1911 Encyclopædia Britannica/Railways/Construction

Construction

Location.—An ideal line of railway connecting two terminal points would be perfectly level and perfectly straight, because in that case the resistance due to gradients and curves would be eliminated (see § Locomotive Power) and the cost of mechanical operation reduced to a minimum. But that ideal is rarely if ever attainable. In the first place the route of a railway must be governed by commercial considerations. Unless it be quite short, it can scarcely ever be planned simply to connect its two terminal points, without regard to the intervening country; in Order to be of the greatest utility and to secure the greatest revenue it must be laid out with due consideration of the traffic arising at intermediate places, and as these will not usually lie exactly on the direct line, deviations from straightness will be rendered necessary. In the second place, except in the unlikely event of all the places on-the selected route lying at the same elevation, a line that is perfectly level is a physical impossibility; and from engineering considerations, even one with uniform gradients will be impracticable on the score of cost, unless the surface of the country is extraordinarily even. In these circumstances the constructor has two broad alternatives between which to choose. On the one hand he may make the line follow the natural inequalities of the ground as nearly as may be, avoiding the elevations and depressions by curves; or on the other he may aim at making it as nearly straight and level as possible by taking it through the elevations in cuttings or tunnels and across the depressions on embankments or bridges. He will incline to the first of these alternatives when cheapness of first cost is a desideratum, but, except in unusually favourable circumstances, the resulting line, being full of sharp curves and severe gradients, will be unsuited for fast running and will be unable to accommodate heavy traffic economically. If, however, cost within reasonable limits is a secondary consideration and the intention is to build a line adapted for express trains and for the carriage of the largest volume of traffic with speed and economy, he will lean towards the second. In practice every line is a compromise between these two extremes, arrived at by carefully balancing a large number of varying factors. Other things being equal, that route is best which will serve the district most conveniently and secure the highest revenue; and the most favourable combination of curves and gradients is that by which the annual cost of conveying the traffic which the line will be called on to carry, added to the annual interest on the capital expended in construction; will be made a minimum.

Cuttings and Embankments.—A cutting, or cut, is simply a trench dug in a hill or piece of rising ground, wide enough at the bottom to accommodate one or more pairs of rails, and deep enough to enable the line to continue its course on the level or on a moderate gradient. The slopes of the sides vary according to the nature of the ground, the amount of moisture present, &c. In solid rock they may be vertical; in gravel, sand or common earth they must, to prevent slipping, rise 1 ft. for 1 to 1½ or 2 ft. of base, or even more in treacherous clay. In soft material the excavation may be performed by mechanical excavators or “steam navvies,” while in hard it may be necessary to resort to blasting. Except in hard rock, the top width of a cutting, and therefore the amount of material to be excavated, increases rapidly with the depth; hence if a cutting exceeds a certain depth, which varies with the particular circumstances, it may be more economical, instead of forming the sides at the slope at which the material of which they are composed will stand, to make them nearly vertical and support the soil with a retaining wall, or to bore a tunnel. An embankment-bank, or fill, is the reverse of a cutting, being an artificial mound of earth on which the railway is taken across depressions in the surface of the ground. An endeavour is made so to plan the works of a railway that the quantity of earth excavated in cuttings shall be equal to the quantity required for the embankments; but this is not always practicable, and it is sometimes advantageous to obtain the earth from some source close to the embankment rather than incur the expense of hauling it from a distant cutting. As embankments have to support the weight of heavy trains, they must be uniformly firm and well drained, and before the line is fully opened for traffic they must be allowed time to consolidate, a process which is helped by running construction or mineral trains over them.

An interesting case of embankment and cutting in combination was involved in crossing Chat Moss on the Liverpool & Manchester railway. The moss was 4½ m. across, and it varied in depth from 10 to 30 ft. Its general character was such that cattle could not stand on it, and a piece of iron would sink in it. The subsoil was composed principally of clay and sand, and the railway had to be carried over the moss on the level, requiring cutting, and embanking for upwards of 4 m. In forming 277,000 cub. yds. of embankment 670,000 yds. of raw peat were consumed, the difference being occasioned by the squeezing out of the water. Large quantities of embanking were sunk in the moss, and, when the engineer, George Stephenson, after a month’s vigorous operations, had made up his estimates, the apparent work done was sometimes less than at the beginning of the month. The railway ultimately was made to float on the bog. Where embankment was required drains about 5 yds. apart were cut, and when the moss between them was dry it was used to form the embankment. Where the way was formed on the level, drains were cut on each side of the intended line, and were intersected here and there by cross drains, by which the upper part of the moss was rendered dry and firm. On this surface hurdles were placed, 4 ft. broad and 9 long, covered with heath, upon which the ballast was laid.

Bridges.—For conveying small streams through embankments, channels or culverts are constructed in brickwork or masonry. Larger rivers, canals, roads, other railways and sometimes deep narrow valleys are crossed by bridges (q.v.) of timber, brick, stone, wrought iron or steel, and many of these structures rank among the largest engineering works in the world. Sometimes also a viaduct consisting of a series of arches is preferred to an embankment when the line has to be taken over a piece of flat alluvial plain, or when it is desired to economize space and to carry the line at a sufficient height to clear the streets, as in the case of various railways entering London and other large towns. In connexion with a railway many bridges have also to be constructed to carry public roads and other railways over the line, and for the use of owners or tenants whose land it has cut through (“accommodation bridges”). In the early days of railways, roads were often taken across the line on the level, but such “level” or “grade” crossings are now usually avoided in the case of new lines in populous countries, except when the traffic on both the road and the railway is very light. In many instances old level crossings have been replaced by over-bridges with long sloping approaches; in this way considerable expenditure has been involved, justified, however, by the removal of a danger to the public and of interruptions to the traffic on both the roads and the railways. In cases where the route of a line runs across a river or other piece of water so wide that the construction of a bridge is either impossible or would be more costly than is warranted by the volume of traffic, the expedient is sometimes adopted of carrying the wagons and carriages across bodily with their loads on train ferries, so as to avoid the inconvenience and delay of transshipment. Such train ferries are common in America, especially on the Great Lakes, and exist at several places in Europe, as in the Baltic between Denmark and Sweden and Denmark and Germany, and across the Straits of Messina.

