Buckle drew his capital distinction, so constantly ignored by his critics, between "European" and "non-European" civilisations. This broadly holds good, but is a historical rather than a sociological proposition. The process of causation is one of life conditions; and the first great steps in the higher Greek civilisation were made in Asia Minor, in contact with Asiatic life, even as the earlier civilisations, such as the "Minoan" of Crete, now being traced through recovered remains, grew up in contact with both Egypt and the East. (Cp. Prof. Burrows, _The Discoveries in Crete_, 1907, chs. v, ix.) The distinction here made between "primary" and "secondary" civilisations is of course merely relative, applying as it does only to the historic period. We can but mark off the known civilisations as standing in certain relations one to another. Thus the Roman civilisation was in reality complex before the conquest of Greece, inasmuch as it had undergone Italo-Greek and Etruscan influences representing a then ancient culture. But the Roman militarist system left the Roman civilisation in itself unprogressive, and prevented it from being durably fertilised by the Greek. Proceeding from general laws to particular cases, we may roughly say that:-- (1) Primary civilisations arise in regions specially favourable to the regular production of abundant food, and lying inland, so as not to offer constant temptation to piratical raids. (Fertile coast land is defensible only by a strong community.) (2) Such food conditions tend to maintain an abundant population, readily lending itself to exploitation by rulers, and so involving despotism and subordination. They also imply, as a rule, level territories, which facilitate conquest and administration, and thus also involve military autocracy.