about 18 feet above the bed of the reservoir, in which apparently a depth of 13 feet of water was retained, the area covered being 652 acres and the resulting capacity 160 million cubic feet. I have met with no reference to any sluice at this reservoir. In one built in 1871, and afterwards replaced by a larger one, there were no stones of the kind that one would expect to find if the materials of an old bisōkoṭuwa had been utilised in it. It is possible that a small brick sluice may have been constructed long after the original works were made. A single escape for floods, about 100 feet wide, was left at the natural ground level on the eastern side, half a mile from the bank, behind some high ground against which the end of the bank abutted. From this, surplus water passed down a depression for three-quarters of a mile, and entered another reservoir now called Yōda-waewa, which appears to be the Duratissa tank of the histories. At the restoration of Tissa-waewa in 1871, as a great part of the top was found to be much worn away the higher parts of the bank were cut down to the extent of three feet, and the depth of water retained was ten feet. After more than a quarter of a century, however, it was found necessary to raise the water level once more to what seems to have been the height originally fixed by the old Sinhalese engineers. This is a high mark of appreciation of the excellence of their designs and their suitability under the conditions which control such works. The reservoir was of such vital necessity to the city that after experience had proved that it often remained unfilled during dry years, important measures were adopted in order to ensure its getting a better supply of water. For this purpose a permanent stone dam was erected across the Kirindi-oya, the river which flowed past the capital, at a distance of two and a half miles from the upper part of the tank. A short shallow channel, with a bed about ten feet wide, was then opened from a point immediately above it in the river, up to a site whence the water conveyed by it could flow into the tank by gravitation, without further works beyond the closing of a hollow which led back to the river. The age of this part of the scheme is unknown, but it must be an extremely early work, and possibly the dam was the first one of the kind built in Ceylon. The stones of the dam had been removed before I visited the place; but a few notes on it, in a Report written in 1858 by Mr. G. D. B. Harrison,1 are of interest. It was then altogether broken down by floods. He stated that it had had a height of fifteen feet, and that it was built of large roughlyhewn blocks of stone, few of which are less than a ton in weight, while many are far more. They appear to have been set dry, or without being imbedded in any mortar. . . . A great volume of water must have passed over the anicut [dam] during the rainy season, and carried with it large trunks of trees, with a force sufficient to destroy anything but the most massive masonry.' Floods, or rather the impact of the great tree trunks that they brought down, did, in fact, eventually destroy the dam, as well as nearly every other work of the kind in Ceylon. The body of water which is approaching a dam built across a river extends considerably below the level of the crest; but 1 Report of the Irrigation Commission, 1867, p. 229. immediately before arriving at the up-stream face of the masonry the lower part of the moving water rises so as to pass over it. With it rise any bodies that were being carried. near the surface, such as large trunks of dead trees; these are tilted obliquely upwards, and at that angle may strike the upper stones of the face of the dam. In that case, when the water is moving with great velocity and the tree trunk is very large (I have seen one of over ten tons stranded on one of these dams)-there is great probability that one of the stones of the top course will be displaced as in Fig. No. 132. It is in this manner that the ancient Sinhalese masonry dams have been breached almost without exception. Among the numerous ancient structures of this kind in Ceylon I have observed all stages of this destruction, from the displacement of the first stone on the up-stream face to the total demolition of the work. The special point of interest in the Kirindi-oya dam is the astonishing fact that instead of being taken across the river by the shortest possible line, as one would expect, it was built at an oblique angle, which, from the traces I saw, I judged to be nearly forty-five degrees from the direct line. There is a possibility that this does not prove that the principle of the oblique dam, and of its greater discharging power than one built square across a river-the knowledge of which was only acquired in comparatively recent years in Europe-was understood in Ceylon in very early times; Mr. Harrison, in commenting on this oblique dam, stated that in India there was an idea that one built at such an angle would be less exposed to the action of the current than one built square across the river. The Sinhalese possessed profound practical knowledge of the best methods of dealing with water, and the illustrations in Fig. No. 133, of typical dams 40 feet wide, the usual size of the larger ancient works, show clearly that they were correct if they believed that such a dam must have much greater stability than one of the same width built square across a river; and especially must be more capable of withstanding violent shocks due to the impact of great tree trunks, than the direct dam. It is evident that in the oblique dam the blow of a log carried by the water would have much less tendency to displace a stone than in the other. Nevertheless nearly all the later dams were built square across the rivers, probably because that was the line of the rocks on which they were founded. It is certain that the dam and its channel are not of much later date than the Duratissa tank next described, which in many years could not be expected to fill without their assistance. THE DURATISSA RESERVOIR This work is first mentioned in the second century B.C.; it is stated that King Saddha-Tissa (137-119 B.C.) built a wihāra there (Mah., i, p. 128). The construction of the reservoir is not referred to in the histories; it must have been made by a previous ruler during that century, or late in the third century B.C. Its purpose was chiefly the irrigation of rice lands. There is little doubt that this is the reservoir now called Yōda-waewa. In the first century A.D., King Ila-Nāga is stated to have executed some works of enlargement at it. There are also later references to it, the last one being in the reign of Parākrama-Bahu I (1164-1197 A.D.), when it is included with other large works which he restored; apparently it was then in a breached state. The embankment, about 3400 feet long, or nearly two-thirds of a mile, was taken in a north-west and south-eastern direction across the mouth of a subsidiary valley to the south-east of the Tissa tank, its south-eastern end abutting against high rocky ground at the point where the valley joins the low lands that stretch for seven miles between the Tissa tank and the sea. The bank was about 14 feet high above the sill of the sluice, and the depth of water retained by it was about nine feet, at which level the area was 1230 acres, and the capacity 336 million cubic feet. The top of the bank was about 15 feet wide, and was doubtless utilised as a cartroad. A single sluice was built at the south-eastern end of the bank. It consisted of the usual short inlet culvert, bisōkotuwa, and two discharging culverts. As restored, probably according to the original dimensions, the inlet culvert was 3 feet 6 inches wide and 2 feet 8 inches high; the bisōkoṭuwa was 13 feet 2 inches wide, in the line of the bank, and 12 feet 6 inches long, in the line of the culverts; and each outlet culvert was 2 feet 6 inches square. These were separated by a pier 2 feet thick. They passed the water into a channel with a base 10 feet 6 inches wide, excavated in decomposed rock. A small dāgaba was built on a rock at the side of this channel; possibly this was at the monastery founded by Saddhā-Tissa. The work in the sluice and culverts was of the usual type of stonework. A place for the escape of floods was left open at the side of the sluice, between it and the rocky hill against which the end of the embankment abuts. It was only about 60 feet wide. High floods apparently were allowed to escape round the other end of the bank. The reservoir received its water-supply partly from some short streams that flowed down from adjoining rocky hills, one being about four miles and another six miles in length; but its chief and unfailing source of supply was from the floodescape of Tissa-waewa, over which the water brought down from the Kirindi-oya dam flowed into Yōda-waewa. After these head-works had been constructed there was little fear of any loss of crops in the lands to which this reservoir supplied water; and it is evident that the prosperity of Magama was largely dependent upon them. Since the restoration of the Tissa tank and Yōda-waewa about 7500 acres of wild forest below them have been converted into rice fields; and the place, instead of possessing, as formerly, one of the most deadly climates of the island, is now fairly healthy. Numbers of healthy-looking children are to be seen about the houses of the cultivators. There is no place in Ceylon where a greater change has been effected by irrigation. YŌDA-KANDIYA RESERVOIR This work was formed in a very shallow valley on the western side of the Kirindi-oya, down which a small stream |