The intermixture is most difficult to account for. No doubt Africa and Europe were then united, and the theory of migration is invoked. The Arctic animals may, it is said, have moved south in winter and the African animals north in summer, and this was doubtless the case to some extent. But there are some facts which militate against this theory; for instance, the hyena caves, which seem to show a continuous occupation by the same African species for long periods. Nor is it easy to conceive how the hippopotamus could have travelled every summer from Africa to Yorkshire, and retreated every autumn with the approach of frost. Such instances point rather to long inter-glacial periods with vicissitudes of climate, enabling now a northern, and now a southern fauna to inhabit permanently the same region. Be this as it may, the fact is certain that this strange intermixture of northern and southern species is found in almost all the European deposits of the Quaternary age, until towards its close with the coming on of the second great glacial period, when the southern forms disappear, and the reindeer, with an Arctic or boreal flora and fauna, become preponderant, and extend themselves over Southern France and Germany up to the Alps and Pyrenees. The Quaternary period is therefore roughly divided by geologists into three stages: 1st, that of the mammoth and cave-bear, there being some difference of opinion as to which came first, though probably they were simultaneous; 2nd, the middle stage of the mixed fauna; 3rd, the latest stage, that of the reindeer. Now to these stages there is an exact correspondence in the character of the human implements found in them. In the earliest, those of the oldest deposits and of the oldest animals, we find the rudest implements. They consist almost exclusively of native stones, chipped roughly into a few primitive shapes: celts, which are merely lumps of flint or other hard stone with a little chipping to supplement natural fractures in bringing them to a point or edge, while the butt-end is left rough to be grasped by the hand; scrapers with a little chipping to an edge on one side; very rude arrow-heads without the vestige of a barb or socket; and flakes struck off at a blow, which may have served for knives. As we ascend to later deposits we find these primitive types constantly improving. The celts are chipped all over and the butt-ends adapted for haftings, so also are the other implements and weapons, and the arrow-heads by degrees acquire barbs. But the great advance occurs with the use of bone, which seems to have been as important a civilizing agent for palæolithic as metals were for neolithic man. This again seems to have been due to the increasing preponderance of the reindeer, whose horns afforded an abundant and easily manipulated material for working into the desired forms by flint knives.