PRESSURE ON RESISTANCE OF METALS. 603 



temperature coefficient of Beckman's gold between 0° and 100° was 

 0.00390, nearly as high as the above. 



The general character of the results is much the same as for silver. 

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

 resistance, the curves at higher temperatures become slightly steeper 

 but the curvature becomes less at an accelerated rate. There is also 

 the same march toward higher pressures of the pressure of maximum 

 deviation, but here it is more pronounced than for silver. 



Copper. I am indebted to the Bureau of Standards for this material; 

 it was supplied from the stock of materials which had been used there 

 in an extensive series of tests on the conductivity and temperature 

 coefficient of copper from diiferent sources. This was furnished by 

 them in the form of wire 0.06 inch diameter. It had been forged from 

 a bar of electrolytic copper without melting after the electrolytic 

 refining. They were not in a position to supply the chemical analysis, 

 but stated that it had proved to be of very high conductivity, and was 

 in all probability purer than another sample which they offered me 

 which showed by analysis 99.995% Cu, trace of S, and no Ag, CU2O, 

 As, or Sb. From 0.06 inch I reduced the wire to 0.046 inch by etching 

 with nitric acid, to remove any contamination of iron introduced in 

 the drawing. From this size it was drawn to 0.003 inch through 

 diamond dies, annealed, and single silk covered by the New England 

 Electrical Works. For the high pressure measurements it was wound 

 into a coreless toroid and seasoned for temperature and pressure at the 

 same time as cadmium. Its initial resistance at 0° was 35.5 ohms. 

 Connections were made with silver solder. 



The smoothed results are collected into Table X and the experi- 

 mental values of average coefficient and maximum deviation from 

 linearity are shown in Figure 12. This substance was one of the first 

 investigated, before all the details of manipulation had been perfected. 

 Pressure was transmitted by kerosene at all temperatures instead of 

 by petroleum ether at 0° and 25°, so that the error from viscosity is 

 larger than necessary. In spite of this, however, the maximum depar- 

 ture from a smooth curve, except the one most viscous point at 0°, 

 was only 0.3% of the total effect, and the numerical average was 0.05%. 

 The zero drift at 75° and 100° was larger than usual, amounting to 2%, 

 because the thermostatic control, which is more difficult at higher 

 temperature, had not been perfected. At 50° the drift was 0.8%, at 

 25° 0.07%, and at 0° 0.00%. The curves of deviation from linearity 

 are distinctly not symmetrical, and the pressure of maximum deviation 

 progresses regularly with rising temperature. The curves are repro- 

 (kiced in Figure 12. 



