REESE LIBRARY UNIVERSITY OF CALIFORNIA. Deceived Accession No. / 5~U jy, / . Cla&s M STREET-RAILWAY ROADBED. BY MASON D. PRATT, Assoc. M. Am. Soc. C. E., AND C. A. ALDEN, Assoc. M. Am. Soc. C. E. NEW YORK: JOHN WILEY & SONS. LONDON : CHAPMAN & HALL, LIMITED. 1898. Copyrighted, 1898, BY THE STREET RAILWAY PUBLISHING COMPANY, NEW YORK. ROBERT DRUMMOND, PRINTER, NEW YORK. PREFACE. THE subject-matter of this little book is mostly made up from contributions of the authors to the Street Railway Jour- nal, Engineering News, and Am. Soc. C. E. Proceedings during the past two years. Some matter has been added and the whole brought up to date. It is not supposed to cover the entire field of street-railway track construction, but to present in compact form the main point of the best practice of to-day. A few hints are given which we hope will be found useful to engineers, managers, and trackmen. M. D. P. AND C. A. A. STEELTON, PA. , August 1898. iii CONTENTS. CHAPTER I. EARLY TYPES OF RATLS 1-12 CHAPTER II. THE DEVELOPMENT OF THE GIRDER RAIL 13-18 CHAPTER III. WHAT GOVERNS THE SHAPE OF RAILS 19-33 CHAPTER IV. THE T RAIL ADAPTED TO STREET RAILWAYS 34-42 CHAPTER V. TRACK FASTENING AND JOINTS 43-59 CHAPTER VI. SPECIAL WORK 60-71 CHAPTER VII. GUARD-RAILS 72-83 CHAPTER VIII. ADVANTAGES OF SPIRAL CURVES AND TABLES FOR SAME. . . . 84-110 CHAPTER IX. DESIGN OF SPECIAL WORK 111-121 CHAPTER X. SURVEYS AND LAYING OUT WORK 122-125 CHAPTER XI. SPECIFICATIONS 126-131 INDEX 133 v UNIVERSITY STREET-RAILWAY ROADBED. CHAPTER I. EARLY TYPES OF GIRDER RAILS. AN esseutial feature of a well-equipped street railway is a good track. This is a fact that has been brought home with force to most managers of electric street railways particularly. And while there have been long strides in the direction of better construction since the advent of rapid transit and heavy cars, we can hardly say that perfection has been attained in this, any more than in other phases of our mundane existence. In the following pages it is the intention to review the ex- perience of the past fifteen years the era of most rapid devel- opmentand to bring together and illustrate the various types of track material and construction, indicating their good and bad features. The most important part of the track is the rail, and we will therefore first follow its development. To America the United States belongs the honor of in- troducing the street railway, or " tramway," as it was first called. The section of rail adopted on the first line laid, that in Fourth Avenue, New York City, was of the flat type, it being nothing more than a simple bar of iron, with a groove formed in its upper surface to receive the flange of the wheel. From that time to the beginning of the present era, a period of nearly fifty years, this type of rail, though modified in every conceivable way, was adhered to. The weight ranged from thirty pounds to eighty pounds or more per yard. 2 STREET-RAILWAY ROADBED. In nearly every modification the rail was dependent on some other continuous and longitudinal support for vertical stiffness, in which respect it differed materially from the modern mil. In America a small lip or flange was added to the under side to keep the rail from slipping off the stringer. In England a second flange was added, and the two increased in depth, thus adding materially to the vertical stiffness of the rail. This feature probably reached its greatest development in the sec- tion used by James Livesey in Buenos Ayres (Figs. 1, 2). His rail had a total depth of two and three-eighths inches, and he FIG. 1. THE LIVESEY RAIL. did away with the longitudinal stringer, supporting the rail on cast-iron chairs placed at three feet centers. There were FIG. 2. SECTION OF THE LIVESEY RAIL. other systems where the two side flanges were replaced by a single flange under the center of the section. As nearly all sections were used with wooden stringers, the fastenings consisted mainly of spikes, staples, or lag-screws passing through the rail. The joint was nothing more than a EARLY TYPES OF GIRDER RAILS. plain flat bar of iron, three or four inches wide and eight to ten inches long, let into the stringer, and gave but a feeble support to the loose rail ends. The flat, or tram, rail was lacking in vertical strength even for the comparatively light traffic of those early days. Engineers realized this, and tried to find a remedy in the T or " Vignole" rail. The difficulty of paving to it and of maintaining the pavement proved too great an obstacle to its general use, except in suburban lines. FIG 3. FIG. 4. : y * . . 2" ; ' 1 T j|M FIG. 5. FIG. 6. STRINGER RAIL SECTIONS. Even a modification of it, in which the center of the base was placed to one side of the center line of the web, thus allowing the paving-stones to rest against both head and bottom flange, though tried, does not seem to have been an entire success. Since street-railway tracks were laid along the lines of other vehicular traffic, it is but natural that this traffic should seek to follow the path of least resistance the rails. But the con- sequent wear and tear on the adjacent pavement was consider- able, and the effect on the track from this street traffic was probably even' more injurious than that from the legitimate wear of the cars. It may have been with the idea of self-protection, or it may have been under pressure from the city government, that the street railways of Philadelphia adopted a flat tram and a wide gage for the accommodation of vehicles. But however it may 4 STREET-RAILWAY ROADBED. have been, the step was in the wrong direction. It was an in- vitation to greater concentration of traffic along the street railway, the difficulty of turning out from the track almost compelling vehicles to remain, and to set the pace of any car that might be following a matter of very serious moment where rapid transit is concerned. No one city has been brought to a greater realization of this fact, probably, than the one that introduced it. The acceptance of this condition has been, strange to say, almost universal in this country, and by far the greater number of rail sections are found to have the side flange or tram for the exclusive use of vehicular traf- fic. In this, as in many other matters, our practice has become directly the reverse of that in European countries, where the use of a grooved rail is universal. Such a rail gives an un- broken surface to the pavement, thus insuring a greater free- dom of movement to the general traffic, with less obstruction to the cars. It cannot be denied, however, that there is some slight ad- vantage to the street railway in a flanged rail over the grooved, as often made. It is more free from an accumulation of dirt in summer and ice in winter, which in the grooved rail ob- structs the free passage of the wheel flange to such an extent as to increase, in some cases, the force required to move the car as much as fifty per cent. In cities where the streets are paved with Belgian blocks, brick, or asphalt, and are kept rea- sonably clean, there can be little objection to the grooved rail. The greater freedom of movement for the cars, due to a less obstructed track, and the longer life of the pavement, with fewer repairs, consequent on the distribution of the street traffic over a larger area, more than compensate for a possible increase in motive power. Before proceeding further it might be well to give the nomenclature of the various parts of a rail. There is no little confusion in these terms as commonly used, and we have given those most approved by general usage. The list of names is given on page 5. The ordinary sections of T rail are not well adapted for use on paved streets, but considering their greater stiffness over EARLY TYPES OF GIRDER RAILS. \ FIG. 7. FIG. 8. FIG. 9. B FIG. 11. DIAGRAMS SHOWING NAMES OF PARTS. # Head 6? Groove. T Tram, or tread. F Flange, any projection from body of rail. B Base, or lower flange W Web. E Fillet, or rounding of any corner. L Gauge-line. N Neck. C Lip or flange. A Flange-angle. A' Guard-angle. C' Joint-plate, splice bar, channel-plate, angle-plate, or fisli-plate. This latter term properly belongs to the joint-plate used wilh flat rails. K Track-bolt, splice-bar bolt. Bearing. The surface of contact between splice-bar and rail. 8 Shoulder. Throat, applied to guard or full-grooved rails. Side-Bearing and center-bearing. Terms applied to rails to indicate position of head, with reference to center line of rail. Fig. 8 is a centre- bearing, all others are side-bearing. 6 STREET-RAILWAY ROADBED. the old flat rails and the superior advantages their shape offers for making joints and fastening to the tie, it is not strange that the early efforts made for a better rail for street railways were in the direction of a modified form of this rail. The first of these special sections actually rolled was section No. 72 of the Cambria Iron Company, of Johnstown, Pa. (Fig. 12), and was made in 1877 for the Clay Street Hill line FIG. 12. FIRST GIRDER RAIL ACTUALLY ROLLED. in San Francisco. It may also be a matter of interest to know that it was made of steel steel rails being at that time by no means common. The design has some remarkably good features notably, that the combined width of the head and tram is the same as that of the base, and the sides are in the same vertical lines, thus affording a good rest for paving- stones. The flange-angles, except that under the head, were small (7 degs.), less than the common practice of to-day. The bearing-surface for fish-plates is ample. The head is broad, and the point of contact between wheel and rail is brought nearly over the center of the web. The upper flange is evidently intended only to act in connection with the pavement to form a groove for the passage of the wheel flange, no attempt being made to provide a track for street vehicles a most commendable feature. It is a notable fact that there are but few great achieve- ments of science or invention brought to public notice that have not been discovered or invented before, and the fact comes to light only when the later and more energetic inventor makes them a success. Hence the old saying, "There is no new thing under the sun." The successful inventor is none the less worthy of his reward. The idea of the " girder rail/' EARLY TYPES OF GIRDER RAILS. 7 so-called, was not new in 1877, when the first rail was rolled, for we find on the Patent Office records a patent granted to Jfailroadffail . Patented My io, 7yitjis& / s'st/*' J/jt OS^t^TH. t^/C FIG. 13. BEERS' EARLY RAIL PATENT. Sidney A. Beers, in 1859, on "An improvement in railroads for streets," which shows the girder rail almost exactly as we know it to-day. That any such rail was ever made or used at that time does not appear, but the inventor certainly antici- pated the idea of our girder rails. It is to a later and very energetic inventor A. J. Mox- 8 STREET-RAILWAY ROADBED. ham, of the Johnson Company that we owe, in a large mea- sure, the successful development of the modern girder rail. His first efforts in rolling such a rail were made in 1881, at UNITED STATES PATENT OFFICE. SIDNEY A. BEERS, OF BROOKLYN, NEW YORK. IMPROVEMENT IN RAILROADS FOR STREETS. Specification forming part of Letters Patent No. 33,891, dated May 10, 1859. able form which ina be intended for or ap- plied to the purpose of a track or train for the accommodation of ordinary vehicles. Letter c is the body of the rail. "Letter d is a bracket planted upon the side of the rail at intervals and extending from the base to the tram to give additional sup- port to the latter, as wellas increased strength to the rail as a whole. Letter e is a base of any convenient width to strengthen the rail and increase the bear- ing. What I claim as my invention, and desire to secure by Letters Patent, is The construction of uprightself-sustaining rails of .cast or other iron, with car and ear- riage track combined, as set forth in the ac- companying specification and drawings to be laid in public streets and highways and for no other purpose. SIDNEY A. BEERS. JOHN C. SMITH, JOSEPH P. To all whom it may concern: Be it known that I, SIDNEY A. BEERS, of the city of Brooklyn, in the county of Kings and. State of New York, have invented a new and useful Improvement in the Construction of Railroads; and I do hereby declare that the following is a full and exact description thereof, reference being had to the accompa- nying drawing, making part of this specifi- tion, and to the letters of reference marked thereon. The nature of my invention consists in the construction of upright self-sustaining rails of cast or other iron, with the head or track expanded in width o as to form a car and carriage track in combination of such width and form as may be desirable to accommo- date buoh purpose when laid in public streets or highways. The figure is a transverse view or section of the rail of sufficient depth and strength to support the travel without the aid of a wooden string-piece. Letter o is the crown or car-track. Letter b is the carriage-track of any desir- FIG. 13a. BEERS' EARLY RAIL PATENT. Birmingham, Ala., afterward at Louisville, Ky., and later at Johnstown, Pa., where in 1883 rails of this type were first rolled to any great extent. The principal early sections of the Johnson Company are shown in Figs. 14 to 26. They re- ceived the nickname of "Jaybird" rails. It was thought, and with good reason, that a very great advance had been made in producing a rail which could be jointed by means of splice-bars, and'which, being in the form of a beam or girder, would have sufficient vertical strength in itself. There was a demand, of course, for a grooved rail, and we see it supplied in Figs. 22 and 25. These have, in place of the broad base, a "bulb," providing only scanty purchase for the fish-plates. These were called " bulb sections," and were doubtless the result of efforts to decrease the difficulties en- countered in rolling the flanged sections. They were exceed- EAKLY TYPES OF GIRDER RAILS. 9 ingly unmechanical in design, though not so much so as the "Wharton" or " Butterfly" rail (Figs. 27 and 28), which was devoid of either flanges or bulb. FIG. 16. FIG. 18. FIG. 19. EARLY JOHNSON RAILB. To get the strongest section there should be an approximate equality in the amount of metal in head and base, which feature very few of these early sections possessed. 10 STREET-RAILWAY ROADBED. Although the necessity for stringers was thus done away with, there was a deficiency in height, and recourse was had FIG. 20. FIG 21. k FIG. 22. FIG. 23. FIG. 24. FIG. 25. EARLY JOHNSON RAILS. FIG. 26. to " chairs," which were made either by forging from flat steel bars or of cast iron, to make the construction suitable for paving. EARLY TYPES OF GIRDER RAILS. 11 A noticeable feature of these early Johnson sections is the shoulder under the head on the side-bearing rails. Its use enabled both splice-bars to be alike, thus effecting a slight economy in manufacture. It increased the thickness of the neck, and apparently added not a little to the strength of the FIG. 27. FIG. 28. WHARTON "BUTTERFLY" RAILS. section and some additional wear. But possibly the main reason was the introduction of a distinctive feature which would be of value in developing patents. So energetic was this company in obtaining patents and aggressive in maintain- ing them, that it practically had a monopoly of the girder-rail business for several years. The attractive profits, however, were too great, and other manufacturers soon entered the field, the principal one being William Wharton, Jr., & Company of Philadelphia, whom we find offering a series of rails, without bottom flanges or base, and later the Lewis & Fowler Company of Brooklyn, N. Y., with its "box-girder" rail (Figs. 29, 30, 31). This latter was the old double-flanged rail of Livesey, brought to a greater refinement of design. Then came in Gibbon with his "duplex" rail (Figs. 32, 33). The head and tram of this rail are in separate parts, and each is provided with a vertical flange or web, which was, like the Wharton rail, devoid of lower flanges. All of these sections,, while proving good substitutes for the old tram-rails when used on horse-car lines, had to succumb to the onslaught of the electric motor. Their weak points were lack of vertical strength and poor joints the fishing space being so narrow that the two joint-plates together were far from having the- 12 STREET-RAILWAY ROADBED. same strength as the rails, and the meager bearing allowed them soon to wear loose. The multiplicity of parts also at 4__Ji FIG. 29. n frJL^ FIG. 30. FIG. 81. LEWIS & FOWLER "Box GIRDER" RAIL. 4k' ^\ 4 ... fi" i ^\ f FIG. 32. FIG. 33. GIBBON DUPLEX RAIL. involved in the use of chairs was a bad feature. The tendency was constantly toward deeper, stiffer, and heavier sections. CHAPTEE II. THE DEVELOPMENT OF THE GIRDER-RAIL. AT the beginning of the present decade the six and seven inch sections shown in Figs. 35 to 48 were the most approved rails in use, and indeed the seven-inch sections con- tinue to be largely used on electric roads laid in asphalt or Street Ry.Jourual FIG. 34. THE CRIMMINS RAIL. brick pavements, or even in shallow Belgian-block pavements where the ties are embedded in concrete. The latter con- struction, while not common in this country, is coming more into vogue. These seven-inch sections are also the ones most used on cable and conduit electric roads, where the rail is supported on cast-iron yokes and the pavement rests on a con- crete base. Fig. 34 shows a seven-inch rail adopted in 1895 by the Metropolitan Street Eailway Company of New York 13 14 STREET-RAILWAY ROADBED. 6* c FIG. 35. FIG. 3(5. FIG. 37. FIG. 38. FIG. 39. 5'- FIG. 40. FIG. 41. LATER GIRDER-RAIL SECTIONS. THE DEVELOPMEBT OF THE GIRDER-RAIL. 15 FIG. 46. FIG. 47. LATER GIRDER-RAIL SECTIONS. 16 STREET-RAILWAY ROADBED. City for use on its lines, most of which will eventually be cable or conduit electric. It is peculiar in having an extended lip attached to the guard, the idea of which is that it will carry the street traffic which tracks along the rails, and thus decrease the wear to some extent on the adjacent pavement. It was designed by Mr. John D. Crimmins, the builder of the Broadway Cable and other lines of the Metropolitan Street Railway Company. Fig. 49 shows the rail used on the new conduit electric roads FIG. 49. WASHINGTON RAIL. in Washington, where the streets are paved entirely with as- phalt. There is a very serious objection to these seven-inch rails on roads laid in granite-block pavement on an ordinary sand base in that the ties, having little or no sand over them, form a solid bed for the pavement, while that portion between the ties, having a more yielding foundation, sinks, and the track soon presents the appearance of a "corduroy" road. To overcome this defect and to meet the conditions where even heavier pavement is laid, still deeper rails were required. Solid rails nine and ten inches high were suggested and called for, but it was not until about six years ago that they were pro- duced. Their manufacture presented many difficulties, and the rail-makers met with many failures in attempting to roll THE DEVELOPMENT OF THE GIRDER-RAIL. 17 them. Much time, thought, and money have been expended in experiments, with the result that to-day these rails are placed on the market at a price but slightly in advance of or- dinary T rails. So great seemed to be the difficulties in the way at first that many devices were brought forth to accomplish the purpose without making a solid rail. The most ingenious of these was the so-called " electric rail/' which consisted of an ordinary " bulb "section and a J_-shaped base rolled separately, FIG. 50. GROOVED RAIL WITH ELECTRICALLY WELDED FEET. the latter being cut into short sections of ^from four to eight or nine inches and electrically welded to the head portion at intervals suited to the tie spacing. By thus rolling the rail in two separate parts a very broad base could be produced, and a large economy effected in the omission of the entire lower half of the rail between the ties. This rail was fully developed and a quantity of it laid, but the inconveniences of handling and laying it proved to be many and great. During the two years following its introduction a reduction in the price of steel rails of nearly fifty per cent took place, and rapid strides were made in the art of rolling solid, deep sections, which, to- 18 STREET-RAILWAY ROADBED. gether with the difficulties above mentioned, rendered it a commercial as well as a practical failure. Many other schemes for the production of a deep construc- tion without resorting to the solid rail have been devised. The idea of a combination rail the head portion to be re- newablehas been worked out (on paper) in many different ways. But no such schemes which have been offered may be considered practical in the light of experience. In the first place they are objectionable on account of multiplicity of FIG. 51. SIDE-BEARING RAIL WITH ELECTRICALLY WELDED FEET. parts a condition which should be avoided, in track work particularly. In the second place the renewable feature their fundamental idea is valueless, from the fact that the perma- nent parts become so much worn that it is impossible to secure a good fit on renewing the wearing portion. Then, again, the cost of this renewal would probably amount to as much as the laying of an entirely new track. For these reasons no other rails than the solid, deep sections are seriously considered to- day. CHAPTER III. WHAT GOVERNS THE SHAPE OF BAILS? THER"E is a wide variation in ideas as to the proper form these deep rails should have, as a glance at the illustrations will show. Of course, as explained before, local conditions govern to a considerable extent. Some city governments specify a full, narrow groove; others a broad, flat tram; while FIG. 52. FIG. 53. 