76 BRIDGMAN. 



cases), the resistance of the soUd was extrapolated to the melting 

 curve at 202.5° and 8430 kg., and at this point the ratio of resistance of 

 liquid to solid found to be 1.49. The accuracy is not high, and 

 probably the only conclusion justified is that the change of the ratio 

 along the melting curve is not large. 



Sodium. The material was from the same sample as that previ- 

 ously used in determinations of the change of melting point under 

 pressure.* No chemical analysis has been made, but the sharpness of 

 the melting is evidence of the high purity. Two series of measure- 

 ments of the resistance were made. The first, in the spring of 1919, 

 was on the resistance of the solid below the melting point at atmo- 

 spheric pressure. The second series, in the fall and winter of 1919-20, 

 was on the resistance of both solid and liquid at temperatures above 

 97.6°. 



For the measurements of the resistance of the solid below 97.6° the 

 sodium was extruded into wire through a steel die, and the wire 

 wound bare on a bone core. As in the case of Lithium, I am not aware 

 that measurements have been made previously on the resistance of the 

 bare wire, but the material has been enclosed in a glass envelope. 

 The details of manipulation were essentially the same as for lithium, 

 except that sodium is much softer, and consequently harder to handle. 

 Two sizes of wire were used, at first 0.015 inches in diameter, and 

 later 0.030. The larger size is considerably easier to handle, and the 

 resistance is high enough to give the requisite accuracy. The initial 

 resistance was of the order of 0.5 ohms for the small wire, and 0.2 ohms 

 for the larger wire. Connections were made to the sodium wire with 

 spring clips, as with lithium. The springs were helical springs of 

 piano wire, 0.009 inches diameter, wound on a mandrel 0.06 inches 

 diameter. The springs were pulled out so as to separate the spires, 

 and the sodium wire dropped in between the spires. The wire was 

 protected between the spires by -RTapping round it a covering of gold 

 or silver foil 0.001 or 0.002 inches thick; otherwise the soft wire is 

 pinched off by the spring. Due to chemical action, the resistance at 

 the spring contacts sometimes became so high as to make trouble. 

 Contact could always be restored by passing through the contact a 

 current from a small magneto. 



Experiments were made on six different samples before a complete 

 set of readings was obtained. Repetitions were necessary because of 

 the effect of corrosion, and also because of accidents to the insulation 

 of the three-terminal plug. In addition to complete runs at 0°, 25°, 

 50°, and 75°, one fragmentary run was made at 0°, and tliree at 50°. 



