292 PROCEEDINGS OF THE AMERICAN ACADEMY 



slab, that tlie motions of the external atmosphere can do nothing 

 towards i)roducing the current. It shows, too, that at the only place 

 "wherc'an upward flow could afloct a candle, there the deflected flame 

 renders the rush of air ai)parent. We dwell more })articularly on 

 this point, because Messrs. Jackson and Blake, in their communication 

 to the Academy,* have denied the existence of any current in the 

 well, since the flame of their candle was not observed to deviate from 

 its natural course ; and because, from the confusion of ideas exhibited 

 by those residing at the locality, it would seem as though there had 

 been entertained a vague notion that a draught or current ought 

 necessarily to be horizontal. In reality, the direct upward course is 

 what would naturally result from the combination of numerous oj)pos- 

 iug rills of air, coming in from all sides through the loose gravel and 

 cracks in the frozen deposit, and, while seeking the point of least 

 resistance, gradually curving towards the perpendicular direction, even 

 before reaching the stones of the lining wall. If there are, indeed, 

 any single streams of more than the average force, they might be 

 found by running an impervious vertical partition down the middle 

 of the shaft, and then making a smoke close to the joints of the stones 

 in the wall. But the fixct that the owner of the ground has only 

 this well to supply water for the daily use of his family, precludes 

 the carrying out of such experiments. 



We had hardly begun to make close observations, before it occurred 

 to us that we were dealing with a case of compressed air, which 

 might be accumulated by some natural subterranean water-trunipet 

 ( Wasscrtrommcl), or " Catalan blower," and which, expanding as it 

 approached the surface of the earth or escaped into this artificial 

 outlet, would absorb and render latent a large amount of heat, and 

 could thus etfect the gradual refrigeration and actual freezing of a 

 considerable body of wet gravel.f Tiie flow of air — wliicli would 

 otherwise be so diffused as to have no appreciable force at any part 

 of the surface of the drift bed — should become especially apparent 

 in this vent bored down througli the overlying clay and fine soil into 

 the very coarse and uncommonly pervious gravel. That the gas ex- 



* ri'occedings, IV. 270. 



I For a valuivl)le series of experiments mnde by Dr. John Gorrie, on tlie absorp- 

 tion of heat l)y comlcnsed air in the act of expanding:, — with partiruiar reference 

 to the economical production of ice, — see Am. Jouni. Sci. fou" 1850, [2.] X. 39, 

 214 ; Ann. Sci. Disc, 1S51, p. 57. 



