CRYSTALLISATION OF WATER. 71 



compressed laterally by the expansion of the surrounclinf^ water as 

 it passed into the solid state. Lar^^er bubbles are less frequently met 

 with in lake-ice ; these are probably filled with f^ases emitted from the 

 decaying matter at the bottom, which we know is continually rising to 

 the surface in bubbles when the water is unfrozen. 



The ice made in freezing machines, or on a small scale in a test 

 tube, is generally cloudy, being quickly formed. The reason why 

 slowlj'-formed ice is transparent is that the crystals are in perfect 

 optical contact, but it is said to be possible, by carefully exposing a 

 block to the action of heat, to cause it by a smart blow to fall into 

 pieces which are roughly hexagonal prisms. This, however, I have not 

 been able to do when I tried, but I once did something like it involun- 

 tarily. The winter before last, when I was skating at the Edgbaston 

 Reservoir, I lost my balance and fell. The ice cracked ominously, and 

 on rising I saw the spot where I had come in contact with it marked 

 by a large six-rayed star, the arms of which were from twelve to 

 eighteen inches long, and arranged almost exactly at equal angles of 

 sixty degi-ees, hke the main rays of the crystals (Figs. 1 and 2.) I also 

 produced the same effect last winter in a water-tub by striking 

 the ice vrith a hammer, the rays being longer or shorter according 

 to the strength of the blow'. It seems probable that the ice cracked 

 in these directions because these were the planes of least I'esistauce, 

 like the planes of cleavage in crystaUised minerals. 



It is interesting to see ice thus analysed, and its component parts 

 demonstrated, but it is still more interesting to observe the synthetic 

 process, by which the crystals are formed. This is in general so masked 

 by various circumstances that they rarely assume their proper shapes, 

 but under special conditions perfect crystals may be obtained. They 

 have been noted by several observers, but are undoubtedly of com- 

 paratively rare occurrence. I am pleased therefore to record the fact 

 that, one morning in December, 1879, I found floating freely on the 

 surface of a basin of water, in my house, three thin plates of ice, the 

 smallest but most regularly formed of which is represented in Fig. 1 of 

 its natural size, about two inches in diameter. It will be seen how 

 closely it imitates the form of the snow-crystal which is sketched 

 above (Fig. 2.) 



The main conditions necessary for the formation of such crystals 

 seem to be intense cold, combined with slow freezing. Professor 

 Tyndall, to whom I aiu indebted for many of these facts and explana- 

 tions, observed the formation of little six-rayed stars of thin ice in the 

 vessel of an artificial ice machine, in which the action was proceeding 

 very slowly. He believed the observation to be then new, and it gives 

 me, therefore, great pleasure to quote the following account of a similar 

 occurrence described in a letter to me by Mr. W. H. Wilkinson. He 

 says that it happened on a Christmas-da>-, some years ago, when the 

 thermometer in BiiTuingham fell below zero : — 



