PEOFESSOE TTjSTDALL ON THE VEINED STEUCTEEE OE OLACIEES. 299 
together by the crystallizing force. By the slow abstraction of heat from water its par- 
ticles build themselves into these little stars, and by the introduction of heat into a 
mass so built the architecture is taken down in a reverse order. In watching the 
formation of artificial ice, by the machine of Mr. Hareison referred to in my paper, I 
have seen little solid stars formed, by freezing, which were the exact counterparts of the 
little hquid stars formed by melting. So far as I can see, the complementary character 
of the phenomena is perfectly natural, and presents no difiiculty to the mind in con- 
ceiving of it. 
MTien the beam is intense, and its action continued for some time, the flowers expand, 
so as to form liquid plates within the mass. Looked at edgeways, these liquid spaces 
appear hke fine lines ; which proves that the melting is not symmetrical laterally and 
vertically, but that the ice melts in the planes of freezing much more readily than at 
right angles to these planes. 
If an air-bubble exists within ice, and if the ice melts at the concave surface of this 
bubble, as might be expected from the foregoing facts, the ice will so yield that the 
composite cell of air and water will not be spherical, even though the bubble of air may 
originally have been so. In the planes of freezing the mass yields most readily, and the 
cavity containing the air and water will appear as if flattened hy a force acting perpen- 
dicular to these planes. This is not a deduction merely, but an observation which I 
have made in a hundred difierent cases. 
What I have here said applies to ordinary lake ice ; but glacier ice has no definite 
“ planes of freezing.” The substance is first snow, which sometimes, it is true, falls 
regularly in six-rayed crystals, as observed by myself on the summit of Monte Eosa ; but 
it is usually disturbed by winds, while falling, and whirled and tossed by the same 
agency after it has fallen ; the mountain snow is often melted, mixed with water and 
refrozen. Even after it has become consolidated it is often shattered in descending pre- 
cipitous slopes. In such ice definite planes of crystallization are, of course, not to be 
expected. 
If tve supjpose a mass of lake ice to he broken up into fragments^ and these fragments 
thrown together confusedly and regelated in their new positions to a continuous mass, 
we have an exact image of the character of the glacier ice in which this flattening of the 
hubbies in different directions has been observed. 
In the paper already referred to, I have given a sketch of a piece of ice composed of 
such segments, and have described the effects obtained with it. That ice was sold to 
me as Norway lake ice. I am not aware whether glacier ice is ever imported into this 
country from Norway; but if it be, the piece in question must, I think, have belonged 
to it. It is so like all the glacier ice that I have examined since that time, and so 
unhke all the lake ice, that I feel little hesitation in saying that it belonged to the 
former*. No matter how coherent and optically continuous a mass of ice may be, a con- 
densed sunbeam would at once tell us whether it belonged to a lake or to a glacier. 
* Perhaps formed from the counecting together of confused fragments. 
