The changed relations of the steel are not appreciated in the bearing surface to the wheel-contact pressures in the stiff sections, as the efficient engineering structures which have empowered the present wheel and total loads of the locomotives and cars. By the design of the section for a given weight, its mechanical properties have been increased, to transport heavier loads. To complete it as an engineering structure, with the requisite efficiency, the physical properties must also be augmented, in sound metal, to raise the limits of its cubic elasticity in the bearing surface proportionately to those required by its enlarged mechanical properties as an engineering structure. This is a problem of mechanics involving a metallurgical solution.

Iron rails failed in the bearing surface when the wheel loads increased to 10,000 pounds owing to the low limits of cubic or elasticity of volume of the metal. The two or three per cent. of slag made it a bundle of fibers only for the wheel pressures, though adequate to distribute the loads as a girder. Increase in stiffness in the iron rails to carry larger bending moments hastened their destruction.

The limber steel rails first rolled, the grain or texture of the metal was fine and compact. The elastic limits in reference to its cubic elasticity were high, and could sustain the wheel contact pressures without distortion. The conversion of the steel was less rapid, the ingots smaller, and liquation was slight. The present conversion, in larger vessels, teemed in larger ingots, are not as sound, for the entire length, as the smaller ingots. The tendency of the upper portion to form a pipe or become porous, or from the longer time in cooling, allows the metalloids to separate from the bath and become concentrated in an upper central core of the ingot. This does not produce homogeneous metal in the head. It is not capable of sustaining the wheel contact pressures by its cubic elasticity, and fails rapidly in service.

These are important problems in the manufacture of rails for the present traffic. It is not alone a question of heat treatment, a necessary, yet an over-estimated panacea for defective steel, but to secure a sound ingot, so that the metal in the entire head will be homogeneous.

The steel in the rail section should be sound and have sufficient physical properties and rigidity of structure to preserve the

shape of the section under the traffic, except the loss by wear of the wheel treads in the bearing surface, and the wheel flanges on the side of the head. Sound metal of 56,000 or more granulations per square inch, and elastic limits of 56,000 to 60,000 pounds, has sustained the present wheel loads without distortion of the heads. The facing ends of the rails should not flatten in sections having moments of inertia of more than 25 4th power inches. More limber sections are not fished sufficiently secure to prevent the wear of the facing ends of the rails under the shocks of the present wheel loads.

I have under observation rails made at several different mills and in those without effort to control or check liquation, the splitting and the piping of the rail heads is pronounced. In rails rolled from metal where attention was paid to checking liquation, the piping and splitting of the rail heads is practically unknown. The problems of checking liquation in large ingots are not always easy of solution, and must be solved in reference to the practice of each mill.

The metal for the stiff rails, the efficient engineering structures, must be sound, for any section to stand in the track without distortion under the traffic. If the steel contains a central core of harder material than the outside, or is porous, containing occluded gases, then it does not have sufficient toughness and tenacity of structure for the requisite limits of cubic elasticity to sustain the wheel load effects without distortion, and fails mechanically as an engineering structure.

INFLUENCE OF METHODS OF PILING STAYBOLT

IRON ON VIBRATORY TESTS.

BY H. V. WILLE.

Staybolt iron, more than any other material, has been sold upon the reputation which certain brands have established rather than any upon particular physical properties. The reason for this is not far to seek for the life of a staybolt is so largely dependent upon conditions other than the quality of the iron, that it is hard to trace by service tests the effect of the quality of the iron upon the life of the bolt.

It is obviously impossible by such a test to determine the value of the material until after it has been used, and it can then only demonstrate that a particular brand of iron gives the best results at the time and under the specific conditions of the test.

It is a well-known fact that staybolts fail because of fatigue from the bending stresses induced by the expansion and contraction of the firebox. It would, therefore, appear to be logical to apply some fatigue test to material to be employed for this class of work, particularly since there is very little variation in the tensile strength, elongation and reduction of area of all good grades of iron universally employed for staybolts.

A large number of such tests have been made on the various brands of staybolt iron and, while there is sometimes considerable variation between the tests of two samples from the same bar of iron, the results conclusively show that the manner of piling exerts a profound influence upon the number of vibrations which a bolt will withstand.

Staybolts fail from fatigue and always close to the outside sheet. The fracture is in detail, starting from the base of a thread and gradually extending inwardly. The remedy lies in the employment of material of a sufficiently high elastic limit to prevent such a fracture from starting, and the use of a high steel naturally suggests itself. Such material has, however, found little favor

with users of this class of material because it is not adapted to a class of service in which stresses are localized at the base of a sharp thread, and because it does not permit of being readily hammered cold, for after staybolts are screwed into a sheet they must necessarily be headed in this manner.

For these reasons iron is almost universally used for the manufacture of staybolts, and it is obvious that an iron piled and worked with a view of breaking up the extension of a fracture, once started, should give the best results. Reasoning along these lines, a num

TABLE I.

RESULTS OF VIBRATORY TESTS OF STAYBOLT IRON.

4,000 pounds load at 3-32" deflection.

ber of manufacturers were induced to pile an iron with a central core of small square piles and with covering plates on each of the four sides. The covering plates ensured a good sound thread and prevented the origination of a fracture due to bending the bolts at right angles to the fibers. The large number of central bars is designed to prevent the continuation of the fracture after it once

starts.

These irons were subjected to the usual tests, and showed no difference from the results obtained from tests of standard makes of staybolt iron. The last two columns in Table I show that there