1891.] MICROSCOPICAL JOURNAL. 269 



-at least with ordinary sections having deep heads and thin bases. Phos- 

 phorus exists in the pig iron, or the ores from which the pig iron is 

 made, and to such an extent in many ores that they are unsuitable for 

 Bessemer metal, and it requires some care to select ores which will run 

 from .08 to .10 of I per cent, of phosphorusr, about the limit to be com- 

 bined witli .35 to .40 of I per cent, of carbon. While rails, many years 

 since, containing .35 to .40 in carbon were suitable for the traffic at that 

 time, they wear too rapidly under the present traffic. To increase the 

 carbon for better wearing qualities, it is first necessary to introduce a 

 section for heavy rails which would cool more uniformly, reducing the 

 coarseness of the texture, and at the same time keep the phosphorus 

 down to avoid brittleness. 



In 1883 I designed an So-pound section, which makes the structure 

 of the metal in the head much finer than usual. The section was put 

 into service in 1884, the manufacture of which became the type of 

 modern sections. Over 200,000 tons of this section have been put into 

 service, some of the rails having .50 in carbon. Even with so much 

 carbon the elastic limits of the steel are below what is necessary for 

 modern traffic, and I am novv making rails with .60 carbon, the phos- 

 phorous being down to or under ,06. Specimen No. 4 shows a 

 piece of steel from such a composition which is very fine-grained for a 

 large rail, tough, and has a tensile strength of 120,000 to 130,000 in 

 the head, the elastic limits ranging from 60,000 to 65,000 pounds. 

 *Such rails can be produced commercially, the cost only being increased 

 about one-tenth above the cost of ordinary rails. 



Without microscopic examination it is difficult to see why it is so 

 important to make the rails of fine texture and high elastic limits. If 

 a rail simply had to perform the functions of a girder, we could in- 

 crease its dimensions so that it would have ample strength, even though 

 the elastic limits of the metal were low. But the upper surface of the 

 rail must also act as the infinitesimal rack by which the drivers secui"e 

 their adhesion for locomotion. Tracing these matters out more fully, 

 we find tile metal, in the head of the rail under a driver, in compres- 

 sion to the vertical axis of the section, while the metal under the neu- 

 tral axis would be in extensioti^ which would reach to each tie, beyond 

 which, as far as affected by the weight of that driver, the base would 

 be in compression at the head in extension. These strains would be 

 reversed as the driver or wheel reached the next tie space. The metal 

 in the head directly under the wheel must not only bear the weight 

 upon the driver, but also all the traction the driver is exerting to draw 

 the train. From the small areas in contact, the ratio of pressure is 

 from 60,000 to 80,000 pounds per square inch, while the traction often 

 amounts to one-half as much for the surfaces in contact as longitudinal 

 strain upon a thin layer of surface metal in the rail head. Examining 

 the rails in the track with the microscope, we find not only small por- 

 tions of the metal torn out, but a series of minute cracks, showing that 

 the metal has been strained upon the surface beyond its elastic limits, 

 and surface wear of metal rapidly occurs. To check this wear we 

 need high elastic limits of the metal for the surfaces in contact. The 

 metal in the tires of wheels abrades on the surface and also drops out 

 in patches, as may be seen in specimens Nos. 3 and 4. 



It will be readilv understood that, while it is desirable to have suffi- 



