437 
1910-1 l.J Energy in Torsionally Oscillating Wires. 
After heating to redness, there seemed to be in this case no tendency 
to change the form of the curve, and again a pair of straight lines was got, 
with values not far different from those got for the original wire. The 
values found were 
A: a= 0, ?i = 4‘00, 6 = 753. 
B: a = 25, ra=l-25, 6 = 120. 
Heating to the limit of temperature reached has thus only a slight 
effect. It tends to increase the angle between the pair of lines by a small 
amount. 
On stretching the wire at the ordinary temperature, the form of curve 
is totally altered. Now all the points fall on one straight line (diagram 9), 
the values of the constants being 
a =18, ii = T05, 6 = 252. 
Another set of values obtained were 
a = 20, n=\ - 05, 6 = 270. 
Stretching has lowered the value of n below either of the values got with 
the original or the heated wires. 
German Silver Wire. 
Similar experiments were carried out with German silver wire, of 
which a length of 1 foot was used, the diameter being 1 mm. Tested at 
the ordinary temperature, results were got which resembled those got for 
brass in the same condition. With a = 110 a perfectly straight line was ob- 
tained, the values being 
a = 1 10, w=l*05, 6= 1080. 
On heating to redness and testing, however, the resemblance to brass ceases, 
as one straight line is still obtained, the values being 
a = 20, 71=1-70, 6=1092. 
Whilst a has decreased considerably and n increased, b has remained 
practically stationary. 
This heated wire was now pulled till rupture occurred, and half the 
original length, of diameter *85 mm., oscillated. A rather unusual result 
was here obtained, as two lines were got with the values 
A: a = 50, n=' 66, 6 = 223. 
B: a = 60, n=- 94, 6 = 292. 
These values of n are lower than either of the former values. The sup- 
position might be made that the original unheated wire, if pulled, would 
