644 PROCEEDINGS OF THE AMERICAN ACADEMY. 



signs at least of the Moreau formula are correct in this case, ^Te and 

 p being both + and j^Te and s both — . 



dP 



The quantitative argument is very simple. "Writing -jj- for the 



potential-gradient which maintains the current C through a plate of 

 width IV, thickness t, and specific resistance p, we have, using absolute 



units, C = -TT X~, and accordingly 

 ai p 



w wt \ w dl ja ^ ' 



Taking, as before (equation (3)), 



and observing that the static Thomson-effect potential-gradient is 



de dP 



we have ^^^= \^ -^ ll ) W ^^> 



If, now, we can assume that the equipotential lines of the electric 

 current in the Hall effect are by magnetic action rotated just as far as 

 the static thermo-electric equipotential lines of the heat current in the 

 Nernst effect, we shall have the parenthesis in equation (6) equal to 

 the parenthesis in (7), and then, dividing (6) by (7), we shall get 



eTe ^ uTe — p-^ S, OV e^e -^ p = h^e ^ S, (8) 



which is the ec^uation of Moreau. 



Before proceeding further it will be well to assemble in one taole the 

 various coefficients under discussion, each evaluated, as accurately as 

 our knowledge of the rates of change with temperature will enable us 

 to evaluate it, for various temperatures. We have, the last line re- 

 lating to Plate 1 and the other lines to Plate 2, 



Temp. eTo eTn i.Th hT* p s S 



20° 86lXlO-'i -505X10-10 545X10-9 -815X10-6 12550 -S51 1429 



40° 1048 " 538 " 588 " 897 " 13730 OGl 1370 



60° 1235 " 572 " 630 " 980 " 14910 1079 1.307 



80° 1422 " 604 " 672 " 1062 " 16090 1203 1240 



100° 1609 " 6.38 " 714 " 1144 " 17270 1333 1171 



45° 963 " -971(?)" 430(?)" -935 " 



(I.) 



