234 



56 



in these measurements, for the results of 3 out of the 7 measurements are given 

 with a note of interrogation. But the 4 remaining measurements may safely he 

 considered as reliable as those of the other 3 series. 



I have made a graphic diagram of all the measurements, (Fig. 13), and con- 



Rl 

 Fig. 13. 



20 m 



nected by means of straight lines the temperatures belonging to each series of 

 observations. The roman figures denote the order in which the experiments suc- 

 ceeded each other. "I" shows the temperature immediately after a large eruption. 

 IV shows the temperature a little before the subsequent eruption, and V the 

 temperature after it. 



