a 40-pound section, 3|- inches high, 2^ inches wide on top and 3|- inches on the base, in 16-foot lengths. In 1844, rails of this pattern were rolled in the United States at Dan- ville, Pa. In 1846, T-rails were substituted for strap-rails between Balti- more and Washington. By 1850, 9021 miles of track had been laid with this pattern, 'and 30,628 miles by 1860. The T-rail was re-invented or rather introduced into England, in 1836, by C. B. Vignolles, who was perhaps the Vignolles who practiced surveying at an early date in South Carolina and published an excellent map of that State. Although the T-rail has been generally adopted elsewhere, it has not replaced the bull- head rail-in Great Britain. In Ireland, however, the "Vignolles" or "flat- bottomed" section, weighing 95 pounds per yard, is the standard pattern on two of the principal lines. Up to 1860, the rails in common use in the United States weighed 50 pounds per yard. By 1870, the usual weight was 60 pounds. With the introduction of steel rails and the requirements of heavier traffic, there was a further increase to 65 and 70 pounds. In 1883, Dr. P. H. Dudley designed an 80-pound section, 5 inches high, which was 66 per cent, stiffer than the 65-pound section, 4-J inches high, with only 23 per cent, more metal. In 1892, this rail was replaced on the New York Central lines by another section of his design weighing 100 pounds per yard and 6 inches high. In 1915, the Pennsylvania Railroad Company adopted a standard section weighing 125 pounds and 6^ inches high. On the Central Railroad of New Jersey, rails of the same height, weighing 135 pounds, have for some ROADWAY 219 time been in use. Recently, the Lehigh Valley Railroad Company has laid rails, on mountain-divisions with heavy grades and sharp curves, weighing 136 pounds and 7 inches in height.' The length of rails was increased in the United States, about 1859, from 16 to 30 feet. The standard length as prescribed by practical con- siderations is 33 feet. Longer rails are more Hable to defects in rolling and in straightening them. It is inconvenient to transport rails over 33 feet in length on a flat car of standard length ; and a larger force of men is required in loading, unloading and distributing them. Notwithstanding these disadvantages, there is an increasing tendency to the use of 45-foot lengths, in order to diminish the number of rail-joints, and rails are even rolled in 60-foot lengths. In laying these longer rails, there should be a somewhat greater allowance for contraction and expansion, for which, 33-foot rails usually require a space between rail-ends of f inch in frosty weather and ^ inch or less on a hot summer day. As long as the stringers were merely intended to preserve the road from being cut into ruts by cart-wheels, they were laid flush with its surface. But as the cast-iron flanged plates and the flat tire were replaced by the edge-rail and the flanged tire, the track was raised upon the surface of the road. It was then held in Une by loading or "ballasting" it with the broken stone that was in common use for building macadamized roads and which still perinitted . of horse-traction. With the heavier loads accompanying steam-traction, the track was raised upon the ballast for a better foundation. Rail and Joint Fastenings A suitable connection of the rail-ends and a sufficient support of the track at these weakest points, have presented the most serious problems in track-construction. In Great Britain, the conformation of the bull- head rail required an additional base for its attachment to the sleeper. This was provided in the heavy chair and wooden key which are essential to the use of this pattern of rail. In the United States, the flat-based pattern in general use at first rested at the joints only on plates through which the rails were spiked to the ties. These narrow plates were supei*- seded by a square plate of wider dimensions, with the middle part of the edges turned inward into lips. Subsequently the plate was transformed into a chair with a continuous lip rolled on its surface. A track laid in this manner soon begins to work loose at the joints. R. L. Stevens sought to remedy this defect, as early as 1830, by the invention of a splice, or fish- plate, bolted through the rails at the joint.^ The hght rail-section then in use was not well suited for an iron splice to fit in its web with a depth ' For Standard Rail Sections, see Appendix V, Table I. 2 The invention of the hook-head spike is also attributed to Stevens, as well as the rail of inverted "T" section. \ 220 EFFICIENT RAILWAY OPERATION sufficient to give the necessary strength at the joint, and the value of this device was not at first appreciated. With the joint supported on a tie or sleeper, the rail-ends are battered by the passing wheels. To minimize this effect, resort was had in Great Britain, in 1847, to the "suspended" joint, in which the rail-ends projected beyond the chairs and were connected by Stevens' fish-plates filling the space between the chairs. Fish-plates thereafter began to supplant chairs in the United States, though not entirely until about 1870, when the 60-pound rail came into general use. A fastening was at one time in use, known as the "Trimble splice," the joint being suspended and spliced by a wooden bar outside, fitted to the web and flush with the rail head, resting on two ties and bolted through the rails to a fish-plate. Several varia- tions of this splice have since been devised in metal. Intermediate tie- plates made their appearance at a later date. Experience with battered rail-ends has led to the abandonment of the supported joint. But in this country, as in others where the flat-base rail is used, a different arrangement of the suspended joint has been adopted, known as the "bridged joint," in which the fish-plates extend over the two adjacent ties. The fish-plates are of an angular cross-section, or "angle-bars," with the base of the bar extending over the base of the rail to a bearing on the ties. Some patterns are further stiffened by extending the base downward in a vertical flange between the ties. Others are returned under the rail and bolted together, to carry a base-plate covering the ties, as an additional support to the rail- ends. With a supported joint, the fish-plates are usually about 20 inches in length, or as much as 48 inches with a bridged joint. They are fastened through the rail with either four or six bolts and with washers which may be of helical form and of tempered steel, to prevent the nuts from working loose. The holes in the fish-plates are somewhat oblong to allow for expansion in the rails. On European roads, it is customary to prepare a seat for the flat-base rail with a slight cant inward. In Great Britain, where the bull-head rail ■yests in a chair on each sleeper, 'the cant is given in casting the chair. In rthe United States, rails are generally fastened directly to the intermediate tjies by hook-head spikes, though tie-plates have come into use with heavy rjiils, and especially with ties of soft wood. It was formerly quite generally tihe practice to use on curves a heavy cast-iron rail-brace spiked close against the outer rail on each tie. The advantage of tie-plates on all the ties was /not at first appreciated. By thus doubling the metal bearing-surface, the cutting of the wood-fibers by the rail-flange is prevented, and the life of the tie, in connection with preservative treatment, greatly lengthened. At the same time, tilting of the rail, with consequent widening of gauge, is obviated, and it is found that, by the use of well-designed shoulder tie- plates on curves, not only is the super-elevation better maintained, but it is also possible, in most cases, to dispense with rail-braces. ROADWAY 221 Ballast The ballasting of a track differs from the metaling of a macadamized road, inasmuch as the latter includes the surface of the road while the former is intended to bind the surface to the subgrade. The surface-water, which flows over the surface of a macadamized highway, drains through a ballasted railway, and, though similar materials are used, they are dif- ferently disposed. In England, this matter is as carefully looked after as the surfacing of a highway. The subgrade for double-track varies in width from 30 to 33 feet in cuts, and 28 to 30 feet on embankments, with a crowned siu-face falHng away six inches to the edges. The drain- pipes in the cuts are laid with open joints a foot below the subgrade, and the trench is filled with broken stone. The drains are connected with bricked-up pits every hundred feet. The stone should be so hard as to preserve its angular fracture. When crushed, the maximimi size of the fragment is fixed by a ring of 1^ to 2 inches in diameter and the minimum at •!■ to f inch. The bottom-ballast should be in cubes of 3 to 3^ inches, laid by hand, 9 inches deep on embank- ments, and, in cuts, 6 inches at the center line and 12 inches at the edges, to preserve a level surface. It is from 24 to 26 feet in width on embankments and extends over the drainage in cuts. The smaller-sized top-ballast is spread in a layer of about 12 inches, in which the sleepers are bedded to their upper faces. The earlier practice of filHng in the ballast to the under- side of the rail, carried on heavy cast-iron chairs, has been abandoned. The top-ballast is spread well out beyond the ends of the sleepers to pre- serve the aUgnment against horizontal thrusts, and is trimmed to the level of the bottom-layer. Furnace-slag compares favorably with broken stone as ballast. It costs nothing except for crushing and carriage, while the handling and packing of broken stone adds materially to the expense of using it. Cinder is used in the factory-districts. It more readily absorbs the surface- water than broken stone, especially on a clay subgrade, but is objection- able on account of dust. River-ballast, being composed principally of water-worn pebbles, does not make as steady a foundation as broken stone. Gravel is screened in meshes, 12 to the inch. On account of the dust, the use of gravel and sand is commonly restricted to sidings and yards. A ballasted roadbed, as understood in England, was rarely to be found in the United States until long after the Civil War. The track was usually laid directly upon the subgrade. If that happened to be clay, sand was distributed along the track from the nearest cut and packed under the "low joints" caused by insufficient drainage. A gravel-bed was a treasure mine and broken stone was an unheard-of luxury. By degrees, the fact became obvious that a well-ballasted roadbed is essential to good track, 222 EFFICIENT RAILWAY OPERATION and that nothing less than 12 inches of suitable material beneath the ties will secure a proper foundation.^ Steel Rails. The Bessemer Process As above described, the railroad track has long since attained its present development, except as to the dimensions and details of its several parts and the character of the materials of which they are composed. The most important change in this latter respect has been in the substitution of steel for iron rails. The rail-sections had been enlarged to meet the require- ments of increasing traffic, but the heavier rails proved to be short-lived and caused serious dissatisfaction as to their rehability and the cost of fre- quent replacement. In 1863, the Pennsylvania Railroad Company made a trial of rails of crucible steel. The cost of this material, and the difficulty of producing a uniform product in masses of sufficient magnitude, rendered its general use impracticable. At this critical period, Bessemer's inven- tion for converting pig-iron directly into steel was applied to rail-manu- facture. The first rails from Bessemer steel were laid in England in 1857. Rails of this material had been rolled in Chicago in 1865, but the first rails on a commercial order were rolled in Johnstown, Pa., in 1867.^ By 1880, renewals with iron rails had virtually ceased.' Viewed as a metallurgical product, the rail, in its early stages, was simply a ball of puddled iron. Its mass was limited to the weight of the ball which could be manipulated by the puddler, and, therefore, to the pro- ■ ' American Standard Track Rails — 85 to 100 pounds per yard. Ties — 6 by 8 inches by 8 feet, or 7 by 9 inches by Si feet ; spaced 18 per 30-foot rail. Ballast — 12 inches under ties. Ballast required per 1000 feet of Single Track : 6 inches deep . .... 300 cu. yds. 12 inches deep 523 cu. yds. 24 inches deep 1005 cu. yds. Subgrade — 18 to 20 feet wide for single track and 13 feet between centers for two or more tracks. — M. L. Byers, ' ' Proceedings International Railway Congress," Berne, 1910. 2 "Track," A. B. Corthell. Journal Am. Soe. Mechanical Engineers, July, 1914. 'Mileage op Steel and Iron Track (Poor's Manual). Yeab Miles, Steel Miles, Iron Per Cent. Steel 1880 33,680 81,967 29.1 1885 98,102 62,495 61.0 1890 167,606 40,697 80.4 1895 206,546 28,662 87.8 1900 238,464 19,389 92.5 1903 271,013 15,247 94.7 ROADWAY 223 portionate weight per yard to the length of the rail. This difficulty was measurably overcome by the invention of the mechanical puddler, but this permissible increase in length and weight of section was next restricted by difficulties in rolhng heavy sections, which were obviated by increasing the power and strength of the rolling-trains. It was at this point that the manufacture of rails was transformed by the invention of the Bessemer converter in 1856. The magnitude of this change and its effect upon rail- way transportation can be adequately appreciated only by those whose experience covers that period of transformation. Rail-making then ceased to be empiric. It had become a scientific process and, with the change, the iron-master had become a metallurgist. The fundamental principle of Henry Bessemer's invention was the re- duction of the carbon in cast-iron to the point at which the treated metal would acquire the malleable property of wrought-iron, without losing its characteristic plastic property. In other words, it was the production of steel directly from cast-iron by the diminution of its carbon percentage, instead of producing it indirectly by the addition of carbon to wrought- iron in kiln-made steel or "blistered steel." As it lacked the property of the molecular change of tempering in cooling, "pecuUar to crucible steel, it was distinguished as "converted steel." Bessemer undertook to reduce the carbon content of cast-iron to the requisite proportion for steel by depositing the molten metal in a receiver and there "converting" a part of its carbon into carbonic-acid gas by oxi- dation with air blown violently through its mass. It was found imprac- ticable, however, to reduce the varying percentage of carbon in cast-iron to just the right proportion to constitute steel by this means, as a commer- cial product. The whole process was accordingly threatened with dis- aster until Bessemer conceived the idea of completely eliminating the carbon from the melted metal, and then restoring a fixed percentage. This result was attained by the addition of a specific alloy of iron with manganese, known as ferro-manganese or "spiegel-eisen." The reactions thus produced pertain more especially to the chemistry of metallurgy. With this change, the process became a practical triumph. Iron-making was revolutionized in all its phases, as also the manufacture of iron prod- ucts and the design and construction of engineering works. Steel Rails. The Open-Hearth Process Wrought-iron produced from sulphurous ores is known as "red-short," because it is brittle at a red heat. The sulphur contained in iron smelted from such ores is eliminated as a gas in the converter. The phosphoric acid present in iron smelted from phosphatic ores causes "cold-short"; that is, wrought-iron from such ores is brittle when cold. As the Besse- mer process does not affect phosphoric acid, phosphatic ores must be mixed 224 EFFICIENT RAILWAY OPERATION in the blast furnace with a sufficient proportion of non-phosphatic ores to render negligible the quantity of phosphoric acid' present in cast-iron from such ores intended for conversion. As most of the iron produced in this country is from phosphatic ores, much of the non-phosphatic ores for ad- mixture is from foreign sources with consequently increased cost of the converted product. The Bessemer process has, therefore, been largely superseded by the Thomas-Gilchrist process, in which iron ore or steel scrap, as well as pig-iron, may be used. The melted metal is converted in open receptacles, known as the Siemens-Martin open-hearth, which revolve slowly and horizontally on a slightly inclined axis. Oxidation is effected by the action of gas-flame upon the continually changing surface of the metal, due to this axial rotation. Consequently, the gradual reduction of carbon is secured, which was found impracticable in the Bessemer process. The progressive reduction can be ascertained by successive tests and ar- rested at any stage of the process. Steel with any desired carbon per- centage can therefore be produced. Alloys of tungsten or of other metallic elements can be introduced during the conversion, and high-grade steel produced, which is substituted for crucible steel for many purposes. This process is well suited for the conversion of cold-short iron, as the phos- phoric acid is eliminated by combination with the lime in the converter linings into phosphate of lime. The process is, therefore, known as the "basic" process, to distinguish it from the Bessemer or "acid" process, and its product is termed "open-hearth" steel.' The ingot of converted steel has now taken the place of the ball of puddled iron in the mechanical processes of rail-making. It is at this point that defects in the molecular disposition of the metal originate, which per- sist throughout the subsequent process of manufacture to the serious detri- ment of the finished product. The defects due to occluded gas bubbles rising to the top of the ingot in pouring and cooHng are measurably excluded if the upper portion of the ingot be scrapped. But others are still concealed within the remaining mass, and only disclosed after the rails are in the track. The percentage of such defects is greater with the steel ingot than formerly with the puddled ball of iron, because the metal in the latter was more thoroughly incorporated into a homogeneous mass and at a lower temperature. The increased length and section of the rail has compelled the sub- stitution of mechanical apphances for manual labor, which has made it ' possible to so quicken the operation of a roHing-mill train that rails may be projected from it almost as if squirted from a syringe. In furtherance of such productive activity, the ingots are reheated to a temperature of ques- tionable desirability for the future endurance of the rails. Differences of opinion on this point have induced extensive investigations into such heat- ' For relative output of Bessemer and of open-hearth rails, see Appendix V. Table VII. ROADWAY 225 treatment. The subsequent operations of straightening the rail and of punching it for the fastenings have to be performed with, such care, as well as its subsequent handling, as to give an impression that the steel rail of to-day is in a far more unstable state of equihbrium, as to its internal structure, than was the iron rail of former days.^ Rail-Failuebs. Standard Types and Requirements Rail-failures may be considered, for the most part, as divided into crushed or spUt heads, clear breaks and broken flanges. There are other less ■numerous defects, as cracks through the web and internal fissures in the head of the rail. The American Railway Engineering Association pub- lished statistics for the year ending October 31, 1911, from which it ap- peared that, for every 10,000 tons of new rail laid, the failures averaged 31.0 tons for Bessemer steel, 20.7 tons for open-hearth steel and, for all rails, 29.0 tons. There was an average of one failure for every 891 Besse- mer rails, one for every 1234 rails of open-hearth steel and one for every 941 rails of both kinds. In 1915, out of 634,898 tons of rails, 13,295 were of Bessemer steel. In that year, the comparative failures were reported as 142 of Bessemer to 100 of open-hearth steel.* Rail-failures from internal defects have drawn the attention of metal- lurgists to the importance of the heat-treatment from the time of reheat- ing the ingot until the last passage of the rail through the finishing-rolls ; also to the prevention of surface-defects becoming incorporated in the body of the rail. And further, it is desirable that there should be a recognized classification of the factor of safety in the relation of wheel-loads to the respective dimensions of rails, considered as girders. The rail, which in its inception was but a metallic surface applied to a wooden stringer, intended only to lessen the rolling friction of passing vehicles, has supplanted the stringer as the weight-bearing element of the track, and fulfills the function of a girder. For one of these purposes it must be hard, for another it must be strong. The hardened surface re- quired of the rail-head must at some point in the composition of the rail merge into the tensile toughness required of the web. The wheel-loads have been continuously increased until they approximately exceed the molecular cohesion of the rail-surface, while the service which the rail itself is thus called on to perform as a beam, induces internal stresses that are developed in lines of cleavage between the texture of the rail-head and the web, and that too often result in broken rails. Then, too, the impact of the swiftly revolving wheel-flanges against the sides of the rail-head brings a fearful horizontal shock upon it. The resulting tendency to turn the rail over is proportionally increased in leverage with the added height ' For Specifications of American Railwa-y Association for Carbon Steel Rails, see Appendix V, Table IV. => For Statistics of rail failures, see Appendix V, Table VIII. 226 EFFICIENT RAILWAY OPERATION of the rail. It may be said that no such combination of services is required of any other metaUic appliance and that, taking all of its functions into consideration, the relative sufficiency of the rail for these purposes controls the standards of construction in every other department of railway opera- tion. The chemical composition of the metal, its heat-treatment in the ingot and its physical treatment in the manufacture, all have a bearing upon the efficiency of the rail when placed in the track ; yet as to neither of these matters is there a satisfactory agreement among metallurgical experts.^ Nor is there an agreement between railroad managers as to the outlines or dimensions of rails of equal weight per yard as to the height, or to the width of base, or of the relative distribution of material respectively in the head, the web and the base of the rail. At one time, the mills had no less than 119 different patterns of 37 different weights per yard. The cost of production might be sensibly decreased by the general adoption of fewer types of standard sections. Efforts have been made to this end by engineering associations. The American Society of Civil Engineers appointed a committee to report upon standard sections from 40 to 100 pounds per yard, varying by incre- ments of five pounds per yard. Its report was adopted in 1910 and it has been estimated that, up to 1914, 75 per cent, of the product of American mills was in sections of that type. In 1912, the American Railway As- sociation provisionally approved two types of sections varying in weight from 60 to 100 pounds per yard, for which it was claimed that they could be finished at a lower temperature than was practicable with the types recommended by the Arfierican Society of Civil Engineers. The American Railway Engineering Association has since proposed a different type of sections varying from 89.96 pounds to 138.52 pounds per yard. A com- parison of these types as to dimensions and distribution of metal is given in Appendix V, Tables I and II. ' Specifications as to Constituents op Converted Steel for Rails Cahbon Content Feb Cent. pHOBPHOans Maximum Per Cent. British Engineering Standards Committee, 1907 0.35-0.50 0.55-0.65 0.65-0.75 0.40-0.55 0.53-0.75 0.45-0.55 0.75-0.85 0.37-0.56 075 Am. Soo. C. B., 1907. Bessemer steel . . Am. Soc. C. E., 1907. Basic steel . . . Am. Soc. C. E., 1914. Bessemer steel . . Am. Soc. C. E., 1914. Basic steel . . . Am. Soc. for Testing Materials, 1907 . . Am. Ry. Engineering andMaint. of Way Asso. Am. Railway Association, 1908 .... 0.085 0.050 0.10 ^ •0.04 0.10 0.03 0.10 Proportion of carbon to vary with weight per yard. ROADWAY 227 The difficulty in arriving at a common standard for rail-sections arises principally from a difference in the conditions to which the track is sub- mitted. On a line with heavy traffic and considerable short curvature, a rail is required with more metal in the head than is required under less strenuous conditions, and, if the head is to be thicker, the base of the rail must be thicker also, to insure a sound rail in its manufacture. It was to meet this variation in conditions that the American Railway Association recommended two standard types. The sections for different weights per yard differ so little in these respects that it would seem feasible to arrive at such a common understanding concerning them as would enable the mills to keep rails of standard sections in stock, and the fish-plates and other accessories as well. The general specifications recommended by the American Railway Association represent a possible compromise. Track-work in England. Ties or Sleepers The care taken in ballasting an EngHsh railway is also given to the other elements of which the track is constituted. The rails are supported on each sleeper in a cast-iron chair from 15 to 16 J inches in length and from 7 to 8 inches in breadth, weighing from 45 to 56 pounds ; in some cases, resting on a layer of felt. The inside-jaw bears well up under the head of the rail, and the rail-seat is so molded that, when the rail is keyed against the inner jaw, it is slightly tilted inward and will keep its position if the key should work out. The inside-face of the outer jaw may be corrugated, to give a better grip to the key. The chair is secured to the sleeper by two |-inch spikes and by two treenails of oak, tapering from 1^ to 1^ inches, alternating diagonally in the chair, or |-inch screw-bolts are substituted for the treenails ; all being driven in holes previously bored in the sleepers. The bolts screw into nuts with fangs or projections forced into the bottom of the sleeper, to prevent the nut from turning. The wooden key is from 6 to 7 inches long and from 2 to 2| inches thick, varying in depth from 2^ to 3^ inches according to the pattern of the chair, the sides of the key being parallel or slightly tapered. The keys and tree- nails are of oak or teak, usually compressed in dies. Some use is also made of a key of steel-plate, pressed in a double-arched fold with the edges turned on the back, so as to form a tapering slot. By driving a steel wedge into this slot, the key is expanded in the chair and against the rail, forming a powerful double spring. On sharp curves, where a "check- rail" or guard-rail is laid, chairs of a special pattern hold the two rails firmly in place. The rails are generally laid with suspended joints, opposite each other, and but rarely with "broken" or "staggered" joints. The joint-ties, which in some cases are of the exceptional width of 12 inches, are spaced as closely together as will admit of proper tamping. The fish-plates, fitting between the chairs, are from 18 to 20 inches in length, from 1 to 1| 228 EFFICIENT RAILWAY OPERATION inches in thickness, and weigh from 22 to 36 pounds per pair. They are punched for four bolts, f to || inch in diameter. The bolt-holes, which were formerly for square or for round-necked bolts, are now more uni- formly for bolts with necks of a pear-shaped or oval section. Allowance is made for expansion in the fish-plates as well as in the rails. The plates are rolled of a mild steel, toughness being its leading characteristic.^ Sleepers in England are usually of Baltic pine or fir, 9 feet long, 5 inches thick and 10 inches wide. They are sawed to a length, bored for the fastenings, machined for the rail-seat and then treated with preservative solutions. The spacing of sleepers at the joints varies between centers of chairs from 23 to 28 inches ; otherwise it varies with the length of the rail. Variations in these respects are noted in Appendix V, Table IX.^ Pot Sleepers. Steel and Concrete Ties An important change in track-material, known as the " pot sleeper," was introduced, in 1854, into Egypt, where timber was scarce and costly. This form of track-support is simply an enlarged cast-iron chair like an in- verted tray ; its shape being intended to keep it in position in the desert sand. The sleeper is formed of two of these chairs kept to gauge by a tie-bar. The flat-bottomed rail is keyed to the pots, which are tamped through large holes in their sides. Sleepers of this kind are used also in Hindustan. The pots are usually 25 inches long, 13 inches wide, 5^ inches deep under the rail and from ^ to f inch thick. The substitution of a metal sleeper or tie for wood really began when converted steel came into general use on the European continent. The steel is rolled in plates from 12 to 14 inches wide, by ^ inch thick at the edges and ^^ inch at the center line, pressed into the shape of an inverted trough from 3 to 3f inches in depth. A tilt is given to the rail-seat and I For Standard Specifications of Fish-plates in the United States, see Appen- dix V, Tables V and VI. ' Usual Dimensions OF Sleepers Country Length, Feet Width, Inches Thickness, Inches Great Britain . . France . . . Crermanv .... 9 8i 8i 8f 91-10 8-8i 10-12 81-91 9i 11 10 8-9 ' 5 5i-6 6i 5i 6 .6-7 Belgium .... India (5? ft. gauge) United States . . In the Southern part of the United States, while the track was of 6-foot gauge, the ties were nine feet in length and from ten to twelve inches in width, where pine timber was abundant. ROADWAY 229 the ends are curved downward to insure a firm bed in the ballast. The flat base of the rail is secured by keys to projections punched out of the plate. Metal sleepers were introduced in 1881 into Switzerland, where 70 per cent, are now of metal, 9 feet in length and weighing 160 pounds. 26,000,000 had been laid in Germany up to 1905. On the government 1-ailway system in Cape Colony 700,000 have been used. The most extensive experience with steel ties in the United States has been on the Bessemer & Lake Erie Railroad, where about 850,000 were in use up to 1913. The most recent type is rolled of an inverted "T" section, 5-^ inches high, 8 inches wide at the base, 4|- inches at the top, 8-^ feet in length and weighing 180 pounds ; it is secured to the rail by bolts through its upper flanges and through clips extending over the base of the rail. Each pair of 30-foot rails is supported by 20 ties. To prevent corrosion, they are dipped in hot tar. The insulation of the track-circuits gives rise to no difliculty.* The value of steel ties as a general substitute for wooden ties is as yet a matter of first cost, which in Europe is about double. This state- ment of itself indicates that an enormous increase of capital would be required to accompHsh such a substitution ; so that the life of a metal tie should be at least double that of a wooden one to justify the change, even disregarding the intermediate interest-account on the additional investment. The economic Ufe of the metal sleeper is as yet undeter- mined. The effects of corrosion are to be balanced against those of vege- table decay, as is the case with steel car-bodies. There is a greater comparative cost for handhng and storing, with less uniformity as to standard dimensions and a greater multiplicity of parts. In countries where suitable timber is scarce Or is exposed to the ravages of the ter- mite, the steel tie may have the preference; but in Great Britain, where the raw material is abundant and they are manufactured on an extensive scale for export, they have not superseded the imported wooden sleeper, although the latter is subjected to an expensive pre- servative treatment. Experience with beams of reinforced concrete has suggested the use of that material for railroad ties. Its component elements may be found in abundance where timber is scarce and may be molded in the immediate vicinity of the railway work, with a saving of transportation and handling. It offers the same resistance to the ravages of insects that sheet-steel does and is not subject to corrosion, but its susceptibihty to disintegration is yet to be ascertained. On the whole, it would appear that wooden ties must long continue in use and that economy must be sought rather in prolongation of their economic life than in their replacement by the substitution of other materials. 1 For further information as to details of traek-eonstruction in Great Britain, see "Modern Permanent Way," C. J. Allen, London, 1915. 230 EFFICIENT RAILWAY OPERATION Tie-timber and Pbesebvative Treatment The Census Bureau reported that, in 1907, the consumption of tie- • imber in this country amounted to 153,700,000 ties, of which 23,500,000 were f oi'' new track. About one-half of this number were of oak, about one-fourth of Southern yellow pine and the remainder was chiefly of chest- nut, cedar, Douglas fir, cypress, tamarack, Western pine and loblolly pine ; with a local use of lodge-pole pine, gum, beech and spruce. Assuming the average yield from an acre of forest as 240 ties, the total consumption of timber for this purpose, in 1907, deforested about 1000 square miles of forest lands. Attention has been paid by a few railroad managements to experimental tree-planting but, as it takes at least twenty years for trees of even the softer woods to attain the proper size for tie-timber, it is plain that the time is approaching when ties will have to be imported into the United States at a cost that will equal that of metallic ties. One-seventh of the cost of maintenance-of-way is for ties, and the expenditure for tie- renewals is about double that for the replacement of rails.^ To a considerable extent there has been a wastage from hard treatment, and it is only in recent years that such attention has been given to this matter as has long prevailed in countries where timber is less abundant and not so cheap. Under ordinary conditions, ties of oak or pine of stand- ard quality may, for the most part, remain in the track from five to eight years, varying with the effects of traffic. A greater economic life is claimed for certain species of oak, chestnut and cedar that are not in general use. In 1908, 82 per cent, of the ties in use were of hewn timber. There is a preference for those hewn singly from the cut, and it is only with the scarcity of timber that sawed ties have become acceptable. Recourse to preservative treatment has become necessary from the progressive deforestation in civilized countries. The purpose of preserva- tive treatment is to arrest decay ; which is itself a decomposition of cel- lular tissue by parasitic fungous growth. The germs from which this growth is developed may either have been dormant within the timber and only have acquired activity with the death of the tree, or they may have . been introduced superficially through the agency of moisture. Protec- tion from the latter cause had long been practiced by ship-builders and by other timber-workers. Such superficial treatment was greatly impaired by any subsequent exposure of unprotected surface by mechanical opera- tions upon the timber, nor did it affect the action of germs already present within its cellular structure. In the early railway era, the comparative scarcity of suitable timber 1 In ten years, 1865 to 1876, renewals of rails amounted to 28.9 per cent, of the total cost of maintenance of roadway and structures, and renewals of ties to 11.7 per cent. In the three years, 1907-1909, rail-renewals made up 5.1 ppr cent, of this account and tie-renewals 14.7 per cent. "Railway Development in the United States," W. D. Taylor. Trans. Am. Soc. C. E., December, 1911 ROADWAY 231 of native growth in Great Britain soon directed attention to the impor- tance of prolonging the usefulness of sleepers by superficial treatment for their preservation. Pitch had been the only protective material available for this purpose until the production of coal-tar as a by-product in the manufacture of illuminating gas. In 1838, Burnett patented the use of oil of coal-tar, known as "creosote oil," for the preservation of timber by steeping it in open tanks. He soon adopted the "Bethell" method of placing it in closed tanks, forming a partial vacuum by steam-condensa- tion to remove the sap, and then injecting the liquid under pressure. By this improvement, the oil not only served for superficial protection but also as a germicide for the destruction of internal vegetable fermentation. A "vulcanizing" process was also tried, which consisted in sterilizing the timber by submitting it, in retorts, to heat sufficient to char its surface, but this process- was found to be less effectual than creosoting, besides diminishing the strength of the timber. About 1837, the French treated timber by expelling the sap under hydrauUc pressure and using sulphate of copper as an antiseptic. About 1878, this process was abandoned in favor of oil which, for a time, the State Railway management mixed with chloride of zinc. Now, both in France and in Great Britain, all sleepers are creosoted. Prior to the treatment, they are machined for rail-seats and for spike-holes or bolt-holes, as the effect of the treatment is comparatively superficial, and the exposure of a fresh surface to parasitic growth is thus diminished. Screw-bolts damage the timber less than spikes do, and fang-bolts less than either. Old spike- holes are plugged up, and handhng the sleepers with picks is forbidden. The rail-chairs rest upon a layer of felt, to protect the sleepers from abrasion. On the British railways, from 28 to 30 pounds of oil are used per sleeper, with an average life of 16 to 20 years. The French inject from 30 to 40 pounds, and even 60 pounds, and claim from 25 to 30 years of life. In France, about 2,600,000 sleepers were used in 1900, while, for the years 1878-1885, the renewals averaged 3,000,000 per annum, with less track- mileage. A saving of 3,000,000 annually is attributed to improvements in the processes employed. In a single year, 14,126,000 cubic feet of tim- ber were treated for railway purposes. Mineral salts in solution are also injected by the Bethell process, and very generally in Germany, where railroad ties or sleepers were treated either with corrosive sublimate or with sulphate of copper. Only chloride of zinc is now used, and recently that has been abandoned on the State Railways for creosote oil. The methods employed are either impregna- tion with chloride of zinc and tar-oil, or creosoting after seasoning and drying in ovens, or after desiccation in hot tar-oil. Ties treated by the zinc-creosote process last from 12 to 18 years, and creosoted ties from 24 to 28 years. "Burnettizing" was introduced into the United States in 1850 at 232 EFFICIENT RAILWAY OPERATiON Lowell, Massachusetts, and was there worked by the "Bethell" process until about 1862 ; but apparently with indifferent success, as it was aban- doned for a time for the "Kyanizing" process with chloride of mercury. Now, however, it is much used with chloride of zinc and, to some extent, in combination with other substances, principally with creosote oil. These several combinations are respectively known as: the "Rutgers-" process, but Uttle used, in which the two hquids are injected in a mixture; the "AUerdyce" process, in which chloride of zinc is injected first and then the creosote oil; and the "Card" process, patented in 1906, in which the two liquids are mixed and maintained under pressure by a centrifugal pump. There is also the "Wellhouse" process, a combination of chloride of zinc, tannin and glue ; and the "Thelmany " process, in which sulphate of copper is used. Chloride of zinc has supplanted other mineral salts as an antiseptic because of its cheapness but, being soluble in water, it eventu- ally leaches out of the ties. It is for this reason that, in the modified pro- cesses, it is used in combination with creosote oil, which is insoluble in water. The same end is sought in a cheaper way by mixing chloride of zinc with tannin and glue. The relative economic value of chloride of zinc and of creosote oil de- pends upon the average annual cost of renewal during the life of ties so treated. In 1899, it was estimated that, in the United States, the average life of ties treated with chloride of zinc was about 10 or 12 years, and of ties treated with oil from 15 to 30 years, but at three or four times the cost. The wide range of life given by the oil-treatment, as here stated, makes this comparison of little value. The relative value of the modified process can only be established after ties treated by them have been in service at least as long as were those that have been treated with chloride of zinc. In any comparison of the value of oil as compared with mineral-salts, account should be taken of the character of the timber used for ties ; as some kinds take more oil than others. Red oak, pin oak, beech and tam- arack absorb less than 22 per cent, in volume ; sweet gum, chestnut, syca- more and poplar, 23 per cent, to 30 per cent. ; tupelo gum, short-leaf pine, cypress, birch and cottonwood, over 30 per cent. The success of the oil- treatment depends also upon the character of the oil. Real " creosote oil " is the heavy distillate of coal-tar, obtained at a temperature ranging from 480° to 760° F., and the proportion between its chemical constituents varies with the temperature at which it is distilled. It is assumed that it should contain at least 5 per cent, of tar-acids and 25 per cent, of naphthalene to secure its insolubility in water ; naphthalene being the efficient germicide. Proper seasoning of the ties is of importance in their preservative treat- ment. In Great Britain, they are thoroughly dry-seasoned, but in the United States they are usually only steam-seasoned. It would seem that dry-seasoning is in itself a preservative, by preventing checking and the admission of extraneous germs "with the moisture. In treating freshly ROADWAY 233 cut timber with oil, it would be well to season it with oil instead of steam ; but, if steam be used, then not for so long a period or at so high a tempera- ture as to diminish its strength.^ Upon the Chicago & Northwestern Railway, there are two very consider- able plants for treating ties. One of these has a capacity of 800,000 ties per annum and the other of 600,000 ; the timber being of pine, spruce or fir. In the more recent plant at Riverton, Wyoming, the retort is 6 feet in diameter with a track of 24|-inch gauge, admitting at one time a train of 16 cars, each containing from 30 to 32 ties. Live steam is first admitted and provision is made for discharging the condensed sap. A 20-inch vacuum is obtained in 30 minutes and held for 45 minutes. The vacuum pump is then shut down ; the solution of .05 chloride of zinc fiows in and is kept at a temperature of 150° F. by heating-radiators. When the re- tort is full, an additional amount of 2.5 solution, containing half a pound of dry chloride for each cubic foot of timber, is forced in from the pressure tank and held under 150 pounds' pressure for about four hours. The air in the pressure-tank is then used expansively for forcing the unabsorbed solution back into the working tank. With four charges in 24 hours, the average output is about 50,000 ties per month.^ The Atlantic Coast Line has an extensive plant at Gainesville, Fla., for the treatment of yellow-pine timber with creosote oil, which is considered preferable for the harder woods. Ten pounds of oil per cubic foot are usually required for pine ties and twelve pounds for softer woods. Treated ties should be laid sap-side up and a dating-naU driven in each tie as it is laid. According to a census report in 1911, in 401,653 miles of track, the annual renewal of ties amounted to 135,053,000, or an average life of some- what less than nine years. Of this number, 23 per cent, had been preserva- tiyely treated, and for the most part with chloride of zinc. Upon the Atchison, Topeka & Santa F6 lines, ties have been treated since 1885. Up to 1916, with 30,422,416 ties in the tracks of the parent system on a mile- age of 9552 miles, over 80 per cent, had been treated. As a result, the annual renewal is less than 200 ties per mile, indicating an average life of fifteen years. Spike-holes are bored and rail-seats are adzed. Each tie is stamped with the date and kind of timber, and the weight of rail for which it was bored.' Track Efficiency. Wheel Loads. Tie-spacing Although the substructure of a well-constructed road deserves the name of "permanent way," its track or superstructure may as appropriately be • "Preservation of Timber," F. A. Kummer. Trans. Am. Soc. C. E., Decem- ber, 1900. "Preservation of Timber in Europe," O. Chanute. Ibid., June, 1901. "Timber Preservation," W. Buehler. Ibid., March, 1911. 2 "New Tie-treating Plant on the Northwestern Railway," L. J. Putnam. Engineering News, April 20, 1916. ' Railway Age Gazette, September 22, 1916. 234 EFFICIENT RAILWAY OPERATION termed, in contradistinction, a "temporary way" ; for it undergoes inces- sant deterioration and restoration. Its distortion and disintegration vary with variations in the character and extent of the train-service. The ballast is ground into dust as the ties rock in their beds ; the ties themselves are shaken and split and their decay thereby hastened. The rails become bent and loosened in their fastenings, their wearing surfaces deformed and their ends battered down ; while fractures occur unexpectedly from inter- nal molecular disintegration. All the refinements proposed in the details of rail-joints and in the spacing of ties, apparently add no material advan- tages to a track of standard construction with bridged joints on a deep foundation of suitable ballast and well drained. At last, it is a question of maintenance, of incessant labor and vigilance on the part of an army of trackmen. What are the shocks and strains by which the efficiency of the super- structure is thus diminished? The recurring waves of deflection under passing trains rack the ballast, ties and joints. The rolling-friction of the wheel-treads wears away the heads of the rails, and the hammering effect of foot-pounds of energy beats down their ends. The swerving of trains from side to side throws the track out of line on tangents. At every curve, this pressure is intensified by the centrifugal force which exerts a tendency in the wheels to climb the outer rail, or to overthrow it, or to force it out- ward by shearing the spike-heads. Under such conditions, it is not sur- prising that the' superstructure of a railroad should be virtually recon- structed within a decade ; and the character and extent of each factor in this combination of destructive elements should be carefully weighed in counteracting their effect upon each component part of that superstructure. Primarily, the track should support the weight of the trains without being thereby deformed. This weight is distributed among the whe^s under a train ; and the depth of the ballast, the dimensions and spacing of the ties and the weight and form of the rail-section should be conditioned by the heaviest weight borne by any passing wheel ; that is, by the driving- wheels of the locomotive. In this respect, the railroad superstructure is submitted to severer tests in the United States than elsewhere. Here, the axle-load of 50,000 to 60,000 pounds is to be supported by a tie with a bear- ing surface of 5.5 to 6.4 square feet, which is a static load of 7800 to 10,"900 pounds per square foot. In Great Britain, the average axle-load of about 38,000 pounds is supported by a sleeper with a bearing surface of 7.5 to 9 square feet with a consequent static load of 4300 to 5000 pounds per square foot, or about one-half of the weight that a tie is expected to support in the United States; and on a ballasted road-bed far more carefully prepared than is customary in this country! There is a similar disparity as to the load concentrated upon the rail as a girder. In the United States, upon a rail-length of 32 .to 40 feet, there is a load ranging from 230,000 to 330,000 pounds, and in the case of an ROADWAY 235 articulated locomotive, in a length of 40 to 67 feet, of 468,000 to 616,000 pounds on the opposing rails. In Great Britain, for a length of 32 to 40 feet, the average load would be about 145,000 pounds, and in no case more than 190,000 pounds.^ As ties are usually spaced in the United States, 2640 to the mile or 24 inches between centers, a rail-length of 32 to 40 feet is supported by 16 to 20 ties with a bearing surface of 88 to 132 square feet, and, with the ordinary locomotive, would sustain a distributed load of 1700 to 3700 pounds per square foot. As sleepers are spaced in great Britain and with the locomotives there in use, the same rail-length would be borne by 12 to 16 sleepers, carrying a distributed load of about 1200 pounds per square foot and not exceeding 2000 pounds. The additional thickness of one inch in the American tie should, however, be considered in this comparison. Upon the Atchison, Topeka & Santa F6 Railroad system, the spacing of ties leaving a uniform distance between them is viewed with favor. Squared ties, 7 by 9 inches, are spaced slightly more than 11 inches between their edges, not centers, so that there are twenty ties under a 33-foot rail. Ties with an 8-inch face are spaced 13.8 inches. The spacing is regardless of joints, where no reinforcement extends below the base of the rail. Where the spacing is uniform with ties of 9-inch face, the joint-fastening will be supported either by two ties near its ends, or by one tie near its center, thus insuring either a suspended or a supported joint.^ Rail Stresses and Wear The track does not merely support a load at rest. It should also retain its stabiHty under a train in motion ; and here are encountered some of the most serious difficulties in railway operation. A wave of deflection, preceding each train, causes undulations in the rails which lift the track from the road-bed, to be rammed back by blows repeated from each wheel in the train. These undulations are not entirely transmitted from rail to rail. They are partially arrested at each rail-end ; part of their energy being dissipated in racking the rail-fastenings. Much ingenuity has been exhibited in designing fastenings that would either resist the disturbing effects of the wave of deflection or that would facilitate its passage from rail to rail. There has been an evolution from a simple bearing-plate to the lip-chair and to the fish-plate to meet these requirements. The fish-plate has been stiffened in design and enlarged in dimensions until it has become an appliance more appropriately called a j oint-bar . The same attention is given to its metallurgical composition and treatment as is given to the manufac- ture of the rails that it connects. It has been widened until it covers the entire surface of the web between the head and the rail, and has been ex- tended over the base to a support on the ties. Still, the wave of deflec- ' For Wheel Loads of American Locomotives, see Appendix V, Table X. 2 Railway Age Gazette, September 22, 1916. 236 EFFICIENT RAILWAY OPERATION tion racks the fastenings at their weakest point, the bolts ; as is shown by the difficulty in keeping the nuts from working loose. An army of men is daily employed in tightening them up, despite the use of jam-nuts and spring-lock washers. The attention given to the design and manufacture of the elements of construction and the thoroughness with which the work is done, have given the English standard track an enviable reputation. The same thorough- ness characterizes the drainage, the sodding of slopes and the other features of an English roadway. Railway practice on the continent of Europe has followed Enghsh methods in track-construction, as in other matters ; ex- cept as affected by the general preference for rails of the inverted "T" or flat-base pattern. In some respects, the bull-head rail presents an apparent superiority to the T-rail. Its joint-connections combine the advantages of the fish-plate and of the chair. In further combination with a heavy chair on each sleeper and the wooden key, there is a resilience in the track constructed after the British fashion which American track does not pos- sess in the same degree. And yet, even in Great Britain, with compara- tively lesser wheel-loads, the problem of transmitting the wave of deflec- tion from rail to rail unimpaired, has not been satisfactorily solved. In the solution of this problem, other issues are presented with refer- ence to the relative position of the joint in opposite rails and to the manner in which they are supported by the sleepers or ties. It has long been cus- tomary in the United States to lay the rails with broken or staggered joints ; probably because the track stood up better on inferior ballasting than when the opposite joints were on the same tie. In Great Britain, however, rails are usually laid with even joints, and rails of shorter lengths, or "make- up" rails, must be frequently introduced, in order to preserve the even joints. On either plan, low joints are common on track that is not care- fully maintained. Nor is it possible, even with the most approved forms of construction, to prevent the rails from acquiring a permanent set, consequent upon the incessant rolhng of heavily-loaded wheels upon them. Track-inspection by Dr. P. H. Dudley, with an autographic track-indicator, upon 10,000 miles of track on the main hnes of the New York Central and the Pennsyl- vania railroads, established the fact that rails in service from three to five years become permanently set. With even joints, they are low at the joints and high at the centers ; with broken joints, they are low at the joints and centers and high at the quarters. It is estimated that a steel rail of standard composition and manufac- ture, weighing from 60 to 80 pounds per yard, will withstand the passage of 300,000 to 500,000 trains, with a loss of ten to fifteen pounds of metal and a wear of f-inch to |-inch on the rail-head, though many rails exceed this minimum.! -q^^ long before the rails are in this condition, they are re- 1 ' ' Railway Location, ' ' WelHngton. Page 1 19. ROADWAY 237 moved to sidings and to industrial branch-lines, when the incessant ham- mering by passing trains has effected a permanent depression in their sur- face at the joints, which intensifies the general deterioration of the track. A test was made of the resistance of track to flange-pressure on a sec- tion laid with new, dressed-chestnut ties, and a rail of 100-pound section spiked to tie-plates on each tie. Under these conditions, it was found that a side-pressure of 10,000 pounds would overturn the rail. In fact, the rail does not overturn, because it is prevented from doing so by the weight of the locomotive. For the rail to be in equiHbrium, the side-thrust must equal one-half of the load on the driving-wheels plus the resistance of the spikes. When this limit has been reached, the rail will overtm^n. As the weight of the. locomotive increases, the percentage of thrust on the wheels necessary to overturn the rail decreases. Theoretically, the limit of safety will have been passed when a side-thrust of 20,000 pounds has been at- tained with a weight of 35,000 pounds on a driving-wheel, and this fact has been sustained by practical tests. Because the rib, or other projection on the under side of the tie- plate, sinks under the weight into the tie, the lateral motion of the rail is resisted before the spikes become loosened. On a test with a wheel-load of 35,000 pounds the rail turned over, and the tie-plate was forced into the tie, crushing and splitting it. In practice, however, there is not' a wheel over each tie, and with ties spaced 21 inches apart under a locomotive with driving-wheels six feet in diameter, the wheel is supported by three ties. The crushing effect would not, therefore, be as severe as in the experimental tests, and, with seven spikes to resist the side-thrust, the resistance to over- turning is 35,000 pounds instead of 10,000 pounds. The thrust in practice, however, is produced by quickly applied loads or shocks, while, in the ex- perimental tests, the thrust was effected by steady pressure. Therefore, when a rail is subjected to side-thrust from train-momentum with a 35,000- pound wheel-load, the factor of safety does not exceed two for the 100- pound rail with standard tie-plates ; whereas the theoretical factor of safety is placed at ten for iron or steel subjected to sudden shock from a live load.^ • For various wheel-loads, the side-thrusts are as follows : Wheel-loads, Lbs. SiDE-THRnsT, Lbs. Pebcentaqe 15,000 17,500 116.5 20,000 20,000 IQO.O 25,000 22,500 90.0 30,000 25,000 83.5 35,000 27,500 78.5 40,000 30,000 75.0 As the weight of the locomotive increases, the side-thrust necessary to over- turn the rail becomes a decreasing percentage of the wheel-load. "The Actual Service of the Track Spike and Tie-plate," W. D. Wood. Rail- way Age Gazette, February 20, 1914. See also Appendix V, Table X. 238 EFFICIENT RAILWAY OPERATION The underlying fact in track-deterioration is that the dynamic action of the train exceeds the static resistance in the track. It is a matter of momentum. The effects vary in intensity with the load and the speed. The force or momentum, accumulated in a train of twelve heavy pas- senger cars drawn by a ponderous locomotive at sixty miles an hour, has been estimated at 224,000,000 foot-pounds, but this applies to move- ment in a horizontal direction. It is surprising that any track, however well constructed, can fulfill the conflicting requirements for rigidity and for elasticity under such conditions. It may be remarked, however, that rigidity of track is a requirement for effective resistance to longitudinal or lateral movement or displacement, while elasticity is a requirement for a perfect reaction to the wave of deflection accompanying a moving train. The track should possess resiliency, acting with a cushioning effect for absorbing the shock of impact. The shght yieldiog under the moving weight of a fast train is a protective agency, offering the graduated resistance of a spring, not that of an anvil. As the track thus yields under the moving train, the foremost wheels are continually operating against the elevation caused by the preceding wave of deflection, which induces a gradual movement of the track in that direction. This movement is counteracted on a single-track line by a similar movement in the opposite direction. But there is a resulting balance in the direction of the heavier traffic which effects a positive move- ment in that direction. This effect is merged at curves in the line into a slight distortion of the curvature. On long tangents, including grades descending in the direction of the movement, it may result in "buckling" the track, and its extent is much increased in very hot weather by the effect of expansion at the closed joints. Although the track may be heavily ballasted, the joint-ties will move with the rails, and even in a double- spiked track, the rails will slip over the intermediate ties, leaving a shaving of metal behind each spike. This "creeping" of the rails becomes far more serious on a double-track line, with the traffic always in one direc- tion on each track, and is facilitated by the disconnection of joints at the switches. British track-construction yields more readily in this respect ; as the bull-head rail is not spiked directly to the sleepers but is held in position by wooden k«ys. In a stretch of 352 yards, there was an aggre- gate creep of 56 inches in the course of two years ; the rails having been pulled back, fourteen times during that period. The use of joint-fasten- ings as road-anchors, by slotting them for spikes, imposes an additional burden on the joint, which is the weakest point in the track-structure, while the elimination of slots and punch-holes adds to the strength of the fastenings. Joint-ties are not then slewed out of place, angle-bars are not stripped by this slotting nor is there an extra strain upon the joint- bolts. Separate road-anchors should be used in suflacient number to re- sist creeping, without relying on joint-fastenings. ROADWAY 239 Problems of Curvature The momentum of a train develops centrifugal force around curves, varying with the speed and weight of the train.' The consequent shock to a train upon entering a curve at a high rate of speed is lessened by the intervention of a transition-curve of greater radius at the points where a tangent is merged into the principal curve. This effect of curve- resistance is negUgible in a modern passenger-train at a speed of 18 miles an hour on a ten-degree curve, and at 13 miles an hour on a curve of twenty degrees. The centrifugal force induces a tendency for the wheel-flange to mount the outer rail, which is counterbalanced by the super-elevation of that rail. Definite formulas have been devised for determining this super-elevation, but the conditions vary so widely with the speed of the train, that it is rather an empiric than a mathematical problem. In practice, the maxi- mum super-elevation is eight inches. The effect of the train-momentmn is perceived in the distortion of the curve, which may be counteracted by laying a check-rail or guard-rail, firmly fixed at a proper distance from the gauge-side of the inner rail ; or by rail-braces against the outer rail. The effect of obhquity of traction at the couplers is now considered as unim- portant. The resistance in a curve to the passage of a train is independent of the momentiun or the length of a train, and is a reaction to the slipping of the loaded wheels. The supposed advantage of coning the wheel-tread soon disappears as the tread wears away. The greater length compara- tively of the outer rail compels a constant readjustment of the progress of the wheels on the same axle, that is only made possible by the occasional slipping of the outer wheel, as its flange strikes against the rail. The 1 Centrifugal force in pounds per ton of 2000 pounds on various curves at various speeds. A. M. Wellington. "Railway Location," Ed. 1914, p. 270. Degree of Curvature Speed, Miles per Hour 1 5 10 15 20 10 2.33 11.67 23.35 35.02 46.70 20 9.34 46.70 93.39 140.09 186.78 30 21.01 105.07 210.13 315.20 420.26 40 37.36 186.78 373.57 50 58.37 291.85 60 84.05 420.26 70 114.40 80 149.43 90 189.12 100 233.48 The computation ends at an assumed maximum limit of speed for safety; that is, when the centrifugal force equals one-fourth of the weight. 240 EFFICIENT RAILWAY OPERATION momentum of the train is transmitted to the wheels from the bodies of the several units in the train through the center-pins and side-bearings of the trucks, somewhat modified by the action of the springs. The rigid wheel-base of the truck is the principal factor in the effects produced by the slipping of the wheels. There is a normal play of | to f of an inch between the wheel-gauge and the track-gauge, which affords some ease by the lateral slipping of the forward truck wheels, permitting a rigid wheel-base of 12 feet on curves up to 3 degrees, and a wheel-base of 5 feet on a 17-degree curve. The longitudinal flipping occurs irregularly either with the forward or the rear pair of wheels, and its effect is ex- hibited in the wear of the rail-heads around a curve from the flange- friction. The curve resistance is directly as the degree of curvature, with corresponding rail-wear. The sharpest curves on the main tracks of the Trunk Lines are as follows : Pennsylvania Railroad, 8 degrees ; Baltimore & Ohio Railroad, 9^ degrees; Erie Railroad, 10 degrees. There is no curvature of as much as eight degrees on the main tracks of the New York Central Railroad. Switches, Frogs and Crossings The effect of train-momentum is especially severe where the continuity of the rails is broken at switches, frogs and crossings. At these points, the track should be maintained in first-class condition, to insure the passage of trains at high speed. With the original "stub-switch" this was im- practicable, as both switch-rails break connection completely in the running-track. Their movable ends are only secure while the switch-lever holds them against one lip of the chair in which they move laterally; and their other parts are kept in position solely by the tie-bars which hold them to gauge. Consequently, the stub-switch has been replaced by the "split-switch" which, if carefully maintained, is safe for the passage of trains at any speed. In the spUt-switch, the position of the switch is reversed, with its fixed ends toward the stock-rail leading to the frog. The running-track rail next to the siding leads without a break into the siding ; the continuity of the main line being preserved or broken by a movable rail on that side in the running-track. The opposite rail in the main line is also unbroken. The switch-rail on that side is connected by tie-bars with the movable end of the rail in the running-track and, as they are moved laterally, the en- trance into the siding is opened or closed. The movable end of each of these "split-rails" is planed to a thin vertical edge, which fits' into the web of the permanent rail for protection and leaves the flange-way free along one or the other of the permanent rails, accordingly as the switch is moved. By the interposition of a stiff, coiled spring in the pull-rod of the switch- lever, a train running from the fixed end of the switch may force its way either from the main line or from the Siding while the switch is set for the ROADWAY 241 other track, without the intervention of a switchman. A switch facing the direction of a train on the running track, is termed a "facing-switch" ; and in the opposite direction, a "traiUng-switch." Where two sidings diverge to opposite sides of a running-track at the same point, a "three-throw" switch is used. The three-throw switch is a double switch with the outer rail on each side permanently connected with the siding and the inner split-rails moving together in opening either sid- ing. These rails intersect at a crotch-frog in the center line of the running- track. As the three-throw switch, like the stub-switch, breaks both rails of the running track at the same point, it is rarely used in the main line. Where double-sidings are necessary, it is preferable to place one switch ahead of the other. Bumping-posts are necessary where a stub-track ends on a bank or trestle, or against a wall or fence. Where the end is clear and the ground is level, less damage is done by careless switching if the cars drop off the ends of the rails. The levers to ground-throw switches should throw paral- lel with the track and should lock automatically. All yard-switches should be numbered. Where two running-rails cross, provision must be made for the passage of wheels on either Une by a "flange-way." ' The angles at the intersec- tion are preserved by planing the two rails in one direction to an acute angle, firmly bolted together, and by bending the rails outward in the opposing angle as "wing-rails," for bearing the wheel over the gap at the intersection. This intersection is known as a "frog," from a fancied re- semblance to the frog in a horse's hoof. The wheels are prevented from taking a wrong direction at this point by guard-rails fixed against the opposite continuous rail. Many plans have been devised for diminishing the exceptional wear at the gap thus formed in the track. In some of them, the wing-rails are held closely to the point of the frog by springs which are forced open by the passage of a wheel on one side or the other. In other cases, a riser is placed in the throat of the frog which relieves the wear of the point by taking the weight from the edge of the wheel-flange. The objection to having so many parts connected by bolts and chairs, at a point where solidity of structure iS all important, is obviated by consolidating them in a steel casting specially smted to the angle of intersection. The Wharton switch, an American invention, is operated without a break in the main track, either for switch or frog. When out of position, it is entirely clear of the Hne but when a train is to take the siding, it is moved 'up laterally so that the switch rails fit closely to the running-rails and gradually lift the wheels ; one wheel rising on the tread until it clears the running rail at the point of contact and the other rising on the edge of 1 In early track-construction, the flange-way at this intersection was provided by a movable piece of rail, pivoted in the track like a switch-rail, which could be thrown over to the siding-rail by a separate lever. 