AND THi; ILLATION OF SPECIFIC HEAT TO ATOMIC WKIGIIT. 



249 



The mean specific heats of the two metals within three successive ranges of 

 temperature now stand as follows : 



Temperature. 



From 100 to 15 

 15' -78-4* 

 15 , -182-5 



Cobalt. 



10303 



0939 



0822 



Nickel. 



10842 

 0975* 

 0838 



and by calculation from the last two results 

 From -78-4 to - 182'5 -0712 -0719 



Mere inspection shows that the change in value proceeds regularly in both cases, 

 but that the value for nickel declines more rapidly than for cobalt, and the conse- 

 quence is that the specific heats of the two metals steadily approach each other. If 

 the numbers given above for the specific heats are multiplied by the atomic 

 weights, the products are very nearly identical. 



Thus -0822 X 58-55 = 4'81 atomic heat of cobalt, and '0838 X 58'24 = 4'88 the 

 atomic heat of nickel. 



Or -0712 X 58-55 = 4-169 and '0719 X 58'24 = 4'187. 



* Other recorded attempts to determine specific heats at low temperatures are as follows : 



The falling off is in each case less than in my experiments, but the probability is that in these old 



experiments the metals were not really at the low temperature assumed when introduced into the 

 calorimeter. (See note, p. 250.) 



VOL. CXCIV. A. 2 K 



