638 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



drift of such a character as to indicate that we should have got 

 larger Ettingshausen effects by waiting longer for these effects to be 

 established. 



K 



Figure 9. 



Figure 10. 



The Nernst Effect. — With Plate 1 we at first used the iron trough 

 already described, during the thermomagnetic tests. We now reject 

 all the Nernst-efifect observations made with this trough between the 

 poles. The remaining results are 



1910. 



June 4 

 " 10 



Means 



— 944 X 10-« 

 928 



—936 



Arm 0. 



45.7° 

 44.8° 



45.3° 



de 



dl' 



11.3' 

 11.3' 



11^ 



H. 



8400 

 8700 



8550 