Gradients.—The gradient or grade of a line is the rate at which it rises or falls, above or below the horizontal, and is expressed by stating either the horizontal distance in which the change of level amounts to 1 ft., or the amount of change that would occur in some selected distance, such as 100 ft., 1000 ft. or r m. In America a gradient of 1 in 100 is often known as a 1% grade, one of 2 in 100 as a 2% grade, and so on; thus a 0·25% grade corresponds to what in England would be known as a gradient of 1 in 400. The ruling gradient of a section of railway is the steepest incline in that section, and is so called because it governs or rules the maximum load that can be placed behind an engine working over that portion of line. Sometimes, however, a sharp incline occurring on an otherwise easy line is not reckoned as the ruling gradient, trains heavier than could be drawn up it by a single engine being helped by an assistant or “bank” engine; sometimes also “momentum” or “velocity” grades, steeper than the ruling gradient, are permitted for short distances in cases where a train can approach at full speed and thus surmount them by the aid of its momentum. An incline of 1 in 400 is reckoned easy, of 1 in 200 moderate and of 1 in 100 heavy. The ruling gradient of the Liverpool & Manchester railway was fixed at 1 in 900, excepting the inclines at Liverpool and at Rainhill summit, for working which special provision was made; and I. K. Brunel laid out the Great Western for a long distance out of London with a ruling gradient of 1 in 1320. Other engineers, however, such as Joseph Locke, cheapened the cost of construction by admitting long slopes of 1 in 80 or 70. One of the steepest gradients in England on an important line is the Lickey incline at Bromsgrove, on the Midland railway between Birmingham and Gloucester, where the slope is 1 in 37 for two miles. The maximum gradient possible depends on climatic conditions, a dry climate being the most favourable. The theoretical limit is about 1 in 16; between 1 in 20 and 1 in 16 a steam locomotive depending on the adhesion between its wheels and the rails can only haul about its own weight. In practice the gradient should not exceed 1 in 22½, and even that is too steep, since theoretical conditions cannot always be realized; a wet rail will reduce the adhesion, and the gradients must be such that some paying load can be hauled in all weathers. When an engineer has to construct a railway up a hill having a still steeper slope, he must secure practicable gradients by laying out the line in ascending spirals, if necessary tunnelling into the hill, as on the St Gothard railway, or in a series of zigzags, or he must resort to a rack or a cable railway.

Rack Railways.—In rack railways a cog-wheel on the engine engages in a toothed rack which forms part of the permanent way. The earliest arrangement of this kind was patented by John Blenkinsop, of the Middleton Colliery, near Leeds, in 1811, an an engine built on his plan by Mathew Murray, also of Leeds, began in 1812 to haul coals from Middleton to Leeds over a line 3½ m. long. Blenkinsop placed the teeth on the outer side of one of the running rails, and his reason for adopting a rack was the belief that an engine with smooth wheels running on smooth rails would not have sufficient adhesion to draw the load required. It was not till more than half a century later that an American, Sylvester Marsh, employed the rack system for the purpose of enabling trains to surmount steep slopes on the Mount Washinton railway, where the maximum gradient was nearly 1 in 2½. In this case the rack had pin teeth carried in a pair of angle bars. The subsequent development of rack railways is especially associated with a Swiss engineer, Nicholas Riggenbach, and his pupil Roman Abt, and the forms of rack introduced by them are those most commonly used. That of the latter is multiple, several rack-plates being placed parallel to each other, and the teeth break joint at ½, ⅓ or ¼ of their pitch, according to the number of rack-plates. In this way smoothness of working is ensured, the cog-wheel being constantly in action with the rack. Abt also developed the plan of combining rack and adhesional working; the engine working by adhesion alone on the gentler slopes but by both adhesion and the rack on the steeper ones. On such lines the beginning of a rack section is provided with a piece of rack mounted on springs, so that the pinions of the engine engage smoothly with the teeth. Racks of this type usually become impracticable for gradients steeper than 1 in 4, partly because of the excessive weight of the engine required and partly because of the tendency of the cog-wheel to mount the rack. The Locher rack, employed on the Mount Pilatus railway, where the steepest gradient is nearly 1 in 2, is double, with vertical teeth on each side, while in the Strub rack, used on the Jungfrau line, the teeth are cut in the head of a rail of the ordinary Vignoles type.

Cable Railways.—For surmounting still steeper slopes, cable railways may be employed. Of these there are two main systems: (1) a continuous cable is carried over two main drums at each end of the line, and the motion is derived either (a) from the weight of the descending load or (b) from a motor acting on one of the main drums; (2) each end of the cable is attached to wagons, one set of which accordingly ascends as the other descends. The weight required to cause the downward motion is obtained either by means of the material which has to be transported to the bottom of the hill or by water ballast, while to aid and regulate the motion generally steam or electric motors are arranged to act on the main drums, round which the cable is passed with a sufficient number of turns to prevent slipping. When water ballast is employed the water is filled into a tank in the bottom of the wagon or car, its quantity, if passengers are carried, being regulated by the number ascending or descending.

