The place at which air gains access to these passages need not be a single opening, but consists, in all probability, of numerous small apertures, covered possibly by a thin coating of moss, loose shale or other porous substance.
In the summer time the warm outside air entering these apertures comes in contact with the rocks which have been chilled by the reverse currents of the preceding winter and in doing so gives up its heat to them, becoming specifically heavier. It then forces its way on down, displacing the warmer and lighter column of air above the pit.
It is evident that the rapidity with which this circulation takes place depends upon the difference in temperature of the two air columns. That is, the cold outward current is much more noticeable on hot days than on cool days in summer, and in winter the strongest inward current is noticed on the coldest days.
This fact accounts for the common belief that the freezing takes place more rapidly and that the mine is colder on hot than on cool days.
The temperature of the mine, or, in other words, of the air as it issues from the crevices, remains practically constant throughout the summer, which is proved by thermometer readings. However, the difference between this constant temperature and the temperature prevailing outside the mine is obviously greatest on the hottest days and therefore, as one enters the mine, the contrast is more noticeable. This causes one to believe that the mine is colder when it really is not. It is true, however, that the ice is formed most rapidly during the hottest weather. This is not because the temperature of the mine is lower, as is generally supposed, but is due to the fact that the circulation of air is more rapid; that is, a greater quantity of cold air issues from the numerous apertures and consequently a greater amount of "cold" is available for the formation of ice.
As soon as the supply of "cold" in the rocks is exhausted the internal and external air columns become gradually equal in temperature and weight, the circulation ceases and the ice begins to melt. This generally occurs about September of each year.
If this is the true explanation of this phenomenon, we may say, with truth, that in this particular instance it is the heat of summer which causes the ice to form, but, at the same time, we can not disregard the fact that it is the severity of the preceding winter and the natural arrangement of the rock strata which make it possible for the heat of summer to produce this peculiar phenomenon.