44 



PROCEEDINGS OF THE AMERICAN ACADEMY 



absolute temperature r be of the supposed form i] = ct', vvliere c is a 

 constant, we may take the value of log t] and log t fi-ora our experi- 

 ments, and expect upon plotting them to get a straight line making an 

 angle whose tangent is x. This method I have applied to the results 

 of Meyer, and the extreme points are shown at the points marked 

 D and £ in the accompanying cut. 



Fig. 1. 



3.^ 



4.40 



.00 



The single experiment at zero centigrade gives the point 0. All 

 the other experiments furnish points scattered between Z) and E. The 

 absolute values of the coefficient in these cases are : — 



For the line CD, x : 



