Chapter III: 182 (2)

With a view to greater permanency and durability, stone sleepers were tried. These consisted of square blocks of good hard stone, measuring about 2 feet wide each way and 12 inches thick. Holes were cut in the stone, and plugs of hard wood inserted. The cast-iron chairs were then placed on the top of the blocks, and the iron spikes driven through the chair-holes into the wooden plugs. The elements of permanency were there certainly, but a rougher road it would be impossible to conceive. The stone was solid and unyielding, there was a total absence of softness and elasticity, and the harsh noisy effect produced when running over the stone-block road very soon became intolerable. Stone-block sleepers were found to be a failure, and were all removed. On some of our old lines, numbers of them, with the chair marks plainly visible, may be still seen in loading banks, buildings, sea walls, and other works for which they were never originally intended, but for which their size and weight render them very appropriate.

Wooden sleepers are used in two forms, transverse and longitudinal. In the former, as in Fig. 301, the sleepers not only carry the rails, but also preserve the gauge; in the latter as in Fig. 302, the longitudinal sleepers only support the rails, additional timbers and strong fastenings being necessary to maintain the gauge.

Longitudinal sleepers have been used to a large extent for bridge rails, it being supposed that with the broad continuous sleeper a lighter and shallower rail could be adopted, which would be equally efficient as a heavier rail on cross-sleepers. Excellent running roads have been made with longitudinal sleepers, notwithstanding the difficulty of making a good bridge-rail joint; but it is well to bear in mind that almost all the lines which originally adopted this form of permanent way have since reverted to the ordinary cross-sleeper road. The longitudinal sleeper road is an expensive road to lay down and maintain. The main pieces are of large scantling, must be of good quality of timber, and are consequently costly. The cross-pieces, or transomes, must be carefully fitted and secured with heavy ironwork. Where there is much traffic, the removal and renewal of one of the long timbers is much more difficult than the renewal of several ordinary 211 cross-sleepers. Again, decay may take place on only one portion of a main timber, but there is no alternative but to remove the entire piece.

For gauges varying from 4 feet 8½ inches to 5 feet 3 inches, cross-sleepers are cut to the length of 8 feet 11 inches, and are generally rectangular in section, as in Fig. 303, measuring 10 inches in width by 5 inches in thickness. On some of the lighter railways with small traffic, sleepers are often used only 9 inches wide by 4½ inches thick, while occasionally on some lines, and in places where there is exceptionally heavy and constant traffic, sleepers 12 inches wide by 6 inches thick are adopted.

Half-round sleepers, as in Fig. 304, are used on many lines because they are cheaper. In some cases the flat side of the sleeper is placed downwards, and the rail or chair is fastened into an adzed seat cut in the round side; and in the others the round side is placed downwards, and the flat side of the sleeper carries the rail or chair. Triangular sleepers, as in Fig. 305, have also been used, made by cutting the blocks diagonally, so as to obtain the greatest possible width. They were laid with the flat side upwards, and the apex downwards. They were difficult to keep packed, and have not been adopted to any great extent.

With the exception of a limited number of larch and fir sleepers grown in the country, most of the sleepers for our home railways are imported from the Baltic. They are brought over in logs, or blocks, each 8 feet 11 inches long, some square and others circular in section, and when sawn down the middle, each block forms two sleepers.

The preservation of timber from decay is a subject that very early occupied the attention of engineers and all those interested in railways. A railway sleeper is particularly exposed to deterioration the lower portion being surrounded with moist ballast, whilst the top portion is more or less uncovered--two different conditions in the same piece of timber. Several processes have been tried, such as Kyanizing, Burnetizing, Boucherizing, etc., but the system which has given the best results, and is now almost universally adopted, is that known as creosoting. This method consists of forcing liquid creosote, under considerable pressure, into sleepers or railway timbers which have been prepared or dried by ordinary natural seasoning or by special artificial means. Creosote is a dark, oily liquid, distilled 212 from coal tar, varying in its composition according to the quality of the coal from which it is obtained, and ranging in its specific gravity from 11·08 to 10·28.

Creosote oils of light specific gravity were at one time in favour, but experience proved that, to some extent, the light oils were volatile and also soluble in water, and that heavy rains washed out the constituents which were essential for the preservation of the timber. On the other hand, by heating the heavy oils and using high pressure the napthaline which is dissolved only by the heat, is forced into the wood, fills the pores, and solidifies.

Creosote is obtainable in large quantities, at prices varying from twopence to fourpence per gallon, according to the demand and cost of production. Newly delivered sleepers or railway timber contain so much sap or water that it is impossible to force a sufficient quantity of creosote into them until they are properly seasoned or dried.

The seasoning is generally arranged by sawing each block into two sleepers, and then stacking the sleepers on edge in tiers, leaving a space of four or five inches between each of them for a proper circulation of air. The sleepers should then be left for nine to twelve months to season, although more may be necessary in some cases if the blocks were particularly wet at the time they were sawn.

