AND THE RELATION OF SPECIFIC HEAT TO ATOMIC WEIGHT. 



255 



In all probability these formulae give fairly correct values of K from 180 C. to 

 100 C C. In spite of the danger of extrapolation, we give in the following table 

 some values beyond this range, to show that the formula expresses the sort of change 

 known to occur in metals generally at high temperatures. Close to the zero there 

 must be greater doubt as to its suitability : 



[April 27, 1900. The following formula reproduces Dr. TILDEN'S results with 

 greater accuracy. It is probably of a nature to suit all metals which exhibit no 

 recalescence effects. 



when the values of the constants for nickel and cobalt are 



