354 PROCEEDINGS OF THE AMEEICAN ACADEMY. 



Le Roux * in 1867 confirmed the results of Lord Kelvin. Working 

 with a more sensitive apparatus, he obtained relative values of the Thom- 

 son Etfect in several metals, and was able to show that, with a given 

 gradient of temperature, the effect, as expected, was directly proportional 

 to the current. 



Tait, in deducing a formula to express the relation between the.e. m. f. 

 of a thermo-electric couple and the diflTereuce in temperature of the junc- 

 tions, assumed that the Thomson Effect was directly proportional to the 

 absolute temperature. The relation deduced in this way agrees very 

 well but not exactly with the results of most experimenfs. In some 

 recent experiments in which a Callendar Platinum Resistance Thermome- 

 ter was used, Mr. H. M. Tory | has found a regular deviation of the 

 experimental results from the results as given by Tait's formula. 



Batelli t has made absolute measurements of the Thomson Effect for a 

 number of metals, and has found that in most cases Tait's assumption is 

 approximately verified. As Batelli's is the only work I have found 

 on the absolute measurement of the Thomson Effect, it is worth while 

 to examine his method closely and to point out the possible sources of 

 error. 



Batelli prepared two rods of the metal in which the Thomson Effect 

 was to be measured, about 30 cm. long and 0.5 cm. in diameter. One end 

 of each rod projected into a heating bath H, the other end into a cooling 

 bath, C, Fig. 1. There was then a flow of heat through the rods from 

 the hot to the cold bath. The rods passed through iron boxes Bj and 

 Bg, 4 cm. long, containing mercury, which served as calorimeters. 



When an electric current was passed up one rod and down the other, 

 that is with and against the flow of heat in Bi and Bo, respectively, the 

 Thomson Effect liberated heat from one bar and absorbed it in the other. 

 This caused the temperature of one calorimeter to rise above, that of the 

 other to fall below, the mean temperature of the calorimeters when no 

 electric current was flowing. From the difference in temperature of the 

 calorimeters measured with a thermo-electric couple and the water 

 equivalent of the calorimeters, Batelli calculated the sum of the heat 

 liberated and absorbed in a given time by the Thomson Effect. 



The method of working was as follows. The temperature of the 

 rods was allowed to reach a steady state before the electric current was 



* Annales de Chiniie et de Physique. 1867. 



t Britisli Association Report. 1897. 



t Atti della Reale Accademia delle Scienze di Torino. 1886. 