When ready for the process the sleepers are placed in the creosoting cylinder, which is generally about 60 feet long by 6 feet in diameter with semi-spherical ends. One of the ends is fitted with strong hinges and fastenings, and forms the doorway. The sleepers are packed carefully inside, and the doorway made tight. The machinery is then set to work to exhaust the air from the cylinder and allow the creosote to flow in amongst the sleepers. When the cylinder is full the force-pumps are started to force in more creosote up to the pressure prearranged and regulated by the safety-valve, in some cases 100, 110, or 120 lbs. per square inch. The creosote should be heated to 112° or 120° Fah., to dissolve the napthaline and reduce all the component parts to a thoroughly fluid condition.

The success of creosoting depends almost entirely upon the effectual seasoning of the timber. Only a very small quantity of creosote can be forced into wet or unseasoned sleepers, even with the best machinery and exceptionally high pressures, while a thoroughly dry sleeper will 213 readily absorb from 2⅓ to 3 gallons. More could be forced into the dry sleeper if necessary, but a little consideration will show there would be no advantage in doing so. In railway sleepers there are two elements of destruction at work--one the decay of the timber, and the other abrasion or wearing away of the wood itself from the constant pounding of the passing loads.

More particularly does this wearing-away take place with the flange, or bridge, rails, their distributed bearing surface on the sleeper being less than the cast-iron chairs.

A thoroughly well-creosoted 5-inch sleeper laid originally with a thickness of 4-¾ inches in the centre of rail-seat, as in Fig. 306, will wear down 1½ inches, the timber remaining quite sound.

The writer has had to take out thousands of sleepers where the seats of the flange, or bridge, rails had been pounded or worn down so deep into the wood as to leave too small a thickness of timber to carry the rail with safety. These sleepers had to be taken out of the road, not on account of decay, but because they were actually worn down too thin to be of service. They had done their work well for a long series of years, and were perfectly sound when taken out. No increased quantity of creosote would have made them last longer, and any increased quantity of creosote would have been waste.

Two and three quarter gallons of creosote is a very good and suitable quantity for a 10 inch by 5 inch rectangular sleeper, but not more than half this quantity can be forced in if the sleeper is wet or unseasoned.

Sleeper-blocks are generally cut from the upper part of the tree, and do not therefore consist of the best portion of the timber, yet sleepers made from the soft, coarse-grained Baltic wood, properly creosoted, will last from twelve to eighteen years in the line in this country, while uncreosoted they would perish from decay in six or seven. The benefit is great when, by adding from eightpence to a shilling for the cost of creosoting, the life of the sleeper may be doubled or trebled. Of course, there are countries, like the far west of America, where the lines pass through vast forests, and where sleepers may be had for the mere cost of cutting. Creosoting in those places would be out of the question, and would cost four or five times the value of the plain sleeper. It is found, also, that in tropical countries and in dry climates at high altitudes creosote loses its 214 efficiency, and in those districts the best creosoted soft-wood sleeper perishes from a species of dry rot in three or four years. Where wood sleepers have to be used in tropical climates it is better to obtain them from the timber of the district, although in many cases suitable trees are difficult to procure and the cost of land transport is very heavy.

The soft cushion-like effect of a sound, properly packed wooden sleeper contributes so largely to form an easy, smooth-running road, that so long as they can be obtained at a moderate cost, and are fairly durable, wooden sleepers will always be preferred to those of any other material. The great question will be the supply. Creosoting and other wood-preserving processes have done much to prolong the life of sleepers, but the rapidly increasing extent of mileage throughout the world, together with the enormous number of sleepers required annually for maintenance or renewals, must before very long severely tax the powers of supply.

In the great timber-producing territories the axe is often heard, but the planter is rarely seen. Vast forests are cleared away, and their sites transformed into busy towns or cultivated lands; and unless some great change takes place, and planting be carried out on a large scale, some other material will have to be adopted for this important item of our permanent way.

Appearances would indicate that at no very distant date iron or steel will take a conspicuous part in the formation of future railway sleepers.

More than thirty years ago several descriptions of cast-iron sleepers were introduced into notice and tried on some of our leading home railways. Cast-iron was at that time considered more suitable for the purpose than wrought iron, as it was very much less costly in price, and could be readily worked into any desired form or size, with the advantage that the castings would all be duplicates of one another.

Figs. 307 to 313 show some of the types that were designed and laid down in the road. In Fig. 307 the sleeper and chairs were all cast together in one piece; the rail was held in its place by wooden keys, and the gauge of the line was maintained by transverse wrought-iron tie-bars. The sketch represents one of the sleepers used at the rail-joints, and has three chairs, the larger one in the centre being for the support of the ends of the rails. This arrangement was the 216 same as was then in use on the ordinary wood-sleeper road, where an extra large chair was placed at the rail-joints, and was the most approved method for many years before fish-plates were introduced. The intermediate sleepers were shorter, and had only two chairs.

Fig. 308 represents a long, flat, cast-iron sleeper made in two halves, bolted together just below the under side of rail at each of the three chair-seats. The rail was gripped and held in position without the use of wooden keys. This being a joint sleeper, three chairs were used, as in Fig. 307. Only two chairs were used on the intermediate sleepers.

Figs. 309 and 310 are somewhat similar, but the circular one is higher and more cup-shaped than the other of oval form. The oval pattern has two small recesses for holding two small hard-wood cushions. The circular holes shown in the sides of the sleepers were intended to facilitate the packing, or tamping, of the light sandy ballast.