10 AND 10i INCH RAILS. those who leave the matter to the railroad companies find rails laid in their streets having all variations between the two. The question as to what is the proper form for the exposed 19 20 STREET-RAILWAY ROADBED. upper surface of the rail is one to be carefully considered in all its relations to motive power, density and character of street traffic, pavement, etc. When these circumstances are consid- ered, no one can say that any one section is the proper one for all cities, or even for all the lines in any one city. We will first consider the various influences which should determine the form of that portion of the rail exposed to wear, as follows : I. Motive Power. This may be divided into two classes : that which is applied to the axles, as with the trolley electric FIG. 54. CENTER BEARING RAIL. FIG. 55. NEW ORLEANS RAIL. system, together with gas, compressed air, or other similar motors ; and that which is applied externally, as in cable or horse traction. With the former it is far more important to have a rail which shall be free from dirt. Particularly is this so with electric roads using the rails for a current conductor, and some form of half -grooved or tram rail should be selected. A center-bearing rail is by far the most desirable, if only the interests of the railroad are considered ; but there has been a strong and growing dislike to it on the part of the public, owing to the two grooves formed in the pavement along WHAT GOVERNS THE SHAPE OF RAILS ? 21 each rail, and which render it doubly annoying to carriage traffic. There are few cities to-day that permit its use. A rail which approaches the center- bearing rail in freedom from dirt is shown in Fig. 55, the standard section adopted by the New Orleans Traction Company. With externally applied power it is not so important, though still desirable, to have the rail as free from dirt as in the former case, for, under like conditions, the resistances are not so great, and a full-grooved rail may be used. II. Street Traffic. This can have but little influence in towns where it is light and of a miscellaneous character, but in cities where it is more dense and heavier it must be con- sidered. The ideal condition obtains, as before stated, where FIG. 56. FIG. 57. SECTIONS SHOWING WEAR OF RAILS. FIG. 58. the surface of the street presents an unbroken face. This is to be had only with a full-grooved rail. With any other sec- tion there is a guiding shoulder for wagon-wheels, compelling them to follow the track. The next best section is of the half-grooved type, in which this shoulder is a minimum, and offers less obstruction to vehicles turning out. The guard or lip should be made substantial to resist bending as well as wear. The full flat tram is very objectionable from the street-traffic point of view, for while it offers a smooth, easy track to travel on, it is most severe on a vehicle when turn- ing off. 22 STREET-RAILWAY ROADBED. The width of tires on light carriages is about one and one- quarter inches, and on deli very- wagons about one and one- half inches ; trucks and heavy wagons have much wider tires. The usual thickness of flanges on street-car wheels is one inch, and as it is customary to place the wheels on the axles to gage about one-quarter inch less than the track, it will be seen that the least width the full groove in a rail may have is one and one-eighth inches. This is less than light carriage tires, and is therefore safe for all vehicles. Indeed, when it is con- sidered that it is only the heavy wagons and trucks that fol- low the street-car tracks to any great extent, it would be perfectly safe to make the groove one and one-quarter inches wide. This gives the car-wheel flanges greater freedom of movement and offers less resistance. Now, if a rail is used having a groove not wider than one and one-quarter inches and the guard brought up level with the head, and the pavement laid on a solid foundation with its top surface level with the head of the rail, it will be exceedingly difficult for the wagon traffic to keep to the car tracks, and the wear will be distrib- uted over the whole surface of the street, thus making the life of the pavement so much greater. While it may be seen from the foregoing that the ideal shape for the top surface of the rail is the full groove (Figs. 35 and 49), no city which does not specify a first-class pavement and does not keep its streets clean should compel its street railways to lay such a rail, for dirt soon fills the groove, render- ing the operation of cars very much more difficult, if not dangerous. III. Pavement. With asphalt there seems to be but one section the full-grooved. Any other which attracts street vehicles is detrimental, in that this pavement is more yielding to a concentrated traffic than any, and is the most costly to repair. Broad tram rails have been laid with this pavement, and they are as much out of place as a square peg in a round hole. They practically defeat the purpose for which the fine pavement was laid, namely, a smooth street. The question as to whether a grooved rail should be laid in granite pavements is one to be settled by the probability of the streets WHAT GOVERNS THE SHAPE OF RAILS ? 23 being kept clean, as well as by the other considerations mentioned above. The depth of the rail is also governed to some extent by the character of the pavement. There is ample stiffness and strength in a seven-inch rail for all ordinary traffic, and there FIG. 59. FIG. 60. SECTIONS SHOWING WEAR OF RAILS. is no occasion for using any deeper section in macadam or brick pavement. Nor does asphalt require a deeper section, except in the cities north of Mason and Dixon's Line. It has been found that in these Northern cities with their cold win- ters the jar of traffic on the rails breaks up the asphalt when laid directly against the rail. As a precaution against this action it is customary to lay a line of " toothing " along each side of each rail, the toothing consisting of an alternate header and stretcher of granite blocks. This, as shown pre- viously, requires rails of about nine inches depth. In macadam pavements T rails are without doubt the most satisfactory sections, for the reason that the street traffic on macadamized streets is light, and usually follows its own tracks at the sides of the street. A grooved rail would be out of the question on account of the dirt, and a tram girder rail would only attract, the street traffic, and ruts would rapidly form along the rails. IV. Wear. Now that we have in the deep sections, rails that will make when properly laid, a most substantial track, the question of wear becomes of even more importance than 24 STREET-RAILWAY ROADBED. formerly. Indeed so substantial may the track now be made in other respects that its life is determined almost entirely by the amount of abrasion the rails will withstand, and with this the form of the rail head has something to do, as well as the material of which it is made. In this connection are shown five sections of worn rails, Figs. 56 to 60. They were taken FIG. 62. FIG. 63. FIG. 64. MODERN RAIL SECTIONS. twelve or fourteen inches from the ends. In the last four are shown, in the lighter lines, the original sections as rolled. WHAT GOVERNS THE SHAPE OF RAILS? 25 The actual amount of metal lost is but a small percentage of the whole section, but there is in each case, and in one case particularly, a noticeable amount of distortion of the section which must have come from a very heavy vehicular traffic. These rails were removed because of their inability to stand up under this, as well as the car traffic, and the wear shown must not be taken as indicative of the life of heavier and FIG. 65. FIG. 66. MODERN RAIL SECTIONS. deeper sections now being laid exception possibly being made to the section, Fig. 60, which is given here to show to what extent rails are sometimes allowed to wear. This rail had about reached the end of its usefulness when removed. An- other noticeable feature in four out of five of these worn sections, and one which will be seen in practically all worn street-railway rails, is the decided inclination of the top sur- face of the head from the gage line up. This, no doubt, comes from the coning of the wheels. The majority of rails hereto- fore made have either been rolled flat on the head or with an inclination in the opposite direction, which has been given them to facilitate rolling. That it is not impossible to make 26 STREET-RAILWAY ROADBED. rails having an inward slope to the head will be seen by a glance at some of the sections shown in connection with this and the previous chapter. Aside from all questions of better electrical contact, better traction, etc., to be obtained from a full bearing for the wheel tread, the rail head is bound to assume this shape early in its career, and if made so in the first place it is manifest that the life of the rail is increased thereb}''. That this is coming to be considered the proper form, the increasing number of sections to which this feature is being applied would seem to indicate. In this connection the following remarks made by the writer before the Am. Soc. of Civil Engineers * are quite pertinent : " In studying the subject of the best form of rail and track construction for city streets where traffic is heavy, some aid may be had by looking carefully into the existing conditions, and seeing what effect the several sections of rail have had on the pavements as laid. This may lead to modifications of the sections which will tend to prevent any injurious effects. " Fig. 67 shows the rail of the old horse-car tracks in New York City, which have by far the greatest mileage to-day. FIG. 67. OLD HORSE-CAR RAIL ON STRINGER. The time for their reconstruction is rapidly approaching, as New York is far behind all the other large cities in the matter of improved traction. The drawing very clearly shows the objectionable features when the pavement is laid to line with the top surface of the rail. In many cases the pavements have gone down, and the entire head of the rail projects above the general street surface. * Vol. xxxvii., Transactions Am. Soc. C. E. WHAT GOVERNS THE SHAPE OF BAILS? 27 " Fig. 68 is a partial section of the Broadway cable road. It is not peculiar to any one portion of the road, but fairly rep- resents the conditions for the major part of its length. The paving-blocks rest on a concrete bed, with but an inch of sand between, giving a practically unyielding bed. The illustration clearly shows the adhesion to an old practice in paving, where FIG. 68. HALF SECTION OF THE BROADWAY CABLE ROAD. the foundation was not so well made and was of a more yielding nature ; reference is made to the height of the paving-blocks above the head of the rail. In pavements as ordinarily laid it is expected that they will sink more or less under the action of traffic and weather, and it was customary to set the stones -J in. or more above the rail. The wisdom of so doing was realized when the blocks did sink, often not stopping at the rail level. This will be found to be the case with many of the old tracks, like Fig. 67, which shows the pavement in its best me FIG. 69. HALF SECTION OF THE THIRD AVENUE CABLE ROAD. condition. In the case of the Broadway road, and all others where there is a concrete base, the pavement has remained just where it was put, with the result that there is a deep rut along each line of rails, the bottom of which is on a level with the rail head or tram. The same is true of the track of the Third Avenue cable road, Fig. 69, and will be so of the tracks 28 STREET-RAILWAY ROADBED. now being laid in First Avenue in asphalt with granite tooth- ing block, Fig. 70. FIGS. 70, 71. RAILS LAID IN ASPHALT IN NEW YORK. " Fig. 71 shows part of a track laid with the improved rail mentioned by Mr. North, in asphalt pavement in One Hun- dred and Sixth Street. This is a very broad street ; no cars have ever run on these tracks, and the street traffic is very light, so there is no guide as to what the results would have been under the conditions existing in most of the streets in the city. " Fig. 72 shows a part of a cable track in asphalt pavement in One Hundred and Sixteenth Street. The hollow on the FIG. 72. SECTION OF CABLE TRACK IN ASPHALT PAVEMENT. inside of each track rail shows where the wear, due to vehicu- lar traffic, has been concentrated. " It will be admitted, of course, that the most desirable con- dition for the surface of a street is one uninterrupted by ruts, grooves, or other irregularities. It will also probably be ad- mitted that the greatest life will be had from a pavement where the traffic is distributed over its entire surface, and that the pavement will suffer most when the traffic is concentrated along certain portions. The greatest damage is done to pave- ments by trucks and other heavy vehicles, and it is they that are constantly seeking the path of least resistance. This is, WHAT GOVERNS THE SHAPE OF RAILS ? 29 of course, along the rails; and where the rail is of such form that the tram or guard portion is lower than the head, no matter how little, there is a guiding shoulder which keeps the wheels in the tracks without any effort on the part of the driver. Comparatively few of the heavy trucks and drays have the same gage of wheels as the railway tracks ; but this in no way prevents their following the tracks, for one wheel will ride on the shelf of the rail, while the other follows a par- allel path just outside the other rail. As there are many trucks of each class, it will be found that this concentration of wear has its effect in an appreciable trough along the outer rail and about 6 to 10 inches away. " The head of a perfect rail should be level, with a groove or flangeway only sufficiently large to pass the wheel flange freely, and of such shape that dirt, ice, and similar obstruc- tions will be easily worked out by the flanges themselves. The groove in the rails (Figs. 34 and 75) does this admirably. Fig. 49 is the section of rail adopted in Washington, D. 0., where all the streets are paved with asphalt and kept clean. The wheel flanges are necessarily smaller than those in gen- eral use elsewhere." Before leaving the subject of proper form for the upper sur- face of the rail, attention should be called to its importance in relation to the form and size of the wheel flange and tread. The variety in design of wheels is as great as in rail sections themselves a fact which is of great annoyance to the manu- facturer of special work, though not of great importance in connection with straight track, except so far as the depth of the flange is concerned. The difference between the depth of flange and the height of rail head above the tram or bottom of groove represents, theoretically, the amount of wear possi- ble before the rail must come up. Flanges are made from five- eighths of an inch to one inch deep generally about three- quarters of an inch; and since this vertical flange space in the rails varies from one inch to one and one-quarter inches, the amount of wear available is from nothing to five-eighths of an inch, depending on which combination of wheel flange and rail section come together. 30 STREET-RAILWAY ROADBED. The question as to just what the life of a rail is when this matter of wear is considered is a most interesting one. Man- ifestly it must be gaged first by the number of cars which pass over it, not by years. The other influences are grades, curvature, motive power, character of streets clean or dirty, and the amount of street traffic. The rail shown in Fig. 60 had carried about one million cars; it was on a heavy up grade and in a busy street. The rail on the parallel track carrying the downhill traffic, which was equal to the uphill, showed but half the wear, and was in fairly good condition when the worn-out section was removed. There are points in Boston where the rail heads are being worn down at the rate of about one quarter of an inch per year. On the Broadway Cable road in New York the total wear since the rails were laid (1892) is not much over one quarter of an inch. The amount of traffic in these two cases is not materially different, but the cause for the great difference in wear must be found in the motive power the Boston cars are self-propelled, the cable cars are externally propelled. The difference in weight between the cable and electric cars no doubt has some influence also. If electric cars were placed on the Broadway road to-day it would be a safe prediction that new rails would have to be laid within a year. And now what is the proper shape for the lower portion of our rail that portion which is not exposed, but which has to withstand all the shocks and strains produced by the traffic over and upon its head ? It mast have sufficient area of head to allow for wear and be of a form to remain rigid and unyielding under the strains of traffic ; it must have a broad base to provide ample bearing on the tie. All of these con- ditions are fully met in the deep sections shown. The web and bottom flanges are as thin as it is practical to roll them, and yet are of ample strength. The surfaces against which the flanges of the channel plates find a bearing have a uniform inclination a necessary condition in manufacture and one equally necessary in securing a proper fit for the joint-plates. There is little room for variation in the form of this portion of the rail. v ,The distinctive feature of the early Johnson WHAT GOVERNS THE SHAPE OF RAILS? 31 rails the shoulder under the head has disappeared, it being an insurmountable obstacle in the way of securing a good joint-plate fit and a broad bearing at that point. The 10-in. and 10^-in. rails shown (Figs. 53 and 54) were among the earliest designs of deep girder rails, but were ies 9" FIG. 73. FIG. 74. MODERN RAIL SECTIONS. never rolled. The 7-in., 8J-in. and 9-in. rails are all stand- ards. Fig. 74 is the standard of the West End Street Rail- way Company, of Boston the form of head being designed by the former Commissioner of Streets, Mr. Carter. The joint shown on this section is peculiar in having a rib ex- tending along the center of the joint-plate, in order to pro- vide a bearing and to prevent the plate from being drawn in against the rail, whereby the fit of the joint is destroyed. This form of joint was designed by the writer over six years ago, when deep rails were introduced, and it is now coming into general use. Since the above was written there have been added several new sections to our collection, three of which are shown here because they possess some interesting improvements over STREET-RAILWAY ROADBED. FIG. 75. STANDARD BAIL OP METROPOLITAN STREET RAILWAY, NEW YORK. l ' r 2*----. " FIG. 76. FIG. 77. STANDARD RAILS OF BROOKLYN HEIGHTS RAILROAD. WHAT GOVERNS THE SHAPE OF KAILS? 