Curves.—The curves on railways are either simple, when they consist of a portion of the circumference of a single circle, or compound, when they are made up of portions of the circumference of two or more circles of different radius. Reverse curves are compound curves in which the components are of contrary flexure, like the letter S; strictly the term is only applicable when the two portions follow directly one on the other, but it is sometimes used of cases in which they are separated by a “tangent” or portion of straight line. In Great Britain the curvature is defined by stating the length of the radius, expressed in chains (1 chain = 66 ft.), in America by stating the angle subtended by a chord 100 ft. long; the measurements in both methods are referred to the central line of the track. The radius of a 1-degree curve is 5730 ft., or about 86⅔ chains, of a 15-degree curve 383 ft. or rather less than 6 chains; the former is reckoned easy, the latter very sharp, at least for main lines on the standard gauge. On some of the earlier, English main lines no curves were constructed of a less radius than a mile (80 chains), except at places where the speed was likely to be low, but in later practice the radius is sometimes reduced to 40 or 30 chains, even on high-speed passenger lines.

When a train is running round a curve the centrifugal force which comes into play tends to make its wheel-flanges press against the outer rail, or even to capsize it. If this pressure is not relieved in some way, the train may be derailed either (1) by “climbing” the outer rail, with injury to that rail and, generally, to the corresponding, wheel-flanges; (2) by overturning about the outer rail as a hinge, possibly without injury to rails or wheels; or (3) by forcing the outer rail outwards, occasionally to the extent of shearing the spikes that hold it down at the curve, thus spreading or destroying the track. In any case the details depend upon whether the vehicle concerned is an engine, a wagon or a passenger coach, and upon whether it runs on bogie-trucks or not. If it is an engine, particular attention must be directed to the type, weight, arrangement of wheels and height of centre of gravity above rail level. In considering the forces that produce derailment the total mass of the vehicle or locomotive may be supposed to be concentrated at its centre of gravity. Two lines may be drawn from this point, one to each of the two rails, in a plane normal to the rails, and the ends of these lines, where they meet the rails, may be joined to complete a triangle, which may conveniently be regarded as a rigid frame resting on the rails. As the vehicle sweeps round the curve the centre of gravity tends to be thrown outwards, like a stone from a horizontal sling. The vertical pressure of the frame upon the outer rail is thus increased, while its vertical pressure on the inner rail is diminished. Simultaneously the frame as a whole tends to slide horizontally athwart the rails, from the inner towards the outer rail, urged by the same centri fugal forces. This sliding movement is resisted by placing a check rail on the inner side of the inner rail, to take the lateral thrust of the wheels on that side. It is also resisted in part by the conicity of the wheels, which converts the lateral force partly into a vertical force, thus enabling gravity to exert a restoring influence. When the lateral forces are too great to be controlled “climbing” occurs. Accidents due to, simple climbing are, however, exceedingly rare, and are usually found associated with a faulty track, with “plunging” movements of the locomotive or vehicle, or with a “tight gauge” at curves or points.

From consideration of the rigid triangular frame described above, it is clear that the “overturning” force acts horizontally from the centre of gravity, and that the length of its lever arm is, at any instant, the vertical distance from the centre of gravity to the level of the outer rail. This is true whatever be the tilt of the vehicle at that instant. The restoring force exerted by gravity acts in a vertical line from the centre of gravity; and the length of its lever arm is the horizontal distance between this vertical line and the outer rail. If therefore the, outer rail is laid at a level above that of the inner rail at the curve, overturning will be resisted more than would be the case if both rails were in the same horizontal plane, since the tilting of the vehicle due to this “superelevation” diminishes the overturning moment, and also increases the restoring moment, by shortening in the one case and lengthening in the other the lever arms at which the respective forces act. The amount of super elevation required to prevent derailment at a curve can be calculated[1] under perfect running conditions, given the radius of curvature, the weight of the vehicle, the height of the centre of gravity, the distance between the rails, and the speed; but great experience is required for the successful application of definite formulae to the problem. For example, what is a safe speed at a given curve for an engine, truck or coach having the load equally distributed over the wheels may lead to either climbing or overturning if the load is shifted to a diagonal position. An ill-balanced load also exaggerates “plunging,” and if the period of oscillation of the load happens to agree with the changes of contour or other inequalities of the track vibrations of a dangerous character, giving rise to so-called “sinuous” motion may occur.

In general it is not curvature, but change of curvature, that presents difficulty in the laying-out of a line. For instance, if the curve is of S-form, the point of danger is when the train enters the contra-flexure, and it is not an easy matter to assign the best super elevation at all points throughout the double bend. Closely allied to the question of safety is the problem of preventing jolting at curves; and to obtain easy running it is necessary not merely to adjust the levels of the rails in respect to one another, but to tail off one curve into the next in such a manner as to avoid any approach to abrupt lateral changes of direction. With increase of speeds this matter has become important as an element of comfort in passenger traffic. As a first approximation, the centre-line of a railway may be plotted out as a number of portions of circles, with intervening straight tangents connecting them, when the abruptness of the changes of direction will depend on the radii of the circular portions. But if the change from straight to circular is made through the medium of a suitable curve it is possible to relieve the abruptness, even on curves of comparatively small radius. The smoothest and safest running is, in fact, attained when a “transition,” “easement” or “adjustment” curve is inserted between the tangent and the point of circular curvature.

For further information see the following papers and the discussions on them: “Transition Curves for Railways,” by James Glover, Proc. Inst. C.E. vol. 140, part ii.; and “High Speed on Railway Curves,” by J. W. Spiller, and “A Practical Method for the Improvement of Existing Railway Curves,” by W. H. Shortt, Proc. Inst. C.E. vol. 176, part ii.

