270 BRIDGMAN. 



properties would be desirable in attempting to picture the complete 

 electrical mechanism. I stated in the previous papers that I hoped 

 to obtain additional measurements on thermo-electrical properties. 



In this paper these thermo-electric measurements are carried out. 

 The change in the thermo-electric properties of nearly all the metals 

 whose electrical resistance was there measured is here given over the 

 same range of pressure and temperature; a few only of the metals 

 were not adapted to these measurements and could not be used. In 

 addition, new data are given for the effect of pressure on both electrical 

 resistance and thermo-electric properties of the alloys manganin and 

 constantan. In most cases measurements of the thermo-electro- 

 motive force were made either on identical pieces of wire, or on pieces 

 from the same specimen as were the resistance measurements. In the 

 few cases where new specimens of metal had to be used, the tempera- 

 ture coefficient of resistance at atmospheric pressure has also been 

 measured, in order to define the electrical character of the metal as 

 completely as possible. 



The results of this paper, unlike those of the previous paper on 

 resistance, are almost entirely novel; the nature of the results to be 

 expected was not known, and accordingly these effects, so far as affected 

 by pressure, were not available for any theoretical considerations. 

 Previous measurements on the effect of pressure on thermal e.m.f. 

 are very few in number, and cover a very restricted range. The maxi- 

 mum pressure reached has been by Wagner,^ 300 kg., over the same 

 temperature range as that used here, 0° to 100°. Over a pressure range 

 so small, the effect is in most cases so minute as to be on the limits of 

 the readily measurable. Previous measurements have been almost 

 entirely confined, of necessity, to determining the total effect of the 

 maximum pressure and temperature. It was not possible to find at 

 all accurately whether the effect was linear with pressure and tempera- 

 ture within the range employed. On a few metals Wagner made 

 measurements of the effect of varying temperature, and found that 

 within his limits of error the effect was linear. The effects are so 

 much larger over the much greater pressure range of this paper that 

 I have been able to make accurate measurements of the variation of 

 the effect with pressure and temperature within the range. This is 

 important, especially the variation with temperature. For it is well 

 known that the Peltier heat is determined by the first derivative of 

 the thermal e.m.f., and the Thomson heat by the second. Previously 



2 E. Wagner, Ann. d. Phys. 27, 955-1001, 1908. 



