ELECTRICAL RESISTANCE UNDER PRESSURE. 149 



increasing pressure. Measurements were not made on the variation 

 with pressure of the temperature coefficient of the soUd 



Comparison cannot properly be made between liquid and solid 

 bismuth, because the solid is abnormal and the lic[uid is normal. It 

 is interesting, however, that numerically the coefficient of the solid is 

 greater than that of the licjuid. This may mean that some of the 

 tendency to abnormality still persists in the liquid, making its coeffi- 

 cient lower than it would otherwise be. 



Lithiiun is abnormal in both liquid and solid. If the data for the 

 solid are extrapolated from 100° to the melting temperature at 180° 

 the figures given would intlicate a pressure coefficient of the solid 

 numerically less than that of the liquid. The difference would be still 

 further accentuated if the unknown correction for the compressibility 

 of the solid is applied so as to make the coefficients of both solid and 

 liquid the coefficients of specific resistance. Although the coefficient 

 of the liquid is greater than that of the solid, its variation with pres- 

 sure is nuich less, and in fact is opposite in sign, the coefficient of the 

 licpid becoming smaller at higher pressures, and the coefficient of the 

 solid becoming greater. The temperature coefficient of liquid lithium 

 is independent of pressure to 12000 kg., as is that of the solid also. 



Summarizing the relations between the coefficients of the liquid 

 and the solid, except for the temperature coefficient of the liquid being 

 less than that of the solid, there does not seem to be a tendency to 

 any one type of behavior. It is noteworth;)-, however, that in many 

 cases the resistance of the liquid responds more sluggishly to changes 

 of pressure than does that of the solid, the coefficient of the liquid being 

 actually less than that of the solid, or else the change of coefficient 

 with pressure being less for the liciuid. 



Alloys. The above data on alloys are entirely unsystematic and 

 fragmentary, so that it is not possible to draw any conclusions as to the 

 behavior of alloys in general. It is interesting to notice, however, 

 that the pressure coefficient of all the alloys, with the exception of 

 that of " Comet," is less numerically than would be computed by the 

 law of mixtures from the coefficients of its components, and in the 

 case of "Therlo" this tendency to a lower value may go so far as to 

 rcA^erse the sign. In making this statement I have assumed that the 

 pressure coefficient of pure Chromium and Manganese is negative, 

 a conclusion which has not been checked by experiment, but which 

 seems very probable from the behavior of similar metals. 



