360 BRIDGMAN. 



The thermo-electric behavior at atmospheric pressure against 

 lead is given by the formulas: 



E = (1.594 /+ 0.01705^2) x 10-« volts, 



P = (1.594 + 0.0341 1) (t + 273) X lO'^ volts, 



a = 0.0341 (/ + 273) X 10-6 volts/°C. 



For the pressure measurements it was annealed once to 12000 kg. 

 at room temperature. At 25°, 50°, and 75°, the readings under pres- 

 sure went as smoothly as could be desired, but at 98° there was greater 

 irregularity; here there was a permanent change of zero of 7% of 

 the total effect. This was also the maximum departure of any point 

 from a smooth curve. The average arithmetical departure of all 

 points from smooth curves was 0.72%. In passing from curves of 

 e.m.f. at constant temperature to those at constant pressure a maxi- 

 mum readjustment of 1% of the total effect was necessary. 



The numerical results are shown in Tables XXXVII and XXXVIII 

 and Figures 38 and 39. The effect is positive, increasing regularly 

 with pressure and temperature, and is rather large, being 20 times as 

 great as that of molybdenum. This was a surprise; the pressure 

 coefficients of resistance of these two metals are nearly equal. The 

 Peltier heat is positive, rising with pressure and temperature. The 

 Thomson heat is positive, at the two extremes of the pressure range 

 rising with temperature at constant pressure, but at intermediate 

 pressures it falls with rising temperature. 



There are no previous measurements for comparison. 



Bismuth. This was electrolytic bismuth, prepared at the same 

 time as the resistance specimen, but it was from the deposit on another 

 electrode, so that the two specimens are not of necessity exactly alike. 

 From the electrode deposit, the metal was made into wire by the 

 same steps as the previous wire; the die tlirough which this was 

 extruded was 0.0285 inches in diameter, somewhat larger than the 

 previous specimen. After extrusion, it was annealed by slow heating 

 and cooling from room temperature to 100°, being maintained at 100° 

 for 30 minutes. Much care was necessary in getting it in place in the 

 pressure apparatus, but this was successfully accomplished without 

 a single mishap. 



At atmospheric pressure the average temperature coefficient of 

 resistance between 0° and 100° of this specime'n was 0.004372, which 

 is somewhat lower than that of the previous specimen, 0.00441; the 

 difference is no greater than has been formerly found to be due to 

 differences of handling. The curve of resistance against temperature 