242 EFFICIENT RAILWAY OPERATION the flange until it is carried above the same rail at the point of intersection. This switch is well suited for entering sidings that are but little used, and at low speeds. "Cross-overs" between running-tracks are properly laid with a trailing- switch in each track, and should have longer leads and frogs with smaller angles than for siding-connections. The length of the leads from switch to switch varies with the spacing of the running-tracks and the angle of the frogs. The minimum length is 122 feet 6 inches for a No. 6 frog (9° 31' 38") and 11 feet between track-centers. The maximum length is 381 feet 3 inches for a No. 15 frog (3° 49' 06") and 16 feet between centers. The crossing of one track by another is effected by a combination of four frogs in opposite pairs, two at an acute angle and two at an obtuse angle, or all at 90° if the crossing is rectangular. These crossings become complicated where double-tracks cross, and still more when they include connections between main lines by "slip-switches." ^ Such crossings re- quire specially constructed combinations of rails and castings. A crossing at fight-angles is objectionable because the width required for the flange-way in both lines of rails leaves little bearing for the wheel tread on the wing-rail and throws additional stress on the frog-points, which should, therefore, be made of manganese-steel and be removable for replace- ment. The standard width of flange-way as established by the American Railway Association is If inches between the main rail and the guard- rail and through the throat of the frog, measured at the gauge-Hne. All guard-rails and frogs should be filled in to the head of the rail with an iron block to prevent the feet of employees from being caught in them in the face of an approaching train. The dimensions of frogs are indicated by angular measurements and they are classified by standards to facihtate their manufacture as stock material. In Great Britain, this classification is determined by the rela- tion of the altitude to the base of the isosceles triangle thus formed, and varies from 1 to 4 to 1 to 12. The obtuse angle at a crossing is limited by the Board of Trade to 1 to 8. In the United States, frogs are classified by numbers. The number refers to the relation of the length of the frog to the distance between the gauge-Hnes at its heel; the length being the distance in feet from the theoretical point of intersection of the gauge-lines to the point at which they are one foot apart. The num- bers range from No. 4 with an angle of 14° 15' 00" to No. 24 with an angle of 2° 23' 13".2 The No. 8 frog is ordinarily used in yards and No. 10 in running-tracks. 1 A slip-switch is a switoh-eonneotion inserted in a crossing in such a manner as to provide connection between the two lines of rails. A double-slip provides for communication in either direction from either track. 2 For dimensions of standard frogs and switches, with length of leads and switch-rails, see "Freight Terminals," J. A. Droege, 1912, p. 47. For length of cross-overs, ibid., p. 50. ROADWAY 243 Maintenance of Way Notwithstanding the improvement in processes of manufacture and in the designing of details, the standard track is still unsatisfactory with respect to its stability and endurance under heavy traffic. Like the macadamized road, its surface suffers continual deterioration from roUing- friction. But the macadamized road has only to offer a dense and coherent resistance to wheel-treads whose loads are transmitted normally to the sub- structure without internal disturbance. In this respect, it somewhat re- sembled the primitive stringer-track, in which the strap-rail was only in- tended to resist wheel-wear. This relation disappeared when the rail was required to act also as a girder. The incessant disturbance thus occasioned to the structure of the track, by the passage of trains, is only counteracted by an incessant reconstruction of it in every part. The fastenings are to be tightened up ; the ties are disturbed in their beds to restore the align- ment ; there is another disturbance to renew the defective ties and defi- cient ballast and, at periodic intervals, a general upheaval with the rail- replacement. In addition to the labor required to keep the track in effi- cient condition, the cost of rehandUng stone-ballast, of repairs to fencing and road-crossings, of keeping down the undergrowth on the right-of-way and of the general policing of the roadway, is no small part of the expense incurred in the Roadway Department. The extent to which the traffic affects the relative deterioration of the principal elements in track-con- struction may be inferred from the distribution of the items of cost in road- way maintenance, derived from the experience on a number of railroads in the United States, for periods of two to eight years and covering 10,127 miles of main line and 3192 miles of branch-lines, with an average daily traffic ranging from 3 to 11 passenger-trains and from 7 to 44 freight- trains.* Efficiency in track-maintenance has been increased by the use of mechanical appHances, of track-jacks, ditching machines, tamping machines, ' Distribution of Items of Cost OF Roadway Maintenance Peh Cent. Per Cent. 11.7 12.6 24.6 5.4 8.5 8.3 28.9 Track and Roadbed . . Ties Earthwork and Ballast Surfacing Track Switches, Frogs and Sidings .... St.riiptiirp*? .... .... 54.3 Bridges Buildings ... 16.8 "Railway Location," Wellington, p. 120. 244 EFFICIENT RAILWAY OPERATION unloading-plows for distributing ballast, steam-shovels and steam-cranes.* The gasoUne-engine is being gradually utilized to diminish or to reinforce manual labor in track-maintenance. The gasoline motor-car may well replace the lever-car, which was itself a relief to the gangs that worked their way to and from their daily tasks on pole-cars. Gasoline weeding and mowing machines have also been devised. An army of men is daily employed in tamping track with primitive implements, and here seems to be an opportunity for using mechanical appliances to advantage. An electrically operated machine weighing 1900 pounds tamped a mile of track in fifteen days at a cost of $157, against hand-tamping in 26 days at a cost of $299. A pneumatic machine weigh- ing 37 pounds (censuming 19 cubic feet of free air at 70 pounds' pressure), with several hundred feet of |-inch hose, is coupled to a gasoline-engine of 12 horse-power. Two of these machines, the entire outfit weighing 2495 pounds, and operated at 70 pounds' pressure by 21 men, tamped a mile of track at a cost of $196, against hand-tamping at a cost of $282. The average settlement after six months was 0.33 foot in the machine-tamped track and 0.67 foot in that which was tamped by hand.'' The substitution of machinery for manual labor has also been applied on work-trains. The outfit consisted of a Ught locomotive, two dmnp-cars of 20 cubic yards' capacity each, a power-ditcher, a spreader and a crew of eight men. The ditching-machine was a small revolving steam-shovel with a half -yard bucket, mounted on a portable track on a flat car between the dmnp-cars. The machine had a hose connected to the tender and sup- pUed its own tanks while the train was in motion. The loads were dmnped clear of the ballast by compressed air from the locomotive. • The cars were loaded from the ditches in about half an hour and, with a haul of one to four minutes, made ten to twenty trips a day, handhng more material than a train of fifteen flat cars of ten-yards' capacity with twelve men. No time was lost in shifting a ditching-plow and plow-cable at each trip. The train requires only a short siding and is conveniently unloaded for filhng bridges or widening banks..' The maintenance of track will doubtless be still further facihtated by such appliances, but, as long as the track is required to resist the wave of deflection by a combination of elasticity and rigidity, so long will the never-ending processes continue of structural disintegration arid restora- tion. When we consider the stresses to which a track is subjected and the disastrous consequences that may result from a failure of a single fasten- ing, or from neglect in some minor respect, it is not surprising that catas- trophes do occasionally occur. In ten years, including 1915, the derail- ments reported by the Interstate Commerce Commission as due to defec- ' Upon the Lehigh Valley Railroad, -with four locomotive-eranes, 4.07 miles of track were relaid with 100-pound rails in six hours ; the old rails being thrown out at the same time. 2 Engineering News, March 11, 1916. a Ibid. ROADWAY 245 tive roadway have averaged 1477 per annum. In 1915, they numbered 1507, or about four or five a day, distributed over some 264,000 miles of line.* Gauge of Track The gauge of a railroad affects its economic efficiency as to the stability of the vehicle in use upon it or as to the cost of construction. Neither of these matters, however, was considered in the establishment of the "stand- ard" gauge now predominant in the railway mileage of the world. This standard-gauge originated when the flange was put upon the tires of the colliery wagons to enable them to run upon an "edge-rail." The resulting relation between the wheel-gauge and the track-gauge happened to be 4 feet 8^ inches. This relation accompanied the development of the col- liery tramway into the commercial railway in England, whence it extended to the continent of Europe, to the United States and elsewhere with the introduction of the EngUsh locomotive. The first departure from the standard-gauge was with the gauge of the Charleston & Hamburg Railroad in South Carolina, which, in 1830, was fixed at five feet by its engineer, Horatio F. Allen.^ As the standard gauge extended from England to the continent of Europe, so did the five-foot gauge extend from Charleston, northward to Wilmington, N. C, along the coast, along the eastern slope of the Blue Ridge to Richmond and Wash- ington, and to the Ohio and the Mississippi rivers. An instance of the fortuitous circumstances under which the gauge of track has been established by law, was shown in the estabUshment of a standard gauge in the State of Ohio. The first railroad in that State was begun in 1835 by the Mad River & Lake Erie Railroad Company, now merged with the New York Central Lines. In 1837, the president of that company witnessed the trial trip of the first locomotive built by Rogers, 1 Derailments Due to Defective Roadway Dueing Ten Years, Including 1915 Causes Annual Avbbaoe Ten Yeabs In 1915 Broken rail . . . Spread rail .... Soft track .... Bad ties .... Sun-kink . . Irregular track . . Miscellaneous defects Total .... 274 182 218 45 24 326 408 1477 272 90 354 61 32 415 283 1507 ^ Allen went to England in 1827 to purchase rails for the Delaware & Hudson Coal & Canal Company's railroad and had then the opportunity to study railway construction in its earlier stages. 246 EFFICIENT RAILWAY OPERATION Ketchum & Grosvenor in Paterson, N. J., and was so much pleased with the sound of the whistle, which was then a novelty, that he bought the locomotive on the spot, and laid his track to suit its gauge, which happened to be 4 feet 10 inches. At the next session of the legislature, that gauge was estabhshed by statute as the standard-gauge in Ohio.' The next departure from the present standard-gauge originated with the construction of the Erie Railroad begun in 1839 on a six-foot gauge, in fulfillment of a provision in its charter forbidding a connection with rail- roads leading to other seaports than New York. The Erie Railroad was the parent of a broad-gauge system extending westward to Cleveland, to Cincinnati and to Chicago ; and, in 1864, to St. Louis, making it the first line of uniform gauge to connect the Atlantic seacoast with the Missis- sippi River. Up to that time, the diversity of gauges in the rapidly ex- tending railway system of the United States was compelling frequent transfers, with serious inconvenience to travelers and delay to freight- shipments. In 1852, connection was established from Buffalo along the southern shore of Lake Erie by the completion of the Painesville, Ashtabula & Geneva Railroad. Continuous train-service was, however, broken at the borders of Pennsylvania and the adjacent states of New York and Ohio by an intervening link of twenty miles, which had been built on the six-foot gauge as a connection of the Erie Railroad. The traffic between Buffalo and Cleveland was, however, of so much greater value that, to secure it in competition with the steamers on Lake Erie, preparations were made to change this line in Pennsylvania to standard-gauge. In anticipation of the impending loss of business by the hotels and other interests in the city of Erie, which were profiting by the transfer at that point, a mob, led by the mayor, tore up six miles of line. Five attempts to relay the track were resisted by violence and only after a lapse of two months was the "Erie: War" terminated by the proclamation of martial law. A similar "Battle of the Gauges" took place in Great Britain in conse- quence of the adoption of the seven-foot gauge in the construction of the Great Western Railway, which was opened from London to Bristol in 1835,. and which led to a mileage of 1456 miles of that gauge. Other lines had been built on a five-foot gauge and the consequent inconvenience to traffic aroused an agitation in favor of a uniform gauge. This battle was, however,, fought in Pariiamentary committees by extravagantly-feed lawyers and was only terminated by the appointment, in 1849, of a Royal Commis- sion to investigate the matter. The resistance to innovation was too strong; to be overcome. In 1872, a third rail was laid on the broad-gauge lines to admit of the passage of standard-gauge equipment, but it was not until 1892 that the broad-gauge lines were entirely changed. As a result of this contest, the gauge of the Irish lines was established at 5 feet 3 inches, that ! 1 "When Railroads were New," C. F.Carter, p. 222. ROADWAY 247 being an alleged compromise between the standard-gauge and the broad- gauge. There was yet another gauge that gained notoriety in the United States about 1870, the "narrow-gauge" of three feet. It had nothing to recom- mend it except its low cost, which was largely due to cheap construction and equipment. It gradually disappeared as the hnes of that gauge were absorbed by standard-gauge lines or were reconstructed. At the present time, there are probably not more than 1600 miles of hne of three-foot gauge remaining in this country, otherwise than on industrial lines. While the narrow-gauge was in vogue, several plans were devised for exchange of freight-traffic with standard-gauge lines without breaking bulk. Of these, the drop-pit was most in use. It was not applicable to the exchange of passenger-car bodies of standard-gauge. It worked fairly well, however, with lines of five-foot gauge, especially with Pullman sleepers, as their trucks were built to standards and could readily be inter- changed. Transfers of trucks were usually made within fifteen minutes, and with little inconvenience to the occupants. In the Act of Congress, approved July 1, 1862, chartering the several Pacific Railroad companies, it was provided, "That the track upon the entire line of railroad and branches shall be of uniform width to be deter- mined by the President of the United States, so that, when completed, cars can be run from the Missouri River to the Pacific Coast." By a further Act, approved March 3, 1863, it was established, "that the gauge of the Pacific Railroad and its branches throughout their whole extent, from the Pacific Coast to the Missouri River, shall ,be, and hereby is, established at four feet eight and one-half inches." This action of Congress sealed the fate of the broad-gauge on every trunk line, from the Atlantic Coast to the Missouri River, which aspired to transcontinental traffic. By 1882, that gauge had yielded on the Erie Railroad and its connections to the demands for uninterrupted intercommunication. . ' Change of Gauge on Southern Railroads The increasing importance of the traffic crossing the Potomac and the Ohio rivers constrained the principal north and south lines of five-foot gauge to take similar action. The organization for a change to standard-gauge on the entire Southern system of nearly 15,000 miles was deputed to a committee of general managers, which appointed sub-committees of offi- cials representing each operating department to recommend, as to details, the course to be pursued in effecting the change of gauge. For some two years this matter was in hand and preparations for making the changes in motive power, rolling stock and roadway were made in accordance with the committees' recommendations. New locomotives were fitted with dished driving-wheel centers, so that the change could be made by merely reversing the wheels on the axles. 248 EFFICIENT RAILWAY OPERATION New truck-wheels and axles under all equipment were fitted with a wheel- seat of sufficient length for each wheel to be pressed back one-half of the difference in gauge, or If inches, this space being filled in by a collar next to the journal-box. A railroad-track is lined continuously along the same side, known as the "line side," and the other side, or the "gauge side," is brought into line by the track-gauge. In preparing for the change, the Une-side was not disturbed, but inside spikes were par- tially driven on the gauge-side at the prescribed distance to suit the- stand- ard-gauge. Spare switches of that gauge were distributed at all sidings connected with the main line. Advantage was taken of the opportunity to increase the equipment by additional locomotives and cars of standard- gauge, parked at convenient points on long stretches of sidings of standard- gauge, which, in many instances, were afterwards utilized as passing-sidings or as sections of second track. The transportation officials collaborated in adopting provisions for running the five-foot gauge equipment off the line on to temporary sidings as the change of gauge progressed, and many sidings were changed in advance. An army of extra laborers was re- cruited for the special occasion, distributed in gangs with men of the regu- lar track-forces, and all were provided with necessary tools and food- rations. The date for beginning the general change was fixed for May 31, 1886, and the day before that date every alternate inside spike was drawn from the ties on the gauge-side of the track. Upon the completion of these preparations, word was passed by telegraph throughout the five-foot gauge system. As the order was extended to the roadway department, the gangs distributed along the line drew the remaining inside spikes, lifted the rails on that side out of the outside spikes, still connected together by the fish- plates, and threw them over with crow-bars against the inside spikes pre- viously set to standard-gauge. These spikes were then driven home, the switches and frogs on the main line were readjusted and each gang made its report to headquarters, where the general manager and his staff remained, through the day and night, recording the progress of the work on maps of the Une, and ordering changes of forces to points where the work was being delayed. The change was effected with such promptness that, on the morning of June 1, 1886, the Florida Express ran over the changed hne from Wilming- ton, N. C, to Jacksonville, Fla., a distance of 498 miles, at a speed of nearly 50 miles an hour, preceded by pilot-locomotives, and arrived at Jackson- ville on time. On the Charleston & Savannah Railroad, 128.7 miles of main line and sidings were changed in six to eight hours by 359 men, dis- tributed in 2k gangs. On the Louisville & Nashville Railroad, 1806 miles of main hne and sidings were changed in a single day by a force of 8763 men. Upon the completion of this remarkable undertaking, the standard- gauge became virtually universal in the United States, and was made so ROADWAY 249 formally by a resolution of the American Railway Association, adopted April 7, 1897, as follows: "That 4 feet 8^ inches shall, hereafter, be the standard gauge of all tracks owned by the railroad companies forming this Association. This gauge shall be the distance between the heads of the rails at right angles thereto, at a point five-eighths of an inch below the top of the rail." There is still a considerable mileage of 4 feet 9 inches gauge, which slight difference does not interfere with through-train service. Different Standards of Gauge Although the standard-gauge is, by far,- the more general in North America, yet in Great Britain and in most of the countries of Continental Europe, there are still more than twenty different gauges. In Great Brit- ain and Ireland alone there are twelve gauges.^ Those wider than "stand- ard-gauge," or of "broad-gauge," range from 5 feet in Russia to 5 feet 3 inches in Ireland and South America, 5 feet 5|- inches in Spain and Portu- gal, and 5 feet 6 inches in British India. "Narrow-gauges" range from 1 foot 11|- inches to 3 feet 6 inches. The narrow-gauge lines built with Brit- ish capital are generally 3 feet 6 inches, or of the meter-gauge where capital has been obtained in Continental Europe. In AustraUa, as in the United States, the differences of gauge have caused dissatisfaction. Out of 18,979 miles of line in 1914, 18,035 miles were owned by the several states and, of this mileage, 10,217 miles were not of standard-gauge. In Queensland and South Australia, there were 6066 miles of 3 feet 6 inches gauge and, in Victoria, 3525 miles of 5 feet 3 inches gauge. Since the unification of government in AustraUa, uni- formity of gauge has become a political issue. In South America, a uniform gauge is a commercial as well as a political question. In Brazil, the narrow-gauges predominate, though there is a considerable mileage of standard- and of broad-gauge lines. The standard- gauge prevails in Uruguay and in Paraguay, and in the adjacent provinces of the Argentine Republic, though in that country important territory is occupied by other gauges. In Chile, there are seven different gauges on 6738 miles of line. In Bolivia, the meter-gauge prevails, though on an important line, the Antofagasta Railway, it is 2 feet 6 inches. Peru is committed to the standard-gauge, but in the states northward as far as Mexico, the railways are, for the most part, narrow-gauge, with the excep- tion of the Panama Railroad, which is five-foot gauge. In South Africa, the gauge is 3 feet 6 inches, and this will probably be the gauge of the projected "Cape-to-Cairo" line, as 1500 miles of line in Egyptian Sudan is also of that gauge. For the unification of the line to the Mediterranean, it will be necessary to change the standard-gauge of the extension through Lower Egypt. 1 See Appendix V, Table XI. 250 EFFICIENT RAILWAY OPERATION The five-foot gauge was introduced by Colonel Whistler into Russia from the United States, with the construction of the line from Petrograd to Moscow, opened in 1851, and has become the established gauge through- out the empire from political and military considerations. Through-train service is maintained with Western Europe by change of trucks. In British India, the mileage of 33,000 miles is nearly divided between the 5 feet 6 inches and the meter-gauge. In New Zealand, Tasmania, Japan and Newfoundland, the gauge is 3 feet 6 inches. In Belgium, in 1910, there was a meter-gauge of 2371 miles on 148 lines of "light railway," of which one-half was laid on public highways.' Turnouts and Sidings The location and arrangement of sidings becomes an important matter with increasing traffic. In early railroad operation in the United States, the single siding at a way-station was used indiscriminately for standing cars and for passing trains. When there were cars in the siding, trains had to "see-saw," with the standing cars coupled ahead of one of the loco- motives, and with consequent delay. Often these "turn-outs" were "blind sidings," with but one switch, and it required considerable clever- ness on the part of the trainmen to pass trains that were too long for the sidings. Blind sidings are now rarely used except as a temporary expedi- ent for work-trains. With sidings at way-stations, where the station- building is used for both passenger and freight service, it is preferable for the building to stand between the two lines, so as to leave the way unob- structed between the trains and the station-building. The passing-siding at a way-station should be on the opposite side of the line. Passing-sidings, however, should not be located where their use might incommode the station-service or block a road-crossing. The clearance-posts at sidings are usually placed at the side of the track and painted white to make them distinct. A preferable plan is to ' Proportion op Mileage of Different Gauges Europe . . North America South America Asia . . . Africa . . . Australasia . The World . Standabd-oauos Per Cent. 71 98 14 7 17 20 71 Nahroweb Gauges Per Cent. 7 1.99 50 50 83 58 14.5 Broader Gattqes Per Cent. 22 0.01 36 43 22 14.5 > Railroad Gazette, International Issue, July 8, 1910. See also Appendix I, Table I. ROADWAY 251 use short stone-pillars, placed in the middle of the siding-track with the tops whitened and made even with the surface of the ties. When so placed, men can not stumble over them, and it can be seen at once whether cars on the siding are clear of the main line. On single-track roads, as the freight-traffic increases, there will occur a normal congestion of meeting and passing trains within certain limits, due to the exigencies of the service. Wherever these conditions prevail, passing- sidings should be specially arranged for relief. The spacing of these sidings should be carefully considered with reference to the points on a division at which there is a congestion of train-movements. A location should be selected away from any way-station, preferably at a summit and free from road-crossings. At this point, there should be a siding on each side of the main line, long enough for at least two freight-trains with the switches "lapped"; that is, so arranged that the entrance- switch of each right-hand siding could be reached by an approaching train in advance of the switch by which a train from the opposite direc- tion would leave the siding on the opposite side. Such an arrangement lessens the probability of collisions at passing-points. Water-cranes should be so placed that, when necessary, the tenders can be filled while the trains are on the sidings ; and, in connection with the signal-cabin, there should be a comfortable room where the trainmen can receive their orders.^ Industrial sidings must be located to suit the industries which they are to serve, but they should be entered from a siding parallel with the main line, and on this siding trains should stand while at work there, in order to keep the running-tracks unobstructed. "Cross-overs" in double-track should be laid with trailing-switches in each track, so that they can only be entered in reverse direction to the usual train-movement, in order to avoid facing-switches in the running-tracks and the possibility of accident from an "open" switch. With increasing traffic, passing-sidings cannot afford the necessary relief .to a single-track road. Further relief may be found by extending passing-sidings into running-sidings, or virtually stretches of second track, but the line would still be operated under single-track conditions. The only permanent remedy is the construction of sections of second track long enough for operation as double-track. Just when the time may be said to have arrived for that expenditure, in the increasing traffic of a single-track road, is a debatable question, which is generally deferred until the resulting congestion becomes unbearable. Relief is meanwhile sought by bunching the freight-trains in sections. In this way, delay to passenger-trains may be prevented at passing-points, but the freight-service is thereby seriously affected, while a hot box or a broken axle may throw the entire train-move- ment out of joint for hours at a time. ' See "Economics of Railway Operations," Byers, p. 636. 252 EFFICIENT RAILWAY OPERATION DOXJBLB-TKACK IN THE UNITED StATES AND GrEAT BhITAIN There has been very Httle original construction of double-track line in the United States. The road from Philadelphia to Columbia, 82 miles long, was opened in April, 1834, with a single track and a second track was completed in the following October. The second track of the Pennsyl- vania Railroad Company was completed across the state in 1877, and that company now has a four-track line from New York to Pittsburgh. The New York Central line is four-tracked between Albany and Buffalo, and the Harlem line, which parallels the main line, virtually constitutes a four- track line between New York and Albany. The West Shore Railroad, in connection with the road from Syracuse through Geneva and Batavia to Buffalo, also provides an auxiliary route between New York and Buffalo- In 1908, on 230,000 miles of railroad in the United States, but eight per cent, was double-tracked. In 1914, on 256,000 miles, the second-track mileage had been increased to ten per cent. The third and the fourth track mileage had been increased in the same period from 3490 to 4767 miles. There has been a remarkable development of yard tracks and sidings which, in 1908, amounted to 79,453 miles and, in 1914, to 98,285 miles ; an increase of 18,832 miles, or 23.7 per cent, while the line-mileage increased but 11.3 per cent, during the same period. In 1914, the mileage of yard tracks and sidings amounted to 38 per cent, of the line-mileage and to 25 per cent, of the total track-mileage. In the Eastern Territorial Dis- trict, with but 25.3 per cent, of the total line-mileage, the yard and siding mileage constituted 40.8 per cent, of the total mileage of that kind in the whole country. In the several territorial districts, the proportion of yard and siding mileage to that of running-tracks was, in the Eastern District, 47 per cent. ; in the Southern District, 21 per cent. ; and in the Western Dis- trict, 28 per cent.^ Double-track construction was common in England from the inception of railway development, as the early lines were intended to provide for a heavy coal-traffic. Yet a large part of the railway system of Great Britain and Ireland is still single-track. The proportion to line- mileage in 1910 was, in England and Wales, 38 per cent. ; in Scotland, 58 per cent. ; in Ireland, 80 per cent. The statistics of yard tracks and sidings are indicative of the relative importance of the manufacturing interests in the several countries forming the United Kingdom. In this respect, it is of interest to compare the track- age in the Eastern Traffic District in the United States in 1914 with that in England and Wales in 1910, as a measure of the railway facilities pro- vided in each region for commercial intercourse. Though the total track- age per square mile in the Eastern District is only about one-half of that in England and Wales, it is double in proportion to the population.^ ' For details of Track Mileage in the iTnited States, see Appendix V, Tables XII to XIV. ' See Appendix V, Tables XV and XVI. ROADWAY 253 Electric Traction Requirements. Third Rail and Overhead Systems The details of track-construction have been somewhat affected where electricity has been substituted for steam as a tractive force. The direct connection of the motors with the driving-axles of the tractors, resulted in an increase of weight not spring-borne which aggravated the hammering upon the rail-ends. The accompanying lowering of the center of gravity caused a greater horizontal stress upon the rails that increased the diffi- culty of keeping the track in line. The passage of a single train at high speed sometimes undid a whole day's work of a track-gang. These de- fects were somewhat remedied by placing a four-wheel truck at each end of the tractor, and by raising the motors above the springs and by re- turning to the intervention of connecting-rods and cranks for transferring the rotary motion to the driving-axles. With the direct-current system, the current is transmitted through an insulated "third rail," placed upon the ties along the outside of one of the running-rails. This rail is generally of the inverted "T" pattern, though on the Pennsylvania Railroad a rail-conductor has been adopted of a special section and composition weighing 150 pounds per yard, with con- ducting activity equivalent to that of 1.9 square inches of copper. The third rail is usually protected from accidental contact by a wooden cover- ing or/ "roof," except in some cases on branch-lines within a fenced right- of-way. The current is either transmitted to the motors by contact with the upper surface of the rail or, as on the New York Central Railroad, the contact-shoe comes in contact with the lower surface of the feed-rail, and the rail is so protected that a person could not come in contact with it unless he knelt down and put his hand beneath the roof. The rail is also better protected from snow or sleet.^ To prevent encroachment upon the space necessary for operation, and to facilitate the interchange of equipment, the American Railway Associa- tion has approved specifications for standard location and clearance in third-rail construction which prescribe that any device attached to the permanent way may not project more than "2^ inches above the top of the track-rail in the space from a point 19f inches from the gauge-line to a point 6 inches above the gauge-line." The clearance in all equipment had also to be carefully defined so that no part should come witjiin a space ' With top-contact, the third rail rests upon insulators bolted to the ties. It is gauged 26 inches from the gauge-line of the running-rail, and its upper sur- face is 2i inches higher than the top of that rail. The top of the roof is about 6 inches higher than the top of the running-rail. With under-contact, the third rail is insulated in brackets bolted to the ties, leaving its under surface free. It is gauged 27 inches from the running-rail and its under surface is 2f inches higher than the top of that rail. The total height is 82 inches higher than the top of the running-rail. The third-rail system requires accurate alignment. 254 EFFICIENT RAILWAY OPERATION 6 inches above the top of the rail and 18i inches outside the gauge-line, which includes an allowance of 2 inches for horizontal variation and 4 inches for vertical variation due to wear of journals and brasses, for compression of springs and for variations in construction. With the usual voltage of 600 volts, third-rail insulation is well main- tained and, experimentally, with 1200 volts, though with greater difficulty under the usual conditions of track-maintenance. The freezing of snow and ice on the rail causes sparking by loss of contact. There has been some difficulty experienced from the third rail creeping on steep gradients, causing a fracture of the insulation. In yards and terminals, it is not practicable to secure continuous contact at switches between the rail and the motors. In such cases, auxiliary contact must be supplied, except with multiple-unit trains of more than two cars. Electric traction with alternating current at high voltage is not prac- ticable with the third rail as a conductor. Recourse is then had to the overhead system, as developed from the street-railroad trolley wire. In the operation of heavy trains at high speed, the conductor must be held in a uniform plane with sufficient flexibility, and compensation provided for the expansion and contraction due to changes of temperature ; other- wise, there will be sparking with rapid wear of the collectors. This pur- pose is attained by the catenary suspension of an auxiliary steel wire, to which the copper wire is secured, by hangers, varying in length in each span to maintain the copper conductor in a uniform horizontal plane; danger from a fallen wire is also obviated. Deviations from the normal line of contact are admissible to a distance of thirty inches, horizontally and vertically. There is no disturbance from snow-drifts nor inter- ference with track maintenance.' The catenary system is attached by insulators to wires stretched across the track at intervals, and supported by iron posts. An auxiliary wire is sometimes interposed between the main suspension-line and the con- ducting-wire in order to diminish the undulations in operation, though experience has proven that the simple catenary is more reliable. The spans may extend to as much as 100 yards. At this distance, the lateral deviation from the middle of the track around a three-degree curve is not over 8^ inches, which is admissible with the pantograph or bow-collector. Greater deviation is corrected by wire-guys attached to posts beside the track. * On the New Haven line of four or more tracks, the overhead system is of a more expensive character than has been adopted elsewhere, on account of the very high voltage carried by the conducting-wire. Steel bridges 1 For the safety of trainmen in giving lamp-signals from car-roofs, the wire should be not less than 24 feet above the rails in the open. The American Rail- way Association has recommended that, in freight-oar construction, there should be a limit of 15i feet from top of rail to brake-staff and wheel, to allow cars to pass over roads electrified by the overhead system. ROADWAY 255 span the tracks at intervals of one hundred yards, supporting a cable over each track upon porcelain insulators, from which the conducting wire is suspended. At intervals of two miles, the bridges are of heavier construc- tion on concrete foundations. To these anchorage-bridges, the cables are attached by strain-insulators. The conducting-wire is attached to the suspension-wire every ten feet by triangular hangers of varying length, so adjusted that the line is maintained in a horizontal plane six inches below the middle of the catenary. At overhead-bridges, there is a specially con- structed insulation of long porcelain corrugated tubes mounted on an iron pipe attached beneath the bridge, to protect the insulation from dirt and moisture. Both rails of all tracks are bonded with compressed-terminal bonds placed around the fish-plates. Where tracks diverge, a section- insulator is inserted in the conducting-wire and the diverging-wire is con- nected by a frog. Deflecting-wires are placed in the angle between the two conducting-wires, so arranged that the collecting-bows can not catch in the frogs. Two feed-wires constitute auxiliary lines to the main conduc- tors, with which they are connected at each anchorage-bridge, through a circuit-breaker, by which different sections may be isolated in case of accident. The anchorage-bridges also carry shunt-transformers for oper- ating the circuit-breakers, besides the lighting-circuit and the wires and conduits for operating the auxiliary-control circuit and the signal-appara- tus. The semaphores are mounted beneath the bridges, which are pro- vided with footways protected by hand-rails. There has been a further development of this overhead system on the New York, Westchester & Boston Railway, which is a four-track subsid- iary of the New Haven line. Here, the spacing of the spans around curves has been reduced to a minimum of two hundred feet. Where the curva- ture exceeds four degrees, pull-off poles of steel lattice-work, between the bridges, bring the conductors over the median line of the tracks. The four catenary cables are combined in a single-system at points 75 feet on each side of the bridges by cross-bars, or stretchers, of 3-inch "I "-beams. Intervening catenary cables are attached to these stretchers by porcelain insulators, and from each cable the copper conductor is suspended by hangers, 10 feet apart. Below each conductor is a steel contact-wire to take the wear of the pantograph-bows. It is held If inches from the con- ductor by clips placed midway between the catenary hangers. On the Butte, Anaconda & Pacific Railway, which is operated with a direct current of 2400 volts, a 4-0 copper wire is suspended from a steel catenary on loop-hangers that permit the wire to ride up and down under pressure of the collector, independently of its catenary support. The collector has a 5-inch steel-tube roller, giving a service of nearly 30,000 miles at a maximum speed of 50 miles an hour. A similar overhead system is in use on the Chicago, Milwaukee & St. Paul Railway, with two wires, side by side, alternately suspended from the same catenary. As the col- 256 EFFICIENT RAILWAY OPERATION lector passes beneath the clip of one hanger, the other wire is hanging free, and there is no tendency to sparking. The danger to persons or property on the railroad right-of-way from the crossing of electric-light and power-transmission lines, and also the trouble to telegraph, telephone and signal wires occasioned from induction, has caused the American Railway Association to issue regulations on the subject covering twenty printed pages. The more important of them, as affecting roadway efficiency, are as follows : I. That the poles or towers supporting a crossing-span shall not be less than 12 feet from the nearest rail on the main line nor less than 7 feet from sidings. II. That the wires or cables shall not be less than 30 feet above the top of rail, at least 4 feet above bare telegraph, telephone or signal wires and 2 feet above insulated wires. III. That lightning-protectors shall be thoroughly grounded at each crossing-support. IV. That poles shall be properly guyed or braced. The regulations also include specifications as to materials and methods of construction, and provisions as to tests and inspection. The observance of standard clearance-hmits requires frequent and careful attention. The most efficient means for detecting inconspicuous encroachment is secured on the Pennsylvania Railroad by the passage of a "clearance car" over the line. The apparatus for this purpose is at- tached to a cabin at one end of a flat car. It consists of sets of templates, to which are attached wire fingers or feelers. The main template has a width of 10 feet, extending from 2 feet to 12 feet above the top of the rail. Immediately in front of it is another template for measuring the clearance at bridges and in tunnels, at elevations of 17 to 20 feet above the rail, wliich may be illuminated by electric lights. The feelers are 2 feet long and 6 inches apart, hinged to the sides and top of the templates and held in position by friction. A board with a set of feelers one inch apart meas- ures the projection of the cornices of adjacent roofs. Graduated scales indicate the distance of objects touched by the feelers. An attachment at the rear of the car indicates the degree of curvature on a scale within the cabin, and another shows the elevation of one rail above its opposite. With this apparatus, it is practicable to take measurements at a speed of 40 miles an hour, with an observer reading the scale and another person recording his readings.' Monuments for Curves and Right-of-way Points of curves should be indicated by monuments of stone or of con- crete, placed between the ties on single-track and in the clear-way on double- track, and, at intervals, long tangents should be similarly preserved. 1 The Railway and Engineering Review. ROADWAY 257 Changes of grade should be indicated by bench-marks. The same care should be taken of the boundaries of the right-of-way and of real estate. It is important that such boundaries should be clearly defined by substan- tial walls where such property is valuable, and elsewhere by fences of strong galvanized-wire netting and hardwood posts creosoted below the surface. These facts, together with the location of sidings, stations, road- crossings and water-tanks, should be shown on maps, on a sufficiently large scale, with symbols showing whether the right-of-way is held in fee-simple or by easement. On an accompanying profile of the line should also be indicated the natural contour, the location, size and character of all open- ings and of stretches of rail of different size and pattern. The same in- formation should also be compiled in note-books of pocket-size for use out on the line. By such means, much time and expense will be saved in maintenance and in connection with proposed changes and improvements. Signals and Interlocking Plants The construction and maintenance of fixed train-signals is a function of the Roadway Department. With their more general use and the intro- duction of electric service, including light and power equipment, the signal engineer has become an important member of the Roadway staff. As the character and interpretation of signals for controlling the movement of trains pertains directly to the Transportation Department, attention here will be given only to the construction and maintenance of signal-apparatus. The earliest fixed signals on railroads in the United States were station sign-boards, mile-posts and crossing-warnings. To these were added whistle-posts for crossings and sign-boards for approaches to stations. It was not until the control of train-movements other than by time-table, passed, with the advent of the electric telegraph, from the train-conductor to the train-dispatcher, that fixed signals were established at the telegraph- stations to stop trains for orders. When the frequency of trains compelled the construction of a second track, fixed signals between such stations be- came track-accessories in connection with the block-system, which was introduced from England by the Pennsylvania Railroad management in 1845, and by 1864 it was extensively used on its lines.^ Fixed signals for controlling the movement of trains were developed from the principle of the indications given by the movable arms of the semaphore-telegraph which was introduced in England by the Admiralty between London and Portsmouth during the Napoleonic Wars. That principle has persisted in railway-signaling apparatus. It was first ap- plied to indicate the movement of switches, and then the semaphore-arm was so connected with the switch-lever as to move with it. The coUoca- 1 The block system now covers the four-track Unes from New York to Pitts- burgh and from Philadelphia to Washington, and in three years to September, 1914, $6,000,000 had been expended in its extension. 258 EFFICIENT RAILWAY OPERATION tion of several switches led to their being controlled by a single switchman, stationed in an adjacent cabin and actuating them by connections com- posed of wires, rods, pulleys and levers. From such appliances was evolved the " manual control " system of interlocking switch-points with semaphore- signals known as the "Saxby" system, which came into general use in Great Britain after the year 1856. This system was further developed in the "block" system, with provision for signals to trains entering or leav- ing a block, which were operated from cabins adjacent to the signal-posts. Even at an earlier date, in 1844, the electric telegraph was employed in connection with the block-system on the Yarmouth & Norwich Railway. By 1873, there were 13,000 interlocking levers installed on the London & Northwestern Railway. A Saxby & Farmer machine was placed in serv- ice on the New York Central line, in 1874, at Spuyten Duyvil ; but the first extensive installation in the United States was upon the, New York Elevated Railway in 1877. The interlocking of switches and crossings with signal-indications is an essential feature of the block-system. The purpose is to provide for the safe passage of trains over routes that may conflict with other routes ; such as tracks crossing at grade, or the junction of two or more tracks, or at a draw-bridge. An interlocking-plant consists of an interlocking- machine, centrally located at a height affording, a cleaf view of the line ; signals to control trains passing over the several routes ; and connections from the machine for operating and locking the signals and switches. The operating levers are so interlocked that it is mechanically impossible to give conflicting signals, or to give any signals until the switches have been properly set and locked. Interlocking-machines operated solely through mechanical connec- tions are more generally in use because of the comparatively low cost for their erection and maintenance. The lines of wire used in the older plants for connecting the levers with signals and switches, have been superseded by one-inch pipe traveling on anti-friction carriers. Expansion and contrac- tion are provided for by a "compensator" in eyery seven hundred feet of line. Turns are made either by bell-cranks or by deflecting-stands sup- ported upon concrete or cast-iron foundations, as are also the signal- masts of pipe set in sockets. Electricity is used in mechanical interlocking for the operation of dis- tant-signals, which are usually located on high-speed tracks at least half a mile from the home-signals whose indications they repeat ; also for the announcement of approaching trains; for repeating the indications of signals invisible from the cabin ; for locking certain routes when occupied by a train, and for other purposes. An electro-mechanical system has been devised in which, while the switches are manually actuated by mechanical means, they are locked and released electrically, and all signals are elec- trically operated. ROADWAY 259 The operation of signals and switches through mechanical means, is usually limited to distances of 800 feet for a switch and 2000 feet for a sig- nal. Difficulties in operating mechanical connections beyond those Umits, and in keeping lengthened lines in adjustment, have led to the intervehtion of inorganic forces in power-interlocking. The mechanical system is suffi- cient for operating a set of twelve to sixteen levers by a single attendant. Even with the aid of electric appliances, it works too slowly for a busy terminal. Under such conditions, power-interlocking becomes indispen- sable. Power-interlocking is accomplished either by hydraulic, hydro-pneu- matic, pneumatic, electro-pneumatic, electro-gas or all-electric power. Hydraulic action was combined with compressed air in the United States in 1884, but was superseded about 1891 by the electro-pneumatic system, in which compressed air, as the motive power, is controlled by electrically operated valves. About 1900, an electric motor system, or all-electric, system came into use. Automatic Signals and Track-circuit The first of the electric power-systems was developed in the United States in connection with the first automatic signal-system, known as "Hall's disk system," which was invented in 1871, and was introduced on the Eastern Railroad in Massachusetts, in 1872. In this system, the wheels of an approaching locomotive pressed upon a treadle beside the rail, which released a disk in the facing signal-instrument. For this pur- pose, the electric track-circuit was first employed ; that is, an electric current that includes the rails as part of its path for changing the position of signals. The continuous or "closed" track-circuit was invented by an American, William Robinson, in 1872. Mr. Westinghouse subsequently utilized compressed air to actuate a semaphore-arm in connection with the track- circuit in the "electro-pneumatic" system. An automatic signal-appa- ratus was patented in England by W. R. Sykes in 1872, but was not ap- plied practically until the opening of the Liverpool Overhead Railway in 1893. An "all-air" or low-pressure pneumatic system was introduced from the United States on the London & Southwestern Railway in 1901.^ Automatic signals controlled entirely by track-circuits were installed on the Lancashire & Yorkshire Railway in August, 1906. Up- to February, 1907, there was no method for actuating track-circuits by alternating currents. On electrically-operated railways, which used the running-rails for the return traction-current, automatic working was accomplished by "depression-bars" beside the rails, operating through solenoids upon the signals. The first automatic signaling on the London District Railway was opened in June, 1903. Automatic signaling on the HaU "normal 260 EFFICIENT RAILWAY OPERATION danger" electro-gas system was installed on the Northeastern Railway, in August, 1904, on a branch-line of 10^ miles.^ The characteristic features of the automatic block-system are con- tinuous track-circuits throughout the whole length of the track and the operation of each signal automatically by electric mechanism. The rails at both ends of the block are electrically insulated. To one end of the insulated section is attached a battery of one or two volts. The opposite end is connected to a resistance-coil which opens and closes the controlling circuit of a power-operated signal. The relay thus energized closes the local circuit of the signal, which is thereby raised to the inclined or "clear" position. As soon as the first wheels of a train enter the block, the electric current is short-circuited through the wheels and axles from one rail to the other, instead of continuing to the relay which, being thus deprived of electrical energy, allows its contact to open. This action de-energizes the magnet which holds the signal-arm in the "clear" position, and the arm then falls by gravity to the horizontal position and so remains until the last pair of wheels in the train has left the block, when the signal again assumes the inclined or clear position. By the use of special circuit- controllers, or "switch-boxes," the opening of the points of any switch in a block instantly causes the signal at the entrance of the block to assume the horizontal position; also, the breaking of a rail under a train, by similarly interrupting the continuity of the track-circuit, prevents the signal from returning to the inclined position, even after the train has left the block. The maximum length of a track-circuit is usually about a mile, except under favorable local conditions ; but two or more sections can be relayed in succession to any required distance, as may be demanded by traffic-conditions. Electric Interlocking By the electric-pneumatic apparatus, which was successfully intro- duced in 1891, valves, electrically controlled by levers in the machine, admit or release compressed air in the working-cylinders and the result- ing movements of switches and ' signals are accompanied by switch and signal indications, to insure that each switch or signal shall follow the movement of its appropriate lever or, failing to do so from any cause, that no unsafe condition shall ensue. The electric mechanism pulls "clear" as soon as there are no wheels in the insulated block-section, and releases the danger-signal as soon as wheels enter the section, or when, from any cause, the track-current is interrupted. Air, at a pressure of 80 to 100 pounds per square inch, operates single-acting cyUnders by means of the electrically-controlled valves. The switch-and-lock movement consists of a double-acting cylinder, an electric valve and the movement itself, secured to a base-plate and 1 See C. H. EUison, Railway Gazette, London, March 9, 1917. ROADWAY 261 bolted to the ties. In one complete movement of the piston, the switch is first unlocked, then moved and finally relocked in the opposite position. The movement is connected to a "detector" bar 50 feet long, or to a series of such bars, lying alongside the rail in brackets. Every time a switch is shifted, the detector-bar must first be raised and lowered. When a train stands upon or moves over the rails, the bar attached to them is held from rising by the tread of the wheels and thus prevents the switch from being moved. This safety device is also used in mechanical plants. Most of the great terminals in the United States are equipped with the electro- pneumatic interlocking-apparatus. In the electro-gas system, liquid carbonic-acid gas, instead of compressed air, is used to move the semaphore-arm from the horizontal to either the inclined or the vertical position, and held in position by an electrically- controlled latch. The electric-motor system, or "all-electric" system, was first used in the United States in 1898, and is installed upon the electrically operated lines of the Pennsylvania. Railroad and the Long Island Railroad in and around New York City. In March, 1910, there were 34,696 of these appliances in service. All the operations are performed electrically, though on the same principle as in the electro-pneumatic system, using 110-volt direct-current motors for operating switches and high signals, and solenoid-magnets for low or "dwarf" signals. Current from storage- batteries in the cabins drives a train of gears, engaging in the ends of levers, or "slot-arms," that are connected to the semaphores by vertical rods extending upward through the hollow mast, and the semaphore-arm correspondingly moves by means of levers and electro-magnets. But one motor and set of gearing are required for any number of iarms on a single post. The electric mechanism pulls "clear" as soon as there are no wheels in the insulated block-section, and releases the danger-signal as soon as wheels enter it, or when, from any cause, the track-current is interrupted. One rail is used solely for operating the signals. The other rail is used in common with other circuits ; the relays for the signal-current being operated at a higher voltage than the return track-current. An alternating-current may be substituted for signal-operation in conjunction with a relay that is irresponsive to direct-current. A signal-system on this principle was installed on the Boston Elevated Railway in 1900. In several plants, working-models in the cabin automatically indicate the presence of trains in each of the sections into which the tracks are divided. By this means, the lever-man can operate the plant without actually seeing either the tracks or the trains. In another electric system, metallic conductors, or "ii^ductive bonds," placed at each end of a block or sub-section, carry the return traction- current around the insulated joints, but prevent the passage of the alter- nating-current from operating the signal-circuit. By this means, both 262 EFFICIENT RAILWAY OPERATION rails can be used for that circuit and for the return traction-circuit, whether that be on the direct or the alternating system. The alternating-current is also used for the operation of signal-motors and electric locks, and for other appUances of automatic block-systems, as also for lighting the signals.' Interlocking machinery has made it possible to operate extensive terminal-yards with facility, economy and safety. The interlocking- plant at South Station, Boston, includes nine signal-bridges and a tower containing an electro-pneumatic plant of 165 levers. In this terminal, there are about 850 train-movements a day; and almost 100 per hour during the rush-hours. At the Hoboken terminal of the Delaware, Lackawanna & Western Railroad the traffic is handled from three interlocking-plants, controlling 627 signal-units. From April, 1913 to April, 1914, there were nearly 29,000,000 switch-and-signal movements, of which but 36 were imperfect. The switches and signals at the Grand Central Station are operated by five all-electric plants, with a minimum of 80 levers and a maximum of 400 levers in the several plants. There are also twelve sub- stations for locaUzing and housing relays and track-apparatus.^ Position-light Signals Colored light signals, composed of an ordinary incandescent lamp in a box behind a lens, have long been in use on trolley-roads for both day and night indications, and on steam-roads in tunnels. But a remarkable inno- vation in track-signaling has been introduced in the "position-light" signal-system, in which the form aspects of the semaphore-indications are given either in the day or the night by a series of uncolored lights. In 1914, Dr. Churchill, of the Corning Glass Company, while working on electric headUghts, found it possible to secure a long range from a small source of light, if it were placed at the focal point of a suitable lens ; and, in collaboration with Mr. A. H. Rudd, signal engineer of the Pennsylvania Railroad Company, arranged a combination of such Ughts in rows, having the effect of a semaphore-arm and doing away with the color scheme. This arrangement was carried into practical effect by the installation of a " position-Ught " signal-system on a section of the Philadelphia Division of the Pennsylvania Railroad, which was then being electrified. The system is now in use from Overbrook to Paoli on 15^ miles of four-track line, and at other localities on that railroad, where it has met with such favor that its use is being constantly extended. Each unit consists of a container, or box, painted dead-black inside, holding a 12-volt, 6-watt, horizontal helical-filament tungsten-lamp, in the exact focal point of a toric ' inverted lens of 2^ inches' focus and 5f 'For more detailed information, see "Railway Signalling in America," J. S. Hobson. Cassier's Magazine, March, 1910. Fully illustrated. 2 "Passenger Terrdinals," J. R. Droege, pp. 49, 55, 114. ' Toric. — A surface generated by the revolution of a conic (especially a circle) about an axis lying in its plane. — Century Dictionary. roadWay 263 inches in diameter, in front of which is a special convex cover-glass of the same diameter. Over the lens, a 4-inch spherical mirror is placed at such an angle that no light can be reflected outward, and that no unlighted unit can appear to be lighted. The cover-glass is of a light-yellow tint, which is much easier to the eyes than white light. For dwarf -signals, the cover-glass is "frosted." The entire range of aspects on the two-arm system, as required in standard practice, may be displayed by sixteen lamps in two groups respectively of ten and six lamps each. The inter- " locking signals are given by a group of three units.^ The signals can be operated with half the current consumption required for the usual colored- light signals, and consequently at half the cost of operation. By doing away with all moving parts except relays,^ the chances of false "clear" indications are reduced to a minimum. For indications in a fog or snow- storm, these signals are thought to be better than any others.' Signal Statistics On January 1, 1901, the manual-controlled system was in use on 24,013 miles of Une in the United States, and automatic-systems on 2295 mUes, being a total mileage of 26,308 miles under block-systems. On January 1, 1908, there were 58,768 miles of line under block-systems, and 86,731 miles on January 1, 1914. At the earlier date, 39 per cent, of the passenger-line mileage was so operated and 46 per cent, on January 1, 1914. In 81,736 miles of this latter mileage under block-systems, there were 60,167 miles operated under the manual system and 26,569 miles under automatic systems, the electro-motor system being in general use. On January 1, 1908, out of 44,165 miles of line under the manual system, the telephone was in use on 3287 miles in place of the telegraph. On Janu- ary 1, 1914, out of 60,125 miles under that system, the telephone was in use on 26,241 miles. At that date, there were 848 interlocking-machines in use, with 17,951 working levers. Of these machines, 632 were mechani- cal, 122 electric, 48 electro-pneumatic and 46 were electro-mechanical. There were 5884 switches interlocked and 4059 derailers.* In 1890, 98 per cent, of the line mileage in Great Britain and Ireland was equipped with block-signals. In 1915, with about 24,000 miles of line, there were about 310,000 levers in use in the manually- controlled system; including working-points, facing-points, locks and signals.^ > See Appendix VII, Note VII. 2 Relay. — ^An electro-magnet designed to repeat the effects Of an electric current in a second circuit. — Signal Dictionary. Railway Age Gazette, Publishers. s See "First Position Light Signal Installation," C. E. Goings. — Signal En- gineer, March, 191§. , * See Appendix V, Table XVII to XX. ' For progressive development of block-signals and interlocking, see Appendix V, Table XX. 264 EFFICIENT RAILWAY OPERATION Snow-sheds The exigencies of railway operation at high altitudes in an inclement winter climate have led to the development of snow-shed construction on the lines crossing the Continental Divide. Experience with this form of protection on the lines of the Southern Pacific Company is set forth in Appendix V, Table XXI. On the Canadian Pacific Railway, seven miles .of snow-sheds, containing 26,000,000 feet of lumber were built at a cost of $3,000,000. On the Great Northern Railway for ten miles down the west- ern slope at the end of the Cascade Tunnel, 76 per cent, of the distance has been protected at a cost of nearly $1,500,000. These sheds have concrete retaining walls and a timber roof designed for a load of 1500 pounds per square foot. At two points, concrete galleries which are practically double-track tunnels have been built in open side-hill cuts, for protecting approaches to rock tunnels. Watee-stations and Tanks. Track Water Trough Among the structures connected with the Motive Power Department, water-stations come more directly under the control of the Roadway De- • partment. There has been great improvement in these stations since the days when the water-tank was a mere tub held together by light iron hoops and unprotected from the summer sun and winter's frosts, save in very cold climates. Thus exposed to the vicissitudes of the seasons, the staves shrunk apart above the average water-level; the hoops expanded and contracted until the lugs by which they were bolted together gave way; the outlet-valve leaked and water constantly dripped along tjie track from the canvas hose. A hand-pump supplied the leaking tank from a well that frequently went dry, and then information went along the line by word of mouth that no water was to be had at that station. This was not an unusual condition as to the water-supply in the early days of rail- way operation in the United States. With the general improvement which has taken place in railroad service, this situation has been greatly changed. Steam-pumps have supplanted manual labor and horse-power. The sources of water-supply are of a more rehable character, frequently remote from the station and connected with the tank by a main and standpipe. Tanks up to 50,000 gallons' capacity and 30 feet in diameter are still made of staves. The staves are of a uniform width from 6 to 8 inches, 20 feet long and 3 inches thick, with bottom planks 12 inches wide. The staves and bottom are joined by one-inch dowel-pins. The hoops should be of extra tensile strength, as the failure of wooden tanks is princi- pally due to defective hoops. In some cases, the tank rests upon a steel substructure, but it is usually supported upon a frame of closely spaced joists, borne by twelve posts of 12 by 12-inch timber, preferably of ma- terial that has been preservatively treated. ROADWAY . 