182 



THE CIVIL ENGINEER AND ARCHITECT'S JOURNAL. 



[June, 



solid ami stable, neitlior retain tlie mail in {;au(;e, nor secure the 

 correct continuous elevation of the dirt'erent points of support in 

 the same line of rail; and as from this circumstance their use is 

 chiefly confined to cutting-s where the substratum is hard, and the 

 ballast Rood, their hardness gives a peculiar harsh and grating 

 feeling to the carriages passing over them. 



The situations now are comparatively few in which stone blocks 

 can be procured for cost to surpass the wooden sleepers, especially 

 when the labour of jumping and plugging the holes for the cliair 

 pins are taken into account. It is liowever to be remarked, that 

 in cases wher; horse power is used they have the advantage of 

 leaving a clear way for the horses' feet. 



Cross sleepers, which have, on the narrow-gauge lines, been so 

 extensively adopted, whilst presenting a very sufficient bearing 

 surface on the ballast, unite in themselves a cross tie to preserve 

 at every ])oint of support uniformity of gauge, and are readily 

 packed and adjusted. Triangular-sectioned sleepers present theo- 

 retically a very large bearing surface to the ballast to resist down- 

 ward pressure; it is however doubtful whether, except in ballast 

 of a very firm and binding character, this eftect is got from them, 

 as in coarse and open ballast the sharp edge of the sleeper has a 

 tendency to work downwards into the ballast with the motion of 

 the trains, and to cant with the driving forward of the rails, a 

 defect to which all cross sleepers are more or less liable. These 

 last sleepers ha\e the great advantage of being surface packed, so 

 that repairs can be eifected without the removal of a large quan- 

 tity of ballast. 



For side stiifness between the points of support both with stone 

 blocks and the various kinds of cross sleepers the lateral strength 

 of the rail is alone depended on; and in this respect the double T 

 rails so extensively used seem ojien to some objection, for to 

 their deflection sideways with an engine slack in gauge, or tra- 

 velling at a high velocity, may l)e attributed much of the side 

 oscillation so observable at times, which a variation of speed will 

 often check. 



The writer is aware that it has not been usual to attach much 

 importance to tliis question; and in Professor Barlow's valuable 

 work, in commenting on the small amount of side deflection as in- 

 dicateil by experiments in what were then considered most unfa- 

 vourable circumstances, Professor Barlow says (p. 421) : — 



"Tlje whole ol these experiments" (on the lateral deflection of railway 

 bars) " have a tendency to show that the stress which the bars have to sus- 

 tain in tills direction is not such as to require to be more amply provided for 

 than the iiicreabed thickness the bar must have to meet the grea'er vertical 

 strain due to a longer bearing. In other words, the additional strength given 

 to the bar for the purpose of resisting the vertical strain will be amply suffi- 

 cient to meet and resist the lateral strain." 



That is to say, the rails then experimented on were deemed 

 strong enough laterally, and it was held that further increase of 

 strength vertically, necessary for a longer bearing, or we may add, 

 to su]iport heavier loads, would suffice to impart the requisite 

 lateral strength to the rails. 



The weight of the engines since Professor Barlow conducted 

 these experiments has been doubled and even trebled, the weight 

 on the driving-wheels more than doubled, and the speed, no unim- 

 portant element in producing side oscillation, has been almost con- 

 stantly doubled, and on special occasions quadru]iled; the weight 

 of the rails has crept up fnmi 45 or 50 lb. per yard to 75, 80, 90, 

 and even 100 lb. per yard, hut the side stiffness has by no means 

 pnqxirtionally increased. In cases where even 70 lb. rails have 

 been tried with long bearings, the side oscillation has been found 

 so constant and violent as to necessitate a recurrence to the 

 shorter bearings most in use, from 3 feet to .'i ft. 6 in. 



'I"he longitudinal hearers have the advantage of presenting a 

 continuous hearing surface to the ballast, and of giving with the 

 rail great and unifornr side stiffness to each line of rail, so that 

 conqiaratively few cross ties are needed to keep the line in gauge, 

 tliose on the broad-gauge lines being 15 feet apart. To these two 

 main features, continuity of bearing on ballast, and continuity and 

 amount of side stifliiess, are to be attributed the great ease and 

 evenness of tlie motion of the engines and trains on the Great 

 A\'estern Railway. 



This system of construction is open to the following objections: 

 — The expansion and contraction of the rails tends to loosen the 

 fastenings, especially at the joints; and from this cause, with the 

 comparatively complex nature of the cross ties, the maintenance 

 is more expensive than on ordinary roads laid on cross sleepers, 

 and it seems difficult to lift this road without great care and atten- 

 tion. When a cross sleeper is used under the longitudinal bearer, 

 securing at the same time correctness of gauge and of the cant of 

 the rail, this objection vanishes; and the cross sleeper being raised 



and packed, the longitudinal timber may he packed subsequently. 

