that there is no room for the Terra Australis of classical and mediaeval cartographers. Nevertheless he was convinced that there was a nucleus of land in the middle of the ice-pestered sea, for he accepted the view that thick ice is not formed on the open sea. This view was based on the argument advanced by de Brosses in 1756, that as sea ice is sweet it must be formed on land, until in 1776 Nairsie explained that ice formed by the freezing of the sea water is fresh because the salt is extruded as brine. Cook, however, was no doubt quite correct in the view that the great fiat-topped Antarctic bergs could not be formed by the direct freezing of the open sea, but must have been formed on land.
Hence in spite of the comparatively narrow limits within which Cook's work had restricted the possible existence of Antarctic land, the search for it was still continued. Islands were found south of the Atlantic; but it was not till 1840 that any extensive land area was discovered south of the Pacific and Indian Oceans. Then almost simultaneously a French expedition under Dumont d'Urville, and the American expedition under Wilkes discovered the long coast line or chain of islands known as Wilkes Land.
Wilkes' work was not only important because he traced this coast line at intervals for 60 degrees of longitude; but the geological collections made by his expedition showed that the land is formed of granites, massive sandstones and other rocks of continental types.
Two years later the extension of Wilkes Land to the east and the south was proved by the famous expedition of Sir James Clark Ross, which circumnavigated the Antarctic area and passed all previous