33 former practice. Section Fig. 75 is the present standard of the Metropolitan Street Eailway Company, of New York, and is used on all of their new conduit electric lines. It retains the lines of the head and tram of their seven-inch section shown on page 13, except that the tram is not quite so wide and is dropped one-eighth of an inch below the head. The head is also made thicker. The depth is increased to nine inches to give greater stiffness, found necessary under the tremendous traffic on their lines. It is the design of Chief Engineer F. S. Pearson. Sections Fig. 76 and Fig. 77 are the standards of the Brooklyn Heights Railroad, Brooklyn, N. Y., and are the design of Chief Engineer J. C. Breckenridge. They are pecul- iar in having the web brought more central under the head a desirable feature where the car traffic is so dense as to almost exclude the wagon traffic from the tracks. The load is then transmitted more directly to the base and without much tendency to cant the rail out. It also gives greater thickness of neck and consequently much more wear when used on curves. Section Fig. 77 follows somewhat the Metropolitan idea in the tram. It is contended that trams of this type hold the asphalt in place. CHAPTER IV. THE T KAIL AS ADAPTED TO STREET RAILWAYS. THE rail sections considered in the two previous chapters are called " girder rails " to distinguish them from the old flat rail, which was in no sense a girder. The rails commonly used on steam roads in this country are called T rails, and have the properties of the girder, as much as the special street-railway rails to which that term is applied. In fact, they more nearly resemble a girder, in that they are symmetrical about a vertical axis, a property which only the center-bearing type of girder rails possess. The term "T rail" is, strictly speaking, a no more accurate one than "girder rail," but both are accepted by common usage. The T rail is used exclusively on the 175,000 miles of steam railroads in the United States, and so far superior is it to any other section that there is little wonder we find it used in an ever-increasing percentage on the 13,000 miles of street railways. The number of sections and the variety in designs far exceed those of the girder rail, which fact is the more remark- able when it is considered how much simpler the sections are, and the almost uniform conditions under which they are employed. Figs. 79, 80, 81, 82, 83 show sections that are standard on several of our most important steam roads, and Fig. 84 one of the sections of the American Society of Civil Engineers. For the rake of comparison I have selected those of about equal weight, and which would be suitable for a well-built electric rail-R ay carrying a heavy traffic. The difference in design is quite marked, even to the unpracticed eye. 34 THE T KAIL AS ADAPTED TO STREET RAILWAYS. 35 \ 2 it' FIG. 83. STANDARD STEAM-RAIL SECTIONS. FIG. 84. STANDARD SECTION OF THE A. S. C. E. 36 STREET-RAILWAY ROADBED. The conditions on steam roads differ materially from those on street railways, mainly in the fact that the track is exposed, so that joints and all other fastenings are accessible, and also in that they are subject to heavier traffic at higher speeds. On street railways the conditions met with in subur- ban and interurban lines differ from those in the cities and more populous districts, and approach more nearly to those of the steam roads, in so far as the construction of the roadbed is concerned. They are often such as to admit of the use of the T rail, and since this section possesses many advantages over the girder rail, the privilege of using it is gladly embraced. The T rail is held in such great favor by street- railway men that many successful efforts are being made to use it, though in a modified form, in city streets in a manner shown later on. Naturally, therefore, the use of T rail on street railways should be considered under two heads, that in suburban and interurban lines, or in unpaved track, and that in city lines, or in paved track. In this order, then, they will be taken up. The points of superiority of the T over girder rails, briefly stated, are as follows : 1. It is cheaper, for the price per ton is less than the girder rail, and owing to the absence of the tram or groove, and also owing to its symmetrical section, a much lighter rail may be used than in a girder track under similar conditions, at the same time obtaining an equally substantial track. 2. It is easier to lay, for, the section being symmetrical, the application of the joints becomes a simpler matter, although this requires careful attention in all track-laying. Easy curves may be sprung in, and sharper ones be laid by means of a portable bender, with a greater certainty of the track keeping its alignment than with girder rails. Then there may always be found a greater number of trackmen familiar with the lay- ing of T rails than girder. These points are of vast impor- tance, for on them depends the durability of the track, more than upon the shape of the rail or fastenings or any other feature. THE T KAIL AS ADAPTED TO STREET RAILWAYS. 37 3. It has a cleaner head, owing to the absence of -the tram and groove of the girder rail. It offers no attraction to street traffic, and if properly laid in a well-macadamized, brick-paved, or asphalted street, the track will usually be found in a much better condition than with any girder rail. > The last two conditions contribute, of course, to a better track in every way for operation, for track frictions are reduced to a minimum. And since the track surfacing is not hampered by street grades, the outer rail of curves may have its proper elevation, and cars may run at the maximum speed at all points. Steam -road practice may be followed very closely on inter- urban lines where the track is exposed, and in Figs. 85 and 86 are given standard cross-sections of single- and double-track construction on the Pennsylvania Railroad, which prob- ably represent the best practice. These sections show in a complete and comprehensive manner the method of grading, draining, ballasting, etc. Below are given extracts from the Pennsylvania Railroad Company's general specifications gov- erning the construction, all of which could be followed to advantage on that class of electric track under consideration. P. R. R. SPECIFICATIONS FOR LAYING ROADBED. Roadbed. The surface of the roadbed should be graded to a regular and uniform sub-grade, sloping gradually from the center towards the ditches. Ballast. There shall be a uniform depth of six to twelve inches of well-broken stone or gravel, cleaned from dust by passing over a screen of one-quarter-inch mesh, spread overthe roadbed and surfaced to a true grade, upon which the ties are to be laid. After the ties and rails have been properly laid and surfaced, the ballast must be filled up as shown on standard plan ; and also between the main tracks and sidings where stone ballast is used. All stone ballast is to be of uniform size and the stone used must be of an approved quality, broken uniformly, not larger than a cube that will pass through a S^-in. ring. On embankments that are not well settled the surface of the roadbed shall be brought up with cinder, gravel, or some other suitable material. Gross-ties. The ties are to be regularly placed upon the ballast. They must be properly and evenly placed, with ten inches between the edges of bearing surface at joints, with intermediate ties evenly spaced ; and THE T RAIL AS ADAPTED TO STREET RAILWAYS. 39 the ends on the outside on double track, and on the right-hand side going north or west on single track, lined up parallel with the rails. The ties must not be notched under any circumstances ; but, should they be twisted, they must be made true with the adze, that the rails may have an even bearing over the whole breadth of the tie. Line and Surface. The track shall be laid in true line and surface ; the rails are to be laid and spiked after the ties have been bedded in the ballast ; and on curves, the proper elevation must be given to the outer rail and carried uniformly around the curve. This elevation should be commenced from 50 to 300 ft. back of the point of curvature, depending on the degree of the curve and speed of trains, and increased uniformly to the latter point, where the full elevation is attained. The same method should be adopted in leaving the curve. Joints. The joints of the rails shall be exactly midway between the joint-ties, and the joint on one line of rail must be opposite the center of the rail on the other line of the same track. A Fahrenheit thermometer should be used when laying rails, and care taken to arrange the open- ings between rails in direct proportion to the following temperatures and distances: at a temperature of deg., a distance of T 5 F in.; at 50 degs., / in.; and in extreme summer heat, of, say, 100 degs. and over, -fa in. must be left between the ends of the rails to allow for expansion. The splices must be properly put on with the full number of bolts, nuts, and nut-locks, and the nuts placed on inside of rails, except on rails of sixty pounds per yard and under, where they shall be placed on the outside, and screwed up tight. The rails must be spiked both on the inside and outside at each tie, on straight lines as well as on curves, and the spikes driven in such position as to keep the ties at right angles to the rails. Switches. The switches and frogs should be kept well lined up and in good surface. Switch signals must be kept bright and in good order, and the distant signal and facing-point lock used for all switches where trains run against the points, except on single-track branch roads. Ditches. The cross-section of ditches at the highest point must be of the width and depth as shown on the standard drawing, and graded par- allel with the track, so as to pass water freely during heavy rains and thoroughly drain the ballast and roadbed. The line of the bottom of the ditch must be made parallel with the rails, and well and neatly de- fined, at the standard distance from the outside rail. All necessary cross- drains must be put in at proper intervals. Earth taken from ditches or elsewhere must not be left at or near the ends of the ties, thrown up on the slopes of cuts, nor on the ballast, but must be' deposited over the sides of embankments. Berm ditches shall be provided to protect the slopes of cuts, where necessary. The channels of streams for a consid- erable distance above the road should be examined, and brush, drift, and other obstructions removed. Ditches, culverts, and box drains should 40 STKEET-RAILWAY ROADBED. be cleared of all obstructions, and the outlets and inlets of the same kept open to allow a free flow of water at all times. Road Crossings. The road-crossing planks shall be securely spiked; the planking on inside of rails should be in., and on outside of rails it should be in., below the top of rail, and 2^ in. from the gauge line. The ends and inside edges of planks should be beveled off as shown on standard plan. A fine example of suburban electric-track construction following steam-railroad practice is that of a Baltimore road built by D. B. Banks, C.E. Fig. 87. It often happens, however, on suburban lines, that the track is laid in or alongside of the public highway, and it is required to fill the track and keep the surface to its level. In these FIG. 87. T-RAIL CONSTRUCTION ON SUBURBAN LINE IN BALTIMORE. cases the sub-construction should be the same as above de- scribed, and the ballast be brought up to the level of the top of the rail. This construction offers less obstruction to travel, THE T KAIL AS ADAPTED TO STREET RAILWAYS. 41 and permits teams to cross at any point; although it is not in- tended that the track itself shall be used for street traffic. When T-rail track is laid in macadamized streets, the only change in construction necessary is the use of a layer of finer FIG. 90. T-RAIL SECTIONS. material for the top course, with a free use of a cavy road roller. It is a gradual transition from the open track to the track in paved streets. With the latter comes not only a change in the general style of construction, but one in the rail itself- 42 STREET-RAILWAY ROADBED. All pavements, save asphalt, require a deep rail, and the man- ner in which this change is brought about in the T rail is well shown by Figs. 88, 89, 90, of five-, six-, and seven-inch rails. Eails of even eight and nine inches have been pro- posed. These all have the characteristic square head. The base increases in width with the height. The fillet at the lower outer corner of the head is made as small as practicable, in order to obtain all the bearing possible for the splice-bars, a very essential feature. There is one point, however, wherein these high T's show up to a disadvantage, and that is in the alignment. Not having the lateral stiffness of the girder rail, with its tram or lip, the track is apt to present a "wavy" ap- pearance which braces and tie rods will hardly prevent. This is a matter which appeals more to the eye of a good trackman than to the operator. These rails have been very successfully laid in the smaller cities in connection with well-macadamized roads and brick pavements. Even the shallower granite block pavements may be laid with this track with success. This construction is becoming more popular as the prejudice to T rails wears away. New England has many examples of well- laid T-rail track in city streets, and in connection with nearly all kinds of pavement notably at New Haven, Bridgeport, and Waterbury. There is also a notable example of good road- bed in the West at Terre Haute, where, however, the construc- tion is unusual in that a concrete base is employed and steel ties used a construction that would do full justice to any rail. CHAPTER V. TRACK FASTENINGS AND JOINTS. RAILS cannot be made, like wire, in indefinite lengths; nor could they be shipped or handled conveniently in lengths of over sixty feet, the longest rails laid to-day. The usual lengths are thirty, thirty-two, forty-five, and sixty feet. There must be some good connection between these separate rails when laid, to insure a continuity of track surface, and the question of joints comes forward as one of greatest moment. As a matter of fact, the joint has received fully as much con- sideration as the rail, since the advent of the electric motor. What constitutes a good joint ? Simply a fastening which, when properly applied to the abutting rail ends, will hold them in as good line and surface as the body of the rail. It is a well-established fact that rails laid in paved or mac- adamized streets and covered to their full depth, as is usually the case on street railways, may be " butted/' i.e., laid without any opening between the ends for expansion. This can be done without danger from the effects of changes of tempera- ture. This fact simplifies to a great extent the problem of making a good joint; for any opening, even as small as in., will cause a " pound " on the passage of a wheel. The jar resulting from this pound will cause nuts, clips, and other fastenings to loosen in time, and consequently produce a defective spot in the track. There is a considerable strain produced in a line of rails by the changes of temperature,* but there is no appreciable move- * The coefficient of expansion for steel due to a change in temperature of 1 F. is .00000688 (Ganot). The rate of elongation of a bar of rail-steel when subjected to a tensile strain within its>lastic limit is, according to the average of a large number 44 STREET-RAILWAY ROADBED. ment because the whole effort is absorbed by the elasticity of the metal and the vise-like grip of the pavement and sur- rounding material. Kails are usually ( ' hot-sawed," i.e., sawed to length as they come from the rolls at a bright red heat; and as it is quite im- possible to thus make a perfectly smooth and square cut, the practice is to slightly undercut them. The amount of thi s undercut is about r ^ in., leaving an opening at the base of about ^ in. when the heads abut. The most common form of joint is that made with two plates in the shape of shallow channels, placed one on either side of the rail, and taking a bearing on the inclined surfaces of the head and base. They are held in position by bolts passing horizontally through both plates and the rail. The bolts used are called track- bolts, and have a button head. That portion of the shank next the head, for a distance equal to the thickness of the joint-plate, is of oval form, and as it fits a hole of the same shape, the bolt is prevented from turning when the nut is be- ing put on. The nut is either square or hexagonal. Where there is room for it to turn, the square nut is preferable in giv- ing more bearing against the joint-plate and a better grip for the wrench. The hexagonal, or "hex," nut is used where there is less clearance, as is often the case on angle-joints on T rails and on the deep girder rails where there are two rows of bolts ; for in these latter cases it is desirable to get the bolt as close as possible to the edge of the plabe. The most essential feature of a joint is that the plates shall have as much bearing as possible. There is little gained of tests made by the Pennsylvania Steel Company, about .00006 in. per 1000 Ibs. per square inch. Dividing the temperature coefficient by the latter, we get 114.6 Ibs. as the strain per square inch produced in a rail due to a change of 1, provided the ends are rigidly held and that there can be no lateral or vertical bending when compression takes place. Assuming a change in temperature of 100, we get a strain of 11,460 Ibs. per square inch, which is about one-fifth of the elastic limit and one. ninth of the ultimate strength of rail-steel. As a matter of fact, however, rails are not usually laid with the entire end surfaces abutting perfectly, and there is some chance for a small movement by compression of the small ridges produced by the saw. Then, again, it is highly improbable that a rail will be under no strain at one of the extremes of temperature, but that the point of no strain will be at an average temperature. So that, under these considerations, it is not probable that the strain in a rail will ever exceed one-third the figure given above, or about 4000 Ibs. per square inch a strain which is absolutely without danger of any kind. TRACK FASTENINGS AND JOINTS. 45 in making any of the flanges wider than others, for a joint is like a bridge in this respect, that the weakest part determines the strength of the whole. That portion of the rail which generally determines the width of the joint-plate flanges is under the head, and whatever width of bearing may be obtained here should be used at the other points. It is ex- tremely important that the joint-plates should fit the rail, and that when drawn up by the bolts the bearing surfaces shall be in contact with the rail with a uniform pressure over their entire surface. In order that the plates may not bend under the strain of the bolts and destroy this bearing, they are always made convex. This is a proper feature, but even an arch when not of sufficient thickness to carry its load will fail, as do many joint-plates by being pulled in against the web of the rail. Fig. 91 clearly shows the result of such FIG. 91. BUCKLED JOINT-PLATES. action. The bearing instead of being distributed over a sur- face is concentrated along a line. The effect of this is to rapidly wear away the parts of the rail and joint in contact, and thus loosen the joint. One cause for this lies in the in- creased size of the bolt used, without increasing the thickness of the plates proportionately. Larger bolts are used because 46 STREET-EAILWAY ROADBED. of their tendency to remain tight, due to the frictional resist- ance of a largely increased thread area. Plain or channel plates for 6-in. rails should be not less than T 9 T in. thick at the center, those for 7-in. rails -f- in., and for 9-in. rails not less than } in. Even with these heavy plates there is danger of their being bent in sufficiently to destroy the fit unless some care is exercised in tightening up the bolts. Plates of less thickness than those given above have ample vertical stiffness, and in order to prevent this in- ward bending on deep girder rails having a double row of bolts the writer devised the " ribbed " plate shown in Fig. 92. (See also Figs. 52, 55, 65, and 68.) The 8|-m. rail (Fig. FIG. 92. RIBBED JOINT-PLATES. 