265 Tanks of 50,000 to 100,000 gallons' capacity are built entirely of steel. At first they were flat-bottomed, but since 1894 they have been built either with conical or hemispherical bottoms, supported by steel columns, attached directly to the sides of the tank, and with a mud-drum in which matter suspended in the water is precipitated. They are provided with inside and outside ladders and a water-gauge, and the other accessories are well thought out and substantially made. Tanks should be covered with a conical roof, having a pitch of 1^ inches to the foot and 14-inch eaves, and covered with composite roofing-material. In designing steel tanks, it is usual to assume the weight of water at 8^ pounds per gallon and 7^ gallons per cubic foot or, more conveniently, at 62^ pounds per cubic foot ; allowing foundation-pressure on dry sand or clay of 3 tons per square foot, and -^inch minimum thickness of cyUn- drical sheets. The progressive demand upon a water-station which accompanies in- creased train-service is illustrated by the experience of the Illinois Central Railroad Company at Centralia. At this station, the demand increased from 72,000,000 gallons in 1895 to 141,000,000 gallons in 1905, and to 280,000,000 gallons in 1914. Efforts to diminish the waste of water re- sulted, in 1915, in a reduction of the demand by 42,000,000 gallons. In 1855, the supply was furnished from a creek by horse-pQwer, for which a steam-pump was substituted in 1858. In 1859, the creek was dammed to form a reservoir. In 1865, two 40,000-gallon tanks were erected. In 1909 in combination with the city authorities, a reservoir of 1,000,000,000 gallons was formed by building a 650-foot dam across a valley, eight miles from the pumping-station, to which the supply was delivered through a 20-inch main of wooden staves.^ To avoid stopping high-speed trains for water, Mr. Ramsbottom de- vised a plan for "picking up" water from long and shallow troughs be- tween the rails, which was introduced upon the London & Northwestern Railway in 1857, and on the New York Central & Hudson River Railroad in 1870. It is now in use on all the trunk lines for express-trains operated at very high speed. The troughs must be placed upon a level piece of track, but not necessarily on a tangent. They are from 1000 to 2000 feet in length, and even 2500 feet for watering two locomotives in a double- header. They are made of sheet-steel, ^ inch thick, 6 to 7 inches deep and 19 to 24 inches wide, with the top of the trough level with the top of the rail. The trough is anchored to the ties about the middle, to allow for expansion. The bottom slopes upward toward the ends, which are protected externally by inclined planes. Water is supplied through a main with several branches. On double- track, the troughs are connected by piping, and, with three Scinch pipes, ' "The History of a Water Station," C. R. Knowles. Railway Age Gazette, June 10, 1916. 266 . EFFICIENT RAILWAY OPERATION a double-tank may be filled in six minutes. The connection to the troughs is made either by rubber-hose or by an expansion-joint to provide for changes of temperature. In severe weather, it is necessary to pipe steam into the troughs to prevent freezing. Drainage should be well cared for. The water is taken into the tender by a hinged scoop which is dropped into the trough by a lever. Signals near the trough indicate the points at which the scoop is to be lowered or raised. The scoop is 12 inches wide with a goose-neck spout, oblong in section, to fill the tender. A stop in the scoop prevents it from touching the bottom of the trough, and water can be taken at a speed of seventy miles an hour. On the Baltimore & Ohio Railroad, between Baltimore and Philadelphia, 92 miles, track- tanks are placed 30 miles from each end.' The Pennsylvania Railroad has every locomotive equipped with a scoop, and slow as well as fast freight- trains and passenger-trains all take water without stopping. Locomotive Houses The building, or " roundhouse," for housing locomotives, should have stalls for at least one-fourth of the number of locomotives normally passing into and out of the terminal yard. There should be a smoke- jack over each stall, with Uberal provision for the exit of smoke and a m9nitor-roof for 'light and ventilation; which diminishes also the cor- rosive effects of coal smoke upon an iron roof. Perhaps a floor of con- crete is preferable to one of brick. A portable gantry crane is useful in a house not fatted with hoists. For housing less than ten locomotives, the simplest arrangement is a set of ladder-tracks entering on a skew into a rectangular shed, and re- quiring neither a turntable nor a transfer-table ; but a greater number must be housed around a turntable, which should not be less than 75 feet in diameter. ^ The angle of the adjacent tracks should be an even factor of 180°, so that they will be in lirfe across the table. The maximum number of stalls in a complete roundhouse depends upoA the arrangement of the tracks ; whether they are laid independently to the edge of the table or intersect between the table and the house. The maximum number of stalls in the former .case would be about thirty, and sixty in the latter. For a greater number, the plan of the house would be two semicircles spaced apart, with a table in the center of each. The distance between the inner and the outer walls of the roundhouse should not be less than 100 feet, to provide space for working around modern locomotives, which vary in length of engine and tender from 78 feet up to 88 feet for an articulated locomotive, with an engine wheel-base of 58 feet. The house should be 1 See "Notes on Track," W. M. Camp, 1903. 2 At Towanda, Pa., on the Lehigh Valley Railroad, seven locomotives are housed in a shed 63 by 183 feet, with tracks 13 feet between centers and 18 feet on a skew of 46°. "Freight Terminals," J. S. Droege, p. 388. ROADWAY 267 lighted through the closed doors as well as through the outer wall. It should be heated by hot blast and ventilated through a monitor roof. The smoke-jacks should be at least 30 inches in diameter, with suction- ventilators and hoods. The minimum clear opening of the doors should be 12 feet in width and 17 feet in height. The locomotives should head toward the outer wall to give space and hght at the front end, where they are most needed. Each stall should be provided with hot and cold water, compressed air and electric current. The pits should be at least 60 feet in length, with a convex bottom sloping from 2^ feet to 3 feet in depth and drained into the turntable-pit, the walls of which should be of con- crete with a wooden coping, 6 inches thick. The circular track and the pivot should rest on concrete foundations. Between the yard-entrance and the turntable there should be access to water-columns, coal-chutes, sand-bins, ash-pits and inspection-pits, with separate tracks for incoming and outgoing locomotives and standing-tracks- for as many of them as there are stalls in the roundhouse. The increasing length of locomotives has, in several instances, necessi- tated the substitution of longer tables and even the reconstruction of roundhouses.^ On this account, it would seem that a house of rectangular plan, in connection with a "Y" track and a transfer-table, would be preferable to a roundhouse and turntable. For an equal length, the trans- fer-table may be of simpler and lighter construction, as the weight may be borne on two or more tracks instead of being concentrated on the center- bearing. Rectangular bays may be of any desired width with less waste- space, and would afford readier access to all parts of the locomotives stand- ing in them. The rectangular roof would be of simpler design, and the building could be extended as additional standing-room was required. If the shop for running-repairs were attached to the rear of such a building, several bays could have independent access to it. Coaling-stations Important coaling-stations are now substantial structures, designed to meet local conditions, with conveyors to elevate coal into pockets, whence it is discharged through chutes, and with accessory sand-bins. Such a station has recently been built of reinforced concrete at Proctor, Minn., on the Duluth, Missabe & Northern Railway. It has storage capacity for 1000 tons, serving a track on each side and receiving coal from a track beneath the storage-room. Sand is contained in a separate concrete ' On the Atchison, Topeka & Santa F6 . Railway it was required to provide for Mallet locomotives measuring 89 feet 2 inches between centers of end-wheels. The 85-foot table was replaced by one 974 feet in length, eounterweighted at the motor-end, at which aU loads entered, in place of a tilting arrangement. This table has handled 51 locomotives in three hours. (Railway Age Gazette, August 18, 1916.) Turntables 105 feet in diameter are in use on the Buffalo, Rochester, and Pittsburgh Railway. 268 EFFICIENT EAILWAY OPERATION structure, outside of a coaling-track. Wet sand is elevated by the coal- conveyor to the top of the tower and descends by a chute into the dry- ing-house. Thence it is returned by compressed air into a bin in the coaling-station.i On the Atchison, Topeka & Santa F6 Railway there are coaling-stations of reinforced concrete, designed as a cylindrical tower, with a shell 9 inches thick and capacity from 250 to 300 tons. The tower is 23 feet outside diameter and 89 feet in height from footing to roof. The coal-bin is in the upper 29 feet, with a concrete bottom and a chute 35 inches in diameter. There is also a sand-drying plant.^ Water-stations are usually spaced from 15 to 20 miles apart, and coaling-stations from 60 to 70 miles. Coal should not be shoveled from a car to a platform and thence into the tender, requiring the labor of from five to ten men. Even where the consumption is small, it would be more profitable to use a coaling-crane and buckets. With increasing consumption, a location should be se- lected where the coal can be imloaded from an elevated track to a plat- form at such a height that it can be dumped from barrows into the tender. Such barrows are usually 6 feet long, 30 inches wide and 30 inches deep, with a capacity of 38 cubic feet. Each barrow-load should be weighed. Better still, the coal may be discharged into pockets and delivered through chutes into the tender. The approach to the elevated track is ordinarily from 400 to 600 feet in length with a rise of 3.5 to 5 per cent. The pockets have a slope of not less than 45° for bituminous coal, or 35° for anthracite. Where a reserve of coal is stored on the ground, it may be reloaded with a steam-shovel. Coal-handling and Coaling-plants In the transshipment of coal, the loaded cars may ascend by inclined planes to the height required for delivery into "pockets." At points of delivery for water-transportation, either on the Great Lakes or on the Atlantic Coast, this operation calls for engineering works of great magni- tude. On the Atlantic seaboard, there are four harbors of importance where coal is discharged directly into shipping, — New York, Philadelphia, Baltimore and Norfolk. In 1913, there were in use thirteen coal-ports in New York harbor with 29 plants; in Philadelphia, three with 11 plants; in Baltimore, four with 8 plants ; and in Norfolk, three with 8 plants ; there were two additional plants under construction at Philadelphia, and two at Norfolk. In that year, about 1,050,000 cars were unloaded at these ports. As the average car-load is about 42 tons, the total tonnage was about 44,000,000 tons. From 1903 to 1911, both included, the ratio of maximum output per i"A Large Reinforced Concrete Coaling Plant." Railway Age Gazette, September 29, 1916. 2 "New Type of Coaling Station." Railway Age Gazette, August 26, 1916. ROADWAY 269 annum at these ports to the estimated capacity of the coaling-plants averaged in New York, 51 per cent. ; in Philadelphia, 29 per cent. ; in Baltimore, 31 per cent. ; and in Norfolk, 30 per cent. At their average estimated capacity, these ports could have discharged the actual output of 1911, in 222 days, at New York; in 185 days, at Philadelphia ; in 195 days, at Baltimore ; and in 141 days, at Norfolk. From this statement, it appears that the then existing plants were capable of supplying a greater demand for coal than was required. The necessity, however, of being in a position to furnish large cargoes in a short time has led to the construction of piers of greater capacity than is usually required. The capacity of many of the piers is restricted by lack of yard- facilities. At several of them, the height is insufficient to admit of a free discharge of coal from the pockets to the vessels. In New York, there is an almost continuous run of cargoes of from 50 to 3500 tons for local con- sumption, which is not nearly as great in other harbors. The total capacity of these several coaling plants in 1911, per day of ten hours, was 2700 cars of anthracite, and 1650 cars of bituminous — total 4350, in New York; 2640 cars in Philadelphia; 3180 bituminous in Nor- folk ; and in Baltimore, 260 cars of anthracite, and 830 cars of bitumi- nous — total 1090; or 11,260 cars at the four ports. The piers with the greatest capacity per day of ten hours were the Pennsylvania Railroad piers in New York, 300 cars, and in Philadelphia, 350 cars ; the Baltimore & Ohio Railroad piers in Baltimore, 300 cars ; the Norfolk & Western Railroad piers at Lambert's Point, Norfolk, 400 cars ; also in Norfolk the Chesapeake & Ohio Railway piers at Newport News, 380 cars, and the Virginian Railway piers at Sewell's Point, 300 cars. The plants under construction for the Norfolk & Western Railway Company and for the Chesapeake & Ohio Railway Company are each to have a maximum capacity of 600 cars. The Pennsylvania Railroad Com- pany constructed, in 1916, at Canton Wharves, Baltimore, a pier 66 feet wide and 942 feet long, in connection with a yard having trackage for 1000 cars and a capacity for loading 20,000 tons daily. At South Amboy, one pier of the Pennsylvania Railroad coaling-plant has a maximum ca- pacity of 140 cars of anthracite or of 300 cars of bituminous coal. In New York harbor, sea-going vessels are usually coaled from lighters. In Philadelphia harbor, the greater portion is in barges for domestic use; though tramps are loaded with from 5000 to 7000 tons. In Baltimore and in Norfolk harbors, tramps and colliers are loaded up to 12,500 tons ; a large part of these cargoes being for New England ports or for foreign countries. Coaling-plants may be classified as gravity-plants, mechanical plants or " combination " plants. In the gravity-plants, coal-trestles are equipped with pockets for receiving the coal and with chutes for discharging it into vessels. Some of them are on grades that permit cars to drift over the 270 EFFICIENT RAILWAY OPERATION deck-tracks and then return empty to the yard by gravity. In others, power is required for these purposes. Mechanical plants are equipped with car-dumping machinery, elevators and conveyors for dehvering the coal without the use of tracks or trestles. The car-dumpers pick up the cars and dimip the coal on an apron whence it flows into adjustable chutes which are kept full to reduce the fall and breakage. In some of the plants, the cars are dumped into conveyor-buckets, which lower the coal to the bottom of the hatch, in order to lessen the breakage. Where the space is restricted, the cars are hfted vertically, and then dumped sidewise into the chutes. Cars may also be hauled up a cable-incline to be turned side- wise into the dump or into a self-propelHng transfer-car having a sloping bottom, which enters a tower that is moved on tracks along the pier, and is then hoisted to the proper height to be emptied into the chute. Where the tracks are at right-angles to the piers, the coal is dumped into a hopper and elevated by endless-chain conveyors to a height sufficient to flow into the pockets. Car-dimiping machinery is combined with gravity-plants in a variety of ways. There are 50 gravity-plants, 7 mechanical plants and 3 combination plants in operation on the Atlantic seaboard. The longest coal-pier is at South Amboy, 1800 feet from bulkhead to end of slip. Other piers have slips from 735 to 1200 feet in length. Two piers under construction at Lambert's Point and at Newport News, in Norfolk harbor, are to have a width of 91^ feet, with a maximima height of 47 feet from the discharging end of the chutes to the water-surface, and a minimum height of five feet. To coal freely into coastwise shipping, the piers ought not to be less than 65 feet above mean tide, when the pier is not provided with storage-bins. For bins storing a car-load of coal each, the height should be 70 feet above mean tide. The tracks leading to the piers and yards should be arranged to require as little shifting of cars as practicable. To sort cars with different kinds of coal, a number of short tracks is desirable ; preferably a gridiron with a ladder at each end. The approach-tracks to locomotive-inclines should have sufficient length to allow a good run for ascending the incline. The approach-tracks to a cable-incline should be down-grade, so that loaded cars may drift down to it, as they are wanted. The grades on locomotive- inclines vary from one to five per cent., according to the length of the pier, the power of the locomotives and the discharging-capacity of the pier. As a general rule, the grades should not exceed three per cent. The grades on cable-inclines may vary from 16 to 18 per cent., and on mechanical dumps from 10 to 12 per cent. The grades on pier-decks should be just sufficient to allow cars to move slowly by gravity, say from 0.5 per cent, to 1.5 per cent. Grades for the return-tracks should not exceed 2 per cent, for hand-brake control. All changes of grade should be eased by vertical curves for a distance of 35 to 50 feet. The delivery-track is sometimes connected with the retupn-track by ROADWAY 271 a switch-back, constructed with adjustable blocking, so that the cars will move by gravity at suitable speeds in either warm or cold weather. Where there is not room for a switch-back, counter-balanced transfer-tables, pivoted at one end, are i^ed. As a car is run on to the table, the table swings in line with the return-track to which the car runs by gravity. The table is then returned to its normal position by a counter-weight. The equipment connected with a gravity-pier consists of bins for stor- ing coal, chutes and pockets for delivering coal directly from the cars to the chutes, track-scales and thawing-plants. The bins are built beneath the approach-trestle and are so constructed that the coal is dropped from the car-bottoms through openings in the deck, and may be discharged through hopper-bottoms into other cars beneath or by chutes into barges. Pockets are placed under the deUvery-tracks on the pier with their bottom- outlet at the side of the pier. For the pocket to receive the full discharge of coal from either a hopper-car or a gondola-car, its top should be at least 12 feet long and 9 feet wide, with the sides and bottom tapering toward the opening at the side of the pier. There are several types of chutes, either simply hiaged or adjustable at either end, or with a telescopic leg at the lower end ; but the simple hinged chute is preferable for its simplicity, speed of delivery and low cost of installation and maintenance. This, however, depends also upon commercial conditions ; as the cost of trimming in the ship's hold and the damage to the coal from excessive breakage, may offset these advantages to a large extent. It is, after all, a question of the quick release of vessels of large daily tonnage and a consequent reduction of expense in operation. Track-scales are connected with coaling-plants where required by local conditions. They should be so located as to require the least shifting of cars and be convenient for observation, with length suitable for the stand- ard length of cars, their moving speed while being weighed, and as to whether they are weighed singly or while coupled. They are from 34 to 64 feet in length, if fitted with automatic weighing devices, or from 32 to 60 feet, if without them. On the new piers in Norfolk harbor, the automatic scales are 68 feet long, with weighing capacity from 160,000 to 300,000 pounds. There are several plans for thawing frozen coal, either by inserting steam-jets from a locomotive or from a stationary boUer, or under cover by hot air; but where thawing is only occasionally required, it is cus- tomary to break the frozen coal by hand.' Many of the appliances here described for the transshipment of coal from rail to water, originated at ports on the Great Lakes. In their development, remarkable ingenuity has been displayed in the substitu- tion of machinery for manual labor. Their general use brought about the construction of coal-carrying ships of large tonnage, with hatches and '"Coal Piers on the Atlantic Seaboard," J. E. Greiner. Trans. Am. Soc. C. E., December, 1914. 272 EFFICIENT RAILWAY OPERATION holds specially planned for the receipt and discharge of cargoes by these appHances. The effect of these various improvements has been shown by ' the reduction of Lake rates from 90 cents per ton of 2000 pounds, in 1887, to 30 cents in 1912. The car-dumper, introduced in 1892, has been so developed that it will now seize a 50-ton loaded gondola-car, weighing (gross) 140,000 pounds, in- close it in a cage, raise it on high, swing it out over the dock and capsize its contents into a vessel's hatches at the rate of forty car-loads an hour. The breakage of coal in' this process is lessened by its delivery through a tele- scopic chute extending to the bottom of the hold, which is gradually with- drawn as the cargo accumulates. In the seven months of open navigation, it was necessary to accumulate at Lake Superior ports, the fuel required in the Northwest for the ensuing winter season. Here, again, American engineering genius devised the means for accomplishing this duty, and, for this purpose, large expendi- tures have been made by the railroad companies, with more regard to social efficiency than to direct profit. The means devised with this object in view, are the grab-bucket to unload coal from vessels, and the bridge- conveyor for distributing it in a spacious storage-area. The conveyors are built to cover an over-all length of 200 feet, and are equipped with a seven-ton self-righting tub for distributing the coal, and a four-ton shovel-bucket for reloading it into cars ; each handled by one man and all operated electrically. Intermediately, each load is auto- matically registered. Some of these plants can unload a cargo of 5000 tons in ten hours, and distribute it over an area 275 feet wide and 900 feet long, with a storage capacity of 150,000 tons. In 1906, a covered anthra- cite-plant was installed, in which 200,000 tons are piled 60 feet in height. The shipment of coal from lake ports into the Northwest is largely in re- turning grain-cars. A special loading-appliance has been devised for loading such cars directly from the storage piles.^ Qkain Handling The Atchison, Topeka & Santa Fe Railway Company has an elevator at Chicago with facilities for receiving and delivering grain at the rate of 15,000 bushels an hour. The plant consists of an elevator-house, a storage- annex, a drying and bleaching equipment and a marine tower for delivery to shipping ; the whole being operated by 1500 horse-power. The elevator house is 225 feet by 83 feet and is 170 feet to the top of the cupola. It has a timber-frame closed in by brick-walls 80 feet to the top of bins of 400,000 bushels' capacity. The cupola above is covered with corrugated steel. The bins are raised 40 feet above the ground floor, on which is placed the machinery for handUng grain from five receiving-tracks. Eighty cars are ' See "The Handling of Coal at the Head of the Great Lakes," G. H. Hutchin- son. — Journal Am. Soc. Mechanical Engineers, August, 1914. ROADWAY 273 set on these tracks at one switching operation, and these cars are placed in position by a car-puller, twenty cars at a time. The grain is discharged into hoppers beneath the tracks ; each containing a full car-load. Each group of four hoppers discharges into a conveyor for delivery to the re- ceiving-plant beneath the bins ; here it is elevated into the cupola, where it is weighed and dropped into the bins ; or spouted into cars for transship- ment ; or delivered to the marine tower ; or conveyed to the annex. The annex has a capacity of 1,000,000 bushels, stored in 35 concrete-bins, each 23 feet in diameter, and in the interspaces. At Victoria Harbor, Georgian Bay, grain is transferred from lake vessels to the Canadian Pacific Railway by means of two movable towers of steel, traversing the dock-front on a double-track. Each tower, 170 feet high, is carried on ten freight-car trucks and is moved by a cable- pull. There is storage-capacity for 2,000,000 bushels in 32 concrete-bins, 35 feet in diameter, with storage also in the interspaces. This plant is operated by electricity and has a working-capacity of 20,000 bushels an hour.i Ore Handling The shipment oT ore from ports on Lake Superior to Lake Erie ports has reached 50,000,000 tons in seven months of open navigation. The appliances originally devised for handling coal have been modified for handling this vast tonnage. At one of the ports on Lake Erie, there is a storage-yard covering 50 acres, with a concrete dock-wall 800 feet long, furnishing space for 1,000,000 tons of ore. Four unloaders, each carrying a 17-ton bucket, span four tracks, and a cantilever-extension enables them to discharge ore 110 feet inward from the dock. The outreach of the bucket-arm is sufficient for the bucket to be lowered into the hold of a vessel 65 feet from the face of the dock. The bucket-shells, when open, have an expansion of 21 feet 3 inches and a telescopic motion of three feet. It takes 50 seconds for a complete movement, which is equivalent to a capacity of 1000 tons per hour for each of the four unloaders. Ten- thousand-ton cargoes have been unloaded in less than three hours. The ore is distributed over the storage-space by a traveling-gantry with a main span of 266 feet and a cantilever at each end of 170 feet. The ore is grabbed from the vessel's hold in the unloading-bucket, elevated by the rocking-arm and discharged into a 60-ton hopper, from which it passes into a 50-ton scale-car, and then into cars or into a temporary storage-pile, from which the bridge-conveyor transfers it to the main storage-space.^ • "Freight Terminals," Droege, pp. 269, 272. 2 An interesting use is made of electro-magnetism in connection with handling pig-iron. At the dock of the Buffalo, Rochester & Pittsburgh Railway in Buffalo, a locomotive-crane mounted on standard freight-trucks is equipped with an electro- magnet with a carrying load of 3250 pounds. The boom is over 40 feet in length, with capacity for a load of 31,500 pounds and 24 per cent, overload in a radius of 13 feet, or a load of 6500 pounds with 23 per cent, overload in a radius of 44 feet. 1 274 EFFICIENT RAILWAY OPERATION Dock and Harbor Facilities. Miscellaneous Freight The proper alignment of yard-tracks with refer'ence to the bulkhead- line of a dock, depends upon the surrounding conditions, as well as upon the character of the commodities for transshipment. Where the harbor is spacious and the access to it is level and free from encumbrances, the piers may project at right-angles to the bulkhead line, the switching-tracks approaching perpendicularly ; though with long trains and several piers, such a track-plan may interfere with the general arrangement of the terminal yard, besides requiring sharp curvature in the leads to the outer switching-tracks. Where the harbor is contracted, as in the channel of a river, and it is practicable to locate the approach to the harbor on an alignment approximately parallel to the bulkhead-line, sharp curves in the switching-tracks are avoided by projecting the piers at an acute angle to the bulkhead-line, on what is known as the "saw-tooth" plan. On this plan, the tracks switch on to the piers from the same side of the approach- track, and all with the same degree of curvature. The approach-track may be continued indefinitely to serve any number of additional piers, and the return-track can be parallel to it with cross-overs. The side of the yard away from the piers is then favorably disposed for a connecting grid- iron, or for general switching purposes. Such an arrangement is suitable for shipping lumber in large quantities, as the lumber can be stacked more conveniently for loading through the bO;W-ports of a vessel. In a narrow river, the slips should open toward the entrance, which makes it easier for a vessel to enter or leave them without blocking the main channel. The Dominion of Canada has undertaken the establishment of a harbor- terminal at Halifax in connection with its system of transcontinental railways. The estimated outlay of $30,000,000 includes the construction of an entirely hew railway-entrance to the city and six miles of approach- tracks, with a passenger-station adjacent to a quay 2000 feet in length, and six piers, each 1250 feet in length in 45 feet of water, providing berths for twenty to thirty pf the largest steamships afloat ; also a protective break-water and all the equipment required for economical transshipment between railway and steamship. Covered piers are required for the transshipment of miscellaneous and package freight. The deck of such a pier should be level with the car- floors for convenient trucking. Where there is a considerable rise of tide, this requires that slip-ways should be cut in the side of the pier to suit the side-ports of a steamship and to avoid hoisting through the hatchways. At low tide, the ascent of the slip-way is a serious draft upon the strength of a truck-hand. This diflaculty is overcome by the introduction of the moving ramp or "escalator," now in common use for passenger-service between different levels. As modified for this purpose, it consists of an endless steel chain, revolving about sprocket-wheels that are driven by a ROADWAY 275 motor. It can be started, stopped or reversed at any point during its travel. The truck-hand brings his loaded truck to the foot of the ramp, whence man and truck are transported from the ship's deck to the pier- deck, quickly and without physical effort. It is asserted that the "single- file" machine has a capacity of 600 trucks per hour, and that the "duplex" machine will double that duty, with the ascending and descending ma- chines both in operation. On a pier now under construction in Norfolk harbor, there will be 35 of these machines in service. The transshipment of miscellaneous freight at New Orleans has been recently facilitated by the construction of a terminal yard on the bank of the Mississippi River by the State of Louisiana at a cost of $3,500,000. The terminal covers 150 acres, of which 100 acres are occupied by 22 miles of tracks. Twenty-three acres are under roof, with storage-capacity for 2,000,000 bales of cotton, and other provision for storing sugar, rice, to- bacco, coffee, corn and wheat. It is equipped with 4^ miles of overhead and floor-level runways and 50 miles of runways for traveling-cranes; also a cotton-compress with a ten-hour capacity of 1000 bales and three others with a total capacity of 1400 bales. The wharf is 120 feet wide on the first-floor and 100 feet on the second, with a total length of 2000 feet. It is of reinforced concrete and steel, on a foundation of nearly 10,000 creosoted piles, in clusters 20 feet between centers, filled in with 2,000,000 cubic yards of river-sand. Along its full front, there is an apron-wharf carrying two tracks, and depressed tracks in the warehouses permit of unloading freight from cars within reach of the ship's tackle. The ware- house-sheds are roofed in spans of 35 and 45 feet. Motors, travehng on runways, transfer articles between the compartments and to the wharf- front. There are also large receiving and sorting yards. The reduction effected in the cost of transportation is estimated at 40 per cent. Freight Sorting Yards. Gravity Switching Yards The increasing production of commodities consequent upon industrial development in the United States caused a gradual tendency to congestion of traffic upon the principal railway thoroughfares, which called for im- provement in methods of concentration and distribution of the freight- equipment employed in such service. It became imperative that these operations should be carried on elsewhere than on the sidings required for the normal movement of trains, and certain principles became recognized in the construction of switching-yards, or sorting-yards. It is obvious that the breaking-up and the making-up of trains should not be undertaken on the rurmjng-tracks nor by crossing from one side of such tracks to the other ; but that operations of this kind should be carried on apart, and away from interference with regular train-service. With this separation of trackage, more efficient track-plans were devised 276 EFFICIENT RAILWAY OPERATION for the distribution of cars as a train is broken up, and for their concen- tration in other trains according to their respective destinations. A series of sidings was arranged in the form of a parallelogram, and known as a "gridiron," on which arriving trains were placed on one set of tracks and the cars in them were distributed on another set, in trains ready for de- parture. The approach and return tracks, connected with these parallel tracks at their ends, are known as "ladders." The usual arrangement is a receiving-yard with long tracks, a classification-yard and a departure- yard, with accessory sidings for caboose-cars, for crippled cars, for live-stock and for icing perishables; with conveniently situated roundhouses and repair-shops. The forwarding of cars in transit was greatly facilitated by such track- arrangements, but there was a loss of time in the double movement of the switching-locomotives; first in pushing each car, or cut of cars, into its appropriate track and then in backing out with the remainder of the train as yet undistributed. The unnecessary movement of this part of the train was prevented by the introduction of the "poling-track," parallel with the "body-track," on which the train is placed for distribution. The switch- ing-locomotive has a pole attached to the breast-beam, which is placed in contact with a "poling-pocket" on the corner of the last car in the cut. Each cut is then pushed into its proper track without disturbing the re- mainder of the train. A safer method has been provided by the interven- tion, ahead of the locomotive, of a flat car specially equipped with four poles ; two on each side that can be used in opposite directions. ^ On the poling plan, the switcher works close to the front of the train, where the clearance of each track can be more distinctly seen and signals received' at shorter range than at the rear of a long train, which is an ad- vantage in thick weather. The operation of a poling-yard is thought to be severe on equipment, on account of the necessarily quick and heavy starting and reversing movements. Where the tracks could be laid on a descending grade, this objection has been obviated; for the train could be gradually dropped ahead by the brakes, as each cut was removed. A grade of at least 0.4 per cent, was required and, in very cold weather, even an 0.8 per cent, grade was not found objectionable.^ The advantage of distributing cars on a descending-grade was further utilized in their classification solely by gravity. The SAvitching-locomotive was replaced at the rear of the train and pushed it to the summit of the grade by successive movements, as each cut of cars was released ; so that there was no unnecessary reverse-movement of either the train or the locomotive. In the approach to the freight-terminals of the London & Northwestern Railway at Liverpool, there was an inclined plane at Edge • In such a yard for distributing coal-trains, 1428 loads were passed through in 10 hours, 20 minutes; an average of 138 cars per hour. "Freight Terminals," J. A. Droege, p. 62. ROADWAY 277 Hill belonging to the early period of railway construction, upon which the custom originated of cutting each coal-wagon loose at the top of the plane and "riding" it down to the foot and into a siding; its speed being con- trolled by a hand-brake or by a skid. At Edge Hill, in 1873, a gravity switching-yard was laid out on the gridiron plan, and the "push-and-puU" process was abandoned. From 2000 to 2500 wagons are collected here daily from the several receiving- stations in Liverpool and are sorted on a group of twenty-four storage- sidings, from which they are re-distributed on a set of shorter tracks in station-order for departure. There are but few places where the oppor- tunity is afforded, as at Edge Hill, for sorting by gravity from a higher ' to a lower level. One was provided near Dresden, as early as 1846, by making a fill at the upper end, seventy feet high. This yard is 1^ miles long and half a mile wide. There is another near St. Etienne, on the Paris, Lyons & Mediterranean Railway, dating from 1863. On the Pennsylvania Railroad lines, there are such gravity-yards at Greenville, N. J., Sheridan, Pa. and Logansport, -Ind. ; and on the Norfolk & Western Railway at Bluefield, W. Va.^ The obvious merits of gravity-switching were limited by local conditions until its fundamental principle was artificially reproduced in the "summit" yard, by the intervention of a double inclined-plane — a "saddle-back" or " hiunp " in railway parlance — between the reception-tracks and the classi- fication-tracks. Over this hump, the trains were pushed and the cars distributed, as in the gravity-yards operated under natural conditions. The first simamit-yard in the United States was constructed in March, 1883, on the Pennsylvania Railroad, near Greensburg, Pa. In the United States, in 1910, out of 510 classification-yards, 82 were himip-yards. The most extensive yard is at Gardenville, N. Y., on the New York Central Railroad for traffic in the Buffalo territory, which has capacity for 21,000 cars.^ A sorting-yard should be carefully designed with reference to the character and volume of the traffic to be handled there, for it is difficult to remedy mistakes resulting in unnecessary switching. By comparatively slight changes in the track-arrangement, the service of a yard-locomotive and crew may be dispensed with at an accompanying saving of about $14,000 per annum, equivalent to the interest at five per cent, on over $275,000. The receiving-yard should lead directly into the classification- yard and should have track-capacity for the trains arriving during one hour of maximum traffic. The tracks should each accommodate the longest train and none of them should be less than half the length of that train. 1 For gravity-yards on Norfolk & Western Railway, see Appendix V, Table XXII, and for sorting-yards in United States and in Continental Europe, see Table XXIII. 2 See Appendix V, Table XXIV. For statistics of principal gravity-yards in the United States, see "Freight Terminals," Droege, pp. 74, 112. 278 EFFICIENT RAILWAY OPERATION The arrangement of tracks in the classification-yard should conform to the purpose of classification. For classification by districts, the cars are to be arranged in station-order for delivery. For arrangement as to com- modities, a further classification is required on other tracks. The capacity of the yard is determined by the average number of cars to be forwarded daily to the several districts, with allowance for emergencies. The min- imum capacity of any hiunp-yard in this country is stated at ten trains of 65 cars each in 24 hdurs, with maximum hourly arrival of three trains, and provision for classification at the rate of eighty cuts per hour. Pref- erably, there should not be more than twenty tracks in one set. A greater number requires longer ladder-tracks and correspondingly unnecessary length of switching-movements. By the use of a "V" ladder, with leads to a gridiron on each side of it, a maximiun number of 36 tracks may be conveniently served. A departure-yard should be so arranged that cars from the classification-tracks can be moved directly into its storage-tracks ready for the roads. Where entire trains are held for orders, separate tracks should be provided for their storage, so as not to interfere with the switching-service. Body-tracks, for holding trains, should be spaced 11-|. to 13 feet between centers and, if possible, without curves. At intervals of five tracks, there should be a wider space to provide for drainage and for piling track-material. Ladder-tracks should be 15 feet between centers, to give room for throwing switches ; and no yard-frog should be of a less angle than the No. 7 frog of 8° 10'. There should also be a clear space of at least 15 feet between the yard-tracks and the running-tracks to afford room for signal-posts and for water-columns. The track-scales should be in the lead to the classification-tracks, with "dead" rails for their protection. Icing-tracks should be placed between the receiving and the classification yards. Tracks for caboose-cars should be situated where they can be readily shifted from an arriving-train, and dropped to the rear of a train departing in the -op- posite direction. "Bad-order" tracks should be so connected with the receiving-yard that cars may be placed on them as they are being classified . These tracks should also be ready of access to the repair-tracks. The repair-tracks should be in two sections, with switches locked at the end of each section. As one section is filled, the switch-key should be held by the foreman of car-repairs until the work on the cars in that section has been completed and the work- men have been transferred to the other section. Repair-tracks should be of a length sufficient to hold fifteen cars with space between the cars for working upon them. The tracks should be in pairs, 16 feet between centers and 25 feet between each pair for handhng materials. A narrow-gauge track in this space should be connected by a turntable with a cross-track to the material-yard and the supply-stores. A track should be reserved for Ught repairs and wheel-renewals to fast-freight and stock trains. ROADWAY 279 Approach-tracks should be provided at the entrance to a freight-yard with an interlocking-plant and telephone-connection, in order that arriving- trains may be removed promptly from the main line. There should be a similar arrangement at the outlet of the yard for departing-trains. Their departure will be faciUtated by testing the brakes before the road-locomo- tive is attached. For this purpose, a Hne of piping should be run across the yard from an air-compressor, with a hose-connection between each pair of tracks. The yard-master's office should be estabhshed at a central point and at a height giving a good view of the yard. There should be a building adjacent, with rooms for the yardmen with wire-screened lockers for their clothing and, in some comparatively quiet spot, a rest room for the road-crews. With a natural inchne, all trains must enter the yard at the upper level. In the case of trains from the other direction, this may require a useless journey of several miles, which can be avoided in the hump-yard by having a hmnp for each direction. The gradient on the descent should be gradu- ally decreased so that the cars will stop in the classification-yard. The practice in Europe is a gradient of 20 to 25 feet per 1000, or of 2 per cent, to 2.5 per cent, from the top of the hump, changing to 15 feet per 1000 at about- 150 feet from the entrance to the sorting-tracks, which have a gradient of five feet per 1000. In the Un ited States, with cars of heavier loading-capacity, the initial grade at the svimmit varies from one per cent., with a fall of 0.7 feet in 70, to a 3.9 per cent, grade with a fall of 5.8 feet in 150. In the former case, there is a further fall of 5.25 feet in 300, or of 1.75 per cent., to the bottom of the ladder, with a velocity at the foot of the initial grade of 4 miles an hour, and, at the bottom of the ladder, of 14.6 miles an hour. In the latter case, the further fall is 13.8 feet in 1500, or 0.92 per cent., with a velocity at the foot of the initial grade of 12.2 miles an hour and of 27 miles an hour at the bottom of the ladder. In the Chicago Clearing Yard, the ascent and descent of the gravity-lead is accompHshed in about 4500 feet. The ascent is on a 1.25 per cent, grade for 1800 feet to an elevation of 22 feet. The descent of 2700 feet is for a short distance on a one per cent, grade and thereafter 0.9 per cent. The velocity at the foot of the initial grade is 10.9 miles an hour and at 2500 feet it is 28.5 miles an hour in the summer.* The wide variation of practice in this important niatter results from difference in local conditions and in the character of the traffic. The effect of the weather is also distinctly perceptible in accelerating or retard- ing the descent of cars on the incline. At Altoona, Pa., in the severe winter of 1903-1904, cars did not run to the end of the yard, although the initial grade from the summit was 2.55 per cent, and from its foot there was a grade of 0.7 per cent., well into the yard. The simimit was raised on ' For additional statistics on this subject, see "Freight Terminals," Droege, p. 74. 280 EFFICIENT RAILWAY OPERATION crib-work to a grade of four per cent, for 200 feet to meet this emergency. A mechanical hump was devised on the Pennsylvania Lines for scale- tracks, which is speedily adjustable by jack-screws through a range of five inches, with a corresponding range in the gradient from one to four per cent. Gravity-yards have been still further developed with interlocking- mechanism controlled from a cabin at the summit, in which the successive movements of the switches are indicated on a miniature-plan of the tracks, which is electrically connected with the interlocking-apparatus. The Chicago Clearing Yard was thus constructed, some fourteen years ago, at a cost of $8,000,000. It is now operated as a union freight-terminal by twelve railroad companies, and serves physically the same purpose with the traffic that a railway clearing-house does in collecting and distributing the revenues accruing from that traffic.^ Experience in lighting freight-yards for night-work has led to the sub- stitution of pole-lights for tower-lights. The high-voltage, mercury- vapor lamp on a tower 90 feet high, is an expensive installation, which can not be as readily changed to suit changes in the lay-out of a yard, as can be done with lighting-units on poles 35 feet high and from 100 to 200 feet apart, with spot-Ughts for switches on the ladder-tracks. Where practicable, yards for traffic in opposite directions should not be located beside each other, but should be connected at their ends and on opposite sides of the running-tracks, and there should be repair-shops and roundhouses near their junction; also a separate entrance for the yard- locomotives to the coal and water supplies, to avoid delay to them from road-locomotives standing on the roundhouse tracks.^ 1 An electro-pneumatic mechanism was installed on the Chesapeake & Ohio Railway at Russell, Ky., at a cost of $14,000. It operates 21 switches, with an increase in the daily average from 851 to 1139 cars and a saving of $645 in the monthly cost of operation. 2 Notes on Gbavity-tards In a gravity-yard used largely for mineral-traffic, the maximum number of trains arriving in one hour was seven, and in twenty-four hours, thirty-three. Average number of cars in a train fifty ; and over the hump, sixty-eight cars with 3800 tons of load. Hump in actual operation, 15i hours per day. Switches operated by compressed air. Classifying capacity, as follows : Hours Eabtbound Cuts Cars Westbound Cuts Cabs 1 5 10 24 81 291 580 900 1,318 2,151 108 389 688 1,205 1,396 2,488 ROADWAY 281 Dblivehy-tards The importance of having delivery-yards well planned and constructed is shown by the magnitude of some of them. The team-dehvery yard in Providence, R. I., occupies 21.8 acres and has capacity for 675 cars, with 458 accessible to teams.^ Tracks for team-dehvery should not hold over fifteen cars each. They should be stub-tracks, ending against the main drive-way and laid in pairs, 12 feet between centers, with a 50-foot road- way between each pair, to give room for teams to back up to the car-doors. Experience has shown that twenty-five cars for team-delivery will occupy about an acre. The driveways should be well paved and drained, and accessible to wagon-scales. Track-scales should be on or near the lead into. the yard, and also cranes for handling heavy articles. Where there is much of such traffic, the crane should be operated by power, and cover several tracks with a bridge. An incline at the end of one of the team-tracks is convenient for loading vehicles on flat-cars. The planning for handling hve-stock ■ on an extensive scale requires specific treatment to meet local conditions. Pens for loading and unload- ing in a small way, are usually of rude construction. A more important matter, from the roadway point of view, is the arrangement for detraining and reloading stock in transit. For this purpose it is advisable that the pens should be ranged along a siding of sufficient length to hold a complete train ; as the shock in switching it in parts, causes injury to the animals and is a foundation for annoying damage-claims. There should be a runway the full length of the pens, fenced on the track side, and^portable chutes for access to the upper deck of double-deck cars. Passenger-tahd Tracks A different arrangement of tracks is required for the temporary storage of passenger-cars while they are being prepared for re-assembling in trains. The "coach-yard," as it is called, should be conveniently adjacent to the train-shed, so that cars may be speedily taken from, or added to, trains as may be required. There should be no dead-end tracks ; it is necessary, also, to have either a turntable or a "Y" track for turning certain cars, A series of switching-tests was made by Mr. C. L. Bardo, with a train of 60 cars in 50 cuts with the following results : Mode op Switching Tail Pole Summit Time consumed .... Distance traveled . . . 2 hours 24,750 ft. 1\ hours 24,750 ft. Half an hour 6,000 feet 1 "Freight Terminals," Droege, p. 183. 282 EFFICIENT RAILWAY OPERATION drop-pits for inspecting trucks, tracks for making up trains and other tracks for storing surplus equipment. The tracks for making up trains should be long enough to stand two trains, say about 1050 feet. The wash-tracks should be spaced alternately 16 feet and 25 feet between centers, with platforms on a level with the rails. The wider platforms are covered by an umbrella-shed extended over the lower roofs of the cars. The storage-tracks need only be 14 feet between centers. The tracks for the wheel-pits should hold five cars, with 25 feet between track-centers. The wash-tracks should be piped for compressed- air, steam, water and gas, with plugs every 100 feet, and with connections for electric current. There should be a carpet-shed with platforms 15 feet by 75 feet and four feet above the ground, with a vacuum cleaning appara- tus. Connected with the yard, there should be provided store-rooms for Pullman bedding and linen, for carpets and curtains, for plumbers', painters', carpenters' and air-brake materials, and for commissary-stores and ac- companying offices, also for ice, oil and waste. ^ The arrangements for the- reception and delivery of ' miscellaneous freight at a local station vary materially from those for handling cars in trains at a transfer-point. Such arrangements also differ with the volume and character of the principal business done at a station. There must be tracks with access from the street to transfer commodities directly be- tween the cars and wagons; facilities for handling, feeding and watering live-stock, and tracks available for switching without interference with the running-tracks. Industries dealing in bulk-freight gather around such stations and demaftd tracks for direct transfer. At important stations, overhead cranes usually span one or more tracks and a driveway. The bridge may be fixed, or move on an elevated track, or on towers moving on surface-tracks. Traveling electric cranes are in use of 40 to 50 tons' capacity, with auxiliary hoists of 5 to 7^ tons' capacity for light weights. The location and plan of the freight-warehouse should accord with the neighboring streets and will probably dominate the general lay-out of the yard-tracks. Freight Houses. Freight Handling. Warehouses Freight-houses were formerly built with a receiving and a delivery platform under the same roof and two or more tracks between them. Now, there are separate buildings for inbound and for outbound freight with several tracks between them and platforms on the outer sides accessible 1 The coach-yard at the Union Station, St. Louis, has a capacity for 665 cars, allowing 80 feet to each car, and is to be enlarged to a capacity of 1071 cars. The Sunny-side yard at Long Island City covers a tract two miles long and 1500 feet wide, with capacity for 340 Pullman cars, 375 coaches and 700 motor-ears. The ladder- tracks are made continuous by loops. During 24 hours, the maximum number of cars handled at this yard has reached 656, with 450 ears on the tracks at one time. "Passenger Terminals," p. 241. ROADWAY 283 for wagons. The spacing for the doors on the track-side should be ad- justed for cars of standard length when coupled together, and cars standing on outer tracks parallel with and close to the house-track can be reached from the house directly through the doors of cars on the intermediate tracks. Platforms for street-delivery should be of ample width for the temporary storage of goods. The platform on the house-track and those between the tracks for transferring directly from one car to another should be wide enough for loaded trucks to pass each other readily and all plat- forms should have roof-protection. The platform-floors should be not less than seven feet from the center of a siding and there should be a clearance of 16 feet from rail to eaves of the building. The house for outbound freight should not be over 30 feet in width, in order to diminish the truck-haul between the receiving-door and the car-door. The scales should be placed near the receiving-doors with the scale-beam against thfe wall. The house for inbound freight should be wider, as more storage-room is required. A slight slope in the house-floor in the direction of the trucking, facilitates the work. The door-j ambs should be protected and the edge of the team-platform by a timber wheel-stop. Part of the house for inbound freight is usually set apart for handling small lots in transit, but at transfer-stations this requires separate accommoda- tions. At one transfer-station, 32 cars are brought to the platform on each trip, and their contents are distributed into 135 cars standing on nine tracks. The transfer-platform at Waverly, N. J., on the Pennsylvania Railroad is 1000 feet in length, with placing-room for 196 cars and an average outward movement of 230 car-loads daily. The elevation of tracks entering larg^ cities has led to the construction of freight-houses with tracks on the second floor. The desire to utilize the valuable area thus occupied, has likewise induced the addition of several stories for storage-purposes, with- special regard to economy and dispatch in moving freight between the different floors and to its storage and preservation from loss and damage. These houses are built in fire- proof compartments with automatic sprinklers for extinguishing fires. The floors are of creosoted yellow-pine resting on sleepers imbedded in sand, tamped-clay or ballast and, for protection in trucking, are covered with tongued-and-grooved maple flooring, 8-inch by ^inch. The reduction of labor with hand-trucks is an important consideration, apart from the lessening of contingent damage to freight. The cost of transferring a car-load of miscellaneous goods from the car-door to the delivery-platform may, on an average, equal the ton-rate for 200 miles; and it has been estimated that the average haul in trucking increases 27.8 feet for each additional hundred feet in the length of the house. Much attention is being given to shortening this haul and to accelerating the movement by the use of capacious trucks operated by electricity; or by such mechanical appliances as horizontal conveyors (either roller, chain or 284 EFFICIENT RAILWAY OPERATION platform), or by hydraulic or electric elevators, with a speed of 20 feet per minute. Overhead cranes are also used in warehouses and, in some instances, the telferage system of steel runways, on which are operated electric travelers, with hoists. These overhead apphances require a con- siderable addition to the ordinary height of walls, to clear the work on the floor, which adds materially to the cost of the building. A freight-warehouse recently completed for the Lehigh Valley Railroad at Buffalo, is 577 feet long and 60 feet wide. It is built of concrete and is divided into three sections by fire-walls with automatic steel doors, and with roof-trusses spanning the entire building. The interior is lighted through continuous wire-glass windows over the doors, and artificially by three 200-watt lights in each section. There are nine scales in the freight- house, each of 3500 pounds' capacity. On the track-side there is a con- crete platform, 10 feet wide, with a canopy roof. The station is provided with a 40-ton electric crane, operating through a longitudinal distance of 148 feet and capable of serving four cars at a time.' Station Accessories. Scales On European roads, a clearance-frame for loaded open cars and a crane operated by hand are usually to be found at every station ; and the use of such station-appliances is becoming more general in the United States. This is also true as to track-scales which came into use about 1850. The importance of having such scales substantially constructed, and the necessity for careful maintenance, in order to preserve their accuracy, is evinced by the attention that has been given to these matters by the American Railway Association. Its rules on the subject were revised at great length June 26, 1916. They include the selection, installation and location of scales, and their maintenance and use. The requirements for track-scales relate to their capacity and length, to cover the maximum capacity and wheel-base of the heaviest loaded cars, and their suitability for weighing cars either at rest or in motion. For verification, a scale should be capable of adjustment to within half a pound per 1000 pounds, when new, and should be considered as inaccurate when it can not be adjusted to within two pounds per 1000 pounds. When cars are to be weighed in motion, the speed should not exceed four miles an hour ; each car to be entirely alone on the scale for three seconds. There should not be less than 50 feet of tangent-track at each end of the scale-rails. The foundation should either be of concrete or of cut-stone laid in concrete, and the walls should be extended to sup- port the approaches for a full rail-length at each end of the scale-pit. An eflicient transfer-rail should be introduced to prevent the impact of cars at the joint between the approaches and the scale-rails. Both of these 1 "New Lehigh Valley Railroad Station at Buffalo." — Railway Age Gazette, Sept. 15, 1916. ROADWAY 285 rails should be securely anchored to prevent creeping, and should be kept in proper line and surface. The scale-pits should be water-proof and be drained into a cistern. They should be properly lighted for cleaning and testing. Locomotives that are not to be weighed, should be passed over the dead- rails. All scale-pits should be cleaned at least twice a month and tested every three months, preferably by a test-car or by test-weights up to 10 per cent, of their capacity. The test-car should weigh between 30,000 and 60,000 pounds for general tests. The United States Bureau of Standards has provided traveling-equipment for testing master-scales which, by ar- rangement with the American Railway Association, is moved over the entire railway system annually in accordance with a pre-determined time- table. The itinerary for 1916 provided for tests at nineteen places. Freight-cars should be stenciled at intervals of time, in order to test the correctness of the weights originally stenciled upon them. The loss of weight, after some years of service, may seriously affect the correct billing of bulk-freight. In a lot of 1500 box-cars of 80,000 pounds' capac- ity, that had been in service about two years, a test of one hundred, made at random, showed that these cars were actually carrying, on an average, about 1500 pounds per car more than the billing called for. Such loss in weight occurs also in steel cars. In a lot of 1000 cars of 100,000 pounds' capacity, one-haK of which had been in service for about two years, the average depreciation in weight was from 900 to 1000 pounds. There is a further loss in billing weights in drop-bottom cars from neglect to clean them out thoroughly. Under weights in billing from this cause have been found to amount to between 500 and 1000 pounds.^ Design of Station Buildings In planning a freight-station, the primary consideration is economic efficiency. In planning a passenger station, the primary consideration should be social efficiency. The provisions to meet this requirement must, of course, be suited to the extent and character of the traffic. At flag- stations, a mere landing-place is frequently the only provision, yet this alone should not be deemed sufficient. There should, at least, be a shelter, walled in on the back and sides, with a bench against the wall. At regular stations, the accommodations must vary with local condi- tions. At minor stations, it is practicable to house both passenger and freight business in the same building, with less expensive service and with better supervision than if they were conducted separately. A building 75 by 25 feet, with broad eaves will answer this purpose. It should be divided transversely ; one-half for the freight-room and one-third for the waiting-room, with the office and baggage-room between them. The office should extend out on the platform, for giving train-orders, for displaying 1 Association of Transportation and Accounting Officers. Minutes of Meet- ing, No. 11, June, 1909. Page 1526. 286 EFFICIENT RAILWAY OPERATION signals and for a better view of the tracks, with a ticket-window in the waiting-room wall. The two lavatories should be on the end-wall of the waiting-room. At such stations, it is advisable to provide dwelling-ac- commodations for the station-agent, that he naay be on hand in an emer- gency. There should be a short stub-track and trucking-platform at the end of the house, with a ramp leading to the freight-room, as well as a door at the front. For places with sufficient traffic to require separate accommodations, the passenger-station can be designed on standard floor-plans, with dimen- sions corresponding to the volume of business, and the elevations varied architecturally. With a frontage of 50 to 75 feet and a width of 20 to 25 feet, half of the floor-space should be given to the general waiting-room ; and, with a frontage of 100 to 125 feet and a width of 30 to 40 feet, about two-thirds. The agent's office and the baggage-room could be at one end of the building. The women's room could occupy the front at the other end with a smoking-room at the rear and the two lavatories between them. The general waiting-room should have an opening into the baggage- room, with hard-wood benches, having arms, at intervals, to separate the occupants in groups. These benches should be fixed to the floor by fas- tenings easily removed for washing the floors, which should be preferably of granolithic or similar material, as well as the outer-platform floors. If the interior walls above the wainscot are of rough plaster, they can not be disfigured by scrawls or scratches ; and a whitened ceiling will aid the illumination. In a cold climate, hot-water heating is better than stoves, with radiators back of the benches, and the heater under observation in the baggage-room. The ground-platform should have a roof-covering and be not less than 15 feet in width, with a slight pitch toward the edge, which should be 5^ feet from the center of the track, with the support- ing posts 7 feet farther in. The end-extensions of the platform should be protected by an umbrella-shed. Near each end, station-signs in large lettering should be conspicuously placed and illuminated. Station-build- ings of this character were formerly framed structures but, of recent years, concrete construction is in more general use.^ At junction-points, the floor-plan of the station-building must conform to th& lay-out of the tracks, and more ample space is required in the waiting-rooms. Pro- vision must also be made for refreshment-rooms, news-stands and sanitary requirements for persons waiting between trains. Arm-chairs and rocking-chairs in the women's waiting-room are appreciated by weary women with children. On double-track lines, consideration must be given to the safe transfer of passengers between the waiting-rooms and the trains. In some cases, a covered platform is built on the opposite side of the line, for access to' 1 For standard plans recommended by the American Railway Eneineerino- Association, see Passenger Terminals," J. S. Droege, p. 254. " ROADWAY 287 trains on that side. Where this plan is adopted, there should be a fence between the tracks, for the length of the platform, with a crossing provided in front of the waiting-room through a gate, which should be locked when trains are due. Where trains are so frequent as to render this arrangement hazardous, the tracks should be crossed either by an overhead passage- way or by a subway. On a four-track line, to reach the middle tracks there must be width between them for an island-platform, accessible from the crossways. The terminals of all lines entering a city should be concentrated in a union-station. This is usually a problem difficult of a satisfactory solu- tion ; and especially in cities on the sea coast and the Great Lakes, where the railway lines radiate from stations widely separated by thronged streets and valuable real-estate. Yet union-stations are obviously so necessary to the general welfare that great efforts have been made to provide them. It is not essential to the convenience of the traveling public that a union- station should be located adjacent to the heart of a great city. The area for its efficient operation can there be acquired only at an enormous cost. If it be situated in the suburbs, means are soon provided for access to it locally. On a single-track line there should be a double-track approach to the station-tracks in a large station, with a cross-over, to avoid obstruc- tion to train-movements. Where there are many station-tracks, these approach-tracks should be again divided into "throat-tracks"; one for every two to* six station-tracks.