ELECTRICAL RESISTANCE UNDER PRESSURE. 105 



The actual work of purification was done by ]Mr. S. Boyer, under the 

 direction of Professor Richards. The final product had less than 

 0.01% total impurity. Professor F. A. Saunders was kind enough to 

 make a spectroscopic analysis for me, and was able to detect traces 

 of zinc as the impurity. There may also ha^-e been some indium 

 present. 



There was available for my measurements about one gram. This 

 was ample for a determination of the effect of pressure on resistance, 

 and also for a determination of the variation of freezing temperature 

 with pressure and an exploration for other allotropic modifications 

 (which had not been pre\'iously done) but was not sufficient for a 

 determination of the change of volume on freezing. The complete 

 freezing data are, therefore, not yec determined. 



The measurements here described include the specific resistance of 

 the liquid, temperature coefficient of resistance of solid and liquid 

 at atmospheric pressure, change of resistance when the solid melts 

 to the liquid, effect of pressure on resistance of both solid and liquid, 

 and variation of freezing pressure with temperature. 



The determination of the freezing curve and the exploration for 

 new modifications was the first task, in order to fix the range over 

 which the resistance measurements were to be extended. Gallium 

 is, of course, abnormal in that it expands when it freezes, and the 

 freezing temperature is accordingly depressed under increased pres- 

 sure. It would not be unnatural to expect new modifications at high 

 pressures like the other modifications of ice. Because of the limited 

 quantit}' of material available the method of exploration for new 

 modifications had to be an electrical one. It was my original inten- 

 tion to form the material into wire by extrusion, and to measure the 

 resistance of the solid as a function of pressure at different tempera- 

 tures. A change from one modification to another would be shown 

 by a discontinuous change of resistance, and melting by open circuit- 

 ing. Unexpected difficulties were found in the extrusion. If the 

 extrusion is peiformed at room temperature, the metal melts under 

 the one-sided stress instead of extruding (the melting temperature 

 is 29.85°), and if this effect is avoided by lowering the temperature, 

 the metal becomes so exceedingly stiff and brittle that extrusion is 

 very difficult. After an unsuccessful attempt at room temperature, 

 1 tried extrusion at the temperature of ice and salt. At this tempera- 

 ture the metal spit out of the die in short pieces. However, by careful 

 work, I did get a few inches of wire at this extrusion temperature. 

 Perhaps some intermediate temperature would be more successful. 