Fig. 311 represents a rectangular cast-iron sleeper, as used for the flange rail. The rail rests on cast-iron cross-ribs, bevelled to give the proper cant, and is held in position by the tie-bar bolt and clip-piece, as shown. The small projecting lug, formed on the under side of sleeper, fits into a corresponding notch in the tie-bar, and keeps the sleepers to gauge. The tie-bar passes through the loop end of the same bolt which secures the rail, and is held up tight against the under side of sleeper.

Figs. 312 and 313, both the same in principle, possessed features which appeared to give great promise. They were simple in construction; the rail was kept well down, and did not come in contact with the cast-iron at any point. The long wooden wedges, which fitted into the rough or serrated sides of the casting, acted as a cushion to the rail, and were intended to sink deeper into the recess as the super-imposed weight increased, or the wood became thinner from shrinkage. In practice, however, it was found that these sleepers were not the success that was anticipated.

It was soon observed that sand and fine particles of gravel from the ballast worked their way into the lower part of the recess, and became so compact as to prevent the wooden wedges working further down to increase their grip on the rail. Even when the recess was kept free and clear of sand, the enormous pressure exerted by the wooden wedges broke the iron at A, although an extra thickness was given to that 217 part of the section. The cast-iron was exposed to the greatest strain at the point where it was the least capable of offering resistance.

Much ingenuity was displayed in many of the patterns brought forward, but in dealing with a hard unyielding material like cast-iron, it is difficult, if not impossible, to impart any soft, elastic effect; and the different systems of cast-iron sleepers failed to become popular on our home railways, on account of the noise and vibration when trains passed over them. Another objection was the great multiplicity of parts required in many of the types, and the constant and severe strain produced on the fastenings on the passing of every wheel. The bolts might be made tight at first, but the incessant shaking would work them loose, the threads became stripped, and the rails ceased to be held in a proper and secure position.

The cast-iron sleeper road was considered unsuitable for the heavy and fast traffic of our home lines, and was ultimately all taken up and replaced with wooden transverse sleepers. At the same time, there is no doubt that cast-iron sleepers have been of great value in India and tropical climates, where timber sleepers were not only scarce, but perish very rapidly. Very large numbers of them have been laid down abroad of patterns very similar to those shown in Figs. 309, 310, and 311, and have done good service for many years. They are not affected by rain or heat, but, unfortunately, being castings, are liable to considerable annual loss from breakage.

Improvements in plate-rolling machinery, and in appliances for bending and stamping wrought-iron, have materially assisted in developing the introduction of wrought-iron and steel sleepers. Cast-iron and wrought-iron are, in the abstract, hard and non-elastic as compared with wood; but whereas cast-iron can only be made into fixed, unyielding shapes, wrought-iron and steel can be worked into forms that possess a certain spring-like effect, which not only enables them almost entirely to resist fracture, but also imparts a measure of elasticity to the permanent way.

The simplest form of wrought-iron sleeper would be a plain, flat plate, to which the chair, or rail-bracket, would be attached; but as this form would have bearing surface only, without any lateral hold on the ballast to keep the rails to line, it could not be adopted.

During the last few years very many types of wrought-iron and steel 218 sleepers have been introduced, and nearly all of them of the transverse-sleeper pattern, formed out of rolled plates; the sides, and in some cases the ends also, are bent, or turned down to obtain a hold in the ballast. Where bull-head or double-head rails are used, cast-iron chairs, or wrought-iron bracket chairs, are bolted, or otherwise secured to the upper surface of the sleeper, a layer of felt, tarred paper, or other soft material being placed between the two metal surfaces. Where flange rails are used, they are fastened to the sleepers either by bolts, clamps, or clips raised up out of the iron sleeper, and bent over to hold tightening keys. Rolled transverse sleepers can readily be bent, or set in the centre to give the proper cant at the rail-seat; and in some types the sleepers are pressed in the machines, so as to be narrower towards the centre, and with a deeper turnover, to obtain increased stiffness.

In Figs. 314 to 319 are shown some of the patterns which have been brought out, laid down in actual practice, and in use at the present time.

From the fact that wrought-iron and steel sleepers have been laid down in so many places where cast-iron sleepers were discarded or refused a trial, it is evident that the former are considered to have qualities which the latter did not possess. Rolled iron or steel sleepers are coming more and more into use, especially on foreign or colonial railways. So long, however, as good, well-creosoted timber sleepers can be obtained for our home railways at prices from 3_s._ 8_d._ to 4_s._ 8_d._ each, and last from fourteen to twenty years, there is little probability that they will be supplanted by iron sleepers at double the cost. But abroad the circumstances of cost and durability are different, and there the rolled iron or steel sleepers, which will outlive two or three sets of wooden ones, must claim advantages which cannot be overlooked. The difficulty will be in the fastenings, the mode of attaching the rails to the sleepers. The constant hammering of metal upon metal, resulting from the vibrations of every passing load, will quickly wear or loosen bolts, rivets, or wedges, and the fastenings which will prove the most efficient will be those that are the simplest and most readily adjusted.

