1908] on Ice and its Natural History. 267 



outflow the crowding is relieved. The disarticulated groups become 

 disengaged, the smaller groups and individual prisms are able to 

 assume their attitude of stability and to float on their sides. All then 

 drift towards the outlet. The ice breaks up, and the lake is cleared 

 in an astonishingly short time. 



If it were not for the law that even impure water in freezing 

 always forms pure ice, the impurity remaining in the liquid and 

 generally entangled in the interstices of the grains, and that the pure 

 ice which is in contact with this impure liquid, melts at a lower tem- 

 perature than that which is in contact with nothing but the water 

 formed by its own melting, the ice covering a lake would be a 

 continuous sheet offering no points of weakness, and it would have to 

 melt as a whole. It is doubtful if lakes such as those met with 

 in the upper Engadine would get rid of their ice covering at all. On 

 the Silser See the ice is usually over 60 centimetres thick when the 

 thaw sets in, but when once the ice begins to break up, the lake is 

 cleared in a day. Sixty centimetres of ice would take a long time to 

 disappear on the basis of surface melting alone. 



While the winter lasts the ice on the lake shows no crystalline 

 structure. This develops only after removal from the water and 

 exposure to the sun. The ice then splits up into prisms in a vertical 

 plane. These are at first of irregular section, and as sun-weathering 

 proceeds the thicker prisms split up into thinner. When a block 

 has lain exposed to the February sun and cold, it may fall to pieces, 

 each piece being a long, thin triangular prism, with some resemblance 

 to a razor blade. When the ice is cold and dry, the outlines of the 

 grains are lines ; when the ice has a temperature of 0° C, it melts 

 perfectly round the grain, forming troughs in which the water 

 collects, and the aspect is that of a dark polygon, surrounded by 

 light-coloured canals. In one piece, which was much weathered, 

 I counted 24 such grains in an area of 9 square centimetres. In 

 a slab which had not been lying long, I counted 23 grains in an area 

 of 150 square centimetres, giving an average area of 6*5 square 

 centimetres per grain ; the largest had an area of 12 square 

 centimetres. In another slab there was a very large grain which 

 measured 7 centimetres in one direction and 4 centimetres at right 

 angles to it. In a slab in which the sun-weathering had proceeded 

 very far, I counted 113 grains in a disc of 5 centimetres radius, 

 which gives 0*69 square centimetre as the average area per grain. 



In the absence of actual experience, one is apt to expect a slab 

 of lake ice, when subjected to sun-weathering, to be disarticulated 

 into hexagonal columns ; but this expectation is quite gratuitous. 

 Ice may crystallise in a form bounded by plane faces, according to 

 the laws of its crystallographic system if it has the freedom which 

 it possesses when crystallising out of an independent medium such 

 as a saline solution or air. But the foreign matter dissolved in 

 fresh water is present in so small quantity that what we have before 