^ For practical purposes, it is preferable that trains should be able to pass through a station without reversing their direction. In a through-station, the currents of passengers are more easily kept apart, each with its own station-faciUties ; communication between them being provided either below or above the tracks. This arrangement is often practicable for a union-station where the lines of railway pass by it, instead of terminating there. But where the lines of railway approach the station from widely different directions and terminate there, the only practicable plan is that of a head-house with dead-end tracks. Such a terminal requires a reverse- movement of every train that enters it. In European countries, where the locomotives and carriages are shorter and lighter than in this country, some of the terminals are equipped at the dead-ends with electrically-operated transfer-tables. At the St. Louis Union Station, trains back into the station over "Y" tracks. Plans have been devised for connecting the dead-ends by a loop. This requires con- siderable radial space beyond them for a loop with 180° of curvature, and can only be adopted where the tracks are on a level beneath the waiting- rooms. In fact, such a plan is of little value either for long-distance trains, 1 In the Union Station at St. Louis, there are two sets of throat-traoks of three each to thirty-two station-tracks. The South Station, Boston, has eight throat- tracks to twenty-eight station-tracks. 288 EFFICIENT RAILWAY OPERATION which must be broken up for the cars to be cleansed and provided with water and fuel, and the storage-batteries recharged, or for suburban- service where cars with electric tractors need not be reversed. Train-service in a busy station may be accelerated by the interposi- tion of three tracks between two platforms of sufficient length for each to hold two trains, with a double cross-over, or "scissors-crossing," midway, and the middle track used as a running-track. This plan is not favored in the United States on account of the normal length of trains, which would require a platform of 500 to 900 feet in length to accommodate two trains and to allow between them 160 to 175 feet for the cross-over. The height of ground-platforms in the United States is about ten inches above the rails. In European countries, except in Switzerland, the platforms are on a level with the side-doors in the carriages. This arrangement admits of more speedy departures with less effort to passengers, also of convenient access to the piping under the platform; but it requires the use of lift- decks on all cars with end-platforms, as in vestibuled cars. Track-plat- forms should be amply lighted with lamps on several separate circuits, so that a part of them can be cut out when the platforms are not in use. Where there are frequent train-movements, the service will be expedited, and with less annoyance to passengers, if it be practicable to have a sep- arate platform forljaggage-trucks between each pair of running-tracks; with a ramp to a crossing at the outer end, where the platforms are ele- vated. Union and Terminal Stations The problems connected with the planning of a union-station vary profoundly with its environment. In Boston, the lines approaching the city from one direction were concentrated in the North Station, and those from the other direction in the South Station. At St. Louis, the concen- tration was rendered morq complete by the construction of two bridges over the Mississippi River. The immense area covered by the city of Chicago has proved an insuperable barrier to the establishment of a single union-station, but there is a gradual concentration of lines in the same terminal which have a common interest in passenger-traffic or which are conveniently located for such a purpose. The Union Station in Wash- ington, D. C, is, in many respects, a model for a complete union-station as to track and platform arrangements, in its accessibility to waiting- rooms and exits, in its relation to environment, and in its architectural effect. In New York City, there is a partial development of two union-stations, as in Boston. The Grand Central Station accommodates the New York Central, the Harlem and the New Haven lines, and the Pennsylvania Railroad Station provides also for the Long Island Railroad trains^ and is now also accessible for trains from the New Haven lines, by the comple- v/ ROADWAY 289 tion of the New York Connecting Railroad over the entrances to Long Island Sound. This station is virtually a "through" station, made so by the depression of its tracks. If its capacity were greater, it would not be a difficult engineering problem to make it a union-station for all lines ter- minating on the opposite shore of Hudson River.' As with plans for ter- minal facilities for freight-traffic, those for passenger-terminals have not always been conceived with sufficient forecast. In New York City, the Grand Central Station has been both enlarged and reconstructed within a period of forty years. The new Pennsylvania Railroad Station had been opened but a few days when it was found necessary to remodel the facilities for Long Island Railroad trains. Even in smaller cities, there has been the same experience, with the demand for more extensive premises and more pretentious structures. In Philadelphia, the Pennsylvania Railroad Company found it necessary to remove its terminal station from the through route between the East and the West and, at great expense, to bring it into the heart of the city, in order to accommodate its suburban traffic. The conditions as to terminal passenger-stations in London are con- trolled by the original location of the lines as they entered the metropolis. These are eight in number and they severally occupy independent terminals, though the different lines are connected with each other by a maze of suburban-tracks. Little attention has been given to external architec- tural effect, and the accommodations for the convenience of passengers do not compare favorably with those provided in the great stations in Con- tinental Europe and in the United States. They are, however, well ar- ranged for expeditious train-service and for handling a large number of passengers daily.^ The planning of passenger-stations culminates in the design and con- struction of terminals where provision is to be made on an extensive scale for the receipt and disposition of throngs of travelers, with due regard to their safety, convenience and comfort between the trains and the station portals. A passenger terminal station in a great city belongs in the class of public buildings, and deserves monumental architectural treatment. Yet its practical purpose is to provide comfortable and convenient facil- ities for receiving and dispatching travelers. The use of classical motives and principles of design which prevail in buildings of a public character, should not affect the dimensions or terminal arrangements of a terminal ' For a description of the Pennsylvania Railroad Station, see Appendix V, Note XXVI. * The Liverpool Street Station of the Great Eastern Railway has a traffic averaging 176,000 persons a day. The St. Lazare Station of the Western Rail- way in Paris is the largest station in Europe, with a traffic, in 1909, of 52,800,000 passengers and a daily average of 144,000, which, in the summer, exceeds 200,000. The daily transfer of passengers from one train to another averages 20,000. The daily average in the Grand Central Station, New York City, is about 60,000 pas- sengers ; and in the Pennsylvania Railroad Station, about 47,000. In the South Station, Boston, it is about 105,000. 290 EFFICIENT RAILWAY OPERATION station, as regards its economic and social efficiency; nor should its ar- chitectural treatment ignore or conceal the features which indicate its purpose. Classical fa§ades may diminish the openings for air and light and affect the proper internal division of rooms. Sculptural groups are hidden at lofty heights and their existence is forgotten. Gaudy mural decoration is unnoticed by the passing throngs, and becomes obscured by dust and faded in the sunlight. There should be, preferably, a harmonious color treatment of broad surfaces, emphasizing the lights and shadows of the architectural details. Prominence should be given to entrances and exits and to the locatioi^ of the accessories necessary for the convenience and comfort of passengers. There is no rational objection to the space above the waiting-rooms being utilized to a suitable height for offices; thereby minimizing the expense for the maintenance of a structure which is otherwise an unproductive charge upon railway revenues. The traffic at such terminal stations is either suburban or long-distance, and the difference in the corresponding requirements is fundamental. The chief desire of dwellers in the suburban zone is to pass through the station expeditiously and with the least possible inconvenience. They move in large bodies, with little hand-baggage and less heavy luggage. They are commuters, and seldom have occasion to visit ticket-offices. They have no need of other conveniences than lavatories, for they do not linger in waiting-rooms but want ready access to the street-cars with protec- tion from the weather. Long-distance travelers have other requirements. They have more need of ticket-offices and Pullman reservations, of tele- phone-booths and telegraph-offices. There is yet another class of travelers who neither begin nor end their journeys at these terminals, but who pass through them on their way to another destination.^ These requirements, with proper ventilation, sanitation, lighting and heating, include the primary conditions for social efficiency in a large terminal passenger-station. It is further advisable to separate the arriving and departing throngs of passengers. To maintain the even flow of these opposing currents, there should be no intervening interruption to break their steady march in a straight line and on a uniform gradient between the train platforms and the station portals. This requirement is fulfilled by a vaulted con- course directly facing the train-shed, with the waiting-rooms, ticket- offices and accessory conveniences ranged around it and lighted through the exterior walls or from the ceiling.^ The station-signs in the con- » The Union Station recently built in Kansas City is probably the largest station in the world that is devoted almost entirely to this class of passenger- traffic. For a description of this station, see Appendix V, Note XXVIII. 2 The principal concourse in the Grand Central Station has an area of 81,122 square feet ; that in the Union Station at Washington has an area of 98,800 square feet and is supposed to be the largest unbroken floor space in any room in the world ; that in the Pennsylvania Raikoad Station in New York City has a broken floor- space of 131,400 square feet. ROADWAY 291 course should be well arranged and of good size, and the track-numbers and train-designations conspicuous and distinct for the instant and ac- curate guidance of the hurrying passengers. The coloring of the extensive wall and ceiling surfaces in a concourse affects the illumination. A white surface reflects 50 per cent, of the intensity of the Hght striking it, red or dark green but 15 per cent, and dark brown but 2^ per cent. Artificial lights should be so distributed as to give a uniform illumination over the floor-area. Electric lights should be shielded by ground-glass or opal- escent shades, and heavy shadows avoided. Train Sheds Passenger-stations in large cities require special treatment to suit each environment, physically and socially. . The buildings may be of greater dimensions, more architecturally ornate, or more diversified in arrange- ment, but there are no additional principles in which they differ, as a class, from other station-structures, except where the track and platform lay-out is amplified and roofed over. The train-shed proper is usually a lofty vault of arched iron ribs covered with glass and sometimes extends over as many as thirty-two tracks.* It is an imposing but costly structure, resonant with conflict of noises, the iuterior stifling at times with coal-smoke which settles upon the glass overhead, dimming the light and corroding the frame- work. The artificial lighting is necessarily concentrated at an elevation at which much of the illumination is useless. The air is draughty in winter and oppressive in sunmier, and the overhead structure requires frequent attention, under diSicult conditions, to keep it in repair. These defects in the vaulted train-shed a^ffect more or less the conditions in the adjacent station-building. The objections to a vaulted train-shed have led to the introduction of individual sheds over the several train-platforms, known as "umbrella" sheds. This type of train-shed, supported by a central line of posts, has the merits of cheaper construction, better ventilation and easier illumina- tion; but with the disadvantage of less protection from bad weather. The "butterfly" shed is a variation of the umbrella-shed, with the pitch of the roof reversed to form a midway gutter, drained through the cen- tral posts. It therefore protects passengers from the drip at the eaves when entering or leaving the train, and gives better light and ventilation when several trains are at the platforms. The butterfly shed was further developed by Lincoln Bush, in a design installed in 1905 at the Lackawanna Terminal in Hoboken. In the " Bush " shed, the inverted eaves of the butterfly shed are extended to connect over the center-line of the intervening track in a series of low vaults. The peculiar feature of this type of shed is the introduction of a duct along each vault over the center-line of the track, so near to the top of the loco- 1 The shed of the South Station, Boston, covers 28 tracks. 292 EFFICIENT RAILWAY OPERATION motive stack that the smoke escapes directly into the open air. The trains and platforms are entirely protected from the weather and the farther sides are also iaclosed. The shed is lighted by continuous skylights in the vaulting. The Bush shed at the Jersey City Terminal of the Central Railroad of New Jersey, completed in 1914, covers 20 tracks and 7;07 acres. There are two tracks between platforms under a vault, with a smoke-duct over each track and three rows of skyUghts in each vault ; the middle row being in a low monitor roof. The rainfall is carried through the central line of posts into an outfall drain. There are now eleven sheds of this type in America.' Fekry Terminals The isolation of great cities from railroad cormnunication by water- courses, too broad to be readily spanned, has caused the construction of railroad ferry-terminals. Terminals of this character, on an extensive scale, are situated on the western shore of Hudson River at New York and, at San Francisco, on the opposite shore of the bay. A distinctive feature of some of these terminals is the location of the waiting-rooms and accessory conveniences for passengers in the second story of the building to afford easy communication with the upper-deck cabins of the ferry-boats. The terminal buildings of recent construction have steel frames sheathed with copper. They rest on piling, and the roof projects over the slips sufficiently to protect the passage to and from the boats. Buffer platforms are'placed in the floor at the head of the slips to diminish the shock to the terminal structures from impact. As the ferries are also used for local street-traffic, driveways to the boats and passenger-accommodations are necessarily required on the ground floor, separated from the railroad terminal. A large suburban-traffic is to be provided for, in addition to the long-distance travel ; also the terminal is necessarily duplicated on the city side.^ Special Requirements for Elevated and Depressed Tracks IN Large Cities Track-elevation through large cities necessitates special arrangements of passenger-stations. Where the environment admits of driveways on suitable gradients, the station-building may be on the same level as the tracks, or even on a somewhat lower level and connected with the train- platforms by a slight incline. If the space is not available for this plan, the station-accommodations must be at the street-level, and the several platforms reached by subways in connection with stairways and elevators or, more efficiently, by moving inclined planes or "escalators." ' See "Passenger Terminals," Droege, p. 40. 2 See description of the Jersey City Terminal of the Central Raib-oad of New Jersey, Appendix V, Note XXVIII. ROADWAY 293 The entrance into a city by tracks below the street-surface permits of a different arrangement of station-accommodations and of an enlargement of space within the surface-area prescribed by local conditions.^ The additional space is obtained by extending the tracks and platforms beneath the lateral street-surface with a concourse on the same level but lighted and ventilated within the surface-area of the railroad property and connected by stairways with waiting-rooms and exits at the street- level. Such stairways should be planned to facilitate the free movement of throngs of passengers. The upward movement will average 20 to 30 persons a minute per foot of width and 18 to 25 persons downward ; all moving in one direction over stairways not less than four feet in width ; they should be at least twice that width for a movement both ways simultaneously. The treads should be 11 inches wide with risers 6^ inches high, and the top and bottom steps should be brightly lighted. Where the relative levels and heights will admit of the use of inclines, or ramps, they are preferable to stairways, permitting a much more rapid move- ment. In the Grand Central Station, they have been introduced generally and on an extensive scale, on gradients of between three feet and eight feet to 100 feet. The area occupied by the principal waiting-room in large terminal stations in the United States varies between 25,000 square feet in the Chicago Station of the Chicago & North Western Railway and 33,000 square feet in the Pennsylvania Railroad Station in New York City. The Liverpool-Street Station of the Great Eastern Railway in London has four waiting-rooms with a combined area of 3000 square feet. Space in bag- gage-rooms varies considerably : it is 27,794 square feet in the South Sta- tion, Boston ; 41,683 square feet in the Union Station in Washington ; 50,000 square feet in the Pennsylvania Station in New York, and 74,048 square feet in the Kansas City Union Station. ' The baggage-rooms in the Grand Central Station are in a separate building. The several offices for fur- nishing information, tickets and Pullman reservations, and for checking baggage and parcels, should be arranged in successive order adjacent to the concourse and the general waiting-room. Carriage and baggage-trans- fer offices, telegraph-offices and telephone-booths should also be con- veniently sought in that vicinity. Refreshment-rooms should be con- spicuously placed, with a clock prominently in view and notices of train departures.^ ' The Pennsylvania Railroad Station in New York City has double the space of the Liverpool-Street Station of the Great Eastern Railway in London ; nearly twice that of the recently-completed Leipsie terminal, the la-rgest in Germany, and about three times that of the South Station, Boston, or of the St. Louis Union Station. For descriptions of recently-constructed terminals, see Appendix V, Notes XXV and XXVIII. ^ Much of this information has been obtained from the work on "Passenger Terminals," by J. A. Droege, 1916; which contains illustrated descriptions of the principal passenger-stations in the United States and in Canada. 294 EFFICIENT RAILWAY OPERATION M Reconstruction of Grand Central Terminal, New York City The reconstruction of the Grand Central Station and Terminal was a re- markable engineering feat, because of the unusual conditions under which it was accomplished. The original terminal occupied about four acres. The train-shed, built in 1871, covered 15 tracks and accommodated a daily service averaging, 88 trains. It was 530 feet long and 100 feet in height, with a roof of 200 feet span. In 1884, an annex was added with four addi- tional tracks. In 1900, the station was enlarged at a cost of $2,500,000 ; the number of tracks was increased and three stories added to the building. The walls contained 750,000 bricks and the material in the train-shed in- cluded 1350 tons of wrought-iron, 350 tons of cast-iron, 90,000 square feet of corrugated iron and 60,000 square feet of glass ; and there were 15 miles in length of piping for steam-heat. In consequence of an accident in the tunnel-approach in January, 1902, the legislature passed an act requiring trains to be operated through the tunnel by electric traction. The sub- stitution of electricity for steam as a motive force made it possible again to reconstruct the station on a far more extensive scale. The old structure was removed without interruption to the train-serv- ice within it. To prevent the material of the train-shed from falling upon passengers, a movable platform was operated upon wheels, beneath which there was a wooden hood shielding the platforms and covering two acres of glass. Meanwhile, the excavation for the new terminal was carried on around the old one and beneath its very foundations, perhaps at double the cost of such work imder ordinary conditions. Before a part of the train-service was removed to a temporary location to make way for the excavation, there was provided a complete system of train-signals, with interlocking and electric power for train-movements ; and a similar system was installed at the permanent location before the trains could be trans- ferred back to it. The maximum number of trains in and out of the terminal was 833 in 24 hours, and the minimum was 750. The work was prosecuted night and day. The men engaged in moving the tracks had to drop their tools every few minutes for passing trains. No blast- ing was done in daylight hours ; only between 9.30 p.m. and 6.00 a.m., when train-movements were fewest. It was only practicable to work 500 men by day and 350 at night. The adoption of electric traction made it^ possible to depress the road- bed below the surface on each side of the old road-bed, beginning at the outer end of the four-track tunnel at 57th Street. Here the suburban tracks diverged and dipped gradually down to the lower level, reaching their greatest depth at about 45th Street. On this lower level, the full width of Park Avenue is utilized. By this plan, 178 per cent, of additional space was added to the terminal yards. Formerly, all trains returned empty for five miles to Mott Haven for yard-room. There were 18 trains ROADWAY 295 daily to the East and North and a train every hour to or from the West. These trains had to be made up again in reverse-order and the Pullman cars placed on wired tracks for recharging their storage- batteries while they were being cleansed. All this work is now provided for within the limits of the terminal, without interference with the main tracks. The cross-streets from 45th to 57th Street, formerly cut in two by the terminal yard, and Park Avenue, which was discontinued south of 57th Street, have now become continuous thoroughfares in a part of the city where the space covered by a ten-car train is appraised at a value of $280,- 000. The terminal yards are covered over for the full width of Park Avenue, 140 feet, from 57th Street to the north end of the station-build- ing, and a similar covering of VanderbUt Avenue and Depew Place af- fords a roadwaiy around each side of the station to a plaza in front of it at an elevation sufficient to continue Park Avenue southward by a viaduct over 42d Street. This improvement greatly diminishes the un- productive area of the railroad property by making 46 acres of it avail- able for building sites. The construction over the tracks has been so designed that the noise of the trains beneath is not perceptible in the buildings above them. This final reconstruction of the Grand Central Station epitomizes the development of station-service in the United States during a period of forty years. In 1910, on an average, 60,000 passengers passed through the station daily. The estimated capacity of the present station is for 70,000 passengers and 200 outbound trains per hour. During the period of reconstruction, in the eight days from August 30 to September 6, 1912, 944,000 passengers passed through the station, and 4826 trains were han- dled with an average delay of 21 seconds. The real-estate covered by the new terminal is valued at $50,000,000. The cost of excavation and put- ting up the steel to carry the train-service and the street-traffic overhead was estimated at approximately $50,000,000; and the various buildings connected with the terminal have cost between $60,000,000 and $80,000,- 000 more.i '■ Difficulty of Providing for Future Growth and Expansion Few enterprises are carried out even as fully as first conceived and still less is adequate provision made for their subsequent development, often far exceeding the original design. Even in old-settled countries, it has not been possible to furnish in advance the increased facilities needed for the volume of trafiio quickly following upon the early period of railway construction, and, in a newly-settled country, this is a far more difficult matter. In no other country has this been the case to such an extent as ' For a description of the Grand Central Station, see Appendix V, Note XXVII. 296 EFFICIENT RAILWAY OPERATION in the United States. Even where a railroad has been constructed in ac- cordance with the best methods at the time in vogue, it has hardly been opened for traffic before the accompanying development of the region that it serves has begun to press upon its facilities for performing that service. Villages spring up along the line and require accommodation ; industries are established that demand siding-facihties. Branch-lines and exten- sions feed the trunk-line with increasing volume of traffic which requires heavier locomotives, and these compel the strengthening or reconstruction of the bridges. There is, also, the accompanying necessity for heavier rails and track-accessories. Increased train-service requires improve- ment in signal-apparatus; passing-sidings become numerous; many are lengthened as running-tracks and into stretches of double-track, until at length, whole divisions have a second track. At division-points and at interchange-junctions, cars are concentrated for classification and dis- tribution, calling for more switching-facilities and increased yard-room, until the accompanying rise in the price of adjacent property reaches a figure which compels removal to a situation where a greater area may be available at lesser proportionate cost ; while roundhouses, shops hnd other accessory structures must follow the switching-yards. There is a continual pressure from public opinion and by legislation for the elimination of highway- crossings at grade, and the growth of cities along the line brings with it the necessity for heavy expenditure in elevating or depressing the tracks for the protection of street-traffic. This is the ordinary, everyday experience of the management of any. railroad that is in a dividend-paying condition. There is a continual de- mand for the means to meet the requirements of increasing traffic and to conform to advancing standards in operation. The cost of much of such betterment is met out of operating-revenueSj and that of additional equipment by equipment-trusts, until the time arrives when the re- quired expenditure for second track and for terminal facilities can only be provided for by an increase of capital or of bonded debt. Indeed, the main railway thoroughfares in the United States have suffered so much financially, as well as in efficiency, from inability to n^eet the demapds of increasing traffic that their managements have learned by experience that they must provide in advance for its further development. To this end, they have even undertaken the relocation of considerable portions of their lines. The Pennsylvania Railroad Company experienced such difficulty in operation, from congestion in its terminals in and around Philadelphia, that a "cut-off," 45 miles in length, was built from.Morrisville, near Tren- ton, N. J., to the Philadelphia Division at Glen Loch, near Downingtown, to get the through-freight trains around Philadelphia instead of passing through the terminals there. This line was built in 1889-1891 at a cost of $3,142,000, or about $70,000 per mile. ROADWAY 297 The Great Salt Lake cut-off, on the Central Pacific Railroad, is an example of relocation on a yet more extensive scale, by which many miles of heavy grades and curvatures were avoided, and which is as remarkable for rapidity of construction as for its economic efficiency .^ An instance of expensive relocation, merely to increase the tonnage- capacity of trains, was undertaken on the Delaware, Lackawanna & Western Railroad .^ The reconstruction of the Erie Railroad, with the resto- ration of the corporation upon a firmer financial basis, has been prose- cuted for thirteen years so vigorouslj' as virtually to constitute a re- habilitation of this great system.' No feature of railway reconstruction is more impressive than the . / conception and organization of important terminals. Inadequate fore- ' cast in providing terminal facilities has compelled subsequent expen- ditures, enormous in amount. This h^s been the usual experience wherever there is an interchange of a large volume of traffic. With the rapid expansion of railway transportation, the arrangements and under- takingsfor the original reception and for the final disposition of traffic at such points, and also for the intermediate exchange of cars, have become an extensive field for capital investment, and have also given rise to an important branch of roadway engineering. Terminals differ in requirements for handling freight in bulk from what , is required for handling freight in packages. In the former case, the commodity is initially stored at an elevation, from which it can be dis- charged by gravity. If it be grain, it is lifted by elevating-apparatus into bins. The terminal arrangements for transferring grain between rail and water at ports on the Great Lakes, are remarkable for efficiency. There is such a terminal at Buffalo with a capacity for 1,000,000 bushels. The structure is of concrete, with a dock 811 feet in length and space for docking either two large or three small vessels. It has facilities for discharging the largest cargo in ten hours, for unloading cars at the rate of 40,000 bushels an hour and for loading into cars at the same rate, while, at the same time, discharging into canal-boats at the rate of 20,000 bushels an hour. This chapter concludes the study of the instrumentalities of railway transportation. Motive Power is the dynamic force. Rolling-stock is the medium, and Roadway is the static element of railway service. Con- struction by vital energy with pick and shovel, with barrows and carts, has been largely supplanted by mechanical appliances energized by steam, compressed air or electricity ; steel and concrete, as materials, have been substituted for timber and stone. With these more efficient means, human intelligence has transcended the limitations prescribed by topographical environment; sinking foundations through unstable marshes, burrowing 1 See Appendix V, Note XXIX. ' Appendix V, Note XXXI. 2 See Appendix V, Note XXX. 298 EFFICIENT RAILWAY OPERATION beneath the beds of rivers, broad and deep, or crossing them in a single stretch, and even penetrating insurmountable barriers crowned with eternal ice. By comparing the railroads of half a century ago with those of to-day, we can form some conception of what has been accomplished by two successive generations of engineers in bringing our railway system into its present state of efficiency and in developing, from our natural resources, the traffic which has made us a wealthy and powerful nation. CHAPTER VI TRAFFIC Pheliminary Definitions In the establishment of any standard of efficient railway operation, due consideration should be given to the characteristic features of the traffic which is properly the subject matter of Transportation. As here con- sidered, Traffic includes all operations connected with the reception of the persons and commodities to be transported, with their delivery after ar- rival at destination, and with their condition intermediately. This field of railway efficiency is not covered merely by the transporta- tion from the point of reception to destination, but is occupied by a serv- ice for which no direct charge may be made, though it adds materially to the cost of operation. It is not measured by ton-miles, nor by passenger- miles, by locomotive-mileage nor by car-mileage. It may be considered as including the various and separate services rendered before, during and after, the actual service of transportation, in gathering together the per- sons and commodities to be forwarded, and in disposing of them after their arrival at destination. It also covers their proper care during transit, as well as the assessment and collection of charges for the trans- portation-service, and the responsibility for covering the resulting revenue into the railway treasury. With the latter exception, it is almost exclu- sively a service of social efficiency which, from the difficulty of defining it otherwise, may be distinguished as Traffic Efficiency. There are three elements of Traffic Efficiency in any mode of trans- portation — expedition, convenience and safety. These requirements are superficially valued in the order here named ; though, as safety is really the essential element, it should always be the first consideration as an evidence of traffic efficiency. This is particularly true as to passenger- traffic, in which the new slogan of "Safety First" is especially applicable. "Safety First" consists in precautions against the occurrence of accidents, in contradistinction to provisions for mitigating their consequences. When our sensibility is shocked by a thrilling newspaper account of some railway catastrophe, we jump to a conclusion as to the inefficiency of railroad man- agement that is not sustained by a sober investigation of statistics bearing upon the subject. If it were possible to compare the meager passenger- 299 300 EFFICIENT RAILWAY OPERATION mileage in Great Britain during the stage-coach era with the thousands of millions of railway passenger-miles that are annually accomplished in that country, the foregoing statement would find ample warrant in the com- parison. Safety in Railway Travel. Accident Statistics There are no passenger-mile statistics in Great Britain ; only passenger- journeys, exclusive of those by season-ticket holders. In 1908, out of 1,278,000,000 of such journeys, there was not a single passenger killed, and but 283 were injured. In the following- year, there was one killed and 390 were injured. In the decade from 1901 to 1910, there were two years in which not a passenger was killed, one year in which but one was killed, and in the remaining seven years the proportion varied between one in 21,385,000 journeys and one in 199,758,000 journeys. The annual number of injuries to passengers in this decade varied between 283 and 1111 ; and the proportion as to journeys, from one in 1,176,000 to one in 4,515,000. The total number killed in the decade was 176, or about 17 persons per annum, out of an annual number of journeys varying between 1,172,000,000 and 1,306,000,000 ; not including the millions of journeys made by season- ticket (holders. During this ten-year period, only 112 employees were killed. As comparisons have been published of the safety of railway service in Great Britain that rather disparage railway efficiency in the United States, -reference may be made to the statistical information on this subject contained in the reports of the Interstate Commerce Commission. It is preferable to consider only the period 1906-1914, .both years included, as during this period the statistics are on a more uniform basis than in pre- vious years. In this period, the fatal casualties to passengers from train- accidents varied annually between 71 and 367 in a passenger-mileage vary- ing between 25,000,000,000 and 35,000,000,000 miles. In 1914 the record was made of 71 fatal casualties, or a proportion of one to 494,778,436 passenger-miles . To bring these statistics into some sort of comparison with British statistics, it should be noted that these casualties occurred on a line-mileage increasing from 224,603 miles in 1906 to 252,230 miles in 1914 ; the British mileage being about stationary at 23,000 miles.^ In 1914, 315 out of a total of 1112 operating-companies in the United States, had a clear record in this respect. These companies operated a line-mileage of 113,333 miles, which is almost equal to the combined mileage in Austria-Hungary, France, ' In the United Kingdom, whicli boasts of an occasional year without a single fatahty to a passenger from a train-accident, the record for the past forty years has averaged 21 a year. Compared on the basis of units of risks, of mileage and trafQe in the two countries, this would be equivalent to over 210 a year in the United States. — Railway Statistics of the United States for 1914. Bureau of Railway News and Statistics, p. 117. TRAFFIC 301 Germany and the United Kingdom. The traffic of these companies in- cluded 43.5 per cent, of the total passenger-mileage of the country, and 50.9 per cent, of its total ton-mileage. There were 104 managements operating a mileage equal to that of the United Kingdom which, in the four years 1910- 1914, had no.fatal casualties. There were also 23 managements operating 34,826 miles of line on which but one passenger lost his hfe in a train- accident in 1914 on each line. There were 338 managements operating 148,159 miles of line, or 58 per cent, of the total mileage of the United States, on which but 23 passengers were killed in train-accidents in 1914. The total number of passengers carried over this mileage in that year was 573,800,000, with a mileage of 20,234,000,000 miles, equivalent to over half a billion journeys of 35 miles each. The statistics as to non-fatal casualties to passengers in train-accidents cover larger numbers, but among them are included many of minor impor- tance and, to some extent, of questionable origin. Such casualties on Brit- ish railways numbered 683 in 1912 and 723 in 1913. In the United States, similar casualties amounted to 7515 in 1913 and to 5993 in 1914, which is about equivalent to the numbers reported in the United Kingdom, in pro- portion to the Hne-niileage. There remain to be considered the casualties to passengers from other causes than from train-accidents ; such as falling from trains, or in attempt- ing to get on or off a train in motion, or in crossing tracks. From the com- parison in the note below, the fatalities to passengers from these causes were far less, proportionately, in the United States.^ The greater num- ber of non-fatal casualties is in some measure due to the stricter reports required by the Interstate Commerce Cormnission, yet, even in this matter, the number is less than three times as much with eleven times greater mileage. The result, in general, is remarkable, in view of the greater freedom of movement at stations which is permitted to passengers in this country. The casualties to employees are to be considered apart from those to passengers. In their hazardous occupation, they are daily and hourly exposed to dangers from which passengers are exempt. This is clearly shown in the accident-statistics given in Appendix VI, Table I. In ^Casttalties to Passbngebs Other Than in Train Accidents United Kingdom United States 1912 1913 1913 1914 KiUed .... Injured . . 100 2,843 117 2,918 195 6,892 152 7,047 Bureau of Railway News and Statistics. 302 EFFICIENT RAILWAY OPEEATION comparing statistics of casualties of this character with others covering previous years, they should be considered in connection with the train- mileage in the periods under comparison/ which was as follows : 1907-1909, average 1,171,242,407 miles 1913, average 1,327,749,456 miles 1914, average 1,293,629,513 miles On this mileage, the casualties to employees per train-mile were as follows : Killed Injuked 1907-1909, one in 687,620 miles 1913, one in 861,056 miles 1914, one in 993,570 miles One in 57,366 miles One in 46,141 miles One in 62,244 miles It is remarkable that, while the proportion of train-mileage to one death of an employee in service, in the periods under consideration, should have increased by 44 per cent., the proportion of non-fatal injuries should have remained about stationary. This difference may partially be accounted for by the stricter demands for information in the reports to the Interstate Commerce Commission as to minor injuries, and to the effect of legislation for insuring compensation to injured employees. The virtually complete equipment of the rolling-stock of the country with automatic couplers has not effected the diminution in coupling-acci- dents that might have been expected. The casualties from this cause have been as follows : ' Killed iNJtTRED 1907-1909 (average), 1913 1914 202 196 171 3,150 3,360 2,692 As such casualties occur mainly in switching, they may be considered in proportion to the estimated mileage of switching-engines, as follows : • Casualties to Employees. Average for 1907-1909 Causes Killed Injured C!oi]T>liTie' oars 202 659 513 359 70 3,150 5,401 10,081 883 902 Falling or jumping from trains Stnip,!?' V>v trains . Overhead obstructions Total 1,703 20,417 TRAFFIC • Switching Mileage 303 Killed Injttbed 1907-1909 (average), one in 1,426,286 miles 1913 (average), one in 1,776,122 miles 1914 (average),. one in 1,996,007 miles one in 91,430 miles one in 103,086 miles one in 124,883 miles There is a progressive diminution in coupling-accidents indicated in this comparison, however, while there will always exist an irreducible minimum due to the recklessness that accompanies habitual exposure to danger. In examining these accident-statistics, attention is drawn to the pre- ponderance of casualties of employees due to falling or jumping from trains or from getting on trains in motion, in proportion to the total nimiber of the casualties to employees.' From these statistics, it appears that about one-third of the deaths and one-haK of the other casualties to employees in railway service occur from causes not directly affected by train-accidents but are probably due, to a great extent, to the indifference to danger which has contributed to coupling-accidents. From the proportionately larger mmiber of non-fatal casualties, it may be assumed that they were mainly sUght injuries. The Interstate Commerce Commission's accident-statistics for 1916 exhibit a total of 9366 deaths and 180,380 injuries. This muster-roll of casualties, appalling as a whole, admits of very considerable reductions when considered solely from the standpoint of railway operation. It may be separated into its component parts, as follows : Casualtieb To Killed Injured No. Per CeDt. No. Per Cent. Passengers Employees Others, not trespassers . . Trespassers Shop-hands 283 2,272 1,478 4,847 486 .030 .243 .158 .517 .052 8,379 43,152 4,444 5,109 119,296 .047 .239 .025 .029 .660 Total 9,366 180,380 1 Casualties to Employees in Falling or Jumping prom Trains Killed Injured No. Per Cent, of Total No. Per Cent, of Total 1907-1909 1913 1914 513 567 640 30 43 41 10,081- 15,110 16,565 49 60 55 304 EFFICIENT RAILWAY OPERATION From this analysis, it appears that over half of the persons killed were trespassers, and but 27 per cent, were passengers or train-employees. Of the other casualties, two-thirds were to shop-hands and about 29 per cent, were to passengers and to employees in railroad operation. In estimating the comparative safety in train-operation, a distinction is to be observed between casualties which result directly from train- accidents, those occurring from other causes, and those not connected with railway operation, as follows : Casualties on Railroad Property in 1916 Pebsonb Train Accidents Other Causes Total Killed Injured Killed Injured Killed Injured Passengers . . . Employees . . . Non-trespassers Trespassers . . . 141 313 11 84 3,850 3,412 92 119 142 1,959 1,467 4,763 4,529 39,740 4,352 4,990 283 2,272 1,478 4,847 8,379 43,152 4,444 5,109 Total . . . Shop hands . . . 549 7,473 8,331 53,611 8,880 486 61,084 119,296 All casualties . . 9,366 180,380 The great disparity between these classes of casualties induces a further analysis of those occurring from other causes in railway operation proper than from train-accidents, which may be divided as follows : Caitses Killed Injured Coupling Overhead obstructions Falling from trains 123 64 441 2,194 1,323 12,488 Other Causes, classified Other Causes unclassified 628 7,703 16,005 37,606 Total from Other Causes . . . 8,331 53,611 As to classes of persons, the unclassified casualties were divided as follows : Employees Non-trespassers Trespassers Total . . Injured TRAFFIC 3C5 A further discussion of official accident-statistics seems to be war- ranted by the impression that they have made upon public opinion as to the characteristic recklessness attending railroad operation in the United States. Such an impression may well result from a statement, without further comment, that in the year 1916, 9366 deaths and 180,380 lesser casualties occurred in operating the railway system of this country ; when, in fact, from train-accidents, there were but 549 deaths and 7473 other casualties, of which only 141 deaths and 3850 other casualties were suffered by passengers. Our railway managements may well take exception to statistics which include 486 deaths and 119,296 other casualties to shop-hands, or about two- thirds of the total, with those actually occurring in railway operation. This is of special importance in a comparison with railway casualties in European countries, in which the statistics are confined to accidents occurring in train- service, and which do not include casualties due to the sufferer's own fault or mischance. A special committee appointed by the American Railway Association to consider this matter, has requested the Interstate Commerce Commission to adopt a more definite classification of casualties as to causes and persons, with especial reference to persons neither passengers nor em- ployees, and as to a clearer segregation of industrial casualties and of cas- ualties to trespassers. The committee has also suggested that a relative measure of traffic-efficiency in these respects should be indicated by the number of accidents in proportion to the locomotive-mileage and, as to industrial accidents, in proportion to the number of working-hours. The committee has further endeavored to induce the Railroad Commissions in the different states to conform to the requirements of the Interstate Commerce Commission with reference to accident-reports. If "Safety First" consists in precautions against accidents, in contra-, distinction to provisions for mitigating their consequences, a discussion of their causes should precede a search for remedies. As affecting railroad operation, an accident is an unforeseen occurrence which interrupts the normal train-movement. Other accidents may be properly separated into two classes or groups — those for which railroad managements may justly be held responsible, as resulting from defects in track or equipment or reg- ulations ; and those beyond their control, as when caused by malice, or by negligence, or by disobedience of orders. Against this latter class of accidents, precautions are of no effect. They can only be diminished in number by the intervention of the strong arm of the Law. Collisions and Derailments Accidents to trains are not always accompanied by casualties, and their number can not be ascertained from casualty-statistics. They are either collisions or derailments. Collisions are of four kinds : butting- collisions, parting-collisions, "side-swipes" and rear-collisions. 306 EFFICIENT RAILWAY OPERATION Butting-collisions can only occur on a line, operated under single-track rules, either when those rules are not strictly observed or when they have been superseded by train-orders. A failure to deliver such orders to all trains thereby affected is the most frequent cause of butting-colhsions, though they may occur from a misinterpretation of orders. They are but rarely due to errors in the orders themselves, as the authorized forms to govern special train-movements are in such general use that errors of this kind are only likely to occur when the presence of some train within the affected district has been overlooked. Butting-collisions also result from misplaced switches. Switches may not only be left wrong after having been used ; in two instances on the same line, a switch was opened in front of an approaching train. Both trains had been some- what delayed ; the man sent to let his train out of the siding sat down near the switch, became drowsy and, when aroused by the noise of the approaching train, half-consciously changed the switch before the train had passed it. Parting-collisions occur from the separation of a moving train in sec- tions and are caused by deficient draw-gear. They have become less fre- quent since the introduction of standard couplers, and serious consequences are averted by the quick action of the air-brakes on the rear section of the parted train. Side-collisions are caused by cars standing on a siding so near the main line as to be struck by a passing train. On lines with two or more tracks, the derailment of part of a train while meeting a train in the opposite direction has occasionally resulted in a side-collision. Rear collisions have caused most of the catastrophes in train-service. They have occurred, both in Europe and in this country, on lines pro- vided with the most approved methods of operation. In the majority of instances, the original cause has been a delayed train encroaching on the time of a following train, or making an irregular stop between stations. Unexpected Stops at Unusual Places The Standard Code, of Train Rules, as approved by the American Railway Association, requires that, "When a train stops under circum- stances in which it may be overtaken by another train, the flagman must go back immediately with flagman's signals a sufficient distance to insure full protection, placing two torpedoes, and, when necessary, in addi- tion, displaying lighted fusees." "When signal 14 (d) or (e) has been given to the flagman, and safety to the train will permit, he may return. When the conditions require he will leave the torpedoes andalighted fusee."' There are no specific regulations in the Standard Code prescribing the duty of either the conductor or engineman of a train in such an emergency. ' Signal 14 (d) — four long whistle-blows to return from west or south, (e) — five long blows to return from east or north. TRAFFIC 307 Among the whistle-signals, there is a signal of one long and three short blows as an " indication " that the flagman "should protect the rear of train," and there is a rule in the Code that, "Both the conductor and the engineman are responsible for the safety of the train and the observance of the rules, and, under conditions not provided by the rules, must take every precaution for protection." When a train stops unexpectedly in an unusual place, the engineman has no other specific responsibility placed upon him by the rules, than to give the prescribed whistle-signal as an "indication" that the flagman "should protect rear of train," and there is no responsibility placed upon the con- ductor other than the general requirement that a conductor and an en- gineman "are responsible for the safety of the train." That this respon- sibility is perfunctory in accidents from such rear-colhsions, is shown by the fact that it is almost invariably stated that the flagman did not, or could not,' go back far enough to protect the rear of the train, even on double-track lines operated under a block-system. Under the Standard Code, upon hearing the whistle-signal, the flagman, upon his own motion, and without waiting for an order from his conductor, is required to leap f rdm the rear of the moving train as soon as he can safely do so, and when he has gone back as far as he thinks that it is necessary, he plants his torpedoes and listens with eager ear for the signal for recall. Unless he has reason for supposing that another traifi is following closely, he will probably wait for his train to come to an actual stop and will linger for a minute or two, hoping for the recall before he starts to the rear. Often, he is required to plunge into the daTkness of the night, burdened with lantern, fusees and torpedoes, perhaps facing rain, snow or sleet. It requires a stout heart to hasten toward the glare of an approaching head- light, as he feels his way over slippery cross-ties upon some lofty bridge or long trestle. If, through inadvertence or undue haste, the train moves off before he can return to it, he may pass the night in solitude, perhaps wet, cold and hungry, until some following train stops at his signal and picks him up. Such are the conditions under which a flagman may be required to pro- tect the rear of a train, when it stops unexpectedly at an unusual place, and it takes pluck and endurance to meet them fully. It also takes intel- ligent judgment to determine promptly just when a flagman should go back, how far he must go, and what he should do when he gets there. Yet this important duty is intrusted entirely to a novice in training for a con- ductor's berth, or to some sturdy brakeman accustomed, it is true, not only to the hardships of train-service, but to avoiding them as well. Either through ignorance or doubt, or from fear of being left, the flagman may disappear in the darkness only just around a curve, or near enough to be handy when recalled, and taking the chances as to whether a train is following or not. 308 EFFICIENT RAILWAY OPERATION This is a fair statement of the conditions under which a flagman is re- quired to protect the rear of a train that is stopped unexpectedly at an un- usual place, and of the manner in which he may be expected to fulfill the requirements of the Standard Code. Railroad managements may well be asked whether the Standard Code also fulfills the requirements neces- sary in such cases ; or, if it be admitted that the Code does not fully do so, whether it be not practicable to take some further precaution in such emer- gencies. Space Interval, How Best Obtained Since the necessary interval of space is not practically insured by the block system, as is proved by the frightful catastrophes that have occurred upon lines equipped with its most approved appliances for securing safety in train-operation, it may be well to consider whether some additional means can not be suggested for securing this interval, other than by relying solely upon the intelligence and the devotion of a flagman. One means may be suggested that does not call for more intelligence or for greater devotion to duty on the flagman's part, but which seeks to obtain both from another source — from the locomotive engineer. He is generally the most experienced man in the train-crew ; the best acquainted with the curves, grades, bridges, cuts, embankments and other physical features of the line; the best informed as to the trains passed and to be passed ; and, when a stop is made or the train is slowed down at an unusual place, he knows the cause and what the probable detention will be, not only after it has occurred but often before, and can usually select a suitable place to stop. It is he then, and not the flagman, who should determine when and how the rear of his train should he protected. If the burden were plainly put upon the engineer to determine, and upon the flagman to act, his action would be controlled by the best-informed man in the crew. With such a modification of the Standard Code, the space-interval between trains moving in the same direction might be more securely pre- served in an emergency, by a more extended recognition of the usefulness of the fusee, which at present is only permissive as a part of the flagman's equipment. How much more valuable in the hands of the engineer! Whenever he is about to stop or to slow down his train at an unusual place, let him be Required to drop a lighted five- or ten-minute fusee by the side of the track one mile before the stop is made and the interval between that train and a following one will have been positively secured by a sentinel that will not desert its post, by a signal whose unmistakable light will illumine its surroundings, let the wind blow and the rain fall as they may. It will indeed be a cloud of smoke by day and a pillar of fire by night ! This statement is not hypothetical, but is founded on ample experience. Such a requirement does not do away with the protection afforded by the flagman, but rather increases it. As the rear of the train passes the blazing TRAFFIC 309 fusee, the conductor will have warning to see for himself that the flagman goes back. As the flagman crosses a bridge on his way to the rear, he will feel secure against an approaching train when he sees that purple light blazing in the distance. The lighted fusee is as valuable by day as by night, for the smoke from it is so distinctive and even its light, as to be readily recognizable by a following engineer, and its presence is made evident, even around curves. On double-track, the fusee should be dropped out- side of the track upon which the train is running, and on divjpions of four or more tracks, it can be dropped in the middle of that track by hand, or perhaps more conveniently through a tube. In the substitution of mechanism for human agency in prevention of railroad accidents, much attention has been given to automatic train-con- trol, and particularly to devices for stopping a moving train independently of the action of the engineman. Such automatic stops ate actuated mechanically or electrically in setting the air-brakes. They are useful adjuncts in preventing a train from over-running a block-signal but are not appUcable otherwise for the protection of a train making an unex- pected stop at an unusual place.^ A device in use on the New York ■ Subway and on the Hudson and Manhattan tunnel-lines, is connected with the controller and with the brakes, and operates the moment that the motor- man releases his hold of the controller, as he can only keep the train in motion by pressing down a button in the top of the controller handle.^ Neither automatic blocks nor trip-signals are in use on the New York Elevated lines. It is asserted that their introduction on those lines would require increased headway, and lessen the number of trains by 25 per cent. Out of 3,000,000,000 passengers carried by these lines in the past decade, there have been but three fatal accidents. The few collisions that have occurred have been due to failures on the part of employees.' The requisites of installation for an automatic train-control, as approved by the American Railway Association, are given in Appendix VI, Note II. Responsibility foe Railway Accidents While railroad managements should not be held morally responsible for train-accidents occurring from causes beyond their control, they should be for such accidents when due to insufficient rules and regulations or to defects in track or equipment. Defects of the latter character are the causes of derailments, and the reports of the Interstate Commerce Com- ' The automatic stop or "trip-signal " in use on the New York Subway in con- nection with the block signals has failed to operate in a case where a collision oc- curred between two trains approaching a junction. 2 On July 5, 1911, a motorman on a subway train was killed by his head strik- ing a signal-post. In three oar-lengths, the train was brought to a stop with the current cut off and every brake set. ' From May 11, 1908, to Dec. 9, 1914, there had been 37 collisions, three passengers and an employee killed and 327 persons injured. 310 EFFICIENT RAILWAY OPERATION mission furnish appropriate statistics on this subject. In the year ending June 30, 1916, derailments from defective equipment were 4073 and from defective track, 1673. The character of these defects, as classified, is given in Appendix VI, Table III; and also the annual average for ten years which, as summed up, were From defects of equipment 3,378 From defects of track 1,516 Total 4,894 While no considerable diminution is shown in the number of derail- ments from these causes, it should be noted that there has been a marked increase in equipment and in train-mileage during this decade. In the same decade, the annual number of casualties from collisions and derail- ments averaged 12,789, and the average annual damage to road and equip- ment to $10,356,715. Automobile traffic has added materially to accidents at railroad cross- ings, and to casualties from other causes than from train-accidents. For the year ending June 30, 1914, 1147 persons were killed and 2935 were injured at grade-crossings. By far the larger number of such casualties were incurred by a disregard of the injunction to "Stop, Look and Listen" before attempting to cross the line. On the Baltimore & Ohio Railroad, in a record of crossing by vehiclies and pedestrians, out of 32,079 instances, in but 298 was this simple injunction observed. More than 18,000 persons took no notice of warning signals, even at crossings where trains were passing at intervals of five to ten minutes. Crossing-gates have not been sufficient for prevention of such accidents. On the Long Island Railroad, in the first eight months of 1915, there were 85 instances of automobiles that had been driven through closed gates. This situation became so serious that the American Railway Association appointed a Special Com- mittee on the Prevention of Accidents at Grade Crossings. This committee has united with a committee from the National Association of Railway Commissioners and the Executive Board of the American Automobile Association in efforts to establish uniformity of warning-signs and signals by State legislation, and for infliction of penalties for the disregard of such, warnings. It was further recommended that approach-signs be erected by the highway authorities at a distance not less than three hundred feet from grade-crossings. Gbade-crossing Difficulties. Electric-traction Dangers The prevention of casualties at grade-crossings has been given, promi- nence in the demand for the elimination of these crossings ; notwithstanding the evidence that such casualties occur entirely from disregard of warning- signals and closed gates. A more rational demand is based upon the interruption of highway traffic at such places. This interference increases. TRAFFIC 311 with the increasing volume of that traffic, and with the greater frequency of passing trains, until it becomes intolerable at street-crossings in large cities. Popular sentiment has been aroused against railroad manage- ments, and mandatory legislation has been sought for the eHmination of grade-crossings, without consideration of the conditions and circumstances under which they originated and stiU exist. In European countries, where an extensive highway system had preceded the introduction of railway transportation, provision could be made in advance for avoiding level crossings yet, even in England, they exist to this day in greater numbers than is generally thought to be the case. In the United States, railroad construction preceded the existence of high- ways in extensive areas, and roads were opened indiscriminately at grade, without regard to future exigencies. Under such conditions, it is mani- festly unjust that the entire cost of grade-separation should be borne by the railroad company. Where the highway had preceded the railroad, the claim is stronger than where the precedence had been the other way. In either case, it is properly a question of proportionate contribution at the public expense, and this view has generally prevailed where legislative action has been taken. The expense of grade-separation varies with the topographical con- ditions, as is shown by experience on the Long Island Railroad. Ten grade-crossings on the Winfield Cut-off were separated at a cost of $1 ,500,000. In the elimination of 95 street-crossings on Atlantic Avenue, Brooklyn, an expense was incurred of $6,400,000. The elimination of a single crossing in Pittsburgh cost $750,000. The estimated cost of eliminating 203 street crossings in Queens Borough was $12,000,000 ; the work to be completed in six years. The cost of eliminating all grade-crossings in New Jersey was estimated at $250,000,000, and of eliminating those still remaining on the lines of the Pennsylvania Railroad Company at $600,000,000. With this information, a conception may be formed of the capital in- vestment required to eliminate all the grade-crossings in the United States. In 1915, there were 255,606 grade-crossings, including 14,913 of one rail- road by another. In that year, there were eliminated 61 crossings of railroads and 466 of streets and highways. The unprotected crossings of steam-railroads numbered 40 per cent, of the total of such crossings ; of electric railroads, 46 per cent., and of streets and highways, 91 per cent.^ Every crossing of steam-railways will sooner or later be the scene of a collision, unless it be protected by interlocked signals. Every crossing of electric railways should be protected either by such signals or by a self- acting gate across one of the lines. Where a public highway crossing is not covered by a watchman, there should be a crossing-alarm. Farm-road crossings should be closed by self-acting gates. In view of the appalUng casualties on railroad tracks, such precautions are of far greater importance ' For Grade Crossing Statistics, see Appendix VI, Table IV. 312 EFFICIENT RAILWAY OPERATION to the general welfare than the expensive elimination of 240,693 highway crossings at the rate of 466 per annum. Some reference may be made to accidents in railroad operation by electric traction. The greater number of injuries to persons and damage to property have been caused by a short circuit of electric current. They occur more frequently, where the third-rail is used as a conductor, from some piece of metal falling across the rails. The concurrent establish- ment of a short circuit has added to the consequences of collisions and derailments.! Apprehension has been expressed as to the effect of the accidental charging of the metalhc surfaces of all-steel equipment, though there are no official reports of casualties from this cause. Contact with overhead-wires by train-employees on car-roofs has been a specific cause of fatal injuries. On lines operated in this manner there is less protection from overhead-obstruction, as "telltale" guards can not be used with overhead trolley-wires.