The Assouan dam contains many abrupt changes of section, both on account of considerable steps in the foundation and from the sluices through the masonry. While much of it was built during very hot weather, the annual range appears to be no greater than at the Boonton dam. Six or seven cracks appeared, the largest occurring at one of the most sudden changes of section and extending from the top of the dam down almost to the rock foundation. It was about 1/4 in. thick at the bottom in cold weather, rather more at the top, and extended through the dam from side to side. The leakage amounted to only about 5 gal. or 10 gal. per hr. and decreased very much after the reservoir had been full for some time.

The paper from which the foregoing is abstracted contains progress profiles of the New Croton and Boonton dams showing the location of the cracks. A study of the cracks shows, as would be expected, that the number and extent of the crack is governed largely by the atmospheric temperature during the construction period as compared with the winter temperature. The conclusion is most plain that in order to minimize the cracks the masonry built during hot weather should be covered with a thick layer built during cool or cold weather. In short, construction should proceed steadily as far into the winter months as practicable, and the top of the dam should be finished during a period of low temperature.

At the Roosevelt dam permanent provision was made for keeping the exposed faces wet during hot weather after the completion of the dam as well as during the latter part of the construction period. At this dam the minimum temperature would be not far from 20 deg. Fahr. and the maximum in the sun certainly over 160 deg. Fahr. A pipe, 2 in. reducing to 1 in., was run along the downstream face about 25 ft. below the top. A small hole was bored every 10 ft., sufficient to keep the face of the dam continually moist. (See Plate VIII, Fig. B.) The pipe was supplied with water by gravity from a canal slightly above and independent of the reservoir level although not high enough to cover the crest of the dam. Through another pipe water can be supplied by pumping to serve similarly all of that portion above the gravity system, including the roadway on top. The thin film of moisture on the face of the dam was not intended to convey to the masonry the temperature of the water, but to produce an even lower temperature from the effect of the rapid evaporation which is a feature of that arid climate. The temperature of the face masonry under these conditions should be many degrees, quite possibly 100, below the maximum that it would otherwise attain. The top of the dam was constructed, except for portions of the concrete roadway and stone parapet walls, during the summer of 1910. Above the level of the gravity system, it had not received such a beneficial effect from sprinkling. During the latter part of October cracks averaging perhaps 1/16 in. in thickness appeared across the top at quite regular intervals of 40 ft. to 50 ft. In but one case (and that at a pronounced change in section) did the crack extend more than a few feet down the faces. After having tried reinforcement in the tops of the Cross River and the Croton Falls dam, without however much apparent effect on the cracks, recent practice has been to govern the number and location of cracks by introducing expansion joints (or rather to substitute joints for cracks). Expansion joints are provided at intervals, varying from 40 ft. to 80 ft. in different dams. They consist of smooth joints constructed across the dam, the first face constructed being covered with a plastic waterproof material and the second being built against it. See Plate XI, Figs. A and B, also Fig. 22. At one or more points along this joint the masonry upon one side contains a rectangular recess which is filled by a corresponding projection in the masonry of the other side, forming rectangular bonds, such that movement in the direction of the length of the dam will not tend to open or close the joint upon two sides of the rectangle. In the Kensico and Farn-ham dams a metal strip crosses the joint with an edge embedded in the masonry upon each side. The joint immediately downstream from the metal strip is intercepted by the drainage system previously described. (See pages 106 to 109).

The method of constructing the joint is to build up a wall of smoth-faced concrete blocks, carefully pointing the joints. This wall is kept one or two courses in advance of the cyclopean concrete of the section being built, i.e., of the section of which the blocks are a part. The section may be carried up almost any desired distance in advance of the adjoining section. (See Plate XI, Figs. A and B.) There is no apparent reason why the section face could not be built against a form instead of being built of concrete blocks; the form would certainly be much cheaper. Say that the materials entering into the concrete of the blocks costs the same as a like quantity of the materials entering into the cyclopean concrete. Then against the cost of mixing, transporting and dumping of the mass concrete, we have the cost of making blocks (which, including curing, storing and handling is about $1.00 per cu. yd.), transporting them to the dam, setting them with derricks in the same manner as face work and finally pointing the joints. Certainly this means a cost for the blocks which is many times the cost of a form to accomplish the same result.