Gauge.—The gauge of a railway is the distance between the inner edges of the two rails upon which the wheels run. The width of 4 ft. 8½ in. may be regarded as standard, since it prevails on probably three-quarters of the railways of the globe. In North America, except for small industrial railways and some short lines for local traffic, chiefly in mountainous country, it has become almost universal; the long lines of 3 ft. gauge have mostly been converted, or a third rail has been laid to permit interchange of vehicles, and the gauges of 5 ft. and more have disappeared. A considerable number of lines still use 4 ft. 9 in., but as their rolling stock runs freely on the 4 ft. 8½ in. gauge and vice versa, this does not constitute a break of gauge for traffic purposes. The commercial importance of such free interchange of traffic is the controlling factor in determining the gauge of any new railway that is not isolated by its geographical position. In Great Britain railways are built to gauges other than 4 ft. 8½ in. only under exceptional conditions; the old “broad gauge” of 7 ft. which I. K. Brunel adopted for the Great Western railway disappeared on the 20th–23rd of May 1892, when the mainline from London to Penzance was converted to standard gauge throughout its length. In Ireland the usual gauge is 5 ft. 3 in., but there are also lines laid to a 3 ft. gauge. On the continent of Europe the standard gauge is generally adopted, though in France there are many miles of 4 ft. 9 in. gauge; the normal Spanish and Portuguese gauge is, however, 5 ft. 5¾ in., and that of Russia 5 ft. In France and other European countries there is also an important mileage of metre gauge, and even narrower, on lines of local or secondary importance. In India the prevailing gauge is 5 ft. 6 in., but there is a large mileage of other gauges, especially metre. In the British colonies the prevailing gauge is 3 ft. 6 in., as in South Africa, Queensland, Tasmania and New Zealand; but in New South Wales the normal is 4 ft. 8½ in. and in Victoria 5 ft. 3 in., communication between different countries of the Australian Commonwealth being thus carried on under the disadvantage of break of gauge. Though the standard gauge is in use in Lower Egypt, the line into the Egyptian Sudan was built on a gauge of 3 ft. 6. in., so that if the so-called Cape to Cairo railway is ever completed, there will be one gauge from Upper Egypt to Cape Town. In South America the 5 ft. 6 in. gauge is in use, with various others.

Mono-Rail Systems.—The gauge may be regarded as reduced to its narrowest possible dimensions in mono-rail lines, where the weight of the trains is carried on a single rail. This method of construction, however, has been adopted only to a very limited extent. In the Lartigue system the train is straddled over a single central rail, elevated a suitable distance above the ground. A short line of this kind runs from Ballybunnion to Listowel in Ireland, and a more ambitious project on the same principle, on the plans of Mr F. B. Behr, to connect Liverpool and Manchester, was sanctioned by Parliament in 1901. In this case electricity was to be the motive-power, and speeds exceeding 100 m. an hour were to be attained, but the line has not been built. In the Langen mono-rail the cars are hung from a. single overhead rail; a line on this system works between Barmen and Elberfeld, about 9 m., the cars for a portion of the distance being suspended over the river Wupper. In the system devised by Mr Louis Brennan the cars run on a single rail laid on the ground, their stability being maintained by a heavy gyrostat revolving at great speed in a vacuum.

Permanent Way.—When the earth-works of a line have been completed and the tops of the embankments and the bottoms of the cuttings brought to the level decided upon, the next step is to lay the permanent way, so-called probably in distinction to the temporary way used during construction. The first step is to deposit a layer of ballast on the road-bed or “formation,” which often slopes away slightly on each side from the central line to facilitate drainage. The ballast consists of such materials as broken stone, furnace slag, gravel, cinders or earth, the lower layers commonly consisting of coarser materials than the top ones, and its purpose is to provide a firm, well-drained foundation in which the sleepers or cross ties may-be embedded and held in place, and by which the weight of the track and the trains may be distributed over the road-bed. Its depth varies, according to the traffic which the line has to bear, from about 6 in. to 1 ft. or rather more under the sleepers, and the materials of the surface layers are often chosen so as to be more or less dustless. Its width depends on the numbers of tracks and their gauge; for a double line of standard gauge it is about 25 ft., a space of 6 ft. (“six-foot way”) being left between the inner rails of each pair in Great Britain (ng. 8), and a rather larger distance in America (fig. 9), where the over-hang of the rolling stock is greater. The intervals between the sleepers are filled in level with ballast, which less commonly is also heaped up over them, especially at the projecting ends.

Fig. 8.—Half of English Double Track.


Fig. 9.—Half of American Double Track.

Sleepers, called ties or cross-ties in America, are the blocks or slabs on which the rails are carried. They are nearly always placed transversely, across the direction of the lines, the longitudinal position such as was adopted in connexion with the broad gauge on the Great Western in England having been abandoned except in special cases. Stone blocks were tried as sleepers in the early days of railways, but they proved too rigid, and besides, it was found difficult to keep the line true with them. Wood is the material most widely used, but steel is employed in some countries where timber is scarce or liable to destruction by white ants, though it is still regarded as too expensive in comparison with wood for general adoption. Steel sleepers were used experimentally on the London & North-Western, but were abandoned owing to the shortness of their life. In Germany, where they have met with greater favour, there were over 26½ millions in use in 1905,[2] and they have been tried by some American railways. Numerous forms of ferro-concrete sleepers have also been devised.

In Great Britain, Germany and France, at least 90% of the wooden sleepers are “treated” before they are laid, to increase their resistance to decay, and the same practice is followed to some extent in other European countries. A great number of preservative processes have been devised. In that most largely used, known as “creosoting,” dead oil of tar, to the amount of some 3 gallons per sleeper, is forced into the wood under pressure, or is sucked in by vacuum, both the timber and the oil being heated. In the United States only a small percentage of the ties are treated in any way beyond seasoning in the open air, timber, in the opinion of the railway officials, being still too cheap in nearly all parts of that country to justify the use of preservatives. Some railway companies, however, having a long mileage in timberless regions, do “treat” their sleepers.

Typical dimensions for sleepers on important British railways are:—length 9 ft., breadth 10 in., and depth 5 in. In America 8 ft. is the most common length, the breadth being 8 in., and the depth 6 or 7 in.

Fig. 10—A, Section of British Bull-Headed Rail, 90 ℔ to the yard, showing also chair and fastenings. B, Plan of Chair.

Fig. 11—British Rail and Rail Joint.