Fastenings.--Figs. 320 to 335 illustrate some types of the principal fastenings used in connection with the chair road, and with flat-bottomed or flange rails.

The fish-bolts, Figs. 320 and 321, are of a form which is in very 220 general use both for steel bull-head rails and steel flange rails. By making the neck square or pear-shaped, to fit into corresponding hole in the fish-plate, the bolt is prevented from turning round when the wrench or spanner is applied to tighten the nut. A channel or groove is sometimes rolled on the outside of fish-plate to grip bolts made with square heads. Some engineers adopt two nuts, others prefer one nut of extra depth. Washers are used in some cases, but are not universal. With a deep rail it is preferable to place the nuts inside, so that the platelayer inspecting his length can see both rows of nuts as he walks along between the rails. With shallow rails the nuts must be placed outside and the cup-heads inside, to give ample clearance to the wheel-flanges.

Fish-bolts are subject to very severe work. Heavy rolling loads passing over the rail-joints--frequently at very high speeds--bring into play all the gripping power of the fish-bolts to maintain a firm support of the fish-plates to ends of rails, and the constant action of pressure and release produces a loosening or unscrewing motion in the bolts which is very difficult to counteract. Loose fish-bolts cause clattering joints and uneven road, and unless promptly remedied, the screw threads are soon destroyed and bolts rendered useless. Many devices have been invented to prevent or check this loosening of the bolts; one of the methods, and a very simple one, consists of a plain steel bolt with a steel lock-nut, made as shown in Fig. 322. As will be seen from the section, one-half of the nut is tapped of the same size as the bolt, and the remainder with deep-locking threads. The first half of the nut is readily screwed on to the bolt, but considerable force must be exerted to screw on the portion having the deep-locking threads; practically the second half of the nut has to cut a new or deeper thread for itself when screwing round the bolt.

The slits or grooves at the angles of the nuts form four distinct cutting edges for shaping the deep threads. As the upper part of the lock-nut is divided by the grooves into four separate or detached segments, these segments will be forced slightly open or outwards during the action of cutting the deep thread on the bolt, and from their natural tendency to return to their original position they must exercise a strong gripping power on the bolt. This combined operation of cutting the deep threads and of forcing open the upper or detached 222 segments, give an enormous holding and retaining power to the lock-nut, and enables it to withstand the train vibrations for a very long time without any perceptible slackening. In case of line repairs the nut can be readily unscrewed, and taken off the bolt.

Round iron spikes, as in Figs. 323 and 324, and round wooden trenails, as in Fig. 325, are both used for fastening cast-iron chairs to the sleepers. The spikes are made with a slightly taper neck, of size rather less than the hole in the chair, to avoid risk of breaking the casting when driving the spike down. Trenails are made out of well-seasoned hard wood, and are compressed by machinery. When driven into the sleeper, they expand by exposure to the atmosphere, and hold the chair very securely in position; but being only wood and of very small scantling, they are subject to early decay. The head, which is the only part in sight, may be perfectly sound, while the part between the chair-seat and top of sleeper may be quite rotten and useless. It would be very risky to depend upon trenails alone; one spike at least should be used to every chair. In some cases an extra large trenail is used with an augur-hole down the centre, through which either an iron spike is driven or a bolt is passed and screwed into a crab-nut on the under side of the sleeper. This arrangement will work well for a time, but there will be a great deal of play in the spike or bolt when the trenail becomes much decayed.

The spikes represented in Figs. 326, 327, and 328, are much used with flange rails. They are square in section, and finished with either blunt or sharp points, as shown. The top of spike is made with a doghead and side-lugs to facilitate the easing or withdrawal when necessary for renewals of sleepers, or alterations in line. By inserting the curved double claw end of a platelayers’ crowbar, the spike can be raised without injuring the sleeper; but if it is required to be driven into the same sleeper again, a new hole must be bored, as the old hole will be too slack to be of any service. Augur-holes must be bored in the sleepers for the above spikes. For new roads, these holes can be bored by machinery when cutting the grooves for rail-seats; but when carrying out alterations or repairs, a large number of spike-holes must be bored by hand-augurs, an operation both slow and laborious. With the hand-boring there is the danger that the hole may not be made deep enough, owing to the workman’s 223 endeavour to avoid damaging the point of his augur by forcing it entirely through the sleeper, and bringing it in contact with a stone. Augur-holes bored wide to gauge will remain out of gauge, and although the spike may be driven down firm in its position, a space will be left for play between the rail-flange and spike.

Fig. 329 is a sketch of a dog-spike for flange rails which the writer has used for many years both abroad and at home, and which can be driven without any boring at all. The back of this spike is made perfectly straight, half of the front side is made parallel to the back, and the remainder is tapered down to a chisel point not exceeding 1/16 of an inch thick, the entering edge on the face being narrowed down to 3/8 of an inch in width. Three jags or spurs are cut on each side of the tapered portion, or twelve in all, and add greatly to the holding power. Not only can this spike be driven without any boring, but it possesses the additional advantage that in driving it down its taper or wedge-like shape causes it to drift hard up to the edge of the flange of rail, an element of great value in securing the exact gauge of line. With these spikes permanent-way laying can be carried on very rapidly, and they are especially valuable when making alterations, as augurs for spike-boring can be dispensed with altogether.

