"2d. I took twelve darning needles of good quality, 3 in. long, 1/24 in. thick. The ends of these were placed against steel props, 2-1/8 in. asunder. In making an experiment, a wire was fastened to the middle of a needle, the other end being attached to a spring weighing-machine. This was then pulled until the needle gave way. Six of the needles, taken at random, were tried at a temperature of 55 deg. F., and the remaining six in a freezing mixture which brought down their temperature to 12 deg. F. The results were as follow:-- Warm Needles. Cold Needles. 64 ounces broke 55 ounces broke 65 " " 64 " " 55 " " 72 " " 62 " " 60 " bent 44 " " 68 " broke 60 " bent 40 " " --- --- Average, 58-1/3 Average, 59-5/6 "I did not notice any perceptible difference in the perfection of elasticity in the two sets of needles. The result, as far as it goes, is in favor of the cold metal. "3d. The above are doubtless decisive of the question at issue. But as it might be alleged that the violence to which a railway wheel is subjected is more akin to a blow than a steady pull; and as, moreover, the pretended brittleness is attributed more to cast iron than any other description of the metal, I have made yet another kind of experiment. I got a quantity of cast iron garden nails, an inch and a quarter long and 1/8 in. thick in the middle. These I weighed, and selected such as were nearly of the same weight. I then arranged matters so that by removing a prop I could cause the blunt edge of a steel chisel weighted to 4lb. 2oz., to fall from a given height upon the middle of the nail as it was supported from each end, 1-1/16 in. asunder. In order to secure the absolute fairness of the trials, the nails were taken at random, and an experiment with a cold nail was always alternated with one at the ordinary temperature. The nails to be cooled were placed in a mixture of salt and snow, from which they were removed and struck with the hammer in less than 5"." The collective result of the experiments, the details of which need not be given, was that 21 cold nails broke and 20 warm ones. Dr. Joule adds, "The experiments of Lavoisier and Laplace, of Smeaton, of Dulong and Petit, and of Troughton, conspire in giving a less expansion by heat to steel than iron, especially if the former be in an untempered state; but this, would in certain limits have the effect of strengthening rather than of weakening an iron wheel with a tire of steel. "The general conclusion is this: Frost does _not_ make either iron (cast or wrought), or steel, brittle. Mr. Spence, in his experiments, decided on having some lengths of cast iron made of a uniform thickness of 1/2 in. square, from the same metal and the same mould.