It is also known that water freezing very slowly is transparent.
Whenever, by the melting of the lower portion of any part of a glacier, a piece of it cracks and falls to a lower level, the friction of the broken sides will produce heat, and melt a small portion of water. This water, trickling down very slowly, will form a thin layer on the broken surface, and a portion will be retained in the narrowest part of the crack. But, since the temperature of a glacier is very near the freezing point, that water will freeze very slowly. It will, therefore, become transparent ice, and will, as it were, solder together the two adjacent surfaces by a thin layer of transparent ice.
But the transparent ice is much stronger and more difficult to break than opaque ice; consequently, the next time the soldered fragments are again broken, they will not break in the strongest part, which is the transparent ice: but the next fracture will occur in the opaque ice, as it was at first.
Thus, by the continued breaking and falling downward of the fragments of the glacier, as it proceeds down the valley, a series of vertical, rudely-parallel veins of transparent ice will be formed. As these masses descend the valley, fresh vertical layers of transparent ice will be interposed between those already existing until the whole takes that beautiful transparent cerulean tint which we so frequently see at the lower termination of a glacier. Another effect of this vertical fracture at the surfaces of least resistance will be alternate vertical layers of opaque and transparent ice shading into each other. This would, in some of its stages, give a kind of ribboned appearance to the ice. Probably traces of it would still be exhibited even in the most transparent ice. Speaking roughly, this ribboned structure ought to be closer together the nearer the piece examined is to the end of the glacier. It
