70 CRYSTALLISATION OF WATEE. 



removing tlie slab to a colder place and examining it tlirougli a lens, we 

 shall see niimbers of these ice-flowers within the slab, each with a 

 "bubble" in the centre. Their diameter is usually about 1-lOth to 

 l-20th of an inch, and they are only dimly visible, because the 

 refractive indices of ice and water, at the freezing temperature, are 

 nearly the same. 



But what are these "bubbles?" Let us appeal to experiment for 

 the answer. If we place a piece of ice containmg them in warm water, 

 we shall find that when tlie ice surrounding them is melted they 

 suddenly collapse and disappear utterly. If we treat a bubble of air 

 contained in ice in the same way, we shall see the bubble rising tlirough 

 tlie water as soon as it is liberated. Hence we infer that this apparent 

 bubble, which occupies the centre of the ice-flower, is really a vacuum. 

 Now observe how unexpectedly a law of nature steps in. Ice is less 

 dense than water, as is manifested by its floating, and, when a certain 

 quantity of ice is melted, the resultant water occupies a smaller volume : 

 so the formation of each flower is attended with the formation of a 

 vacuum. 



This contraction of the volume of water, however, takes place in a 

 peculiar way. When the cavity is small, and its sides close together, 

 the molecules of water are able to bear the strain put upon them, and 

 separate by a minute interval, so as still to till the space. Thus the 

 discs are at first bubbleless, but as the quantity of ice melted increases, 

 the strain becomes greater, until at last the molecules yield to the 

 influence of their mutual attractions, and rush together suddenly into 

 a smaller space. Thus the formation of each vacuum is attended with 

 a "clink" which is clearly audible, and resembles, on a smaller scale, 

 the metallic clink heard when water deprived (as this ice-water is) 

 of its dissolved air is shaken in a tube, as in Denny's well-known 

 experiment. 



The planes of these " negative " crystals, as they have been termed, 

 lie in the planes of freezing, that is, in ordinary cases, parallel to the 

 surface of the water. We may detect this direction in any block of 

 lake-ice, taken at random, either by developing the flowers or by 

 observing the bubbles which are almost invariably entangled in it. We 

 shall see layers cloudy -with bubbles, separated by layers of clear ice, 

 the plane separating the two being, often, perfectly distinct on the 

 lower side of the clear layer, but less so above. The explanation of 

 this appearance is that the clear ice is that produced by slow freezing, 

 e.fj., during the daytime, when the molecules of water, as they fall 

 into their places in the crystal, have time to push out of the way the 

 particles of air entangled among them. These collect into bubbles on 

 the lower surface, and if, e.g. after sunset, a sudden fall of temperature 

 takes place and the freezing becomes more rapid, they are unable to 

 extricate themselves and become locked in their icy prison. It will be 

 noticed that the bubbles, which would be normally nearly spherical, 

 are usually pear-shaped or conical, with their longer axis vertical and the 

 nari-ow end pointing downwards as if they had been gradually 



