580 BRIDGMAN. 



of the purity, careful measurements were made of the temperature 

 coefficient. In many cases the materials seems purer than any 

 on which measurements have been previously published. Detailed 

 data from which the purity may be judged are given under the in- 

 dividual substances. The resistance at 25° intervals in terms of that 

 at 0°C as unity is tabulated in the following. The average tempera- 

 ture coefficient between 0° and 100° may be read directly from the 

 resistance at 100°. 



In comparing these results with those of others it is necessary 

 to keep in mind that the relation between resistance and tempera- 

 ture is not linear, but for most substances the resistance increases 

 more rapidly at higher temperatures. This introduces a slight correc- 

 tion. For instance, Jaeger and Diesselhorst ^ made measurements at 

 18° and 100°, extrapolated linearly to find the resistance at 0°, and 

 tabulated the coefficient obtained in this way as the average coefficient 

 between 0° and 100°. The value thus found is evidently too high, 

 and so is not strictly comparable with the results found here. The 

 correction, which is always slight, may be computed in any case from 

 the tables given below. 



After winding the coils they were in most cases seasoned by ex- 

 posing to several changes of temperature between 0° and 140° or 150° 

 in an air bath. In addition to this, the coil was frequently seasoned 

 for pressure by several times exposing to 12000 kg. This seasoning 

 for pressure turned out to my surprise, however, to be hardly necessary. 

 A number of substances showed no perceptible change of resistance 

 after their first exposure to pressure. This is an interesting point, as 

 it indicates with a high degree of sensitiveness the perfect elasticity of 

 volume under hydrostatic pressure. Further details of seasoning are 

 given under the individual substances. 



Measurements of the change of resistance were made at intervals of 

 1000 kg. at 0°, 25°, 50°, 75°, and 100°. Two readings were made at 

 the maximum and two zero readings, one before and one after the run. 

 At each temperature, therefore, 15 readings were made. The straight 

 line connecting the mean of the two points at 12000 with the mean 

 zero was computed, and at each pressure the difference between the 

 observed and computed ^'alue was found. These differences were 

 then smoothed graphically. 



Readings of resistance were made with increasing and decreasing 



8 W. Jaeger und H. Diesselhorst, Phys. Tech. Reichsanstalt, WLss. Abh. 3, 

 269-425 (1900). 