65) was the first section made having this type of joint. It was laid on the Atlantic Avenue Railroad in Brooklyn, N. Y., in 1892-3. With this center bearing it is readily seen that bending is prevented and the true bearing of the joint-plate flanges insured, even when bolts are tightened up to the limit of their strength. In this connection it might be well to explain how rolled joint-plates should be applied. Rails fresh from the mills are covered more or less with a thin coat of black oxide of iron. TEACK FASTENINGS AND JOINTS. 47 Much of this falls off during the process of straightening, loading and unloading ; but there is always some adhering to the rail when placed in the track. This, to my mind, is one of the worst enemies of the joint, for after the latter is applied and the track is used the jar of passing wheels reduces this scale to a thin powder. This powder, working its way out from between the rail and joint-plate, leaves the 'latter loose* or well started in that direction. This coating of oxide or ''scale" is also found on the joint-plates. Therefore the first thing to do is to remove it from the bearing surfaces, which may be done with a light hammer, a file, or a scraper. By the time the rail reaches its destination this scale will be found only in patches, and probably the first tool mentioned is the best for the purpose, as the scale easily crumbles off after a few light blows on the spot. This is a matter which is of considerable importance, and yet is often, if not always, over- looked. The next step is to place the plates in their proper position and, putting in all the bolts, screw them up only sufficiently tight to hold the rails snugly. Care should be exercised here, as well as at each subsequent operation, to see that the plates go on evenly. After the spiking has been done and the track surfaced, all bolts should be gone over carefully, pulling every nut up tight. This may be done most effectively with a two- foot wrench ; and while pulling on the wrench, tap the head of the bolt with a one-pound hammer. A few blows on the plates and on the head of the rail with a light sledge during this proceeding will have a beneficial effect. Again, after the track is finally lined and surfaced, every bolt should be gone over with wrench and hammer. A final inspection before filling in will do no harm. If plain channel joints have been used, too much care can- not be exercised, in drawing up the bolts, not to bend the plates, for they will do more good when bearing evenly against the rail-flanges, even if the bolts are not as tight as they might be. Channel joints are used from 20 in. to 38 in. long and with four to twelve bolts. The bolts should be either in. or 48 STREET-RAILWAY ROADBED. 1 in. in diameter. The writer is of the opinion that for 6-in. and 7-in. rails the joint should be 26 to 36 in. long, f in. thick, and have six or eight 1-in. bolts; on rails deeper than 7 in. a joint of about the same length, or even shorter, with two rows of six bolts each. Plain channel bolts should be -J in. thick; "ribbed" plates may be ^ in. and the bolts 1 in. in diameter. The spacing of the bolts is a matter about which there seems to be considerable diversity of opinion, but that in which the length of plate is divided evenly will give as good results as any. Nut-locks are frequently used on street-railway tracks; but while admitting that there are many excellent devices intended to hold the nuts up to their work which may be of service on exposed tracks, it is my opinion that their use is an unnecessary expense on tracks that are covered in. In these cases there is quickly formed a coating of rust which, with the grip of the surrounding gravel and sand, holds the nut as in a vise, and if the joint becomes loose it is from other causes. The joint made as above described and properly applied on rails of seven inches and over will give very satisfactory re- sults. The deeper the rail, however, all other things being equal, the better the joint is apt to be, for with the stiffer rail the tendency to a movement between the parts of the joint is lessened. With the nine-inch girders as laid to-day this movement of track is practically nothing. There are many forms of the bolted and keyed joints, some of which possess considerable merit, and the question of their use is one to be settled by the manager in each case. We give below descriptions of a few of the most important ones. The girder joint (Fig. 93) is of that class of joints which grip the base of the rail. In addition to this it performs another office that of a chair and is therefore best adapted to rails of six inches and less in height, although they are used on seven-inch rails to some extent. There is no doubt that joints of this type would be used much more extensively than they are but for the fact that solid deep-rail track can now be bought for about the same price as the shallow and lighter rails with chairs. The girder joint is not so well TKACK FASTENINGS AND JOINTS. 49 adapted to deep rails: first, because it does not hold the rails in strict alignment ; and second, because the ties are thrown so far below the surface. FIG. 93. GIRDER JOINT. The Wheeler rail-joint (Fig. 94) is made of malleable cast iron in two parts, and is without bolts. One of the parts, the larger one, called the " housing," has a bearing surface extending under the entire width of the base of the rail, and has formed under this "shelf a tapered pocket which receives a wedge FIG. 94. WHEELER RAIL- JOINT. formed in the lower side of the other part. The housing is provided with lugs engaging holes in the rail-web which pre- vent its slipping from a central position when the wedge por- tion is driven home. The whole is well braced by ribs. The manufacturers say: "There is no attempt to assist the web in holding up the rail-head, as we believe all T and tram rails are or should be stiff enough to carry the traffic, and are as 50 STREET-RAILWAY ROADBED. stiff at the ends as at any other portion of the rail length, and that the province of a rail-joint is to prevent motion of rail ends by keeping the bases and webs in perfect alignment and immovable, thus insuring a permanent alignment and surface of the head and tram." The Weber joint (Fig. 95) is peculiar in having, in addition FIG. 95. WEBER HAIL- JOINT APPLIED TO GIRDER-RAIL. to the two well-fitting channel joints, an angle which has one leg extending under the rails. The space between the vertical leg and one of the joint-plates is filled with a piece of sound Georgia pine. The bolts pass through the two plates, the rail, the pine filler, and the angle. Of course such a joint is more expensive, first cost considered, than the ordinary joint ; but it is claimed that the addition of the angle not only stiffens the whole joint greatly, but also maintains the rails in good surface. The elasticity of the pine filler keeps the whole joint tight by taking up all loosening effect of wear. The Weber joint has made a remarkable record on open track, where it not only maintains a good surface when applied to new track, but on old track with " low " joints it has brought the rails up to line and surface, which it is not possible to attain with the ordinary angle-joints. Two other joints made on the principle of base support are TRACK FASTENINGS AND JOINTS. 51 used to some extent: the "continuous" joint (Fig. 96) and the "Churchill" joint (Fig. 97). With the knowledge of the fact that rails may be laid with FIG. 96 "CONTINUOUS" RAIL-JOINT. FIG. 97. " CHURCHILL" RAIL- JOINT. butted joints naturally comes also the question: Is it not possible to make and apply a permanent joint which will 52 STREET-RAILWAY ROADBED. make the track practically two continuous rails ? There are two methods now in use which endeavor to reach this end, namely, the process of electrically welding the rail ends together and that of "cast-welding." The former process consists in fusing a piece of metal on each side of the web at the joint by passing through them, when held tightly against the rail, a current of low voltage and great volume. Some attempts have been made to unite the rail ends directly. The process of electric welding is not welding in the ordinary sense, but a melting of the sepa- rate pieces together. To do this properly with steel requires the expenditure of about fifty horse-power per square inch. The operation is completed so quickly that a few inches from the point of melting steel the rail is quite cold. The heated portion on slowly cooling passes through an annealing process and leaves a distinct line of demarkation between two conditions of steel which is also a line of weakness. This was proved by a considerable percentage of breakages taking place at this point when the track became subject to the strains produced by changing temperature. The apparatus used is necessarily cumbersome and expensive, and it is questionable if the results attained are commensurate with the expense. The " cast-welded " joint consists simply in a mass of cast iron poured around the abutting rail-ends, uniting through holes in the web. It is possible to make a very close union between the cast iron and the rail, and with the proper amount of iron a very strong and substantial joint can be produced. The results from this method of joining rails have been quite satisfactory and several roads particularly in the West have adopted this type of joint exclusively. The use of shallow girder rails required the use of some support between the rail and tie in order to secure sufficient depth for the paving. The first provision made with this ob- ject in view was simply the old construction for flat rails, with a longitudinal timber stringer dropped to a depth sufficient to accommodate the rail. It was claimed for this construction that a continuous support was provided TRACK FASTENINGS AND JOINTS. 53 under the rail and joints. As the timber rapidly decayed under the joints, the latter claim soon proved its weakness. The desire for a cheaper construction, and also the laud- FIG. 98. Fm. 99. VARIOUS FORMS OF CHAIRS AND THEIR FASTENINGS. able wish to avoid the use of so much timber in the street, led to the use of metallic chairs. They were at first made of cast iron, but the bad fits produced with this mate- rial, combined with the many breakages in applying and 54 STREET-RAILWAY ROADBED. use, led to its_'abandonment in favor of rolled or forged chairs. If the metal was properly distributed in these, a fairly stiif FIG. 100. FIG. 101. VARIOUS FORMS OF CHAIRS AND THEIR FASTENINGS. structure was obtained, especially in forged chairs in which a bracket was struck up, as shown in Fig. 103 and to a more TRACK FASTENINGS AND JOINTS. 55 notable degree in Fig. 104. Examples of various forms of chairs and their fastenings are shown in Figs. 98-105. FIG. 103. VARIOUS FORMS OP CHAIRS AND THEIR FASTENINGS. In order to support the joint in chair construction the 56 STREET-RAILWAY ROADBED. FIG. 104. FIG. 105. VARIOUS FORMS OF CHAIRS AND THEIR FASTENINGS. FIG. 106. CONSTRUCTION WITH TIE-PLATES AND TIE-RODS. TRACK FASTENINGS AND JOINTS. 57 girder joint shown in Fig. 105 was introduced, as previously mentioned. The lower girder rails were difficult to spike properly on account of the overhanging tram. The use of a tie-plate such as is shown in Fig. 107 obviated this difficulty to some extent, as well as serving to protect the tie from wear. The gage is prevented from spreading by several methods; the earliest and yet most common is by the use of a tie-rod between the webs of the rails. The most usual size is a flat bar 1-J" by f " with the ends forged to f " round and threaded, FIG. 107. SECTION WITH MALLEABLE IRON BRACE. a nut is placed on each side of the web and tightened up as required, thus providing a convenient method of adjusting the rails while laying track. The rod is of considerable length and is liable to stretch somewhat, thus widening the gage. As it is necessarily applied at some little distance below the head, a leverage is developed which tends to pull the inner spike and thus widen the gage. Altogether, it is far from a perfect fastening. The construction with tie- plates and tie-rods is shown in Fig. 106. A malleable iron brace is sometimes used, such as is shown in Fig. 108. It provides a support directly under the head, and thus resists the outward thrust most efficiently, except 58 STREET-KAILWAY KOADBET). that it depends for its own stability almost entirely upon the hold of the spikes in the tie. A further advance is recorded in brace tie-plates, such as is shown in Figs. 109 and 110. This has the advantage of the FIG. 108. SECTION WITH MALLEABLE IRON BRACE. FIG. 109. VARIOUS FORMS OF CHAIRS AND THEIR FASTENINGS. support under the head, and also utilizes the load on the rail to retain the brace tie-plate itself in position. "With the substitution of a metallic tie, such as is shown in Fig. Ill, the use of many of these fixtures will be dispensed with. At present prices there is little difference between the cost of a first-class wooden tie and a steel tie suited to street-rail- way use. With any pavement or track construction using a TRACK FASTENINGS AND JOINTS. 59 concrete foundation the writer would recommend their adop FIG. 110. VARIOUS FORMS OF CHAIRS AND THEIR FASTENINGS. FIG. 111. CONSTRUCTION WITH METALLIC TIES. tion. The very slightly increased cost will be more than re- paid by the saving in repairs and renewal of ties. CHAPTER VI. SPECIAL WORK. CURVES. IT would be safe to assert that there has never been built a street-railway system that has not had a piece of track that required some special preparation other than that given to plain, straight track before it could be laid in place. Most systems have a considerable percentage of their trackage made up of curves, crossings, switches, etc. In nearly every case these curves and crossings have to be made specially to fit given locations, and hence the term "special work." In tracks made with rails of five inches or under, all curves over 500 ft. radius may be " sprung in " as the construction proceeds; and if the track is otherwise well laid the alignment may be depended on to remain good. But with all heavier rails, particularly girder rails, no curves under 1000 ft. radius should be laid without first curving the rails with a portable bender; and for those under 300 ft. radius the rails should be put through a power machine. In no other way is it pos- sible to avoid angular joints. The writers are familiar with several cases, and one in particular, where a piece of track was laid with seven-inch girder rail, in which there are several curves, varying from 400 ft. to 1000 ft. radius, which were " sprung in." It was laid by a skillful trackman and engineer and paved in brick. The alignment when new was fine, but after one year's traffic under a five-minute headway the joints began to show themselves by a slight angle in the line and a perceptible jerk of the car in passing. It would there- fore seem the better practice to avoid the habit of "spring- ing in " light curves. Street-railway curves are always designated by the radius, and not by the degree of curvature in hundred-foot chords as on steam roads. The chord method is not generally used 60 SPECIAL WORK. CURVES. 61 in laying them out, except where they approach the dimen- sions of steam-road curves. And aside from other incon- veniences, it is manifestly impossible to designate curves by that method when the radius is under fifty feet a 180-deg. curve. With the higher speeds that have come with mechanical power, it is desirable to have easier -running curves than the simple circular curves heretofore commonly used. This may be obtained by "compounding," i.e., starting with a long radius curve and increasing the curvature, or shortening the radius at intervals till the desired curvature is reached for the central portion of the curve. This is done at each end, making usually a curve which is symmetrical about the radial line at its center. Theoretically the most satisfactory curve is a spiral with constantly increasing curvature such, for instance, as the hyperbolic spiral or the logarithmic spiral. But practically a compound curve as described above is better; for it is much easier to figure, and if the compounding is properly done, and the curve properly laid, it will be irnpos sible to detect any difference in the motion of the car in passing either. As to the method of compounding curves, there has been a considerable improvement within the last three years. At first three-center and five-center curves were used, all arcs being about the same length. But as the demand for greater refinement arose, a close approximation to the true spiral was obtained in the adoption of compound curves made up of arcs of five feet or less. It will be shown later that on three- center curves and the same is true of compound curves having arcs of greater length than the wheel base, combined with large changes in radius the ends of the car follow a peculiar path which imparts a jerky motion rather unpleasant to the passengers. Of course the motorman, with one hand on the controller and the other on the brake, has a consider- able influence over the manner in which the car passes around a curve; and with a little effort or the lack of it he may knock into a cocked hat the greatest refinements of the engineer and track-layer. STREET-RAILWAY ROADBED. On a double-track road it is well to have tht curves so laid out as to allow cars to pass each other on them, although on most roads there is a rule against it, both on account of the greater liability to accidents, as well as to prevent a heavy FIG. 112. DIAGRAM SHOWING OVERHANG ON A PAIR OF SIMPLE CURVES. drain on the power station. But to meet the cases where rules are not always followed, as well as where there are none, col- lisions of cars may be prevented by a little more care in de- signing the curves. In working out an easement for any given case, the outside dimensions and wheel base of all cars to be used must be determined. The most convenient method of plotting the curves is to have the outline of a car's horizontal SPECIAL WORK. CURVES. 63 projection cut from cardboard or transparent celluloid, with the position of the center of the axles shown, or the center of trucks in the case of double-truck cars. Having laid down a curve with its center line, the space that the car will occupy FIG. 113. DIAGRAM SHOWING COMPOUND CURVES. will be found by placing the template at successive positions on the curve and marking the outer corners and the inner side at the center. Fig. 112 very clearly shows the overhang on a pair of simple curves. It will be noted that there is clearance at the center, due to the fact that the curves are not concentric a very common way of laying curves. But at the ends the cars overlap, and if they attempted to pass each other at those points there would be a collision. In Fig. 113 it is shown how 64 STREET-RAILWAY ROADBED. it is possible to compound the curves to obtain clearance all the way. In this diagram note that the compound curve has the same position at the center as the simple curve, and will therefore fit the same location. To obtain this it was neces- sary to cut down the center radius but 5 ft. The car used is a 33-ft. body on a six-wheeled "radial" truck. But the principle involved is the same with any cars. If cars of more than one kind are to be used, of course they should all be tried and a curve found that will suit all. This method may also be used to great advantage in laying out car-house curves, locating posts and poles, or to clear any fixed obstruction. The several factors that enter into the problem of overhang and car clearance are: The length and width of cars and the shape of their ends; The wheel base ; The distance between track centers on tangent ; The curvature; The elevation of one rail above the other. In addition to the first two, which relate to the car, should be mentioned the rigidity with which the body is attached to the trucks laterally. If there is any swing of car body there will be danger where small clearances have been figured on ; but this is largely under the control of the motorman. The third item, or distance between trucks, plays a very important part, not only on curves, but on straight track. A very common distance on standard (4 ft. 8| in.) gage is 4 ft. from back to back of head, giving about 9'-0|" center to center. With ordinary cars this gives a clearance on straight track of 12 in. to 18 in., and reduces possible clearances on curves to a minimum. Cars measuring 8 ft. and over in width are coming into use, while open cars with running boards are even wider; and with people standing on the step, there would be a clearance of about 6 in., which is entirely too small. It is suggested that a distance on centers of not less than 10 ft. be used. To be sure, this adds nearly 550 sq. yds. of pavement per mile to be laid and maintained, but the com- pany will be fully repaid in the added security to its passen- gers and cars. In cases where the railway company is made SPECIAL WOKK. CURVES. 