gives a distinct metallic ring when struck. Balls of iron, lead, tin, ivory, &c., thus cooled, exhibit an increased rebound when dropped from a height; an indiarubber ball, on the other hand, becomes brittle, and is smashed to atoms by a very moderate fall. Tables XII. and XIII., which give the mean results of a large number of experiments, show the increased breaking stress gained by metals while they are cooled to the temperature of liquid oxygen. TABLE XII.--_Breaking Stress in Pounds of Metallic Wires 0.098 inch in diameter._ +15° C. -182° C. Steel (soft) 420 700 Iron 320 670 Copper 200 300 Brass 310 440 German silver 470 600 Gold 255 340 Silver 330 420 TABLE XIII.--_Breaking Stress in Pounds of Cast Metallic Testpieces; diameter of rod 0.2 inch._ +15° C. -182° C. Tin 200 390 Lead 77 170 Zinc 35 26 Mercury 0 31 Bismuth 60 30 Antimony 61 30 Solder 300 645 Fusible, metal (Wood) 140 450 In the second series of experiments the test-pieces were 2 in. long and were all cast in the same mould. It will be noticed that in the cases of zinc, bismuth and antimony the results appear to be abnormal, but it may be pointed out that it is difficult to get uniform castings of crystalline bodies, and it is probable that by cooling such stresses are set up in some set of cleavage planes as to render rupture comparatively easy. In the case of strong steel springs the rigidity modulus does not appear to be greatly affected by cold, for although a number were examined, no measurable differences could be detected in their elongation under repeated additions of the same load. No quantitative experiments have been made on the cohesive properties of the metals at the temperature of boiling hydrogen (-252°), owing to the serious cost that would be involved. A lead wire cooled in liquid hydrogen did not become brittle, as it could be bent backwards and forwards in the liquid.