which represents a longitudinal section of a tire of recent steel, polished and prepared, and magnified about 30 diameters.

Fracture usually occurs between the crystalline groups, not through such a group. In rolling a tire or rail, a group of these small crystals is elongated and flattened in a plane parallel with the surface. The crystals are first thinned, then loosened, by wear; and finally they flake out, leaving a very uneven surface on the tread of the tire or the head of the rail. The depth of these depressions is, as I

have said, about

of an inch. They are found to range in number from 6 to 15 per square inch. Figs. 3 and 4 are reproduced from photographs of two steel tires, representing right and left car-wheels, which have thus flaked out.

Now the early steel rails did not flake out, but wore smooth. It is evident that the chemical composition of two rails may be nearly identical, and yet the difference in the grain of crystallization may give them different physical properties.

It was the consideration of these features which led me to design the broad and shallow head of the 80-lb. rail shown in Fig. 1. My first object was to lessen the pressures per unit of contact, thereby reducing the rate of wear of the rail-head and the tires.

The rails of which I had diagrams weighed 56, 58, 60, 65 and 67 lbs. per yard and ranged in height from 3 to 4 inches. In the track of the Boston and Albany Railroad there were several miles of 72-lb. rails, 4 inches high.

The diagrams of the condition of track showed conclusively that none of the rails were stiff enough to carry the traffic for any length of time and keep their surface without constant attention.

The rails examined in the tracks, with the aid of the track deflection apparatus of my car, to test the condition of their surface, were

found to have definite forms of permanent set, which could be directly traced to want of care in the track, or, in some cases, want of care in manufacture.

Rails which were low at the joints and high in the center, being the most frequent, I called the first form of permanent set; rails low

at the joints and center but high in the quarters, I called the second form; rails, the surface of which was a series of short waves, I called the third form.

Figs. 5, 6 and 7 illustrate these three forms.

Combinations of the first and third, and the second and third forms were frequent. Fig. 8 shows the long waves which are found to take place in the best tracks under the weight of the cars, from want of stiffness of the rails and equal tamping of the ties. Fig. 9 shows the destructive effect of car-wheels upon rails of the first form of permanent set. One-half the length of the rail being an up-grade and the other half a down-grade, the receiving-end of the up-grade part is hammered down by the descending wheels. Figs. 10 and 11 show the effect upon the first and second forms respectively of trains running in the direction shown in Fig. 9. Trackmen are unable to surface such injured rails unless they are removed from the track, the ends sawed off and redrilled.

As a rule, rails laid with opposite joints show predominantly the first form of permanent set. In rails laid with alternate joints, on the other hand, the second form predominates, provided the section of the rail, joint-fastening, and labor expended in maintenance are sufficient to maintain the track under the traffic. If any one of the three requisites be lacking, rails of the second form sooner or later pass to the first form of permanent set. This occurs very soon in planked or paved stations and at road crossings. If the joints deflect of an inch or more under passing trains, the receiving ends of the rails will be cut out as shown in Fig. 10.

The deflections of rails under the wheels of trains (see Fig. 8) are more extensive than is generally supposed. The subject has not received the attention due to its importance, in view of our present wheel-tonnage. In low and light sections the weight of each wheel is not distributed over five or six ties, but is concentrated, for the instant, upon two or three-two, if the wheel is between them; three, if the wheel is over a tie. The second or third tie, as the case may be, is relieved of pressure, the base of the rail being under compression instead of tension, as would be the case with a rail of proper stiffness. The concentration of the wheel-weight over a small area of ballast quickly disturbs it; the spikes are started, the ties become loose, and are rapidly abraded under the rails, while the labor to maintain the track is much greater than would be the case with stiffer rails. This has been well demonstrated by experience with the 80-pound rail. In its design the stiffness of the section was an important feature, for the reasons above stated. Up to the present date it is stiffer than any 85-pound section yet produced. The following table gives its deflections under loads: