CHAPTER III.
SOUTH-WEST AFRICAN ISLANDS.
NSULAR groups are rare in the Guinea waters, although until recently the marine charts were liberally dotted over with phantom lands, which seafarers had observed on the horizon and mistaken for islands or reefs. It was even supposed that continental coastlines formed a southern fringe to the ocean furrowed by vessels sailing from Portugal towards India. Ptolemy's hypothesis of a "great Austral land," connecting South Africa with an eastern extension of Asia, had been revived by the cartographers of the sixteenth century, and modified in accordance with more recent discoveries. This Austral region was traced by them from South America along the southern limits of the oceans round the whole periphery of the globe. Then, with the progress of southern exploration, this coastline became broken into fragments, and on Homann's chart, published in 1722, a "Land of Life," fringed by a whole archipelago of islets, forms the southern boundary of the South Atlantic, under the latitudes where navigators had discovered the island of Gonçalo Alvarez. Fresh surveys thrust farther south these real or imaginary shores, which have now become the seaboard of the "Antarctic" Continent, and when Bouvet in 1739 discovered the island and group of islets now bearing his name, he called them "Cape Circumcision," supposing these snowy rocks to be a headland of the polar continent. But this region, lying between 54° and 55° south latitude, is already far removed from the African waters, being washed by the southern seas, which with their masses of floating ice form, so to say, a continuation of the southern glacial zone.
With the "Land of Life" have also disappeared several islands, the existence of which seemed firmly established by the circumstantial statements of navigators. Thus modern explorers have vainly sought for the island of Saint Matthew, which long figured on the old maps about the latitude of 2° 30' south of the equator. Yet the commander of a squadron of seven vessels had landed and remained fourteen days on this island in 1525. The description which he gives of it corresponds exactly with that of Annobon, which was probably the land visited by him, although he was out of his reckoning by some 600 miles, no unusual error in the history of navigation at that time. Another island, Santa-Croce, or Santa-Cruz,
