ELECTRICAL RESISTANCE UNDER PRESSURE. 95 



The final results for resistance as a function of pressure and tempera- 

 ture are given in Table VII. The striking and unexpected result, as 

 in the case of lithium, is the positive coefficient. The coefficient is 

 greatest at the lowest temperature, having the average value + O.O4I 292 

 at 0°. Both the instantaneous and the average coefficient increase 

 with rising pressure. In both these particulars the behavior is much 

 like that of bismuth. The absolute value of the coefficient is of the 

 order of half that of bismuth. 



Bismuth is abnormal in so many other respects than its pressure 

 coefficient, that it seemed of interest to find whether calcium has the 

 same sort of abnormalities. The large increase of resistance of bis- 

 muth in a magnetic field is one of the well known abnormal effects. I 

 tried the effect of a field of approximately 27000 Gauss on the resistance 

 of a coil of calcium, the wires of the coil being everywhere at right 

 angles to the field, and found a decrease of resistance of only 1/1600. 

 It is evident, therefore, that any parallelism between the conduction 

 mechanisms of calcium and bismuth cannot be very complete. A 

 more exliaustive investigation of the various properties of calcium 

 than has yet been made seems well worth while. 



In a previous paper it has been shown that the abnormal positive 

 pressure coefficients of resistance of iMsmuth and antimony are asso- 

 ciated with values of the thermal expansion abnormal as compared 

 with the other properties. The same comparison for calcium appeared 

 of interest, but the thermal expansion of calcium has apparently not 

 been previously measured, and I accordingly made special measure- 

 ments of it. 



The thermal expansion was determined by ordinary methods, 

 using a glass dilatometer. The calcium was the same si^ecimen as 

 that from which the resistance sample was cut. It was turned over 

 its entire surface in the lathe, and finished to a cylinder about 6 cm. 

 long and 20 cm^ volume. The surface was smooth, without blow^- 

 holes. In the glass dilatometer the calcium was surrounded with 

 Xujol. The calcium and Xujol had been previously heated together 

 in another vessel to 100° to avoid as far as possible chemical action 

 during the dilatometer readings. Discoloration of the surface of the 

 calcium I^y this preliminary heating was very slight. Air bubbles in 

 the dilatometer were avoided by filling in vacuum. The Nujol is 

 so viscous that without special precaution error may be introduced 

 by sticking of the Nujol to the walls of the capillary. This error was 

 aAoided l\v heating the capillary before the readings with a small gas 

 flame, and by making readings with increasing temperature. A pre- 



