PRESSURE ON RESISTANCE OF METALS. 605 



sure. The valuf 0.004293 is a trifle higher than Jaeger and Diessel- 

 horst's ^ highest value, 0.00428 for copper with less than 0.05% of Zn 

 or Fe. The initial value of the pressure coefficient at 0°, found graph- 

 ically from the deviation curves was — O.O52OO8. This is to be com- 

 pared with — O.O5I87 of Lisell.^ He does not give the temperature 

 coefficient of his Cu, but states that chemical analysis showed no trace 

 of any foreign metal. The difference in our values cannot be due to 

 difference of thermal or mechanical treatment, because Lisell found 

 no perceptible difference between the pressure coefficient of the same 

 material when in the annealed or hard drawn state. 



If the resistance-pressure curves are scaled to the same initial 

 resistance, the curves for higher temperatures are less steep, and of 

 much less curvature. There is the same abnormal march of the 

 pressure of maximum deviation shown by silver and gold. 



Nickel. This was obtained from the Electrical Alloys Co. of 

 Morristown, N. J., and was stated by them to be of high purity, 

 although they gave no analysis. The purity was not as high as it 

 should be, as will be seen later from the temperature coefficient. It 

 was provided by them in the form of wire 0.003 inch diameter, double 

 silk covered. For the measurements under pressure it was wound 

 into a coreless toroid of 176 ohms resistance at 0°. The initial resist- 

 ance was chosen so high because the pressure effect is very small. It 

 was seasoned by many times heating and cooling between 0° and 140°, 

 and by a single exposure to 12000 kg. at 25°. Connections were made 

 with silver solder. 



A few readings were made on another sample of nickel, kindly fur- 

 nished by Leeds and Northrup. It is of the grade used by them in 

 resistance thermometers, and is of high purity. It was probably not 

 so pure as that used in the final readings, however, because its tempera- 

 ture coefficient was 3% lower. The results obtained with it are only 

 of orienting value, because the temperature during annealing was acci- 

 dentally allowed to get so high as to slightly discolor the silk insula- 

 tion. At 25° the average pressure coefficient was 4% less than that oF 

 the purer sample. 



The smoothed results are collected in Table XI and the experi- 

 mental values of mean coefficient and deviation from linearity are 

 shown in Figure 13. The results for nickel are not quite as regular as 

 for many other metals, because the pressure coefficient is unusually 

 small compared with the temperature coefficient. The zero drifts 

 were not large, however, being respectively 1.1%, 0.16%, 0.2%, 0.1%, 

 and 0.3%o of the total pressure effect at 0°, 25°, 50°, 75°, and 100°. 



