Finally, a curved dam can resist changes of temperature better than a straight dam. The expansion and contraction of masonry is something like 1 inch per 100 feet per 100° F. change of temperature, and a drop in the temperature of the dam of 20° would cause a tension of something like 500 lb. per sq. in. Since such a change might occur—particularly when the reservoir was partly empty,—and since no masonry could stand that tension, it is wise to prevent the possibility of any such stress by building a gravity dam convex up-stream. The face of a curved gravity dam is likely to be in compression due to the overturning effect, and the back in compression due to the arch action; that is, the tendency of the overturning moment is to produce compression in the down stream face, while the tendency of the arch action is to produce compression in the up-stream face. However, there is not likely to be much of either action above the surface of the water—the portion exposed to the greatest changes of temperature,—but nevertheless the possibility of some such action is worth a little additional cost.
The limitations of space prevent a discussion of the matter.
There are forty-six masonry dams in the world over 100 ft. high; and of the forty-four whose plans are known, two are of the pure arch type, twenty-eight of the curved gravity type, and fourteen of the straight gravity type.
The theory of the best cross section for an overfall dam is not fully settled; but Fig. 105 shows the cross section of a dam near Atlanta, Georgia, which is representative of good practice.