There are two main ways of attaching the rails to the sleepers, corresponding to two main types of rails—the bull-headed rail and the Vignoles or flange rail. In the first method, which is practically universal in Great Britain and is also employed to some extent in France and India, the rails have rounded bases and are supported by being wedged, with wooden keys, in cast-iron chairs which are bolted to the sleepers. In the second, method the rails have flat flanged bases which rest directly on the sleepers (fig. 10). The chairs on the British system weigh about 45 or 50 ℔ each on important lines, though they may be less where the traffic is light, and are fixed to the sleepers each by two, three or four fastenings, either screw spikes, or round drift bolts entered in holes previously bored, or fang bolts or wooden trenails. Sometimes a strip of felt is interposed between the chair and the sleeper, and sometimes a serrated surface is prepared on the sleeper for the chair which is forced into its seat by hydraulic pressure. The keys which hold the rail in the chairs are usually of oak and are placed outside the rails; the inside position has also been employed, but has the disadvantage of detracting from the elasticity of the road since the weight of a passing train presses the rails up against a rigid mass of metal instead of against a slightly yielding block of wood. The rails, which for heavy main line traffic may weigh as much as 100 ℔ per yard, or even more, are rolled in lengths of from 30 to 60 ft., and sleepers are placed under them at intervals of between 2 and 3 ft. (centre to centre), 11 sleepers to a 30 ft. rail being a common arrangement. On the London & North-Western railway there are 24 sleepers to each 60 ft. rail. A small space is left between the end of one rail and that of the next, in order to allow for expansion in hot weather, and at the joint the two are firmly braced together by a pair of fish-plates (fig. 11). These are flat bars of iron or steel from 18 in. to 2 ft. long, which are lodged in the channels of the rail, one on each side, and secured with four bolts passing through the web; sometimes, to give additional stiffness, they extend down below the lower table of the rail and are bent round so as to clip it. Occasionally the joints thus formed are “supported” on a sleeper, as was the practice in the early days of railway construction, but they are generally “suspended” between two sleepers, which are set rather more closely together than at other points in the rail. Preferably, they are so arranged that those in both lines of rails come opposite each other and are placed between the same pair of sleepers.

Fig. 12.—French Rail, 90¼ ℔ to the yard, showing rail joint and seat in the sleeper.

Flat-bottomed rails are fastened to the sleepers by hook headed spikes, the heads of which project over the flanges. In the United States the spikes are simply driven in with a maul, and the rails stand upright, little care being taken to prepare seats for them on the sleepers, on which they soon seat themselves. The whole arrangement is simple and cheap in first cost, and it lends itself admirably to fast track-laying and to repairs and changes of line. On the continent of Europe the practice is common of notching the sleeper so as to give the rail a slight cant inwards—a result obtained in England by canting the rail in the chairs—and metal plates or strips of felt are put under the rail, which is carefully fastened to the sleeper by screwed spikes (fig. 12). This method of construction is more expensive than the American in first cost, but it gives a more durable and stable track. Such metal plates, or “tie-plates,” have come into considerable use also in the United States, where they are always made of rolled steel, punched with rectangular holes through which the spikes pass. They serve two principal purposes: they diminish the wear of the sleeper under the rail by providing a larger bearing surface, and they help to support the spikes and so to keep the gauge. On all the accepted forms there are two or more flanges at the bottom, running lengthwise of the plate and crosswise of the rail; these are requisite to give proper stiffness, and further, as they are forced into the tie by the weight of passing traffic, they help to fix the plate securely in place. The joints of flanged rails are similar to those employed with bull-headed rails. Various forms, mostly patented, have been tried in the United States, but the one most generally adopted consists of two symmetrical angle bars (fig. 13), varying in length (from 20 to 48 in.), in weight and in the number of bolts, which may be four or six.

Fig. 13.—American Rail, 90 ℔ to the yard, showing rail joint.

The substitution of steel for iron as the material for rails which made possible the axle loads and the speeds of to-day, and, by reducing the cost of maintenance contributed enormously to the economic efficiency of railways, was one of the most important events in the history of railways, and a scarcely less important element of progressive economy has been the continued improvement of the steel rail in stiffness of section and in toughness and hardness of material. Carbon is the important element in controlling hardness, and the amount present is in general higher in the United States than in Great Britain. The specifications for bull-headed rails issued by the British Engineering Standards Committee in 1904 provided for a carbon-content ranging from 0·35 to 0·50%, with a phosphorus maximum of 0·075%. In the United States a committee of the American Society of Civil Engineers, appointed to consider the question of rail manufacture in consequence of an increase in the number of rail-failures, issued an interim report in 1907 in which it suggested a range of carbon from 0·55 to 0·65% for the heaviest sections of Bessemer steel flange rails, with a phosphorus maximum of 0·085%; while the specifications of the American Society for Testing Materials, current at the same period, put the carbon limits at 0·45 to 0·55%, and the phosphorus limit at 0·10. For rails of basic open-hearth steel, which is rapidly ousting Bessemer steel, the Civil Engineers’ specifications allowed from 0·65 to 0·75% of carbon with 0·05% of phosphorus, while the specifications of the American Railway Engineering and Maintenance of Way Association provided for a range of 0·75 to 0·85% of carbon, with a maximum of 0·03% of phosphorus. The rail-failures mentioned above also drew renewed attention to the importance of the thermal treatment of the steel from the time of melting to the last passage through the rolling mill and to the necessity of the finishing temperature being sufficiently low if the product is to be fine grained, homogeneous and tough; and to permit of this requirement being met there was a tendency to increase the thickness of the metal in the web and flanges of the rails. The standard specification adopted by the Pennsylvania railway in 1908 provided that in rails weighing 100 ℔ to the yard 41% of the metal should be in the head, 18·6% in the web, and 40·4% in the base, while for 85 ℔ rails 42·2% was to be in the head, 17·8% in the web and 40·0% in the base. These rails were to be rolled in 33-ft. lengths. According to the specification for 85 ℔ rails adopted by the Canadian Pacific railway about the same time, 36·77% of the metal was to be in the head, 22·21% in the web and 41·02% in the base.

