[Footnote 194: Daubuisson estimated the depth in question at from 46 to 61 feet, while Kupffer put it at 77 feet.] [Footnote 195: De Saussure found a variation of 2°·25 F. at a depth of 29·5 feet; but this was in a well, where the influence of the atmosphere was allowed to have effect. Naturally, the fissures which there may be in the rock surrounding a cave will increase the annual variation of temperature, by affording means of easier penetration to the heat and cold. Sir K. Murchison's cavern in Russia would seem to be entirely _sui generis_.] * * * * * CHAPTER XVIII. ON THE PRISMATIC STRUCTURE OF THE ICE IN GLACIÈRES. It was natural to suppose that the prismatic structure which I found so very general in the glacières was the result of some cause or causes coming into operation after the first formation of the ice. On this point M. Thury's visit to the Glacière of S. Georges in the spring of 1852 affords valuable information, for at that time the coating of ice on the wall, evidently newly formed, did not present the _structure aréolaire_ which he had observed in his summer visit to the cave. He suggests that, since ice is less coherent at a temperature of 32° F.--which is approximately the temperature of the ice-caves during several months of the year--than when exposed to a greater degree of cold, its molecules will then become free to assume a fresh system of arrangement.[196] On the other hand, Professor Faraday has found that ice formed under a temperature some degrees below the ordinary freezing point has a well-marked crystalline structure.[197] M. Thury suggests also, as a possibility, what I have found to be the case, by frequent observations, that the prismatic ice has greater power of resisting heat than ordinary ice; and on this supposition he accounts for the fact of hollow stalactites being found in the Cavern of S. Georges.[198] At the commencement of the hot season, the atmospheric temperature of the glacières rises gradually; and when it has almost reached 32° F., the prismatic change takes place in the ice, extending to a limited depth below the surface. The central parts of the stalactites retain their ordinary structure, and are after a time exposed to a general temperature rather above than below the freezing point; and thus they come to melt, the water escaping either by accidental fissures between some of the prisms, or by the extremity of the stalactite, or by some part of the surface which has chanced to escape the prismatic arrangement, and has itself melted under increased temperature.[199]