552 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



These data yielded the following values : 

 Mean total e. m. f. = 2807 X 10" 6 volt. 



" e. m. f. per degree, between 99°.9 and 217°.9 = 2379 X 10- 8 volt. 



Examining this summary as we have examined that of July 21, we 

 find: 



First, that, as before, practical equilibrium of temperature was main- 

 tained throughout the course of observations. 



Second, that in each of the two groups, " parallel " and " crossed, " 

 the differences between the deflections yielded by different couples are 

 considerably less than were the corresponding differences on July 21, 

 the greatest difference now found being rather less than 1 per cent 

 of the mean deflection. This greater uniformity is probably due to 

 the greater interval of temperature used on July 25, 118° instead 

 of the 40.3° of July 21. Taking a summary of the deflections yielded 

 by the four couples on the four days when they were tested with tem- 

 peratures near 100° and 218° we get: 



A 2 B a C 2 D 2 Mean. 



"Parallel" 19.60 19.56 19.63 19.62 19.60 



>0.02 >0.13 >0.13 >0.03 >0.02 



" Crossed" 19.62 19.69 19.50 19.65 19.62 



19.61 19.63 19.57 19.64 19.61 



