It was the practice until comparatively recently to make the discharge outlet by laying pipes in a trench under the dam, generally at the lowest point in the valley, or constructing a culvert in the same position and carrying the pipes through this, and in the earlier works the valves or sluices regulating the outflow were placed at the tail of the down stream bank, the pipes under the bank being consequently at all times subject to the pressure of the full head of the water in the reservoir.  An instance of the first mentioned method is afforded by the Dale Dyke reservoir, Fig. 2, where two lines of pipes of 18 in. diameter were laid in a trench excavated in the rock and resting upon a bed of puddle 12 in. in thickness, and surrounded by puddle; the pipes were of cast iron, of the spigot and faucet type, probably yarned and leaded at the joints as usual, and the sluice valves were situated at the outer end of the pipes.  As the failure of this embankment was, as we all know, productive of such terrible consequences, it may be of interest to enter a little more fully into the details of its construction.  It was situated at Bradfield, six or seven miles from Sheffield, and at several hundred feet higher level.  Its construction was commenced in 1858, the puddle trench was probably taken down to a depth of 40 ft. to 50 ft., a considerable amount of water being encountered.  This trench was 15 ft. to 20 ft. broad at the top, and of course had to be crossed by the before mentioned line of pipes; and although the trench was filled with puddle, and the gullet cut in the rock already mentioned for carrying the pipes under the site of the dam was “padded” with a layer of 12 in. of puddle, we can imagine that the effect of the weight of the puddle wall and bank upon this line of pipes would be very different at the point where they crossed the puddle trench to what it would be where they were laid in the rock gullet and partially protected from pressure by the sides of the latter.  At the trench crossing there would be a bed of puddle 50 ft. in thickness beneath the pipe, in the gullet a bed of 1 ft. in thickness.  So much as regards the laying of the pipes.

The embankment had scarcely been completed when, on March 11, 1864, a storm of rain came on and nearly filled it up to the by-wash, when the bank began slowly to subside.  The engineer was on the crest at the very time, and remained until the water was running over his boots; he then rushed down the other slope and was snatched out of the way as the bank burst, and the whole body of water, about 250,000,000 gallons, rushed out through the trench, carrying with it in the course of about twenty minutes 92,000 cubic yards, or say one fourth of the total mass of earthwork, causing the death of 250 human beings, not to mention cattle, and destruction of factories, dwellings, and bridges, denuding the rock of its surface soil, and, as it were, obliterating all the landmarks in its course.  The greatest depth of the bank from ground level to crest was 95 ft., the top width 12 ft., and the slopes, both on the up stream and down stream sides, 21/2 to 1, and the area of the reservoir 78 acres.