134 BRIDGMAN. 



is a function of the current density. Furthermore, this function is 

 known to be nearly independent of current density for low values. 

 Doubling the total current in a conductor of small section will produce 

 a much greater departure from Ohm's law than doubling the same 

 current in a conductor of larger section. This is the only sort of 

 departure from Ohm's law which we are looking for, or indeed which 

 seems at all likely, and in my opinion Maxwell's experiment is entirely 

 competent to answer this question up to its limits of error. As regards 

 the other sorts of departures from Ohm's law, I believe that Wenner's 

 position is sound. 



Since Maxwell, very few attempts have been made to detect the 

 effect, probably because of the appalling sensitiveness of Maxwell's 

 experiment as usually quoted. Lecher * in 1906 made measurements 

 on platinum and silver wires. He attempted to correct for the tem- 

 perature effect in the platinum by observing the thermal expansion 

 of the wire when carrying a very heavy current, and comparing the 

 resistance to this heavy current with the resistance to a feeble current 

 passing through the wire when heated artificially to have the same 

 thermal expansion as when carrying the heavy current. The diameter 

 of the wire was 1 mm., and the current density was 3.8 ,X 10^ amp/cm^; 

 there was no effect greater than 0.1%. The silver wire was 0.03 mm. 

 in diameter. It was placed in a rapid stream of water, and the 

 resistance measured under a current increasing until it fused. No 

 temperature correction was applied. The apparent temperature when 

 the wire burned out, using the ordinary temperature coefficient of 

 resistance, and assuming that Ohm's law is true, was 130°. It is to 

 be expected that this temperature would be somewhere between 100° 

 and the melting point of silver. Hence within limits of error which 

 may be several hundred percent, the experiment is consistent with 

 Ohm's law. The accuracy is very much less than that of Maxwell, 

 but on the other hand, the current density is very much higher, 

 reaching a maximum of 1.4 X 10^ amp/cm^, 25 fold greater. 



H. Rausch von Traubenberg ^ has attempted to avoid the tempera- 

 ture difficulty by employing a condenser discharge of very short 

 duration. The current densities that piay be reached are higher than 

 previously realized, attaining in one case a maximum of 10^ amp/cm^. 

 But the measurements of potential are inaccurate, being estimated 

 by the break-down of a spark gap in air, and there are other sources of 

 error arising from distributed capacities and inductances and the 

 necessity of a long range extrapolation. I estimate that the error may 

 certainly be as high as 10 or 15%. Within these limits, no effect was 