^ Expedition of Passenger Service The second factor in importance in Traffic Efficiency is expedition. It is superior expedition that gives value to railway transportation, but this value differs with the purposes for which it is desired by those who avail themselves of it. These persons may be distinguished, as suggested in the discussion of station-accommodations, as commuters and as travelers for longer distances. To the commuter, the railway journeys between his home and his place of business are a part of his daily routine. His every movement is collocated to the minute with the railway time-table. Consequently, even a shght derangement of the train-service on which he depends, causes a serious disturbance in his everyday life. This con- dition extends to communities within the commuting zone. Their pros- perity is affected by the. relative frequency of the suburban service, and also by its comparative promptness. This fact is appreciated by the managers of railroads with a heavy suburban traffic, who exert themselves to insure prompt train-service and contribute to the comfort of commuters by specially reserved or "club" cars and, since 1889, by "club trains" on the railways centering in London. The long-distance travel may be subdivided into that which is confined to journeys within the terminals of a line and that which extends to those terminals or beyond them. The latter class of "through" travelers is ordinarily affected, as to expedition, by competitive conditions. Prompt- ' In a rear-oolUsion on the New York Elevated lines, a car caught fire from the third-rail. Two men were killed, and eleven seriously and seven slightly- injured. "As to train-accidents, see also, "American Railway Management," pp. 16, 66 and 227; "Restrictive Railway Legislation," p. 126; "Railway Cor- porations as Public Servants," p. 144; "Problems in Railway Regulations," pp. 279, 305 ; and as to grade-crossings, p. 274. TRAFFIC 313 ness is not so much a matter of minutes, as in suburban service, but of hours, and it is secured for journeys over connecting lines by continuous- train-service. " The traffic within the terminals of a line is characterized as "local," and local passengers usually receive less consideration as to expedition thasn either commuters or "through" passengers do. Transportation- officials endeavor to restrict train-mileage by using the through trains for local service at important way-stations. As a consequence, the business and social relations of such communities with each other are often seriously incommoded by derangements in train-service that occur far away from their vicinity. The minor stations are more fortunate in this respect. Though their ' train-service may be less frequent, it is not affected by such contingencies, as they are served independently by local trains. On a double-track line, this local service is. usually conducted with regularity, but on a single-track line, it is often delayed by the preference given to through trains. These less-favored communities are therefore resorting to the public highways in motor-cars, to an extent which is affecting railroad revenues ; and there should, therefore, be provided a more frequent and more punctual service for this class of travelers. This requirement might be economically fulfilled by a more liberal use of mixed trains between fixed terminals. The trains should be made up of a combination baggage and smoking car and two coaches, attached to six or eight box cars on standard freight-trucks, running on a 25-mile- an-hour schedule. These cars should only be loaded with less-than-car- load lots of merchandise, consigned from one terminal to the other. By this means, the cost of the passenger-service should be sensibly reduced and the loss of business to motor-cars as well ; while the frequent and speedy .transmission of small lots of merchandise would likewise be of benefit to local-business interests. Irregularity in train-service has been diminished by the use of more powerful locomotives, by more thorough inspection of equipment and by more careful lubrication of axle-journals, with fewer delays from hot boxes.* On single-track lines, passenger-trains are often delayed by congestion of freight-trafiic, with consequent loss of time at passing-points, but this should not occur on double-track. There is some statistical information available as to the comparative regularity of train-service. The Public Service Commission of the State of New York reported that, in September, 1916, out of 70,168 trains operated in that State, 76.4 per cent, arrived at destination on time. The average delay for each late train was 24 minutes, and 5.7 minutes for each train run. This statement, however, included suburban-trains to a large extent, with comparatively few and slight delays. ' On the New Haven line, there were but eighteen cases of hot boxes on pas- senger-trains in a week, with a daily car-mileage of 240,000 miles. 314 EFFICIENT RAILWAY OPERATION 46 per cent, of the delays were due to waiting for connections, 3 per cent, to wrecks, 2.7 per cent, to poor track and 2.6 per cent, to locomotive failures.^ High Speed in England and the United States High speed is of less importance than regularity as a measure of traffic efficiency. Continuous high speed is of greater value in long-distance journeys, and is stimulated by competition. The rivalry for the summer- travel between London and the north of Scotland reached a point, in 1895, at which the journey by one route of 523|^ miles was made in 518 minutes, and by another of 539f miles in 512 minutes ; being an average speed respectively of 60.64 miles an hour and of 63.25 miles. The time consumed was abbreviated by long runs without stopping. In 1903, non- stop runs were made of 150, 175 and 185 miles. All previous records were eclipsed by a regular train-service on the Great Western Railway, between London and Plymouth, 245f miles in 233|- minutes without a stop, being at the rate of 63.1 miles per hour. Competition between rival lines has had a similar effect in the United States. In 1910, the trains between Camden and Atlantic City covered the distance of 55^ miles by the Philadelphia & Reading Railroad in 50 minutes, and by the Pennsylvania Railroad, the distance of 59 miles in 52 minutes, being respective speeds of 66.6 and 68 miles an hour. Such speed in regular service has not been attained on longer journeys. In 1910, the 18-hour schedule between New York and Chicago, over the New York Central route of 979 miles, was at the rate of 54.4 miles an hour. The distance by the Pennsylvania Line is 70 miles shorter, but under less favor- able conditions as to intermediate gradients and curvature. The time by both routes has since been lengthened to twenty hours with an additional charge for passage by these trains, which is rebated when a train is over an hour late at destination.^ Continuous-train-service at high speed for long distances over a single track was inaugurated with the "New York and Florida Special," between New York and Jacksonville, Fla., on January 9, 1888. The equipment consisted entirely of Pullman vestibuled drawing-room sleepers, a buffet dining-car and an observation car. The train ran tri-weekly, leaving New York at 9.30 a.m.. Eastern time, and arriving in Jacksonville at 3.40 P.M. the next day on Central time, or virtually in 29 hours and 10 iThe "Broadway Limited" train, on the Pennsylvania Lines between New York and Chicago, is scheduled to make the run of 909 miles in 20 hours. Dur- ing the first six months of 1915, this train was on time on 92 per cent, of the trips. 2 In June, 1905, the run of 525 miles from Buffalo to Chicago, over the Michigan Central Railroad, was made at the rate of 69.69 miles an hour, excluding stops. In March, 1901, a train on the Plant System between Jacksonville and Savannah, made a run of five miles in 2i minutes, or at the rate of 120 miles an hour, which is the highest speed on record for a regular train. For comparative speeds of long-distance trains, see Appendix VI, Table V. TRAFFIC 315 minutes. The distance on that run was 1084 miles, and the rate of speed averaged 37.17 miles an hour, with no second track between Washing- ton and Jacksonville, 860 miles. Now, about ten such trains serve the Florida winter-travel from north of the Potomac and the Ohio rivers. The fastest service of this character is rendered between Washington and Jacksonville. By one route, the distance of 755 miles is made, including stops, in 18 hours and 35 minutes, at the rate of 40.8 miles an hour ; and by another, 792 miles in 17 hours and 55 minutes, at the rate of 44.2 miles an hour. There are now, however, considerable stretches of double-track on each of these lines. American Methods. Sleeping-cak and Dining-car Service Convenience and comfort on railway journeys have been greatly fur- thered since the days when the railway carriage was merely a combina- tion of stage-coach bodies; and nowhere else to such a degree as in the United States. As mentioned in the chapter on Rolling-stock, the American passenger-car was not developed from the stage-coach, but from the long wagon-body in general use in this country. The long passenger- car, with end-platforms and supported on swivelin'g trucks, has been the fundamental principle in the differentiation of American railway operation from that which originated in England and was adopted in other European countries. It has had a world-wide influence in the subsequent develop- ment of methods and appliances for the welfare of the traveling public as well as for economic operation. The car with end-entrances affords f aciUties for free communication through the train, which was impracticable in the English railway carriage. Acts of violence committed in closed compartments are prevented, the comfort of passengers with regard to the common decencies of life is made possible, and methods of lighting, heating and ventilation can be applied more efficiently.' The European class-distinction in railway operation has not been definitely observed in the United States. The difference in names be- tween the railway carriage and the passenger-car is significant of the difference in social environment. The smoking-car, intended as a relief for passengers to whom smoking is offensive, has virtually become occu- pied by those who, in other countries, would have been third-class pas- sengers. There is something of the first-class distinction in the occupants of sleepers and parlor-cars, leaving the ordinary thoroughfare cars, or "day coaches," for those who, in European countries, would be considered as second-class passengers. Competition on some of the Western roads has even resulted in furnishing somewhat similar accommodation of a cheaper character in "tourist sleepers" and "reclining-chair cars." The fourth- class accommodations on German roads have no counterpart in this country, where even the immigrant passengers are far better accommodated. 1 See Chapter IV, pp. 132-139. 316 EFFICIENT RAILWAY OPERATION The introduction, on British railways, of communication through the train by the use of compartment-carriages with a side-passage way, or "corridor-cars," has exercised an influence toward diminishing class- distinctions by doing away with second-class passenger-fares. As a con- sequence, the third-class service has been so much improved that it is quite comfortable even for long journeys ; as between London and Scot- land, where third-class passengers are provided with dining-car service. In democratic Switzerland, the American open car is in general use. Cars fitted with permanent sleeping-berths, originated in the United States in 1858. Poorly equipped at first, they were developed by George M. Pullman in a car suited also for travel by day, and in a style that was never before attempted.' On roads in Continental Europe, the sleeping- cars are corridor-cars, with two berths in a closed compartment and a toilet-room between every two compartments. This arrangement is more objectionable than the open Pullman car, if the occupants of the same compartment are strangers to each other. The standards estab- lished by Mr. Pullman in the construction, decoration and furnishings of passenger-train equipment have conduced to the welfare of travelers by rail throughout the world. The Pullman car-service has also benefited, them in such matters as convenient lavatories and water-closets. The identification of train-employees by a distinctive attire had long been customary on European roads before attempts were made to in- troduce the custom into this country. This innovation, so important to railway travelers, was not generally established until it was made compulsory in the Pullman service. The provision of iced drinking- water with clean glasses, in the Pullman cars, incited the crusade in recent years against the use of drinking-cups in common, and suggested the general introduction of individual cups of water-proof paper, to be used but once and then destroyed. The presence of an attendant in the sleeping-car afforded the oppor- tunity for serving simple refreshments to its occupants. From this be- ginning, there has been developed an elaborate dining-car service which has followed sleeping-car service across the Atlantic ; as well as luxurious parlor-cars, caf^-cars and observation-cars, with the auxiliary service of barbers and stenographers. The first sleeping-car with kitchen and pantry, called a "hotel-car," was placed in service in 1867 on the Great Western Railway in Canada. The first regular dining-car, the "Del- monico," was operated on the Chicago & Alton Railroad in 1868.^ In Great Britain, light refreshments were served in the Pullman parlor- cars, which were introduced on the Midland Railway in 1875. The first restaurant-car was placed on the Great Northern Railway in 1879. These 1 PuUman's first car, the "Pioneer," was used on President Lincoln's funeral- train, in April, 1865. "When Raiboads Were New," p. 175. 2 "When Raiboads Were New," p. 178. TRAFFIC 317 cars served only first-class passengers, as there was no communication through the trains until the corridor-car trains were put in service on the Great Western Railway in 1892. Restaurant-car service was then ex- tended to second-class passengers, and on the Great Eastern Railway, in 1893, to third-class also; but not on the Great Western Railway until 1903, where, in 1914, there were 88 trains with daily dining-car service. In that year, this service was provided on 457 trains in the British Isles, in the summer-season, and on 403 trains in the winter. Many of these cars are operated on routes of less than two hours in duration, with meals including Hght breakfasts and afternoon tea.^ The American plan of "first come, first served" results in irregular and unsatisfactory service. In a well-filled train, the passage-way to the dining-car may be filled with hungry people, impatiently waiting for seats to be vacated by others who are leisurely finishing their repasts. On European roads, there are two dinners in courses, served one hour apart, and passengers are supplied in advance with tickets for reserved seats at one or the other of these hours. The dining-car service is con- fessedly unprofitable in itself. It is still more so, if consideration be given to the additional car-mileage and accompanying cost of operation, in comparison with the small percentage of the total number of passengers who are served by it. A steel dining-car, weighing 160,000 pounds and costing perhaps $40,000, with a crew of twelve persons, will seat from 30 to 48 passengers, and not infrequently there are two dining-cars in a train.^ It is a difficult matter to maintain scrupulous cleanliness in a dining-car that is in regular service and is perhaps dropped at a way-station tem- porarily, to be taken on by a returning train. It is questionable whether extensive catering to luxurious habits does not unreasonably affect economic operation. The equipment of caf6-cars, hbrary-cars and observation-cars is even a more wasteful investment of capital, as they add nothing of substantial benefit to the welfare of the great body of travelers, and are of doubtful value for advertising purposes.^ There are exceptional cases in which local conditions may warrant such service; as over a route devoted principally to high-class season travel. FaciUties for the rest, refreshment and relaxation of travelers on a long railway journey might be afforded by a half -hour stop at a station where all the passengers could obtain meals well served in a comfortable dining- room at moderate prices, including a lunch-counter and basket-lunches. ' During 1915, the Southern Pacific Company ran dining-cars nearly 11,000,000 miles and fed 3,207,000 persons. 2 On the New Haven line, in October, 1914, twelve dining-cars, valued at $ 400,000, made 175 round trips with a total mileage of 88,700 miles. In that time, 33,440 meals were served, exclusive of 7285 free meals to employees ; being an average of 96 passengers per trip. " Passenger Terminals," p. 369. 3 The "Limited" trains between New York and Chicago are said to average about one hundred passengers per trip. On this basis, the weight of the train, exclusive of the locomotive, is about seventy-five times that of its pajring load. 318 EFFICIENT RAILWAY OPERATION This would also afford aft opportunity for a little exercise in the open air while the cars were well aired for a continuance of the journey. Sanitation The occasional freshening-up of a closely occupied passenger-car may be associated with a recognition of the germ theory of diseases, and with the attention that should consequently be given to the sanitation of all travehng-conveyances. It is a fundamental principle of that theory that the germs of disease thrive under conditions of combined heat and moisture, that they are merely rendered dormant by severe cold, and that they are only devitalized by complete exposure to dry, hot air. Sweeping and dusting are worse than useless, as they disseminate the seeds of disease instead of exterminating them. This is also true of the use of a blower with compressed air; which is being superseded by the vacuum-exhaust cleaner, in which the dust is collected to be burned. Perfunctory spray- ing with germicide-solutions has been abandoned in cleaning Pullman cars. The berths are opened up, and bedding and drapery arranged for proper exposure in a temperature of seventy degrees. The cars are then fumigated with formaldehyde in combination with permanganate of potash, heated to a gas in paraffine burners, and the car then kept closed for three hours. "^ In Germany, the carriages in the trains de luxe, which ran to the Russian frontier, were disinfected in a cylindrical compartment of steel tubing, in which a carriage was entirely inclosed. By means of piping along the inner walls, the' temperature was raised until the cushions exposed in the carriage had been heated to 110° F. A partial vacuum of about 28 inches was then produced by an electrically-operated air-pump. When more powerful disinfection was required, connection was made with a chamber containing formaldehyde vapor, insuring complete sterilization. The temperature can be raised to 122° F. without injury to the finish of the carriage or to the textile stuffs. The disinfection by heat alone was accomplished within two hours, or in five hours with formaldehyde vapor .^ Station Conveniences and Lx^xukie^ Provision for the comfort and convenience of travelers is made at many of the terminal stations in the United States on a scale equally as luxurious as on the limited trains. It is well to provide suitable waiting- rooms and lavatories, properly lighted and warmed, with information- bureaus, carriage-agencies, telegraph and telephone facilities, newstands and uniformed porters but the palatial decorations and trivial accessories of some of these terminal stations border upon extravagance. At the Grand Central Station there is a room for women to have their shoes polished by girls in livery, a "hair-dressing parlor" with walls and ceiling 1 "Passenger Terminals,'' p. 250. 2 Bulletin International Railway Congress, Brussels, November, 1912. TRAFFIC 319 of "Carrara glass," and also a manicure parlor, private dressing-rooms for changing from a traveling-dress to evening costume, with attendants, which can be reserved by telegraph. There are similar provisions for men, and a public barber-shop, that is said to represent an investment of $100,000. The emergency hospital, with a physician in attendance, is more worthy of imitation. There are such accommodations also in the Union Station at Washington, and, in addition, private bathrooms, a Turkish bath, a swimming-pool and several "mortuary chambers." Ticket Systems Material changes in the manner of ticketing passengers were brought about with the increasing volume of travel. In early railway practice in England, the famiUar methods of stage-coach traffic had been continued in assigning each passenger by name to a definite seat in a carriage, of which there is still a reminder in the use of the term "booking-ofiice" for ticket-office. As there was no communication between the carriages, tickets were collected from the outside running-boards and, as late as 1895, express-trains to London were stopped for this purpose just before enter- ing the terminal station. In the United States, tickets were at first only issued between principal stations and might be bought or not, at the will of the traveler. Much of the revenue from the passenger traffic was collected by the conductors.^ For continuous journey over consecutive roads, there were no interhne- tickets. Before the consolidation of the line between Albany and Buffalo, a traveler between those cities was required to obtain tickets at Albany, Schenectady, Utica, Syracuse and Rochester ; to claim his baggage at the end of each intermediate change of cars, and to provide for his transfer across the town to the other station. Not until 1853 were through-tickets on sale between New York and Chicago, and not until 1875 could a through ticket be purchased from New York to Jacksonville, Fla. Local tickets were printed on cardboard, with neither number nor date, and were reissued until the ticket-punch was introduced in 1856. Interline-tickets in coupon-form came into general use when the through- fares had been reduced by competition below the total of the intermediate local fares. For the same reason, other forms of tickets were issued, for round-trips, for excursions and for immigrants, also return-tickets, season- tickets, mileage-tickets and half-fare tickets, in such variety that from 500 to 2000 forms are now on sale in the principal ticket-offices.^ Suburban travel was stimulated by the issue of season-tickets, but for auditing pur- poses, the issue of commuters' tickets in strips is preferable, as the season- ' On the Michigan Central Raih-oad, the cash-collections on a round-trip between Detroit and Chicago often amounted to $1400, and even to $1800. 2 Half -fare tickets for clergymen were issued over the Erie Railroad as early as in 1843. "When Raih-oads Were New," p. 88. 320 EFFICIENT RAILWAY OPERATION ticket affords no indication of the mileage made with it and, if it is to be shown each time that it is used, it is of no advantage to the habitual com- muter. For the convenience of commercial travelers, a special form of ticket was issued on a mileage-basis. This privilege has been extended to include several persons in a party and to cover connecting-lines, until the mileage-traffic now contributes appreciably to passage-receipts. The exhibition of tickets at the entrances to train-platforms is of benefit to travelers, where there are frequent train-departures at terminal stations, as it insures their direction to the proper trains. The obstruction by turn- stiles at these entrances is objectionable to passengers with children and hand-luggage, and is not necessary for the exclusion of loiterers. The admission of persons accompanying passengers to the trains can be regu- lated by the issue of platform-tickets. These -tickets produce a consider- able revenue at important stations, and are conveniently distributed from automatic machines adjacent to the platform gates, which were introduced in 1906 on the Metropolitan Railway in London.' Restricting the entrance to platforms to ticket-holders is not as com- mon in the United States as in Europe. Its extension to way-stations in general is unnecessary for travelers, and would require a very considerable increase in the number of station-employees. For this reason, even in Europe, it is not enforced at minor stations and, in Switzerland, only at very busy stations. In Italy, tickets are not taken up on the trains, but are exhibited on entrance to the train-platforms at terminals and to the waiting-rooms at way-stations; and are surrendered on leaving the plat- forms. But wages are low there, and employment on the railways forms an important part of the political patronage. Long-distance travelers have been much benefited by the establish- ment of city ticket-offices in the interest of competing lines. In Europe, this service is more commonly rendered by tourist-agencies, which are indispensable to travelers unfamiliar with foreign languages and customs. The agencies provide tickets for any desired route, with information in several languages with regard to luggage, custom-house inspection and other requirements. The tickets are arranged in consecutive leaflets in book-form, which is preferable to our cumbrous coupon-forms for long journeys. For the establishment of this invaluable service, travelers are largely indebted to the world-wide agencies of Thomas Cook & Son, whose uniformed representatives are at hand at many important railway terminals in Europe, prepared to give assistance to whomsoever may desire it. In this respect, Thomas Cook has rendered a service to travelers similar to that for which we are indebted to George M. Pullman in the matter of car-service. At the principal terminal stations in Germany, seats in the 1 On the Prussian Railway system, 31,000,000 platform-tickets were issued m 1910 ; an average of 85,000 per day, yielding a revenue of $735,000. ' ' Passenger Terminals," p. 286. . TRAFFIC 321 ordinary carriages could be reserved a day or so in advance, at a slight additional charge. The carriage-seats were numbered and then labeled "reserved." Such concession to persons traveling in a party, might be made in this country, instead of confining it to the Pullman service. Baggage Handling Handling and forwarding baggage, like the issue of railroad-tickets, was stimulated in its development by competition between rival Unes for the patronage of travelers on continuous journeys for long distances. English railway carriages were originally built with a baggage-compart- ment between the passenger-compartments, similar to the boot of a stage- coach, into which a porter placed the luggage of a passenger as he took his seat in the same carriage. To those who are accustomed to the Ameri- can method of checking baggage, it seems remarkable that it should be delivered at destination to whomsoever may claim it without further identification; yet, in London, this course is still pursued on local trains, even at terminal stations. It has its advantages in the speedy delivery of luggage, and with little probability of its being wrongfully claimed. In American railway operation, some other means of identification became necessary, since train-baggage is carried in a separate car, identi- fication being provided for by means of brass checks in duplicate, bearing serial numbers and the initial letters of the corporate title of the railroad company. Formerly, one of these duplicates was delivered to the passenger and the other attached to the piece of baggage by a leather strap. This plan required the resorting of the duplicates at the end of the route and their return to the issuing station. With increasing volume of travel, this return became so onerous that card-duplicates were substituted, one of which could be slipped into a receptacle on the strap, and were not re- issued. The use of card-checks made it unnecessary to keep a large stock of duplicates on hand as, by filling in the destination on a blank space, they were applicable to delivery at any station on any route, and also at any hotel or private residence, through the intervention of local-transfer agencies. The strap-check is open to the objection that a duplicate may be mis- matched or attached to the wrong piece of baggage, either through care- lessness or with fraudulent intent. For this reason, when the American system was introduced abroad for continuous interline-journeys, the plan was changed to the so-called "registration" of luggage by the use of a paper label pasted on each piece. These labels are duplicated for iden- tification and are registered on stubs in the books from which they are taken. The same number is repeated on several labels, which may be pasted on a number of pieces belonging to one passenger. The pieces so numbered are placed- together on reaching their destination, to expedite delivery. This plan is an improvement on the strap-check, as it prevents 322 EFFICIENT RAILWAY OPERATION the mismatching of dupUcates. In view of the number of pieces handled at a busy terminal, it is creditable to the baggage-men that so few pieces go astray.^ The free transportation of baggage, which originated in England as a privilege accessory to the stage-coach fare, became also customary in the United States, until the abuse of the privilege brought about a limita- tion as to weight, beyond which a charge is made proportionate to the excess. In Continental Europe, this privilege is not granted, except, to a lesser extent, in Switzerland. As a consequence, the quantity of hand- luggage is often augmented beyond the capacity of the racks in the car- riages, and to the inconvenience of their occupants. With five times the density of passenger-traffic in the United States, the Prussian railways in 1910 carried only 820,000 tons of baggage which, at 150 pounds apiece, is only about four times the annual average at the South Station, Boston. It has also been found necessary in the United States to limit the di- mensions and weight of articles accepted as personal baggage. This limit is fixed at 45 inches in any one dimension, and at 250 pounds as the maxi- mum weight which can be safely handled by one man. Miscellaneous articles, as musical instruments, bicycles and baby-carriages, when as- sociated personally with a passenger, are accepted as baggage, unless by reason of their number or size, they obstruct the regular service. This has become the case with the baggage, properties and scenery of traveling theatrical companies which, like circus-companies, are consequently trans- ported by special trains and at hours suited for making one-night stands. The storage of undelivered baggage at important stations becomes a serious burden upon the railroad companies, whose responsibility as a common carrier is thereby changed to that of a warehouseman. The extent of this responsibility may be inferred from the floor-space provided in some terminal baggage-rooms. -The hmit of responsibility as a common 'At the Union Station, St. Louis, the annual average is about 1,500,000 pieces. During the World's Fair period, ending June 30, 1904, there were handled nearly 2,400,000 pieces, an average of over 6500 pieces daily. At the Kansas City Union Station, the annual average exceeds 2,000,000 pieces, and 2,500,000 pieces at the South Station, Boston. "Passenger Terminals," Droege. Loss AND Damage to Bagqaqb. United States Roads in Class I. — 1912-1914 Yeah Eastern District Southern District Western District United States 1912 1913 1914 $108,781 112,471 124,779 $56,044 68,916 61,993 $130,398 115,335 110,092 $295,223 296,722 296,864 Average $115,344 $62,318 $118,608 $296,270 TRAFFIC 323 carrier is definitely terminated by making a charge for storage at the end of twenty-four hours. The temporary care of hand-baggage at a passenger-station had been a virtual perquisite of the newsstand, until its value as a source of revenue became so apparent that it was rented as a parcel-room privilege, or re- tained by the owner of the station. The parcel-room at the South Station, Boston, is open night and day, with eight employees and a rack-capacity for 1680 pieces. The parcel-room also serves as a depository for lost articles which, at the Pennsylvania Railroad Station in New York City, numbered 3630 during the first seven months that it was in operation.' , Express and Mail Business The transportation of light and valuable articles by passenger-trains was first undertaken in the United States by Alvin Adams, who traveled back and forth between Boston and New York as a private messenger, carrying valuables for banking-houses in his hand-baggage. From this unobtrusive beginning, Adams extended his field of action by contracts with transportation-lines until he founded the Adams Express Company, which, with its subsidiaries, monopohzes the express business on important lines from the New England states to the Mississippi River and the Gulf of Mexico. The success of the Adams Express Company led to the or- ganization of similar companies on other trunk-line routes, in several instances by influential stockholders of the railroad companies. Other railroad companies endeavored to carry on the express-business as part of their regular train-service but, in no instance, has this been successful. The necessity for an organization corresponding to the vast extent of territory served, is shown in the control by three companies of the whole express-service in the United States. The business so far exceeds the capacity of the regular passenger-trains that a large part of it is carried on special trains. The express companies have also entered the field of international banking. In 1907, they issued money-orders and travelers' 1 Area op Baqgage-rooms and of Parcel-rooms Station ClTT Bagqaob-rooh Sq. Ft. Paboel-boom Sq. Ft. Union Penn. R.R. Union Union Union South Detroit New York . Washington St. Louis Kansas City Boston . 50,000 41,683 42,000 74,648 27,794 66,650 750 2,780 1,020 1,475 2,040 2,100 C & N W R. R. Chicago . . 1,456 "Passenger Terminals," p. 22. 324 EFFICIENT RAILWAY OPERATION checks to the amount of $168,000,000, which was equal to 30 per cent, of the postal money-order business in that year. The prosperity of the express companies has been largely due to their making a house-to-house delivery. In England, this had been customary with the carriers on the public highways before the advent of railways. So, when the colliery-railways took over the carriage of miscellaneous goods, they conformed to this custom and entered into a community of interest with the common carriers, which left no field, as in the United States, for express companies. Nor was there the same extensive area for operation over disconnected railroad-lines. Merchandise-trains cover the greater part of Great Britain in a single night, and goods taken from the warehouse of a London merchant in the afternoon are delivered at the shop of his customer in Scotland during the following forenoon. Special provision is made for the carriage of such valuables as are not intrusted to the post. There was developed a considerable mileage of railways in Great Britain before they began to be utiUzed for the dispatch of mails. In 1838, the British government introduced a Bill into Parliament to require railway-companies to carry the mails at such hours and at such speed as the Postmaster-General might direct; authorizing the Post Office De- partment to use its own locomotives and carriages ; to carry passengers on its trains and to keep the lines free from interference with their move- ments ; "a fair remuneration" to be paid for the use of the tracks. This measure was withdrawn, however, and contracts for the transportation of the mails were amicably arranged. In the same year (1838), the United States Post Office Department contracted for railway mail-service between Washington and Philadelphia. The mails were dispatched in locked bags accompanied by a messenger. Afterward, a compartment was assigned to the messenger, or "mail agent," in the baggage-car, which was fitted with pigeon-holes for sorting mails received at^ay-stations^^^^^parate post-office cars were in use between Boston and Al^m5,ajfl852. The rapid express-trains were known as "The Fast^^'ii;^!fid mail-bags from way-stations were caught from mail- cranes as^hese trains passed at full speed. With increasing population, the volume of mail assumed such proportions that, on the trunk lines, the locked bags and paper-mail for terminal destinations were transported in storage-cars and eventually by special train-service.' ' The inclusion of a parcel-post in the railway mail-service is causing an increase in the volume of mail which is likely to greatly delay the regular train-service, unless it is taken by special trains. ' The first special postal-train in England was placed on the London & North Western Railway in 1885. In 1915, in the Third Postal District of the United States, which includes all the States between the Ohio and the Mississippi River with Iowa and Missouri, a daily average of 2862 tons of mail was carried on 4000 trains. At South Station, Boston, 250 tons of mail are handled daily, and from 225 tons to 300 tons daily at the Union Station, Kansas City. TRAFFIC 325 Variations in Volume of Traffic Under certain social conditions, passenger-traffic may be conducted with close approximation to economic efficiency; that is, with vehicles loaded to their normal capacity and with train-resistance in economic ratio to tractive-power. This is especially the case with excursion-travel, which permits of passenger-train service in its most economic aspect. Commutation-travel is another example of possible economic efficiency, though it has the disadvantage of almost empty train-service in one direc- tion, during the morning hours, and in the other direction in the after- noon. Emigrant-travel affords long hauls for fully loaded trains with minimnm requirements as to speed and convenience, but also with empty mileage in one direction. But, as a general proposition, it is more difficult to combine economic with social efficiency in passenger-traffic than in freight-traffic. Empty cars and lightly loaded trains in one direction is the rule rather than the exception, varying even more with seasonal changes than freight-traffic does. In winter, as in summer, there is a rush of travelers seeking a change of climate, like migrating flocks of birds; at the beginning, few in number, but gradually increasing until the maximum volume in one direction has been attained. Then, by degrees, the direc- tion of increasing and diminishing travel becomes reversed. This class of travelers demands facilities for comfort of a very luxurious character. Sleeping-cars, dining-cars, caf^cars, observation cars, all electrically lighted, have to be furnished throughout the vicissitudes of the season, and the ratio of dead weight to paying load is greater than in any other class of train-service. The relative density of both passenger and freight traffic, per mile of line of our railway-system, for the twenty-five years ending in 1914, is graphically illustrated in Appendix VI, Table VI ; in which the effect upon that traffic of the financial disturbances in 1893, 1903 and 1907 are plainly visible. It is interesting to compare the relative situation at the beginning and at the end of that period. The relative increase of traffic, as compared with the increase in line-mileage, is indicative of the marvelous prosperity of our country.' 'Density of Traffic, United States, per Mile of Line 1889 1914 Inc. Peb Cent. Passenger mileage . . Ton mileage .... Miles operated .... 75,325 448,069 157,759 144,278 1,176,923 247,397 91.5 162.6 56.8 Minor Lines (6,695 miles), excluded in 1914. 326 EFFICIENT RAILWAY OPERATION The importance of adequate provision for the safety, expeditious movement and convenience of travelers as an element of Railway Efficiency and as affecting the general welfare, is made evident in the statistics of passenger-traffic on the railway-system of the United States.^ Measured by passenger-miles, the volume of traffic nearly trebled from 1890 to 1912, with an increase of 45 per cent, in railroad-mileage. From 1912 to 1914, it increased about 10 per cent, with an increase in railroad-mileage of .027 per cent. The actual number of travelers was less than the number given in these statistics, for there is a duplication of passengers traveling over two or more fines. But, accepting these figures as a basis for the com- parative density of traffic, it appears that the annual rate of increase in the number of passengers per mile of line has been as follows : 1890 to 1900 754 per mile 1900 to 1910 5,098 per mile 1910 to 1912 835 per mile (decrease) 1912 to 1914 4,954 per mile In the period 1910-1912, the railroad-mileage seems to have been ex- tended so rapidly in proportion to the increase in the volume of traffic as to result in a considerable decrease in its density. This decrease occurred only in the Western Traffic District, where the line-mileage increased by 3276 miles from 1911 to 1912, with a decreased traffic that resulted in decreased density of 5257 passengers per mile of fine. The apparent decrease of railroad-mileage from 1912 to 1914 is due to the exclusion from the statistics for that period of the operation of minor fines with a mileage of 6695 miles from a total mileage of 256,547 miles. The traffic in the Eastern District constitutes about 61 per cent, of the total on our entire railroad-system; although it contains but one-fourth of the fine-mileage. The density of its traffic is two and a half times that in either of the other districts. ' Passenger Traffic, United States No. Passengers (millions) . . Passenger-miles (millions) . Passengers, per mile of line . Railway mileage Passenger train mileage (mil- lions) Passengers per train-mile . Average journey, miles . . 1890 492 11,847 75,751 165,936 1900 576 16,039 83,295 193,345 1910 . 971 32,338 134,279 240,831 549 58.9 33.50 1912 1,004 33,132 132,608 249,852 585 56.5 33.18 Mileage of minor lines excluded in 1914. See also Appendix VI, Tables VI-XI. 1914 1,053 35,258 142,516 247,397 602 58.5 33.61 TRAFFIC 327 The average passenger-train is made up of a mail car, a baggage and express car, three standard passenger-cars and a Pulhnan car, with a seat- ing capacity for about two hundred passengers. In the five years ending in 1914, the average number of passengers per train-mile has not varied materially from fifty-eight persons. On account of the duplication al- ready mentioned, the actual average has been somewhat less. The average seating-capacity of a train is therefore over three times the average train- load of passengers. From 1910 to 1914, the length of the average journey, including du- plications, has been about 33 miles for the whole system. But, confining the average to roads in Class I (with annual operating-revenues above $1,000,000), the average in 1911 and in 1914 was the same, 34.49 miles; being between 26 and 27 miles in the Eastern District, and between 50 and 52 miles in the Western District. This difference is due to the fact that the great volume of suburban travel is in the former district. Its effect is shown in a comparison of the total traffic in the Eastern District with that on the Long Island Railroad.^ With 0.67 per cent, of the total fine- mileage of roads in Class I in the Eastern District, the Long Island Railroad carried 6.8 per cent, of the total number of passengers. As a consequence, its passenger-mileage per mile of line was nearly six times greater than the general average. Density of Traffic From these statistics, it may be inferred that the density of railroad passenger-traffic will hereafter increase but slowly, and will be restricted in the future, as in the past, by the competition of electric roads. In 1908, the electric roads of Massachusetts, urban and interurban, had a mileage of 2233 miles, which was greater than the mileage of its steam-roads. In that year, the electric roads carried about 600,000,000 passengers, or four times as many as the steam-roads carried. They made three times the car-mileage with less than twice as many passenger-cars. Their revenue from passenger-traffic was nearly $30,000,000, or about three-fourths of the steam-roads' passenger-earnings with much longer average haul. The 1914 Long Island Railroad Class 1. Eastern District • Mileage of line . . No. passengers . Passenger-miles Passengers per mile of line Passengers per train-mile Passengers per car-mile Cars in train 398 41,570,612 602,787,853 1,512,718 110 27 4.52 14.5 58,667 608,647,324 16,348,655,263 279,975 65 17.3 5 6 Average journey, miles 26.86 328 EFFICIENT RAILWAY OPERATION passenger-traffic of both steam and electric roads is now being affected by the increasing use of motor-cars upon the public highways. In 1916, there were about 3,250,000 automobiles in use in this country, with an average capacity for three passengers at 5000 miles per annum. This potential passenger-mileage of 48,750,000,000 miles per annum may be compared with electric-road mileage, estimated in 1912 at 38,000,000,000 miles, with an average journey of 4 miles ; and with steam-road passenger- mileage in 1914 of 35,258,000,000 miles, with an average journey of 33.61 miles. Statistics of the density of passenger-traffic in 1914 on the railway- system of the United States are given in Appendix VI, Tables VI to XI, and in Appendix VII, Tables XX and XXII. As the passenger-mile and the ton-milfe are not recognized as traffic units in European railway-opera- tion, there is no basis on which a useful comparison can be made of the rail- way traffic in Europe with that in the United States. Efficiency in Freight Thaffic. Loss and Damage The freight-traffic statistics in this country, from 1890 to 1914, are given in Appendix VI, Tables XII to XIV, and may be summarized as follows : Freight Traffic, United States 1890 1900 1910 1912 1914 Tons, millions . . . 636 1.081 1,745 2,058 1,976 Ton-miles, millions . . 76,207 141,596 255,016 264,080 288,319 Tons, per mile of line . 487,245 735,352 1,071,086 1,078,580 1,176,923 Railway mileage . . . 165,936 193,345 240,831 249,852 247,397 Freight- train mileage. millions ... 635 612 605 Average train-load, tons 401 431 475 Average haul, miles . . 119 138 146 143 146 Minor Unes excluded in 1914, 6,695 miles. "With respect to Traffic Efficiency, the statistical averages of a railway- system so extensive as that of the United States are not so instructive as are those of the several traffic-districts into which that system has been divided in the reports of the Interstate Commerce Commission. An analysis of these statistics shows that the decrease in density of tonnage per mile of line, from 1912 to 1914, was mainly in the Eastern District, partly in the Western District, with a slight increase in the Southern Dis- trict. The average haul on roads in Classes I and II has varied but Httle relatively in the Southern and in the Western Districts, where the hauls are much longer than in the Eastern District. In this latter district, there TRAFFIC 329 has been a material reduction in the average haul on each class of roads. This was probably due to the increase in coal-tonnage, with a relatively- shorter haul than is the case with the tonnage of raw products. It is a striking tribute to the general efficiency of the railway manage- ments that, under the prevaiHng traffic-conditions, the average train-load in Classes I and II should have continuously increased through the four years 1910-1914 from 401 tons to 475 tons, or 18 per cent. In the Eastern District, the increase was from 469 tons to 549 tons; in the Southern District from 365 to 418 tons ; and in the Western District, from 352 tons to 419 tons. Yet, in the same period, with an increase of 9.0 per cent, in passenger-mileage, the passenger-train mileage increased 9.7 per cent. This is plainly a sacrifice of economic efficiency to social efficiency. The statement of tonnage from points of origin, affords an interesting view of the relative magnitude of products and commodities as elements in the railway-traffic of the country. In the period 1912-1914, one-half of this tonnage originated in the Eastern District, one-sixth in the Southern District and one-third in the Western District. Minerals constituted 57 per cent, of the total tonnage, manufactures, 14 per cent, and forest products 10 per cent. Agriculture contributed but 4.8 per cent, of the tonnage in the Eastern District, 8.2 per cent, in the Southern District and 16.5 per cent, in the Western District. The following comparison of the classified tonnage in this period with that in the period 1905-1909, shows how slight has been the relative change between the several classes, since that time : ^ 1912-14 Agriculture Animals Mines . . Forests Manufactures Merchandise Miscellaneous Per Cent. 9.0 2.5 54.0 11.5 14.0 4.0 5.0 Per Cent. 9.2 2.4 57.0 10.0 14.0 3.8 3.6 The fundamental principles of safety, expedition and convenience apply to efficiency in freight-traffic as in passenger-traffic, differing only as they are applied in the transportation of products and commodities as dis- tinguished from human beings. In the conduct of both classes of traffic, the railroad company becomes an insurance underwriter, but to a greater extent with reference to freight than to passengers, as the responsibility of the carrier is somewhat diminished by negligence on the part of the passenger. The carrier, however, assumes a responsibility in caring for human lives which can not be measured by pecuniary considerations. Yet, 1 For TrafiBc Statistics in the period 1905-1909, see "Problems in Railway Regulation," Appendix X. 330 EFFICIENT RAILWAY OPERATION even when a claim for loss or damage to freight has been adjusted, there remains an economic loss which is distributed among the risks assumed in common, with the additional burden of profit to the underwriter. These losses occur from negligence in packing, marking, bilhng, loading and delivering shipments ; from quahties inherent in the commodities shipped ; and fraudulent claims are believed to form a considerable item. In fact, the claims arising from these causes far exceed in amount those occasioned by raUroad-wrecks. This economic waste has been increasing to a remarkable extent. The payments for loss and damage, which amounted to $11,000,000 in 1902, had increased to $27,500,000 in 1908, and to $33,500,000 in 1914, which was equal to 1.59 per cent, of the total revenue from freight-traffic in that year. Of this loss, 18 per cent, was for entire packages, and unlocated dam- ages amounted to nearly 20 per cent. Damages ainounting to 12.5 per cent, were attributed to rough handling of cars, and 4 per cent, to im- proper handling and to unsuitable packages, while 10 per cent, was charge- able to defective cars, of which 60 per cent, was incurred in loss and damage to grain, flour and other mill-products. The result of an investigation, made by the Interstate Commerce Com- mission, of payments for loss and damage in the first nine months of 1914, is stated in Appendix VI, Table XV. From this statement, it appears that ou* of $26,000,000 paid in that period, about $11,000,000 was on account of unlocated loss and damage. It is remarkable that it should have been found impracticable to trace the causes of 40 per cent, of the total amount. Errors of employees and rough handling of cars account for another $4,000,000. Over $1,000,000 charged to improper handling and packing was, for the most part, due to carelessness in packing and to un- suitable packages. Robbery and concealed loss and damage accounted for over $2,500,000. Defective equipment (that is, leaky car-roofs and car-floors), cost another $2,500,000, aiid improper refrigeration and ven- tilation cost over $750,000. The losses actually incurred in transporta- tion were as follows : From delays $1,704,014 From wrecks 1,577,474 From fire 608,753 Total $3,890,241 Apart from these transportation-items, 85 per cent, of the total amount paid out in that period of nine months, was incurred from causes which may be summed up as general inefficiency in the inspection of car-repairs, in billing and loading, in the switching-service, and in watchfulness over the property intrusted to the care of railroad-employees. It is a virtual confession of impotence or of negligence on the part of operating- officials, which it was time to bring to the attention of railway-managers through the American Railway Association, and they are indebted to a TRAFFIC 331 special committee of that body for a lucid exposition of this situation, in which it was stated that, if the working of the operating-departments were perfected, there need be little loss or damage. One-half of this loss was in the following commodities : Fruit and vegetables G 2,056,575 Grain . 2,050,380 Live-stock . 1,789,314 Clothing and dry-goods . . 1,736,512 Furniture 1,297,145 Groceries 1,160,281 Flour and mill-products ... 1,149,231 Meats . . 775,228 Household goods 771,896 Pottery and Crockery 658,138 Total 313,444,700 Payment for unlocated loss or damage ran through this entire list of commodities, to the extent of about one-third of the total amount. Other causes affected certain classes more especially, as the rough handling of cars in switching, for which losses were paid amounting to $1,500,000; and particularly for live-stock, furniture, fruit and vegetables.. Live-stock claims for delays amounted to nearly half of the losses on such shipments. Losses on grain from leaky car-floors amounted to over $1,000,000 and on flour, etc., from leaky car-roofs to $500,000. Improper refrigeration caused a loss of $538,000 on shipments of fruit and vegetables and of $126,000 on meats. Robbery and concealed loss of clothing and dry-goods amounted to $828,000. Upon such an analysis of the causes of loss and damage, the special committee of the American Railway Association, in April, 1915, based its recommendations for their prevention. These recommendations included a specific inspection-certificate for each car to be loaded with any com- modity liable to damage from a leaky car-roof or to loss of bulk grain from a defective floor. Certain improvements were suggested in the methods of interline-billing and in auditing "over and short" reports, as means for prompt detection of irregularities; as also in the loading and billing of shipments in less than car-loads ("L.C.L. " freight), and for greater care in packing, marking and loading such shipments ; and, further, that closer attention should be paid by inspection-bureaus to such matters. As to defective cars, the committee suggested that claims arising from this cause should be chargeable to the company on whose line such cars had been loaded. For rough handling of cars, the yard-men should be held to a stricter responsibility, and more efficiency should be displayed in the pre- vention and detection of robbery. The committee mentioned that, in tracing for delayed shipments, the railroad companies annually transmitted about 5,000,000 telegrams and 3,000,000 letters, at an expense of over $1,000,000 ; much of which could be avoided by stricter observance of its 332 EFFICIENT RAILWAY OPE^IATION recommendations as to the practice of tracing shipments. Apparently as a result of the efforts of this committee, a comparison of paid claims for 1914 and 1916 on 84 roads operating 134,132 miles of line, showed a re- duction from $19,008,709 to $13,806,280, or 27.3 per cent. Further investigation by this committee developed the fact that a con- siderable proportion of damage-claims were due to qualities inherent in the commodities themselves. For instance, the corrosion of galvanized- steel sheets had been found to be caused simply by a change of tempera- ture. This matter of damages from qualities inherent in the commodities shipped, has assumed paramount importance in the transportation of ex- plosives. Serious accidents, resulting in loss of life, in addition to the de- struction of property, occurred so frequently from this cause that, in April, 1905, this matter was taken under consideration by the American Railway Association. An instance was cited in which a car loaded with powder, in si freight-train standing on a siding, exploded just as a passenger- train was passing. The train was partially destroyed and fifteen persons were killed. Investigation of such occurrences showed great carelessness and even intentional concealment in the packing of explosives, as well as negligence in loading the packages into cars, and in handling the trains to which such cars were a;ttached. Safe Tbansportation of Explosives, etc. A committee of the American Railway Association was appointed to prepare regulations for the safe transportation of explosives, which obtained the valuable assistance of Colonel B. W. Dunn, an experienced officer in the ordnance department of the United States Army. Under his direc- tion, in 1907, a bureau was organized to enforce the regulations recom- mended by the committee. Courses of lectures were given by officials of this bureau to railroad-employees, manufacturers and shippers who, by this educational process, were stirred to personal interest in the enforce- ment of these regulations. This subject was brought to the attention of Congress, which passed an Act, approved May 30, 1908, and amended July 1, 1910, "to promote the safe transportation in interstate commerce of explosives and other dangerous articles, and to provide penalties for its violation." This Act gave authority to the Interstate Commerce Commission, which promptly incorporated the rules of the American Railway Association in a series of regulations made effective October 11, 1908. By virtue of these regu- lations, "the Bureau for the Safe Transportation of Explosives and other Dangerous Articles, of the American Railway Association " was authorized, through its agents, to inspect even the methods of manufacturing and pack- ing of explosives " so far as it affects safe transportation." These regulations of the American Railway Association as to the transportation of explosives have been adopted by the Board of Railway Commissioners of Canada, and TRAFFIC 333 the Canadian railway companies are members of the Association Bureau of Explosives,^ which also includes 8 express companies, 12 steamship com- panies, and 71 manufacturers of explosives and other dangerous articles. It was found impracticable to regulate the transportation of explosives effectually without regulating that of inflammables. The Act was ac- cordingly amended to include "other dangerous articles," and it became necessary to define such articles with precision, to regulate the manner of marking, for identification, the packages and cars containing them, and to prescribe the methods of packing and handhng them. They were classi- fied as inflammable liquids, inflammable solids, and acids. Among the inflammable liquids, the greatest danger to life and prop- erty has been incurred in the transportation of benzine and gasoline, and particularly of "casing-head" gasoline — the condensed vapor of natural gas. In any type of package, whether of glass, metal or wood, there will occasionally occur an accidental leakage and, if the temperature within the car is above the "flash-point" of the liquid, the vapor from it will, in combination with the air, constitute an explosive vapor of which the dangerous flash-point has been fixed at 100° F. As a preventive of such leakage, it was required that tank-cars should be tested by cold water to a pressure of 60 pounds per square inch, and stenciled accordingly. On Sept. 27, 1915, a tank-car, billed as gasohne, exploded on a siding at Ardmore, Oklahoma, causing 41 deaths and 458 cases of personal injury, with a property-loss of over $800,000. The contents proved to have been casing-head gasoline, and the accident was caused by the removal of the dome-cap of the tank while its contents were subjected to internal pressure from vaporization under atmospheric heat. The attention of the consignee had been called to the continual popping of the safety-valves in the tank which, as required by the regulations, should have been sprayed with cold water until the temperature had been reduced. In 1914, 52 per cent, of the total damage occurring in the transporta- tion of "other dangerous articles" was caused by gasoline, matches, char- coal and nitric acid. "Strike-anywhere" matches caused 93 fires, with loss of $24_,886. Fires from spontaneous combustion of ground charcoal, in 392 fires, cost $140,092, from 1910 to 1915. In 1915, besides the Ardmore catastrophe, there were 24 fires from gasoline, causing the death of 12 persons and injuries to 9 others. In 1915, the breaking of a bottle of nitric acid, unlawfully packed in excelsior, caused a fire-loss of $165,000. • A valuable incidental work of inspection, in 1914, was the correction, in 263 cases, of the dangerous location of storage-magazines of explosives. In this useful service, 1596 storage-magazines and 278 factories were in- spected and 5366 packages of explosives were condemned as unsafe. This inspection has been extended to the location of storage-tanks and, in con- 1 "Problems in Railway Regulation," p. 332. 334 EFFICIENT RAILWAY OPERATION sequence of the Ardmore accident, of sidings for loading and unloading tank-cars with inflammable liquids, and with more rigid regulations con- cerning casing-head gasoline. A laboratory has been established for analyzing samples of explosives and other dangerous articles. A further hazard connected with the transportation of articles not con- sidered as dangerous, has been developed in this work. In 1915, in the trans- portation of fuel-oil, 8 persons were killed and 36 injured, with property loss of $74,854. In consequence of accidents from explosives in passengers' baggage, this practice was prohibited, with an exception in favor of cylin- ders of compressed gases used in stereopticon outfits, and also of moving- picture films, when properly packed and labeled. Yet, in 1915, a package of films became ignited in an express-car upon a rapidly moving train. The car had no end-doors and the messengers took refuge on the iron side- ladder, where they rode for seven miles before the train was stopped. The car and contents were destroyed with a property-loss of $50,000. In November, 1914, a paper parcel containing such films was taken into a smoking-car and placed on the floor. After the train was in motion, the parcel burst into flames ; 38 persons were badly injured and 3 lost their lives. On investigation, it was found that more than 7000 persons were employed in this service in the distribution of moving-picture films. In 1914, in the transportation of over 600,000,000 pounds of explosives, there were 11 accidents, in which no lives were lost, though 5 persons were in- jured, with property-loss of $14,106. In the same year, in the transporta- tion of other dangerous articles, there were 470 accidents, in which 11 persons were killed and 109 injured, with property-loss of $257,365. In 1907, there were 79 accidents in the transportation of explosives, in which 52 persons were killed and 80 injured. In the five years ending in 1915, there have been 42 accidents in such transportation, in which 23 persons were injured, but no lives have been lost, while the total property- loss amounted to $46,481. In 1915, the loss was $127.00 ! During the transportation of immense quantities of war-material, no explosion had occurred until July 30, 1916, when there was a serious fire and explosion at the railroad-pier on Black Tom Island in New York Harbor, caused in the loading of a barge, in which three lives were lost, and which has been traced to incendiarism. The success attained in this matter should be an example to operating officials as to what might be done with respect to loss of life and property in other ways, by the exercise of experience, common sense and tact.