Not only however in air, but in water also, figures of great beauty are sometimes formed. Harrison's excellent machine for the production of artificial ice is, I suppose, now well known; the freezing being effected by carrying brine, which had been cooled by the evaporation of ether, round a series of flat tin vessels containing water. The latter gradually freezes, and, on watching those vessels while the action was proceeding very slowly, I have seen little six-rayed stars of thin ice forming, and rising to the surface of the liquid. I believe the fact was never before observed, but it would be interesting to follow it up, and to develop experimentally this most interesting case of crystallization. [Sidenote: DISSECTION OF ICE BY SUNBEAM.] The surface of a freezing lake presents to the eye of the observer nothing which could lead him to suppose that a similar molecular architecture is going on there. Still the particles are undoubtedly related to each other in this way; they are arranged together on this starry type. And not only is this the case at the surface, but the largest blocks of ice which reach us from Norway and the Wenham Lake are wholly built up in this way. We can reveal the internal constitution of these masses by a reverse process to that which formed them; we can send an agent into the interior of a mass of ice which shall take down the atoms which the crystallizing forces had set up. This agent is a solar beam; with which it first occurred to me to make this simple experiment in the autumn of 1857. I placed a large converging lens in the sunbeams passing through a room, and observed the place where the rays were brought to a focus behind the lens; then shading the lens, I placed a clear cube of ice so that the point of convergence of the rays might fall within it. On removing the screen from the lens, a cone of sunlight went through the cube, and along the course of the cone the ice became studded with lustrous spots, evidently formed by the beam, as if minute reflectors had been suddenly established within the mass, from which the light flashed when it met them. On examining the cube afterwards I found that each of these spots was surrounded by a liquid flower of six petals; such flowers were distributed in hundreds through the ice, being usually clear and detached from each other, but sometimes crowded together into liquid bouquets, through which, however, the six-starred element could be plainly traced. At first the edges of the leaves were unbroken curves, but when the flowers expanded under a long-continued action, the edges became serrated. When the ice was held at a suitable angle to the solar beams, these liquid blossoms, with their central spots shining more intensely than burnished silver, presented an exhibition of beauty not easily described. I have given a sketch of their appearance in Fig. 34. [Sidenote: LIQUID FLOWERS IN ICE.] [Illustration: Fig. 34. Liquid Flowers in lake ice.]