Another cause of the greater coldness of the higher atmosphere is the expansion of the denser air of the lower strata when it ascends. The dense air makes room for itself by pushing back the lighter and less elastic air which surrounds it: _it does work_, and, to perform this work, a certain amount of heat must be consumed. It is the consumption of this heat--its absolute annihilation as heat--that chills the expanded air, and to this action a share of the coldness of the higher atmosphere must undoubtedly be ascribed. A third cause of the difference of temperature is the large amount of heat communicated, _by way of contact_, to the air of the earth's surface; and a fourth and final cause is the loss endured by the highest strata through radiation into space. FOOTNOTES: [A] The opposite faces of a thermo-electric pile. [B] See a most interesting paper on this subject by Mr. Hopkins in the Cambridge 'Transactions,' May, 1856. [C] See M. Pouillet's important Memoir on Solar Radiation. Taylor's Scientific Memoirs, vol. iv. p. 44. ORIGIN OF GLACIERS. (4.) [Sidenote: THE SNOW-LINE.] Having thus accounted for the greater cold of the higher atmospheric regions, its consequences are next to be considered. One of these is, that clouds formed in the lower portions of the atmosphere, in warm and temperate latitudes, usually discharge themselves upon the earth as rain; while those formed in the higher regions discharge themselves upon the mountains as snow. The snow of the higher atmosphere is often melted to rain in passing through the warmer lower strata: nothing indeed is more common than to pass, in descending a mountain, from snow to rain; and I have already referred to a case of this kind. The appearance of the grassy and pine-clad alps, as seen from the valleys after a wet night, is often strikingly beautiful; the level at which the snow turned to rain being distinctly marked upon the slopes. Above this level the mountains are white, while below it they are green. The eye follows this _snow-line_ with ease along the mountains, and when a sufficient extent of country is commanded its regularity is surprising. The term "snow-line," however, which has been here applied to a local and temporary phenomenon, is commonly understood to mean something else. In the case just referred to it marked the place where the supply of solid matter from the upper atmospheric regions, during a single fall, was exactly equal to its consumption; but the term is usually understood to mean the line along which the quantity of snow which falls _annually_ is melted, and no more. Below this line each year's snow is completely cleared away by the summer heat; above it a residual layer abides, which gradually augments in thickness from the snow-line upwards. [Sidenote: MOUNTAINS UNLOADED BY GLACIERS.]