^ Through and Local Freight Handling Adequate facilities for reception and loading, and for storage and de- livery, are of far more importance than expedition in transit to the great ' For statistics as to accidents in the transportation of dangerous articles see Appendix VI, Table XVI. ' TRAFFIC 335 volume of traffic in bulk-freight, in minerals and in grain which, in 1914, constituted 58 per cent, of the total tonnage. These commodities move altogether in car-load lots and mostly in train-loads, at rates of speed so slow as to reduce the average speed of freight-trains to about thirteen miles an hour between terminals. They call for ready reception from mines and elevators, as required by commercial demand, but the time in transit is of little consequence compared with continuous movement suffi- cient to maintain the necessary reserves in the great industrial centers and international markets. In the distribution of commodities manufactured from raw products, expeditious transit is desirable in proportion to their greater intrinsic value per ton, which is provided, for by "fast-freight"' service. Perishables and live-stock are given even greater acceleration. Shipments in less than car-load lots receive less consideration, though they are charged relatively higher rates. The stumbling-block in long-distance freight-traffic is the obstruction at intermediate terminals, where trains are broken up and re-arranged for different districts and for divergent routes. To prevent such obstruction, millions have been invested in classification-yards and in cut-offs. Con- tinuous train-movement is impracticable on a single-track line, on which trains must meet or pass at prearranged places and times. On double- track, freight-trains must clear the running-tracks for passenger-trains and, even on four-track lines where the freight and passenger trains move on separate tracks in opposite directions, slow trains must give way for fast trains. The ideal service for heavy freight-tonnage would only be possible on a double-track line devoted solely to that class of traffic, equipped with automatic block-signals and track-tanks, operated by loco- motives of adequate tractive power and loaded to full capacity, in a con- tinuous procession at an economic speed between division terminals just far enough apart for coaling and inspection. Car Interchange and Car Service Expeditious transportation by rail would be impracticable, but for the interchange of roUing-stock between connecting lines. The original railroad-corporations controlled only short and disconnected lines, and all freight was billed locally. Shipments to points beyond the line were de- livered to forwarding agencies that paid the bills and transferred the goods to the receiving-station of the next line, where they were rebilled with ac- cumulated charges. This course was pursued until the extension of lines from rival seaports, into a region common to both, aroused competition. Such rivalry between a route controlled by a single corporation and one composed of several disconnected lines, induced the construction of track- connections between their terminals with an interchange of rolling-stock for business in which they were alike interested. With the rapid extension of rival routes farther into the interior, and stimulated by competition, the 336 EFFICIENT RAILWAY OPERATION increasing volume of interline-traffic absorbed the freight-equipment, without regard to local requirements. In such emergencies, cars were loaded indiscriminately without respect to ownership. Many strayed far away from the proprietary Une, perhaps to be discarded when disabled, or to disappear permanently* by destruction in wrecks. Corporations con- trolling a considerable volume of competitive traffic, demanded some equitable adjustment of the use of their cars by their connections. This was accomplished to some extent by the organization of fast-freight lines, with equipment jointly owned by the companies interested in the traffic in which these lines were engaged. This plan did not cover the entire field of competitive business, and agreements were entered into for equalizing the contribution of cars in such traffic by settlement for their use on a car- mileage basis. This required a special organization on each line for the collection and interchange of data between the companies whose cars were so employed. These agreements were open to misinterpretation and rested entirely on the good faith of the contracting parties. A management, upon whose line a considerable volume of business originated which was common to two or more of these rival routes, could afford to treat such agreements with indifference ; as it might direct business to one or another of them. Other managements, with but little interest in common, refused to enter into mileage-contracts and used and abused, at their pleasure, the cars that came to their lines in the ordinary interchange of traffic. "Lost-car" agents traversed the country from ocean to ocean, searching for missing cars by their numbers, wherever they could be spotted, in passing trains or on sidings, and with very little assistance from delinquents. The car-mileage basis for the interchange of cars had proved to be an insufficient remedy for the misuse of them. It recognized only the wear- and-tear of a car while in motion. It did not recompense the owner for the loss of its earning capacity while the car was away from the proprietary line. An attempt was made to devise a more efficient remedy by the ad- dition of a per-diem charge to the car-mileage charge. This "mixed" basis was to be applied by voluntary agreements through local car-service associations. To insure uniformity in the conduct of these associations, the subject was referred to the only organized body of railroad managers, the General Time Convention, which, in 1889, appointed for this purpose a Committee on Car Mileage and Per Diem Rates. In reporting a plan for uniformity, the committee included the application of a demurrage- charge for the unreasonable detention of a loaded car by its consignee. Efpobts to Incbease Car Efficiency. Car Detention In April, 1891, the General Time Convention was transformed into the American Railway Association, with an increased membership of manage- ments and a wider field of action. In October, 1900, a set of Car Service Rules was adopted, in which the car-mileage basis was replaced by a per- TRAFFIC 337 diem charge of 20 cents a day, made effective from July 1, 1902. United action for enforcing these rules was found impracticable for lack of a central organization. This was provided in 1903 under a Car Service and Penalty Agreement, which included the appointment of a Committee on Car Effi- ciency. The Committee on Car Efficiency represented only the managements that were parties to the Car Service and Penalty Agreement. Although these managements controlled 86 per cent, of the line-mileage and 90 per cent, of the freight-equipment of our entire railway-system, that agreement had been signed by only half of the managements which were members of the American Railway Association. . The parties to the agreement them- selves responded but incompletely to its requirements. The Committee on Car Efficiency was kept busy in hearing appeals for the remission of charges and penalties. The arguments in support of these appeals were, for the most part, so intensely technical as to be suggestive of the pleas in a chancery court. The managements which were not parties to the agree- ment, embarrassed its effective operation. The per-diem basis had the effect of placing a definite penalty upon a foreign road for the unnec- essary detention of a loaded car, which penalty, by the application of the demurrage-rule, was transferred to its consignee. This misuse of foreign cars had hitherto prevailed at competitive points without much question. Thousands of cars were thus held out of service while brokers were seeking customers for their contents, or to save to consignees the cost of handling and storage by selling directly from the cars. The Committee on Car Efficiency reported, in February, 1907, a short- age of 137,000 cars but, in consequence of the non-observance of the de- murrage-rule, it was unable to improve this situation. In the enforce- ment of the demurrage-charge, no support was received from railroad commissions or from public opinion. Several State Commissions formu- lated their own demurrage-rules and one of them, in an. official report, boasted that its rules were "more favorable to the shipper than those of any other State in the Union." In the latter part of 1906, complaints had arisen of embarrassment to commerce, and even of suffering in some communities, from long-continued delay in the transit of necessary commodities. The President of the United States was urged to send a special message to Congress, recommending legis- lation to compel railroad companies to furnish cars within a reasonable time. A bill of this character was introduced, prescribing penalties and authorizing consequential damages. Meanwhile, through the efforts of the Committee on Car Efficiency, the lack of transportation was remedied in the regions Tvhence the most urgent appeals had proceeded. In July, 1907, there was a general surplus of 37,000 cars and, in April, 1908, of 413,000 cars. A more efficient means for the prevention of the unreasonable deten- tion of cars by consignees was proposed in the American Railway Associa- 338 EFFICIENT RAILWAY OPERATION tion, in the recognition of demurrage-charges by the Interstate Commerce Commission, and the collection of such charges through Demurrage Bureaus acting independently of the local railroad agencies. Through the National Association of Railway Commissioners, the Interstate Commerce Commis- sion subsequently secured the cooperation of the State Commissions with a committee appointed by the American Railway Association, and with representatives of commercial organizations, in the formulation of a set of "National Car Demurrage Rules," which was approved by the National Association on November 17, 1909, by the Interstate Commerce Commis- sion on December 18, 1909, and by the American Railway Association on January 27, 1910.i Notwithstanding the formal approval of these demurrage-charges by the lawfully constituted representatives of the public welfare, their appli- cation to cars held for storage-purposes met with determined opposition from those who were profiting by this custom. The Interstate Commerce Commission refrained from extending its authorization beyond a recom- mendation for compliance with them. Consignees regarded with indif- ference these efforts to remedy the misuse of. railroad-equipment. They relied upon the persuasive effect of a diversion of patronage and interposed technical difficulties in counterclaims for delay in placing cars and by pleas of inability to discharge them by reason of inclement weather. As a con- sequence of these adverse influences, delivering-lines enforced the rules in a half-hearted way, and placated consignees by rebating the demurrage- charges. Car Shortage The efforts of the American Railway Association for securing increased efficiency in the use of freight-equipment had proved but partially effec- tive, when the outbreak of the European War created a sudden and un- anticipated demand for such equipment in export-traffic. Cars were ar- riving at the seaports by thousands daily, which were held in waiting for shipping until, by April, 1915, there were 321,000 cars withdrawn from profitable employment. The storage-tracks at the terminals were filled to repletion and the tracks available for switching were encroached upon to such an extent that loaded trains were hauled back to interior yards to make room for the daily train-movements. As a consequence of these conditions, the terminal lines were forced to refuse to accept loaded cars from their connections. These, in turn, took similar action, until the entire freight-traffic of the country was threatened with stagnation. In this emergency, remarkable efficiency was displayed by the railroad manage- ments. Heavy exports to Russia, which were held up by slides in the Panama Canal, were hauled back across the continent to Pacific ports and many cars with weather-proof contents were unloaded on the ground. ' See "Problems in Railway Regulation," p. 335. TRAFFIC 339 By November, 1915, the situation had been sensibly improved, and the American Railway Association took into consideration means for prevent- ing its recurrence. Attention was directed to restricting the use of cars for storage by reducing the period of free time and by increasing the de- murrage-charge after the first day subsequent to the free period. It was also proposed to abolish unlimited free time upon cars loaded with export grain. The indifference of certain managements in the enforcement of the car-service rules led the Association to give notice in May, 1916, that from June 1, the Commission on Car Service would impose penalties for non-observance of these rules, to be paid to the companies whose equip- ment had been thus misused. At the Association meeting in May, 1916, further pressure was put upon consignees' holding loaded cars by the adop- tion of "National Storage Track Rules," imposing a charge for the use of track-room in addition to the demurrage-charges. The Commission on Car Service was further authorized to prepare rules for the issuance and handliag of embargoes. With the return of autumn in 1916, a season of car-shortage recurred which increased in intensity until, by November 1, 1916, it had reached 114,000 cars ; though the situation again changed so rapidly that by Feb- ruary 17, 1917, there were 171,000 cars on the idle list. At the meeting of the Association in November, 1916, provision had been made for the appointment of a Conference Committee on Car Efficiency, composed of operating-officials, to be in permanent session at Washington, for coopera- tion with the Interstate Commerce Commission in the enforcement of the rules and regulations governing car-service. On December 6, 1916, the per-diem charge had been increased, effective May 1, 1917, and, at a spe- cial meeting of the Association on February 2, 1917, the hands of the Com- mission on Car Service were strengthened for a more rigorous enforcement of the rules by the adoption of a "Car Service and Penalty Agreement" over the signatures of responsible officials of all the members of the As- sociation. This account of the car-service situation, during the most serious con- gestion of railway traffic which has ever occurred in the United States, exhibits the conditions which restrict, the efforts of railroad managements for improved traffic-efficiency. The situation in this respect, during the past ten years, has been graphically depicted in a diagram attached to a publication by the American Railway Association on February 6, 1917, and which is reproduced in Appendix VI, Table XVII. The comparison ithere presented between periods of "Car Idleness" and of "Car Shortage" shows that during this decade there have been two periods of considerable car-shortage, in 1907 and in 1916, and that the most serious of these short- ages reached maxima of 120,000 to 135,000 cars, with four intervening maxima of idle cars, varying between 195,000 and 413,000 ; the number of cars in service having increased from 1,840,000 on July 1, 1907, to 2,518,855 340 EFFICIENT RAILWAY OPERATION on July 1, i916. In considering these statistics, attention should be di- rected to the number of cars held for movement or unloading. On March 1, 1917, with a reported shortage of 124,973 cars, there were 123,063 cars thus held out of service. The fluctuations between the maximum and the minimum periods of surplusage and shortage have been frequent and ex- treme, indicating the necessity for remedies which the managements have not been able to supply, nor the railroad commissions to enforce. Means for Preventing Terminal Congestion. Seaport Rivalry The exigencies of traffic, periodically experienced in this country, can be obviated only by such remedial measures as will serve the interests of the producers, no less than those of the middlemen. And this view should not be confined to the ebb and flow of currents of tonnage over a few east- and-west trunk lines, but should be extended to the floods of commerce that sweep over our land from ocean to ocean and from the Great Lakes to the Gulf of Mexico, accumulating in volume, like tidal waves, as they approach the ports, and subsiding in intensity as they recede over the broad intervening area. With increasing prosperity, these waves of export- trafiic are progressively increasing in volume and in value.^ Exports doubled in value from 1896 to 1906, and tripled in the following decade. Upon this business, depends the favorable position of our country in that international commerce which has become of vital necessity to our con- tinued prosperity ; and freedom of movement is of first importance as a measure of traffic-efficiency. In the export-business, the port of New York is far in the lead. Its proportion has increased in twenty years from 40 per cent, to nearly 51 per cent., and it is in the New York terminals that the congestion of traffic has been most severely felt. There should be no attempts to relieve this congestion by the refusal of the trunk-lines to receive freight from their connections by the declara- tion of so-called "Embargoes"; which are as impotent as a measure of traffic-efficiency as they had proved to be as a national policy. They are at best but temporary palliatives. The resort to them is like obstructing a drainage-system in time of freshet, only to overflow the fruitful harvests farther back. The influx and efflux of commerce should be more efficiently 1 Exports. — U NiTBD States. (Millions op Dollars) 1896 1906 1914 1915 1916 United States .... New York New York, per cent. 882 354 40.0 1,798 622 34.6 2,113 833 39.4 3,554 1,797 50.5 5,481 2,790 50.9 TRAFFIC 341 V regulated ; and the first step for relief would seem to be to open wider the outlets at the ports. This was the experience in the port of London, where the greatest volume of international commerce has heretofore been con- centrated. The port of New York is gaining proportionately in that field, and should have the same measures for relief that have been successfully- applied in the port of London. This is not a matter to be left to private enterprise, either of railroad corporations or of individuals. The commerce of the port of London had suffered for centuries from a reliance upon such a policy, until it became recognized that effective remedies could not be applied by the disconnected efforts of conflicting private interests ; but that the situation should be dealt with through an organization which adequately represented the general welfare. The results have fully justified the adoption of this policy, and the communities bordering on New York Harbor should profit by this example. They should not rely upon railroad corporations for anything more than transient storage, nor upon warehouse and dock com- panies for other facilities than may be required for local purposes. These communities should put aside their sectional jealousies, and unite in fur- thering the interests which they have in common, by an ample provision, under a central organization, for storing export-commodities as they ar- rive, with reference to ready access by rail and water and of sufficient extent to accommodate commercial interests awaiting favorable oppor- tunities for export and import.' Without such intelligent measures, the port of New York can not main- tain its primacy on the North Atlantic coast. The Dominion of Canada is already prepared to challenge it with facilities provided at th'e public expense in the port of Halifax. New Orleans has, even now, superior port- facihties, which will draw an increasing proportion of the business of the Mississippi Valley with the greater use of the Panama Canal. The harbor of Norfolk is being provided with spacious pier-accommodation and with modern economic appliances. It is quite as well situated as the port of New York, for either European or South American trade, and is far better prepared to furnish abundant and cheap supplies of coal — the con- trolling element in conducting sea-borne traffic. Relieved of responsibility for export-storage at their terminals, the railroad managements will be freer to concentrate their efforts in that field of traffic-efficiency which more directly concerns them, in the supply of transportation-facilities by rail whenever and wherever they may be re- quired. Their obligation in this respect is two-fold ; to those interests that are entirely dependent on each of them, and to those which they serve in common. The temptation to favor these latter interests should not prevail over their duty to the former. Their obligations with respect to interline-traffic are now discharged through a multitude of individual and ' See Appendix VI, Table XX. ^ 342 EFFICIENT RAILWAY OPERATION disconnected agencies, which exercise directly an influence over the dis- position of transportation-facilities that is not efficiently controlled by superior operating-officials. Division-superintendents and freight-agents, local station-agents, yard-masters and even the foremen of switching- crews can and do neutralize the efficient distribution of empty cars by inertness or by intentional disregard of the regulations which they are ex- pected to observe. \ Need of Central Authority to Regulate Car Distribution Apart from such venality as was uncloaked in an official investigation of the distribution of cars in the coal-traffic, there is one aspect of this under- hand favoritism that may be considered more creditable to those who ex- ercise it. For it is but human nature to give preference to neighbors rather than to strangers, and, when this tendency is indulged in, in order to retain or to gain competitive traffic, it is much more Ukely to be leniently dealt with by those higher in authority. For the reasons here given, it is futile to expect from local agencies a strict compliance with regulations for the distribution of cars in interline- traffic. This duty can only be efficiently performed by a central organiza- tion removed from local influences, with entire control over equipment devoted exclusively to interUne-traffic. That such a course is necessary for this purpose, is admitted in a statement made by the Commission on Car Service that, in an inspection of the car-records of 107 roads, there were detected over 40,000 violations of the car-service rules in a single month ; , and that there was convincing proof that, for years, these rules had been generally disregarded whenever it seemed desirable in the in- terests of individual roads. The responsibiUty for an improvement of this situation is not widely scattered over the whole of our railway-system. It rests primarily with thirty corporations which, in 1914, controlled 76.4 per cent, of the box-car equipment. They have but to agree upon a plan for it to be put in effect.' Action of Interstate Commerce Commission in Car Service Regulation In the case of Missouri & Illinois Coal Company vs. Illinois Central Railroad Company, the Interstate Commerce Commission held that "1. The temporary con;fiscation by carriers of the cars of other rail- roads and the placing of embargoes against cars being sent off the lines of the owners are alike unlawful and the railroads are expected to make such rules for the return of cars as will prevent such abuses. "2. The railroads of the country are called upon to so unite themselves that they will constitute one national system ; they must establish through 1 See Appendix III, Table V. TRAFFIC 343 routes, keep these routes open and in operation, furnish the necessary- facilities for transportation, make reasonable and proper rules of practice as between themselves and the shippers, and as between each other. "3. An embargo may be justifiable because of the physical inability of the carrier for some reason to deal with traffic which overwhelms it, but an embargo placed against connecting carriers because of their failure promptly to return cars is not consonant with the service which the carriers constituting the through route are required by law to give. "4. Railroads are required under the Act to serve the through routes which they have established with other carriers without respect to the fact that in rendering such service their equipment may be carried beyond their own Hnes. "5. Carriers are required to make reasonable rules and regulations with respect to the exchange, interchange and return of cars used upon through routes, and where they have failed in this respect the Commission is empowered to determine the individual or joint regulation or practice that is just, fair and reasonable." Any efficient method for providing against the recurrence of a car- famine must be based upon the exclusive control, by a centralized organi- zation, of sufficient rolHng-stock for distribution whenever and wherever wanted. To this stock of equipment, each company interested in inter- line-traffic should contribute its quota of standard rolling-stock upon some equitable basis. The Commission on Car Service has accepted this view in a special report on "Principles for Study of Car Service Problems," in which it classified freight equipment as either "Special Equipment, e.g., open cars, which necessarily involve an empty return movement, or 'Legal Tender' Equipment, e.g., box cars which are or may be loaded at any time a "" any point in any direction where there is traffic." "To be just to the railroads themselves and to the public gen- erally, this pool" of legal-tender equipment should be regulated, to the end that there shall be secured to every road the use, when it needs them, of its quota of such equipment "or, in the alternative, compensation in money for the difiference. Such regulation can be made effective only by aban- donment of the right to physical return to the owner of its o ,/n cars, and the substitution of the right to possession and use by each line of 'legal- tender' cars in kind equivalent to the cars by it owned and contributed to the pool." Special Facilities for Transportation of Mine Products, etc. The prosperity of extensive regions and of populous communities de- pends largely upon convenient methods and appliances for the transporta- tion of certain commodities. Coal, coke, ores and other products of the mines, in 1914, constituted nearly 57 per cent, of the tonnage of our railway- system. The many millions expended in providing faciUties for loading 344 EFFICIENT RAILWAY OPERATION and unloading such commodities are far exceeded in amount by the invest- ment in rolling-stock devoted solely to their transportation. While the total number of freight-cars increased 35 per cent, from 1905 to 1914, the open-top cars increased 42 per cent, in number and 87 per cent, in capacity. ' The total coal-tonnage in 1914 could have been carried by this equipment in ten trips, and the total product of the mines in sixteen. There is an investment of some $700,000,000 in this equipment, which can be used for no other purpose but must, for the most part, return empty to the mines. Refrigerator Cars. Cold Storage Great sums are invested in special equipment for the transportation of particular commodities classed as "perishables," which require protection from the vicissitudes of climate and from the germs of fermentation and putrefaction. The refrigerator-car was first patented in 1868, though experiments had previously been made on the Pennsylvania Railroad for the transportation of dairy-products and fresh meats in box cars with double sides, roofs and floors, insulated with sawdust and cooled by the insertion of a box of ice in the doors of the loaded cars. Until about 1875, Western cattle were transported to the seaboard to be slaughtered but, with the introduction of refrigerator-cars, this business was gradually transferred to the Western stockyards and the shipment of dressed meats became an important business. The export-trade has been extended across intervening oceans and between the hemispheres by the construction of cold-storage compartments in ocean-going ships. ^ This mode of transportation was greatly furthered by the use of artificial ice, and was adopted by the breweries in connection with the development of summer brewing. The transportation of products fresh from the fish- eries became practicable throughout the country, from the waters of the North Atlantic, the Gulf of Mexico and the Pacific coast. 1 Mineral Prodttcts Transported in 1914 Bituminous coal Anthracite . . Total . . Coke .... Ores .... Stone, sand, etc. Other minerals Total Tons (2000 lb.) 307,875,950 76,006,299 383,882,249 31,345,056 101,975,316 93,982,351 14,890,694 626,075,666 190S 1914 Inobease Peb Cent. Total no. freight oars . . Total no. coal cars . . . Capacity (tons, 2000 lb.) . 1,727,620 632,171 21,529,310 2,325,647 899,314 40,410,665 598,027 267,143 18,881,355 35 42 87 2 From 1880 to 1900, the number of cattle slaughtered in the United States increased from 8,000,000 to 24,000,000. TRAFFIC 345 In no part of the country has the use of refrigerator-cars been of more general benefit than along the seacoast of the South Atlantic States. This region had suffered severely from the ravages of the Civil War. The plantation-buildings had been burned or were in ruins ; the cotton-fields were overgrown with briers and bushes ; the rice-field drainage was useless, and the negroes were leading a shiftless life. In fact, large districts were retrograding into barbarism, until an impulse was given to the growth of perishable fruits and vegetables for the Northern markets, beginning in the vicinity of Norfolk and extending southward into the Carolinas. The discovery of fossil phosphates in this very region provided a necessary fertilizing material. Idle lands were reclaimed ; habits of industry were encouraged, and a change was effected in the condition of the resident population which was a striking example of the civilizing influences of speedy and reliable means of transportation. With the extension of railroads into the peninsula of Florida, the facil- ities for frequent and rapid transportation induced the planting of orange- groves, which became a profitable industry. Here, also, in a land ap- parently devoid of mineral-resources, valuable deposits were discovered of phosphate of lime.^ Fruit and vegetable shipments were made through the seaports untU the change of gauge on the Southern roads in 1886. In the first attempts at all-rail transportation, crude appliances for ventila- tion were of no other benefit than to admit air charged with dust, and the results were so unprofitable to shippers from Florida that the all-rail route would have been abandoned but for the opportune appearance of refriger- ator-cars, loaded with beer and with dressed meats for the winter-resort hotels. These cars were loaded back with perishables, and refrigerator- cars became the recognized equipment for this traffic. The experience in the transportation of perishables from the South Atlantic coast was re- peated on the coast of the Gulf of Mexico and in California. The ship- ment of bananas from Central America through Mobile and New Orleans was successfully established in the same way, and also the trans- portation of cantaloupes and of other perishables from the irrigated regions of Colorado. All these thriving agricultural industries have been made possible by the use of refrigerator-cars, which now number about 180,000, ^ ' In 1914, Florida marketed 2,543,876 tons of phosphate rock, being 82 per cent, of the entire production of the United States. 2 Citrus Fruit prom California. 1896-1897 — 7,350 car-loads 1897-1898 — 16,400 car-loads, 22 per cent, under refrigeration 1904-1905 — 31,422 car-loads, 51 per cent, under refrigeration Citrus Fruit prom Florida. 1914 — 26,435 car-loads Peaches from Georgia. 1895 ■ — 743 car-loads 1896 — 2,500 car-loads 1904 — 4,800 car-loads Strawberries from Southern Coast. 1897 — 425 car-loads 1906 — 2,613 car-loads Early Vegetables prom Southern Coast. 1907 — 467,169 tons 346 EFFICIENT RAILWAY OPERATION In 1885, cantaloupes were grown for market at Rocky Ford, Colorado, and the first car-load shipments were made in 1894. In 1897, the refriger- ator-shipments amounted to 121 car-loads and, in 1904, to 1182 car-loads. In 1897, the total shipments from all points were 400 car-loads' and, in 1904, 6920 car-loads ; the season for car-load shipments having been ex- tended from two months to six months. This remarkable development of production has been due to cold-storage in connection with refrigerator- transportation. Even apples and cabbages are more profitably marketed, when transported, without ice, in refrigerator-cars between November and April to prevent them from freezing, and then, under ice up to September, as a protection from overheating. The traffic in bananas originated in 1872, in shipments by sail from the West Indies to Boston, which led to the organization of the United Fruit Company. In 1903, this company imported from Central America, 30,- 000,000 bunches of bananas, averaging 100 bananas each, distributed largely through Mobile and New Orleans. At these ports, extensive wharf- facilities are provided in connection with speedy train-schedules, convenient re-icing stations and immediate delivery into cold-storage. Among other tropical fruits moved in refrigerator-cars, the pineapple is the most im- portant. The importation of this fruit increased from 1366 car-loads of 341,657 crates in 1900 to 3840 car-loads of 960,000 crates in 1908. Meanwhile, the development in the production of this fruit on the line of the Florida East Coast Railway increased from 5000 crates in 1883 to 690,- 000 crates in 1908. Refrigerator Tra.nsportation for Dairy and Perishable Products Refrigerator transportation has also been extended to shipments from poultry-farms and dairy-farms which, in one section of a Southern State, were in gross as follows : Butter . . . Cheese . . . Eggs .... Dressed .poultry 1893 Tons 64,130 9,040 32,097 16,251 1905 Tons 127,024 7,421 91,167 41,456 Cars in which milk is transported in iced containers, are operated in trains on a schedule speed of 25 miles an hour to New York City from points 400 miles distant ; leaving at 8.00 a.m. and arriving at midnight for next morn- ing's delivery. This business originated on the Erie Railroad in 1875, for a distance of 87 miles from Jersey City. In 1886, the total shipments TRAFFIC 347 were about 5,500,000 cans of 40 quarts each. By 1907, they had increased to 15,000,000 cans ; an increase, per capita of population supplied, from 96 quarts in 1886 to 136 quarts in 1907. Efficiency in the transportation of perishables is sought by rendering dormant the germs of vegetable-fermentation and of animal-putrefaction in the commodities transported, by reducing the maximum temperature within the car below 50°. This purpose is accomplished by insulation. The car is built with double walls lined with felt or other non-conducting material ; thus forming an inclosed air-space which excludes the external temperature. But refrigeration has also to reckon with the moisture in- herent in the commodities themselves, which, condensing on their sur- faces, becomes vaporized, causing the germs of decay to become revitalized on re-exposure. This moisture has to be removed by ventilation ob- tained through screened openings by the motion of the train. The reduced temperature is maintained by the passage of the indraught of air over the ■ ice-bunkers at the forward end of the car and out at the rear through graduated openings controlled by attendants. Under such ventilation, refrigerator service is efficiently rendered from the Pacific coast, across the Rocky Mountains and the arid Western plains, to the humid regions of the Atlantic coast. A device is in operation for removing accumulated gases and heated air from the top of the car, permitting cool air to flow in through the ventilators. In a car-load of peaches from Colorado to New York, the greatest variation of temperature between the 'floor and the roof was nine degrees. Pbe-cooling and Heating Arrangements A further improvement in refrigerator transportation has been derived from pre-cooling in cold-storage and the intermediate protection of the commodity from a higher temperature during its transit and redelivery into cold-storage terminals. The pre-cooling of perishables has secured a reliable extension of marketable territory with saving of ice in transit, with greater loading-space and less necessity for high speed. ' A car-load of pre-cooled fruit, loaded in California at a temperature of 42° and re- iced before starting, received no further refrigeration, and, after a journey of 3255 miles in ten days, arrived at Jersey City with 700 pounds of ice in the bunkers ; having been for seven days consecutively in an external tem- perature between 70° and 90°. A car-load in the same train, loaded at a temperature of 63°, was reduced to 50° on the fourth day out and arrived with a temperature of 48°, having been re-iced seven times during the trip ' In a pre-cooUng plant of the Southern Pacific Company, at Roseville, Cal., air from the storage plant at 32° is blown through an insulated tube into the ice- bunkers at one end of the car and is exhausted at the other end with the moisture and gases from the fruit. It takes from 30 to 60 hours to cool the fruit to 40° in the center of the car. In the pre-cooUng plant of this company at Colton, Cal., there are connections for forty cars. 348 EFFICIENT RAILWAY OPERATION with from one to two tons of ice at each station. Bananas are pre-cooled with air carried over ammonia-piping to the train-shed in insulated tubes, and there injected through canvas ducts into as many as fifty cars at a time, until the temperature is reduced to 68°.' The effect upon the prosperity of an agricultural region of providing protection from atmospheric conditions in the transportation of perish- ables has been signally exempHfied in the northeastern section of Maine, where the Bangor & Aroostook Railroad was constructed, about twenty- five years ago, for two hundred miles into an undeveloped timber-region. The shipment of potatoes from the line of that road increased from 1,500,- 000 bushels in 1894 to 12,300,000 bushels in 1906. The special facilities here provided consisted of heating arrangements, as the average winter temperature ranged between 15° and 20° below zero. The "heater" cars are built with hoUow walls. Underneath the car-floor there is an oil-stove, from which the heated air ascends through a duct into the space between the walls of the car. Fires lighted at the point of shipment require no • further attention for distances of 800 to 900 miles, although an attendant ^ accompanies each lot of five cars. These cars may be used for return- freight, as there are no interior obstructions.^ Transportation of Petroleum, etc. A similar use of special equipment has accompanied the vast develop- ment of the petroleum industries. The average daily production of the Oil Creek district in Pennsylvania, in 1861, was 700 barrels. There was no other oil-region of importance until the discovery of oil, in 1885-1886, in the Lima field of Ohio. In 1889, oil was discovered in West Virginia, and subsequently in other states, particularly in California. In 1901, the oil- region of Texas came into notice and produced 30,000,000 barrels in four years. In 1904, oil was discovered in Indiana and in Louisiana, in 1905, in Kansas, and in 1906 Illinois produced 24,000,000 barrels. ' The early refineries were situated on the railroad-lines in the oil-dis- tricts. The crude oil was piped to them from the wells, and the refined product was shipped in barrels. With improvements in pipe-line trans- portation, the oil was piped to refineries established within the regions of consumption. In 1899, the Standard Oil Company of New Jersey alone operated 35,000 miles of pipe-lines, representing an investment of $50,- 000,000. The total production of crude oil was — in 1874 10,000,000 bbl. in 1903 100,000,000 bbl. in 1915 218,000,000 bbl. in 1916 292,000,000 bbl. ' "Freight Terminals," Droege, p. 375. 2 Most of the information as to refrigerator transportation has been obtained from the Proceedings of the International Railway Congress at Berne, in 1910. Report on "Perishable Goods," by J. M. Gulp, Vice-President, Southern Railway Company. TRAFFIC 349 The production of gasoline has been so stimulated by its use as a source of motive power that, in 1915, it amounted to 1,000,000,000 gallons, and the output in 1916 is estimated at 1,500,000,000 gallons.^ The present con- sumption of crude oil by the refineries in the United States is estimated at about 1,000,000 barrels daily. The distribution of the refined products is dependent upon the special equipment of tank-cars, which are also used for the transportation of other liquids in bulk. In 1911, the total shipments of this character amounted to about 8,500,000 tons and, in 1914, to 11^600,- 000 tons ; an increase of 37 per cent, in four years. The private tank-car lines alone, in 1917, had a capacity of 4,000,000 tons. Special Freight Equipment. Private Car-lines The specialization of both passenger and freight equipment for a par- ticular service has been characteristic of railway operation in the United States. It has, for the most part, originated in the organization of private car-lines in the development of new business-enterprises on a large scale and is, in some respects, a reversion to the original system of separating the ownership of the cars from that of the track. The extent and variety of special freight equipment in organized private car-lines as compared with that owned by railroad companies is as follows : Special Freight Equipment in Use in the United States Compiled from Official Railway Equipment Register. March, 1917 Chabacteb 1917 Pbivate Cab-lines 1914 Railboad Companies Refrigerator cars Refrigerator cars, fast freight lines Total , Tank cars Stock cars Open-top cars , 85,895 43,591 129,486 83,945 36,825 13,852 48,886 8,530 82,971 899,314 Railway Delivery Service in England In the exercise of traffic-eflSciency, the service of a railroad-corporation is, under certain conditions, beneficially extended beyond its raiUterminals. Service of this kind has been co-existent in England with the earliest ap- plication of railway-transportation to the carriage of merchandise. As attention was directed to this traffic, it was necessary to compete with the carriers on the public highways by house-to-house service, which included cartage to and from the railway-terminals. Commercial intercourse be- tween the centers of production and consumption was thereby greatly i"The Age of Oil," by Charles A. Stoneham. 350 EFFICIENT RAILWAY OPERATION facilitated, and to the advantage of dealers trading on a small capital. The necessity for carrying a large stock of goods was sensibly diminished as an order for merchandise received on one day could be delivered early the next morning anywhere in Great Britain. At the London terminals, such freight is received up to 6.00 p.m., and is loaded and dispatched before midnight. From each of the principal stations, there are usually about thirty of these trains every night, of thirty car-loads each, operated at a speed of forty miles an hour. As the cars are of small capacity, full car-loads may be distributed along either the main line or branches without breaking bulk. Overnight delivery is made to Edinburgh and Glasgow, 400 miles away, and, at a later hour, to ports on the East coast of Ireland. ^ The traffic in fresh fish from the seaports is conducted in the same way. The transshipment of household goods from house to house is accomplished by transferring the body of a loaded van from its running-gear to an open wagon, and again, on arrival at the rail- way, transferring it back to other running-gear. The cartage of all goods is included in the railway-service, with the excieption of bulk-freight. On five railways, there are 18,000 horses in this service ; the London & North Western Railway Company alone em- ploys 6000 men and horses. Much of the service of this character is per- formed in the United States by the express companies, where it has been apparently found impracticable to conduct it as satisfactorily by the rail- road companies, because of the far greater area of territory to be covered, the greater capacity of the freight-cars and the more numerous centers of production and distribution. Still, unless railroad-managements with terminals in large cities, adopt some plan for overnight delivery of mer- chandise, motor-trucks will compete for such traffic, as motor-cars are competing for local passenger-traffic. The insular position of the British railways induced the extension of their service across the Irish Sea to Ireland, and across the British Channel and the North Sea to the Continent. The development of this service has been stimulated by competition from the several seaport-terminals. In 1912, 167 steamers were so employed by fourteen railway companies. The Lancashire & Yorkshire Railway Company owned a fleet of 33 ships, of which 24 were in the North Sea trade to Continental ports. Water and Rail Combination Service. Car-floats and Car Ferries The combination of water- and rail-transportation in the United States, originated in the necessity for maintaining communication across estuaries too broad to be bridged, as exemplified at New York City. Here, the situation of the metropolis between the Hudson and East rivers, induced the establishment of steam-ferries in the interest of the railroad companies ' "Freight Terminals," Droege, p. 307. TRAFFIC 351 with terminals on the farther shores, in order to compete with Hnes whose terminals were within the city. This competition was not confined to passenger-traffic. The local- freight-traffic was of equal importance to the railroads, and to place themselves on an equality in this respect, cars were transported upon floats, which were towed to freight-stations on the city piers. Car-floats carrying from eight to twenty-four cars virtually ex- tended the railroads along the entire shore-line of New York Harbor.^ In connection with lighterage and grain-elevators, the whole field of domes- tic and foreign commerce, by river, sound and sea, has been brought into direct connection with the railroads terminating on the west shore of the Hudson River. The water-front of Hoboken and of Jersey City is an al- most continuous railroad-terminal, with ferry-slips and piers of an assessed valuation of $93,000,000. The railroad-lines terminating on Manhattan Island have had also to gain access to the water-side, to meet this exten- sion of railroad-service, by the establishment of a similar system of light- erage and car-floats. The results, of these measures of trafiic-efficiency are to be seen in the commanding position which the port of New York holds in the domestic and foreign commerce of our country. There has been a similar extension of railroad-service by supplementary water-transportation in the harbors of Boston, Baltimore and Norfolk, under systematic provisions as to efficiency that have not been practicable in the port of New York, by reason of antecedent control of wharf -privileges in special interests. At these ports, and elsewhere, wherever the rails reach the water-front, the extent of the water-borne commerce has de- pended upon the efforts of the railroad-managements for its development, as has been amply illustrated in the growth of commerce on the Great Lakes. In July, 1909, enough Lake freight passed Detroit to fill 300,000 cars, equal to a daily movement. of 200 trains of 50 cars each. For the most part, this tonnage was brought by rail to the ports, and had to be moved again by rail, either as raw material or as manufactured products. The car-float, itself an American contrivance, has had a further develop- ment which is also peculiar to this country, in the steam car-ferry, as an intermediate link in an otherwise all-rail line. In 1875, 173,708 cars • The New York Harbor line has an extent of about 448 miles with about 604 miles of dock and wharf-frontage, of which 44 miles of harbor-line and 93 miles of wharf-frontage are along the shores of Manhattan Island. There are, in addi- tion, 30 miles of harbor-line and 96 miles of wharf -frontage on the New Jersey shore between Amboy and Fort Lee, to which railroad-freight is distributed by lighters. About 1000 cars are distributed daily on floats to stations between Canal Street, on the North River, and Jackson Street, on the East River. The railroads bring annually about 13,000,000 tons of freight into New York City. The Pennsyl- vania Railroad operates 60 floats and 8 tugs, with capacity of 700 cars, and transfers about 1000 cars daily. The New York Central Railroad operates 41 floats and 20 tugs, with 485 cars' capacity, and transfers about 760 cars daily. The New York, New Haven & Hartford Railroad operates 46 floats, 19 tugs and 2 steamers, with 771 cars' capacity, and transfers about 2000 cars daily, for distances of six to thirteen miles. "Freight Terminals," Droege. 352 EFFICIENT RAILWAY OPERATION were transferred across Detroit River. In 1909, 735,753 cars were there transferred by nine car-ferry boats, the largest being 351 feet long and carrying 24 freight cars, and the longest distance between terminals being about five miles. Frdm three to eight hours were consumed in crossing the river and in passing through the terminal yards.' Since the construc- tion of the Detroit River Tunnel, the most important ferries of this char- acter are in operation between San Francisco and the railroad terminals on the opposite shore of the bay. Between Vancouver and Vancouver Island, British Columbia, the car-ferry accommodates twenty-five pas- senger-ears. Car-ferries are in operation at several points on the Great Lakes. ^ This plan for the continuance of railway-service across rivers, estuaries and lakes, was applied, in 1915, in connection with the Florida East Coast Railway, which is built for over 100 miles along the Florida Keys and the intervening channels and sounds to Key West. Connection is there main- tained with the Cuban railway system at Havana by car-ferries. These vessels, two in number, are each of 3500 tons, 357 feet long and 58 feet beam, with speed of 14 miles an hour and capacity of 30 freight-cars. The steamers are oil-burners and consume the equivalent of 32 tons of coal on the round trip of 184 miles between Key West and Havana. The success that has attended the establishment of car-ferry service across the Florida Channel seems to warrant the adoption of similar "all-rail" transit across the English Channel, where the voyages are shorter and under no more disadvantageous conditions. Railway Control of Ocean Steamer Service The control of steamboat and steamship service in the interest of rail- road corporations, is to be desired, where such control is of benefit to the public, as a measure of traffic-efficiency ; as where it involves the traffic interchanged between otherwise disconnected railway-systems on the Great Lakes, and on the bays and sounds along our seacoasts. An un- divided responsibility, in connection with continuous service by land and water, is of prime importance in the transportation of passengers and of general merchandise. Such service is conducted more efficiently, as to safety, dispatch and convenience, where the water-line is an adjunct to a great railway-system than where it is the sole source of profit to its owners. This is not so much the case with water-borne commerce in bulk-freight, which is received and distributed under different port-conditions. The well-equipped steamers on Long Island Sound, Chesapeake Bay and the Great Lakes furnish a high-class service that is unequaled elsewhere. Our coastwise commerce is also conducted more satisfactorily in con- » " The Detroit River Tunnel," W. S. Kinnear, Trans. Am. Soc. C. E., Decem- ber, 1911. 2 See Appendix VI, Table XIX. TRAFFIC 353 nection with the railroad-lines, by which the service is continued into the interior, than if it were managed in separate and conflicting interests ; and the business of the terminal ports is benefited by a more frequent, regular and efficient service. To some extent, this statement may be applied to the conduct of commerce across the oceans. A great port, like New York, furnishes profitable business for independent steamship-lines in the car- riage of passengers and general merchandise ; but ports of less local impor- tance can not furnish the basis for an equally high-class service, unless it be provided in the interests of the railroad-lines terminating at such ports. The Pacific coast has been largely indebted to such interests for the main- tenance of their commercial relations with the farther shores of the Pacific Ocean. The political relations of our country with those regions, have also been facilitated by the existence of these steamship-lines under our national flag. It requires no lengthened forecast to conceive of an enormous ex- pansion both of our commercial and of our political relations with the other countries bordering on the Pacific Ocean. If our transcontinental systems are to be debarred from contributing to the maintenance of these relations, not only their interests, but the welfare of our country also, will be handi- capped in the rivalry of the ports on the coast of British Columbia, which will have the backing of the Canadian railway-system, with an unbroken control of steamship and railway service from the shor-es of the British Isles, across the Atlantic ■ Ocean and the North American Continent to thousands of miles of the seacoasts of Asia, Australia and New Zealand. This review of the subject-matter of railway-transportation in the United States, covers an area of territory about equal to that of the con- tinent of Europe, and a mileage greater by 11 per cent, than that of the en- tire railway-system of that continent, but with only one-fourth of its popu- lation. Railway-service in the United States is at present rendered to its population at the rate of about one mile of line to every 389 persons, while that rendered to the population of Europe is at the rate of about one mile to 1900 persons.' This statement affords some conception of the increase in trackage, equipment and terminal facilities which must be provided in advance for the requirements of our growing population, and also of the amount of capital that should be invested for this purpose. ' Area, square miles Railway mileage . Population . . . United States 3,026,789 261,554 101,882,479 Europe 3,872,561 234,625 464,681,000 2a CHAPTER VII TRANSPORTATION Relation op Thansportation to Other Railway Departments The purposes of Transportation are accomplished by means of Motive Power, Rolling-stock and Roadway ; its subject-matter is Traffic. The efficiency of the Motive Power Department is shown in the furnishing of tractive power and in the maintenance of a plant for the same. The efficiency of the Rolling-stock Department lies in the design and con- struction of vehicles suited to the traffic in which they are to be engaged, and in diminishing the effects of external and internal shocks and friction upon such vehicles while they are in motion. The efficiency of the Road- way Department is manifested in the location of a railway line, in con- formity with its physical and social environment, to the greatest advantage and at the lowest possible cost ; in the construction of works to faciUtate coipmunication across and beneath rivers and mountain-ranges ; in re- ducing gravity's resistance to train-motion on an ascending -grade ; in lessening external friction on curves ; in diminishing, through careful maintenance, the effects of oscillation and impact between trains arid tracks ; and in providing adequate means for the protection of train-move- ments and for the receipt and delivery of persons and commodities. How- ever skillfully these several instrumentalities of railway transportation may have been designed, constructed and maintained, it is the intelligent coordination of them by the Transportation Department that secures their profitable application to the movement of traffic with safety, dispatch and convenience. Fundamental Principles The primary office of the Transportation Department is the economical application of traction to train-seryice, or of Energy to Matter in Motion. Hence the physical laws which control this relation should be clearly understood by those who have to do with railway train-service.' The 1 As expounded by. Sir Isaac Newton, in 1685-1686, these laws are I. — Every body continues in its state of rest, or of uniform motion in a straight line, except ?o far as it may be compelled by force to change that state. II. — Change of motion is in proportion to the force applied and takes place in the direction of the straight line in which the force acts. III. — To every action there is always an equal and contrary reaction ; or, the mutual actions of any two bodies are always equal and oppositely directed. — Century Dictionary. 354 TRANSPORTATION 355 distinctive characteristic of matter is inertia.^ Matter can neither start itself nor stop itself. It can only be set in- motion by external impetus ; its motion can only be arrested by external resistance. Matter can only be set in motion by power derived from vital energy or from inorganic force. Once in motion, its onward course can only be arrested or modified by re- sistance originating in one of these forces and appearing either as gravity or as friction, except as due to collision. These propositions are as applicable to railway-transportation as to other aspects of matter in motion. The energy derived from heat is converted, through the mechanism of the locomotive, into power by which the train is set in motion. Once the train is in motion, that power is spent in overcoming the resistance offered to the continuance of that motion by the effect of gravity on ascending grades, by the internal friction be- tween the moving parts of the train, by the external friction of its wheels against the rails, and by the frictional resistance of the atmosphere. The inherent inertia of the train has first to be overcome, either by the action of gravity or of tractive power. The energy required to set a train in motion is, so to speak, stored up in the train before it begins to move; and the amount of that energy, as a quantitative force, is proportionate to the mass of matter in the train. But, as soon as the inertia of the train has been overcome, then the stored-up energy is released and reappears as work in maintaining the train in motion, or as velocity.^ Theoretically, once a body is set in motion, it continues to move on indefinitely with an initial velocity proportioned to the energy exerted in overcoming its inertia. That is, if a train were to be started by a push from a locomotive, not coupled to it, that train would continue to move on indefinitely at the same velocity, upon a track which was uniformly level, straight and smooth, provided that it encountered no resistance. But even upon such a track, its velocity would be retarded by the resistance offered by the friction of the atmosphere and by the journal-friction of the car-axles. These retarding effects would gradually absorb the energy of motion and bring the train to a state of rest again. Therefore, it is neces- sary for a continuance of uniform motion that there should be a continual expenditure of energyby the locomotive in the form of tractive power. Practically, no railroad-track is level for any considerable distance. Its physical environment compels a change of direction vertically, and here there comes into action an energy of a different kind from that de- ' Inertia. — That' property of matter by virtue of which it retains its state of rest or of uniform rectilinear motion, so long as no foreign cause changes that state. Quantitatively, inertia is the same as mass. — Century Dictionary. 2 Velocity is the rate of motion. The velocity of a body is uniform, when it passes through equal spaces in equal times. It is variable, when the spaces passed through in equal times are unequal. It is accelerated, when it passes constantly through a greater space in equal successive portions of time. It is retarded, when a less space is passed through in each equal successive portion of time. — Century Dictionary. 356 EFFICIENT RAILWAY OPERATION rived through the expansive property of steam, as developed in the tractive power of the locomotive. This is the Energy of Position or Potential Energy, as distinguished from the Energy of Motion or Kinetic Energy, and is due to tKat mysterious law of physics, which holds the universe together, the Attraction of Gravitation, the effect of which was accurately ascertained by the experiments of Galileo in 1589-1591.' In accordance with the Third Law of Motion, if the momentum ac- quired by a falling body in a vacuum could be entirely applied to its motion in a directly opposite direction, it would ascend in exactly the same pro- portionate distance, inversely, in each successive second ; and would come to a state of rest at just the place from which it started, and in exactly the same number of seconds consumed in its fall. The initial Energy of Position, or Potential Energy, transformed into Energy of Motion, or Kinetic Energy, in its downward course, would have been completely restored at the termination of its upward course. The ultimate effect upon a mass of matter of the acceleration or of the retardation of its motion, is the same, whether .the motion be directly vertical or in any inclined direction, provided the altitude be the same. Conse- quently, the expenditure of kinetic energy required to raise such a body in any inclined direction would be equivalent to that required to raise it vertically to an equal height ; though the total quantity of energy would have been expended more slowly, in proportion to the relatively greater distance over which that body had passed in its inclined ascent. SOUECES AND CAUSES OF RESISTANCE TO TrAIN MoTION In the application of these principles to train-service, other resistance than that of gravity must be considered. First, as to frictional resistance, which is due either to the sliding or rolling of two surfaces in contact, the effect of which is independent of the velocity or the area of contact, and depends solely upon the nature of the two surfaces and upon the pressure upon them, to which it is directly proportional. Friction may be external or between two bodies independently in motion, or internal, due to motion of parts within the body itself. Internal friction in train-service is principally caused by the revolution of the axles in their bearings, or Journal-friction.^ Journal-friction is due to the interlocking of the molecules in the two surfaces in contact, and this resistance to the revolution of the axles absorbs a portion of the kinetic energy of the train, which is dissipated in heat. There is also an external friction between the wheel-treads and the rails. Atmospheric resistance is of three kinds : the head-resistance, due to pressure by the moving train ; the skin-friction along the sides of the train ; and the resistance due to the " See Appendix VII, Note I. ^ The internal friction in the locomotive is not here considered, being in- cluded in estimating its tractive power. TRANSPORTATION 357 partial vacuum created between the cars and at the end of the train. These several effects of atmospheric resistance may be considered as varying with the speed and with the number of cars in the train. There are yet other causes of train-resistance in the oscillations of the locomotive and of the cars, due to irregularities in the track and to the impact of the wheel-flanges with the rails as they swerve from side to side. The effect of these resistances is proportional to the momentum of the train. Though none of these several resistances to train-motion can. be stated accurately as to quantity, as the effect of gravity can be, yet their total effect may be determined empirically by dynamometer-tests of the tender-drawbar-pull. With this statement of the causes and effects of train-resistance, consideration may be given to their practical application in train-service. Means fob Ovebcoming oe Lessening Fbiction in Train Movement It is the province of the TranspOTtation Department to utilize in train- service the locomotives and cars provided by the Motive Power and the Rolling-stock Departments. The theoretical tractive power or duty of each class of locomotives is estabhshed in pounds and its equivalent in tons of gross train-loads, in accordance with the formulas given in Ap- pendix II, Table XVIII. This equivalent is estimated for a straight, level and smooth track and,, on such a track, the tractive power sufficient to overcome the inertia of a train should, under normal conditions, maintain that train in motion at a speed of between seven and ten miles an hour. The tractive power of each locomotive is further proportioned on the par- ticular line or division upon which the train-service is conducted, in ac- cordance with the "ruling grade," as set forth in the "rating-sheet." The ruling-grade is usually the maximum grade on a tangent as, on a properly located line, any curvature on the maximum grade is duly compensated for by an equivalent reduction in the gradient at that point.^ Assuming that, upon any division, the curvature has been duly com- pensated for, the grade-resistance may be accurately determined in accord- ance with the law of gravitation. As already stated, the retardation of motion is theoretically the same, whether that motion be vertically upward or in any inclined direction; provided that the altitude is the same. Consequently, the tractive power required to draw a train up any grade, in excess of what would be required upon a level at the same speed, would be equivalent to the power expended in raising the train vertically to an equal height. The excess of tractive power so required is the measure of grade-resistance, and amounts to 20 pounds per ton-weight of train on a one per cent, grade, and proportionately on other gradients. ^ The frictional resistance can not be so accurately determined. Its most important element is journal-friction. Any increase of pressure on an axle- 1 See Appendix VII, Note II. ' See Appendix VII, Note III. 358 EFFICIENT RAILWAY OPERATION journal, within car-load limits, does not materially increase the journal- friction. The frictional resistance therefore, in an eight-wheel car, is virtually the same whether it be empty or loaded. Consequently, a train of twenty empty box cars offers the same resistance, as to journal- friction, that it would offer if the cars were loaded. Experimentally, this resistance is estimated at about nine pounds per ton-weight for an empty standard box car, at about four pounds per ton when fully loaded and, for. a mixed train of empty and loaded cars, it is averaged at six pounds per ton.i It is assumed that the journals are in normal condition as to lubrication, that the lubricating material is not chilled below 40° F., and that the jour- nals are not heated above 100° F. With reference to lubrication, it may be stated that the character or composition of the lubricating material seems to be of less consequence than the fit of the bearing, or the manner in which that material is applied. Much more attention is paid to this matter in Europe than in this country. There, the lubrication is made with oil contained in a reservoir beneath the journal, to which it is applied by a pad fed with wicks. The box is dust-tight,, the oil economically supplied, and a hot box is rarely seen. In the United States, this matter has received far less attention than it deserves. The journal-boxes are neither oil-tight nor dust-tight ; the wastage of lubricating material is evident along the track and hot boxes are by no means unusual. Apart from the dangerous delays thus caused, and the possible accidents from broken journals, the draft upon the tractive power of the locomotive is ma- terially increased and to no profit. The track-resistance from the external or rolling friction between the wheels and the rails, may be regarded as proportionate to the weight of the train, and as sufficiently covered by the allowance for journal-friction. But this does not include the effects of oscillation, impact, or the wave of deflection, which vary greatly with the speed and weight of the train, and may be comparatively ascertained by dynamometer tests.^ The effect of atmospheric resistance upon slowly moving freight-trains is almost negli- gible. Probably the most serious resistance is caused by leaving open the doors of box cars. Rating and Tonnage Capacity of Locomotives The several causes of resistance to train-motion which are to be con- sidered in the construction of a locomotive rating-sheet may be summed up as follows : ' The results of recent tests are given in Appendix VII, Note IV. With rolling- stock of modern construction, the frictional resistance of a fully loaded train may be safely assumed at 4.65 pounds per ton, at speeds up to ten miles an hour. '^ For a more extended reference to this subject, see Chapter V, Part II pp 236-238. TRANSPORTATION 359 1. Journal-friction, varying from 4 pounds per ton-weight of loaded train to 9 pounds per ton-weight of empty train, exclusive of locomotive and tender. 2. Track-resistance, covered by allowance for journal-friction in slowly moving trains. 3. Atmospheric resistance, virtually negligible in slowly moving trains. 4. Grade-resistance at 20 pounds per ton-weight of entire train for a 1 per cent, grade, and proportionately for other gradients. The practical allowances for the first three causes of resistance apply under all conditions of train-service, whether on a straight and level track or on a gradient compensated for curvature.' The effect of journal- friction upon the tractive power of a locomotive is greatest in starting a train, which is often only accomplished by "taking up the slack" and thus successively overcoming the inertia of each car in the train. This initial resistance has been found to reach 20 pounds per ton-weight of the train, but is only momentary. ^ As soon as the train is under way, the resistance from journal-friction rapidly diminishes to its normal amount, which is usually attained in a distance of five-eighths of a mile and between the speeds of seven and ten miles an hour.^ Within this range of speed, the normal tonnage-capacity of a locomo- tive is determined by dividing its tractive power, available at the tender- drawbar, by the assumed resistance of a train of fully loaded oars on a straight and level track. From this result, a reduction is to be made for the resistance over the ruling-grade on the division. In practice, it is not safe to utilize more than 90 per cent, of this rating, in order to provide for starting a loaded train when stopped on the ruling-grade. In extremely cold weather, there should be an additional reduction of 10 per cent. Furthermore, locomotives are not always in condition to work up to their theoretical capacity. For all these reasons, the allowance to be made from the rating-sheet for each locomotive in its class must be left, in a great measure, to the discretion of the yardmaster and of the dispatcher, who should be familiar with the local conditions. In fact, the economic conduct of freight-train service is practically dependent upon the efficiency of these officials and of the firemen.* Use of "Pushers" There are several ways of increasing the average train-load over a division. Where there is, an exceptional gradient in excess of the ruling- grade, or where the maximum gradient is concentrated at a summit, the ' For resistance from uncompensated cm-vature, see Appendix VII, Note III. ^ The use of sand in starting increases the track resistance ; it should not be used over an interlocking plant. 