It may be mentioned here that the introduction of such joints operates to reduce the amount of stone which it is feasible to introduce per cubic yard of masonry. The reason is that such stone, in order not to interfere with the construction of the joint, must be kept away from the joint upon one side and from the blocks forming the first side of the joint upon the other. It is probable that the Olive Bridge and Kensico dams would otherwise have contained a percentage of stone nearly or quite equal to that in the Cross River, viz., 35 per cent instead of 25 per cent In addition to taking care of any longitudinal movement due to changes of temperature, these joints obviously afford a perfect adjustment for any possible unequal settlement of different sections of the dam, so that no cracks can be produced from that cause. A dam with such joints therefore may be regarded as a series of dams placed end to end with the leakage between them insignificant in amount and intercepted.

Expansion joints are in reality one of the most significant features of the entire advance in practice of dam construction, not solely nor even principally owing to their function in caring for settlement and temperature movements, but on account of the purely incidental function of making possible and even facilitating certain changes in construction methods. Briefly the change consists in a departure from the method of construction in horizontal sections to or toward a method of construction in vertical sections. This may become a most radical departure although so far it has not been made use of to the extent of modifying previous construction plant or methods.

For the purpose of handling water during construction, dams 40 ft. or 50 ft. high have been built in alternate sections with (in effect) similar joints between sections. If for any reason whatever any marked advantage will result, there seems to be no reason why the method should not be extended to its logical conclusion and be applied to the construction of consecutive if not alternate sections of a high dam. A writer has recently questioned the advisability of such procedure on account of the stresses which would follow possible extreme differences in temperature between the face of a completed section and the fresh masonry being built in the adjoining section. It can be shown, however, that such an objection is based upon conditions which may be entirely eliminated or the importance of which has been exaggerated.

Thus, assume that a vertical face is a certain number of degrees colder than the new concrete that is being deposited against it; still the temperature of each will be affected by the other during the interval between the deposition and the hardening of the fresh concrete. The old concrete will be warmed for some distance on its side of the joint and the new concrete cooled on the other side, so that in effect the transition will not be sudden but gradual. Even if some subsequent adjusting motion along the joint might be necessary, the joint will accommodate transverse motion even better than longitudinal. The purpose of the joint is to allow the masonry to move without rupturing itself in the endeavor. At the worst the conditions would be better than under the usual procedure in building dams. As pointed out in the paragraph on Initial Stresses the same conditions have always been encountered and they should be less inimical if confined to a vertical transverse plane than if occurring in a horizontal one where no joint is provided. If any motion is possible, or is anticipated, that would render two vertical bonds or recesses in the joint inadvisable, it should be remembered that one is all that is necessary and that one is better than two.

If expansion joints at right angles to the axis of a straight dam are feasible, so also would be radial joints in a curved dam. Not that the radial joints would be necessary as expansion joints but they may be advisable because they would adjust settlement and permit the above suggested departure in construction methods. The extent of the revolution in construction methods to which such joints may give rise will be discussed in the chapter on Probable Future Methods.

At the Farnham dam (see page 243) the transverse expansion joints are 80 ft. apart, with 6 in. recesses 8 ft. apart along the joint. Each joint is intercepted by a drainage well on the upstream side of which a copper strip crosses the joint. The first side of the expansion joint was formed with concrete blocks, and the face was covered with paper between two coats of pitch before the cyclopean masonry of the other side was built against it. During the last of February, 1913, with reservoir full, the drainage wells intercepted and carried away leakage amounting to 18,000 gal. per day. During the following summer, with reservoir drawn down 10 ft. or 12 ft., this leakage had decreased to 200 gal. per day.