ELECTRICAL RESISTANCE UXDER PRESSURE. 97 



of the selected pieces of clear metal. The extrusion is considerably 

 easier than that of calcium, and may be successfully performed at a 

 temperature of 230°. At room temperature the metal spits out of the 

 die in small pieces. The wire is fairly soft and pliable and can be bent 

 to a radius of perhaps ten times the diameter of the wire, but it is quite 

 different in mechanical properties from the alkali metals, such as 

 sodium, and shows brittleness if too sharp a bend is attempted. 



Previous measurements on the electrical properties of strontium 

 seem to have been published only by Matthiesen.^ He extruded the 

 metal to wire in much the same way as above. He gives for the 

 specific resistance at 20° the value 25 X 10^^ ohms per cm. cube. I 

 found at 0° for my specimen the value 30.7 X 10~^. This specimen 

 is presumably considerably purer than Matthiesen's, and this value 

 would seem to be preferred. Matthiesen did not attempt to measure 

 the temperature coefficient of resistance, probably because of resist- 

 ance of the contacts. He did not use a potentiometer method, but 

 had to correct as best he could for the resistance of his leads and con- 

 tacts. ^Yith a potentiometer method as used here, there is no such 

 difficulty. There is difficulty, however, in the chemical action 

 accompanying changes of temperature, which produces permanent 

 changes of resistance. Error from this effect was avoided by the 

 same procedure as that used previously for the alkali metals. Four 

 thermostated baths were kept simultaneously in operation, and the 

 specimen transferred rapidly from one to another. The bare wire was 

 immersed in a glass tube of Nujol for the measurements, the Nujol 

 having been previously heated with sodium to remove all moisture 

 and exhaust as far as possible all tendency to chemical action. There 

 nevertheless seems to have been some specific action between the oil 

 and the strontium. About fifteen minutes were required for the 

 attainment of temperature equilibrium after the wire had been 

 transferred from one bath to the next. Readings were made with 

 ascending and descending temperature between 0° and 96°, starting 

 with 0° and ending with 0°. The mean of the ascending and descend- 

 ing readings was taken as the true effect. The permanent change of 

 zero after the excursion was 12% of the maximum effect. The mean 

 coefficient between 0° and 100° found from these readings was 0.00383. 

 The effect is not quite linear w^ith temperature, but the change becomes 

 more rapid at the higher temperatures, as is normal. 



The value of the temperature coefficient is quite normal for pure 

 metals, and in the absence of further information, makes it probable 

 that this material was of satisfactory purity. It is not possible to 



