164 Barker — Thermoelectromotive Forces of Potassium 



heated, generally to boiling, or . which was colder than the air 

 of the room, and corresponding values of temperature and 

 electromotive force read at intervals as the temperature of this 

 junction fell or rose, chiefly through radiation. Inasmuch as 

 the least difference in electromotive force between successive, 

 points at which the potentiometer, as it was constructed, could 

 be set was 10 microvolts, it was the uniform practice to set the 

 potentiometer at the point whose corresponding electromotive 

 force would next be reached, in following the variation in tem- 

 perature in either direction ; and, depressing the galvanometer 

 key shortly before, watch the slowly decreasing deflection until 

 it was reduced to zero, when the temperatures of the hot and 

 cold junctions were read in close succession. No difficulty was 

 met with by reason of variation of zero point, and none from 

 variation of electromotive force of the storage cell used to sup- 

 ply the potentiometer current, this latter variation being very 

 slow and of small extent. The deflection of the galvanometer 

 followed the variations in temperature closely and regularly. 

 Under the circumstances of actual measurement a variation of 

 the setting of the potentiometer of one step of 10 microvolts 

 caused a difference of deflection of the galvanometer of over 

 20 scale divisions (millimeters), each of which was very easily 

 divisible by the eye into at least five parts ; I conclude, there- 

 fore, that I would be quite justified in stating the electromo- 

 tive forces measured to 0*1 microvolt, if this seemed desirable. 



The results obtained are embodied in the subjoined tables 

 and in the accompanying curves. In the tables the column 

 headed E gives the electromotive force in microvolts at which 

 the potentiometer was set ; that headed T gives the tempera- 

 ture difference between junctions, one of which was invariably 

 near zero. The thermometer readings have been corrected for 

 the determined variation in the positions of the fixed points, 

 but in no other way. The value of T given is the mean of all 

 the observed values. 



In each of the three cases the electromotive force developed 

 was such as to cause current to flow from the enclosed metal 

 to the platinum at the heated junction. The electromotive 

 forces are therefore all of the same sign and are treated as 

 positive. 



Under 1ST is stated the number of observations made at the 

 corresponding potentiometer setting. This must not be 

 regarded as indicating the number of separate series of obser- 

 vations made with a given couple, or even with a given couple 

 in a given range. Thus, in the case of the potassium-platinum 

 couple, eight series of observations were made at different 

 times during a period of about two weeks, and four of these 

 embraced the higher range of temperature in common. In no 