'See "Railway Location," Wellington, p. 512. ' For further information on this subject, see Appendix VII, Notes IV to VI. 360 EFFICIENT RAILWAY OPERATION assistance of a "pusher" (or "bank-engine," in British railway-parlance) is of advantage for this purpose. The type of locomotive for a pusher, as to wheel-arrangement, should be suitable for running backward down grade, and its tractive power should be limited to the exceptional grade- resistance to be overcome by a locomotive with the average train-load. The use of pushers with excessive tractive power is not only an unnecessary expense, but it also tends to increase the size and weight of the regular road-locomotives beyond the requirements of the normal freight-traffic. The average train-loads on a division may be sensibly increased by the use of pushers on intermediate maximum grades. This is done on inter- mediate gradients of 0.4 per cent, on the Hudson River Division of the New York Central Railroad, which is on a level for nearly 95 per cent, of its length. The trains should be kept far enough apart to avoid detention in waiting for the return of the pusher, or the unnecessary use of a second locomotive for this purpose.' Where the freight-trains are not sufficiently frequent to keep a pusher profitably employed on an exceptional grade, resort is sometimes had to the expedient of "doubling the gra,de." This practice is of questionable advantage. The presence of part of the train unattached to a locomotive on the main line is an obstruction to the traffic and a probable cause of accidents. The growth of traffic on a single-track line is frequently accompanied by concentration of train-movements on Certain divisions, requiring' ad- ditional passing-points and consequent delays in train-service. This condition may be somewhat relieved by consolidation in heavier trains drawn by two locomotives. This plan calls for long lap-sidings, which are gradually lengthened into running-tracks. Such measures are but palliative in the postponement of double-track operation.^ Double- headers are also used on double-track lines, operated under block-signals, for reducing the number of trains in blocked sections and thereby expedit- ing train-movements. It is however preferable to duplicate trains, where the first section can be efficiently protected against the following section.' Tonnage Distribution in Trains. Tonnage Rating The difficulty in adjusting the tonnage of a train to the rated capacity of a locomotive, may be inferred from the following statement of the factors which enter into its determination ; viz. 1. Rate of ruling-grade. 2. Degree of imcompensated curvature. 3. Car-journal lubrication. 4. Average gross weight per car. 1 See "Railway Location," p. 591. 2 See Chapter V, Part II, p. 251. 8 Double-heading is a current practice in Europe, even for the fastest trains ; being necessitated by the use of light locomotives, or in order to reduce the number of trains in the blocks. TRANSPORTATION 361 5. Condition of rails. 6. Weather conditions. 7. Character of traffic on the line. 8. Running time allowed. 9. Location and extent of passing-tracks.* From this statement it will be seep that the rating-sheet of a locomotive is only of value as a basis for the establishment of its maximum capacity on slowly moving and fully loaded trains, and that the reduction to be made in its application in practical service is governed by conditions which are largely empirical, and that it could be more satisfactorily established, with reference to any particular division or service, by occasional dyna- mometer-tests. At intermediate junctions on a division,, the transfer of tonnage may be provided for with economy in train-service by a judicious distribution of loaded cars among the through trains. Loaded cars taken into a train at one terminal are dropped at the junction and other loads there taken on for the other terminal. There is also what is known as "turn-around" service at intermediate points where a marked change occurs in the ruling- grade. On the lighter part of the line, the tonnage-rating may be increased while in the farther direction it must be diminished ; or lighter locomotives may be employed on the easier grades. The same quantity of draw-bar pull may be absorbed in moving a heavy train at a low rate of speed, or a lighter train at a higher rate. Al- though the tonnage rating-sheets are based on a speed of ten miles an hour, it is not practicable to restrict all classes of traffic to that rate of speed and meet the requirements for social efficiency. Live-stock, perishables and other commodities of a special character are, therefore, moved at higher speeds. There are also economic reasons for a higher rate of speed for freight-trains. On lines with dense passenger-traffic, lighter loaded freight- trains may move at higher speed with fewer passing-points with passenger- trains, and with a saving of time at sidings on the road. This is the wsual practice in Great Britain and, with the increasing pressure for legislative regulation for shorter hours of labor and the frequent occurrence of car- shortage, it is probable that a similar course will be pursued in our own railroad-operation, wherever the passenger-traffic is of commensurate importance. It has been determined experimentally that, "with long and heavy trains it requires less fuel with the same engine to run trains at 18 to 20 miles an hour than at 10 to 12 miles an hour."^ The consump- tion of fuel also depends, to a considerable extent, upon the time that a locomotive is on the road between terminals. The loss 1 See "Freight Termiaals," p. 155. 2 "Experiments on Lake Shore & Michigan Southern Raiboad," by P. H. Dudley. Trans. Am. Soc. C. E., October, 1876. 362 EFFICIENT RAILWAY OPERATION of energy by external radiation of heat is continuous, and is in proportion to the difference between the temperature within the boiler and that of the atmosphere. The percentage of loss between the boiler- pressure and the cylinder-pressure, which is some 15 per cent., and the waste of fuel while standing in sidings at passing-points, are more affected by the lapse of time than by the increased train-resistance, which is also lessened to some extent by the effect of momentum at higher speeds. Other favorable factors are now available in the greater tonnage-capacity of cars, the stronger draft-gear and the general use of quick-action air-brakes. In passenger-train service, economic efficiency becomes subordinated to social requirements. The capacity of the carriages may be established as to the number of occupants, but travelers can not be herded like cattle, nor can the carriages be marshaled in trains to conform to the rated power of locomotives. Speed supplants tonnage as the ruling factor in efficiency, and that property of matter known as momentum assumes far greater importance than in freight-service.^ Speed Requirements. Velocity Resistance The speed required in passenger-train service is only to be attained by tractive power considerably in excess of that required for overcoming the inertia of the train in starting, and, to maintain high speed, the evapo- rative capacity of a passenger-locomotive must be correspondingly greater in proportion to its tractive power than with a freight-locomotive. After the train is in motion and has acquired the speed due to the initial expendi- ture of energy for that purpose, any additional draft of tractive power would have the sole effect of accelerating the speed of the train, on a straight and level track. The effective draw-bar pull, however, decreases rapidly with acceleration of velocity. For this reason, an additional locomotive is frequently attached to a heavy, high-speed passenger-train.'' Conversely, a slight reduction of velocity at high speed considerably in- creases the effective draw-bar pull. The tractive power of a locomotive is not materially affected by the ascent of moderate grades. Where such grades occur as undulations, the average speed is sustained on the ascending grades by the momentum accumulated on the preceding descent, if the average grade of the undula- tions approximate a level, in which case the normal tractive power is efficiently applied in the ascent, and may be economically diminished in I Momentum is the Energy of Motion, as Inertia is the Energy of Position. It is the product of the mass of a body multiphed by its velocity. While the velocity is uniform, the momentum is directly proportional to the mass or quantity of matter in the moving body. While the mass remains unchanged, any change in momentum is due to a corresponding change in velocity. Its quantity may be measured in foot-pounds per second, and its sum increased, maintained or di- minished by an alteration in the factors of which it is a product. ' On a level, straight track, the draw-bar pull of a locomotive of 800 horse-power is virtually exhausted at a speed of 70 miles an hour. See Appendix VII, Note V. TRANSPORTATION 363 the descent. The energy thus developed is measured as work, in foot- pounds per second.! This work is performed in overcoming the resistance offered by ascending grades, in rounding curves by the friction of the rolling wheels against the rails, by inequalities in the track and by the internal friction of the axles in their bearings. Resistance from any of these causes must diminish the momentum of a train, unless its speed is maintained by a continuous absorption of energy in tractive power. It is necessary to keep these propositions in mind in any discussion of the economic apphcation of tractive power in passenger-train service. But other causes of train-resistance, which are negligible in their effect upon slowly moving trains, engender more serious consequences at high speeds; atmospheric resistance, for instance, and the resistance arising from oscillations and concussions on the track. The rolling-friction be- tween the wheels and the rails is materially increased by irregularities in the track, . whether in line or in surface, and proportionately with the increase of momentum at high speeds. It is generally assumed that velocity-resistance varies with the square of the velocity, somewhat in accordance with the formula, R = FV^+C in which i^ = friction, y = velocity and C = curvature. It is not possible to separate this aggregated velocity-resistance into its constituent elements including those due to the resistance of the track and of the atmosphere. Experimentally, it has been shown that the effect of track-resistance is the greater of the two. But at high speeds, the skin-friction of the train and the partial vacuum created at the rear of the train and between the cars increases the resistance proportionately more than the head-resistance is increased. The introduction of vestibuled trains seems to eliminate the partial vacuum between the cars. The effect of vertical alignment varies with changes of gradient, as to degree, in accordance with the immutable law of gravitation and with the ascending or descending direction of the train. Even on a straight, level and smooth track, the alternating and unbalanced impulses in the steam-cylinders induce changes in horizontal direction, which are as in- cessantly restricted by contact of the wheel-flanges with the rail. There is a consequent impact and oscillation that vary with the train-momentum, and with the accompanying wave of deflection in the track-surface. Re- tardation from wheel-and-rail contact increases with increasing curvature until the permissible hmit has been reached. The velocity of the train is further retarded by the friction' of the wheel-flanges against the sides of ' The work done in running a mile by a locomotive with tractive power, of 20,000 pounds is 20,000 X 5280 feet = 105,600,000 foot-pounds. Such a locomotive, in running a mile in four minutes, develops horse-power equal to „„ ' . ' . = 800 oo, UUU X 4 H. P., or, running a mile in five minutes, ' ' ' =640 H. P. 364 EFFICIENT RAILWAY OPERATION the rail-heads and by the slipping of the wheel-treads upon the top-surface of the rail around curves; also by the journal-friction and the internal friction of the locomotive machinery. The rapid decrease of draw-bar pull at high speeds may be neutraUzed to some extent by a judicious use of the momentum stored up in the train, and especially on undulating grades. In "drifting," or descending a grade without steam or braking, a train acquires accelerated velocity in proportion to the degree of gradient. Starting from a state of rest, at the top of a 0.5 per cent, grade, it should attain a speed of 25 miles an hour in the distance of 222 feet.' A train approaching the foot of such a grade at a speed of 50 miles an hour has attained momentum sufficient to hft it vertically 88.75 feet before coming to a state of rest, and to carry it up a 1 per cent, grade for a distance of 8875 feet.^ The initial rate of speed could be maintained for this distance by the locomotive's supplying only the tractive power which had been required to attain the same rate of speed on a level track. There is a degree of slope at which the train-resistance would just balance the acceleration of gravity and a train remain at a state of rest. This is the "Grade of Repose." On such a grade, a train should theo- retically descend continuously at its initial rate of velocity, according to the First Law of Motion. In this state of equilibrium, any accession of tractive power would serve entirely to increase that velocity. Conversely, on an ascent, the tractive power would have to be increased sufficiently to overcome the grade of repose, in order to maintain the initial rate of velocity. It follows that any ascending grade is equivalent, in its retard- ing effect upon the velocity of a train, to the actual rate of gradient plus the grade of repose, and that the accelerating effect of a descending grade is equivalent to that of the actual rate of gradient minus the grade of repose.' The use of the momentum of the train in freight-service to eke out the over-rated power of the locomotive is a practice of questionable value. It encourages excessive speed on descending grades and is a hindrance to train-movements when an emergency-stop is made on an ascent. The effect upon long freight-trains of sudden changes from a descending to an ascending grade frequently causes a train to part because of broken couplings, though this trouble has been diminished by the general in- 1 For velocity of "drifting" trains, see Appendix VII, Note XII. 2 See Appendix VII, Note XI. 'A train resistance of 7 pounds per ton would be equivalent to a grade of repose rising 7 feet in 1000, or of 0.7 per cent., which is about 37 feet to the mile. As the train resistance in pounds per ton varies with the velocity of the train and with its length, the grade of repose varies accordingly. On account of the initial journal-friction, which is equivalent to about 20 pounds per ton, no car will start of itself on a 0.7 per cent, gradient, or of 37 feet to the mile ; but will generally do so on a gradient of 1.1 per cent, or of 58 feet to the mile. See Appendix VII, Note X; and also "Railway Location," p. 341. TRANSPORTATION 365 troduction of closely-coupled vertical-hook couplers, with stouter draft- gear, and can be entirely eliminated by the intervention of vertical curves at the change of grade. Brake Action and Brake Efficiency While the mass of a moving body remains unchanged, its momentum can only be diminished by a decrease in its velocity. With the withdrawal of tractive power, the speed of a train is decreased either by the resistance due to gravity, as in ascending a grade, or in its transformation into heat by impact or by frictional resistance. This negative action may be sup- plemented by brake-friction for the purpose of acquiring positive control over the motion of a train. In the application of brakes, the momentum of the train is affected indirectly, through the transformation of the rotary motion of the car- wheels into heat. As this frictional resistance affects only the wheels to which the brakes are applied, the efficiency of the brakes depends primarily upon the proportion of the weight carried by the braked wheels to the total weight of the train. With increasing pressure of the brakes, a point is reached at which the rotary motion of the wheel is entirely arrested, and its rolling-friction upon the rails becomes changed to sUding-friction, or "skidding," to the injury of both the wheel and the rails. The maximum brake-pressure should, therefore, be limited to two-thirds of the load on the braked wheel. Consequently, the maximum efficiency of the brakes on all the wheels in a train can not exceed that proportion of the total weight of the train, as measured by the pressure in pounds per ton required to skid all the wheels. As the rotary momentum of the wheels varies with the swiftness of their revolution, it was formerly assumed that the brake-pressure sufficient to arrest their revolution would decrease propor- tionately, and that the continued eliiciency of the brakes could only be maintained by a corresponding decrease of pressure upon the wheels. But the experiments made by Mr. George Westinghouse and Sir Douglas Galton, in 1878, proved that the pressure sufficient for skidding depended upon the extent of adhesion to the rails.' The speed of the train is a controlling factor as to the time and distance within which it is practicable to bring a train to rest by brake-friction on a level track. This problem is further affected, on an ascending or descending grade, by the influence of gravity ; and the retarding effect of the brakes may be assisted, in an emergency, by reversing the action of steam in the cylinders of the locomotive. The maximum efficiency of brake-friction, as theoretically established, is far from being attainable by means of the hand-brake. Its application is progressive, according as the brakes are applied by one or more persons. Its efficiency further depends upon the vigor of the brakeman, as well as upon the leverage of the brake-gear, and is empirically established as I See Chapter IV, p. 109. 366 EFFICIENT RAILWAY OPERATION from 2^ to 5 per cent, of the load on the braked wheels. Hand-brakes are not capable of the simultaneous action required of an effective train-brake, for which reason, except as applicable to the movement of individual cars, they have been superseded by power-brakes operated by compressed air. In connection with closely-coupled draw-gear and improved brake ap- pliances, the quick-acting electro-pneumatic air-brake acts upon a train as upon a single mass and, by this means, the maximum efficiency of brake- friction has been closely attained. The air-brake apparatus is used to advantage in moving a train backward, by attaching to the rear end of the air-pipe a section of hose provided with an air-valve to control the speed of the train and a signal-whistle, which are operated from the rear platform.' A train of twelve steel passenger-cars, weighing in all 920 tons, develops a kinetic energy of over 100,000 foot-tons at a speed of 60 miles an hour ; yet with the quick-acting brakes, it can be stopped on a level track in about 21 seconds in a distance of 1000 feet. The transformation of this tremendous energy into heat is the equivalent of about 20,000 nominal horse-power. This statement aids in forming a conception of the work done in stopping such a train, sufficient to lift the whole train vertically 128 feet in 21 seconds, or to maintain it at full speed for six miles on a level track. The use of electric traction has introduced some exceptional problems in connection with transportation-efficiency. These have been, for the most part, discussed in the chapter on Motive Power. One of especial interest relates to "regenerative braking," or the restoring of energy to synchronous motors from the momentum of trains on descending-grades.^ There is said to be a tendency in electric operation to overload the heating capacity of the motors by quickening the schedules, increasing the number of stops or adding trailers to the trains. On the Swiss electric railways, with a permanent draw-bar pull of 22,046 pounds, the maximum speed on faUing gradients of 27 in 1000, or of 137 feet to the mile, is 40.4 miles an hour for passenger-trains. On similar rising gradients, a speed of 31 miles an hour is attained with trains weighing 300 tons, and of 28 miles an hour for freight-trains of that weight. Redttction in Teain-weight. Efficiency in Train-loading The application of tractive power in train-service has now been fol- lowed as a cycle of traction from a state of rest through one of motion to its return to a state of rest. There is yet another aspect of it as a phase of economic efficiency. The profitable application of tractive power is not to be measured by the total amount absorbed in keeping a train in 1 For the development of power-brakes, see Chapter IV, pp. 105-112 » See Chapter III, p. 84. TRANSPORTATION 367 motion. That part which is required to keep the dead-weight of the cars in motion earns no money for the railroad company. To whatever ex- tent the unproductive weight can be diminished, to that extent the pro- ductive weight may be increased by the absorption of the same amount of energy in tractive power and without reduction in the rate of speed. It is therefore an element of the efficient use of tractive power, from a com- mercial point of view, that the dead-weight of a train shall be maintained at the lowest point consistent with the safety of its contents. The foundation of economic efficiency in freight-service is the utiliza- tion of each car to its fullest capacity, and the first step in this direction must be taken at the points of origin of the traffic. In mineral districts, these traffic-points are scattered along the line or division, or on short branch-lines, and are served by special trains which distribute empty cars as required and deliver them loaded at the farther terminal, as the tractive cai)acity is attained. In a similar way, in agricultural districts, grain-cars are loaded at station-elevators, cotton at local presses and perishables at cold-storage stations. Forest-products are likewise taken from the mill- sidings and petroleum-products from the refineries. The above-mentioned products constitute by far the greater part of the tonnage of our railway- system and, in order to insure full loading, a considerable and expensive train-service is required before they become incorporated in the general volume of traffic. This is also the case with the distribution and collec- tion of cars in large cities with several freight-terminals. Even where a charge is made for such service, it is based rather upon the expectation of profit to be derived from the subsequent haul in through-trains than upon its actual cost. The current demand for cars at points of origin and the sources whence they may be obtained must be ascertained sufficiently in advance for the cars to be promptly supplied. This information must cover the ownership and the character of the available equipment. It is desirable to keep the "home" cars upon the company's own line, and to load "foreign" cars in a direction homeward. The suitability of the empty cars for the desired purposes must also be kept in view. As there afe about eighteen kinds of freight-cars, it is necessary to know the character and location of each of thousands of cars, whether they are loaded or empty, to have this informa- tion collected by telegraph or telephone, and to have it collated daily and promptly in the office that controls the distribution of freight-equipment. This result can be secured only by a systematic method under intelligent supervision.^ When the demands for cars have been supplied, the next step is to in- sure that they shall be loaded to their normal capacity. In a committee report of the American Railway Association, it was reported that in 1909 not more than 60 per cent, of the total car-capacity had been utilized, and ' See "Economies of Railway Operation," Byers, pp. 485-492. 368 EFFICIENT RAILWAY OPERATION that the increases in the individual car-capacity had not been accompanied by an increase in the average load per car.' This statement does not apply to the mineral traffic, which constituted 57 per cent, of the total tonnage in 1914. It may be assumed that the owners of the private refrigerator and tank-car lines utilize them to the best advantage with their own products, and that the same may be said of the flat cars which are principally used for lumber. Excluding open-top, refrigerator, tank, stock, and flat cars, as being practically fully loaded at points of origin, there remained in 1914, 1,043,796- box cars, or about 40 per cent, of the total equipment of the railway companies, with a nominal tonnage capacity of 36,365,350 tons, or an average of 35 tons per car.^ The total tonnage from points of origin in 1914 was 1,094,123,895 tons. Deducting the tonnage in products of the mines, of animals and of forests, amounting to 764,092,081 tons, there re- mains a tonnage of 330,031,814 tons, which represents approximately the box-car freight in that year.' This tonnage could have been moved in about nine trips of the total box-car equipment, making a trip about every six weeks. From this statement, a conception may be formed of the extent to which box cars are light-loaded or detained at terminals. In the matter of light-loading, consideration must be given to the character of the commodities to be transported. Many of them fill the capacity of a car for space before its limit for weight has been reached.* The minimum allowances in classification-rates contribute to light-loading and have not been increased with the larger tonnage capacity of box cars. Light-loading is also preferred to starting a foreign car homeward empty. The average car-load varies greatly with the physical or industrial en- vironment in which a railroad is operated, as is shown in the Transporta- tion Statistics for 1914, in Appendix VII, Table XVIII. The maximum of 46 tons was attained on the Butte, Anaconda & Pacific Railway, a mineral road with 26 miles of main line and 36 miles of branches. The minimum loading of 10 tons was on the Northwestern Pacific Railroad, with a main line of 281 miles and 182 miles of branches. Typical car- loadings on important lines in different regions of the United States are given in Appendix VII, Table XVI. Box-car freight, for the most part, passes through the freight-houses to be weighed, billed and sorted for loading. Each shipment involves the services of a receiving-clprk, a weigher, freight-handlers, truck-hands, a tally-man and stevedores. The monthly tonnage at some thirty stations averaged 1,000,000 tons of track-freight and 400,000 tons of house-freight.* At the Wood Street Station in Chicago of the Chicago & North Western ' For other instances, see Appendix VII, Tables VIII and XIV. 2 Appendix III, Table III. ' Appendix VI, Table XIV. * Appendix VII, Note XIII. 5 "Economics of Railway Operation," p. 517. Track-freight is loaded from -wagon directly into cars ; house-freight is placed in, and taken from, storage. TRANSPORTATION 369 Railway, on an average, 130 cars are loaded daily with from 15,000 to 20,000 packages. In the large cities, the requirements of the local traffic induce the establishment of many separate freight-stations. The New York Central Railroad has eight in New York City. The Pennsylvania Railroad has seven in Pittsburgh and thirty-one in Philadelphia. In Boston, the New York, New Haven & Hartford Railroad has twelve and the Boston & Maine Railroad has fifty.^ Light Loading and "L. C. L." Freight The economical handling of "L. C. L." freight (less-than-car-load lots) is a troublesome problem. If such freight for different stations is loaded into one car, the train is delayed in discharging it, even though it be loaded carefully in station order. If the shipments for each station be loaded separately, there is a waste of tractive power and of car-mileage in handling partially loaded cars, a loss of time in setting them out of the train, and a considerable loss in the use of cars so employed. Ship- ments of this character are therefore seldom moved in through-trains. Between way stations on the same division, they are usually handled in and out of the "peddler-car" by the crew of the way-freight train. If destined to or beyond the end of the run, they are loaded together in a "straight car" to be handled at the terminal of the division or, at junction points, at the transfer station.^ Cars with package-freight should be placed next to the caboose, so that the small lots may be unloaded at the freight-house while the fully loaded cars are being set off from the head of the train. It becomes exceedingly difficult to secure maximum loading with minimum delay for shipments of package-freight on roads with many side-lines and junction points. Elaborate instructions are issued for this purpose.' L. C. L. freight is the least profitable class of traffic. This is not only as to train-movements but also with respect to its receipt at the point of departure, its handling and billing and the care required in loading it ; and again, as these operations are reversed at the point of delivery. It is with this class of freight that the claims for loss and damage are propor- tionately great and expensive of adjustment. Special attention to these matters by committees of the American Railway Association has resulted in codifying in great detail the rules governing them.^ The conditions prevailing on the British railways, where the miscella- neous traffic includes much of our perishable and express business, are exemplified in the following statement : 1 "Freight Terminals," p. 276. 2 See Chapter V, Part II, p. 282. s See "Freight Terminals," p. 317. * For further information on this subject, see "Economics of Railway Opera- tion," pp. 517-538; and "Freight Terminals." 2b 370 EFFICIENT RAILWAY OPERATION Loading of Miscellaneous Shipments Total shipments .... ... Total packages Total weight, pounds Average weight, per shipment, pomids Average weight, per package, pounds . Number of cars used Number of destinations .... Average load, tons (2000 pounds) . . Gheat Northeen London & North Railway Western Railwat 985 6,201 4,427 23,607 273,800 2,029,440 278 327 62 86 72 379 53 720 L83 2.67 Car-capacity, 11.2 tons. Empty weight, 6.75 tons. The average loading on the Great Northern Railway was but 16 per cent, of the car-capacity, and 24 per cent, on the London & North Western Railway. The 72 car-loads on the Great Northern Railway could have been loaded in 12 cars, and the 379 car-loads on the London & North Western Railway in 91 cars. The saving, in the former case, of 60 cars and of 405 tons of dead-weight would have amounted to 65 per cent, of the actual gross train-weight ; in the latter case, of 288 cars and of 1937 tons of dead-weight, amounting to 54 per cent, of the actual gross train-weight. Similar experience on railroads in the United States is given in Appendix VII, Table XV. In an inspection of ten cars, taken at random, they were found to be loaded to less than one-half of their total weight-capacity of 940,- 000 pounds, and by no means to their space-capacity. Several plans have been proposed for reducing the proportion of light-loading from this cause. The American Railway Association has endeavored to restrict the receipt of L. C. L. freight for specific destinations on certain days in the week, with the object of concentrating the shipments for each destination in the same car. A more extended use of mixed trains, as suggested in Chapter VI, page 313, seems worthy of consideration for this purpose. Train Make-up and Average Car-load In classifying cars for train-movements, trains may be made up with either of the following objects in view : 1. To move the greatest net tonnage with a locomotive of a given tractive power. 2. To move the greatest net tonnage in a given period of time with the same locomotive. These are economic applications of tractive power. Or, 3. To move certain commodities at higher speed to meet commercial requirements. In making up trains, the tonnage of locomotives, as classified in the rating-sheet, must be modified to meet these varying requirements. The TRANSPORTATION 371 actual condition of each locomotive and the prevailing weather-conditions have also to be considered so that, to insure efficient service, the tonnage of each train must be left somewhat to the discretion of the experienced yardmaster. The frictional resistance of an empty car is about double that of a loaded car, weight for weight. This fact is to be kept in mind in making up a mixed train ; since the greater the number of cars in a train, the greater is the resistance of such a train of the same gross weight. This is especially of importance in making up fast-freight trains of cars that are variously loaded in proportion to their nominal capacity.^ Cars should also be so placed in trains as to facilitate their distribution in accordance with their respective destinations, thus saving much intermediate delay in the re-arrangement of trains. In the movement of time-freight, that is, of commodities to be delivered at distant destinations within a definite period, this becomes an important matter. The relative position of empty cars in a train is supposed by trainmen to affect the train-resistance. They think that empty cars should be at the rear of the train. However this may be, the resulting shock to a long train should be less when brakes are suddenly applied, or when steam is shut off on a descending grade, if empty cars are not next to the locomotive. This effect has been lessened with the general use of close couplers and quick-action brakes. Notwithstanding these im- provements, it is questionable whether the length and weight of heavy freight-trains has not about reached a practical maximum with trains of eighty to a hundred fully loaded cars drawn by locomotives of weight and power sufficient to take such trains up a 1 per cent, grade at a speed of ten miles an hour. It is of the first importance that trains should be "made up" to the full tonnage-capacity of the locomotives to be attached to them. It is with this object in view that the freight-locomotives on a division are classed on a rating-sheet. In the earlier development of this plan, loco- motives were rated by the number of loaded cars, whether fully loaded or not ; two empty cars being rated as one loaded. With the introduction of cars of greater capacity, it might occur that a gross tonnage of 2000 tons would vary between 27 and 65 car-loads. The tonnage-rating was subse- quently made from the way-bills, with closer approximation to the tractive power of the locomotive. The estimated tonnage-rating was based on cars of 40 tons gross weight, so that a locomotive with 36,000 pounds draw-bar pull might be rated, to suit the ruUng-grade on a division, at 2000 tons in a train of fifty of such cars. As any pressure on an axle-journal, within reasonable limits, does not materially increase the friction, the frictional resistance of a car is virtually the same whether it be empty or loaded, though the resistance offered by the car per ton will vary with the gross weight of the car.. The total re- 1 See Appendix VII, Note IV. 372 EFFICIENT RAILWAY OPERATION sistance of the train will vary according to the gross weights of the several cars of which it may be composed and, consequently, as to whether they are fully or but partially loaded. The economic results of hauling trains of cars of various nominal capacities and with full or partial loading, as discussed in Appendix VII, Note XIV, may be stated as follows : With the same draw-bar pull, the greater the capacity of the car, the fewer may be the cars in the train and the greater will be the net tonnage ; while the lighter the cars are loaded in proportion to their capacity, the greater will be the waste of tractive power. In actual operation, there is a wide variation in the make-up of trains, with respect to tonnage and to the relative proportion of loaded and empty cars in a train as will be seen by reference to the Transportation Sta- tistics for 1914, in Appendix VII, Table XVIII. The heaviest average car-load and train-load and the greatest number of loaded cars and of total cars in a train, were in the Eastern District ; with averages of 23.4 tons per loaded car, and of 557 tons per train of 23.8 loaded and 12.0 empty cars. The Southern District averaged 21.3 tons per loaded car and 427 tons per train of 20 loaded cars and 10 empty cars. The Western District averaged 18.4 tons per loaded car and 472 tons per train of 23.2 loaded and 10.0 empty cars. In the Eastern District, the maximum average car-load of 44.7 tons was on the Bessemer & Lake Erie Railroad, and the maximum average train-load of 1162 tons was on the Pittsburgh & Lake Erie Railroad, with an average haul of 63 miles. The maximum average number of cars per train was 34 loaded and 17 empty on the Lake Shore & Michigan Southern Railway. The minimum average in all of these respects was on the Atlantic City Railroad, of 13.2 tons per loaded car and 141 tons per train of 11 loaded and 4 empty cars, with an average haul of 23.8 miles. In the Southern District, the Virginian Railway averaged a car-load of 45.5 tons and a train-load of 1409 tons in a train of 31 loaded cars and 26 empty, on an average haul of 354.77 miles ; a performance unequaled elsewhere in the United States, and probably in the world. On the other hand, the Florida East Coast Railway averaged a car-load of 9.3 tons and a train-load of 158 tons in an average train of 17 loaded and 11 empty cars on an average haul of 164 miles ; these conditions being due to a traffic principally in perishables in refrigerator-cars. The heaviest average car-load in the Western District, and in the United States, was 46 tons on the Butte, Anaconda & Pacific Railway on an average haul of 27 miles, and the heaviest average train-load was 1108 tons on the Duluth, Messaba & Northern Railway on an average haul of 73.68 miles. The Great Northern Railway averaged a train of 32 loaded and 14 empty cars on an average haul of 224.59 miles.^ ' See Appendix VII, Table XIX. TRANSPORTATION 373 Tkain Mileage Statistics In the Eastern District, the greatest mileage per mile of line was on the Philadelphia & Reading Railway, of 8297 freight-train-miles and 6013 passenger-train-miles, a total of 14,310 train-miles, and an average of 22.7 freight-trains and 16.5 passenger-trains, or 39.2 trains per day on a line- mileage of 1120 miles of main lines and branches. This freight-train mile- age was exceeded on the New York, Chicago & St. Louis Railroad, which averaged 9679 miles with 26.5 trains per day on a line-mileage of 567 miles. There was an average of 37.7 passenger trains per day on the Long Island Railroad on a line-mileage of 398 miles. The transportation business of New England is handled almost wholly by two railroad corporations, the Boston & Maine and the New York, New Haven & Hartford, with nearly equal line-mileage of 2302 miles and 2003 miles, respectively; equal freight-tonnage of 34,752 tons and 24,996 tons; but unequal passenger- mileage of 896,081 miles and 1,600,476 miles respectively. As between the four trunk-lines, the train-performance is as follows : Average Train Performance. Trunk Lines. Eastern District. 1914 Baltimobe & Ohio N. Y. Centhal & H. R. R. R. Pennstlvania, R. R. Ebie R. R. Line-mileage 4,478 3,756 4,084 1,988 Per Mile o/ lAne Freight-train mileage . . Passenger-train mileage . . 4,743 3,726 5,414 7,064 7,414 6,211 5,389 4,565 Total 8,469 12,478 13,625 9,954 Per Day Freight-trains Passenger-trains .... 13.0 10.2 14.8 19.3 20.5 16.9 14.5 12.5 Total 23.2 34.1 37.4 27.0 In the Southern District, the Richmond, Fredericksburg & Potomac Railroad averaged a passenger-train mileage of 8756 per mile of line with 24 passenger trains daily, and a total train-mileage of 15,517 miles with an average of 40.8 trains daily; a performance unequaled elsewhere in the United States, though on a line-mileage of but 88 miles. The greatest average freight-train mileage in this district was on the Cincinnati, New Orleans & Texas Pacific Railway, 7958 miles on a line-mileage of 337 miles. As between the principal trunk-lines in this district, the train-performance is as follows : 374 EFFICIENT RAILWAY OPERATION Average Train Perpobmancb, Southern Trunk Lines, 1914 SOTITHBEN Atlantic GOAST Line Seaboabd Air Line Louisville & Nashville Illinois Cen- tral R. R. Line-mileage .... 7,033 4,646 3,084 4,937 4,767 Per Mile of Line Freight-train mileage . Passenger-train mileage 2,240 2,611 1,833 1,843 1,837 1,843 3,645 2,156 3,859 2,765 Total 4,851 3,676 3,680 5,801 6,624 Per D(iy Freight trains .... Passenger trains . . . 6.1 7.2 5.0 5.1 5.0 5.3 10.0 6.0 10.6 7.6 Total 13.3 10.1 10.3 16.0 18.2 In the Western District, the Chicago & Alton Railroad averaged a passenger-train mileage of 3353 miles per mile of line with 9.2 trains daily, and a total train-mileage of 6536 miles with 17.6 trains daily on a line- mileage of 1033 miles. The greatest freight-train mileage was on the Duluth & Iron Range Railroad of 3546 miles with 9.7 trains daily on a line-mileage of 282 miles. As between the Transcontinental lines, the train-performance is as follows : Average Train Performance, Transcontinental Lines, 1914 ) Chicaoo, Milwau- kee, & St. Paul Atchison, TOPEEA, & Santa Fe Great Northern southebn Pacific Northern Pacific 1 Union Pacific Line-mileage . . 9,684 8,346 7,780 6,457 6,325 3,614 Per Mile of Line Freight-train mileage Passenger-train mileage 2,044 1,814 1,697 2,218 1,244 1,603 1,522 3,247 1,453 1,900 2,141 2,808 Total . . . 3,^58 3,915 2,847 4,769 3,353 4,949 Per Day Freight trains . . . Passenger trains . . . 5.6 5.0 4.6 6.1 3.4 4.4 4.2 9.0 4.0 6.2 5.9 7.7 Total .... 10.6 10.7 7.8 13.2 9.2 13.6 TRANSPORTATION 376 • Large Cars and Fast-freight Trains The investment in cars of large capacity and the cost of maintaining them, are relatively less in proportion to tonnage; while they afford increased facility in train-movements, because of the shorter length of trains of equal weight. A train of 30 cars of 50 tons' capacity with gross weight of 2010 tons, as compared with a train of cars of 30 tons' capacity and of equal gross weight, will carry 1500 tons net weight as against 1307 tons, and the train will be shorter by 500 feet. The economic application of power to net tonnage principally affects the transportation of the mineral-products which constitute the larger proportion of railway traffic and which are moved at a slow speed. It does not follow that this is the most efficient mode of transportation when the element of time is taken into consideration ; for, when the volume of traffic is great, a larger tonnage may be moved in a given period by quick- ening the speed of the train. Other reasons for the opinion that the economic rate of speed for heavy trains is over ten miles an hour, have been stated already in this chapter.^ There are commercial inducements for moving certain commodities at higher rates of speed than are warranted by an economic application of tractive power. As these commodities occupy a much greater space than minerals, in proportion to their weight, they are hauled in separate trains ; as fast-freight trains, time-freight twns, live-stock trains and refrigerator- trains. The fast-freight trains originated in the period of fierce competi- tion between the trunk-Hnes for the west-bound merchandise-traffic from the North Atlantic ports, with a separate equipment and organization on each of the rival lines. ^ That class of commodities is still transported by trains, the integrity of which is maintained, as far as" practicable, at divi- sion-terminals. Their average rate of speed is over 15 miles an hour, and they, are given preference on the road over trains loaded with bulk-freight. The time-freight trains are a higher class of fast-freight trains, loaded with commodities of such a character as to require special facilities and care in transportation; or of such intrinsic value that delay in delivery 1 Assuming that, on a certain ruling-grade, a train of 50 cars of 40 tons' gross weight, or of 2000 tons' gross weight and 1150 tons' net load, can be moved at a speed of 10 miles an hour by a locomotive with 36,000 pounds' tractive power, then, accord- ing to the rating on the Pennsylvania Railroad as per Appendix VII, Table XIV, the same locomotive would take a train of the same cars, fully loaded, over the same grade at a speed of 16 miles an hour, provided that the train weighed but 69 per cent, of 2000 tons, or 1380 tons gross and 805 tons net, in a train of 35 cars, each with 23 tons of net load. (See Appendix VII, Note IV.) If, on a division of 100 miles in length, this locomotive could, haul 1150 tons net in 10 hours, at a speed of 10 miles an hour and 805 tons in 6.25 hours, at a speed of 16 miles an hour, then in 50 hours the total net tonnage hauled would be 6760 tons at 10 miles an hour, or 6440 tons at 16 miles an hour ; which would be an increase of 12 per cent. 2 See Chapter VI, p. 335. 376 EFFICIENT RAILWAY OPERATION counts up in interest on the investment in them; as for instance, silk or tea from the Orient through Pacific ports. These trains are run on schedules at rates of speed up to thirty-five miles an hour, passing through inter- mediate terminals unbroken. The cars with time-freight are identified by "symbols" composed of letters and numbers indicating their routeing and destination.^ The symbols also serve to attract attention to cars so carded in case of accident or delay on the road. Time-freight trains are the especial care of train-dispatchers and, to further insure satisfactory service, special locomotives and train-crews are assigned to such trains. Perishables and live-stock are also given preferential service. Time- freight service is an important feature in British railway operation, as described in Chapter VI, pages 349, 350.^ Economy of Full Loads The stress that is laid upon fully loaded trains by railroad managers is evinced in the Proceedings of the American Railway Association, in which it is stated that, in 1909, by increasing the average loading by one ton per car, the available equipment could be virtually increased by 69,457 cars, and an increase of one mile per car per day would mean 79,395 more cars, with daily average earnings for 200,000 cars.' The importance of fully loaded trains is so- strongly impressed upon trans- portation oflacials that it is to be expected that the ambitious yard- master will strive to keep his tonnage-reports up to the rating-sheet re- quirements and that, in consequence, there will be occasional reports of overloaded trains. In such cases the responsibility may rest with the Motive Power or the Rolling-stock Department rather than with the Transportation Department; but the dynamometer-test is the only standard by which this can be fairly determined. Classification Yards, Freight and Passenger As cars are concentrated at division or junction terminals, the marshal- ing of them promptly in trains, in such a way as to insure the economical ' See "Freight Terminals." 2 As to live-stock, see Chapter V, Part II, p. 281, and "Freight Terminals " p. 184. 3 On the Chicago, Burlington &. Quincy Raihoad, from 1900 to 1913, the aver- age gross tonnage per traia, in the principal locomotive-pool, was increased from 1200 .to 3000 tons; and the average amount of car-mileage decreased from 84,000 to 48,000 miles. In 1900, twenty or more trains per day consumed, for 160 miles, an average of 7 hours at a speed of nearly 23 miles an hour. In 1915, the aver- age time was 14 hours, with the rate of speed reduced nearly one-half, and the net load per train increased from an average of 250 tons to 600 tons. In 1910, the Norfolk & Western Railway handled 6,722,495,887 ton-miles of freight with 10,578,541 revenue train-miles. In 1916, 12,131,187 revenue train-miles were required to move 11,795,891,557 ton-miles; being an increase of less than 15 per cent, in train-miles with an increase of 75 per cent, in ton-miles. The revenue train-load was increased from an average of 635 tons in 1910 to 957 tons in 1916, an increase of 50 per cent, in six years. TRANSPORTATION 377 use of tractive power, is an operation that requires suitable yard-facilities, experienced yard-crews and intelligent supervision. It is almost impossible for trains to be made up in proper order, or with full tonnage, in a yard that is badly designed or with insufficient track-room. At such points, the traffic becomes congested, with consequent derangement which affects the operation of the entire line. Some eight or ten years ago, on the Pennsylvania Railroad, it was found that the movement of traffic did not respond to the large expenditure that had been made to facilitate it. The speed of a car from Pittsburgh to New York averaged less than four miles an hour. It was standing still more hours than it was moving on the road. On another line, in three separate investigations, it appeared that from 81 to 84 per cent, of all the cars on the road were not in motion. On one of the largest roads in the country, the average time of all freight-cars in terminals on its line had been 18 hours, which was subsequently reduced to 13 hours. Five hours per day represents 5000 car-hours in a yard handling 1000 cars daily; which is equivalent to an addition of 213 cars in service, representing an investment of not less than $213,000. The saving of time in a well-designed classification-yard is well worth the expenditure of milhons of dollars. In a yard with tail-switching, a train of 60 cars was classified in 50 cuts in two hours, the switching- locomotive having in that time run 4.7 miles. In a "hump" yard, or sum- mit-yard, the same work was accomplished in half an hour, with a total run of 1.14 miles. The yard at Gardensville, N. Y., on the New York Central Railroad, has track-room for 21,000 cars and 260 locomotives, with capacity for handling 10,000 cars daily .^ The opportunity for economy of labor in a great freight-terminal may be understood from the statement that in the Chicago Terminal District of the Chicago & North Western Railway the yard-organization includes 70 locomotives, 24 yardmasters, 40 switch-tenders, 80 yard-clerks and 700 yardmen. In order to secure economic efficiency in such an organiza- tion for the prevention of delays and of possible blockades, everything should be in readiness for instant action as soon as a switching-list is pro- vided or information received of expected trains. For this purpose, a local telephone-system connected with the general yardmaster's office is of valuable assistance- The yard switches and signals should be kept in good condition, and day-and-night work will be greatly facilitated by eight-hour shifts. While switching is going on, the yardman at the head of a cut of cars should always be in view of the engineman. If he disap- pears from sight but for a moment, that should be considered as a signal to stop. By a strict observance of this rule, accidents from sudden colli- sions will be prevented. I For construction and operation of classification-yards, see Chapter V, Part II, pp. 275-280. 378 EFFICIENT RAILWAY OPERATION The service in classification-yards is much simpler at points of origin than at important junction-points and at division-terminals, where trains with differing equipment and variously loaded are arriving at frequent in- tervals, and are to be rearranged for departure at successive hours for separate destinations. Locomotives and cabooses are to be detached, cars and brakes inspected, bad-order cars designated to be set out on repair- tracks, refrigerators to be re-iced, live-stock to be watered and fed, and explosives and inflammables to be carefully set apart. This service is to be performed, in some instances, with 33 trains arriv- ing in 24 hours and a maximum of seven in an hour, averaging 50 cars to a train.i As these trains are broken up, their component elements are to be distributed and rearranged in other trains for prompt departure. Sys- tematic organization and intelligent supervision are necessary for this multifarious service to be efiiciently rendered. Each and every car is to be identified by placards and checked on the way-bills, from which the tonnage is to be abstracted and totalized as the trains are made over and placed on the departure-tracks. There is perhaps no more important factor in the eflicient conduct of freight-train service than the proper make-up of trains. As a general rule, trains should be made up for points as far distant as practicable. Better dispatch will be given, even if cars be held up for 48 hours to make up a solid train, as they will thereby be kept out of intermediate yards, and extra handling will be prevented. The extent to which this general rule can Ije observed depends, however, upon prevailing conditions as to the specific character and total volume of the traffic, and also upon the physical features of the line and the relative importance of the business handled at the intermediate transfer points.^ Yard-service in connection with making up passenger-trains is a simpler matter, as the kind and number of the cars and their arrangement in a train are more uniformly dependent upon the character of the traffic for which the train is intended. The locomotive is to be suited to the train and to its scheduled speed ; and not the train to the locomotive. Still, in large union-stations, the frequency of arrivals and departures, the neces- sity for handling cars between the station-tracks and the wash-tracks without interruption to the regular train-movements, and the incidental maneuvering of the train-equipment, call for careful attention to the requirements of the time-table and for skillful fulfillment of the demand for prompt and efficient service.' Each train must be made up to suit the service for which it is intended ; whether for suburban, excursion, local or through traffic ; and as to mail, express, baggage, combination or smoking cars, or as to day-coaches or sleepers, or the special equipment for limited trains. Where racial dis- 1 See Chapter V, Part II, p. 280. 2 See "Freight Terminals," p. 103. ' See Chapter V, Part II, p.. 281 ; and also "Passenger Terminals," p. 241. TRANSPORTATION 379 tinctions are observed, additional coaches must be provided. The number of coaches in a train must be varied to meet changing requirements. The wooden day-coach may seat 60 persons, the all-steel coaches seat 70, and a parlor car, 30. A Pullman sleeper with twelve sections can accommodate 50 persons, though usually not more than half of that number occupy the car. A train of eight cars in through-service will ordinarily consist of a mail, an express, a baggage and a smoking car, with three day-coaches seating 240 persons and a sleeping-car accommodating 50 at most ; or a total of 290 passengers. A limited train of nine cars would be made up of a baggage-car, a buffet-car^ an observation-car, a dining-car and five sleepers, of which four would be compartment-cars, with total accommoda- tion for perhaps 200 passengers; though ordinarily not more than half- occupied. On lines with terminals in large cities, the efficient conduct of the subur- ban service is of great importance. For persons with homes in the suburbs and business in the city, or "commuters," the arrangement of the time- table as to hours of arrival and departure, as to frequency of trains and as to the time consumed in the journey, profoundly affects their daily lives, and also the general character and the welfare of suburban communities. Commuters may be divided into three classes : those whose daily work takes them early into the city and. who return late, those whose occupa- tions admit of later hours in the morning and earlier in the afternoon ; and the shoppers, whose frequent visits to the city conform to their domestic routine. By far the largest number of commuters constitutes the first class, who can spare no more time than from half to three-quarters of an hour for the journey each way; the second class can afford to take a full hour, and the third class even more. Zones of Suburban Traffic An examination of the local time-tables with heavy suburban traffic shows that the train-service may be separated in zones to suit these re- spective requirements. In the inner zone of 10 to 15 miles radius, the service is performed in 35 to 45 minutes by 20 to 30 trains daily each way, with many stops, and closely grouped between 6.00 and 8.00 a.m. and 5.00 to 7.00 P.M. In the middle zone of 15 to 25 miles radius, the service is performed by 10 to 15 trains within an hour, with as many stops but fewer of them within the inner zone. The service iirthe outer zone, 25 to 30 miles distant from the city, is performed in from an hour to an hour and a half, by trains in the later forenoon and earlier afternoon.'^ A few night-trains serve the three zones for the accommodation of late pleasure-seekers and the night-toilers. Where the trains are frequent and the trains are numerous, the stops are made at alternate stations. With electric traction, the number of stops is increased within the same ' See "Passenger Terminals," p. 324. 380 EFFICIENT RAILWAY OPERATION period of time, but its substitution for steam has not apparently had the effect of lengthening the radii and increasing the area of the several zones. The situation in the community-zones around London does not differ materially from that here described.^ Switching Service Switching service is of three kinds : 1. Industrial and commercial switching, which is largely performed, either directly or indirectly, by the enterprises interested in the production or sale of the commodities handled ; 2. Terminal service at union-stations, which is performed by the station-employees ; 3. The distribution and making up of trains at the stations and in the classification-yards of the railroad companies by tl^eir own yard-crews, and which neither enters into the ton-mile computations nor into the freight- rates. The magnitude of the switching-service of this latter character, as an essential element of transportation-efficiency, is indicated by the propor- tion of switching-mileage to freight-train mileage as follows : Freight Train and Switching Mileage, 1914 MILLIONS OF MILES Mileage Eastern District sotjtheen District Western District United States Freight trains .... Switching Proportion 257 179 70% 117 49 ■42% 217 101 47% 591 329 56% Suburban Service. London, October, 1907 Zone Charing Cross, Miles Area, Square Miles Passengers Number Per Square Mile Workman's Tickets 1 2 3 4 5 6 7 8 4-6 6-S 8-10 10-12 12-15 15-20 20-25 25-30 63 88 113 138 244 560 707 864 2,232,201 3,406,588 2,432,996 843,780 238,252 331,213 151,025 107,614 35,584 38,729 21,512 6,104 977 592 213 124 389,229 1,054,461 512,858 122,456 24,290 24,168 673 819 Total 4 to 30 2,777 9,743,669 3,509 2,128,954 "Passenger Terminals," pp. 274r-275. TRANSPORTATION 381 Equivalent Switching Mileage Estimated at Six Miles an Hour NUMBER OF LOCOMOTIVES Freight Switching . . Proportion ... 17,053 4,926 29% 5,923 1,422 28% 14,429 3,533 24% 37,405 9,881 27% MILEAGE PEE LOCOMOTIVE Freight Switching .... 15,095 36,302 19,703 34,432 15,023 28,640 15,795 33,318 This statement is restricted to roads jn Class I ; that is, to those with annual operating-revenues in excess of $1,000,000; and is abstracted from a computation of transportation statistics for 1914 in Appendix VII, Table XVIII. In that year, the proportion of switching-mileage to freight- train mileage was 58 per cent, for the United States and 70 per cent, for the Eastern District. The proportion of switching-locomotives to freight- locomotives for the whole country was 27 per cent, and the average mileage per switcher was double that of the freight-locomotive. This is not actual performance in switching, but equivalent mileage, estimated at six miles an hour while so employed.^ On this basis, and estimating the freight-loco- motive mileage at 13 miles an hour between terminals, the freight-loco- motives averaged 1215 hours per annum or 3.3 hours per day, while the switchers averaged 5533 hours per annum, or 15 hours per day. This statement emphasizes the comparative magnitude of the switching-service and the attention which should be given to it as an element of transporta- tion-efficiency. Switching-service is not only hindered by inadequate or badly designed trackage, but also by the lack of organization which results in crowding the freight-yards with cars that should either be on the road or in the shop. Systematic supervision is necessary to prevent such a condition from prevailing where hundreds, or even thousands, of cars are daily pass- ing through a yard. For this purpose, it is essential that the yardmaster should know the condition of the yard. By means of a card-index, it is practicable for him to have in his office, at all times, complete information graphically displayed as to the number of cars passing through the yard, grouped as to hours of arrival and their destination, separated as to classes of freight and character of equipment. Where this plan has been intro- duced, there has been a reduction of 50 to 75 per cent, in the average time per car in the movement through the yard.^ ' The actual distance covered in switching averages 2.31 miles for each hour's work in freight-service, and 2.78 miles in passenger-service. American Railway Association Proceedings. October, 1916, p. 125. 2 See Proceedings of the Association of Transportation and Car Accounting Ofacers. No. 10, December, 1908, p. 1283, and No. 11, June, 1909, p. 1472. 382 EFFICIENT RAILWAY OPERATION Wrecking-train Organization At each division-terminal and important junction-point, there should be provided a wrecking-train, consisting of a derrick-car, a tool-car, several flat cars and a passenger-car. The derrick-car should have the means for slow propulsion ; the tool-car should be equipped with the appliances for replacing locomotives and cars on the track. An inventory should be kept with the car, which should be checked off after the car has been used, and missing or damaged articles replaced. One flat car should carry- blocking, guy-poles, and a few rails and joint-fastenings ; and another, a pair of standard freight-trucks. The passenger-car should have, at one end, a kitchen supplied with utensils and dry stores, to be occasionally renewed, and at the other, it should be fitted to provide first aid to the injured. . It should also be equipped with a full set of signals, fuses, tor- pedoes, torches, gasoUne lights and lamps ready for use, two field-telegraph instruments, climbing appliances, extra wire and a telephone-instrument ; and should be fitted with insulated connections. An inventory should also be kept with this car. The wrecking-crew should be drawn from the forces in the car-repair yard and should be composed of an experienced wreck-master, a competent crane-man and at least six of a repair-gang, together with a telegraph- operator and a line-man. The train should stand on a special track near the entrance to the yard in condition for immediate use. Each member of the crew, and the railroad surgeon, should report promptly at the train upon the sound of the prescribed signal. This is a matter to which more attention should be given than it generally receives. The maintenance of such a train, with an efficient crew, will greatly shorten the delay to the service from train-accidents, and be. the means of saving life. On lines exposed to obstruction by snowdrifts, a somewhat similar organization is required for winter use in connectiqp with a snowplow. Where snow drifts into deep cuts, the plow with a rotary cutting-face greatly facilitates the work. Empty-car Mileage Empty-car mileage is a matter to be considered' in a discussion of transportation-efficiency. The transportation-statistics for 1914, in Ap- pendix VII, Table XVIII, show that a freight-movement of 1,843,000,000 tons was accompHshed with 284,924,000,000 ton-miles and 13,507,000,000 car-miles, which is an average of about 21 tons per car. But this loaded movement was accompanied by 6,426,000,000 empty-car miles ; that is, one empty car was moved for every two loaded cars. In consequence, the total car-mileage represented an average loading of 14.3 tons per car. Empty mileage is unavoidable with a heavy mineral-traflSc in cars which are unavailable for other purposes, but much of it is occasioned by TRANSPORTATION 383 efforts to return cars directly to proprietary roads by routes on which there is no return lading for them. The paying freight must not only be burdened with this expense, but also with the cost of the train-movements in motive power and for train-crews. In seasons of surplusage, it would seem practicable to prevent unnecessary car and train mileage by some provision for storing idle cars at unloading points. The relative proportion of loaded and empty mileage in the Eastern and Southern traffic districts in 1914 was as 66 to 34, and in the Western District as 70 to 30. The variations below 60 per cent, and above 70 per cent, are given in Appendix VII, Table XIX. The lines with empty mileage . in excess are those engaged principally in mineral-traffic, whose open top equipment is necessarily returned empty, except in the case of those with terminals on the Great Lakes that handle coal and ore in opposite direc- tions. The lines with loaded mileage in excessive proportion are those whose principal traffic is in bulky commodities in lightly loaded cars. Train Dispatching When cars have been assembled in a train and placed on the departure- siding, the road-locomotive . attached, the brakes tested, the train-crew having reported for duty, and the way-bills having been furnished to the conductor, the responsibility of the yardmaster is at an end and that of the train-dispatcher begins. As train dispatching is now conducted, to the chief dispatcher is dele- gated much of the authority which was formerly exercised by the superin- tendent. He should be familiar with the controlling features of the fine on his division; with the local peculiarities of the grades, curves and sidings, with the rating of the locomotives and their current condition, and with the respective capabilities of the principal yardmen, operators and trainmen. When the weather forecast seems to warrant sucR a precaution, the tonnage-rating should be judiciously reduced in order to prevent a blockade out on the line. In addition to having at hand all this information, and the aid of devices for promptly registering train-move- ments, the chief dispatcher must, furthermore, have a natural aptitude for immediate decision in giving train-orders under the confusing conditions that frequently exist on a busy division. A dispatcher must provide for the occupation of the line by work-trains to 'the greatest possible extent, and yet be ready at any time to facilitate the passage of a high-speed train. Other emergencies arise, as when it becomes necessary to hasten assistance in case of a train-accident, or when from track-obstruction on a double-track line, the conflicting currents of traffic must be accommodated on a single track, or perhaps trains from a neighboring road must be suddenly detoured over his division because of similar obstacles. He must be careful in giving orders not to "bunch" trains at passing-points where th,e siding-accommodation is insufficient to 384 EFFICIENT RAILWAY OPERATION avoid obstruction upon the running-tracks. Precedence is to be given to trains requiring dispatch, as fast-freight or time-freight or perishables or live-stock. Such trains should be brought into terminals ahead of slow freight-trains, in order to insure promptness in getting them out of the yard again, even if it be found necessary to hold other trains out on the road. Yet, in so doing, care should be taken not to have the trains follow each other so closely into the yard that they can not be speedily cleared from the running-tracks. Success in these respects depends upon the dis- patcher's familiarity with the capacity of the terminals for the disposition . of arriving and departing trains, and necessitates intelligent team-work by the dispatcher and the yardmaster. For men who are fitted to cope suc- cessfully with such conditions, there is a broad field of usefulness in railway operation. Time-tables and Train Sheets The first railroads were, for the most part, so short that the same pas- senger-train could cover the line daily, going and returning. Its time of departure from each station was printed on one side, and a few simple rules on the other side, of a bit of card-board ; whence the term of "time-card" has been retained in the railroad vocabulary. On longer Unes, the pas- senger-service was performed by one train a day, each way, leaving each terminal at a convenient hour in the morning. If either train failed to make the schedule meeting-point on time, the other train waited for it from fifteen to twenty minutes and then proceeded cautiously, flagging the curves. So long as it kept within the prescribed time of its schedule, it retained the right of way ; but if it were unable to do so, then the oppos- ing train could proceed, also flagging curves, until the trains met. Be- twe^ the stations, there was a "half-way post," and the train that first passed that post thereby secured preference and backed the other train to a siding ; and "running for half-way" was of frequent occurrence.