The subsoil ice of the tundras of Siberia and Alaska is almost identical with the ice of boulder formations, except that it extends under larger areas. It is the product of a climate where there is a long, rigorous winter and it is not surprising that the ice is found at greater depths than in more southerly latitudes.[59] The depth to which the ice extends is, of course, determined by the depth to which the winter's cold can penetrate the soil. There is no doubt that the causes of this ice are local, that is, that it is due to the long prevailing low temperatures. [59] See Part III.: Alaska, page 166; Klondike, page 167; Kowak River, page 166; Kotzebue Sound, page 166. The freezing wells of which the most conspicuous examples are at Brandon, Owego, Decorah, and in Montana, seem also due to local causes and the ice is never far from the surface, that is, not over twenty meters; and apparently also it forms above the water horizons which supply the wells. The ice sheet on Mount Etna[60] does not seem to be at any great depth. It apparently had a different origin from most subterranean ice masses, in that the snow probably fell first and was then covered by a flow of lava. It is, therefore, almost _sui generis_ in its mode of formation, unless there are similar sheets on other volcanoes, which is probable in a country like Iceland, and which is said to be the case in Tierra del Fuego,[61] But the original cause of the ice sheet on Etna was the same as all other subterranean ice masses, namely the cold of winter. [60] See Part III., page 210. [61] See Part III., page 190. _Evaporation and Movements of Air._--The formation of subterranean ice is sometimes assigned partly to evaporation or to expansion of the air. The theory is an old one, and both scientific and non-scientific men have advanced it.[62] [62] Among those who have written that evaporation is a factor in the formation of subterranean ice, may be mentioned: Pictet, 1822; Scrope, 1827; Reich, 1834; Pleischl, 1841; Murchison, 1845; Olmstead, 1856; Hitchcock, 1861; White, 1870; Kirchhoff, 1876; Krauss, 1895. I have already said that I believe it is the movements of the air which cause a cave to be a glacière cave or a normal cave. When they act in such a way as to permit the cold air of winter to permeate a cave, we find low temperatures and ice; when they do not, we find the temperature about the same as that of the ground and no ice.