The geological evidence that the Mediterranean region has been subjected to oscillations of level during the tertiary period, is clear and decisive. Prof. Gaudry[241] has proved, in his work on the fossil remains found at Pikermi, that the plains of Marathon, now so restricted, must have extended in the meiocene age far south into the Mediterranean, so as to afford pasture to the enormous troops of hipparions and herds of antelopes, the mastodons and large edentata, revealed by his enterprise. The rocky area of Attica, as now constituted, could not have supported such a large and varied group of animals, nor could the broken hills and limestone plateaux have been inhabited by hipparions and antelopes, if their habits at all resembled those of their descendants living at the present time. It may, therefore, reasonably be concluded that Greece, in those times, was prolonged southwards, and united to the islands of the Archipelago by a stretch of land. If Africa were then as now the head-quarters of the antelopes, it is very probable that one of the lines by which they passed over into Europe, and spread over France and Germany, was in this direction. Nevertheless, it must be admitted that the changes of level, which have taken place since the meiocene age in those regions, are so complicated as to render it almost impossible to restore the meiocene geography. In the succeeding, or the pleiocene age, the presence of the African hippopotamus in Italy, France, and Germany, can only be accounted for by a more direct connection with the African mainland than is offered by a route through Asia Minor. It would seem, therefore, that the Mediterranean Sea could not then have formed the same barrier to the northern migration of the animals which it does now. In many regions, however, the present land was then sunk beneath the sea, and marine strata, of pleiocene age, were accumulated in the Val d’Arno, Sicily, and southern France. The physical geography[242] of the Mediterranean in the pleistocene age may be ascertained with considerable accuracy by the distribution of the animals, coupled with the evidence of the soundings. _Bone-caves of Southern Europe._ The mammalia in the bone-caves of southern Europe differ from those of the region north of the Alps and Pyrenees in the absence of the arctic species, and the presence of some which are of a more strictly southern type. Nevertheless, the influence of the mountains in lowering the temperature in their neighbourhood is to be traced in the presence of the remains of certain animals. Thus, in the caves of Gibraltar we find an ibex, which cannot be distinguished from those of the Spanish sierras, and in Mentone and Provence, a marmot, specifically identical with that of the Alps.