^ A daily freight train, leaving each terminal in the morning, ran leisurely over the road, avoiding the schedule passenger-trains and waiting for them indefinitely if they were delayed. This train delivered and loaded the small lots of freight at the way-stations and set off or took on the loaded cars. In the lumber-regions, the saw-mills were placed along the line wherever the situation was most convenient to the forest-timber, and the sawed lumber was loaded upon special trains while they occupied the main line. This was, in fact, the method of train-service on many of 1 Prior to 1845, on the Eastern Railroad of Massachusetts, in case a train was over one hour late in arriving at either Lynn or Salem, " The depot-master will im- mediately start on horseback to learn the cause of the delay." Trains began to be numbered on this road about 1848, each train retaining its number for the round, trip. In 1855, the outward trains had the low numbers and the inward trains the high numbers. In 1872, outward trains bore the odd and inward trains the even numbers. There were no Sunday trains until 1874, nor uniformed employees untill 1881. " The Eastern Railroad," F. B. C. Bradlee, Salem, 1917. TRANSPORTATION 385 the early railroads in sparsely settled regions, and until the mineral-traffic became of importance. In the more populous districts, and more especially in the New England states and along the North Atlantic seaboard, the train-service was more frequent. The station-sidings were not so far apart, and running for half- way was not customary. It was only on the longer lines with through- connections that there was a night passenger-service, usually with a single train from each terminal and, with heavier freight-traffic, there was a clearer track at night for the through-freight trains. With an increasing number of passenger-trains, it became necessary to keep the freight-trains also on schedule, and the time-card was superseded by the more extended time-table. The increased number of passing-points made the construction of the time-tables quite a serious matter and, when a change became necessary, the superintendent would retire into seclusion, with his clerk, until the schedules of the opposing trains had been so fitted into each other as to insure correctness with respect to the passing-points. On a road over a hundred miles in length, and with comparatively frequent train service, when a new time-table went into effect, all those concerned were somewhat anxious until all trains had been satisfactorily accounted for. On some roads the same schedule was kept in effect for years. In one instance of this kind, when it became necessary to change the time of a train a few minutes, this was accomplished by moving the time of all the trains around just that number of minutes. The time-table of a prosperous and progressive railroad grows with its growth, and with cumulative experience as to changes in social requirements and the character of its traffic. The time-table is also affected by improve- ments in alignment and equipment, and by the establishment of additional junction-points. This is especially true of single-track operation, as in- creasing traffic develops additional meeting-points to avoid congestion. The time-table usually includes a supplement containing information as to the location of sidings, water-stations, grade-crossings and drawbridges. A speed-schedule, to indicate the time elapsing for different distances at different rates of speed, is of use in running a delayed train to make a meet- ing-point. It is advisable to equip with speed-indicators the locomotives assigned to high-speed trains with long runs between stops. The time- table may also state the time usually allowed for stops.' ' Allowance fob Train Stops To change locomotives Train of 7 cars . . . Train of 8-9 cars . . Train of 10 cars . . . Train of 11-12 cars . . Regular Stations 5 minutes 2 minutes 3 minutes 4 minutes 5 minutes Flag Stations 1 minute 1| minutes 2 minutes 2i minutes 2c 386 EFFICIENT' RAILWAY OPERATION The construction of time-tables was greatly simplified by the use of the graphic train-sheet, which was introduced in the fifties by superintend- ents who had had the training of a civil engineer, using profile-paper for the purpose. This made the arrangement of passing-points a matter of certainty, in advance of the actual change of time. The number of passing- points increases rapidly with the addition of trains.^ The increasing complexity with heavy traffic on a single-track line is seen by a comparison of train-sheets taken from actual practice on the same division, eighteen years apart (Appendix VII, Plates I and II), show- ing that single-track operation had reached the limit at which it becomes necessary to obtain relief by the construction of sections of second-track. The daily collocation of the passenger-trains on a system covering thousands of miles of main lines and branches, in such form as to keep the chief transportation-official promptly advised of the manner in which the service is being performed, and of the causes and location of delays, is an important feature of transportation-efficiency. On the Southern Railway the information for this object covers a region es^tending from the Potomac to the Mississippi River, and from the Ohio River to the Gulf of Mexico, with continuous through-train service from Washington of 792 miles to Jacksonville, Fla., and of 926 miles to Memphis, Tenn. The causes of delay are classified under appropriate headings, such as waiting for foreign connections, delays caused by freight-trains, meeting trains, derailments, locomotive-failures, mechanical causes, slow-orders, blocks, mail transfers and miscellaneous. These daily statements are abstracted monthly, showing the percentage of trains late and average loss of time. The several divisions are classified on this basis, as the "Blue Ribbon" class, without delays; the "Red Ribbon" class, with delays of less than 15 per cent.; and the "Yellow Ribbon" class, with a greater percentage of delays.^ • See diagram in Appendix VII, Table XXV. Possible number of passing-points on single-track, in a distance of 120 miles, as calculated by P. M. La Bach, speed of passenger-trains, 30 miles an hour, and of freight-trains, 15 miles an hour, shown in the following table : Trains per Day Passenger Fbeioht PAseiNQ Points 1 1 6 2 2 20 3 3 36 4 4 71 '' Passenger Train Performance. Southern Railway. May, 1915. Trains run 13,803 Trains losing time 1,271 Total delays 929 hours, 9 minutes Average delay to late trains 43 minutes TRANSPORTATION 387 With the extension of our railway system and the accompanying ramifi- cations of passenger-train connections, the changes of time-tables to faciU- tate through-service began to be made at meetings of neighboring trans- portation-officials. It was then found convenient to concentrate the dates of such changes by the organization of the General Time Convention, covering the states north of the Potomac and Ohio rivers, and of the Southern Time Convention in the more southern states. At these con- ventions, the schedules for through passenger-train service were agreed upon in general outline. The time of arrival and departure at connecting ter- minals had to be carefully adjusted, by reason of the frequent differences in the local standards of time. This situation was intensified by the wide differences of time at the junctions or intersections of long lines of road, using the local standards of terminals hundreds of miles apart as to longi- tude. For instance, on one line extending for about four hundred miles in an east-and-west direction, there were ten standards of local time to be considered. In its published time-tables, at some junctions and terminals, its trains were timed to arrive after connecting trains had apparently left, although there was actually an interval of twenty minutes ; and it exer- cised trackage-rights over six miles of another line, where a difference ex- isted of several minutes. On one line between New York and Boston, there were three standards of time with a total difference of twelve minutes. Standard Time The discordance of railway standards of time occasioned so great in- convenience to the traveling public that it had attracted the attention of horologists. Greenwich Observatory time had been established as a standard for the British Islands on January 13, 1848, by Act of Parlia- ment. In 1869, Professor Charles F. Dowd of Saratoga, N. Y., proposed that railway-time in the United States should be standardized by plus or minus signs from meridians one hour apart, based originally upon the meridian of Washington, but afterward on that of Greenwich. Between 1876 and 1882, Sir Sanford Fleming, Chief Engineer of the Canadian Pacific Hallway, advocated a similar plan, complicated with a twenty-four hour system and the substitution of letters for the hours, instead of numbers. Suggestions of this character were also made by Professor Benjamin Pierce and by Dr. Thomas Hill, President of Harvard University. None of these proposals were approved by railroad-officials, as the establishment of stand- ards arbitrarily based upon meridians cut across so many lines as to render such a system impracticable. In 1881 Dr. F. A. P. Barnard, with Professor Cleveland Abbe and Professor Ormond Stone, presented the subject to the General Time Conven- tion, by which it was referred for a report to its Secretary, Mr. W. F. Allen. At that time, there were over fifty standards of time in use on our railway system. As manager of the Official Railway Guide, Mr. Allen was 388 EFFICIENT RAILWAY OPERATION familiar with the difficulties to be overcome in any solution of the problem that would be acceptable to railroad-officials. In April, 1883, Mr. Allen presented a report including certain propositions as follows : 1. The division of time over the earth's surface is to be based upon the meridian of Greenwich Observatory and upon each successive fifteenth meridian around the globe. The whole 360 degrees of its circumference would thereby be divided by meridians one hour apart ; and in sections 7J degrees on each side of these meridians, the time on that meridian would be the standard. 2. The boundaries between each section are to be so modified as to avoid cutting arbitrarily across the operation of east-and-west lines, thus meeting the most important practical objections to the establishment of the change of time by even hours at such boundaries. The General Time Convention gave its approval to this plan and au- thorized its Secretary to secure agreements for its adoption. At a joint meeting, in October, 1883, of the General Time Convention and the South- ern Time Convention, Mr. AUen reported that he had secured agreements to put the plan in use from the managements of 75,000 miles of railway and had been assured of the cooperation of the Naval Observatory at Washington, of the Cambridge Observatory and of several municipal gov- ernments. Both conventions then resolved to put Standard Time into effect on Sunday, November 18, 1883, when over fifty standards of time were merged into five on the railroads in the United States and Canada.^ The practical effect of the adoption of Standard Time has been to unify train-movements in this respect, in sections, as follows : 1. Intercolonial Time, based on the 50th meridian, applies only in New Brunswick and in Nova Scotia. 2. Eastern Time, based on the 75th meridian, is in use in virtually all of the Atlantic Coast states down into Georgia, and includes West Vir- ginia. It extends in Canada to the meridian of Fort William on Lake Superior. 3. Central Time, based on the 90th meridian, prevails in the states in the Mississippi Valley, along the Great Lakes and the Gulf of Mexico, in- cluding Florida and the greater part of Georgia. It extends westward into the Prairie states and, in Canada, to the eastern border of Manitoba. 4. Mountain Time, based on the 105th Meridian, extends thence westward to the states on the Pacific Coast. 5. Pacific Time, based on the 120th meridian, governs train movements in the Pacific Coast states and in British Columbia. Virtually, four standards of time cover the railway system of the United States. At 11.55 A.M., 75th meridian, business ceases on the telegraph-lines '"Short History of Standard Time and its Adoption in North America," published privately by W. F. Allen. November, 1903. TRANSPORTATION 389 throughout the country. The master-clock in the Naval Observatory at Washington then commences to beat seconds over the lines, skipping one second before each half-minute and five seconds before each minute, until 11.59 and fifty seconds. The circuit then remains open for ten seconds. When it closes again, it is exactly Noon, Eastern Time ; and all clocks and ■watches are regulated accordingly. At many of the principal stations, the clocks are synchronized electrically with the transmission of time from Washington. Every railroad company requires its train and station em- ployees to have their timepieces regulated periodically by its official watch- inspectors, within a variation of thirty seconds a week. Since the establishment of Standard Time in America, its use has been ex- tended all over Europe, except in Ireland and Portugal ; also in the greater part of South America, in Egypt and South Africa, in Australia, Japan, the Philippine Islands and in Porto Rico. The loss in time in a westerly direction and the gain in an easterly direction from Greenwich Observatory,, in circiminavigating the globe, are neutralized in the Pacific Ocean, in crossing the 180th meridian, by dropping a date one day when going east- ward, and by keeping the same date for two days when going westwards- Over the larger part of the earth's surface, clocks are now regulated by Standard Time. While the hours may differ, according to the control- ling meridian, the minute-Jiand points contemporaneously to the same mark on the dial, around the world. A traveler can cross the ocean from New York, and feel assured on arriving at Naples, for instance, of the interval of time available for taking the train for Rome, by looking at the minute hand on his watch. Though the hour-hand must be adjusted for the proper meridian, the minute-hand remains unaltered.. Yet the advantages of a uniform standard of time were so little appre- ciated by the general public in the United States that its adoption by the railroad companies met with strenuous opposition in communities whose local time differed materially from the railroad standard. Even to this day, the system has been legally established in but four states of the Union. The system of numbering the hours from one to twenty-four, which has long been in use in Italy, has not been generally sanctioned, although it was advocated in Canada by Sir Sanford Fleming. Uniform Train Rules Joint-action in the adoption of Standard Time, and the change on the southern roads from the five-foot gauge to the standard gauge, brought about a merger of the Southern Time Convention with the General Time Convention.! This step afforded the opportunity for harmonizing another 1 The General Time Convention was organized in April, 1875, with 36 dele- gates* representing 24 roads north of the Potomac and Ohio and east of the Mis- sissippi River. The Southern Time Convention was organized in October, 1877, covering the lines southward of the Potomac and Ohio rivers. The two conventions 390 EFFICIENT RAILWAY OPERATION feature of railway operation by the adoption of a code of uniform trains rules. Differences in the details of train-service were many, and often widely divergent, even upon neighboring roads. Rules were variously interpreted, though verbally alike ; nor were there any common forms of train-orders. A dispatcher might issue orders in language at times so in- definite or obscure as to lead to their misinterpretation. Hand, lamp and whistle signals not only differed, but even conflicted, on tracks and at stations used jointly by separate managements. The difficulty in recon- ciling local prejudices could only be appreciated by those who took part in the protracted deliberations by which uniformity was attained in the Code of Standard Train Rules, adoptee! in 1899. The Code, which had originally been prepared only for single-track operation, was supplemented by other rules for train-movements in double- track operation and on three or more tracks.' Signals of every character are illustrated and their specific indications clearly prescribed. The phraseology of rule and train orders is carefully expressed to meet each preconceived situation in train-movements, and the extent of the authority or responsibility of every official and employee concerned in them, from the train-dispatcher to the brakeman, is definitely determined. Since the adoption of the Standard Train Rules, they have had the test of nearly thirty years in practical use. During that period, the Standing Committee on Train Rules has been repeatedly called on to interpret their application under every condition that the ingenuity of dispatchers, trainmasters or trainmen could suggest. From time to time, the rules have been amended to conform to changes in operating-conditions, but al- ways by experienced transportation-officials, who have exercised due dis- cretion in modifying the action of their predecessors. The value of their work has been so generally recognized that a trainman from a New Eng- land road can go to the Pacific Coast with entire confidence in his acquaint- ance with the manner in which the service is conducted there. As the Standard Code is in force to-day, it is contained in 147 octavo pages, covering also the automatic block system, and including interlocking rules. Its use ha's benefited every train-employee in the country, and has been so greatly instrumental in increasing the safety of train-movements that, in the public interest, the Interstate Commerce Commission might well recommend legislation to make its adoption obligatory throughout the United States. were consolidated as The General Time Convention, in April, 1886, with a change of title in April, 1891, to The American Railway Association. — "Railway Opera- tion Associations." W. F. Allen. January 11, 1909. 1 In England, the left-hand track is the running track, following the custom there on public highways. In other countries, for the same reason, the right- hand track is used. It was stated in 1910 that, on the Lake Shore & Michigan Southern Railroad, the change had been made to the right-hand track at a cost of over a million dollars. TRANSPORTATION 391 Train Rights and Telegraphic Dispatching. Train Staff At first, the control of train-movements was directed to establishing a scheduled interval of time between them. But this interval was liable to derangement by unforeseen circumstances, and could only be restored by the action of the train-crews themselves. The next step taken in the re- sumption of external control was the display of signals at fixed points, beyond which a following train was not permitted to pass until a certain period of time had elapsed from the departure of the preceding train ; but this plan did not secure a preservation of the time-interval if the preceding train were delayed between stations ; nor did it provide for the meeting of opppsed trains, if either were delayed. In the latter case, the only provision for absolute safety lay in holding all trains moving in one direction until expected trains from the opposing direction had arrived. To prevent the consequent congestion of traffic, trains in one direction were given the right to proceed from a meeting- point, after an allowance of time for the possible difference of watches. This plan was, however, ineffective when a train with the right of way was itself unable to proceed. In this emergency, after a certain lapse of time, the precedence passed to. the opposing train, so long as it could maintain its schedule. When this could no longer be done, the control of train move- ments reverted to the respective train-crews, and anarchy prevailed until the time-table could be reestablished. The field for train-control was greatly enlarged with the introduction of the electric telegraph system which had been considerably extended in the United States before its value in railway operation was generally recog- nized. Although built along the right-of-way, wherever this was per- mitted, its offices were opened only in towns whose commercial relations afforded them profitable business, and they were usually remote from the railroad stations. The earliest use of the telegraph, as an adjunct to train-movements, is said to have been due to an emergency on the Erie Railroad in 1851, when Charles Minot, the superintendent, held a delayed train by telegraphic order to give the right of road to the train on which he was a passenger. He afterward framed rules for controlling train move- ments by telegraph.^ The establishment of telegraph-oflaces at railroad stations was subsequently made a condition for the use of the right-of-way for telegraph lines. Train-orders were at first only issued over the signature of the super- intendent, and the independent authority of the train-dispatcher over train-service was not customary until about 1870. This innovation was not generally favored by transportation-officials of the elder generation. In 1856, on account of a misunderstanding of orders, a freight-train on the Eastern Railroad of Massachusetts waited all night at Salem for an extra 1 "When Railroads Were New," p. 104. 392 EFFICIENT RAILWAY OPERATION passenger-train, which also passed the night at Ipswich, 11 miles away. Yet there were public telegraph offices in the railroad stations at Boston and Salem. In an investigation of the Revere disaster, which occurred on this road in 1871, in reply to a suggestion that the accident might have been prevented by the use of the telegraph, the Superintendent,, who had been in office since 1855, said that he " could not be responsible for the oper- ation of a road running the number of trains he had charge of in reliance on any such system." At that time, there were 66 trains daily with numer- ous extras, on 218 miles of line, with but 18 miles of second-track.' Train-dispatching, or the control of train-service by telegraphic orders, was developed in the United States in single-track operation upon the roads with terminals in Chicago, and was introduced into the East about 1872. Originally resorted to only in emergencies, it has now become the rule to move trains by telegraph. Fundamentally, train-dispatching is the tem- porary rearrangement of meeting and passing points by special orders, to faciUtate the movement of delayed trains. Its function in this respect, however, has been extended to the control of the entire train-service as regulated by the time-table, by the introduction of trains running under special orders or in sections of a scheduled train. On one important road, there are but two scheduled freight-trains daily out of its terminals, one in the morning and the other at night ; all other freight-trains are run as sections of these trains. Efficiency in train-dispatching requires the sole use of telegraph-wires, with day-and-night offices at all passing-sidings and junctions. In recent years, the telephone has been coming into use; first, as an adjunct, for transmitting telegraph-orders from telegraph-stations to outlying sidings and classification-yards ; and later, as a substitute for the telegraph. By repetition in return, equal accuracy is insured and with greater promptness; The initial cost is somewhat greater, but the cost of maintenance is about the same.2 At the beginning of 1912, the telephone was in use for this purpose on nearly 59,000 miles of line. In 1914, there were 26,241 tele- phones and 33,396 telegraph instruments in connection with the manually- controlled block system. On several roads, trains are provided with a portable apparatus which can be readily connected with the telephone- wire. It is also supplied to bridge and rail-laying gangs, to track-super- visors and work-trains, and to superintendents' cars. Wireless communi- cation with moving trains has been experimentally tested on the Delaware, Lackawanna & Western Railroad. In single-track operation on British railways, the precedence in train movements was secured between signal-stations by the use of the train 1 "The Eastern Railroad." 2 The telephone (invented in 18761 was first regularly used for train-dispatch- ing in 1882 by E. H. Whorf, superintendent of the Boston, Revere Beach & Lynn Railroad. His successor, C. A. Hammond (1883-1893), greatly improved the system, adding several features for securing accuracy and responsibility. TRANSPORTATION 393 staff in the hands of one train-crew. Until that emblem had been delivered to a station-master, no other train could enter the occupied section. Under this mode of operatipn, train-movements necessarily alternated in direc- tion. No train could move without the staff and, if there were any con- siderable delay in returning it by train, there was a consequent conges- tion of trains at the other end of the section, which could only be relieved by sending the staff back by a messenger on horseback. When this con- dition could be foreseen, it was provided for by running trains in a group, with a time-interval between them. Each successive train was furnished by the station-master with written authority to proceed until the last of tl^e group bearing the train-staff had delivered it at the end of the section, which was thereby freed for the passage of opposing trains. A train-staff system operated electrically has been devised by the General Electric Company, by which the staff -instruments at the ends, of a section are connected and synchronized, so that the withdrawal of a staff can only be effected by the joint-action of the two signalmen. The staff is a metal rod, six inches long. Half of its length serves as a handle ; the other end is turned down to rings of different diameters to actuate the locks of the staff-instruments. But one staff can be taken out at a time, though they can be inserted at either end without the cooperation of the other operator. Each instrument may be equipped with from twenty to thirty staffs in grooves from which the upper one must be removed first. A special form of staff-and-tablet apparatus enables a train carrying the staff to be placed clear of the line at an intermediate point, and to restore the instruments at the ends of the section to their nor- mal position. The staff is virtually a key to the block, as it unlocks the signals controlling the entrance to it, and also to any intermediate switch. It can be caught by hand on a train passing at a speed of fifteen to twenty miles an hour. Train-dispatching, in the American sense, has never found favor in British railway operation, nor on the Continent, . where British methods prevail. The Block System On account of the density of the traffic which had previously existed, many of the English lines were originally of double-track construction, and the short distances between the stations facihtated the control of trains by fixed signals. As early as 1851, electric caU-bells for this purpose had been placed at the stations. As the use of fixed signals became more general, the adoption of telegraphic communication was accompanied by the development of a different method of control by establishmg definite intervals of space between trains, instead of a time-intervak In 1854, a block system with visual signals was introduced upon the London & North Western Railway. Manually-controlled signals then came into general 394 EFFICIENT RAILWAY OPERATION use. In 1876, the Metropolitan Railway in London was operated under a block system with 3^-minute intervals. Under the block system, each section between .telegraph stations is termed a "block" and into this section no train is permitted to enter until, by communication between these stations, it is known to both that the block is unoccupied by another train. Under the train-dispatching sys- tem, train-orders are directed to the conductor of a train. Under the block system, the visual indications controlling the movement of an ap- proaching train are communicated to its engine-driver. At first, the entrance-signals merely indicated that there was or was not a train in the block. If the block was occupied, an approaching train could not pass the signal until it indicated that the block had been cleared. But, just as increasing traffic had become congested by a strict observance of the time-interval between stations, so did the same experience follow the institution of the space-interval, and the rule was therefore so modified as to permit a following train to enter a block with the knowledge that it was occupied. This is the broad distinction between the absolute and the per- missive block system. In the United States, the necessity for more direct and immediate con- trol of train-movements was provided for by supplementing telegraphic train-orders with fixed signals. Such a system was put in effect, in 1884, on the Pennsylvania Railroad between New York and Philadelphia. Man- ually-controlled signals were introduced on the New York Central & Hud- son River RaUroad in 1882 ; but the first important use of these signals in America in single-track operation was in 1884 on the Canadian Pacific Railway. The principle of the block system has been extended to a diversified control of train-movements by increasing the number of indications that give information to an approaching train. For this purpose, visual signals may vary in color, form or position. By common consent, red indicates danger and is a signal to stop, whether displayed from a station-master's office or from a signal-cabin, by a track-gaUg or by a casual wayfarer.* The signal may be a disk, or (usually) a semaphore-arm. Formerly, a ball or light at a masthead was sometimes used to give right-of-way at drawbridges or railroad crossings. As at first used, the "two-position semaphore," with its arm or "blade" extended to the right of the signal-post next to the running-track, showed two indications only. When horizontal, it indicated danger, "stop." When inclined downward it indicated safety, "proceed." Placed at the entrance of a block it gave the stop-indication if the block was occupied, or the proceed-indication if the block was clear. Since'the speed and momen- 1 The white light, as an iadication of safety, was superseded by a green light on the Great Northern Railway in England in 1876 ; and by a yellow or amber- colored light on the Pennsylvania Railroad, June 28, 1909. TRANSPORTATION 395 turn of trains have been increased beyond the ability to stop them within the distance at which, under all circumstances, the signal can be seen by an approaching train, it became necessary to place a supplementary signal nearer to the coming train, and at a sufficient distance in advance to permit of a safe stop before the signal at the block-entrance — called the "home" signal — should be reached. The home signal was distinguished by having a square-ended blade, while the advance, or "distant" signal had a notched end like a fish-tail.^ The distant signal simply repeated the position of the home signal, thus giving advance notice of that signal's indication. Afterward the two signals were placed on the same post at the entrance of each block, the square-ended arm placed above the notch-ended arm. Both arms hori- zontal, indicate that the block immediately ahead is occupied ; both arms inclined downward, that two blocks ahead are clear ; and the upper arm inclined and the lower horizontal, that only one block ahead is clear, and ■trains must proceed with caution expecting to find the second block occu- pied. Each arm is counterweighted, to bring it to the horizontal position, in case of a broken connection. A further improvement consisted in using the "upper quadrant" for the safety or "clear" indication; the upper inclined arm showing more distinctly, and avoiding the need of counterweighting and the liability of the arm being frozen in the clear or "proceed" position.^ The single-arm "three-position" semaphore thus combines the functions of the home and distant signal. Its position is horizontal when the first block ahead is occupied ; inclined 45° above horizontal when only the first block is clear ; and pointed vertically upward when both the first and second blocks ahead are clear.^ The general arrangement of a set of signals, as to color, form and posi- tion, is termed its "aspect," — the impression, as a whole, which it makes upon the vision of the engineer of an approaching train. The aspect may permit him to enter an occupied block, or may indicate that the next signal is at "stop," or that a train-order is to be given to him, or that his speed is to be restricted through an interlocked route, or that a drawbridge is open. Three or more signals may be located in the space of a few hundred feet ; all to be read while approaching them, perhaps at high speed. It is g, ques- tion whether this use of signal-apparatus for other purposes than simply to preserve a space-interval between trains is not being overdone, and whether, in the interest of safe operation, the aspect should not be con- fined to two meanings only — "stop," or "proceed." The length of a block section varies with the frequency of train-move- 1 The fish-tail notch was introduced on English railways in .1872, and was made obligatory in 1877. 2 This change was introduced into the United States in 1906, and is now in general use. ' For standard signal indications, see Appendix VII, Note VII. 396 EFFICIENT RAILWAY OPERATION ments. On the New York Central Railroad, there are 225 block towers between New York and Buffalo, a distance of 440 miles. On the Pennsyl- vania Railroad between New York and Philadelphia, the block sections are in no case more than 4000 feet in length, with an allowance of two minutes between trains. In the New York Subway, more than 2000 trains start from the terminals daily and, in the "rush hours," trains fol- low each other at intervals of one minute and forty-eight seconds. The block sections at the stations are shorter than the length of the platforms, which accommodate ten-car trains. Automatic Block System Under the manual system, the information that the block is or is not occupied is conveyed to the engineer of an approaching train through human agency, and the possibility of error is tripled by the intervention of two other persons. The engineer may either misinterpret or disregard the signal indications ; the receiving operator may misinterpret or disre- gard the information which is to control the display of signals ; the sending operator may either transmit that information incorrectly or not send it at all. In these respects, the manually controlled block system was orig- inally defective. The effort to eliminate human fallibility began with the control, at the entrance-signal, of the action of the operator at that point when it should be necessary to inform an approaching train that the block was or was not occupied. By placing the appliances for that purpose under the control of another operator at the outlet of the block by means of an elec- tric circuit, the receiving operator at the block entrance could not display a signal improperly, and there was one mind the less to make a mistake. At the outlet of the block, it was required to guard against the transmission of incorrect information and of failure to send information at all. The latter contingency is not so important where the position of the entrance signal can be changed only by the act of the sending operator, but pro- tection against incorrect transmission is more difficult to secure. The operator at the outlet determines the condition of the block, first from a notification that a train has entered, next by actual observation of its passage. It is not sufficient for him to know that a locomotive has passed out, but also that every car which was attached to it when it entered had passed out with it. If this operator knows that the block is clear, then the assurance should be returned to him that this information has been correctly transmitted and that the entrance-signal is in proper position. When rules and appliances have successfully brought the system to this stage, the next step is to eliminate the operator at the outlet of the block. This object has been attained under the automatic block system, which had its origin in the United States with the invention of Hall's automatic disk-signal. This signal was experimentally tested on the Eastern Railroad TRANSPORTATION 397 in Massachusetts, when train dispatching was introduced there in 1872. It was, however, only made apphcable as an accessory to the block system by its association with William Robinson's electric closed track-circuit in 1879, on the Fitchburg Railroad. Automatic apparatus was afterward applied to semaphore signals, but this date may be fixed as the era of double-track operation under the block system, as distinguished from signal-track opera- tion by telegraphic orders. Electro-pneumatic signals were introduced in 1885, but the first extensive automatic-signal block system was installed in 1891 on the Cincinnati, New Orleans & Texas Pacific Railway. The next installation was on the Chicago & Alton Railroad in 1879. In 1915, this system was in general use on 4466 miles of line on the Southern Pacific and Union Pacific railways. Notwithstanding its advantages, automatic signaling has not been generally adopted on steam-operated roads. The manually-controlled system is usually preferred in Europe. In the United States, in 1914, out of 86,738 miles of track under block signals, 52,032 miles were under some one of the non-automatic systems.' Under the automatic system, devices actuated by the train, as it passes the entrance and the outlet of the block, simultaneously operate a display of the signals in the positions required to block the interval of space into which the train is entering, and to clear that which it is leaving. This effect can also be extended to the entrance of the block next behind it, so that the engineer of a following train may thereby be informed in advance, • not only as to the condition of the block ahead of him, but also as to the block ahead of that. By the use of a closed track-circuit, the entrance or home signal can not give the indication that a block is clear, so long as a single pair of wheels is in the running-track within the block, or while any intervening switch is misplaced. The characteristic features of the automatic block system are contin- uous insulated track-circuit and the operation of the signals at the ends of the blocks by electric mechanism which holds them in the clear position. As soon as the front wheels of a train enter the block, the electric current ii short-circuited through the wheels and axles. The signal, being de- prived of electrical energy, falls by gravity into the horizontal position, indicating that the block is occupied, and so remains until the last pair of wheels has left the insulated rails. The electrical energy then restores the signal to the position indicating that the block is unoccupied. The opening of a switch or the breaking of a rail, by interrupting the conti- nuity of the current, likewise releases the signal from the clear position. Electric Signaling Electric traction requires the uninterrupted use of the rails in the run- ning-track ; therefore it becomes difficult to divide the track into insulated 1 The development of the several methods of operating fixed signals is described in Chapter V, Part II, pp. 257-260. 398 EFFICIENT RAILWAY OPERATION sections for block signaling. On electric lines operated by direct current, it was necessary to employ alternating current for the signal track-circuit. For this purpose, "inductive" or "impedance" bonds are substituted for the insulated joints on steam-operated roads. The alternating-current signal track-circuit was first developed satisfactorily in 1902. But for its use in connection with the inductive bond, it would not have been prac- ticable to operate either the Grand Central Station or the Pennsylvania Railroad Station in New York City. Upon the Chicago, Milwaukee & St. Paul Railway, the electrified sec- tion of 440 miles over the Continental Divide is operated under colored signals only, in place of the semaphore. There are eight signals between sidings that are at an average distance of about seven miles apart. The circuits are arranged for single-track operation under permissive three- light signals ; red, for stop ; green, cautionary ; white, proceed. A pilot- lamp gives the indications, if the main lamp should burn out. A small red marker-lamp below the number-plate, staggered as to the main lens, locates the signal at night if the indication should not be displayed. The range of the signal-lights in the daytime is 3000 feet under normal conditions, and 2000 feet under the most unfavorable conditions, with the sun shining directly on the lens. On curves, the lens is provided with deflecting prisms, which give the beam of light a wide spread. The indications can be seen during a driving snowstorm in the daytime much farther than it is possible to see a semaphore blade. ^ Automatic Stop A further step has been taken for the protection of trains against the neglect or misinterpretation of signals by enginemen, by the introduction of appliances connected with block signals which shall strike the cab-gong, or sound the whistle, or apply the brakes, or even close the throttle-valve of an approaching locomotive, before it passes within an occupied block. Such appliances are in use on several electric railways, where the trains are light and of the same character. But the application of automatic control to the general train-service of a steam-operated road involves many serious considerations, as will be seen by reference to the requisites for such an installation as prescribed by the American Railway Association.^ An automatic stop in connection with an electrically-operated block system has been in use for some years on 107 miles of double-track on the Chicago & Eastern Illinois Railroad, which acts upon the air-valve of the locomo- tive. It also closes the throttle-valve on locomotives attached to passenger- trains, but not on those upon freight-trains. The permissive block-system, which is of great value on lines with heavy traffic, is inadmissible in con- nection with an automatic control. The primary difficulty in the correlation of train-movements directly ' Railway Age Gazette, September 8, 1916. " See Chapter VI, p. 309. TRANSPORTATION 399 with the signal-indications of the block system is due to the operation of trains by motive power of an entirely different character from that by which the signals are actuated, and which is derived from a source borne inde- pendently by the train. This difficulty does not exist in providing for the automatic control of train-movements in electric operation, where both the train-movements and the signal-indications are actuated by the same motive power and from a common source. A megaphone-horn connected with the stop-signal and actuated by a treadle beside the running-rail would sound an alarum, when operated by the wheels of a passing train, that would not only arouse a drowsy engineman but would also give a wide- spread notice whenever the stop-signal had been disregarded. The pur- pose of the automatic stop is to provide against the inattention of the engineman to the signals. As this purpose should be attained more simply by close supervision of the manner in which the signal-indications are re- spected, the introduction of automatic train-control would seem to be un- necessary.i Interlocking Improvements Proceeding on a line parallel with the development of the block system, there has been an introduction of appliances for greater security against open drawbridges or misplaced switches and at railroad crossings, in con- nection with interlocking apparatus controlling several of such points from one station. This control has been made interdependent with the control of the block signals by marvelously ingenious mechanism to prevent a route from being altered, unless the change can be safely made. The lock- ing-bar, or "detector-bar," used to prevent a switch from being changed under a passing train, has been replaced in automatic signaling by specially designed relays operated by alternating current. "Approach-locking" is used on full-speed tracks by which a signal, having been accepted by a train, is locked until the train has passed. In Europe, the interlocking is still largely under manual control.^ The intervention of mechanical appliances in the movement and con- trol of signals, switches, drawbridges and crossings, and the interlocking of such appliances, have induced the substitution of other motive power for muscular energy and, at the present time, electricity, compressed 1 On the Pennsylvania Raiboad, in the first six months of 1915, 2,000,000 efficiency tests show^ed 99.9 per cent, of correct observance of rules by 135,458 employees. There were 23,390 tests in the use of signals with 99.4 per cent, of correctness. Out of 10,000 tests with signals set at "stop,'' in only 13 cases did the train pass the signal by as much as one foot. Railway and Locomotive Engineer- ing, September, 1915. See Appendix VI, Note II. 2 The periscope has been utilized on the Northwestern Elevated Railroad in Chicago, placed on, the top of a switch tower at a sharp curve where the view is obscured by buildings. The towermau is thus enabled to have a view of trains ap- proaching from either direction. 400 EFFICIENT RAILWAY OPERATION air and other gases are employed in this operation, either singly or in combination. 1 On the Pennsylvania Railroad, at Greensburg, Pa., an electro-pneu- matic interlocking plant controls from a single cabin, 155 switch signals and other functions with but 48 levers, in a territory 2450 yards in length. The east home signal is 3700 yards distant from the west home signal one, and the distant signals are over 5000 yards apart. In the Pennsyl- vania Railroad station in' New York City, the interlocking cabin at the west approach, spans five tracks and covers a machine which controls 124 signals, 30 double-slips and 47 switches. In one building at the Grand Central Station, the interlocking machines contain 400 levers on one level and 360 on the other, all actuated electrically ; and the movement of the trains is shown by electric indicators.^ In important passenger-stations, electric communication between the gateman and the conductor of a departing train actuates light-signals that inform the gateman when the train is ready to receive passengers, and advise the conductor that the gate to his train has been closed. By similar means, the station-master can send the number of the outgoing train's track to the interlocking cabin from which the locomotive-engineer receives a signal that his routeing is fixed. From that moment, the route can not be changed until it has been cleared by the train. At a converging junction where trains may approach at nearly the same moment, it becomes necessary to give one of them precedence. With heavy freight-trains or on a descending gradient, this order of preference once established should never be changed, and one of the trains should be held at the previous signal. With lighter trains on an ascending gradient, the service may sometimes be expedited by changing the precedence as the trains approach. Where the junction arrangements are automatically controlled by the track-circuit, this can not be done under ordinary con- ditions though it may be accomplished by the use of time-relays. In each converging track there is a length of insulated track-circuit in connection with a relay controlled from the signal-cabin. This relay is set so as to require for its operation more time than would be consumed by a train passing over the insulated track at a greater speed than 1^ miles an hour, which is equivalent to a train slowing down for a stop. At any higher speed, the locking is automatically released, and the clear signal may be changed from the other track. This arrangement is however inoperative on either track, if the signal on that track is already in the " stop " position. Immediately, as the train passes beyond the junction, the block automatically clears behind it, and ' The development of the different methods of operating interlocking apparatus, in connection -with the block system has been described in Chapter V, Part II, pp. 257-260. 2 B. B. Adams. London Times, June 28, 1912. TRANSPORTATION 401 the time-limit on each track becomes inoperative until the blocked section beyond the junction has been cleared. To insure that the time-relay is acting properly, a " proving-contact " is inserted in the track-circuit, which retains the. signal in the rear in the "stop" position. An illuminated diagram in the signal-cabin shows the condition of the tracks, and small red lights at the junction indicate whether an approaching train has come to a stop or is operating the time-relay.^ The degree of efficiency that has been attained in the control of train- movements by the combination of the block system with interlocking apparatus may be inferred from experience on the New York Central Railroad. In March, 1907, 31,440 freight-cars were handled in a single day between New York and Buffalo, in addition to about 1000 passenger- trains. Forty-nine thousand men were employed in this service, exclusive of the cleric^,l force and the men employed on construction-work. In the total distance of 440 miles, the operators in 225 towers controlled, on an average, 110 trains per mile of road.^ The sureness, or freedom from failure, in the operation of interlocking apparatus of this character, is shown in the following table : ' ' General Signal Co., Roeliester, N. Y. ' New York Times, March 24, 1907, p. 79. « ' Switch and Signal Failxtbes, Hudson and Manhattan Railroad year ending december 31, i9i5 Cattse Delays MiNTJTEB Average Operation PER Failube Signals Automatic stops . . . Switches 31 7 15 89 13 41 2,011,804 4,567,113 289,983 Total 53 143 • 6,868,900 Railway Age Gazette, April 28, 1916. 2d CHAPTER VIII WAR-TIME Military Transportation by Land in Ancient Times and Previous TO THE Railway Era Five years ago, there would not have been available material sufficient for a chapter on railroad efficiency in war-time. But to-day, railway efficiency for warlike purposes has undergone a wonderful development, and in no other country on a grander scale or under more favorable auspices than in the United States. To appreciate these facts in all their aspects, reference may be made to the earUest association of transportation with military operations. From pre-historic sources, we obtain the impression that an invading army was not supplied to any great extent from the rear, but that it "hved on the country," ravaging it as the army marched and leaving a wake of desolation' as it advanced. Its numbers were, therefore, limited by the resources of the country which it traversed, and its strength was otherwise restricted by the magnitude of the droves of asses or camels that formed its supply-columns. For the roads were but trails, impassable for any wheeled vehicles other than the light Egyptian chariots, and streams that could not be forded could only be crossed by ferries, for lack of bridges. The Pharaohs were able to conduct victorious campaigns because. they had the use of water-transport, either upon the Nile, or on the Red Sea, or along the Syrian coast. The first invasion of Greece by the Persians was by sea, with a fleet of 600 triremes, or galleys with three banks of oars, and 400 horse-trans- ports, which landed 110,000 men for the Battle of Marathon. The second invasion was organized on a far more extensive scale. The resources of the vast empire of Xerxes were drawn upon from the banks of the Indus west- ward to the jEgean Sea and the waters of the Danube, from the Black Sea to the Red Sea and the Indian Ocean, and, in Africa, from the Egyptian outposts on the borders of Ethiopia. For a whole year, this host was assembling at Sardis, on the coast of Asia Minor. For two years, the women of Thrace were grinding corn for bread, and the flour was stored at fortresses on the line of march, with other supphes forwarded from Egypt and Syria by transports. A pontoon-bridge, 4500 feet in length, with two roadways, was built across the Dardanelles at Abydos, and the peninsula of Mount Athos was 402 WAR-TIME - 403 separated from the Thracian coast by a canal of 4000 feet, to permit the passage of the fleet collected from Egypt, Phoenicia, Cyprus and the coasts of Asia Minor. For seven days and nights, the bridge at Abydos was thronged with soldiers on one roadway, and the baggage-trains on the other. After the crossing into Europe, the army numbered 1,700,000 footmen, 80,000 cavalry, 20,000 desert Arabs on camels and a force of chariots. There were 1500 triremes, 3000 fifty-oared vessels and unnumbered supply- ships and horse-transports, which kept along the coasts to supply the marching troops. The host of Xerxes was probably the most numerous body of men that was ever organized in a single army. It took four months to march from* the Dardanelles to Athens, a distance of about 480 miles, — an average of four miles a day. After the disastrous naval engagement in the Bay of Salamis, Xerxes covered the same ground in forty days in his homeward flight, which was an average of twelve miles a day. These two figures represent the maximum rapidity of movement of large armies and of in- dividual travelers over the thoroughfares as they existed up to the time of the Roman Republic. After an assumed existence of two hundred and forty-four years, the dominion of Rome, at the time of the expulsion of the Tarquins and the establishment of the Republic, had only extended over an area of about sixteen miles square, bounded by the sea, the Tiber and the mountains. For two hundred years thereafter, the Romans were almost constantly engaged in warfare with their neighbors. They might drive these from the plains but, in their mountain fastnesses, they awaited a favorable opportunity to resume hostilities. The Romans could gain no dominion over them nor even a permanent peace, only a truce for a term of years. In 309 B.C., a Roman consul constructed a carriage-road out of the city across the adjacent marshes to the highlands. This was the beginning of the famed Appian Way, which was gradually extended southward into the enemy's country. The Romans had grasped the idea of securing their hold upon their conquests by establishing strategic lines of communication, which eventually radiated from the gates of the Eternal City to the farther- most bounds of their empire. Over these substantial highways, which bridged the smaller streams, armies moved with certainty and celerity, accompanied by siege-trains and by supply-trains of heavy wheeled vehicles, instead of pack-animals. Long after the Empire had been overrun by hordes of barbarians, the Roman roads facilitated communication until they were neglected beyond repair, though to this day portions of them are still passable. The lack of roads in the Middle Ages prevented the movement of large armies with their supplies. Medieval warfare was carried on principally by small bodies of armored, knights, who preyed upon the countries which they invaded. More important campaigns were based upon communica- 404 EFFICIENT RAILWAY OPERATION tion by sea, as were the Crusades upon the Mediterranean, while the long- continued warfare of England against France was sustained by transports from the Channel ports to Bordeaux. It was not until the middle of the Sixteenth Century that carriages began to supersede saddle-horses and horse-Htters, and wagons to supplant pack-animals. As late as 1776, it required six weeks for a wagon with eight horses to make the round trip of about 420 miles between London and Edinburgh at an average rate of ten miles a day. At the time of the French Revolution, the maximxim rapidity of move- ment of large armies did not much exceed that of the army of Xerxes. •The first carriage-road across the Alps, between Switzerland and Italy, was the military road over the Simplon Pass, constructed by Napoleon's orders in 1800-1807. At that time, fourteen days were allowed for a regiment to march from London to Liverpool, a distance of about 200 miles. About that period, a new era was inaugurated in road-making by Telford and Macadam, which was followed by extensive improvement in the means of communication in France. Rapidity of movement was, however, still limited by the muscular power and vital energy of man and beast, until the advent of transportation by rail and steam. Early European Use of the Railway for War Purposes Railroad construction in Europe was developed by the commercial activity that ensued upon the cessation of warfare after the Battle bf Waterloo. As early as 1833, Harkort, a Westphalian, pubUshed a scheme for the construction of a railway along the right bank of the Rhine between Mainz and Wesel, which contained the following statement : "Any cross- ing of the Rhine by the French would then scarcely be possible, sin"ce we should be able to bring a strong defensive force on the spot before the attempt could be developed. These things may appear very strange to-day ; yet in the womb of the future, there slumbers the seed of great develop- ments in railways, the results of which it is, as yet, quite beyond our powers to foresee." ' In 1833, in the French Chamber of Deputies, General Lamarque declared that the strategical use of railways would lead to "a revolution in military science as great as that which had been brought about by the use of gun-powder" ; and M. de B6rigny asserted that, "An army, with all its material, could, in a few days, be transported from the north to the south, from the east to the west, of France." In 1842, M. Marschall, in advocating the construction of a line from Paris to Stras- burg, said, "the German Confederation is converging a formidable system of railways from Cologne, Mayence and Mannheim. Twenty-four hours will suffice to concentrate on the Rhine the forces of Prussia, Austria and the Confederation, and on the morrow an army of 400,000 men could 'See "The Rise of Rail Power in War and Conauest, 1833-1914." E. A. Pratt. London, 1916. WAR-TIME 405 invade our territory by that breach of forty leagues between Thionville and Lauterburg, which are the outposts of Strasburg and Metz; Three months later, the reserve system organized in Prussia and in some other of the German States would allow of a second army being sent of equal force." In the same year, Ponitz published a work in Saxony on the use of railways in war, in which, in a time of profound peace, he proposed the construction of a system of strategical lines for the protection of the Ger- man frontiers against France and Russia. In this work, he said, "We have to look to these two fronts; and, if we want to avoid the risk of heavy losses at the outset, we needs must also at the outset be prepared to meet the enemy there with an overwhelming force." ' Notwithstanding these repeated warnings, the French government at- tached so little importance to the strategical value of railways for national defense that, in 1844, von Moltke wrote to his brother that whilst Germany was building railways, the French Chamber was only discussing them ; and when Germany had 3300 miles of railway in operation, France had only about 1000 miles. Nor were the miUtary authorities in Germany in agree- ment as to the usefulness of railways in actual warfare. In 1847, a leading German military writer declared that the best organized railway could not carry 10,000 infantry for 60 English miles in 24 hours. "As for the con- veyance of cavalry and artillery by train, that would be an impossibility." At the opening of the Liverpool & Manchester Railway, in 1830, a regiment was conveyed 34 miles in two hours, which would have required a journey of two days on foot. In 1846, a Prussian army-corps of 12,000 men, with horses, guns, road-vehicles and anomunition was moved to Cracow on two lines of railway. In 1849, a Russian corps of 30,000 men and its equipment was taken by rail from Poland to Moravia, to a junction with an Austrian army and, in the winter of 1850, an Austrian army of 75,000 men, 8000 horses and 1000 vehicles was transported from Hungary and Vienna to the Silesian frontier. In the following year, a division of 14,500 men, 2000 horses, 48 guns and 464 vehicles was taken by rail from Cracow for a distance of 187 miles in two days. Allowing a march of twelve miles a day and one day's rest in seven, this movement would have occupied fifteen days. In the ItaUan campaign of 1859, railways were conspicuous in actual warfare, both strategically and tactically. "Thousands of men were carried daily through France to Toulon, Marseilles or the foot of Mont Cenis; injured men were brought swiftly back to the hospitals. The railway cuttings, embankments and bridges presented features of impor- tance equal or superior to the ordinary accidents of the ground, the posses- sion of which was hotly contested."^ In 86 days, the French railways ' Pratt, p. 5. ' Major MiDar, R. A. V. C. Journal Royal United Service Institution, Vol. V, pp. 269-308. London, 1861. 406 EFFICIENT RAILWAY OPERATION transported 604,000 men and 129,000 horses with a total of 2636 trains, including 253 military trains. In ten days, the Paris-Lyons Railway moved a daily average of 8421 men and 512 horses, without interruption to the ordinary traffic and with a maximum of 12,138 men and 655 horses. 75,966 men and 4469 horses were transported by rail in ten days from Paris to the Mediterranean, or to the frontiers of the Kingdom of Sardinia ; it would have taken sixty days for them to perform the same journey by highways. This rate of transit was about twice as fast as the best achievement recorded up to that time on the German railways. The men were described as leaving the station at Turin with none of the fatigue or reduction in numbers which would have occurred in marching. On the Austrian side, the First Army Corps of 40,000 men and 10,000 horses occu- pied fourteen days on the journey by rail from Vienna to Verona, having to march 83 miles from Innsbruck to Botzen, including the Brenner Pass over the Alps. The same journey would have taken 64 days, marching all the way. Railway efficiency in these European wars was unfavorably affected by the lack of sufficient second track, station-accommodation and equip- ment, and by the want of cooperation of the military staff with the railway officials ; but far more by a failure to work out a detailed scheme of or- ganization, in advance of its being put in effect. From a combination of several of these causes, trains were blocked or delayed, stations were con- gested with supplies that should have gone forward, and needed reinforce- ments were kept waiting because of the delay in the return of empty equip- ment. It frequently happened that the transportation of troops occupied more time than if they had marched, and that they reached the front unaccompanied by necessary supplies. As a consequence, important military movements were retarded or frustrated. Development of Railroad Strategy during the American Civil War The use of railroads in warfare, on an extended scale, really began with the opening of the Civil War in the United States, in 1861. The northern frontier of the field of action was formed by the Chesapeake Bay, the Potomac and the Ohio rivers. This frontier line of over 1200 miles was paralleled for the whole distance on its northern side by railroads and, with the exception of about 350 miles west of Washington, it was also covered by navigable waters which were dominated by the federal forces. The Mississippi River separated the eastern and the western portions of the seceding states. At a short distance above Cairo, on the Ohio River, the Cumberland River gave access to Nashville, and the Tennessee River to the northern border line of the states of Mississippi, Alabama and Georgia, although these navigable waters into the heart of the Confederacy were defended by fortifications. WAR-TIME 407 The continuous line of defense for the South was defined by the railway lines through Virginia and along the southern border of Tennessee to Memphis for nearly 900 miles; and from Memphis to New Orleans for 400 miles. Toward the seacoast, sealed by blockades, the railroad line extended from Fredericksburg through Richmond, Petersburg, Wilmington, Charleston and Savannah for 650 miles ; and thence through Southern Georgia and Middle Florida for 280 miles to St. Marks, on the Gulf of Mexico. In a general way, this region was thus surrounded on three sides by a continuous line of railroads for over 2000 miles, a situation which had never before existed in warfare.^ Toward the south, peninsular Florida extends for over 400 miles, its eastern border then accessible from the interior only by light-draft vessels for 125 miles up the St. John's River. This peninsula was covered from Fernandina, on the Atlantic coast, to Cedar Key on the Gulf of Mexico, by a railroad line of 150 miles, which was crossed by another railroad of 81 miles from Jacksonville to Live Oak, where it connected with the Une from Savannah. Otherwise, this vast territory was for the most part an unexplored region of lakes and impenetrable swamps, surrounded by lagoons and sand-bars. Virtually, the same conditions prevailed along the Gulf coast to the vicinity of Mobile, whose fortifications protected the interior navigation in the State of Alabama, as those at New Orleans did the ascent of the Mississippi River ; and there was an interior navigable route between those ports. The only points at which the Southern railway system approached that in the Northern States were at Columbus, Ky., on the Mississippi River, at Louisville, on the Ohio River and by the Orange & Alexandria Railroad at Alexandria, eight miles below Washington ; but at each place on the five-foot gauge.^ The Shenandoah Valley Railroad lay west of the Blue Ridge for 51 miles from Harper's Ferry, on the Baltimore & Ohio Railroad, to Strasburg. Trains also ran into Strasburg from the Orange