Fig. 14.—Points and Crossings.

FP =  Facing points.
TP = Trailing points.
a = Stock rail.
b = Switch rail.
 V = Single or V-crossing.
 D = Diamond crossing.
c = Check rails.
d = Wing rails.
e = Winged check rails.
f = Diamond points.

Points and Crossings.—To enable trains to be transferred from one pair of rails to another pair, as from the main line to a siding, “points” or “switches” are provided. At the place where the four rails come together, the two inner ones (one of the main line and the other of the siding), known as “switch rails” (b, fig. 14), are tapered to a fine point or tongue, and rigidly connected together at such a distance apart that when one of the points is pressed against the outer or “stock” rail (a) of either the siding or the main line there is sufficient space between the other tongue and the other stock rail to permit the free passage of the flanges of the wheels on one side of the train, while the flanges on the other side find a continuous path along the other switch rail and thus are deflected in the desired direction. The same arrangement is employed at junctions where different running lines converge. The points over which a train travels when directed from the main a branch line are called “facing points” (FP), while those which it passes when running from a branch to a main line are “trailing points” (TP). In Great Britain the Board of Trade requires facing points to be avoided as far as possible; but, of course, they are a necessity at junctions where running lines diverge and at the crossing places which must be provided to enable trains to pass each other on single-track lines. At stations the points that give access to sidings are generally arranged as trailing points with respect to the direction of traffic on the main lines; that is, trains cannot pass direct into sidings, but have to stop and then run backwards into them. In shunting yards the points are commonly set in the required direction by means of hand levers placed close beside the lines, but those at junctions and those which give access from the main lines to sidings at wayside stations are worked by a system of rods from the signal cabin, or by electric or pneumatic power controlled from it and interlocked with the signals (see Signal: § Railway). Crossings are inevitable adjuncts of points. Where a branch diverges from a main line, one rail of the one must cross one rail of the other, and a V-crossing is formed (V). Where, as at a double-line junction, one pair of rails crosses another pair, “diamond” crossings (D) are formed. At both types of crossing, check rails (c) must be provided to guide the wheel-flanges, and if these are not accurately placed the safety of the trains will be endangered. At double-line junctions trains passing over the diamond crossings evidently block traffic going in the opposite direction to that in which they are travelling. To avoid-the delay thus caused the branch line which would occasion the diamond crossing if it were taken across on the level is sometimes carried over the main line by an over-bridge (“flying junction”) or under it by an under-bridge (“burrowing junction”).

Railway Stations.—Railway stations are either “terminal” or “intermediate.” A terminal station embraces (1) the passenger station; (2) the goods station; (3) the locomotive, carriage and waggon depots, where the engines and the carrying stock are kept, cleaned, examined and repaired. At many intermediate stations the same arrangements, on a smaller scale, are made; in all of them there is at least accommodation for the passenger and the goods traffic. The stations for passengers and goods are generally in different and sometimes in distant positions, the place selected for each being that which is most convenient for the traffic. The passenger station abuts on the main line, or, at termini, forms the natural terminus, at a place as near as can conveniently be obtained to the centre of the, population which constitutes the passenger traffic; and preferably its platforms should be at or near the ground level, for convenience of access. The goods station is approached by a siding or fork set off from the main line at a point short of the passenger station. In order to keep down the expense of shunting the empty trains and engines to and from the platforms the carriage and locomotive depots should be as near the passenger station as possible; but often the price of land renders it impracticable to locate them in the immediate vicinity and they are to be found at a distance of several miles.

In laying out the approaches and station yard of a passenger station ample width and space should be provided, with well-definedPassenger station. means of ingress and egress to facilitate the circulation of vehicles and with along setting-down pavement to enable them to discharge their passengers and luggage without delay. The position of the main buildings—ticket offices, waiting and refreshment-rooms, parcels offices, &c.—relative to the direction of the lines of rails may be used as a means of classifying terminal stations. They are placed either on the departure side parallel to the platform (“side” stations) or at right angles to the rails and platforms (“end” stations). Many large stations, however, are of a mixed type, and the offices are arranged in a fork between two or more series of platforms, or partly at the end and partly on one side. Where heavy suburban traffic has to be dealt with, the expedient is occasionally adopted of taking some of the lines round the end in a continuous loop, so that incoming trains can deposit their passengers at an underground platform and immediately proceed on their outward journey. Intermediate stations, like terminal ones, should be convenient in situation and easy of approach, and, especially if they are important, should be on the ground level rather than on an embankment or in a cutting. The lines through them should be, if possible, straight and on the level; the British Board of Trade forbids them being placed on a gradient steeper than 1 in 260, unless it is unavoidable. Intermediate stations at the surface level are naturally constructed as side stations, and whether offices are provided on both sides or whether they are mainly concentrated on one will depend on local circumstances, the amount of the traffic, and the direction in which it preponderates. When the railway lies below the surface level the bulk of the offices are often placed on a bridge spanning the lines, access being given to the platforms by staircases or lifts, and similarly when the railway is at a high level the offices may be arranged under the lines. Occasionally on a double-track railway one platform placed between the tracks serves both of them; this “island” arrangement, as it is termed, has the advantage that more tracks can be readily added without disturbance of existing buildings, but when it is adopted the exit from the trains is at the opposite side to that which is usual, and accidents have happened through passengers alighting at the usual side without noticing the absence of a platform. At stations on double-track railways which have a heavy traffic four tracks are sometimes provided, the two outside ones only having platforms, so that fast trains get a clear road and can pass slow ones that are standing in the station. In Great Britain, it may be noted, trains almost invariably keep to the left, whereas in most other countries right-handed running is the rule.

