ELECTRICAL RESISTANCE UNDER PRESSURE. 



71 



ness of Dr. A. W. Hull of the Research Laboratory of the General 

 Electric Co. The purity was not known. It was prepared by 

 electrolysis of the fused chloride with graphite electrodes ; the chloride 

 was pure, so that any impurities were introduced in the electrolysis. 

 There were inclusions in it of some slag-like substance which had to be 

 cut out as well as possible. This specimen was used for a determina- 

 tion of the melting cur^*e. The method was that of the discontinuity 

 of volume, which I have previously used in determining melting or 

 transition curves, and has been fully described elsewhere.^ The 

 volume of the specimen was about 5 c.c. It was placed in an iron 

 container, and pressure transmitted to it with Nujol. It was evi- 

 dently somewhat impure, for the corners of the melting curve were 



TABLE II. 



Melting Curve of Lithium. 



considerably rounded, melting being detectible at least 1000 kg. be- 

 fore the end of melting. This corresponds to a spreading of melting 

 at constant pressure over a temperature range of 3.5°. What the 

 depression of the freezing point is at the conclusion of melting, that is, 

 how much the observed curve should be raised in order to correct for 

 the impurity, it is not possible to state from the data. Points on the 

 freezing curve were obtained at three temperatures. A curve was 

 drawn through these points and the data from the smooth curve are 

 given in Table II. The curve and the observed points are also shown 

 in Figure 2. The points lie on the curve as drawn, but it is seen that 

 there may be some uncertainty about the extrapolation to atmospheric 

 pressure. The value of the melting temperature at atmospheric 



