368 



BRIDGMAN. 



sure is a difference of a difference, the residual effect being the differ- 

 ence between the e.m.f s. at the hot and cold ends brought about by 

 the difference of temperature produced by compression between the 

 inside and the outside of the cylinder. When one considers that the 

 instantaneous rise of temperature by compression might be 10°, it 

 does not seem surprising that there should be such oscillations with a 

 galvanometer sensitive enough to be thrown off the scale by 6 X 10~® 

 volts. The e.m.f. between iron and constantan is 50 X 10~^ volts 

 per degree. 



Because of the complications involved in any interpretation of the 



TABLE XLI. 



Constantan. 



Thermo-electromotive Force, volts X 10^. 



instantaneous effects, I gave up the attempt to observe them after a 

 few metals had been measured. The effects may change sign with 

 the same metal with pressure and temperature. I did not happen to 

 make observations like those on constantan with any other substance 

 except manganin, also an allo5'. This is probably only a coincidence. 

 The pressure measurements of e.m.f. went fairly smoothly. Except 

 for two points at 96° with discrepancies of 4.0 and 2.3%, the maxi- 

 mum departure of any point from a smooth curve was 1.3% and the 

 average numerical departure was 0.26% of the total effect. The 



