It often happens, however, that it is necessary to leave a low place in the dam for a considerable time, during which it is desirable that work on the masonry should be in progress over the greatest possible length of the dam. This necessity will usually be in connection with the river control and should be a most compelling one because such a break in the continuity of the work is to be avoided if possible. At the Wachusett dam two flumes were carried across the work and were maintained until the general level of the masonry was considerably higher than the flumes. A small flume which carried water to supply the Metropolitan District was operated until the permanent pipes and valves were installed and the foundations of the lower gate chamber built. A large flume carried the flow of the river until such time as progress on the dam and a low stage of the river combined to present a favorable opportunity for discontinuing it.

At the Roosevelt dam the river was turned over the north end of the masonry at elevation 9 while permanent gates were installed in the outlet tunnel at elevation o. Again a gap was left at elevation 106 to carry the flow of the river while repairs and alterations were made to the gates, and while the tunnel lining was being completed. The first gap was 50 ft. below the top of the masonry when the necessity for it had passed, and the second one was 90 ft. Another situation may occur which would in effect result in a gap, or rather in one side of a gap. Namely, when excavation for the dam up the hillsides interferes with plant or with progress on the masonry, there may be a tendency to defer it for a time and bring the dam up instead of extending it. This situation is entirely avoidable as the masonry should run out horizontally to the foundation instead of stepping down to it.

When a gap is left in the masonry, its form (inclination or rack of the sides) is a matter of some moment, the more so the deeper the gap, especially as when a gap is discontinued it is usually desirable, if not absolutely necessary, that it be filled with masonry as quickly as possible. Hitherto accepted practice has been to rack the masonry up away from the gap in a series of steps equivalent to a 1 to 1 slope or flatter if possible, and if the gap is to be a deep one. In the 50-ft. gap above mentioned the rack attained that height in 85 ft. horizontally, the 90-ft. one in 170 ft. horizontally. In the latter case much of the rack was equivalent to 1 to 1, a wide step being left in the middle.

Aside from the fact that the flatter rack is desirable on account of the character of the work, consideration must be given to the question of how derricks can be placed to fill the gap later. If the bottom of the gap is so narrow that derricks cannot set in it and readily build the masonry, it will be very convenient if not absolutely necessary to leave a wide step part way up the rack.

However, assume that a gap has been filled. The result will be that a joint following the rack from top to bottom will open along the junction of the fresh masonry and the older masonry. On the faces the cracks will appear distinctly, showing thin at the outer corners of the rack and thicker in the re-entrant corners. Although the character of the interior masonry is different from the regularly coursed face in that the different layers are bonded together, still no one can doubt that these cracks penetrate the entire mass. In fact, for several reasons, they may be much more pronounced in the interior than would be apparent from the appearance of the finished face because more settlement occurs in the early stages. As the face joints are raked out the crack is not so easily detected and observed, and it is only the settlement subsequent to pointing that causes the crack that shows in the completed dam. It is not argued that such a crack is liable to be a serious matter. Still it is undesirable and should be eliminated if practicable. The following method would appear to be better practice: Make the sides of the gap nearly vertical in steepness and eliminate the steps entirely as they interfere with the settlement, and it is the interference that causes the crack. Make the sides straight, with forms if necessary, so that the intersection of any vertical plane (across the gap) with the side would be a straight line. Of course, any horizontal plane which cuts the side could show as many skew backs or bonds as desired. This would allow the masonry to settle without any rupture and provide ample break in the continuity of the joint. In fact, this scheme would be analogous to the expansion joint introduced in so many recent dams.

The procedure suggested might be criticised in the small particular that it was not adaptable to the usual method of building the faces, i.e., in regular courses. True, if the face courses are stepped up or toothed up the same cause for rupture exists and the same crack will appear at the face, but it will be no worse and the whole interior will be sound. Or, unless the gap is one that could not be foreseen and planned for, pilasters could be built on the face to cover the joint. Building vertical sides to a gap would also result in one great advantage in the direction of not cutting down the area available upon which masonry could be built except for just the width necessary for the bottom of the gap.