!M. Arago on Graham or Julia Island. 207 



even which, falling down vertically, unceasingly augmented the 

 apparent size of the islet, were cold. And besides, it is well 

 known that for two whole months it was scarcely possible to 

 walk upon the island, on account of the heat of the scorias 

 and sand of which it was composed. 



If the submersed portion of the new isle had been formed by 

 the superposition of incandescent materials, or at least of mat- 

 ters which were very hot, as was the portion which protruded 

 from the wave, it could not have failed to have heated the sea, 

 at all events to a certain distance ; thus in approaching the 

 island, a thermometer introduced into the sea should have gra- 

 dually risen. But this is precisely the reverse of what took 

 place ; for the dhninution of temperature observed by Dr Davy 

 on the 5th of August in approaching the isle amounted to 5°fi 

 Centigrade. 



Dr Davy, struck with this great diminution, conceived it 

 must be attributed to the floating powder with which the sea 

 was covered on the occasion. According to his view the dust 

 projected in the vertical column by the crater, must have ac- 

 quired, when it fell upon the water, the low temperature which 

 it had acquired in the more elevated atmospheric strata. This 

 explanation, however, seems liable to two serious objections : 

 Jirst, it is not very evident why each particle of dust in again 

 traversing the atmospheric strata from above downwards, should 

 not have re-acquired all the heat which it lost in ascending ; 

 and, secondly, we must remark that the total height of the co- 

 lumn was not above 400 feet English, an altitude which, accord- 

 ing to the known law of the decrease of atmospheric tempera- 

 ture, would only produce a difference of two-thirds of a de- 

 gree of the Centigrade scale. 



The 5°,6 of refrigeration observed by Dr Davy surpass by a 

 great deal every thing which hitherto has been found on approach- 

 ing the islands or shallows of the Mediterranean, or even the 

 islands and shallows of the ocean. It is not sufficient then to 

 reject the hypothesis which would have implied an augmenta- 

 tion of temperature ; it moreover remains that we should ex- 

 plain how the refrigerating influence of the islet should have 

 been so great. And nothing can be more simple, for we have 

 only to suppose that the ii-land was at first formed in the M'ay 



