ELECTRICAL RESISTANCE UNDER PRESSURE. 75 



The pressure coefficient of the solid Hthium was measured in the 

 capillary at 171°. The most important result of this measurement was 

 the verification of the positive coefficient. The numerical value of 

 the coefficient cannot be accepted, however, because of the restraining 

 effect of the capillary. A similar effect had already been met in the 

 case of Gallium. The results for the resistance of the solid were very 

 irregular, there being deviations of 10% from the mean curve. The 

 value found for the coefficient was + O.O4I59, nearly twice the value 

 found at low temperatures for the bare wire. It is quite possible that 

 part of this large difference is real, since we in general expect the 

 coefficient to become larger at the higher temperatures. It was not 

 possible to definitely state any variation of the coefficient within the 

 range 0° to 100° for the bare wire, since chemical action cut down the 

 accuracy of the measurements. It is also possible that some of the 

 difference may be due to difference of the material, this latter speci- 

 men not being so pure. The pressure coefficient of the solid in the 

 capillary was also measured at 94°, but the irregularities of the data 

 were much greater than at 171°. One would expect the effect of con- 

 straint to become greater farther from the melting point. The best 

 value of the coefficient at 94° was 2-5% higher than at 171°, but the 

 accuracy was so low that it is not at all certain that there was any real 

 difference. 



The temperature coefficient of the solid in the capillary was also 

 measured between 0° and 171°. Here again we should not expect the 

 results to be very accurate because of the effects of constraint. The 

 mean coefficient o^er this range was 0.0039. This is considerably less 

 than for the unconstrained wire; part of the difference may be due to 

 greater impurity. 



The change in the resistance on melting at atmospheric pressure 

 could be computed from the measurements on solid and liquid separ- 

 ately. The specific resistance of the liquid is thus found to be 1.68 

 times that of the solid at the melting point at atmospheric pressure. 

 The only other published value is by Bernini,^ who found 2.51. 

 Bernini's value of the melting point was 177.84°, somewhat lower than 

 the value given above, so that it is not probable that the difference is 

 to be ascribed to greater purity of his sample. I also made an attempt 

 to find the variation of the ratio of specific resistance of solid and liquid 

 along the melting curve, but this cannot be very accurate, because of 

 the constraining effect of the capillary. Using the value found for the 

 pressure and temperature coefficients of the solid in the capillary, and 

 assuming that the temperature coefficient is not affected by pressm-e 

 (which has been proved to be true by direct measurement in most 



