115 
On the Loxo Temperature of T articular Caverns. 
perature, it would imply no extravagance of supposition to take 
it for granted that these subterraneous recesses are connected by 
means of horizontal crevices, running inwards to a great extent, 
with vertical fissures ascending in some adjacent mountain, to a 
height above the line of perpetual congelation. The great dis- 
tance under ground to which several of the caverns in calcareous 
rocks have been traced, permits us to indulge the conjecture, 
that, in many instances, they may proceed far beyond the limits 
to which they have actually been explored ; and, as we are cer- 
tain that, in some cases, they penetrate several miles through the 
solid strata, it is easy to imagine that rents proceeding from them 
may branch off‘, in a vertical direction, to the very summit of the 
mountain in which they occur, or at least to a height beyond the 
regions of perpetual congelation. This state of things would be 
quite consistent with the fact, affirmed to be observed in all the 
icy caverns, that more ice is formed in them during summer 
than in winter ; as the current of air flowing downward ought to 
be most powerful, and consequently the cold induced greatest, 
when the difference between the temperature of the atmosphere 
at the mouth of the cave, and at the opening of the vertical cre- 
vices, where the air is supposed to enter, is greatest. Nor does 
this hypothesis exclude the influence of evaporation in contri- 
buting to the effect ; on the contrary, it admits its operation in 
the only circumstances under which it could possibly act, name- 
ly, when the humidity of the air, before it entered the crevices 
which conduct it to the cavern, is less than what it possesses at 
the freezing point. In no other case could evaporation reduce 
the temperature of a moistened surface to that point. 
