There is one condition which necessarily accompanies the construction of every large dam which is never taken into account in any discussion of the principles of design. It may be that this neglect is because of the difficulty, if not impossibility of assigning a correct value to it, or of suggesting a remedy. Nevertheless the condition may be recognized. The condition is that the construction of a large dam extends over several seasons. Now while in localities of severe winter climate construction is suspended during cold weather, it is easily possible that portions of the dam are built during mean temperatures 40 deg. Fahr. or more different from that prevailing during the building of other portions. It follows then that certain layers of masonry built in cold weather actually contain more masonry per unit of length than others built during warm weather. The measure of the amount is L X coefficient of expansion; and whatever the subsequent mean temperature of the mass, stress will exist between those layers. In Fig. 24, though a dam is not built in such regular layers, the principle is correctly illustrated.
Fig. 24. Showing possible range of temperature during construction, and its effect.
Suppose the assumed dam is built under the temperature conditions shown. Then the curve of effect shows the differences in length at the various elevations, whatever may be the subsequent temperature of the masonry, and stresses sufficient to produce those differences should exist. It may be argued that the maximum stress acts between two levels 60 ft. apart, and that the comparatively fresh masonry will tend to accommodate itself to some mean temperature, being assisted thereto by the force of gravity and resistance of side walls. This is undoubtedly true to a certain extent. However, the conditions assumed in this case are far from being the most severe that it is possible to encounter. The assumed range of temperature is moderate. Masonry is often built at 10 deg. or 15 deg. lower temperature. Further, the temperature for periods, even if short, might be over 80 deg. with a corresponding change in the curve of effect. Again some interruption to the work might sharply accent the curve. Thus a suspension of work for two months, say April and May or September and October, would bring three-fifths of the maximum effect to act between planes but a few feet apart. As an extreme case conceive of a 6-in. granolithic surfacing being built in January on top of masonry which had been built in July. Then subsequently the surfacing will be a detached mass, simply laying on the masonry, unless the temperature cracks in the masonry are carefully duplicated by expansion joints in the surfacing. Carefully devised expansion joints might relieve much of the stress, but only in the direction of the length of the dam; it is a remedy that could not be applied of course, to the stress across the dam. It is only intended here to call attention to conditions which seemingly must exist in some degree. Any intelligent measure of them must be prefaced by more accurate knowledge of the temperature changes in a mass of masonry.