HALL. — THERMO-ELECTRIC HETEROGENEITY IN ALLOYS. 553 



Here, too, the greatest difference observed in either of the two groups, 

 "parallel," and "crossed," is a very little less than 1 per cent of the 

 mean, 19 parts in nearly 2000. 



Third, that, as before, a couple which gives a deflection larger than 

 the mean with the " parallel " arrangement, is likely to give a deflection 

 smaller than the mean with the " crossed " arrangement, and vice versa, 

 so that differences between the individual means of " parallel " and 

 " crossed " taken together, yielded by the various couples, are small, the 

 greatest being 9 parts in nearly 2000 for July 25 and 7 parts in nearly 

 2000 for the four-day summary. 



Fourth, that the mean " parallel " deflection for July 25, 19.68 +, is a 

 little greater than the mean " crossed " deflection, 19.64, for the same 

 day. In the four-day summary, however, we find the mean " parallel " 

 deflection to be 19.60, while the mean " crossed " deflection is a very 

 little greater, 19.62. 



On one day, July 5, and on that day only, observations were made 

 with the interval of temperature from 100° to 136°, approximately. 

 The summary for the individual couples with this interval for this one 

 day is as follows : 



Aj B 2 C 2 D 2 Mean. 



"Parallel" 15.21 15.19 15.34 15.27 15.25 



>0.09 >0.16 >0.07 >0.02 >0.04 



"Crossed" 15.30 15.35 15.27 15.25 15.29 



15.26 15.27 15.31 15.26 15.27 



It seems hardly worth while to discuss this summary in detail at pres- 

 ent, though it will be referred to later. 



Taking the five-day summary for the interval 140° to 180° with the 

 four-day summary for the interval 100° to 218°, and still postponing dis- 

 cussion of the absolute values of electromotive force found with these 

 intervals, we reach the following conclusions from the data and the dis- 

 cussion already given : 



1. That with the four couples A 2 , B 2 , C 2 , D 2 , made in the same way 

 from the same materials, a maximum range of electromotive force amount- 

 ing to about £ per cent of the mean electromotive may be expected in a 

 thorough test with any interval of temperature between 100° and 218°. 

 The one day's work with the interval 100° to 136° is in complete accord 

 with this conclusion. 



