Evolution of Heat by Pitchblende. f>l 



larger as time goe^ on, finally becoming so large that the 

 experiment was discontinued. The moan for the last 20 days, 

 which, however, is not very reliable, is 6*25 scale-divisions, 

 corresponding to a heat evolution of 6*5 x 10~ 5 calorie per hour 

 per gram. 



It is evident that in all previous experiments the first 

 fortnight during which the ice temperature is fairly steady 

 is expended in the cooling of the pitchblende. So after the 

 conclusion of the last experiment, before the pitchblende had 

 warmed up more than o, 2 C, the ice vessel was repacked. 

 As will be seen (B, fig. 2) the temperature became steady 

 four days after packing, and remained so for seven days, 

 after which irregularities again made their appearance. The 

 mean 5*9 scale-divisions corresponds to a heat evolution of 

 6'2 x 10~ 5 calorie per hour. 



This figure agrees very well with the mean of the previous 

 results, 6"1 x 10" 5 . In this experiment the pitchblende had 

 been in the wax nearly eleven months and had been approxi- 

 mately at 0° C. for about two months. The steadiness of 

 the final temperature attained also lends weight to the 

 result. 



The variations in the temperature of the ice may be due to 

 traces of zinc oxide which, however, is almost insoluble in 

 water, but, as coating the zinc with vaseline, though it 

 stopped all apparent action, did not effect any improvement, 

 it seems more probable that the variations are in some way 

 caused by regelation. The ice when put in is in very small 

 pieces and almost resembles snow in appearance; but when 

 the vessel is opened the ice is always found to be frozen into 

 a glassy cake evidently pierced by numerous small air spaces. 

 It is hard to say whether these are isolated cavities or whether 

 they may form connected channels through which air might 

 circulate. No connexion could be traced between the tempe- 

 rature variations and the occasions on which the outer ice 

 vessel was repacked. 

 The three experiments recorded give for the heat evolution — • 



(1) 8' 15 X 10~ 5 calorie per hour per gram. 



(2) 6 5 xlO-5 „ 



(3) 6-2 x 10-^ „ 



Mean 6*95 x 10~ 5 ,, 



The mean obtained in the previous paper was G'l xlO - *; 

 so if we take the mean of all the results we obtain b'*5 x 10"" 6 . 

 The discrepancies between the various results are so large 



