It seems that, when the ancient lake of the Limagne first began to be filled with sediment, no volcanic action had yet produced lava and scoriæ on any part of the surface of Auvergne. No pebbles, therefore, of lava were transported into the lake—no fragments of volcanic rocks imbedded in the conglomerate. But at a later period, when a considerable thickness of sandstone and marl had accumulated, eruptions broke out, and lava and tuff were deposited, at some spots, alternately with the lacustrine strata. It is not improbable that cold and thermal springs, holding different mineral ingredients in solution, became more numerous during the successive convulsions attending this development of volcanic agency, and thus deposits of carbonate and sulphate of lime, silex, and other minerals were produced. Hence these minerals predominate in the uppermost strata. The subterranean movements may then have continued until they altered the relative levels of the country, and caused the waters of the lakes to be drained off, and the further accumulation of regular fresh-water strata to cease. Lower Miocene Mammalia of the Limagne.—It is scarcely possible to determine the age of the oldest part of the fresh-water series of the Limagne, large masses both of the sandy and marly strata being devoid of fossils. Some of the lowest beds may be of Upper Eocene date, although, according to M. Pomel, only one bone of a _Palæotherium_ has been discovered in Auvergne. But in Vélay, in strata containing some species of fossil mammalia common to the Limagne, no less than four species of Palæothere have been found by M. Aymard, and one of these is generally supposed to be identical with _Palæotherium magnum,_ an undoubted Upper Eocene fossil, of the Paris gypsum, the other three being peculiar.