435 
1910-11.] Energy in Torsionally Oscillating Wires. 
The values of n are seen to have increased greatly, whilst the average 
value of b has remained nearly the same. The value of a required for the 
first line has not altered much, but considerably for the second. 
The effect of stretching the wires was then observed. The original 
wire broke in each case before any lengthening was obtained, but the 
heated wire pulled easily, and a length of 6 inches was got, whose dia- 
•10 -20 *30 *40 -50 -60 70 *80 -90 
found to have the same effect as in the case of the brass in the sense of a 
reversion to the non-heated condition, but, as is seen in the diagram, the 
actual type of effect is quite different. 
The second line again breaks away from the first, and a totally different 
value of a is required. The condition here is seen to be intermediate 
between those of the original unheated and the heated conditions. The 
actual values were 
A: a= 1, n = 2*38, V = 336. 
B: a = 50, n= 70, V = 130. 
In no condition, then, has it been found possible to obtain one straight 