Wood screws with square heads similar to Fig. 330 are sometimes used for fastening flange rails to wooden sleepers. They are passed through holes punched or drilled in the flanges of the rails, and are intended to preserve the gauge as well as secure the rails to the sleepers. Experience has shown that these wood screws possess very limited holding power. The screwed portion of the bolt cuts but a very imperfect and weak holding thread in the soft wood of an ordinary sleeper, moisture insinuates itself into the bolt-hole, rusting the bolts and decaying the surrounding timber, and in a very short time the bolts become loose and incapable of holding the rail down firmly. As permanent-way fastenings wood screws are very inferior to crab bolts.

Crab bolts, as in Fig. 331, may be made either with square or hexagonal heads, and with three spur-nuts or four spur-nuts, as in A or B. The length of the bolts will depend upon the thickness of the sleeper or timber-work through which they have to be inserted. The bolt is pushed down through the hole bored in the sleeper, and the crab-nut put on from underneath. With a few turns of the bolt, the 224 crab-nut is brought close up to the under side of the sleeper, the spur-points become embedded in the wood, and hold the nut firmly in position during subsequent tightening of the bolt. Crab bolts are extensively used with flange or flat-bottomed rails, and also in switch chairs and in crossings. A large number of flange rails are used with one hole through the flange at each end of rail, and a crab bolt passed through the hole and through the sleeper next to the joint, as shown in Fig. 332. This system checks the creeping of the rails by effectually securing or anchoring each rail to two of the sleepers. As there is always a tendency for these rails to crack through to the outside at the flange-holes, it is very desirable to have as few holes as possible. The two above described will be found sufficient for all practical purposes. To avoid punching or drilling more holes in the flanges of the rails, additional or intermediate crab bolts can be used by means of the fang clips shown on Fig. 333. The crab bolt is passed through the fang clips and through the sleeper close up to the flange of rail, and by screwing it round in the crab-nut under the sleeper the fang-clip is pressed down until the two spurs are driven into the timber, and the rail held securely in its place and to gauge. Intermediate crab-nuts and fang-clips should always be used in pairs, one on each side of the rail. Possessing more holding-down power than ordinary spikes, they are particularly valuable on sharp curves.

In some cases flange rails are laid in small cast-iron saddles, or chairs, as shown in Fig. 334, one end of the rail-seat having a recess to prevent the rail tilting upwards and outwards. An ordinary spike may be used for the inside end of chair, and a crab bolt with bent washer for the other. Unless the fastenings can be kept always tight, the above arrangement makes a very noisy, clattering road, as there are so many metal surfaces in contact, and so little to deaden the vibration. For narrow flange rails carrying heavy rolling load, chairs may be necessary to increase the bearing surface on the sleeper, but with rails having flanges five inches wide and upwards, it is better to let the flange rest direct on the wood of a properly grooved sleeper, and thus obtain a smoother and less noisy road.

On exceptionally sharp curves, wrought-iron or steel tie-bars, as in Fig. 335, are sometimes used to maintain the line to gauge. They may be made out of bars 3 inches wide by ½ an inch thick, turned over at 225 the ends to grip the outside flanges. Being made to exact template, they have to be threaded on to the rails before spiking down, and are placed between the sleepers at distances from 7 to 10 feet apart.

Laying Permanent Way.--To preserve a good line and level to the permanent way, it is absolutely necessary that the road-bed should be kept thoroughly drained. If provision be not made for quickly carrying away the rain-water, and if it be allowed to collect under and around the sleepers, the action of the passing trains will work the finer particles of the packing into the consistency of soft mud, which will be gradually squeezed away, leaving the sleepers imperfectly supported and insecure. A loose sleeper involves a depression in the rails, and a corresponding lurch in the vehicles of the train, and a series of these depressions may produce such an oscillation in the train as to cause it to leave the rails.

The height or space from formation-level to rail-level is generally about 1 foot 9 inches for a flange railroad, and about 2 feet for a chair railroad.

Figs. 336 and 337 show cross-sections of both descriptions of road as laid down for a double line in cutting. The same arrangement applies to similar roads laid down in embankment, merely omitting the side-drains or water-tables. The bottom layer of ballast or road-bed should consist of good hard, quarried, or broken stones, each 6 inches deep, set on edge, firmly and closely hand-packed, forming a foundation through which the rain-water can be quickly carried away. On the top of this bottom pitching should be placed a 6-inch layer of broken stone ballast or strong clean gravel, of which none of the stones should be larger than will pass through a 2-inch ring. When the sleepers and rails have been laid on this second layer, and properly packed to line and level, the top ballasting, or boxing, of either broken stones or strong clean gravel, should be filled in to the form and extent specified. Where broken stones are used for the top ballasting none of them should be larger than will pass through a 1½-inch ring.

Broken stone ballast should only be made from the hardest and soundest description of rock or boulders, so that, however small the particles, they will remain sharp and clean.