The arrangement and appropriation of the tracks in a station materially affect the economical and efficient working of the traffic. There must be a sufficient provision of sidings, connected with the running tracks by points, for holding spare rolling stock and to enable carriages to be added to or taken off trains and engines to be changed with as little delay as possible. At terminal stations, especially at such as are used by short-distance trains which arrive at and start from the same platform, a third track is often laid between a pair of platform tracks, so that the engine of a train which has arrived at the platform can pass out and place itself at the other end of the train, which remains undisturbed. At the new Victoria station (London) of the London, Brighton & South Coast railway—which is so long that two trains can stand end to end at the platforms—this system is extended so as to permit a train to start out from the inner end of platform even another train is occupying the outer end. One of the advantages of electric trains on the multiple control system they economize terminal accommodation, because they driven from either end indifferently, and therefore avoid the necessity for tracks by which engines can change from one end of the train to the other.

The platforms on British railways have standard elevation of 3 ft. above rail level, and they are not now made less than 2½ ft. in height. In other countries they are generally lower; in the United States they are commonly level with, or only a few inches higher than, the top of the rails. They may consist of earth with a retaining wall along the tracks and with the surface gravelled or paved with stone or asphalt, or they may be constructed entirely of timber, or they may be formed of stone slabs supported on longitudinal walls. They should be of ample dimensions to accommodate the traffic—the British Board of Trade requires them to be not less than 6 ft. wide at small stations and not less than 12 ft. wide at large ones—and they should be as free as possible from obstructions, such as pillars supporting the roof. At intermediate stations the roofs are often carried on brackets fixed to the walls of the station buildings, and project only to the edge of the platforms. At larger stations where both the platforms and the tracks are covered in, there are two broad types of construction, with many intermediate variations: the roof may either be comparatively low, of the “ridge and furrow” pattern, borne on a number of rows of pillars, or it may consist of a single lofty span extending clear across the area from the side walls. The advantage claimed for roofs formed with one or two large spans is that they permit the platforms and tracks to be readily rearranged at any time as required, whereas this is difficult with the other type, especially since the British Board of Trade requires the pillars to be not less than 6 ft. away from the edges of the platforms. On the other hand, wide spans are more expensive both in first cost and in maintenance, and there is the possibility of a failure such as caused the collapse in December 1905 of the roof of Charing Cross (S.E.R.) station, London, which then consisted of a single span. Whatever the pattern adopted for the roof, a sufficient portion of it must be glazed to admit light, and it should be so designed that the ironwork can be easily inspected and painted and the glass readily cleaned. For the illumination of large stations by night electric arc lamps are frequently employed, but some authorities favour high-pressure incandescent gas-lighting.,

At busy stations separate tracks are sometimes appropriated to the use of light engines and empty trains, on which, they mayLocomotive depots. be run between the platforms and the locomotive and, carriage depots. A carriage depot includes sheds in which the vehicles are stored, arrangements for washing and cleaning them, and sidings on which they are marshalled into trains. At a locomotive depot the chief building is the “running shed” in which the engines are housed and cleaned. This may be rectangular in shape (“straight” shed), containing a series of parallel tracks on which the engines stand and which are reached by means of points and crossings diverging from a main track outside; or it may take a polygonal or circular form (round house or rotunda), the lines for the engines radiating from a turn-table which occupies the centre and can be rotated so as to serve any of the radiating lines. The second arrangement enables any particular engine to, enter or leave without disturbing the other; but on the other hand an accident to the turn-table may temporarily imprison the whole of them. In both types pits are constructed between the rails on which the engines stand to afford easy access for the inspection and cleaning of their mechanism. Machine shops are usually provided to enable minor repairs to be executed; the tendency, both in England and America, is to increase the amount of such repairing plant at engine sheds, thus lengthening the intervals between the visits of the engines to the main repairing shops of the railway. A locomotive depot further includes stores of the various materials required in working the engines, coal stages at which they are loaded with coal, and an ample supply of water. The quality of the last is a matter of great importance; when it is unsuitable, the boilers will suffer, and the installation of a water-softening plant may save more in the expenses of boiler, maintenance than it costs to operate. The water cranes or towers which are placed at intervals along the railway to supply the engines with water require similar care in regard to the quality of the water laid on to them, as also to the water troughs, or track tanks as they are called in America, by which engines are able to pick up water without stopping. These consist of shallow troughs about 18 in wide, placed between the rails on perfectly level stretches of line. When water is required, a scoop is lowered into them from below the engine, and if the speed is sufficient the water is forced up it into the tender-tanks. Such troughs were first employed on the London & North-Western railway in 1857 by John Ramsbottom, and have since been adopted on many other lines.

Goods stations vary in size from those which consist of perhaps a single siding, to those which have accommodationGoods stations. for thousands of wagons. At a small roadside station, where the traffic is of a purely local character, there will be some sidings to which horses and carts have access for handling bulk goods like coal, gravel, manure, &c., and a covered shed for loading and unloading packages and materials which it is undesirable to expose to the weather. The shed may have a single pair of rails for wagons running through it along one side of a raised platform, there being a roadway for carts on the other side; or if more accommodation is required there may be two tracks, one on each side of the platform, which is then approached by carts at the end. In either case the platform is fitted with a crane or cranes for lifting merchandise into and out of the wagons, and doors enable the shed to be used as a lock-up warehouse. In a large station the arrangements become much more complicated, the precise design being governed by the nature of the traffic that has to be served and by the physical configuration of the site. It is generally convenient to keep the inwards and the outwards traffic distinct and to deal with the two classes separately; at junction stations it may also be necessary to provide for the transfer of freight from one wagon to another, though the bulk of goods traffic is conveyed through to its destination in the wagons into which it was originally loaded. The increased loading space required in the sheds is obtained by multiplying the number and the length of lines and platforms; sometimes also there are short sidings, cut into the platforms at right angles to the lines, in which wagons are placed by the aid of wagon turn-tables, and sometimes the wagons are dealt with on two floors, being raised or lowered bodily from the ground level by lifts. The higher floors commonly form warehouses where traders may store goods which have arrived or are awaiting despatch. An elaborate organization is required to keep a complete check and record of all the goods entering and leaving the station, to ensure that they are loaded into the proper wagons according to their destination, that they are unloaded and sorted in such a way that they can be delivered to their consignees with the least possible delay, that they are not stolen or accidentally mislaid, &c.; and accommodation must be provided for a large clerical and supervisory staff to attend to these matters. British railways also undertake the collection and delivery of freight, in addition to transporting it, and thus an extensive range of vans and wagons, whether drawn by horses or mechanically propelled, must be provided in connexion with an important station.

