54 BRIDGMAN. 



much flatter and moves in toward smaller tensions at higher tempera- 

 tures. The initial magnitude of the change, for small alterations of 

 tension, becomes larger at the lower temperatures. Here again we 

 find a temperature coefficient with a sign the reverse of what we 

 would expect. 



The changes of resistance as a function of tension are reproduced 

 in Figure 1. The range of tension is the same at the three tempera- 

 tures, namely 1 kg., which corresponds to about 1900 kg/cm-. This is 

 not far below the elastic limit. 



Cobalt. An examination of the behavior of cobalt was of interest 

 because of its close relationship to nickel, and the abnormal behavior 

 of nickel. 



There was available a piece of the same cobalt wire, 0.0030 inches 

 diameter, as that on which the previous determinations of the pressure 

 coefficient of resistance were made. ^^ It was annealed by heating to 

 redness. Measurements were made to a maximum load of 100 gm.; 

 the breaking load is about ISO gm. Within the range of 100 gm. the 

 effect is normal with respect to sign, and the relation is perfectly linear 

 within errors of the single readings of not more than 2 or 3%. There 

 is no trace of the abnormal behavior shown by nickel. 



The tension coefficient is +9-94 X 10"'^ at 30°, the tension being 

 measured in kg/cm-. 



Young modulus was not determined for this wire; because of the 

 small diameter a special procedure would have been necessary. I 

 shall in the following computations assume that it is the same as for 

 iron and nickel, namely 2.0 X 10^- C. G. S. I shall also assume that 

 Poisson's ratio is 0.30. 



Manganin. The interest of this alloy lies in the fact that its pres- 

 sure coefficient of resistance is abnormal in being positive. ^^ 



Determinations of the tension coefficient were made on samples from 

 the same spool as that from which the pressure coefficient samples were 

 obtained and from which the pressure gauges used in all my work were 

 cut. Measurements were made on the wire in both the hard drawn 

 and the annealed condition. 



Thre^runs were made on the hard drawn specimen. The wire was 

 0.0054 inches in diameter. The breaking load was about 900 gm. 

 The maximum load applied before the measurements was only 300 gm., 

 and the runs themselves reached a maximum of only 250 gm. The 

 wire was seasoned for this range of tension by a number of applica- 



12 Reference 4, p. 607. 



13 P. W. Bridgman, Proc. Amer. Acad. 53, p. 370, 1918. 



