1912-13.] Application of the Law of Torsional Oscillation. 187 
required, giving two straight lines with different value of n. The actual 
values were : — 
Temp. 
1st Value of a. 
2nd Value of a. 
1st Value of n. 
2nd Value of n. 
15° C. 
0 
105 
1-34 
•20 
150° C. 
0 
60 
1*40 
*45 
200° C. 
7 
20 
1-40 
•47 
243° C. 
2 
50 
1-48 
•47 
280° C. 
1 
20 
1*85 
•67 
315° C. 
2 
12 
1-90 
•92 
325° C. 
1 
0 
1-90 
1*25 
330° C. 
4 
0 
1-90 
1-70 
350° C. 
4 
2 
1-90 
1-80 
373° C. 
1 
2 
1-90 
1-90 
380° C. 
1 
1 
1-90 
1-90 
390° C. 
3 
6 
2-20 
1-80 
420° C. 
2 
12 
1-70 
1-50 
450° C. 
0 
12 
2-80 
2-05 
475° C. 
0 
40 
2*90 
•95 
565° C. 
0 
65 
3-70 
1*40 
The first value of a remains in the vicinity of zero, whilst the second 
one, although fluctuating occasionally, tends on the whole to decrease to 
zero and increase thereafter. The first value of n tends to increase 
uniformly, whilst the second increases till a temperature in the neighbour- 
hood of 370° C. is reached, and thereafter fluctuates. It is noticeable, 
however, that both values of a and both values of n tend to become 
identical at that temperature. That was also the temperature at which, 
in the case of brass, the one value of n, which till then had increased 
uniformly, split up into two values. Both metals, then, would seem to 
undergo a molecular change at that temperature, a change which is not 
noticeable in the larger oscillations, but whose effects come into play after 
the swing has died down to about half its original value. 
Diagram II. shows the relationship between temperature and the value 
of n, and shows how the curves approach and finally touch at a tempera- 
ture of about 370° C. Thereafter the values fluctuate so much that no 
conclusion can be drawn regarding the form of the curves, but they 
certainly tend to separate again after this point. 
A length of 30 cms. was immersed in liquid air for 30 minutes, and 
subsequently tested, but no change in the constants was evident. 
Aluminium Wire — Effect of Fatigue. 
The same wire was then fatigued by means of a repeated rotational 
strain * for a period of two hours, but again, except for a slight lowering 
* Proc. Boy. Soc. Edin vol. xxxi. p. 440. 
