Conditions indicating such a scheme of diversion may merge by any number of gradations into others which would properly call for-what is here termed Type 2 of diversion works.

The chief difference would be that the width of the valley is considerably less; in fact, narrow to the point that an artificial channel must be constructed along one side. The matter of desirable working area on the foundation, and necessary width for the channel, may require considerable excavation from the side hill and also that the channel be crowded into the side hill as far as possible.

Cofferdams below the channel inlet and above the outlet, respectively, turn the stream into the channel and prevent it from backing into the pit from below.

To keep down the width of the channel it may be given a steeper grade than the river, requiring of course that the water above the upper cofferdam be raised above the normal. Where the channel crosses the dam the foundation may be previously prepared, and the permanent masonry put in up to the level of the bottom of the channel. The channel may at this point have permanent masonry sides, with dividing piers to facilitate final closure; may be arched so that masonry construction may proceed above it, and have, in fact, whatever best meets the scheme of subsequent procedure. The river side of such a channel may be formed of the material excavated from the channel; depending upon the character of the material it may be necessary to put in a core of sheet piling to reduce the leakage; also to protect the channel from erosion, as by means of rip rap. A minimum width of channel would involve sides, and probably bottom, of timber or masonry, which at once conduce to and withstand high velocities. This may be particularly necessary for that portion of the channel opposite the pit, where an embankment of loose material would an objectionable or prohibitive degree upon the working area of the pit. (See Plate IV, Figs. A and B, also Fig. 17.) We may thus pass by degrees to the extreme case of this type, i.e., to a case in which the entire artificial channel is a flume. (See Plate IV, Fig. C, also Fig. 16.) Such a flume would probably be constructed of timber, with the possible exception of the section through the dam where the plan for subsequent diversion or final closure might indicate that it should be of masonry. At the Cross river dam, with a small flow to divert, the water was carried across the dam in pipes which were built into the masonry and subsequently filled with concrete. (See Plate VI, Fig. D).

Other things being equal the narrow channel will result in deeper water, requiring a higher cofferdam. This fact should be considered not only in the design of the cofferdam but in its effect upon leakage and pumping. If the capacity of the channel or flume is to be probably exceeded it will be well to provide for the filling of the pit at some other point and in some other manner than over the upper cofferdam, such as through a gate in the channel or flume. If the pit is full when the upper cofferdam is topped, the danger of washing out the cofferdam and of filling the pit with loose material is probably minimized. If it is possible for a wooden flume to be submerged its flotation should be estimated, and it should be adequately anchored to its foundation.

Type 3, in the United States, is met with in the rocky canyons of the West, such as for the Roosevelt, Pathfinder, Shoshone, and Arrowrock dams of the U. S. Reclamation Service. (See Plate III, Figs. A and C.) Eastern dams, being in comparatively spacious valleys instead of abrupt canyons, have used Types 1 and 2 with a great number of variations in detail to suit particular conditions.

Wherever this scheme is employed it is probable that the tunnel is a feature of the completed works as a reservoir outlet, and should not be charged entirely to stream diversion. Considered purely as a reservoir outlet such a tunnel should be viewed very critically for two reasons.


It seems to require or lead to, or at least so far to have been accompanied by, the installation of high-pressure gates. Such gates and their operating mechanism are complicated and costly and are not certain to give entire satisfaction in operation.


The leakage by or around such gates, although it may still be insignificant in amount, will be much greater than the leakage in connection with gates installed at or near the upstream end of an outlet through the body of the dam. The reason is that while the path of the leakage through masonry may be as long in one case as in the other, the areas over which the leakage is applicable or effective are vastly different.

In the case of gates on the upstream face of the dam, only a very limited area immediately around the gate can pass leakage by a short path through the masonry. The path rapidly becomes longer as area away from the gate is considered. In any event, the masonry, if not stone with thin joints and hence impervious, is laid in the open under conditions conducive to high-class work.

In the case of gates in a tunnel, leakage is applicable and effective over the entire inside area of the tunnel upstream from the gates. Passing through 1 ft. or 2 ft. of concrete lining the water reaches comparatively open channels, either through rock which is very likely to have been affected by blasting during the tunnel excavation, or else between the lining and the rock. Indeed, it would be a rare job of tunnel lining if no channels were there. Plenty of weep holes should be left in the tunnel below the gates in order that the leakage returning to the tunnel will exert no inward pressure on the lining.