PRESSURE ON RESISTANCE OF METALS. 635 



pressure at the higher temperatures means that at higher tempera- 

 tures the instantaneous pressure coefficient decreases more rapidly- 

 with increasing pressure than at lower temperatures. This comes 

 to much the same thing as the statement that the curvature is less at 

 the higher temperature, although the two statements are not entirely 

 coextensive. The rate at which the pressure of maximum deviation 

 is displaced at higher temperatures is very different for various 

 substances and seems to have no obvious connection with other 

 properties. 



To sum up: different metals show minor irregularities in behavior, 

 but they are alike in several general features which must be the first 

 task of any theory to explain. These general features are the approxi- 

 mate constancy of pressure coefficient with temperature, and the 

 accompanying constancy of temperature coefficient with pressure; 

 contrasted with this the pronounced decrease of instantaneous pres- 

 sure coefficient with rising pressure, and, of less compelling importance, 

 the decrease in the curvature of most of the resistance-pressure curves 

 at higher temperatures. 



Bearing of the Results on the Question of the Metastability 



OF the Metals. 



In the last few years a great deal of work has been done by Cohen ^* 

 and his pupils on the variation in the behavior of metals after different 

 kinds of heat treatment. They have found very small discontinuities 

 in various physical properties which have been interpreted as indicat- 

 ing that a number of the metals occur in several polymorphic forms. 

 Similar discontinuities have also recentlv been found bv Janecke,^^ 

 but almost always at higher temperatures. The existence of such 

 modifications is important both from the practical point of view and 

 because of the intimate relation to the theory of allotropy of Smits.^^ 

 Cohen's result for copper has recently been called in question by Bur- 

 gess and Kellberg,^^ who could find no discontinuity in the electrical 

 resistance in the expected place. 



The results above throw light on the same question. If there are 

 polymorphic transitions, there should be discontinuities in the resist- 



24 E. Cohen, Numerous papers in Proc. Amst. Acad, since 1913. 



25 A Smits, Proc. Amst. Acad., numerous papers 1910-15. 



26 G. K. Burge.s8 and I. X. Kellberg, Jour. Wash. Acad. Sci. 5, ()57-(562(191.'5). 



