THERMO-ELECTRIC QUALITY UNDER PRESSURE. 279 



the current in the potentiometer circuit varied by means of a suitable 

 resistance until balance is obtained. This has the advantage over the 

 ordinary slide wire form of avoiding all possible e.m.f's. at the sliding 

 junctions. In order to provide the very small e.m.f's. required, the 

 known source of e.m.f. was stepped down twice. The comparison 

 cell was a battery of ten acid AYeston cells connected in parallel. The 

 advantage of the acid cell is that its e.m.f. does not fluctuate, even 

 when a comparatively large current is drawn. This was most kindly 

 loaned to me by Professor H. N. Davis, by whom it had been person- 

 ally constructed. I am also indebted to him for a number of the other 

 pieces of standard apparatus used. The e.m.f. of the acid cell varies 

 somewhat with the external resistance; it was measured by standard 

 potentiometer methods against a standard Weston cell for external 

 resistances from 1000 to 10000 ohms. The e.m.f. at the terminals 

 varied under these conditions from 1.0410 to 1.0425 volts. This was 

 measured several times in the course of the work, and the voltage 

 under any given external resistance remained constant within 1/100%. 

 The temperature effect was negligible for temperature fluctuations 

 within the range of this work. 



The working cell was connected through a reversing switch with a 

 10,000 and a 10 ohm. coil in series. The 10 ohm coil was tapped by 

 a second circuit containing a variable Leeds and Northrup decade 

 box reading to 10,000 ohms, and guaranteed to 1/10%, and three 

 other coils, of seasoned copper, of 10, 1 and 0.1 ohm. According to 

 its magnitude, the e.m.f. in the pressure circuit could be tapped across 

 either the 10, 1, or 0.1 ohm coil. Readings were made by a null 

 method by the method of reversals. Reversing switches in the pres- 

 sure circuit, but not in the galvanometer circuit, and in the second 

 stage of the potentiometer circuit allowed the reversal of both the 

 e.m.f. being measured and the balancing e.m.f. The decade box was 

 set until no throw was obtained on reversing both these circuits. 

 The range covered with the apparatus set up as above was from 

 approximately 10~* volts, with 1000 ohms in the decade box and 0.1 

 ohm tapped, to 10"^ volts with 1000 in the decade box and 10 ohms 

 tapped. For larger voltages the working cell could be directly con- 

 nected wdth the potentiometer circuit, increasing the range in this 

 way from 10"^ to 10"^ volts, still with never less than 1000 ohms in 

 series with the working cell. E.m.f's. less than 10~* volts were read 

 by the deflection of the galvanometer without attempting to balance. 

 Of course the e.m.f's. corresponding to the stated resistances are not 

 exactly those given, but there are slight corrections. The corrections 



