Within the limits of an occasional paper it would be impossible, and if possible, excessively wearisome, to give an exhaustive view of the whole subject. For the sake of your patience I will condense as much as possible the many considerations involved.

First, to explain why I prefer to use the Americanism "Track," rather than the usual phrase "Permanent Way." The latter was originally used as distinctive from the "temporary roads" laid by the contractors in building the line, and not as expressive of a positive quality of permanence in the road. Because under any circumstances this requires to be renewed at intervals, and cannot, therefore, be said to be "permanent" in the same sense as a bridge or culvert. So that "track" better expresses the nature of the thing, and it is shorter.

Before we can form an opinion on the relative merits of different systems, we must have clearly before us the essential conditions to be fulfilled by any track. The simple statement of the work to be done is, that trains are to be carried. But there are very important variations in the degree of that work. According to the nature of the traffic, the trains may be light or heavy; may have to run at a low or at a high speed. For light trains at a low speed it matters little, so far as safety merely is concerned, what kind of track be used. For heavy trains at a low speed, we have only to see that we have sufficient bearing-surface and sufficient strength of rail and fastenings. The true final test, then, of a good system of track is, that it will carry heavy trains at high speed with assured safety. It is obvious that a system which will bear this test may be adapted for all gradations of traffic, down to a minimum, by merely reducing the dimensions, or sections, or strengths, which are different expressions for the same thing, of the various members. What, then, are the conditions a track must fulfil in order to do this work?

Assuming that we have a stable and even surface, or "formation," on which to lay the track, we find the first essential to be-that the actual bearing surface be kept constantly dry.

If the wet accumulate, in however small quantities, under

the bearer of the rail, the hollows in which the wet lies are constantly increased by the motion of the trains, involving one of two alternatives; either constant expense in "lifting and packing," or the destruction of the stable foundation on which we started. This condition of dryness is usually fulfilled by the use of "ballast," consisting of an unyielding, yet free and porous material, which allows the water to pass. through and drain away on the prepared surface beneath. But we shall see presently that there is another mode of attaining the required result.

The second desideratum is-as near an approach as practicable to continuity of bearing under the rail.

It is self-evident that if there be much variation in the resistance of the track to the weight of the train, a series of concussions must occur as it passes over the alternately yielding and unyielding portions, destructive of the rolling stock, and, by reaction, of the track itself. The absolute continuity of bearing attained by a wooden sleeper under the full length of the rail, as on the Great Western and Metropolitan, causes the train to glide over the track with a smoothness which is delightful to the traveller, and reduces wear and tear to a minimum. For various reasons, however, to some of which I shall hereafter incidentally refer, the method has never been largely adopted.

The third essential is nearly related to the second, and subsists in a certain amount of elasticity, or partial yielding under the load with subsequent recovery.

The vibration of the train is due partly to the friction and vertical blows of the tire on the track, partly to the remittent impacts on the rail, arising from the play in the gauge. If the track be perfectly rigid, that is, non-elastic, it cannot respond to or absorb the vibration of the load.* The con

Euler says at the close of one of his demonstrations-" This may be contrary to all experience, but is nevertheless true." So, the statement in the text may appear to be (though in fact it is not) contrary to all theory, but is nevertheless true. The theory of "a perfect railway" involves absolute rigidity and regularity of surface-assumptions necessary to the analyses of friction and traction in the abstract. But these assumptions are impossible of realisation in practice. Therefore the theory leaves out of view the conditions which are practically inherent. The fact of practice is, that the more

sequence is a closer contact between the surfaces of rail and tire, and an increase of friction, causing a hard grinding action between these surfaces, destructive of both. In illustration of this, we may remember when our northern lines were laid with stone-blocks, the very perceptible difference between the jar and grating in passing over them, compared with the easier motion when we changed further south to track laid on wood sleepers. The weight and mass of the blocks rendered the road comparatively non-elastic, and, by so much, incapable of absorbing the vibration or transferring it to the road-bed. Or, taking another kind of illustration, nobody prefers to walk on a non-elastic flagged pavement, when there is a choice of an elastic and much more agreeable asphalt pavement.

The bed or foundation itself is a remarkable element in the elasticity of the track, being sometimes rigid and nonreceptive of the vibration, sometimes yielding and receptive. The train is exceedingly sensitive to different conditions of foundation; as when running through a rock-cutting or over a peat-moor, the increase or decrease of vibration is most perceptible. It ought, therefore, to be a feature in a well-designed system of track, that provision can be easily. made for obtaining a compensatory elasticity where the foundation is unyielding.

This important desideratum of elasticity was only slowly arrived at. We used to aim rather at rigidity. Attention was directed to it in observing the action of traffic on the railjoints. Before the adoption of fish-plates for lashing the rails together, the joint was of course the weakest point, and apparently demanded, in a cross-sleeper road, an extra amount of bearing surface. It was found, however, that the rail-ends underwent rapid destruction, particularly the after ends, which received concussion from every coming wheel. A host of expedients was tried to obviate this. When finally the practice, now universal, of fishing the rails was adopted, and instead of laying a sleeper under the joint, it was left suspended between two sleepers, the full amount of elastirigid the road, even when the bearing is "continuous," the harder" the running of the train.