That the climate should be considered as cool and arid rather than comparable with the bitter-cold climate of the 'upper loess' period, when a true steppe fauna entered Europe for the first time, is further indicated by the fact that late Acheulean implements are more frequently found in the centre and north of France than in the south. To the far north, before the close of Acheulean times, the Scandinavian ice-fields had again begun to advance southward; the region bordering the glaciers was cold and moist and favored the migration from the tundra regions of the woolly mammoth and woolly rhinoceros to the locality still frequented by the Acheulean flint workers, for it is said(39) that Acheulean flints are occasionally associated even with the remains of these tundra mammals. At the very same time the Acheulean flint workers along the Somme may have enjoyed a more genial climate. It is only through this interpretation of the various climatic and life zones in western Europe that we can explain the survival on the River Somme, or return to this river from the south, of a warm temperate fauna, hippopotami, rhinoceroses, and elephants, in the Mousterian period, which is even subsequent to the close of Acheulean times. The valleys of the two great river-systems of southwestern France, the Dordogne draining the central plateau, and the Garonne draining the eastern Pyrenees, were now sought by the Acheulean flint workers. The valley of the Vézère, a northern tributary of the Dordogne, cuts through a broad plateau of limestone in which the streams have hollowed out deep beds with vertical sides. Here the landscape of late Acheulean times bore the same general aspect as at present.(40) Evidences of a change of climate are observed even in the sheltered valleys where the flint workers were seeking the warmer and sunnier river-slopes. The river channels were the same as they are to-day, and the quarries of the early Acheulean flint workers are found quite close to the streams; but as the period progressed they moved up nearer to the cliffs and shelters. Here, too, there is evidence that a dry continental climate prevailed. On the upper levels of the old plateaus of Dordogne we still find the _Quercus ilex_ occurring quite frequently, a tree which belongs to relatively dry regions and which in southern Russia is reckoned with the flora of the steppes. Yet the greater aridity toward the close of the Acheulean stage was probably not such as to prevent the growth of forests along the borders of the streams. Thus, in the mammalian life of the period there was, perhaps, a division between the more hardy forms which frequented the dry plateaus above and the forest-loving and less hardy forms which frequented the river-valleys. [Illustration: FIG. 86. "Valleys of the two great river-systems of southwestern France, the Dordogne draining the central plateau and the Garonne draining the eastern Pyrenees." After Harlé.]