65 responsible for the pavement over the entire width of street this question of additional cost will not arise. Where center- n o o Q n, to * in (On ^ pole construction is used, the center-to-center distance should be not less than 12 ft., and then care should be exercised not to have a pole anywhere near a curve. 66 STREET-RAILWAY ROADBED. A large percentage of the short-radius curves are required for angles approximating 90 degs., and the accompanying diagram (Fig. 114) will be found of much value in selecting the proper curve to be used. The curved lines indicate the inside rail of a single-track curve, and the edge of the diagram the center lines of track. To illustrate its use, suppose we have to pass around a corner with a double-track curve; the streets are 40 ft. and 60 ft* between curbs, and the tracks are 10 ft. between centers, making the center of the track in one case 15 ft. and the other 25 ft. from the curb. Following the 15-ft. and 25-ft. lines to their intersection, we find it to be just inside of a 60- ft.-radius curve; showing that to be the largest radius that could be used in the given case. In designing curves there are numerous other things to consider, chief among which are sewer and water manholes. The position of the former may often be changed by going down three or four feet and building up on a slant. It is seldom possible, however, to alter a water manhole or stop- cock plug, and they sometimes prove annoying, and require some nice work in compounding. Then there are sewer intakes, lamp-posts, telegraph, telephone, and electric-light poles, and the shape of the curb corner itself, and the question of dodging or removing them to be settled. The direction of flow and amount of surface drainage should also be considered, for it will not do to obstruct or divert it to the damage of abutting property. All of these points require the careful consideration of an engineer. On curves of 300 ft. radius, and under, it is not safe to de- pend entirely on the bearing of the flange aga;nst the gage line of the outer rail to keep the cars on the track, and a guard-rail on the inner or short side of the curve should be used. The guard is from in. to | in. higher than the head of the rail, and with its broader bearing against the back of the inside wheel-flange prevents derailment. Theie are many who think that curves from about 100 ft. radius down should have a guard on the outer or long side as well, on the ground that the rear wheels have a tendency to run off on the inside of SPECIAL WOKK. CURVES. 67 the curve. There is such a tendency, but it is so slight and so nearly overcome by other forces that it does not require the services of an outer guard to keep them on. An experi- ence with several thousand curves built with single guard and used by cars of all conditions and gages has tended to con- firm this latter proposition. There are two decided benefits to be derived from the use of but one guard, namely, a sav- ing in first cost of the curve, and a continual saving in power required to move cars around the curve, due to an avoidance of additional flange friction. Fig. 115 shows very clearly the position a rigid four-wheel FIG. 115. DIAGRAM SHOWING POSITION OF RIGID FOUR WHEEL TRUCK ON CURVE. truck assumes in traversing a curve. The shaded portion of each flange represents that which is below the top of the head of the rail, and the black portion of same the part of the flange in contact with the rail, the arrow indicating the direc- tion of travel. The truck is guided almost entirely by the inner front wheel (1), and the flange of its mate (2) is in con- tact only when the gage of track and truck permit. Wel- lington shows that when not restrained by the flanges, the rear pair of wheels will follow the forward pair in the manner shown in "Fig. 116, i.e., with the rear axle on a radial line. On a 35-ft.-radius curve the distance of the rear wheels from the gage line would be about 7 in., and on a 100-ft. curve 2 in. It is apparent, therefore, that the duty imposed on the 68 STREET-RAILWAY ROADBED. rear-wheel flange is simply to keep that portion of the truck a few inches away from its normal position a condition that may be illustrated by a pendulum held to one side by the pressure of a finger. As the angular distance is small, so is the force required but a small percentage of the weight of the FIG. 116. DIAGRAM SHOWING POSITION OP RIGID FOUR-WHEEL TRUCK ON CURVE. pendulum. Now if the pendulum be swung in a circle about its normal position, the centrifugal force will sustain it away from the vertical. So with a car, while the tendency of the rear wheels to climb the inner rail is comparatively small of itself, it is counteracted to a greater or lesser extent, depend- ing on the speed, by the centrifugal force acting through the car. All of which goes to show that an outer guard-rail is more of a hinderance than a necessity. Within the last three years the form of the groove in guard- rails has undergone a decided improvement. There are shown herewith the three sections of solid guard-rails in use at the present time (Figs. 117, 118, 119). The idea in all is to have a form of groove that will best fit the wheel-flange and present its full face to the wear of the flange. Since the exact shape of the groove depends upon the size and shape of the wheel- flange, the diameter of the wheel, the wheel base, and the radius of the curve all variable factors it is manifestly im- possible to have a different guard-rail to suit every condition. SPECIAL WORK. CURVES. 69 The manufacturers have therefore settled on a form which is best suited to the average conditions. A careful comparison FIG. 118. SECTIONS OF SOLID GUARD-RAILS. of the figures will show that there is practically no difference in the contour of the groove. As the shape of the groove was 70 STREET-RAILWAY ROADBED. determined in a different way by each, it is interesting to note the closeness of the results. The method pursued by No. 3 was to fill the groove of an old curve with plaster of Paris or clay and run a car around, noting the form of groove made by FIG. 119. SECTION OF SOLID GUARD-RAIL. both front and rear wheels, and a number of trials were made under different conditions. The manner in which No. 2 made the determination was by the use of quarter-size models which were pivoted to a fixed center an'd run over a clay track. Sec- tions of the groove formed were carefully cut and dried, and the final section was the result of a large number of observa- tions. The test is more fully described in the Street Railway Journal for June, 1895, page 399. No. 1 was determined by observation of numerous worn guard-rails, from which wooden templates were taken and compared. It was noticed that, no matter what the original form of the groove had been, the worn groove assumed a definite shape, depending largely on the radius of the curve. On each section is shown in dotted lines the portion worn away in service. The shape of the worn groove was taken by a template from a curve in use, and it is assumed that when the groove is worn to a width of 2 in., and the head cut down T 6 ^ in., it is time to renew. A study of these three sections will show that there is a great waste of SPECIAL WORK. CURVES. 71 metal in the guard of No. 2 and No. 3. When the wear has reached the point indicated by the broken line the guard-rai has about served its usefulness and should be renewed. They are all of about equal strength as to the guard turning out, and any more metal than shown by No. 1 adds nothing but so much more scrap to be thrown away. After all it seems rather absurd to use many refinements in designing a guard rail groove when one of the main factors the shape of the wheel flange is so variable, as is clearly FIG. 120. DIAGRAM SHOWING DIFFERENCES IN SECTIONS OF WHEEL-FLANGES AND TREADS. shown by Fig. 120 a composite of sixteen wheel treads and flanges placed with their gages coincident. These sections were taken by templates from wheels in service. The largest and smallest were found on the same road. CHAPTER] VII. GUARD-KAILS.^[SPECIAL~"WORK. IT is a comparatively simple matter tcTdetermine graphi- cally the shape of groove in a guard-rail, having given the FIG. 121. SECTION OP RAIL AND WHEEL-FLANGE. flange, diameter of wheel, radius of curve, and wheel base. By reversing the operation the maximum flange may be found that will pass through any groove, having given the radius, 72 GUARD-BAILS. SPECIAL WORK. 73 wheel base, etc. It was by such a process that the flange shown in Fig. 121 was developed. The guard-rail taken is section No. 208 of the Pennsylvania Steel Company, and the broken line shows the limits of the space in the groove occu- pied by the flange on a thirty-five-foot-radius curve. The flange which the groove will allow to pass on such a short- radius curve, it will be noted, is about as large as any in use. The conditions on curves of longer radii being more favorable, the flange will pass them freely. The section of rail and wheel-flange shown in Fig. 121 also gives us a clue to the proper gage of tracks to be used on curves a very much mooted question. It may be well to explain that in this figure the plane in which the wheel section is shown has been turned so as to coincide with the plane in which the rail section is projected, the line of intersection of the two planes being the vertical line from the wheel gage. Therefore the relative position of gage of wheel and gage of rail is the same as on a curve, the distance between them being -fe in. in the case of a thirty-five-foot-radius curve. Keferring now to the diagram (Fig. 122), the distance, ac, between the gage FIG. 122. DIAGRAM FOR FINDING THE GAGE LINES. lines of the rails along the axle is easily figured, and for 4 ft. 8 in. gage on a 35-ft.-radius curve is 4 ft. 8f in. Assum- ing that the wheels have been placed on the axles to gage i in. less than the track, as is customary, or 4 ft. 8 in., and adding T \ in. (twice the distance between gages in Fig. 121) we have 4 ft. 8^J in., or T \ in. less than the distance ac. This would indicate that the track should be that much tight gage on a thirty-five-foot-radius curve. On longer-radius curves the angle formed by the axle and a radial line through 74 STREET-RAILWAY ROADBED. one end is less; therefore the wheel-flange has more play, the distance ac approaches the gage, and there is less reason for changing the gage from that used in straight track. The old theory that the curved gage should be increased in., or any amount, should be abandoned, and practically has been. Besides the various forms of " solid " guard-rails shown in Figs. 38, 46, 117, 119, 121, etc., there are several styles of compound guard-rails, the four principal ways of making them being shown by Figs. 123 to 126. The use of these is FIG. 125. FIG. 126. WAYS;*OP MAKING GUARD-RAILS. limited mostly to T-rail construction. That shown in Fig. 123, the one now generally used, was first devised and used by the writer. Its virtue lies in the ease with which it is adapted to a large variety of T-rail sections, and that the groove has a floor which prevents buggy-wheels going down GUARD-RAILS. SPECIAL WORK. 75 too deep. The section Fig. 124 has the same qualities with the added advantage of a larger base and being more sub- stantial generally; but it is more expensive, as the rail which FIG. 127 PLAIN CURVE. FIG. 128. LEFT HAND BRANCH-OFF. i ^J FIG. 129. RIGHT HAND BRANCH- FIG. 130. CONNECTING CURVE OFF. AND CROSSING. FIG. 131. PLAIN Y. FIG. 132. CROSSING. forms the guard has to be cut out on a planer, i.e., machine fitted for each particular case; while the guard on Fig. 123 may be, and is, rolled in large quantities. None of these com- bination guards are as satisfactory as the solid sections, which 76 STREET-RAILWAY ROADBED. L... FIG 183. THREE PART V. 1 FIG. 184. THHEE-PART THROUGH Y. FIG. 135. REVERSE CURVE. FIG. 136. RIGHT HAND CROSS-OVER. FIG. 137. LEFT HAND CROSS-OVER. GUARD-RAILS. SPECIAL WORK. 77 latter, however, are used only on work of six inches and over. There are a few solid guard sections under six inches, but the fishing space is so greatly reduced on the guard side that it is impossible to make a substantial joint, and they are but little used. Before going deeper into the question of special work it has been thought best to illustrate the various simple layouts, giving the names most generally used. See Figs. 125 to 141. The illustrations show single-track work entirely. Double- track layouts have the same names with the prefix " double track," and it is customary in all cases to describe them by the FIG. 138. DIAMOND TURNOUT. FIG. 139. SIDE TURNOUT. FIG. 140. THROWN OVER TURNOUT. initial letters; for instance, S. T. E. H. B. 0., meaning single track, right-hand branch of. The "hand" is always deter- mined by the side to which the curve turns off as shown to a person facing the point of curve. There are many other pecu- liar and complex arrangements of tracks to which no specific names can be given except the general term "special work." Outside of the curves the pieces which go to make up a job of special work are switches, mates, and frogs. A switch is a piece having a movable part to deflect the car, and the one most commonly used on street railways is called a tongue switch. Split switches, Lorenz switches, and stub switches are also used, as on steam roads, but only in open track or car-houses. A switch is automatic when it is so arranged that by means of a piece of rubber or a spring it will 78 STREET-RAILWAY ROADBED. automatically return to its former position after allowing a car to " trail " through. The tongue switch is almost always placed on the inner or short side of the curve, and when placed in the other position it should be designated as an "outside" tongue switch. An outside switch should not be used on curves of less than 150 or 200 ft. radius unless in connection with an inside tongue switch, or in cases where the curve is used much more than the straight track. A mate is a piece used in connection with a tongue switch on the opposite side of the track and has no movable parts. Corresponding to the tongue switch, its position is on the out- kvwj Uv~J FIG. 141. SPRING FROG FOR OPEN TRACK. side of the curve, and when otherwise placed is called an " inside" mate. A frog is the intersection of any two lines of rails. Except in open track, frogs have no movable parts, the endeavor usually being to make them as rigid as possible. In cross- overs which are used only for emergency and at points where the track in one direction is little used the main line or rail of the frog is usually made ' ( unbroken," i.e., with no flangeway for the crossing rail. The wheel is made to climb over the head of the main rail by inclines on each side. In open track the same thing is accomplished in a better way by a "spring frog" in which the main line is practically unbroken and the wheel in taking the side track opens its flangeway by pressing out the spring rail. (See Fig. 141.) Frogs, switches, and mates are the vital parts of special work, and there are many ways of making them. In the days of the light, slow-going horse-car they were made almost GUARD-RAILS. SPECIAL WORK. 79 universally of cast iron. Now that metal is used but very little, except in a manner to be described below. Until within the last two or three years these parts were mostly built of the rails themselves. This construction is used exclusively on steam roads, but there the conditions are altogether differ- ent from street railways. The tracks being exposed, repairs and renewals are easily and. cheaply made. But with the track buried in pavements in busy streets their renewal is a serious and expensive matter, and the longer life there is in a frog the more valuable it becomes. For this reason there is a constant effort to improve on their construction, with the result that now we have three different types of first-class construction, viz., " Manganese," "Guarantee," and "Ada- mantine Steel." As stated above, cast-iron switch-pieces have been aban- doned. They do not possess the necessary durability for even the lightest kind of electric-car traffic; nor is there any suit- able way of joining the several frogs together or to rails. Built work, if well and properly done, answers the purpose very well for roads of moderate traffic. A built frog should last on a road running fifteen-minute headway about ten years; when it is easy to figure that under a minute headway its life would be less than a year. The writer is familiar with built work that has given very much better service than this. Managers should not complain if they are compelled to renew built frogs within a year or two when placed under heavy traffic. Their steam-road brethren have to renew frogs every three or four weeks in places where the number of movements over them do not exceed those in many places on cable or electric roads. Fig. 142 shows the detail of a built frog for a square or "girder crossing." The parts of a built frog are two or three pieces of rail, four angle-plates, or braces (on small angles two of these are replaced by cast-iron chocks), making seven large pieces, all of which are held together with from ten to thirty bolts and rivets. A renew- able tempered steel floor-plate is often used, and this will pro- long the life of the frog two or three times its normal length if care is taken to renew the floor-plate before the points are 80 STREET-RAILWAY ROADBED. too much worn. There has been used a rail section with solid floor for building frogs which adds something to the life of a frog; but since the floor is of the same steel as the rail, it does not offer much resistance to the cutting action of the wheel-flanges, and when worn there is no way of repairing the frog. A very good tongue switch may be built of rails, and if the radius is not too short it will have a longer life than the frogs in the same job. T rails are particularly well adapted to building switches and frogs. FIG. 142. A " BUILT" FROG. The mate is the most difficult piece to build substantially, from the fact that there is necessarily a considerable distance where the wheels have to be carried on their flanges entirely. These rapidly cut into the floor, with the result that the outer edge of the wheel-tread cuts into the head of the main rail at the point where it leaves, and pounds down the point of the mate. This has resulted in the use to some extent of a "tongue mate," or mate with a movable tongue. This tongue mate is not objectionable on a running-off end, but when used facing it is almost imperative that some form of connection be used between the tongue of the mate and the tongue of the switch so they shall work in unison, usually a simple con- necting rod placed in a cast-iron box extending entirely across the track. This] box without ample drainage is liable GUAKD-KAILS. SPECIAL WORK. 81 to become filled with dirt in summer and ice in winter, and requires constant attention a feature which renders its use objectionable. These latter remarks apply with equal force to all kinds of so-called automatic