256 EAST AFRICAN FLORA AND FA UNA 



The African fish in the Jordan, however, appear to have 

 entered that river at a later time, and they are not found in 

 the Lower Nile. 



As Dr. Giinther says, " Heniichroviis is not represented in 

 the north-eastern part of Africa proper, but chiefly on the west 

 coast and in the Central African lakes." Similarly the Clarias 

 of the Jordan is not the species {C. anguillaris, Linn.) of the 

 Lower Nile, but that {C. inacracantJms, Gthr.) of the Upper 

 Nile.^ It is no use, therefore, to assume the existence of a 

 connection between the Lower Nile and the Jordan, to account 

 for the existence in the latter of fish which do not occur in the 

 former. 



As, therefore, the route proposed for the migration of the 

 fish across the Esdraelon Gap is opposed to the evidence, and 

 useless as an explanation, we are compelled to turn to some 

 other line of connection between the Jordan and the rivers of 

 Equatorial Africa. 



It has been mentioned as certain that in Pliocene times a 

 large area in Palestine was occupied by a lake, the terraces 

 of which may still be seen on the banks of the Jordan and 

 around the shores of the Dead Sea. The fossils from the 

 terraces indicate that the water was fresh, so the lake prob- 

 ably had an outlet. This must have been either westward 

 across Palestine, or southward by the valley of the Wadi 

 Arabah, which leads to the Gulf of Akabah. The only 

 locality suggested for the former is the Esdraelon Gap, which, 

 as we have already seen, is improbable ; but the latter is sup- 

 ported by several arguments. It is true that the Wadi Arabah 

 depression is crossed by a ridge known as El Sate, which rises 

 to the height of 787 feet above the sea, or nearly 500 feet 

 higher than the Esdraelon pass. But the evidence already 

 quoted shows that the latter has been reduced in elevation by 

 denudation since Pliocene times, while there is reason to think 

 that the height of the Wadi Arabah ridge has been increased 

 by elevation. The equatorial part of the Rift Valley is divided 

 into a series of basins by transverse ridges, similar to that of El 

 Sate. Some of these are certainly due to elevation, for the old 

 lake terraces upon their flanks have been tilted from their 

 originally horizontal position (see PI. VIII. p. 94). Similar 



^ Giinther, op. cit. p. 228. 



