1911-12.] Effect of Vibration upon the Structure of Alloys. 317 
copper and 10 per cent, of tin, was cast in an iron mould about 4 inches long 
and | inch square section. This alloy, when in a state of physico-chemical 
equilibrium at the atmospheric temperature, consists of uniform crystals 
of a homogeneous solid solution of tin (or a copper-tin compound) in copper, 
usually designated a. Rapid cooling from the liquid state, as by pouring 
the melted mixture into a cold iron mould, prevents the attainment of that 
condition, and the chill cast alloy is formed of crystal skeletons of inhomo- 
geneous a, each surrounded by a network of a material of somewhat 
indefinite composition and constitution, but richer in tin than the skeletons 
(fig. 2). Chilled castings of other alloys were prepared, but no definite 
results were obtained from them; these alloys were: copper 70, zinc 30; 
copper 90, aluminium 10 ; and aluminium 95, zinc 5 per cent. 
In order to apply the numerous small blows required, the chill cast 
pieces, sawn into f-inch cubes, were put into a wooden box, about 5 inches 
by 5 inches by 7 inches, mounted as shown in fig. 1, and rotated, by means 
of a small water motor, at a speed which averaged 80 revolutions per 
minute. The apparatus was run, as a rule, only during the daytime for 
periods of 6 to 8 hours. In a tumbling box of this kind, if the speed of 
rotation is too low, the pieces slide down the sides of the box and receive 
no appreciable blow, whereas if the speed is too high they remain in 
stationary contact with the sides, on account of centrifugal action. At 
the right speed of rotation the pieces are carried to nearly the top of the 
box, and then drop freely to the bottom ; each piece therefore receives two 
blows per revolution. About two dozen of the metal cubes were put into 
the box together ; but since some were harder than others, the softer pieces 
suffered considerable distortion, especially in their surface layers, by impact 
with the harder pieces. Much wear took place, chiefly through the 
exfoliation of the crushed surface layers ; the aluminium-zinc alloy suffered 
the greatest wear, the copper-zinc and the copper-tin showing less, and 
the copper-aluminium very little. Thus the original intention of imparting 
vibration without distortion was defeated, but the results obtained with 
the copper-tin alloy are not without interest. 
Fig. 2 is a section of the chill cast gun-metal, already described : the 
dark primary a-skeletons are small, and their outlines are irregular. 
Pieces were withdrawn from the rattling box after various periods of 
rotation, cut through the middle, polished, and etched, exactly as the 
original unshaken sample, with a solution of ferric chloride acidified with 
hydrochloric acid. After only \ million revolutions no change was definitely 
detectable, but figs. 3, 4 and 5, taken after f, 1J, and 3 million revolutions, 
respectively, show a progressive increase in the size of the skeletons and 
