At the end of the eighteenth century it had been rather grudgingly admitted that some few animals were actually extinct. Buffon was able to quote as indubitable examples the mammoth and the mastodon. Their occurrence in countries unknown to the ancients, such as Siberia and North America, disposed of the explanation long clung to by the learned—viz., that their bones were the remains of elephants which had been led about by the Roman armies, while their large size and the ease with which they can be recognised rendered it highly improbable that they still survived anywhere on the surface of the globe. It was therefore natural that Cuvier's first study in palæontology should relate to extinct elephants. He compared and distinguished several species, showed that they were distinct from the existing Asiatic and African species, a fact which had escaped the notice of Pallas, and argued from the well-known case of a Siberian mammoth preserved in ice and frozen mud with hardly any decomposition that it must have been overwhelmed by a sudden "revolution of the earth." Whatever we may think of Cuvier's geology, his comparisons of all known elephants, recent and fossil, introduced a new standard of exactness into these inquiries. From this beginning he went on to study all the extinct vertebrates which he could discover in public or private collections. By 1821 he had published elaborate and well-illustrated descriptions of near a hundred extinct animals, an extraordinary output for one investigator. The most remarkable of his palæontological discoveries were made at home, in the lower tertiary rocks which underlie the city of Paris. He proved that in the valley of the Seine a large population of animals, all now extinct, had formerly flourished. None of these discoveries impressed his contemporaries more than the celebrated case of the fossil opossum. The bones were imbedded in a slab of gypsum, and were at first imperfectly exposed. The lower jaw, however, exhibited a peculiarity of marsupial or pouched animals, for its angle had an inwardly projecting shelf, not found in other quadrupeds. The opossums, like all marsupial animals, bear on the front of the pelvis two long bones, which support the pouch. These were as yet concealed, and Cuvier delayed clearing them until he had summoned friends, some of whom may have been sceptical about the possibility of reasoning with certainty from anatomical data. Warning them what to expect, he removed with a sharp tool the film of stone, and revealed the long and slender marsupial bones.[34] The ancient existence of marsupials in France was then a striking and almost incredible fact; increase of knowledge has not lessened its interest, though it has abated some of the wonder.