There are many kinds of rock which appear hard and compact when first excavated, but upon exposure to the weather undergo a complete change, 227 developing into soft masses containing too much clay to allow the water to pass through readily. Where rock is scarce and gravel plentiful, the lower layer may be made of the heavier or coarser gravel, leaving the finer gravel for the upper layer, or boxing; but there is no doubt that the broken stone pitching makes the most efficient bottom layer. No gravel ballast should be used which is not free from clay or earthy sand.

Wherever there are particles of earthy matter, sufficient to furnish nourishment for vegetable growth, weeds will quickly spring up, and once established are most difficult, if not impossible, to eradicate. The presence of weeds checks drainage, and gives an untidy appearance to the line, besides constantly occupying a large portion of the platelayers’ time in their removal.

Clean cinders, free from dust or earth, are much used for upper ballast and boxing, and being lighter than gravel, are specially applicable for soft boggy ground. Burnt clay, broken into small pieces, has been largely adopted in districts where both rock and gravel were difficult to obtain. Chalk, furnace-slag broken small, crushed brick and sand, are frequently used as ballast. Sand is objectionable where there is high-speed traffic, as the finer particles rise in the form of dust and deposit themselves on the vehicles and machinery of the train.

The water-tables, or side drains in the cuttings, should be cut below the formation level, and to a depth or width sufficient to take away all rain-water, or water arising from springs. Where the material of the cutting is of a loose friable nature, it may be necessary to protect the sides of the water-tables with low dry stone walls, as in Fig. 338; or glazed earthenware pipes may be laid, as in Fig. 339, with open joints, or with grate openings at regular intervals. In some cases substantial side-walls and invert are requisite to carry away the flow of water.

Timber sleepers intended for the flange railroad should have the rail-seats grooved by machinery to ensure perfect accuracy in the position of the grooves, and in the angle or inclination of the rail-seats. Fig. 340 is a side view of part of a sleeper grooved to receive a flange rail. The presence of the grooves materially facilitate the laying of the rails to gauge, but must not be allowed to interfere with the constant use of the platelayer’s gauge. In a similar manner the timber sleepers for the chair road frequently have 228 the spike-holes bored to template by machinery, as indicated on Fig. 341. Steel or iron sleepers are delivered with the recesses for rails, and holes for bolts or fastenings formed complete by machinery.

The distances apart of the sleepers will be regulated in a great measure by the weight of the rails and the description of the traffic. Where light rails are intended to carry heavy engines the sleepers must be laid closer together than would be necessary for heavy rails. The joint being the weakest part of the rail, it is usual to put the sleepers closer together at that place, with a view to gain additional support, to assist the fish-plates in preserving as much as possible a firm unyielding surface at the rail-joint.

Fig. 343 shows an arrangement of sleepering largely adopted for steel flange rails 26 feet long, and weighing 79 lbs. per yard. The length of a rail is more a question of convenience of handling, facility of transhipment, and general use, than of actual manufacture. There is no difficulty in rolling rails up to 50 feet in length, or more; but very long rails are extremely ungainly things to move about, and are more exposed to receive permanent bends or kinks in unloading, besides requiring greater spaces at the joints to allow for contraction and expansion.

Fig. 344 is an example of sleepering for a chair railroad, for steel bull-head rails 26 feet long, and weighing 85 pounds per yard.

Line stakes and level pegs must be put in at suitable distances to guide the platelayers in laying the rails to the correct line and level, and on the curves the proper amount must be marked off for the super-elevation of the outer rail.

When the second layer of ballast has been spread for its full width and depth the sleepers can be distributed, and the rails or chairs spiked down to the correct gauge. Before putting on the fish-plates spaces must be left at the ends of the rails to allow for contraction and expansion, the amount depending upon the temperature at the time of laying down the rails. As the rails will expand, or increase in length, with the heat, it is necessary to allow more space for expansion for rails laid down in the cold, or winter months. On our home railways rails are very rarely laid down when the temperature is lower than 25° F., or higher than 125° F., and this range of 100° may be considered as covering all the variations likely to occur in ordinary practice. The greater portion of the permanent-way laying is 229 carried on when the temperature is between 40° and 75°. The results of very carefully conducted experiments show that an increase of temperature of 1° F. will cause an iron or steel bar, or rail, to expand or lengthen to the extent of seven one-millionths of its length. Working this out for a range of 100° F. would give an increase in length of seven hundred one-millionths, which would be equal to an extension of 0·2184 of an inch in a 26-foot rail. For our home railways, therefore, a space of 5/16 of an inch will be found amply sufficient to meet the variations in length between the extremes of winter and summer, for a rail from 26 feet to 30 feet in length. Too much allowance for expansion is detrimental to the rails, because where the spaces are excessively large the wheels drop into the hollow and hammer or spread the ends of the rails.

The fish-bolts should not be completely tightened up until the permanent way is thoroughly set, and packed to its finished line and level.

On straight line the rail-joints should be laid square and opposite to each other. Permanent-way laying with broken joints is rarely adopted, except on curves or station-yards.

On curves the joints of the inner rails gain on the joints of the outer rails to the extent of--

radius + gauge
-------------- × length of rail.
radius

The amount of this gain, or lead, is adjusted by cutting off a portion of the end of the inner rails at certain intervals.

