Thermo-electric Couples 



253 



Errors, which were afterward remedied, were introduced into 

 the first few sets of readings taken, due to the following causes: 



(a) The coil of the galvanometer was not hung in a uniform 

 magnetic field, this was discovered in calibration. 



(b) When the temperature was gradually raised, and after 

 the flame had been removed, the deflection of the galvanometer 

 would reach a maximum, and decrease, before the thermometer 

 had reached its maximum. This "time drag" was due to the 

 fact that glass, being a poor heat conductor, did not transmit 

 heat to the mercury as rapidly as heat was transmitted to the 

 couple. 



(c) A great amount of trouble was encountered in getting a 



Fig. 9. Arrangement for measuring electromotive forces. 



suitable method of joining the elements, i. e., of producing the 

 junctions themselves. The following methods were tried: 



(1) Simple twisting. This scheme proved very satisfactory 

 when the junction was wanted for immediate use; but after a 

 few hours the results varied, due to corrosion, the metals oxidiz- 

 ing even where they were twisted the tightest. 



(2) Soldering the wires. The use of solder in itself doss not 

 necessarily alter the thermo-electric effect, since from the law 

 of intermediate metals, the electromotive force is not changed 

 by the introduction of a third metal. But the flux necessary 

 for soldering has a chemical effect whose thermo-electric charac- 

 teristics are different. The presence of a mass of solder also 

 tends to give a shunt across the junction, which permits the 

 return of current. 



