ELECTRICAL RESISTANCE UNDER PRESSURE. Ill 



any of the observed points from a linear relation was 2.7% of the total 

 effect. 



The constraining effect of a glass capillary on the solid is shown by 

 the low value, — O.O5I9I, found for the coefficient of the soHd in glass. 



The pressure coefficient of the solid is seen to be negative, that is, 

 normal. This was rather a surprise; I had anticipated because of the 

 abnormal expansion on freezing and the fact that bismuth also expands 

 on freezing and has a positive coefficient of resistance that the coeffi- 

 cient of gallium might be positive also. The numerical value of the 

 coefficient of the solid is quite normal, when compared with other 

 metals. It is the value characteristic of a hard metal, which in most 

 other cases also means a metal with a high melting point. The coeffi- 

 cient of the solid is of the order of one half that of the liquid. This 

 again is as one would expect, except for the abnormal volume rela- 

 tions on freezing. However, the solid is less compressible than the 

 liquid in spite of its greater volume ^^^; so that from this point of view 

 the relative magnitudes of the pressure coefficients of liquid and solid 

 do not seem unnatural. 



The temperature coefficient of resistance of the unconstrained 

 solid w^as obtained from readings at 0° and 21.5°. The value for this 

 range is 0.003963, an entirely normal value. Previous values for this 

 coefficient are exceedingly uncertain. Guntz and Broniewski's read- 

 ings were quite irregular, the effect even reversing in sign above 18.6°. 

 This may have been due to the constraining effect of the glass; svich 

 an effect is to be expected, and in the observed direction. 



It may be mentioned that I made measurements on the sul)cooled 

 liquid at 0°, and found the resistance to lie on a regular prolongation 

 of the curve for the resistance above the melting point. Guntz and 

 Broniewski, on the other hand, found the resistance of the liquid to 

 pass through a minimum and to increase again in the unstable region 

 below the melting point. 



The ratio of the resistance of the solid to that of the liquid at the 

 freezing point was found from measurements of the resistance of the 

 solid at 0° in the glass capillary and the resistance of the liquid in 

 the same capillary. The resistance of the solid was extrapolated to 

 the melting point with the coefficient found. This procedure may be 

 open to some question, but it seemed as satisfactory as any other that 

 presented itself. The specific resistance of the solid was found to be 

 1.733 times that of the liquid at the melting point. Notice that the 

 relative magnitude of the volumes governs the relative magnitudes of 

 the resistance; the solid with the larger volume also having the larger 



