being constructed at that time upon the same lines, viz., with culvert outlet under the dam, was, at the advice of Sir Robert Rawlinson, altered to a side tunnel outlet clear of the dam.

Some years previous to the failure of the Dale Dyke reservoir there occurred, in 1852, a failure of a similar character—though, as far as the author is aware, unattended by such disastrous results—at the Bilberry reservoir at Holmfirth, near Huddersfield, which had never been filled previous to the day of its failure, and arose from the dam having sunk, and being allowed to remain at a level actually below that of the by-wash; so that when the storm occurred, the dam was topped and destroyed. An after examination proved that the bank was badly constructed and the foundation imperfect.

Besides the above instances, there have been numerous failures within recent times of earthwork dams in Spain, the United States, Algeria, and elsewhere, such as that which occurred at Estrecho de Rientes, near Lorca, in Murcia, where a dam 150 ft. high, the construction of which for irrigation purposes was commenced in 1755 and completed in 1789, was filled for the first time in February, 1802, and two months later gave way, destroying part of the town of Lorca and devastating a large tract of the most fertile country, and causing the death of 600 people. The immediate cause of failure in this case the author has been unable to ascertain. In Algeria the Sig and Tlelat dams were destroyed in 1865; and in the United States of America, at Williamsburg, Hampshire Co., Massachusetts, in 1874, an earthwork dam gave way, by which 159 lives were lost and much damage done to property. In another case, viz., that of the Worcester dam, in the United States of America—impounding a volume of 663,330,000 gallons, and 41 ft. high, 50 ft. broad at the crest, and formed with a center wall of masonry, with earthwork on each side—which gave way in 1875, four years after its completion; here, as in almost all other instances of failure, the leakage commenced at a point where the pipes traverse the dam. In this case they were carried in a masonry culvert, and the leak started at about 20 ft. on the up stream side of the central wall. The opinion of Mr. McAlpine as to the cause of failure, which agrees with that of the most eminent of our own water engineers, was to the effect that "earthen dams rarely fail from any fault in the artificial earthwork, and seldom from any defect in the natural soil. The latter may leak, but not so as to endanger the dam. In nine tenths of the cases, the dam is breached along the line of the water outlet passages."

The method of forming the discharge outlet by the construction of a masonry culvert in the open has no doubt many advantages over that of tunnel driving through the hill side clear of the dam, permitting as it does of an easy inspection and control of the work as it proceeds; but a slight leakage in the instance of a side tunnel probably means nothing more than the waste of so much water, whereas in the case of the culvert traversing the site of the bank, the same amount or less imperils the stability of the bank, and in ninety-nine cases out of a hundred would, if not attended to, sooner or later be the cause of its destruction. I think the majority will therefore agree that the method of discharge outlets under the site of embankments should not be tolerated where it is possible to make an outlet in the flank of the hill, to one side, and altogether clear of the dam.

At Fig. 9 is a diagram of the Roundwood dam of the Vartry Water Works, supplying Dublin, which is a fair specimen of the class of earthwork dam with the outlet pipes carried in a culvert under the embankment, and which, perhaps, is one of the most favorable specimens of this method of construction, as the inlet valves are on the up stream of the dam, and consequently when necessary the water can be cut off from the length of pipes traversing the dam. A short description will be given. This dam is 66 ft. high at the deepest point and 28 ft. wide at the crest, having to carry a public road. The slope on the inner face is 3 to 1, and on the outer 2½ to 1. The by-wash is 6 ft. below the crest, which is about the average difference. The storage capacity of the reservoir is 2,400,000,000 gallons, or sufficient for 200 days' supply to the city. The puddle wall is 6 ft. wide at the top and 18 ft. at ground level, the bottom of the puddle trench about 40 ft. below the surface of the ground. The culvert was formed by cutting a gullet 14 ft. wide with nearly vertical sides through the rock, and covering it with a semicircular arch 4 ft. in thickness. Through this tunnel are laid a 33 in. and 48 in. main; the former for the water supply, and the latter for scouring or for emptying the reservoir on an emergency. There is a plugging of brickwork in cement under the center of the dam in the line of the puddle wall, and then stop walls built at the end of the plugging, projecting 25 ft. beyond the sides of the culvert and 8 ft. above, the space between them being filled up with cement concrete tied into the rock, and on this the puddle wall rests. This bank, like almost all others pierced by outlet pipes or culverts, was not destined to be perfect. In 1867, four years after the completion, spurts of water showed themselves in the culvert in front of the puddle wall, which began to settle, and the water had to be drawn off to admit of repairs. Diagram No. 10 shows a structure of a different character to any of these already described. This character of work is adopted on the North Poudre Irrigation Canal, in N.E. Colorado. Timber is there plentiful, and a dam of this character can be rapidly constructed, although probably not very durable, owing to liability to decay of timber. That represented is about 25 ft. high.

The author has now concluded the consideration of earthwork dams, and proposes making a few remarks upon those of masonry or concrete, with reference to some of the most important, as shown on the diagrams. Their stability, unlike those of earthwork, may be considerably increased where the contour and nature of the ground is favorable by being curved in plan, convex toward the water, and with a suitable radius. They are especially suitable for blocking narrow rocky valleys, and as such situations must, from the character of the ground, be liable to sudden and high floods, great care is necessary to make sufficient provision for overflow.

When of masonry, the stones should be bonded, not merely as they would be in an ordinary vertical wall, where the direction of the stress is perpendicular, but each course should be knit in with that above and below it in a somewhat similar manner to what is termed "random" work. And lastly, if hydraulic mortar be used, a sufficient time should elapse after construction before being subjected to strain, or in other words, before water is allowed to rise in the reservoir. For this latter reason, and also the liability to damage by sudden floods during the progress of the works, dams of Portland cement concrete, on account of their quick consolidation, possess advantages over those of hydraulic masonry apart from the necessity in the latter instance of constant supervision to prevent "scamping" by leaving chinks and spaces vacant, especially where large masses of stone or Cyclopean rubble are used.

Again, should the dam be drowned by flood during its erection, no harm would accrue were it composed of Portland cement concrete, whereas should it be of hydraulic mortar masonry, the wall would probably be destroyed or, at all events, considerably injured by the mortar being washed out of the joints. Portland cement, however, is only suitable for situations where the foundation is absolutely firm, as, should there be the slightest settlement, fissures would certainly be produced.

As regards foundations, the dam of the Puentes reservoir in Spain is somewhat remarkable—see Fig. 12. Its height is 164 ft., and the profile or cross section is of precisely the same character as that of the Alicante dam,