rapids, which cut off from outward intercourse populous regions whose fluvial systems ramify over many hundred millions of acres. The Nile and Congo rising amid the higher plateaux, where the slope is still undecided, traverse in their upper courses many great lakes, which according to a vague tradition once constituted a single lacustrine basin of enormous extent. In the sixteenth century the Portuguese explorers had some idea of this hydrographic system. But in tracing the outlines of the great equatorial lakes they seem to have rather copied older maps than relied on positive information. But, however this be, they appear to have believed in the existence of a single source for the Nile, Congo, and even the Zambesi. But the streams were also supposed to traverse extensive underground regions, and an Italian map engraved in the middle of the fifteenth century represents a Nile with three heads, separated by a vast space from the emissaries of the chief fountain. This Nile is moreover made to flow in the direction from north to south, a small Egyptian delta corresponding to a much larger delta in South Africa.
The first modern explorers of the same region were also influenced by these traditional ideas. Even Speke traced the course of four rivers issuing from various parts of Lake Nyanza to form the Nile, while Stanley made Tanganyka the source of two effluents, one flowing northwards to the Nile, the other westwards to the Congo. But although these great arteries do not rise in a common source, the water-parting between them is in some places so low and undecided that a slight disturbance of the surface would suffice to change the direction of many affluents. It is even possible that on the dividing line of some basins there may exist lakes or swamps draining in both directions.
The unfinished aspect of the central rivers, the cataracts interrupting their course, the lacustrine reservoirs scattered over the plateaux, produce a certain resemblance between equatorial Africa and the Scandinavian peninsula. But in the northern region, still under ice within a comparatively recent geological epoch, the rivers have scarcely commenced their work of erosion. The climatic conditions are of course entirely different, and although the existence of an old glacial period may be suspected even in the torrid zone, the long ages that have elapsed since that remote epoch must have effaced nearly all trace of glaciers and moraines. Hence the rudimentary character of these fluvial basins is probably due to a different cause. The climate, which was formerly much more humid in the Sahara, may possibly have been correspondingly drier in the south-eastern region of the Nyanza plateau. In the absence of a copious rainfall the rocks would remain uneroded, and the now flooded cavities unfilled by the alluvia of running waters. During its long geological life the earth has witnessed many shiftings of the climatic zones. If the rains are more abundant in some places than formerly, in others they are more rare, and the Igharghar basin, for instance, in North-west Africa, belongs to one of these dried-up regions.
East of the Nile and of the great lakes there is no space between the plateaux and the coast for the development of large streams. From the Egyptian uplands the Red Sea receives nothing but intermittent wadies, and along a seaboard of about 2,400 miles southwards to Mozambique the Indian Ocean is fed only by such