A bed of snow covering the ground acts as a bad conductor, and renders the exchange of temperature between the surface of the ground and the lower stratum of the atmosphere slower than it would be if the snow were absent. This is a result of the porous structure of snow, the interspaces of which are filled with air. In this matter the condition of the snow is of considerable importance; the minute crystals formed by cold are poorer conductors than the larger flakes; but if the snow by alternately thawing and freezing assumes the form called névé, it becomes a much better conductor of heat. Farmers in countries that enjoy cold winters are well aware of the protective nature of the snow-covering, and do not fear for their grain when it is thick.
The presence of snow thus assures a higher temperature to the upper layer of the soil than it would otherwise have, and its thickness is an equally important factor with its structure. But its effect on the air is different, for it separates the air from the warmth which without its presence would escape from the ground. Snow also exercises an important influence through its power of radiation, which is dependent on its whiteness and the extent of its surface, but in which thickness is not a factor.
When air is rarefied, it contains, besides the vapor of water, only a few scattering particles of dust or smoke. Even in the tropics, snow on high mountains does not melt. Although a considerable amount of solar heat is received by the upper surface of the snow, it is all sent back into space, and the air, being very diathermanous, only retains a very small part of it. In the plain countries of high latitudes the air is not rarefied, but, when there is a large extent of snow, the other conditions are similar to those which prevail at great elevations under the tropics, particularly the absence of dust and the small quantity of aqueous vapor. The latter condition is the result of cold, while the former may arise from the fact that there is no dry bare ground near from which dust can be swept up, or from the more or less complete absence of organic life. In this case as in that of high mountains, radiation into space goes on freely, and the solar rays are without power to melt the snow so long as it preserves its light structure, and its surface does not present any dark object to absorb the rays of the sun and cause the snow around it to melt. Thus, we observe that it melts on roofs and immediately around trees, while the clear snow in the vicinity shows no signs of thawing. Snow will melt rapidly under a piece of brown paper lying in a garden; and a thin sprinkling of dust on top of snow will produce a similar effect.
We conclude, from these observations, that if a large extent of continent is covered with snow, that snow will not melt under the rays of the sun. Yet we know that the winter snows in Northern Europe, Asia, and America melt every spring and summer. How does this come about? Observations made in high latitudes show that the
