PRESSURE ON RESISTANCE OF METALS. 611 



The mean temperature coefficient between 0° and 100° given above 

 is 0.006206. This value was found from a coil of the same iron as that 

 above, seasoned for temperature in the same way, but not subjected 

 to pressure. The mean temperature coefficient of the coil used for 

 pressure measurements was 0.006201. The agreement is unusually 

 good, and shows very little internal change produced in this material 

 by the application of pressure. I have adopted the value 0.006206 as 

 more probably that of the pure metal in a state of complete ease. 

 The unusually large departure of the relation between temperature 

 and resistance from linearity is to be noticed. The value found above 

 for the temperature coefficient of iron is exceptionally high, as is to 

 be expected from its exceptional purity. Kaye and Laby give for 

 pure iron 0.0062, essentially the same as the above. Dewar and 

 Fleming ^^ give 0.00625, for an iron said to be very pure, but without 

 analysis. The large effect of impurity is to be inferred from data of 

 Jaeger and Diesselhorst,^ who give for a sample composed of 99.55 

 Fe, 0.1 C, 0.2 Si, 0.1 Mn, trace of P, S, Cu, the value 0.00461, and for 

 a purer iron with 0.1% C, other metals not determined, 0.00539. 



The initial pressure coefficient at 0° has been found to be —0.05246 

 by Beckman,^ against O.Oo2405 found graphically from the deviation 

 curves above. The discrepancy is not as great as usual between our 

 results, and is in the opposite direction. 



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

 resistance at different temperatures the curves for higher temperature 

 become slightly steeper, but are less curved. 



Palladium. This was furnished by Baker and Co. in the form of 

 wire 0.04 inch diameter. It was not etched, but was from this size 

 drawn down to 0.003 inch in diamond dies, annealed to redness, and 

 single covered with silk by the New England Electrical Works. It 

 was wound into a coreless toroid of 68.8 ohms resistance at 0°, and 

 seasoned at the same time as the cadmium. Connections were made 

 by silver soldering. 



The smoothed values ore collected in Table XIV, and the experi- 

 mental values of mean coefficient and maximum deviation in Figure 16. 

 The zero drifts during the runs were respectively 0.3%, 0.07%, 

 0.24%, 0.03%, and 0.9% of the total pressure effect at 0°, 25°, 50°, 

 75°, and 100°. Pressure was transmitted at all temperatures by 

 kerosene; it would be possible to increase the regularity of the results 

 by using petroleum ether at 0° and 25°. The effect of viscosity seemed 

 to be greater than usual. Except at 0°, the greatest departure of any 

 single reading from a smooth curve was 0.2% of the total pressure 