 This last construction, as used on the old Croydon line, stood a 

 very large amount of traffic, although laid on substratum of a very 

 inferior character. 



AV'here a bridge rail is laid down directly on cross sleepers, the 

 rail undoubtedly possesses in itself considerable side stiffness be- 

 tween the points of support; it does not however contain so much 

 vertical strength to resist deflection as the ordinary double r rail, 

 and the expansion and contraction of the rail is apt in cases to 

 split the joint sleepers. The proportions in which, in the Great 

 Southern and Western of Ireland, the bearing surface on the bal- 

 last is varied (from 2'50 feet to 0-93 feet super per foot run of 

 rail), give so great a preponderance to the joint, that it may be 

 doubted whether in practice it will not he found that constant 

 packing is required in the centre of the rail lengthwise; and the 

 amount of bearing surface of the rail on the sleepers is so small 

 that there will he much crushing of the rail into the timber, espe- 

 cially on the curves. 



Having made these general remarks on the various peculiarities 

 of some of the leading modes of constructing permanent way, it 

 remains for us to consider the shape, strength, and quality of the 

 materials used, to prevent them from crushing in themselves. 

 This more immediately applies to the rails, or wearing surface of 

 the permanent way. It is in this respect that most of the systems 

 have alike suffered since the introduction of heavy engines. 



The rails most in use vary from 2j to 2,^ inches in width on the 

 upper table or wearing surface, and are for the most jiart made 

 rounding at the top. Now if we look at the line of contact of a 

 tyre on a rail, it will be found that a comparatively small portion 

 of the width of the rail, in favourable cases not more than ] j inch, 

 and in some instances less than g inch, is in actual contact. Now 

 if it be assumed that on a 3 ft. 6 in. wheel of an engine in work- 

 ing order, a weight of 6 tons has to be carried, and if the strength 

 of the iron in large railway bars to resist compression be taken at 

 8 tons per inch — (and it is doubtful whether more may be taken) 

 — then the line of contact of the tyre on the rail in section being 

 f inch, it is obvious that such line of contact will have to be ex- 

 tended in the other direction into a surface of 1 inch, before the 

 surface of the rail in contact with the tyre becomes sufficient to 

 resist the weight superimposed, and the amount of compression in 

 the rail will be represented by the versed sine of the chord of an 

 arc 1 inch long with a radius of 2 ft. 9 in. The limit of compres- 

 sion of iron, such as is used in railway-bars being determined, it 

 is evident that the amount of bearing surface between the rail and 

 tyre will vary directly with the weight superimposed; that its ex- 

 tent in the length of the rail (or the length of the circumference 

 of the wheel in contact), will vary with the length of the line of 

 contact in section; and the extent of the compression or length of 

 the versed sine, will vary with the radius of the wheel. 



The amount of this permanent compression, or of the motion pro- 

 duced in the particles of the iron beyond the elastic limit, even 

 supposing all the compression to take place on the rail and none 

 on the tyres of the wheels, will evidently be inflnitelv small; but 

 it may be fairly argued that such motion does take place, and re- 

 newed from time to time, from infinitesimal and insensible, becomes 

 palpable and evident in its results. 



It is difficult on other grounds to account for the rapid deterio- 

 ration of rails, — the word deterioration being used in contradis- 

 tinction to destruction, as the rails now removed on some of the 

 leading lines have in many cases lost more than 2 lb. per yard of 

 their original weight, showing that although rendered useless, 

 they have not given out a fair amount of wear to the companies. 

 For this information the writer is indebted to Mr. Dockray, and 

 would take this opportunity of acknowledging the kind courtesy 

 of that gentleman in permitting access to his very valuable report 

 on this subject. 



The cause of the removal of rails when not thoroughly worn 

 out, is their becoming distorted in sliape, such distortion being the 

 result of lamination. Now this effect may be produced either 

 from defective shape, or want of strength in the material itself to 

 bear the superincumbent load. In reg. rd to the usual T-headed 

 rail, the impression very generally prevails that the shape is in 

 fault, and it may very readily be imagined that a rail of this make, 

 shall gradually grow distorted, from the pressure bending down 

 the overhanging portions of tlie top table, without of necessity 

 proving any motion to have taken place from absolute crushing. 

 But bridge rails are found not to wear uniformly down, (as they 

 should do if no crushing took place); the upper corners of the 

 rail turn outwards, and when the wearing part of the rail has been 

 rolled or crushed out sideways, the centre part of the top is driven 

 downwards, and the sides turned completely over. 



