Hitherto, as I have said, these disks have been mistaken for bubbles containing air, and their flattening has been ascribed to the pressure to which they have been subjected. M. Agassiz thus refers to them:--"The air-bubbles undergo no less curious modifications. In the neighbourhood of the _neve_, where they are most numerous, those which one sees on the surface are all spherical or ovoid, but by degrees they begin to be flattened, and near the end of the glacier there are some that are so flat _that they might be taken for fissures when seen in profile_. The drawing represents a piece of ice detached from the gallery of infiltration. All the bubbles are greatly flattened. But what is most extraordinary is, that, far from being uniform, _the flattening is different in each fragment_; so that the bubbles, according to the face which they offer, appear either very broad or very thin." This description of glacier-ice is correct: it agrees with the statements of all other observers. But there are two assumptions in the description which must henceforth be given up; first, the bubbles seen like fissures in profile are not air-bubbles at all, but vacuous spots, which the very constitution of ice renders a necessary concomitant of its inward melting; secondly, the assumption that the bubbles have been _flattened_ by pressure must be abandoned; for they are found, and may be developed at will, in lake-ice on which no pressure has been exerted. [Sidenote: CELLS OF AIR AND WATER.] But these remarks dispose only of a certain class of cells contained in glacier-ice. Besides the liquid disks and vacuous spots, there are innumerable true bubbles entangled in the mass. These have also been observed and described by M. Agassiz; and Mr. Huxley has also given us an accurate account of them. M. Agassiz frequently found air and water associated in the same cell. Mr. Huxley found no exception to the rule: in each case the bubble of air was enclosed in a cell which was also partially filled with water. He supposes that the water may be that of the originally-melted snow which has been carried down from the _neve_ unfrozen. This hypothesis is worthy of a great deal more consideration than I have had time to give to it, and I state it here in the hope that it will be duly examined.