No matter in what direction a solar beam is sent through lake-ice, the liquid flowers are all formed parallel to the surface of freezing. The beam may be sent parallel, perpendicular, or oblique to this surface; the flowers are always formed in the same planes. Every line perpendicular to the surface of a frozen lake is in fact an axis of symmetry, round which the molecules so arrange themselves, that, when taken down by the delicate fingers of the sunbeam, the six-leaved liquid flowers are the result. In the ice of glaciers we have no definite planes of freezing. It is first snow, which has been disturbed by winds while falling, and whirled and tossed about by the same agency after it has fallen, being often melted, saturated with its own water, and refrozen: it is cast in shattered fragments down cascades, and reconsolidated by pressure at the bottom. In ice so formed and subjected to such mutations, definite planes of freezing are, of course, out of the question. [Sidenote: CONSTITUTION OF GLACIER-ICE.] The flat round disks and vacuous spots to which I have referred come here to our aid, and furnish us with an entirely new means of analysing the internal constitution of a glacier. When we examine a mass of glacier-ice which contains these disks, we find them lying in all imaginable planes; not confusedly, however--closer examination shows us that the disks are arranged in groups, the members of each group being parallel to a common plane, but the parallelism ceases when different groups are compared. The effect is exactly what would be observed, supposing ordinary lake-ice to be broken up, shaken together, and the confused fragments regelated to a compact continuous mass. In such a jumble the original planes of freezing would lie in various directions; but no matter how compact or how transparent ice thus constituted might appear, a solar beam would at once reveal its internal constitution by developing the flowers parallel to the planes of freezing of the respective fragments. A sunbeam sent through glacier-ice always reveals the flowers in the planes of the disks, so that the latter alone at once informs us of its crystalline constitution. [Sidenote: VACUOUS CELLS MISTAKEN FOR AIR-CELLS.]