ELECTRICAL EESISTANCE UNDER PRESSURE. 147 



particular substance, which does not change greatlj^ as pressure and 

 temperature are changed along the melting curve. We now have the 

 figures for the ratio of the resistance of solid to liquid for six metals at 

 different pressures and temperatures. For lithium the accuracy was 

 not high enough to permit more than the statement that the ratio does 

 not change greatly in a pressure range of 8000 kg. For sodium the 

 ratio is 1.45 at atmospheric pressure, and has dropped to 1.36 on the 

 melting curve at 12000 kg. The difference of volume between solid 

 and liquid has dropped to half its initial value in the same pressure 

 range, so that the ratio of resistances is evidently more constant than 

 the difference of volume. For potassium the ratio of resistance of 

 liquid to solid is 1.56 at kg., and has dropped only to 1.55 at 9700 kg. 

 Contrasted with this almost negligible change in the ratio of the 

 resistances is a decrease under 9700 kg. of the difference of volume 

 between solid and liquid to 0.31 of its initial value. For mercury, 

 I determined the ratio of resistance of liquid to solid at the melting 

 point at 0° and 7640 kg. to be 3.345. I did not make measurements 

 at any other temperature but there are values by other observers. 

 Onnes ^5 finds 4.22, Bouty and Cailletet ^^ 4.08, and Weber " gives 

 3.8 as the mean of six determinations, all for the ratio at the freezing 

 point at atmospheric pressure. The error is so large that it is not 

 possible to say more than that the change in the ratio along the melt- 

 ing curve is not large, and is in the direction of a decrease with increas- 

 ing pressure. The change is probably greater than the change in the 

 difference of volume between solid and liquid, which is abnormally 

 constant for mercury, there being a decrease in the difference of only 

 1% over the pressure range of 7640 kg. The change in the resistance 

 of the liquid over this range is, however, 19%, which is probably larger 

 than the change in the ratio of the resistance of liquid to solid. 



The behaA-ior of the two abnormal metals gallium and bismuth is 

 similar. At 7000 kg. I found the ratio of the resistance of liquid to 

 solid bismuth to be 0.45, and at atmospheric pressure Northrup and 

 Sherwood ^^ found 0.43. The ratio is probably constant within the 

 limits of error. For Gallium I found 0.58 for the ratio at atmos- 

 pheric pressure, and calculated the value at 12000 kg. to be 0.61. 

 This again is perhaps to be regarded as constant within the limits of 

 error, but it is noteworthy that the little variation there is is in the 

 same direction for both gallium and bismuth, and is toward an increase 

 with rising pressure, whereas what variation there was for normal 

 metals was always in the direction of a decrease wdth rising pressure. 

 We now compare the relative magnitudes of the pressure and 



