As to the stability of structures of masonry for this purpose, as compared with earthwork, experience would seem to leave the question an open one.  Either method is liable to failure, and there certainly are as many cases on record of the destruction of masonry dams as there are of those constructed of earthwork, as instanced in Algeria within the past few years.  As regards masonry dams, the question of success does not seem so much to depend upon their design, as far as the mere determination of the suitable profile or cross section is concerned, as that has been very exhaustively investigated, and fairly agreed upon, from a mathematical point of view, but to be principally due to the correctness of the estimate of the floods to be dealt with, and a sufficient provision of by-wash allowed for the most extreme cases; and, lastly, perhaps the most important of all, the securing a thoroughly good foundation, and a careful execution of the work throughout.

These remarks equally apply to earthwork dams, as regards sufficient provision of by-wash, careful execution of work, and security of foundation, but their area of cross section, supposing them to be water-tight, on account of the flatness of their slopes and consequent breadth of base, is, of course, far in excess of that merely required for stability; but in these latter, the method adopted for the water supply discharge is of the very greatest importance, and will be again referred to.

Before commencing the excavation for the foundations of a dam, it is most essential that the character of the soil or rock should be examined carefully, by sinking a succession of small shafts, not mere borings, along the site, so that the depth to which the trench will have to be carried, and the amount of ground water likely to be encountered, can be reliably ascertained, as this portion of the work cannot be otherwise estimated, and as it may bear a very large proportion of the total expense of construction, and in certain cases may demonstrate that the site is altogether unsuitable for the proposed purpose.

The depth to which puddle trenches have been carried, for the purpose of penetrating water-bearing strata, and reaching impenetrable ground, in some cases, has been as much as 160 ft. below the natural surface of the ground, and the expense of timbering, pumping, and excavation in such an instance can be easily imagined.  This may be realized by referring to Fig. 4, giving a cross-section of the Yarrow dam, in which the bottom of the trench is there only 85 ft. below the ground surface.  In the Dale Dyke dam, Fig. 2, the bottom of the trench was about 50 ft. below the ground surface.

There is one other point which should be mentioned in connection with the form of the base of the puddle trench—­that instead of cutting the bottom of the trench at the sides of the valley in steps, it should be merely sloped, so that the puddle, in setting, tends to slide down each inclined plane toward the bottom of valley, thereby becoming further compressed; whereas, should the natural ground be cut in steps, the puddle in setting tends to bulge at the side of each riser, as it may be termed, and so cause fissures.  It will be noticed that the slopes of these earthwork dams vary from 7 to 1 to 2 to 1.