‘It is impossible,’ says Humboldt, {260b} ’to cross the burning plains’ (of the Orinocquan Savannas) ’without inquiring whether they have always been in the same state; or whether they have been stripped of their vegetation by some revolution of nature.  The stratum of mould now found on them is very thin. . . .  The plains were, doubtless, less bare in the fifteenth century than they are now; yet the first Conquistadores, who came from Coro, described them then as Savannas, where nothing could be perceived save the sky and the turf; which were generally destitute of trees, and difficult to traverse on account of the reverberation of heat from the soil.  Why does not the great forest of the Oroonoco extend to the north, or the left bank of that river?  Why does it not fill that vast space that reaches as far as the Cordillera of the coast, and which is fertilised by various rivers?  This question is connected with all that relates to the history of our planet.  If, indulging in geological reveries, we suppose that the Steppes of America and the desert of Sahara have been stripped of their vegetation by an irruption of the ocean, or that they formed the bottom of an inland lake’—­(the Sahara, as is now well known, is the quite recently elevated bed of a great sea continuous with the Atlantic)—­’we may conceive that thousands of years have not sufficed for the trees and shrubs to advance toward the centre from the borders of the forests, from the skirts of the plains either naked or covered with turf, and darken so vast a space with their shade.  It is more difficult to explain the origin of bare savannas enclosed in forests, than to recognise the causes which maintain forests and savannas within their ancient limits like continents and seas.’

With these words in my mind, I could not but look on the Savanna of Aripo as one of the last-made bits of dry land in Trinidad, still unfurnished with the common vegetation of the island.  The two invading armies of tropical plants—­one advancing from the north, off the now almost destroyed land which connected Trinidad and the Cordillera with the Antilles; the other from the south-west, off the utterly destroyed land which connected Trinidad with Guiana—­met, as I fancy, ages since, on the opposite banks of a mighty river, or estuary, by which the Orinoco entered the ocean along the foot of the northern mountains.  As that river-bed rose and became dry land, the two Floras crossed and intermingled.  Only here and there, as at Aripo, are left patches, as it were, of a third Flora, which once spread uninterruptedly along the southern base of the Cordillera and over the lowland which is now the Gulf of Paria, along the alluvial flats of the mighty stream; and the Moriche palms of Aripo may be the lineal descendants of those which now inhabit the Llanos of the main; as those again may be the lineal descendants of the Moriches which Schomburgk found forming forests among the mountains of Guiana, up to four thousand feet above the sea.  Age after age the Moriche apples floated down the stream, settling themselves on every damp spot not yet occupied by the richer vegetation of the forests, and ennobled, with their solitary grandeur, what without them would have been a dreary waste of mud and sand.