second time, and new strata of gypsum will be precipitated upon the marls containing the marine shells. This condition is exhibited in the delta of the Rhône, where the Lavalduc Lake, which has been isolated from the sea for centuries, has sunk to fifteen metres below its level. It has a stratum of gypsum on the bottom, and above it is a thick formation of mud, that has become a true marl where it has dried, at the base of which, and at a level corresponding with the epoch at which the water was nearly normal sea-water, marine shells are abundant. It may occasionally happen that a marine estuary, in which a deposit of gypsum has taken place, shall receive an accession of fresh instead of salt water, and then we shall have deposits containing freshwater fossils above the gypsum. We can see from this how valueless, as an argument against the theory that gypsums are products of the evaporation of sea-water, is the assertion, so often put forward in that guise, that gypsums are sometimes found covered with fresh-water deposits.
Similar phenomena and another illustration of our theory are exhibited on a colossal scale in the Caspian Sea. On the eastern side of that sea is the Gulf of Karabogaz, relatively small, but having a superficial area of at least twenty thousand square kilometres. It contains no living beings except some of the lower organisms, and its shores are marked by a complete sterility. Its only communication with the Caspian Sea is by a shallow channel which allows water to flow over from the sea, but lets none back. The excessive evaporation always going on in that hot and arid region causes a constant depression of the level of the water in the gulf, and this induces an incessant flow from the sea. In the absence of a counter-current from the Karabogaz to the Caspian, all of the salt that is brought in with the inflowing water remains in the Karabogaz; the amount of salt thus regularly added to the quantity already held there is not less than three hundred and fifty thousand tons every twenty-four hours. It is easy to prognosticate the future of this gulf. If the channel of communication is kept open, the water, now nearly saturated, will continue to deposit gypsum; but the constant accession of water from the sea will prevent its reaching for an extremely long time the degree of concentration that will permit the deposition of salt. There will in this case be produced in the Karabogaz a colossal deposit of gypsum, to which no parallel can be found in the ancient formations. If, on the other hand, the channel becomes obstructed, evaporation in the Karabogaz will go on more rapidly, for it receives no important affluent of fresh water. Then, at the end of a time which will not be prodigious, we shall have a saline deposit identical with that of Stassfurt, having large masses of gypsum at the bottom, and deliquescent salts with boracic acid at the top. The latter ending is more likely to be realized; for the level of the Caspian Sea itself is falling under the excess of evaporation over the supply of water brought by the rivers into it, and will at no very dis-
