Hundreds of miles on the northern edge of the desert, from the Atlantic along the southern foot of the Atlas to the Great Syrtis, are pasture-lands without a tree—an ocean of verdure. At the Great Syrtis the Sahara comes to the shores of the Mediterranean; and, indeed, for 1100 miles between the termination of the Atlas and the little table-land of Barca, the ground is so unprofitable that the population only amounts to about 30,000, and these are mostly wandering tribes who feed their flocks on the grassy steppes. Magnificent countries lie along the Mediterranean coast north of the Atlas, susceptible of cultivation. History, and the ruins of many great cities, attest their former splendour; even now there are many populous commercial cities, and much grain is raised, though a great part of these valuable kingdoms is badly cultivated or not cultivated at all. The base of the sandy parts of North Africa is stiff clay; in Lower Nubia, between the parallels of Assouan and Esneh, red and white granite prevail, followed by argillaceous sandstone; Middle Egypt is calcareous; and, lower down, the alluvium of the Nile covers the surface. It would appear that Southern Africa, though similar in its unbroken surface and peninsular shape to South America, bears no resemblance to it in other respects, but has a great analogy to the Deccan in its triangular form, its elevated platform, and in the position of its encompassing mountain-chains, if, as there is every reason to believe, from the fertile region to the north, either that South Africa descends in a succession of terraces to the low lands, or that the Komri mountains have a real existence, and run directly across the continent. From the connection already mentioned between external appearance and internal structure, as well as from partial information, it is surmised that the mountains surrounding the two triangles in question are of corresponding constitution; that, if any secondary strata do exist in this part of Africa, they must be exterior to these chains, and neither on the summits of the high mountains nor in the interior; and that any tertiary strata on the table-land must, as in the Deccan, have formed the basins of fresh-water lakes.[39]