Shunting Yards.—It may happen that from a large station sufficient traffic may be consigned to certain other large stations to enable full train-loads to be made up daily, or several times a day, and dispatched direct to their destinations. In general, however, the conditions are less simple. Though a busy colliery may send off its product by the train-load to an important town, the wagons will usually be addressed to a number of different consignees at different depots in different parts of the town, and therefore the train will have to be broken up somewhere short of its destination and its trucks rearranged, together with those of other trains similarly constituted, into fresh trains for conveyance to the various depots. Again, a station of moderate size may collect goods destined for a great variety of places but not in sufficient quantities to compose a full train-load for any of them, and then it becomes impossible, except at the cost of uneconomical working, to avoid dispatching trains which contain wagons intended for many diverse destinations. For some distance these wagons will all travel over the same line, but sooner or later they will reach a junction-point where their ways will diverge and where they must be separated. At this point trains of wagons similarly destined for different places will be arriving from other lines, and hence the necessity will arise of collecting together from all the trains all the wagons which are travelling to the same place.

Fig. 15.—Diagram to illustrate use of Shunting Yards.

The problem may be illustrated diagrammatically as follows (fig. 15): A may be supposed to be a junction outside a large seaport where branches from docks a, b, c and d converge, and where the main line also divides into three going to B, C and D respectively. A train from a will contain some wagons for B, some for C and some for D, as will also the trains from a, b, c and d. At A therefore it becomes necessary to disentangle and-group together all the wagons that are intended for B, all that are intended for C, and all that are intended for D. Even that is not the whole of the problem. Between A and B, A and C, and A and D, there may be a string of stations, p, q, r, s, &c., all receiving goods from a, b, c and d, and it would manifestly be inconvenient and wasteful of time and trouble if the trains serving those intermediate stations were made up with, say, six wagons from a to p next the engine, five from b to p at the middle, and four from c to p near the end. Hence at A the trucks from a, b, c and d must not only be sorted according as they have to travel along A B, A C, or A D, but also must be marshalled into trains in the order of the stations along those lines. Conversely, trains arriving at A from B, C and D must be broken up and remade in order to distribute their wagons to the different dock branches.

To enable the wagons to be shunted into the desired order yards containing a large number of sidings are constructed at important junction points like A. Such a yard consists essentially of a group or groups of sidings, equal in length at least to the longest train run on the line, branching out from a single main track and often again converging to a single track at the other end; the precise design, however, varies with the amount and character of the work that has to be done, with the configuration of the ground, and also with the mode of shunting adopted. The oldest and commonest method of shunting is that known as “push-and-pull,” or in America as “link-and-pin” or “tail” shunting; An engine coupled to a batch of wagons runs one or more of them down one siding, leaves them there, then returns back with the remainder clear of the points where the sidings diverge, runs one or more others down another siding, and so on till they are all disposed of. The same operation is repeated with fresh batches of wagons, until the sidings contain a number of trains, each intended, it may be supposed, for a particular town or district. In some cases nothing more is required than to attach an engine and brake-van (“caboose”) and despatch the train; but if, as will happen in others, a further rearrangement of the wagons is necessary to get them into station order this is effected on the same principle.

Push-and-pull shunting is simple, but it is also slow, and therefore efforts have been made at busy yards where great numbers of trains are dealt with to introduce more expeditious methods. One of these, employed in America, is known as “poling.” Alongside the tracks on which stand the trains that are to be broken up and from which the sidings diverge subsidiary tracks are provided for the use of the shunting engines. These engines have a pole projecting horizontally in front of them, or are attached to a “pole-car” having such a pole. The method of working is for the pole to be swung out behind a number of wagons; one engine is then started and with its pole pushes the wagons in front of it until their speed is sufficient to carry them over the points, where they are diverted into any desired siding. It then runs back to the train to repeat the operation, but while it is doing so a second engine similarly equipped has poled away a batch of wagons on the opposite side. In this way a train is distributed with great rapidity, especially if the points giving access to the different sidings are worked by power so that they can be quickly manipulated.

Another method, which was introduced into America from Europe about 1890, is that of the summit or “hump.” The wagons are pushed by an engine at their rear up one slope of an artificial mound, and as the run down the other slope by gravity are switched into the desired siding. Sometimes a site can be found for the sorting sidings where the natural slope of the ground is sufficiently steep to make the wagons run down of themselves. One of the earliest and best known of such “gravity” yards is that at Edgehill, near Liverpool, on the London & North-Western railway, which was established in 1873. Here, at the highest level, there are a number of “upper reception lines” converging to a single line which leads to a group of “sorting sidings” at a lower level. These in turn converge to a pair of single lines which lead to two groups of marshalling sidings, called “gridirons” from their shape, and these again converge to single lines leading to “lower reception and departure lines” at the bottom of the slope. The wagons from the upper reception lines are sorted into trains on the sorting sidings, and then, in the gridirons, are arranged in the appropriate order and marshalled ready to be sent off from the departure lines.  (H. M. R.) 

  1. See The Times Engineering Supplement (August 22, 1906), p. 265.
  2. See a full account of steel sleepers in a paper read by A. Haarmann before the Verein der Deutschen Eisenhüttenleute on Dec. 8, 1907, translated in the Railway Gazette (London) on April 3, 10 and 17, 1908.