118 



BRIDGMAN. 



the kindness of the manufacturers I have since been able to measure 

 the resistance of two samples of "undoped" tungsten of high purity. 

 The ^Yestinghouse Lamp Works ga^-e me a specimen which they esti- 

 mated to contain less than 0.03% total impurity, and the Research 

 Laboratory of the General Electric Co. placed at my disposal a speci- 

 men which they estimated to be even purer. Judging by the test of 

 the temperature coefficient the General Electric sample was appreci- 

 ably purer. The average temperature coefficient between 0° and 100° 

 of the Westinghouse sample was 0.003925, and that of the General 

 Electric sample 0.004209. 



Pressure measurements were made on both samples, but only those 

 on the purer are given in detail here. The treatment of both speci- 

 mens was the same. The wire was 0.002 inches in diameter; it was 

 wountl bare on a bone core, and connections were made by fusing to it 

 pure nickel wire with an arc in hydrogen. This method of making 



TABLE XIV. 



TrXGSTEN. 



connections was taught me by the General Electric Co., and is simpler 

 than the combination of gold and platinum which I formerly used, 

 although the former connection was just as satisfactory electrically. 

 The wires were seasoned by a long preliminary heating to 125° and 

 by an application of 12000 kg. Three runs were made, at 0°, 50°, 

 and 95°. Except for two bad points, the greatest departure of any 

 of the observed points from a smooth curve at any of the three tem- 

 peratures was 0.2% of the total eft'ect, and the displacement of the 

 zero after a run was not greater than the irregularity of the other 

 points. The observed results were smoothed, and a table constructed 

 for the resistance as a function of pressure and temperature by regular 

 methods. The results are shown in Table XIV and Figure 9. The 



