142 BRIDGMAN. 



At high frequencies /(w) °^ — , so that at high frequencies the micro- 



cc 



phone effect is proportional to the reciprocal of the frequency, and it 



may be eliminated by proceeding to infinite frequencies (or zero 



reciprocal frequency) . 



The procedure suggested by this analysis was that followed in the 

 experiment. For a fixed D.C. the difference between the settings at 

 D.C. and A.C. balance (that is AR') was observed over a range of 

 frequencies, AR' was plotted against the reciprocal of frequency, and 

 extrapolated to zero. The residual, if there is one, is the effect due 

 to deviation from Ohm's law at the particular D.C. density in question. 



This procedure was repeated for a number of currents, and so the 

 departure from Ohm's law obtained as a function of current density. 



Before proceeding further with the theoretical discussion it will 

 pay now to describe the experimental details, in order that we may 

 have an idea of the order of magnitude of the quantities involved. 



Experimental Details. 



The bridge was an ordinary four gap alternating current bridge, so 

 constructed that inductive and capacity effects in the bridge were 

 negligible. The resistance Ro, which was kept constant during the 

 measurements on any single specimen, was a coil of heavy manganin 

 wire immersed in an oil bath to carry away the Joulean heat. This 

 resistance was approximately equal in magnitude to the resistance x 

 which carried the high current density. The resistances R3 and Ri 

 consisted of heavy manganin coils connected by a slide wire, which 

 was tapped by a slider. The wire was about 1 meter long, with a total 

 resistance of about 3 ohms. The resistance of the extension coils was 

 five or ten times greater than that of the specimen x, and the genera- 

 tion of heat in them was so small that it was not necessary to immerse 

 them in an oil bath. 



The method of connecting the D.C. and the A.C. sources and of 

 tapping across with the detectors for D.C. and A.C. balance is shown 

 in Figure 3. The direct and alternating current sources are connected 

 to the same terminals of the bridge, with a large inductance L in the 

 D.C. line to prevent the A.C. backing into the battery, and a large 

 condenser C in the A.C. line to prevent the D.C. backing into the 

 A.C. source. D.C. balance was shown by a Leeds and Northrup high 

 sensitivity galvanometer of about 8 ohms internal resistance connected 



