948 
The measurement of y at very low temperatures is one of extreme 
diftieulty. For, at these temperatures, the temperature coefficient is 
so great that even the smallest fluctuations of temperature can greatly 
affect the resistance; in this way a slight disturbance of the tempe- 
rature equilibrium can occasion a variation of the resistance which 
completely obscures the whole phenomenon of variation with pressure. 
With the wire we used, for instance, at 7 — 20°.3 K. a pressure 
of 100 atm. brought about a change in the resistance of only about 
0.001 2, while a change of 0.0003 2 was the result of a variation 
of 0.01 degree in the temperature. And it is pretty obvious that it is a 
matter of extreme difficulty to re-adjust the temperature to within 
0.01 degree of its former value after it has been altered by the heating or 
cooling of the liquid oceasioned by fluctuations in the pressure. 
There is a second factor operating which renders the measurement 
difficult. When the compression has been continued for a long time, 
elastic after-effects occur whieh can also attain a value that is a 
considerable fraction of the magnitude to be observed. Should, the- 
refore, the variation of pressure be distributed over a long period of 
time in order to disturb the temperature equilibrium as little as 
possible, this after-action will give rise to a source of error. 
§ 2. The lead to be subjected to pressure consisted 
of a turning about 2 metres long and about 0.2 m.m. 
in diameter. After some practice these long thin 
turnings could be successfully prepared. Attempts 
to draw wires of this small cross-section did not 
meet with success. The wire, «, was wound upon 
an ebonite cylinder (see fig. 1). To either end was 
soldered a band, c¢,, c,, rolled from a wire of elec- 
trolytie copper; to these bands were soldered the 
255 two pairs of leads, d,, d,, d,, d,. To enable one to 
5 el ive. subject the lead to pressure it was enclosed in a 
= thick-walled copper cylinder, A,, closed “below by 
a heavy cap, A,, (screwed and soldered), and above 
by a cap A, through which two copper capillaries 
A,, A,, pass. One of. these capillaries, A,, is used 
for filling the cylinder with liquefied gas and for 
exerting pressure upon the liquid, while the other, 
A, is connected to a manometer and, at the same 
time, acts as a safety valve, in allowing the liquefied 
gas to escape in the event of the supply tube getting 
frozen. This second tube also admits the wires required 