arrangements or other mechanisms which have to be placed underground. On cable or conduit electric roads where the drainage is well provided for, underground machinery is used without special objection, but on surface roads it should not be used without a good sewer connection and frequent inspection. Special work of the first class as made to-day is comprised in the three types known as "Manganese," "Guarantee," and "Adamantine." The first of these is shown in the typical frog Fig 143, which is composed of four short pieces of rail Street Ry .Journal FIG. 143. A "MANGANESE" FROG. held to a center piece, which is a casting of manganese steel, by means of a mass of cast iron at each joint. It will be seen, therefore, that the frog is composed of six separate and distinct pieces. A late modification of this construction has the four small pieces of rail held together by a single mass of cast iron in which is cored a pocket to receive the center cast- ing, which is much reduced in size, and which is held in place by bolts or wedges. " Guarantee" special work, so called from the fact that the manufacturer sold certain guarantees with it, is made in a similar manner to the later form of manganese work, except that the center pieces are made of tempered steel, and are retained in the pocket by means of zinc. It was originally intended that these pieces should be renewable without dis. 82 STREET-RAILWAY ROADBED. turbing the main body of the frog in its bed in the pavement. But the difficulty of making anything like a good fit after the adjacent parts of the frog had become worn has compelled the abandonment of the "renewable" feature in all special work. Guarantee frogs are peculiar also in a very generous use of cast iron. (See Fig. 144.) FIG. 144. A " GUAKANTEE " FROG. In "Adamantine" or "Solid Cast Steel" special work, as its name implies, the switch-pieces are castings of open-hearth steel, and each frog, switch, or mate is one solid piece of steel. FIG. 145. AN "ADAMANTINE" FROG. Large strides in advance have been made in the manufacture of open-hearth steel during the past seven years, and it is now possible to produce castings of the most intricate shape of a fine tough quality of steel. Fig. 145 is a sample, and shows GUARD-RAILS. SPECIAL WORK. 83 three frogs combined in one piece. The points most subject to wear are protected by pieces of a special steel which is at the same time exceedingly hard and tough. These pieces, or "centers," are placed in the moulds and the main body of the frog cast around them. The contraction of the steel in the cooling holds them with a vise-like grip which never releases. Special work of the first class, while expensive in first cost, is by far the most economical in the end, for not only will work of this class outlast built work from two to four times, but it saves the large cost of more frequent renewals and re- pairs. CHAPTER VIII. DISCUSSION OF THE ADVANTAGE OF SPIRAL CURVES, AND TABLES AND FORMULA FOR THEIR USES. WITH the higher speeds and larger cars which the advent of the electric railway brought in street-railway work, the necessity of some sort of easement curve, for the short radii required, very early became evident. The question as to its form has generally been considered as if the paths followed by all parts of the car were necessarily somewhat similar to the alignment of the track. This has led many engineers to believe that a three-centered curve was sufficient for practical purposes, besides being somewhat easier to design and calcu- late. This assumption as to the path of the car is only true as to the small portion of the car which lies between the two axles or the center pins of a double-truck car. The parts of the car outside of this area describe rather peculiar paths, as is shown on the engravings herewith. Fig. 146 is a case which the writer has seen repeated many times in the last few years. The curve is primarily designed to enable cars to pass each other on a double-track curve, but is also supposed to act as an easement. It will be noticed that there are four changes in the direction of rotation of the point A y and also four abrupt changes in the rate of rotation of the point B. It should be understood that the radii given for the short arcs in these paths are only given in order to convey an idea of the sharpness of curvature at these points, as these arcs would not be exactly circular, although nearly so. These changes of direction in the path of A will occur at any P. C. C. of a three-centered curve, if the first radius is suffi- ciently long to be of any use in preventing the jar of striking the P. C. of the curve. The sudden changes in the rate of 84 ADVANTAGE OF SPIRAL CURVES. 85 rotation of the point B will be found to occur at every P. C. C. of a compound curve unless it is of the form of the railway spiral with chords of a length shorter than the wheel base of the car. Fig. 147 shows exactly the same main curve shown in Fig. 146, but connected with the tangent by a spiral of seven chords having the following radii : 300, 150, 100, 75, 60, 50, and 40. The "spiral" is somewhat forced on the last radius in order to keep the center of the curve in exactly the same position while at .the same time maintaining the same car 86 STREET-RAILWAY ROADBED. clearance. The track curve does not vary in position at the center of the spiral more than 4 in. from that shown in Fig. 146. Note that there are but two changes in the direction of rotation of the point A. One of these is at the point where A leaves the tangent, from which there is a gradually de- creasing curvature until it reaches the point of reverse curva- ture. From this point the curvature increases gradually until the point A attains its maximum rate of rotation parallel to the main curve. The rate of decrease and increase of curva- ture for both these arcs is somewhat less than that for the track curve. The point B also describes a curve with gradu- ally increasing curvature until it reaches its maximum rate of rotation and follows the same path as A. The rate of increase of curvature would be somewhat greater than that of the track curve. The car shown in these figures is by no means an extremely long car, as there are other cars known to the writer which are 30 ft. long, and others proposed which will be 38 ft. long and with only 6 ft. 6 in. wheel base. The overhang of a double- truck car is nearly always as much as that shown in the figures. The sudden changes of motion shown in Fig. 146 exert a severe racking strain on the car framing which is particularly destructive to open cars, as the connection between roof and floor framing is necessarily weak. The effect on the passengers is not pleasant, although this does not appeal so directly to the treasury of the railway. The plan shown in Fig. 146 would cost for material about $12 more than Fig. 146, when caedit is given for the straight track replaced by the extra- curve of the spiral, and would cost that much only for the most expensive track in use for surface roads. The use of spirals has been largely delayed by the absence of any tables and formulae for spirals suitable for street-rail- way speeds and curvature. Those presented herewith have been in use for some two years or more, and are the final results of a number of other forms which have been tried through a period of about five years. This form has given universal satisfaction, and it is believed that the tables present sufficient variety for most cases arising ADVANTAGE OF SPIRAL CURVES. 87 in usual practice. Table I gives a choice of 50 spirals for curves of radii from 30 ft. to 1785 ft. In case a special spiral is desired for any reason, these may be modified by multiply- ing by a suitable factor, in the same way that spirals Nos. 1 and 2 have been obtained from spirals Nos. 4 and 5. Spiral No. 3 was designed as the spiral of shortest length to enable a certain double-truck car to pass on a double-track curve. While the car used for this purpose was one of the largest double-truck cars in use on surface roads in the East, this should be carefully tried with the car in use on the road in question, as indeed it must necessarily be in order to select a radius for the outer curve. Substantially the same chord length was carried throughout the system with a view to ease in giving the necessary information to the track-layers. At switches it is mechanically undesirable to have a greater initial radius than 100 ft. This prevents the use of a theoretical spiral at such points, but, as shown in the first part of this chapter, some easement is necessary between the long- radius switch and the main curve of shorter radius. The most desirable form for this easement appears to be that por- tion of a theoretical spiral which lies between the switch radius and the radius of the main curve. Table II was prepared in this way, and the choice of easements was made in such a manner as to produce the standardization of the crossing frogs. The latter is very desirable for such constructions as require the making of special patterns for frogs. This table gives a choice of forty-eight easements for twelve different radii from 30 ft. to 70 ft. radius of center line. For radii greater than this no easement is necessary between the switch radius of 100 ft. and the main curve. The most satisfactory way of plotting these curves is to cut out thin celluloid templates of each one which is to be used Mark each P. C. C. of the spiral on the curved edge and the direction of a radial line through this point. Fig. 148. Problems 3 and 4 can generally be solved accurately enough on a drawing made to scale and in lesa time than by calcula- tion for "special" curves, as a plan must be made in any 88 STREET-RAILWAY ROADBED. event in order to have the work made by the manufacturer. If only a few curves are to be drawn up, however, it may be more expeditious to figure the curve and plot the points on the spiral by the x and y taken from the table, and connect them with a variable or " French " curve. riu. 1*0. TEMPLATE FOR LAYING OUT CURVES. The problems given in this chapter contain all the prin- ciples necessary for laying out any curve with symmetrical spirals. The most usual form 'of unsymmetrical curve is shown in Problem 9, and the application of the latter will enable any such curve to be solved. It is by no means necessary to lay out on the ground every point on the spiral. If the curve is "special work" and curved by the manufacturer, the point of spiral, a point about in the center of the spiral, and the junction of spiral with the main curve should be laid out. Points for the latter should be laid out from 20 to 30 ft. apart, depending on the radius of the curve and the location of the joints. If the curve is to be "sprung in" by the track-layers, every alternate point on the spiral should be laid out and the track-layer furnished with sufficient middle ordinates for 10-ft. chords. These can be obtained by Problems 10 and 11 and Table III., p. 132. The most expeditious way to lay out a spiral curve, if the final x does not exceed two feet, is to set the transit on the intersection point and lay off the tangent distances, then lay out sufficient points on the spirals by the successive offsets x and the long chords. Then bisect the included angle and lay out a temporary point V, Problem 8. Move the transit to the last point on the spiral, set the vernier to a back reading equal to the spiral angle, set the telescope on an offset from V equal to x inside the intersection point. Deflect to 0, and the line of the telescope should ADVANTAGE OF SPIRAL CURVES. 89 strike V and the distance V'L = R tan (\A S). This will check the preceding calculations and field-work. The circular arc can then be laid out in the usual manner. PROBLEM 1. To select a spiral. (a) The radius of the main curve must be less than the pre- ceding branch of the spiral, must be more than the next branch would be were it produced, Tand should nearly equal the latter. (b) The longer the spiral, the easier the entrance will be. But bear in mind that the main body of the curve should be circular, the spiral simply acting as an entrance to it. (c) A spiral of less than three branches should not be used. PROBLEM 2. PROBLEM 2. Given: A circular curve with symmetrical spirals, to find the tangent and external distances. OG = R + x - ver sine 8R ; GS = y - sine SJK ; 90 STREET-RAILWAY ROADBED. Tangent distance = OG tan %A -f GS; External distance = OG ex sec | A + a; ver sine PROBLEM 3. PROBLEM 3. Given: The tangent distance VS, the inter- section angle A, and the desired length of [spiral, to find the radius of the curve. Approximate R = cotangent %A( VS | length of spiral). Having selected a spiral by this radius, the exact radius may be found, if required, by the following formula: _ cos \A( VS y x tan sine (%A 8). Caution. If the result is enough different from the original radius to require a change in the spiral by Problem 1, a second trial must be made. This rule does not apply for approximate radius to the easements in Table II. PROBLEM 4. Given: The intersection angle A and the external distance VH, to find the radius. Approximate to the radius by finding that for a simple ADVANTAGE OF SPIRAL CURVES. 91 curve passing through the point H, and select a spiral for a radius somewhat smaller. VH cos \ A x /0 , . Then R = -5 ~^~r (Searle.) cos S cos - A Caution. If the result is enough different from the original radius to require a change in the spiral by Problem 1, a second trial must be made. \/ y PROBLEM 5. Given : The x and y for any point on the spiral, to find the deflection from the tangent at the point of spiral. 9* Tangent deflection angle = -. PROBLEM 6. Given: The x and y for any point on the spiral, to find the long chord. (a) Long chord = - : - ~z - r , cosine def. angle or (b) Long chord = Vx* -\- if. PROBLEM 7. Given: x and y for a point on the spiral, to find x' and y' on a line parallel to the spiral, and offsetted the distance SS f inside the spiral. x' = x SS' ver sine S ; y' = y - SS' sine S. STREET-RAILWAY ROADBED. Note. Problems 5 and 6 can then be applied to x f and y' if it is desired to use deflection angles to lay out the curve. t- y PROBLEM 7. As these curves will almost invariably be laid out on an offset varying with the gauge of the road, the deflections are not figured in the table. k PROBLEM 8. ADVANTAGE OF SPIRAL CURVES. 93 PROBLEM 8. Given a circular curve with spirals, to find the distance VV, in order to lay out a tangent to the circular curve, from which the latter may be laid out in the usual manner. VH see Problem 2; VH = R ex secant VV = VH- VH. PROBLEM 9. General solution for unsymmetrical curves OQ = R + x - R ver sine S; OS = y - R sine #; OG' = R + as* - R ver sine # 0/ ; sine A 00' - 00 tan A 94 STREET-RAILWAY ROADBED. Note. in above; -f if A is more than 90, and if A is less than 90. F'Z or V'U = tan WR; VB =^-f F sine ; FZ? = VS (y -\-VL VL ver sine tan C = FF' = F5 ' cosine C S> i(<7 = 46 +- PKOBLEM 10. PROBLEM 10. Given : The middle ordinate for a chord of length AB for R and R', to find the middle ordinate at the P. C. C. From the figure it is evident that D'C' bisects CD. CF+ DF 2 .'. EF = Therefore the middle ordinate at any P. 0. C. in the spiral equals one-half the sum of the middle ordinates for the radii vn each side for the same chord. Note. See remark following Problem 11. ADVANTAGE OP SPIRAL CURVES. 95 _ / i \ / t \ A PROBLEM 11. PKOBLEM 11. Given : That portion of a spiral with equal chords L, L', and L" and angles a b, a, and a -f b, to find the middle ordinate at the center of the chord L in the length D'C'. OF = C'A and DF = D'B. From the figure it is evident that D'C' bisects CD. CF+DF ..EF-- -- . Then C'A = %L tan i + L' sine Ua + D'B = iZ tan Ja + ^" sine ia + ^ and since the sines of small angles are proportional to the angles, But this last equation equals the middle ordinate in the length AB for the radius of the central arc; and since the increment to the angle ~b would be equal if L' and L" were equal, the middle ordinate at the centre of any arc of the spiral, for any length of chord, is equal to the middle ordinate of the radius of that arc in the same length. 96 STREET-KAILWAY ROADBED. Remark. It will be noticed that the solutions of Problems 10 and 11 are slightly inaccurate in not allowing for the increase of length of D'C' over AB, nor for the inclination of the middle ordinate found to the true middle ordinate. Problem 11, if applied to the curving of rails, as intended, also assumes that each rail forms a spiral, whereas they are simply lines parallel to one. Both solutions, however, are sufficiently accurate for any spiral of 5 -ft. chords or greater. Taken in connectio with Table III, the middle ordinates can be easily found for points on the rails from 2^ to 3 ft. apart. The rails being curved to these, exactly what is de- sired will be obtained, i.e., a curve of constantly changing radius, which is only considered a compound curve for the purpose of calculation. These spirals are figured for a length suitable to the ordi- nary * street-railroad speeds. For greater speeds the writer would advise the use of the regular railroad spirals. Tables in convenient form for these have been published by William H. Searle, C.E., and others. TABLE I. SPIKALS. 97 TABLE I. SPIRALS. SPIRAL No. 1. Rad. Angle. X. y- O Ver. Sine. Sine. 210 42' 0.015 2.565 42' .00007 .01222 105 1 24 0.078 5.130 2 06 .00067 .03664 70 2 6 0.219 7.692 4 12 .00269 .07324 52i 2 48 0.469 10.245 7 .00745 .12187 42 3 30 0.860 12.780 10 30 .01675 .18224 35 4 12 1.420 15.283 14 42 .03273 .25376 30 SPIRAL No. 2. Rad. Angle. X. y- so. Ver. Sine. Sine. 300 30' 0.011 2.618 30' .00004 .00873 150 1 0.057 5.235 1 30 .00034 .02618 100 1 30 0.160 7:851 3 .00137 .05234 75 2 0.342 10.463 5 .00381 .08716 60 2 30 0.627 13.065 7 30 ,00856 .13053 50 3 1.036 15.651 10 30 .01675 .18224 42| 3 30 1.587 18.187 14 .02970 .24192 33i 4 2.309 20.703 18 .04894 .30902 SPIRAL No. 3. Rad. Angle. X. y- s. Ver. Sine. Sine. 300 1 0' 0.046 5.236 1 0' .00015 .01745 150 2 0.229 10.468 3 .00137 .05234 100 3 0.639 15.688 6 .00548 .10453 75 4 1.368 20.871 10 .01519 .17365 60 5 2.501 25.982 15 .03407 .25882 50 6 4.118 30.959 21 .06642 .35837 40 7 6.143 35.403 28 .11705 .46947 35 SPIRAL No. 4. Rad. Angle. X. y- s. Ver. Sine. Sine. 420 42' 0.031 5.131 42' .00007 .01222 210 1 24 0.157 10.261 2 6 .00067 .03664 140 2 6 0.439 15.384 4 12 .00269 .07324 105 2 48 0.939 20.490 7 .00745 .12187 84 3 30 1.720 25.561 10 30 .01675 .18224 70 4 12 2.839 30.567 14 42 .03273 .25376 60 4 54 4.352 35.469 19 36 .05794 .33545 52| 98 STREET-RAILWAY ROADBED. TABLE I. SPIRALS (Continued). SPIRAL No. 5. 'Rad. Angle. X, y- s. Ver. Sine. Sine. 600 30' 0.023 5.236 30' .00004 .00873 300 1 0.114 10.471 1 30 .00034 .02618 200 1 30 0.320 15.703 3 .00137 .05234 150 2 0.685 20.926 5 .00381 .08716 120 2 30 1.255 26.130 7 30 .00856 .13053 100 3 2.073 31.302 10 30 .01675 .18224 85 3 30 3.175 36.374 14 .02970 .24192 75 SPIRAL No. 6. Rad. Angle. X. y- s. Ver. Sine. Sine. 900 20' 0.015 5.236 20' 00002 .00582 450 40 0.076 10.472 1 .00015 .01745 300 1 0.213 15.706 2 00061 .03490 225 1 20 0.457 20.936 3 20 .00169 .05814 180 1 '40 0.837 26.158 5 .00381 .08716 150 2 1.385 31.365 7 .00745 .12187 128 2 20 2.125 36.524 9 20 .01324 .16218 112* SPIRAL No. 7. Rad. Angle. X. y- s. Ver. Sine. Sine. 1260 15' 0.012 5.498 15' .00001 .00436 630 30 0.060 10.995 45 .00009 .01309 420 45 0.168 16.492 1 30 .00034 .02618 315 1 0.360 21.987 2 30 .00095 .04362 252 1 15 0.660 27.475 3 45 .00214 .06540 210 1 30 1.091 32.957 5 15 .00420 .09150 180 1 45 1.678 38.424 7 .00745 .12187 157 SPIRAL No. 8. Rad. Angle. X. y- s. Ver. Sine. Sine. 1890 10' 0.008 5.498 O r 10' .00000 .00291 945 20 0.040 10.996 30 .00004 .00873 630 30 0.112 16.493 1 .00015 .01745 472| 40 0.241 21.990 1 40 .00042 .02908 378 50 0.441 27.483 2 30 .00095 .04362 315 1 0.729 32.973 3 30 .00187 .06105 270 1 10 1.120 38.457 4 40 .00o32 .08136 236 TABLE I. SPIRALS. 