Assuming the fish-bolt holes to be spaced as shown on Fig. 342, then, when the inner rail is leading to the extent of 2 inches, a piece 4 inches long is cut off, as shown by dotted lines, leaving the original second fish-bolt hole to serve as first or end fish-bolt hole, and a new or second bolt-hole is drilled by hand at A. This method sets back the joint 2 inches from the square, and the lead is allowed to go on again until it becomes necessary to cut off another piece of 4 inches. Another mode is to have a proportion of the rails rolled 2 or 3 inches shorter for use on the curves.

On curves of a 1000 feet radius and upwards, the rails should be laid to the normal gauge, but on curves of lesser radius the gauge may be slightly increased, and as much as ¾ of an inch allowed on a curve of 500 feet radius.

The amount of cant, or super-elevation, to be given to the outer rail 230 on curves must be regulated by the speed of the train and the gauge of the line. Many formulæ have been compiled to determine the necessary amount of super-elevation, but experience has shown that by some of them the calculated amounts were excessive. Possibly during past years too much cant has been given in many cases. The following simple formula approaches very closely to practical experience--

(velocity in miles per hour)^2 × gauge in feet {the super-elevation
----------------------------------------------- ={ of outer rail
radius in feet × 1·25 { in inches.

For high-speed trains uniformity of cant is of the utmost importance, more so even than the exact amount. Any irregularity in the super-elevation of the outer rail, sometimes high and sometimes low, will produce a dangerous swaying movement in the train, which, if not promptly checked, would lead to derailment.

More injury is done to curves by spreading, arising from rigid wheel-bases of engines and tenders, than from any want of counteraction to centrifugal force.

When a long length of permanent way has been linked in, rails spiked to gauge, and fish-plates bolted together, the platelayers can proceed to the final adjustment to line and level in accordance with the stakes and pegs provided for their guidance. The setting to exact line is effected by means of long pointed round iron crowbars, which are struck forcibly into the ballast alongside the rails, and serve as powerful hand-levers to pull or push the rails to the right or left as directed by the foreman standing some distance back at one of the line-stakes. The men with the crowbars pass from rail-length to rail-length, until a long stretch of road has been pulled into correct line.

The adjustment to rail-level is done by first packing up the sleepers to the correct height at the various level-pegs, and then packing up the intermediate sleepers so that the surface of the top of the rails forms one uniform even line from level-peg to level-peg. On new lines it is usual to pack a little high in the first instance to allow for the subsidence or compression which invariably takes place on the passage of heavy trains over fresh ballast.

The form or contour line of the top ballast will vary according to 231 circumstances. In station-yards it is usual to fill in the ballast almost up to the level of the top of the rails for the convenience and safety of the men who are constantly moving about marshalling the carriages and waggons. Out on the open line between stations, the ballast on some railways is filled in up to rail-level, while on others it is only filled in up to the tops of the sleepers, leaving the rails and chairs quite clear of the ballast. On others, again, the ballast is filled well up to the rails and channelled in the centre, as shown on the sketches Figs. 336 and 337. Channelling the centre of the road reduces the quantity of ballast per mile, ensures good drainage, and also stability by not permitting any central support to the sleepers. By covering up the lower table and sides of rails the noise is reduced to a minimum, vibration is absorbed, and a more silent road is the result. The contact with the ballast also preserves the rail from the extremes of temperature. Where the ballasting is not channelled there is some risk of the sleepers breaking in the middle. The constant packing of the sleepers just under the rails has a tendency to drift some of the ballast inwards towards the middle of the sleeper, forming a hard compact mass, and this mass, acting as fulcrum, throws considerable strain on the middle of the sleeper when the trains pass over and depress the ends. Where the ballast is filled in level with the rails on top of sleepers it should be loosened occasionally in the middle to prevent it becoming too hard.

Connections with the rails of the main line will have to be made in various forms to suit the circumstances of the joining lines or sidings.

Fig. 345 shows a simple double-line junction.

Fig. 346 shows an example of what is termed a _flying junction_, or a junction of two double lines arranged in such a manner as to cause the least interruption to a constant train traffic passing UP and DOWN over both lines. Upon referring to Fig. 345 it will be seen that a train from F, turning off at the points E and proceeding to G, must block, or close for traffic the section ABC during its passage over that line towards G. With a crowded train-service the blocking of both UP and DOWN main lines for the working of one train would cause much interruption, and to obviate such delay the _flying junction_ is substituted. Fig. 346 shows how a train from F is turned off at the points J and proceeds on to K, where by means of a bridge it passes either over or under both main lines, and 233 continues on to G without in any way interfering with the train service on ABC.

Fig. 347 is an ordinary plain siding or _turn-out_, including the necessary throw-off or trap-points and short dead end.

Fig. 348 is an ordinary _cross-over road_ from DOWN main line to UP main line, and _vice versâ_.

Fig. 349 is a double cross-over road, generally termed a _scissors_ cross-over.

Fig. 350 is a simple through cross-over road from DOWN main line to siding alongside UP main line.

Fig. 351 is a similar arrangement of through cross-over road with the addition of a pair of slip points at S to make a connection with the UP main line, thus combining the facilities of the ordinary cross-over and through cross-over road.

