A similar subterranean stream is said to conduct to the sea from another singular well near Tillipalli, in sinking which the workmen, at the depth of fourteen feet, came to the ubiquitous coral, the crust of which gave way, and showed a cavern below containing the water they were in search of, with a depth of more than thirty-three feet. It is remarkable that the well at Tillipalli preserves its depth at all seasons alike, uninfluenced by rains or drought; and a steam-engine erected at Potoor, with the intention of irrigating the surrounding lands, failed to lower it in any perceptible degree. Other wells, especially some near the coast, maintain their level with such uniformity as to be inexhaustible at any season, even after a succession of years of drought--a fact from which it may fairly be inferred that their supply is chiefly derived by percolation from the sea.[1] [Footnote 1: DARWIN, in his admirable account of the coral formations of the Pacific and Indian oceans, has propounded a theory as to the abundance of fresh water in the atolls and islands on coral reefs, furnished by wells which ebb and flow with the tides. Assuming it to be impossible to separate salt from sea water by filtration, he suggests that the porous coral rock being permeated by salt water, the rain which falls on the surface must sink to the level of the surrounding sea, "and must accumulate there, displacing an equal bulk of sea water--and as the portion of the latter in the lower part of the great sponge-like mass rises and falls with the tides, so will the fresh water near the surface."--_Naturalist's Journal_, ch. xx. But subsequent experiments have demonstrated that the idea of separating the salt by filtration is not altogether imaginary; as Darwin seems to have then supposed; and Mr. WITT, in a remarkable paper _On a peculiar power possessed by Porous Media of removing matters from solution in water_, has since succeeded in showing that "water containing considerable quantities of saline matter in solution may, by merely percolating through great masses of porous strata during long periods, be gradually deprived of its salts _to such an extent as probably to render even sea-water fresh_."--_Philos. Mag_., 1856. Divesting the subject therefore of this difficulty, other doubts would appear to suggest themselves as to the applicability of Darwin's theory to coral formations in general. For instance, it might be supposed that rain falling on a substance already saturated with moisture, would flow off instead of sinking into it; and that being of less specific gravity than salt water, it would fail to "displace an equal bulk" of the latter. There are some extraordinary but well attested statements of a thin layer of fresh water being found on the surface of the sea, after heavy rains in the Bay of Bengal. (_Journ. Asiat. Soc. Beng_. vol. v. p. 239.) Besides, I fancy that in the majority of atolls and coral islands the quantity of rain which so small