99 TABLE I. SPIRALS (Continued). SPIRAL No. 9. Rad. Angle. X. y- s. Ver. Sine. Sine. 2730 o r 0.006 5.559 o r .00000 .00204 1365 14 0.028 11.118 21 .00002 .00611 910 21 0.079 16677 42 .00007 .01222 682 28 0.170 22 234 1 10 .00021 .02036 546 35 0.311 27.791 1 45 .00047 .03054 455 42 0.515 33.346 2 27 .00091 .04275 390 49 0.792 38.899 3 16 .00162 .05698 341 SPIRAL No. 10. Rad. Angle. X. y- s. Ver. Sine. Sine. 3780 5' 0.004 5.498 5' .00000 .00145 1890 10 0020 10.996 15 .00001 .00436 1260 15 0.056 16.493 30 .00004 .00873 945 20 0.120 21.991 50 .00011 .01454 7:>6 25 0.220 27.488 1 15 .00024 .02181 630 30 0.364 32.983 1 45 .00047 .03054 540 35 0.560 38.478 2 20 .00083 .04071 472 SPIRAL No. 11. Rad. Angle. X. y- 8. Ver. Sine. Sine. 5250 4' .0035 6.109 4' .00000 .00116 2625 8 .0178 12.217 12 .00001 .00349 1750 12 .0498 18.326 24 .00002 .00698 13121 16 .1066 24.434 40 .00007 .01164 1050 20 .1955 30.542 1 .00015 .01745 875 24 .3234 36.649 1 24 .00030 .02443 750 28 .4975 42.756 1 52 .00053 .03257 656 SPIRAL No. 12. Rad. Angle. X. y- S. Ver. Sine. Sine. 7140 3' .0027 6.231 3' .00000 .00087 3570 6 .0136 12.462 9 .00000 .00262 2380 9 .0381 18692 18 .00001 .00524 1785 12 .0816 24.923 30 .00004 .00873 1428 15 .1495 31.153 45 .00009 .01309 1190 18 .2474 37.384 1 3 .00017 .01832 1020 21 .3806 43.613 1 24 .00030 .02443 892 100 STREET-RAILWAY ROADBED. TABLE II. Explanation. These easements were designed with a view to combining a very easy-running curve with expedition and economy in the manufacture of the frogs. The latter is effected by making the four crossing frogs ex- actly alike for different distances between track centers within certain limits. This is done by adopting a standard distance FIG. 149. (Fig. 149) and varying the easement to meet this con- dition. For the outer curve of a D. T. B. use that easement which corresponds to the distance between track centers (7. If the required distance is not included in the table a new easement must be figured, or the distance of track centers must be changed in the special work to agree with the nearest one given. For the inner curve of a S. T. B. 0. any ease- ment may be used which corresponds to the center radius de- sired. TABLE II. 101 TABLE II. Continued. CENTER RADIUS 30' 0". Rad Angle. X. V. 8, c j 9' ou'< C = \ 9' gj" TF. z. 50 30 10 0' 0" 0.760 8.682 10 0' 0" 3.473 0.304 Rad. Angle. X. y- 8. *=M''' w. z. 50 37i 30 10 0' 0" 5 43 20 0.760 1.593 8.682 12.332 10 0' 0" 15 43 20 4.203 0.470 Rad. Angle. X. y- s. J 9' 4^" ' ~ 1 10' 0^" W. z. 75 45 30 7 50' 0" 4 4145 0.700 1.351 10.222 13.851 7 50' 0" 12 3145 7.343 0.637 Rad. Angle. X. y- s. r j 9' 6H" C - 1 10' 2^" TF. z. 75 50 37| 30 7 50' 3 48 5 0.700 1.260 2.059 10.222 13.490 16.662 7 50' 11 38 16 38 8.075 0.804 CENTER RADIUS 32' 6". Rad. Angle. X. y- s. HW w. z. 50 32| 10 0' 0" 0.760 8.682 10 0' 0" 3.039 0.266 Rad. Angle. X. y- ' W w. z. 102i 82i 66 55 6 30' 3 2 3 49 0.658 1.267 2.139 11.584 15.909 20.217 6 30 9 32 13 21 7.518 0.653 Rad. Angle. X. y- s. c=j 9' 4^" 10' OJ$" w. Z. 102i 96 77 64 W 55 6 30' 2 13 2 46 3 20 0.658 1.149 1.801 2.648 11.584 15.265 18926 22.551 6 30' 8 43 11 29 14 49 8486 0.820 Rad. Angle. X. y- s. c=\ 9' 6fc" 10' 2^" w. Z. 102i 96 77 64 55 6 30' 2 44 3 25 4 8 0.658 1.285 2.156 3.328 11.584 16.121 20.628 25.093 6 30' 9 14 12 39 16 47 9211 0.986 108 STREET-RAILWAY ROADBED. TABLE II. Continued, CENTER RADIUS 60' 0". Rad. Angle. X. y. s. r- -I 9 c - is '0^" '%W w. z. 102* 84 70 60 6 30' 1 45 2 6 0.658 0.987 1.402 11.584 14.129 16.660 6 30' 8 15 10 21 5.881 0.425 Rad. Angle. X. V' s*. c- -I s c ~ \\ ' 2W ' W W. z. 1021 84 70 60 6 30' 3 10 3 49 0.658 1.310 2.246 11.584 16.180 20.748 6 30' 9 40 13 29 6.758 0.592 Rad. Angle. X. y. s. o.\, 9' W 0' OH" W. z. 102 96 80 68^ 60 6 30' 2 21 2 50 3 18 0.658 1.184 1.889 2.798 11.584 15.486 19.379 23.217 6 30' 8 51 11 41 14 59 7.705 0.759 Rad. Angle. X. y. 8. 1 V'W ~ \ 10' 0}2" w. z. 102* 93} 84 76 70 6 30' 3 9 3 30 3 47 0.658 1.379 2.393 3.695 11.584 16.664 21.693 26.539 6 30' 9 39 13 9 16 56 6.151 0.660 Rad. Angle. X. y- 8. c J 9' 6^" - 1 KX SB" w. z. 102J 95 85* 77i 70 6 SO 7 3 36 4 4 29 0.658 1.519 2.770 4.476 11.584 17.490 23.325 29.143 6 30' 10 6 14 6 18 35 6.835 0.826 CHAPTEE IX. DESIGN OF SPECIAL WORK. WHILE the frog and switch work, as previously remarked, is necessarily u special work" in its truest sense, there is a larger opportunity for standardization than is usually taken advan- tage of. Every road should insist that at least the switches and mates from the same manufacturer, of the same " hand " and height of rail, shall be strictly interchangeable. It may be necessary in some cases to use a switch of shorter radius than the standard, but there is no excuse for the medley of switches and mates which are often found even on the smaller roads. It very often happens that part of a piece of work has to be thrown out because badly worn on one track only. If it were interchangeable with a piece in another location which was worn on the other track, both could be used for several years longer. As conditions generally prevail, both go to the scrap- heap. There is more difficulty with frogs than with switches. The customer should insist that only one angle and length of frog be used for crossovers and turnouts. Even this is often not done. Probably 90 per cent of the frogs required for any road are included in a double-track branch-off from straight track. As an illustration of the possibilities of standardization in this direction, it might be mentioned that special work was re- cently designed for an Eastern road which required 162 frogs in 24 different layouts, covering nearly every case which can occur in either single- or double-track work. The latter was of two distances of track centers. Of these 162 frogs, 102 111 112 STREET-RAILWAY ROADBED. were exact duplicates of others, making only 60 different frogs. There need not be more than three different sets of branch-off frogs of each hand to fill the needs of the most ex- tensive system for such frogs. In Table II of the preceding chapter is given a set of switch easements which enables us to take care of different distances of track centers, if such occur, while at the same time insuring the easiest possible entrances to the main curve. If special work be required for a number of different places, the following method is a convenient one to follow. First look over the system and determine the location and radius of the minimum radius curve. Let us assume for the purpose of illustration that this would be one of about 40 ft. if the curve is properly spiralized. We will then assume a radius somewhat larger, say 45 ft., for the radius of the inner curve and such an easement taken from Table II as appears suitable to us. With the usual sizes of cars and distances of track centers car-clearance will be obtained if the outer curve has a radius five feet greater than that of the inner, if the latter be properly compounded at the ends. This fixes the radius of the outer curve at 50 ft., with the easement corre- sponding to the distance of track centers. This completes our first standard double-track branch-off up to the line AA, Fig. 150. It will be noticed that this one layout gives us two standard frogs, No. 13 and No. 19, for use in such pieces as shown in Figs. 128, 129, 130, 133, 134, if these should occur on the road in question. Having established such a standard, make a tracing on cloth and with AA as a P. C. C. line draw the center lines of a number of curves, say from 35 ft. to 55 ft. for the inner curve and from 40 ft. to 60 ft. for the outer curve. Having then a plan of the location where a branch-off is required, showing all obstructions which are to be cleared, put on the center lines of the tracks to be connected. Also sketch, roughly, the position in which it seems possible to lay the center line of the curve. Then place the tracing on top of the plan and make the straight track on the tracing coincide with that on DESIGK OF SPECIAL WORK. 113 the plan. Move the tracing along the straight line until some one of the center lines on the tracing appears to coincide as closely as seems possible with the location desired for the inner curve. Prick through the center for this curve and the centers and P. C. C.'s for the inner curve of the standard branch-off. Proceed in the same way for the outer curve and draw them both in on the paper plan. Finish the free ends 114 STREET-RAILWAY ROADBED. of the curves with spirals, and the final plan will be some- what as shown in Fig. 151. FIG. 151. DIAGRAM SHOWING FINAL PLAN OF CURVE. Checking the car-clearance with the celluloid template described in Chapter IV completes the operation. If all the streets were of about the same width and intersected at about the same angle it would not be necessary to have more than this one standard. It may be necessary to establish one or two more. If the location of trolley wire be desired, it can be easily found by the use of a template of the essential elements of the problem, i.e., the wheel base and the horizontal projection of the trolley pole. Having a plan of the curve, for which the location of trolley- DESIGN OF SPECIAL WORK. 115 wire is desired, to a scale of five feet to the inch or larger, take a piece of thin transparent celluloid and cut it to the length of the wheel base (on the same scale as the plan), and mark with a sharp point a line for the center line of the car, and mark the center of the wheel base. Then cut another piece a little longer than the horizontal projection of the trolley- pole. Mark a line on this at right angles to one end, and mark the length of the horizontal projection of the trolley- pole from this end. Then join the two pieces at the center of the wheel base and center of the trolley base by an eyelet paper binder, loosely enough so that they may turn with some little friction. Then, placing the template upon the plan of the curve so that the wheel base coincides with the center line of the curve, and swinging the "trolley-pole" until the square end is radial to the track curve, we can mark a point which will be approx- imately on the wire curve. Carrying this process through the spiral to the point where the offset becomes constant, we next sketch in the approximate location through the points just found. Going over this again, and making the "trolley- pole " square with radial lines from this approximate location instead of those of the track curve, we can lay down the final location of the wire for cars to run in one direction. If cars are to run in both directions, the location should be found by taking an average of the curves located by the aid of the tem- plate. In the instance shown in Fig. 152 there is from six to eight inches difference in the offsets for cars going in opposite directions. Having plotted the wire curve on the plan, of course the offsets will be taken off by scale and used as cir- cumstances may require. This method can also be used for locating frogs over complicated switch-work, in finding the proper position for the trolley wire in a car-house door, and in other ways which will suggest themselves to the construct- ing engineer. The time taken for the operation, having the plan and template in hand, should be trivial. If the outer rail of the curve be elevated, the trolley-wire 116 STREET-RAILWAY ROADBED. should be set in towards the center of the curve an additional amount equal to Elevation X height of trolley-wire above rail Gage In designing a car-house layout these standard frogs and possibly standard switches cannot be used if more than one entering track is to be provided. There should be as far as UJNlVJiitbJ.TY DESIGK OF SPECIAL WOEK. 117 possible an entering curve for each track inside the house, in order that the cars may be quickly gotten out in case of fire. This would involve a large number of frogs and switches in the main-line track if the curves were to start directly from the latter. It is also desirable to avoid facing switches, while the conditions sometimes require that the curves should leave the main line in such a way as to make the switches face the direction of traffic. The best plan in this case is to have the curves start from a gauntlet track offset about six inches from the main line. This involves only one facing switch in the main line instead of one for each curve, and removes the latter switches and mates from all the wear due to the main-line traffic, thus insuring them a much longer life. If the frogs are made, as they should be, " jump-over " or " unbroken main line," the main line will receive very little wear from the car-house traffic, and the continuity of the rail will not be broken by the throats of the frogs. This plan is shown in Fig. 153. Another plan, which is used in case there is in front of the car-house plenty of room which is not needed for car storage, is to put in a ladder or series of them. This is a simpler and cheaper plan, but is only available in the case of wide streets or very cheap real estate. An illustration is shown in Fig. 154. It is impossible to take up every case which may occur in the design of special work, but the most usual cases have been covered, and the same principles may be applied to nearly any combination which is desired. A word more might be added as to the desirability of avoid- ing facing switches whenever possible. They are especially dangerous in positions in which they are rarely used in one direction, thus leading the motorman to be more careless in approaching them, since they are nearly always set in the di- rection desired. A left-hand crossover is especially dangerous and should be prohibited as a general rule. The position pro- duced by a misplaced switch is shown in Fig. 155. This acci- dent occurred twice within a few days on a road of which the writer had personal knowledge, in one case involving loss of life. The first was with electric cars, and the switch was cov- ered with water and was said to have been opened and left by 118 STREET- RAILWAY ROADBED. FIG. 153. CAR HOUSE CURVES LEADING FROM A GAUNTLET TRACK. DESIGN OF SPECIAL WORK. 119 FIG. 154. CAB HOUSE CURVES, USED WHERE LARGE SPACE is AVAILABLE. 120 STREET-RAILWAY ROADBED. the crew of a snow-plow. The second was with horse-cars and seemed to have been pure carelessness, as the switch was in plain sight. FIG. 155. POSITION OF CARS ON A MISPLACED SWITCH. The position which may occur at a left-hand branch-off is shown in Fig. 156. It is rarely possible to avoid the use of FIG. 156. POSITION OP CARS ON LEFT-HAND BRANCH-OFF. the left-hand branch-off, although it may sometimes be done by the method shown in Fig. 157. This problem generally has to be left for the operating department to handle by always requiring the employes to shift and block the switch for the main line after using the branch-off. In fact this should always be the rule for any facing switch, but as far as the writer has knowledge it is rarely strictly enforced. The requirement that the switch should be blocked is essential, as a wagon may shift a switch enough to trip the wheels, thus DESIGK OF SPECIAL WORK. 121 fully throwing the tongue. Most makes of switches may be easily blocked with a small piece of iron wedged between the FIG. 157. METHOD OF AVOIDING A LEFT- HAND BRANCH- OFF. guard and movable tongue. This crude device is better than an elaborate lever arrangement, which is sure to be clogged with mud or ice. CHAPTER X. SURVEYS AND LAYING OUT THE WORK. THE surveys required for an electric railroad will depend altogether upon the class of road to which it belongs. For the interurban road, owning its own right of way and up- proaching in construction a steam railroad, the surveys and track-engineering features will be so similar to the latter that there is little more to be said than is already covered by some scores of field-books. For laying tracks in city streets it is desirable to obtain a knowledge of all subterranean structures which may lie within some little distance of the bottom of the track structure, in order that their safety and future existence may be provided for. Sometimes this information may be obtained from the city authorities, but more often the first knowledge of hidden pipes is obtained when the ground is opened up for track- laying. For the location of turnouts (if a single-track road) the en- gineer requires some knowledge of the location of grades which are heavy enough to limit the speed of the car, and in order to locate the power house and design the feeder system it is also necessary to have some knowledge of the configuration and grades of the system. In the absence of public maps which will give all this infor- mation, the quickest and cheapest way to proceed is to run a random transit line through the streets on which the road is to be laid. If the location of the track in the street is already fixed by ordinance or otherwise, this line may be the tangents for the final track, but it is not worth while to attempt to do anything towards laying out the shorter-radius curves in the 122 SURVEYS AXD LATINO OUT THE WORK. 123 field. From this line locate all sewer manholes, electric-con- duit manholes, fire-plugs, and any other obstructions which may require to be moved or which should be cleared. Also locate the curbing and note the location of any steam or for- eign electric railway crossings. At locations which require special work make a careful surrey of all obstructions in the street, the shape of the curb corners, and also any telegraph- poles, awning-posts, and so forth which may require to be moved in order to avoid the danger of catching a person be- tween them and the car or of striking passengers on the car. Finally take sufficient levels over the line to establish the gradet* This completes the preliminary field-work, and sufficient in- formation has now been obtained to enable the engineer to make a small scale map for the purposes previously men- tioned. After settling the question of the location of the straight track with a view to avoiding as many obstructions as possible consistent with a desirable line and without departing from the legal location, the next question to be settled is that of the special work required. This is preferably designed on a plan of fairly large scale, say of one inch = five feet. From the survey plot the corner at which the curve or branch-off is required. Also lay down the location of the straight tracks which it is desired to con- nect. Then proceed as detailed in the preceding chapter. It will be noted that all questions of clearing obstructions with track or cars can be settled by scale on a plan of this size without recourse to calculation. For crossings it is desirable to run the final line and measure the angle of intersection, distances between tracks and gauges, on the ground. The section and drilling of the connecting rail should also be ascertained. If a foreign road, it should be ascertained by correspondence with the retjwnsible official whether the crossing is to be built to conform to the existing tracks, or if the track is to be relaid to some different align- ment or change in track centers; also if it is to be relaid shortly with a different section of rail which may affect that 124 STREET-RAILWAY ROADBED. of which the crossing is to be built. The style of crossing for a steam railroad is often changed to agree with the beliefs or fancies of the particular official in charge of the section of track crossed, and it is well to secure a written assent to the style of construction before ordering the material. It may be said here that it is the practice for the various supply companies to furnish plans for all special work re- quired. This practice has much to be said both for and against it. It is almost impossible to compare prices for plans of special work drawn by different engineers, as the amount of material can be made to vary quite largely. This is especially true of work drawn up with spiralized curves, as all should be; and as a general rule the curve which contains the most track will be the easier riding one, while the differ- ence of cost will appear to be much greater than it really is. Thus take two curves of forty-foot-center radius. Let one have spirals on both ends twenty feet long, and the other spirals thirty feet long. The difference in price between the two curves will be about fifteen dollars, but the second one will displace about ten feet more straight track, leaving the real difference about seven and a half dollars for the longer spiral. This is well worth the money, but the difference is enough to insure the purchase of the curve with the shorter spiral. If the road is of some length, an