i40 On the Movemmf of Rocks, 



as it expands from the middle of the pond towards the shore. 

 Carries the rock along with it. The fact, that the ice does 

 expand from the middle towards the borders, in all cases 

 where water is frozen, must be evident to all acquainted 

 with cold climates, and who have observed the circum- 

 stances in which ice is formed. When water is left to be 

 frozen in a vessel, the expansion from the middle to the 

 outside is so strong as to break the vessel. This is "some- 

 times the case even where the vessel is of iron. There is 

 often, also, a considerable elevation in the middle of the 

 ice resulting from the resistance of the sides of the vessel 

 to the outward expansion. But in ponds and lakes this 

 central elevation is never formed, on account of the im- 

 mense weight of the ice,* and the little or no resistance to 



* The expansion of ice, though so great a force, that no known resist- 

 ance can confine it, is always exerted in the direction where there is least 

 resistance. I several times repeated the following experiment before my 

 classes, while Professor of Mathematics at the University in Rhode-Island. 

 I procured a military shell weighing 70 or 80 pounds, and capable of con- 

 taining nearly two quarts of water. The orifice was about an inch in di- 

 ameter. At the approach of a very cold night, I filled it with water, and 

 placed it in a situation favourable for freezing la the morning, all the 

 water was frozen, and a column of ice was driven through the orifice, of 

 the diameter of the orifice, rough in appearance, and 5 or 6 inches long. 

 When water freezes in a vessel of some strength, at first the resistance of 

 the sides causes the elevation of the ice adverted to above. As, however, 

 the ice continues to thicken, and to oppose a resistance continually increas- 

 ing to the expansion upwards, a time arrives at length, when the sides of 

 the vessel present a less resistance to the expansion, than the super-impos- 

 ed ice, and at this point, the vessel is broken. But when ice forms upon 

 a lake, this elevation cannot take place on account of the very great weight 

 of the whole mass of ice, which weight in ordinary circumstances prevents 

 the expansion upwards. Its expansion below into the general mass of the 

 water is hindered by the water being confined on all sides, and thus oppos- 

 ing a resistance scarcely less than that of a solid body. The expansion, 

 therefore, will naturally be directed towards the shore, and a disruption 

 between the main body of the ice and the shore will be made, where the 

 shore is inclined at a moderate angle to the surface of the water, and a 

 projection of the ice will take place. This projection must have often 

 been seen, by every one accustomed to cold climates, when thick ice is melt- 

 ing ; as it frequently lies several feet beyond the edge of the water. And 

 if the fracture of this ice be examined, the appearance indicates that the 

 lower part of the formation has heea forced outwards. 



Whenever the shore is perpendicular to the water, or approaching to it, 

 this projection cannot take place in any considerable degree, and in such 

 circumstances, /I have seen the ice cracked in many places, and nume- 

 rous planes joined at the crack, elevated so as to be gently inclined to each 

 otberlike a very flat roof. This was the natural effect on the mechanical 

 principles which must govern the results. My views on this part of the 

 subject, are very much confirmed by the circumstances of the " Deerfield 



