252 



A. W. Davison 



Since the electromotive force of different couples varies widely, 

 a number of values of >S were used, and the galvanometer stand- 

 ardized for each one. The following values of S were used: 0, 

 350, 600, 2100, 5000, and 6000 ohms. 



Method. The terminals of the thermo-couple were connected 

 to the galvanometer, with the box S in series, as per fig. 9. 



In order to insure that the only thermo-electric effects in the cir- 

 cuit were those to be measured, three junctions were used, Cu-M' 

 (e); M'-M" (f); and M"-Cu (g). The junctions (e) and (g) were 

 maintained at 0°, and (f) was raised to various values of t By 

 recording the temperature of (f ) , and the deflections of the gal- 

 vanometer, the various electromotive forces of the circuit may be 

 determined, and curves plotted. 



Junctions (e) and ^g) were placed in separate kerosene baths^ 



Fig. 8. C is a copper wire to the galvanometer; L, a lead wire, M, a glass tube 

 containing mercury. 



and these baths placed in the same ice-bath. This separation 

 of the baths was effected to overcome any galvanic current that 

 might be set up. Kerosene was used, since it does not corrode 

 the metals. 



Junction (f) was heated in a bath of heavy lubricating oil, 

 which is non-corrosive, and is not volatile at the temperatures 

 used. 



In order to null thermo-effects at the various junctions (or 

 rather connections) of the system, the junction (f) was first 

 maintained at 0° by placing it in a third bath similar to those of 

 (e) and (g), and "setting" the galvanometer at zero. These other 

 thermo-effects were due to a number of causes, radiation from 

 the body, from an incandescent lamp used for lighting the scale, 

 etc. 