Fig. 352 shows a set of three throw-switches with all the sliding tongues placed side by side; and Fig. 353 shows another arrangement of three throws with the sliding-rails of the second set of switches placed just behind the heel of the first set of switches. The latter method works very well where there is sufficient length for the purpose.

Fig. 354 shows a square crossing, where one line of railway crosses another line of railway on the same level.

Fig. 355 shows a connection with a siding by means of an ordinary carriage or waggon turn-table.

Fig. 356 shows a set of “runaway” points which are sometimes placed in the main line at the top of an incline close to a station, the object being to intercept or throw off any portion of a train which may have become detached, and which would, if unchecked, run away back down the incline. By means of a weighted lever or spring the points are set to the normal position of _open_ to the siding, and as they are “trailing” points for the running road they are readily closed by a passing train. One or other of the above forms of connections, or a combination of them, will meet all the requirements which usually occur in railway work.

Fig. 357 is an enlarged sketch of an ordinary cross-over road, and Fig. 358 of a double or _scissors_ cross-over.

Fig. 359 shows a _single-slip_ point connection, and Fig. 360 a _double-slip_ point connection. In places where slip connections can be introduced they add greatly to the facilities for train movements 235 without curtailing the available standing-room for vehicles on the lines and sidings. They are simple in construction, do not require crossings, and in many cases save a complete cross-over road. At the same time slip connections can only be laid down where the angle of the intersecting lines is sufficiently flat to admit of a connecting curve of workable radius.

Fig. 361 is an enlarged sketch of a set of ordinary 15-foot switches or points. By placing them about the middle of the stock rails the joints of the latter are kept well beyond the sliding rails, and the road is held firmly together. It is necessary to place the sleepers closer together at the switches to allow for the reduction in section of the sliding rails, which results from planing them down to the requisite shape. By substituting two long timbers for the ordinary sleepers at the points of the switch rails, as shown on the sketch, a more efficient support is obtained for the switch-box or crank in the case of rod-worked switches, and the working distance from the rails is accurately maintained, irrespective of any packing or pulling of the road. In the sketch a steel bull-head rail is shown on one side, and a steel flange rail on the other, each bolted to an ordinary cast-iron switch chair. Switch chairs are sometimes made of plates of wrought-iron or steel, forged to the correct shape, and riveted together. They are, however, much more costly than cast-iron chairs, and deteriorate more quickly from corrosion.

Fig. 362 is an enlarged sketch of an ordinary crossing similar to the one indicated at C (Fig. 359), and composed of a cast-steel reversible block. The ends and lugs, L, L, are formed to suit the connecting rails and fish-plates, as shown in the cross-sections. The casting is secured to the crossing timbers by bolts passing through the side lugs, S, a cast-iron packing-washer, W, being placed between the lug and the timber to ensure a solid seat and avoid rocking. A very important point in the construction of these block crossings is to have the groove or flange-path sufficiently deep to prevent the striking or touching of the flange of a much-worn tyre. A well-made, carefully annealed steel-block reversible crossing is very smooth in the road, and has a long life. It is all in one solid piece; there are no parts to work loose or spread; the wear of the running surface is very uniform, and when the one side is much worn down, there is the other ready for service. The writer has had many of these steel-block reversible crossings in use under heavy and fast traffic 237 for six and eight years without turning.

Fig. 363 shows an ordinary crossing made of steel bull-head rails secured in strong cast-iron chairs; and Fig. 364 is a similar crossing made of steel flange rails. In some cases the two rails forming the V are welded together at the point B, and in others they are riveted or bolted together. Fig. 365 shows a diamond or through crossing similar to the one indicated at D, Fig. 359, made of steel bull-head rails and chairs.

Crossings are constructed in a variety of forms, whether on the principle of the cast-steel block, or made out of ordinary steel rails; and the above sketches merely illustrate some well-recognized types which experience has proved to be efficient and durable in the road. The angles of the crossings will depend upon the divergence of the intersecting lines to be connected; ordinary crossings, to the angle of 1 in 10, work in for very general use in station-yards, but many are required of angles varying from 1 in 6 to 1 in 14, and in some cases 1 in 16.

As a rule, engineers endeavour as far as possible to avoid using ordinary crossings flatter than 1 in 12, or diamond crossings flatter than 1 in 9, because the gap between the running rails becomes very considerable beyond those angles. At the same time, there are many cases of ordinary crossings of 1 in 16, and diamond crossings of 1 in 12 and 1 in 13 laid down in exceptional places, and which have carried heavy and fast traffic for many years. All crossings should be well protected with wing rails and guard rails, as shown on the sketches.

Fig. 366 illustrates a method of bringing the UP and down lines of a double line of railway close to each other, and passing them over a single-line opening bridge, or a bridge where the works for the second line have not been completed. This arrangement avoids the necessity of any switches, and prevents any accidents which would arise from a misplaced switch. Each set of trains is effectually kept to its own line of rails. With proper signalling or pilot working, the double-line traffic can be worked over the single-line bridge without difficulty. The writer has adopted the above arrangement in many cases when renewing double-line bridges or viaducts where the width for traffic working has been restricted to half of the bridge.

In some instances the same system has been extended to the carrying of 239 four lines of rails over a double-line bridge, as shown on Fig. 367.