engineer will be required to look after the many questions which require his skill, and he should be competent to design the track struc- ture and special work. For the smaller roads there is no question but that they receive better service from the men engaged exclusively in this work by the supply companies than they could afford to engage for themselves. After the plans are finally settled upon for the track, the points should be given for line and grade. Points should be given about fifty feet apart on straight track, and close enough together on curves to insure having at least two points opposite each rail. It is best on paved streets to run this line on an offset outside the outer rail rather than attempt to run the center line. An offset should be chosen SURVEYS AND LAYING OUT THE WORK. 125 large enough to put the points out of danger from being disturbed by the work. At special work a point should be set opposite to the heel and toe of a switch, and enough points around the curve to establish the position of each joint of the adjacent rail. If spiral curves are to be used, a point should be set at the beginning, middle, and end of the spiral portion of the curve. All this should be done, if time permits, in advance of the track-laying gangs. It is a waste of time and energy, which could be better employed, to attempt to set up a transit, or even lay out the work by tape measurements, while the track gang is laying the track. If the engineer is to supervise the work, as he should, he can do so much better while not attempting to carry on sur- veying operations at the same time. If the actual laying out is carried on by a transitman who is not expected to look after construction, he can do much faster and better work by being in advance of the track gang. After the road is completed, a skeleton map should be prepared for the operating department, showing all turnouts, crossovers, and connections, with the distance between them. The latter need not be drawn to scale in order to reduce the plan to some reasonable dimensions. A convenient map, in addition to this, is a small-scale one showing a complete plan of the city or cities, with different rails and dates of laying, shown by different character or color of the lines. Each piece of special work is numbered, and a table, is prepared and placed on plan showing date laid, drawing number, manufacturer's order, and drawing number and class of construction. A column is left for notes as to repairs and renewals on each piece, the whole making a very neat and convenient record of the track structure. CHAPTER XL SPECIFICATIONS. IT is obviously impossible to give specifications which will meet all the conditions found in practice. Those presented are intended to secure, under the usual conditions of roadbed and city paving, a thoroughly first-class track. The construction could be improved by the substitu- tion of a concrete foundation and the use of steel ties. Some of the patented joints, welded or otherwise, would doubtless be considered desirable by many engineers, but in the pres- ence of their vast variety one hesitates to express a personal preference for other than the usual splice-plate, especially for the present purpose. Tie-plates might be added; but as the movement of the rail on the tie is so limited by the paving, their value is much reduced below that for open track. Tie-rods might be sub- stituted for the brace tie-plates, but the arguments in favor of the latter seem so conclusive to the writers that no hesitation is felt in recommending them. The varieties of paving which may be required by the city authorities are almost infinite. For asphalt paving with concrete base the use of steel ties would be recommended. The use of wood in such a position as for railway ties in a buried track is only to be defended on the score of economy, and the cost of wooden and steel ties are fast approaching each other in many localities. The desire to avoid the use of wood as far as possible led to the use of the concrete filling for the recesses of the rail. For this might be substituted specially moulded brick, or second quality or second-hand brick with nearly as good results. 126 SPECIFICATIONS. 127 SPECIFICATIONS FOR STREET-RAILWAY TRACK. Construction. Nine-inch girder rail on wooden ties, broken- stone ballast, and granite-block pavement. 1. WORK TO BE DONE. The work to be done consists of the construction of a single-track railway track on , , and streets, between Street and Street in the City of , State of 2. TOOLS AND LABOR. The contractor is to furnish all necessary tools, apparatus, and other means of construction, and do all the work required for the above construction. 3. MATERIAL. The company will furnish and deliver to the contractor, at its yard located on , Street, all material required for the above construction except such as are not to be part of the finished construction, which will be furnished by the contractor. 4. The street must not be torn up for a greater distance than 500 feet in advance of the finished paving. The con- tractor must so arrange his work and deliver the material upon the street as to obstruct public travel as little as possible, and a roadway must be kept free on at least one side of the street for public travel. The contractor shall use all necessary precautions to pre- vent accidents by maintaining suitable barriers and by keep- ing lights burning at night. 5. GRADING AND EXCAVATION. The roadbed is to be ex- cavated to sub-grade, which will be twenty-four inches below the finished grade of the street. This excavation is to extend feet each side of the center line of track. If any fur- ther width of excavation is required, it will be directed by the engineer in writing, and paid for under clause c, paragraph 17. All material removed from the excavation is the property of the company, and must be promptly removed by the contrac- tor and deposited in such places and in such manner as may be designated by the engineer. It shall not, however, be hauled a distance greater than .... feet, except as provided for under clause/, paragraph 17. 128 STREET-RAILWAY ROADBED. No plowing will be allowed which disturbs the material be- low six inches above sub-grade. 6. SUB-DRAINS. If considered necessary by the engineer, a trench will be dug in the center of the roadway to such depth and grade as he shall prescribe. After thoroughly compacting the bottom of the trench, a inch porous tile-drain shall be laid and such connection made with the sewers or other drains as the engineer may direct. The trench is then to be refilled with clean gravel filling, in layers not exceeding twelve inches in thickness. Each layer is to be thoroughly compacted by ramming before another layer is added. 7. PREPARING SUB-GRADE. The sub-grade shall then be thoroughly rolled to the satisfaction of the engineer with a roller weighing not less than pounds per inch of roller. If any portions of the sub-grade cannot be reached by the roller, such portions shall be sprinkled with water and thoroughly compacted by ramming. If any spongy or vege- table matter, or material which cannot be rolled, is found in the excavation, it must be removed and the space below sub- grade filled with clean gravel filling. The roadbed shall be in a moist condition when rolled, and if dry must be mois- tened by the contractor. 8. BALLAST. Upon the sub-grade, prepared as above de- scribed, there shall be spread a layer five inches thick of broken-stone ballast, composed of stones not larger than two and one-half inches in their largest dimension. This layer shall be thoroughly compacted by rolling with the roller here- tofore described, or by ramming in such places as cannot be reached with the roller. 9. DISTRIBUTION OF TIES. Upon this layer of ballast the ties shall be distributed and spaced at intervals of ... inches on centers. The joint ties will be spaced inches on centers and arranged as shown on plan furnished by the engineer. 11. LAYING TRACK. The rails shall then be placed on the ties and the splice-plates bolted on. Care must be taken not to handle the rails in such a manner as to bend them or mar the heads or flanges. The rails will be spiked with four spikes SPECIFICATIONS. 129 tojthe_tie. Spikes will be staggered at least two and one-half inches in the tie, and driven in such a manner as to hold the tie at right angles to the track, except when otherwise directed. Brace tie-plates will be used and spiked to the tie with three spikes at intervals of .... feet. The rails will be laid with staggered joints, and no joint shall be more than twelve inches from a line drawn at right angles to the center of the opposite rail. Care must be taken to place the splice-plates squarely in position, and any scale or rust must be removed from the bearing-surfaces of plates and rail. The heads of the bolts must be struck with a two-pound hammer while pressure is applied on a thirty-inch wrench to tighten the bolts. The rail ends must be placed in as close contact as possible. The rails must not be bolted up for more than five rail lengths in advance of the finished paving. The gauge of the track shall not vary more than one-sixteenth of an inch from the standard on this road, which is .... feet .... inches. 12. SPECIAL WORK. In laying frogs, switches, and other special work, special care will be taken to maintain line, sur- face, and gauge. The latter will be widened on curves if so directed by the engineer, but not otherwise. The straight- track gauge at switches and mates will preferably be ^" tight. If the special work does not appear to fit, no attempt what- ever must be made to force it except by direction of the en- gineer. 13. KAISING TRACK AND TAMPING. After the preparation of the track as previously described, the entire track must then be raised to the finished grade and aligned to the lines given by the engineer. The space under the ties must then be filled with broken-stone ballast, composed of stones not larger than one and one-half inches in their largest dimen- sions. This shall be tamped under the ties in such a manner as to secure an even, solid bearing throughout the entire length and width of the tie. Care must be taken in raising and tamping the track not to deform the rails or splice-bars. The space between the ties is to be filled with the same ballast and thoroughly rammed. 130 STREET-RAILWAY ROADBED. 14. BONDING. The rails are to be bonded with the bond, applied in the following manner: 15. JOINTS. The joints are to be gone over again and each bolt tightened up, striking the head of each bolt with a two- pound hammer, while steady pressure is applied to the end of a thirty-inch wrench until they cannot be further tightened. 16. PREPARATION OF RAIL FOR PAVING. The recesses under the head and tram of the rail will be filled with con- crete in such a manner as to present a vertical surface for the paving to rest against. This concrete shall be composed of one part Rosendale cement, .... parts sand, and parts of broken stone, no piece of which shall be larger than 1" in its greatest dimension. 16. PAVING. Over the entire portion of the street to be repaved will be spread a layer of clean sharp gravel, not larger than }" in its largest dimension, and thoroughly com- pacted until its upper surface is eight inches below the finished grade. Especial care must be taken to thoroughly compact that portion between the ties. A layer of bedding sand will then be spread over the gravel of sufficient thickness to bring the granite blocks that are to be embedded in it to the proper grade after they are thoroughly rammed. The blocks are to be covered with clean, fine, and dry gravel or coarse sand, which shall be raked and swept until all the joints become filled therewith. The blocks shall then be rammed to a firm, unyielding surface to agree with the section of track as fur- nished by the engineer. No ramming will be done within fifteen feet of the face of the paving that is being laid. The blocks will again be covered with a layer of clean, fine, or dry gravel or coarse sand, and raked and swept until the joints are filled therewith. The blocks shall then be rammed until made solid and secure. Finally, the paving shall be covered with a layer at least 1" in thickness of fine dry screened gravel. 17. MEASUREMENTS. The work will be measured and paid for under the following prices: (a) Per foot of single track, including all excavation, re- filling, preparation of the sub-grade, ballasting, paving, and track-laying; SPECIFICATIONS. 131 r . (b) Special work shall be measured on the center line of track, measuring the center line from the separation of theo- retical center lines to the similar separation or to a point opposite the farthest joint of special work. Price per foot, determined in this manner, including items mentioned in clause (a) (c) Price per square yard for excavation, refilling, ballasting, and paving outside the limit of feet from the center line of track, when required by the engineer; (d) Price per cubic yard of excavating and refilling meas- ured in excavation for tile-drains (e) Price per running foot for laying tile-drains and con- necting to sewer or drains (/) Price per ton per 1000 feet for hauling material from the excavation a greater distance than feet from the excavation. 18. ESTIMATES. It shall be understood and agreed by the parties hereto that due measurements shall be taken during the progress of the work, and the estimate of the engineer shall be final and conclusive evidence of the amount of work performed by the contractor under and by virtue of this agreement, and shall be taken as the full measure of com- pensation to be received by the contractor. The aforesaid estimates shall be based upon the contract prices for the per- formance of all the work mentioned in these specifications and agreement, and when there may be % any ambiguity therein, the engineer's instructions shall be considered explanatory, and shall be of binding force. 132 STREET-RAILWAY ROADBED. TABLE III. MIDDLE OKDINATES, 10' CHORDS. M. 0. Radio*. M. 0. Radius. M. O. Radius. 0" Infinity 2" 75' 1 4" 37' 8" 1 4807' 8" 2399 3 1600 6 1 73 lift 72 9f 71 8J a" 4 3 3 2 37 4* 37 li 36 9J i 1200 9 70 8 t . 4 36 6/ 8 6 960 1 ~'51 69 7 36 3 ft 3 800 3ft 2-3 68 7 1 4? 3 s 35 11 T * / 682 3A 2 3 7 2 67 8f" 4 3 7 2 35 8* I 600 Oi 2i 66 9 t *i 35 5iJ 533 4? 2 3 9 Z 65 10ft 35 2j% X 480 Oft 2i S B 64 11 U 9 B 4_s 34 11, % 11 436 4 T 9 a 311 64 1 5 4 32 34 8i| 1 400 Oft 2f 63 3 S 34 5f y 369 3 2 3 1 62 5 1 4 3 3 34 2|| !' 342 10ft 2l ? B 61 7 I 4 is 33 111 j 320 OT 2^S 60 10 r 8 s 4i 5 33 9^ 300 Oi 24 60 1 - 41 33 6i j ; 282 4ft 266 8ft 253 0{ 2 19 59 4f 58 7J 57 iii 1 33 3i 33 O^S 32 10i 240 0/ B |}i 57 3 32 7* 228 7X 56 7 32 4f- 218 H| 28 55 Hi 4 2 32 2f 209 U* 2i 55 3- t 7 B 31 11 JJ i 200 2ft 192 Of 54 7 54 4f 2 31 9 T 6 8 31 6 i 184 7f 177 9} B It 53 5 52 lOj s Jjf 31 4f 8 1 2ft 171 54 s 52 3, 4j ? 30 llf ! 165 6| 160 Oi 154 10* | 51 8^ 51 2j 50 7 i i II 30 9 TB 30 7 30 4f i 150 Oi 3 2 50 1, 5 8 30 2i 11 145 6 141 2}J 137 2| 8$ $ 49 7ft 49 if 48 7f s 30 Oi 29 10 T V 29 71* ii Is 8 5 133 4J 129 9 T 5 B i 48 1ft 47 7 5i B 2 29 5 T | 29 3U 126 4* 33 47 2, % 5ft 29 1ft 1 123 2| 3 3 7 2 46 8 I 5 3 7 2 28 114 j^ 120 Of 3i 46 3*" 5i 9 28 94 1ft 117 li 114 4| 3?s 45 10ft 45 5ft 28 74 28 54 j 1? 111 8ft 3 3 a 45 5^J 28 34 If 109 1J U 44 7 5| 28 1, B iu ia 104 4y* 102 2i il 44 2| 43 9j 43 4 1 ft 1 27 llf 27 9f 27 7| 11 100 0} 9 43 O a 5* 27 6 j I 98 Oi 96 0* 94 2ft 42 7." 42 3ft 41 10H 5 3 | 27 4f B 27 2f 27 T | i 93 4* 3 2 41 61 5f 26 I(4j i i i ! 90 7 IB 88 114 87 4i 3 | 41 2i 40 10 40 5 | 26 9ft 26 7 T ^ 26 ^^g i 85 9/a 3 40 i; 54- 26 3}| i 84 3 TB 3 39 9}| 5i 26 2j i 82 10 81 5i 39 6, 39 2, 1 5p 26 (,% 25 10}* 8) 0- ls jl 83 10t 7 n M 25 9 r s 5 i i 78 i ;^|| 38 6f 25 7U i i 77 6 76 3i 3iJ 38 3* 37 Hi 5s! 25 4ft INDEX. PAGE American Society of Civil Engineers, Remarks before 26 Standard rail, Section of 34 Baltimore, T-rail construction in 40 Boston, Standard rail in 31 Brace, malleable iron , 57 Brooklyn, Standard rail in 32 Buenos Ayres, Livesey rail in 2 Chairs : First cast iron 2 Types of 10, 53 Chords, Table of 132 Clearance, Car 64 Curves: 60 Advantage of spiral 84 Car clearance on 64 Car-house 116 Compounding 61 Designing 87, 112 Easement 62 Gage on 73 Nomenclature of 77 Problems in laying out 89 Table of 132 Tables and formulae for use of spiral 84, 97 Turnouts 77 Frogs 78 Gage on curves 73 Grades, Effect of... 30 Grading 124 Guard-rails 66, 68, 72, 74 Joints : 43 Churchill 51 Continuous 51 133 134 INDEX. PAGE Joints : Girder 48 Ribbed 46 Weber 50 Welded 53 Wheeler 49, 50 Line- work 124 Maps, Making 125 Mate 78 Motive power 20 New Orleans, Rails in 20 New York City : Broadway rail in 13, 30 Construction of the Broadway cable road 27 Cable track in asphalt in 28 Third Avenue cable road 27 Flat rail in 1 Old horse-car rail in 26 Rails laid in asphalt in 28 Standard rail in 32 Nomenclature : Curves 77 Parts of rails 4 Nut-locks 48 Patent for rails, Early 7 Pavement 22 Pennsylvania R. R., Specifications for laying road-bed 37 Philadelphia, Flat rail in 3 Rails: Boston, in 31 Box girder 11 Broadway 13, 30 Brooklyn, in 32 Center-bearing rail 20 Combination rail 18 Electric rail 17 Flat rail in New York City 1 Gibbon duplex 11 Girder rail : Deep sections 31 Development of 13 First actually rolled 6 Later sections 13 Grooved 8,17 Guard-rail 66, 68, 72, 74 Horse-car rails in New York.: 26 Johnson, Early 8,30 Laid in asphalt in New York 28 Life of. . 24 INDEX. 135 PAGE Rails : Livesey ................................................. 2 New Orleans, in ........................................ 20 New York, in ........................................... 32 Nomenclature of parts of ............................... 4 Patent for, Early ........................................ 7 Side-bearing rail with electrically welded feet ............. 18 T rail : Adapted to street railways ................ ...... 34 Advantage of .................................. 36 Construction in Baltimore ........ ................... 40 Early type of .................................... 3 High ........................................... 42 Standard sections .............................. 34 Vignole ............................................... 3 Washington ......................................... 16,29 Wear of ................................................ 21 Wharton sections, Early ................................ 9 What governs the shape of ............................... 19 San Francisco, Cal. , Early girder rail in .......................... 6 Special Work : ....... ....................................... 60, 72 Adamantine ....................................... 79 Built-up ......... . ................................ 79, 82 Design of .................................... Ill Frogs, Standardization of .......................... Ill Guarantee ...................................... 79, 81 Manganese ..................................... 79, 81 Specifications ................................................ 37, 126 Surveys ...................................................... 122 Switches ....................................................... 77 Tables for spirals ............................................. 97 Temperature, Effect of changes in .............................. 43 Tie-bars ____ ................................................... 57 Tie-plates ...................................................... 57 Ties, Metallic ................................................ 42, 58 Trolley-wire, Locating .......................................... 114 Turnouts . ....... ........................................... 77 Street traffic ................................................. 21, 30 Track construction : Cable track in asphalt ...................... 28 New York, in .............................. 27 Washington, D. C., Standard rail in .......................... 16, 29 Wear ......................................................... 23 Welded joints ................... , ......... .................... 52 Wheel flanges ................................................ 67, 71 Wire, Locating the trolley ..................................... 114 OF THE UNIVERSITY ^LCALIFOR^L UNIVERSITY OF CALIFORNIA LIBRARY THIS BOOK IS DUE ON. THE LAST DATE STAMPED BELOW re 13