370 BEIDGMAN. 



maximum zero correction was 1.1%. In passing from the curves of 

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

 mum readjustment was 0.60%. 



The numerical results are shown in Tables XLI and XLII and Fig- 

 ures 42 and 43. The effect is positive, rising regularly with pressure 

 and temperature to 35.5 X 10"« volts at 100° and 12000 kg. At 

 constant temperature the relation between e.m.f. and pressure is 

 linear. The Peltier heat is positive, rising with pressure and tempera- 

 ture. The Thomson heat is also positive, and rises with pressure and 

 temperature; at every constant pressure it is directly proportional to 

 the absolute temperature. 



Constantan, of unstated composition, was one of the substances 

 measured by Wagner. He gives to 300 kg. and between 0° and 100° 

 31.1 and 26.4 X 10"'^ volts per degree per kg. for two different speci- 

 mens. The value given by interpolation of the data above is 29.7, 

 falling within Wagner's limits. 



Manganin. This was from the same spools as the coils used in 

 measuring resistance, and is of German origin, probably nearly 20 

 years old. The resistance at atmospheric pressure shows a maximum 

 near 30°. The following values of resistance were found: at 0°, 

 1.0000; at 26°, 1.0007; at 50°, 1.0006; at 76°, 1.0002; and at 97°, 

 0.9995. 



The thermo-electric behavior at atmospheric pressure against lead 

 is given by the formulas: 



E = (1.366f-f0.000414<2+0.0000112^3) >< iQ-e volts, 



P = (1.366 + 0.00828 < + 0.0000336 <-) (t + 273) X lO"" volts, 



0- = (0.00828 + 0.0000672 1) (i + 273) X 10-« volts/°C. 



The resistance of this wire was also measured because I now had 

 apparatus for getting more accurately the variation with temperature 

 of the pressure coefficient than formerly. Previously I had measured 

 the same coil successively at different temperatures against an ab- 

 solute gauge, and found only a small variation in the pressure coeffi- 

 cient up to 80°. The change may be slightly different for different 

 specimens, even from the same spool. This procedure was now 

 improved upon by simultaneously measuring two coils, both exposed 

 to the same pressure, but placed in different pressure cylinders and 

 maintained at different temperatures with two different thermostats. 

 The one coil, that with which pressure was measured, was always 

 maintained at the same temperature, 40°, whereas the temperature 

 of the other coil was varied for different runs. The new coil was 



