THERMO-ELECTRIC QIALITY UNDER PRESSURE. 377 



The maximum tension applied 'was 700 kg. /cm.- Within this range 

 the e.m.f. was Hnear with tension. Trustworthy measurements 

 were obtained only at 95°. Here tlie two specimens differed from the 

 mean by 20%. 



e.m.f. = 0.000."i()T X 10"^' volts. 



The direction of e.m.f. is from vmstretched to stretched at the hot 

 junction. This corresponds to from compressed to uncompressed 

 at the hot junction, and is opposite in sign from the pure pressure 

 effect. 



Iron. This was annealed "ingot" iron, from the same piece of 

 metal as that of the pressure measurements. Its diameter was 0.025 

 inches, and it was subjected to a maximum tension of 500 kg./cm.^ 

 Satisfactory readings were obtained at two temperatures; the two 

 wires differed by less than 10%. The results are covered by the 

 formulas : 



At 52° e.m.f. = (0.0120T-0.042.35T2) X 10-« volts, 

 at 95° = (0.022.3T-0.0442T2) X 10-«. 



The effect has a pronounced maximum near the middle of the ten- 

 sion range; this recalls the complicated behavior under hydrostatic 

 pressure, although the range of tension is comparatively much lower 

 than the range of pressure. The e.m.f. is from unstretched to stretched 

 at the hot junction, corresponding to from compressed to uncom- 

 pressed. This corresponds to the sign of the pure pressure effect over 

 the first portion of the pressure range. 



Aluminum. Commercial stock of 0.0195 inches diameter was 

 used. Satisfactory results were obtained at two temperatures; the 

 two wires differing by about 10% from the mean. The results are 

 given by the formulas: 



At 50.5° e.m.f. = (O.OO88T-O.O567T2) X IQ-e volts, 

 at 76° = (O.OOS8T-O.O547T2) X 10-«. 



The departure from linearity, contrary to what one might expect, 

 is only slight. The abnormal behavior under pressure, however, is 

 reflected in the slight change of e.m.f. with temperature. The direc- 

 tion of e.m.f. is from unstretched to stretched at the hot junction, 

 corresponding to from compressed to uncompressed. This corre- 

 sponds to the initial direction of the pressure effect. 



Brass. This was commercial wire of 0.008 inches diameter. The 



