112 BRIDGMAN. 



resistance. Guntz and Broniewski found 2.09 for the ratio of the two 

 resistances. 



Titanium. By the kindness of the Research Laboratory of tlie 

 General Electric Co. I was enabled to make measiirements on a fila- 

 ment of titanium deposited on tungsten, which had been used for 

 experimental work with incandescent lamps. The dimensions of the 

 tungsten core were such that the total impurity of tungsten was only 

 1.8%. The method of deposition of the titanium on the core is not 

 known to me; the surface of the filament was distinctly crystalline 

 in appearance, probably due to recrystallization after deposition. It 

 had been glowed out in yacuum at high temperature after deposition 

 in order to remoA^e impurities of hydrides, nitrides, and oxides, all of 

 which are readily formed with this substance. This glowing out must 

 have produced alloying with the tungsten core, and the alloy so formed 

 is evidently quite different in its properties from the pvire metal. This 

 may be stated with confidence because the General Electric Co. 

 found for the specific resistance of this filament the value 350 X 10"* 

 ohms per cm. cube, which is higher than would be given by the tungsten 

 core alone. The same thing is indicated by the low value of the tem- 

 perature coefficient, w^hich was 0.000221. 



The difficulties of the pressure measurements were very great, and 

 it was not possible to obtain results which were at all regular. It can 

 be stated only that the pressure coefficient is exceedingly small, 

 probably not greater than 10~^ per kg., and that the likelihood is that 

 the resistance increases with pressure. 



Zirconium. Two filaments deposited on tungsten in the same 

 way as the titanium were made available through the kindness of the 

 General Electric Co. The treatment of the filaments had been the 

 same as that of titanium. It is probable that the temperature of 

 glowing out had been sufficient to produce alloying with the tungsten 

 core. This is strongly suggested by the low A'alue for the tempera- 

 ture coefficient of resistance, which, between 0° and 100°, was 0.00004 

 for one specimen, and 0.00058 for the other. The dimensions of the 

 wires on which these filaments were deposited would indicate a total 

 impurity of tungsten of 1.8% and 0.6% for the two samples respec- 

 tively. The impurer sample has the smaller coefficient, as is usual. 

 The exceedingly low value of both coefficients indicates that the 

 impurity has a specific effect, and that any results found for the pres- 

 sure coefficient may not be very close to the values for the pure metal. 



In view of the probably large effect of the impurity, and also of 

 the difficulty of the measurements, a great deal of effort \yas not 



