106 BRIDGMAN. 



Mechanically the extruded wire is very crystalline and brittle, and 

 ^\^ll support almost no bending. When enclosed in a glass capillary, 

 however, and if the capillary breaks across without breaking the 

 gallium, as sometimes happens when the capillary is ruptured by 

 freezing, the metal may be pulled out by a tensile stress and appears 

 as ductile as lead. 



The extruded wire was used in an attempt to find the melting 

 cur\'e. Short lengths were layed horizontally between two copper 

 wires, short-circuiting them. It was expected that when melting 

 took place the system would open circuit. The surface tension of the 

 gallium proved to be so high, however, that this method did not work. 

 The metal, e^'en when melted, only sagged between the supports, 

 and no measurements could be made. Several modifications of this 

 scheme were tried with indifferent success. These preliminary meas- 

 urements made pretty evident, however, that there were no new 

 modifications, and that above the normal melting point the metal 

 remains liquid at all pressures below 12000 kg. 



Measurements were now made on the resistance of the liciuid as a 

 function of pressure above the melting temperature. The liquid 

 gallium was enclosed in a glass capillary, provided with four platinum 

 terminals, and measurements made by the potentiometer method. 

 The capillary was filled in high vacuum to avoid error from air stick- 

 ing to the walls. From a comparison of the resistance of the capillary 

 when filled with gallium and when filled with mercury it was possible 

 to obtain the resistance of gallium in terms of that of mercury, and so 

 the specific resistance of liquid gallium. 



After completion of measurements on the liquid, the temperature 

 of the apparatus was lowered into the region of the solid, and after 

 some trouble, the hquid was induced to freeze. The melting curve 

 was now determined by finding, as a function of temperature, the 

 pressure at which the resistance began to change discontinuously. 

 It was necessary to do this very cautiously. In changing pressure 

 the thermal eft'ects of compression might easily be sufficient to entirely 

 melt the gallium, when long and tedious manipulation would be neces- 

 sary to make it freeze again, because of the well known propertj' of 

 supporting great subcooling. In this way the melting curve was 

 mapped out to 12000 kg. 



INIeasureraents were also made on the resistance of the solid as a 

 function of pressure while enclosed in the glass envelope, but there 

 were irregularities, and it was evident that the solid must be uncon- 

 strained in order to give reliable results. Accordingly a small rod 



