Beside the snow-layers and the sheets of dust alternating with them, there is still another feature of the horizontal and parallel structure of the mass in immediate connection with those above considered. I allude to the layers of pure compact ice occurring at different intervals between the snow-layers. In July, when the snow of the preceding winter melts up to the line of perpetual snow, the masses above, which are to withstand the summer heat and become part of the glacier forever, or at least until they melt away at the lower end, begin to undergo the changes through which all snow passes before it acquires the character of glacial ice. It thaws at the surface, is rained upon, or condenses moisture, thus becoming gradually soaked, and after assuming the granular character of _neve_-ice, it ends in being transformed into pure compact ice. Toward the end of August, or early in September, when the nights are already very cold in the Alps, but prior to the first permanent autumnal snow-falls, the surface of these masses becomes frozen to a greater or less depth, varying, of course, according to temperature. These layers of ice become numerous and are parallel to each other, like the layers of ice formed from slosh. Such crusts of ice I have myself observed again and again upon the glacier. This stratified snowy ice is now the bottom on which the first autumnal snow-falls accumulate. These sheets of ice may be formed not only annually before the winter snows set in, but may recur at intervals whenever water accumulating upon an extensive snow-surface, either in consequence of melting or of rain, is frozen under a sharp frost before another deposit of snow takes place. Or suppose a fresh layer of light porous snow to have accumulated above one the surface of which has already been slightly glazed with frost; rain or dew, falling upon the upper one, will easily penetrate it; but when it reaches the lower one, it will be stopped by the film of ice already formed, and under a sufficiently low temperature, it will be frozen between the two. This result may be frequently noticed in winter, on the plains, where sudden changes of temperature take place. There is still a third cause, to which the same result may possibly be due, and to which I shall refer at greater length hereafter; but as it has not, like the preceding ones, been the subject of direct observation, it must be considered as hypothetical. The admirable experiments of Dr. Tyndall have shown that water may be generated in ice by pressure, and it is therefore possible that at a lower depth in the glacier, where the incumbent weight of the mass above is sufficient to produce water, the water thus accumulated may be frozen into ice-layers. But this depends so much upon the internal temperature of the glacier, about which we know little beyond a comparatively superficial depth, that it cannot at present afford a sound basis even for conjecture.