The provision in the design of dams for a possible thrust from a field of ice is a question which has for some years been recognized, and regarding which considerable controversy has been waged. The board of experts who advised in the design of the Quaker Bridge dam recommended that provision be made for a thrust from the ice, at full reservoir level, of 43,000 lb. per lin. ft. of dam. In the dam as later built some distance upstream at a new location, and known as the New Croton dam, no provision for ice thrust was included. Several dams constructed since then however have included the provision. The dams, and the amount of the allowance in pounds per lin. ft. are as follows: Wachusett 47,000; Olive Bridge 47,000; Kensico 47,000; Croton Fails 30,000 and Cross River 24,000.

These allowances must have been based upon two assumptions, namely, the amount of the crushing strength of ice and the maximum thickness of ice expected in the particular locality. Although the figures for these assumptions have not been stated, it may be interesting to observe that 47,000 lb. per lin. ft. of dam corresponds to assumptions of strength of ice 200 lb. per sq. in., and a thickness of 19 5/8 in. The American Civil Engineers Pocket Book quotes U. S. Engineer Corps experiments showing a crushing strength of 100 lb. to 1000 lb. per sq. in. Doubtless, such wide variation depends entirely, upon the conditions under which the ice was formed.

If it is proper under any given circumstances to include in the design some allowance for the thrust of ice, 200 lb. per sq. in. may be a close enough approximation to crushing strength of ice formed under actual conditions - but against what is the reaction applied? A certain gentleman once called attention to the fact that a fulcrum was as necessary as the lever in order that he might move the world; a perfectly sound principle whatever the weights involved.

In current discussion regarding the matter of ice thrust, there is much that is superficial and ill considered. Thus the question of what is to hold the ice while it is exerting the thrust seems to have been either misapprehended or ignored entirely. Apparently the question of thrust has been confused in some minds with the idea of motion of the sheet as a whole. For example, in the Transactions of the American Society of Civil Engineers, an engineer advances the following as one of the reasons why no provision was made for ice thrust in a recent important dam: "Because the point of rock projecting into the reservoir in front of the dam will protect it in such a way as to make this thrust small." This line of reasoning would seem to lead to the conjecture that a rocky point opposite the entire length of the dam, and but a short distance from it, would afford complete protection. As a matter of fact, it should be obvious that the shorter the ice column and the more unyielding the abutments, the more likely is the ice to exert an effective thrust.

Only one case, that of a low overflow dam, is on record where the failure of a dam was unmistakably attributable to ice thrust. Opposed to the upstream side, and at a short distance, was a masonry foundation wall. A considerable thickness of ice was formed, and the fluctuations of water level were such that the ice acted with a toggle-joint effect between the dam and the foundation wall.

On a pond the ice will expand and crowd into or slide up onto the shores. The movement may be considerable, and the causes and process may be analyzed as follows: As the underside of the sheet of ice is in contact with the water, it takes its temperature from the water and is constant at 32 Fahr. While the ice is thin the temperature through its entire thickness may be approximately the same, but as the ice grows thicker the top surface of the sheet gets more and more away from the influence of the water and is subjected to that of the air. As the air gets colder the upper surface contracts and cracks. Water enters the cracks and is immediately frozen.

Then when, as during a sunny day, the ice expands from the rise in temperature, it undoubtedly opens the same or additional cracks on the underside of the sheet which are in turn filled with water to be subsequently frozen. Thus the process may repeat itself with each marked fluctuation of temperature. The movement of the ice and the unequal resistance offered by different portions of the shore doubtless acts to increase the width of the cracks, or a portion of them, and this effect is in turn the cause of additional filling, freezing and movement.

The Boston Metropolitan Water Board made observations for some two months one winter on the motion of the ice around the shore of a pond. The pond was about 1/2 mile in diameter, the ice 12 in. to 15 in. thickness, the water level practically constant and the surface generally was free from snow. At seven or eight places around the pond distances were carefully measured between points on the shore and points opposite on the ice. During the period of about two months the points on the ice all moved toward the shore by amounts varying from 4 in. to 7 in., and apparently depending principally on the steepness of the bank and the amount of resistance that it offered. No attempt was made to measure the force exerted.

Disregarding ice which is in motion as a sheet, which is subject to wave action or violent fluctuations of elevation, and considering only ice under such conditions that it may produce thrust as here discussed, it would be a fair statement of the case to say that if ice can ever exert a crushing thrust against a dam it could crush against itself. Yet this phenomenon was never observed and never will be. Ice moves into or slides up onto a shore simply because the shore offers less resistance to the ice than the ice itself offers to compression.

The experiments, as to crushing strength, noted above, also yielded figures varying from 6 per cent, to 30 per cent, as to amount of compression before crushing. Assuming any intermediate value we like as being applicable to an actual case, and assuming the worst possible location for the dam, it is almost impossible to conceive that the ice can exert more than a small fraction of its crushing strength as thrust against the dam. It should not be difficult to devise nor expensive to conduct under actual working conditions a series of experiments to ascertain: first, what the expansion of a field of ice would be; second, the amount of thrust that the ice could exert when confined between immovable opposing shores. In all probability the results of such experiments, though doubtless interesting, would show that an allowance for ice thrust is very seldom, if ever, necessary in the design of a masonry dam.