our ideas are based essentially on the synthesis by Thys (1963) on the
tilapias pf Zaire and adjacent regions, and on Beadle (1974) for the paleo-
geograp$y. The maps (Figures 1 and 2) show that:
a. The genus Tilapia has a very wide distribution except for the eastern
slope of the eastern Rift valley and river basins facing the Indian
Ocean. They are abundant in west and central Africa. Speciation is
slow; they are 'stable', riverdwelling, close to the original stock. The
species are separated not by geographical or hydrographic barriers, but
rather by ecological or behavioral barriers.
b. The genus Sarotherodon is rare in west Africa, absent from the Zaire
basin, but abundant, diversified and specialized in small local popula-
tions in the Rift lakes. The subgenus Oreochromis (S. mossambicus, S.
mortimeri, S. andersonii and S. spilurus niger) occurs in the basins
facing the Indian Ocean. The various Sarotherodon species are segre-
gated by geographical and hydrographical barriers and generally have a
small range of distribution, except for S. galilaeus, S. niloticus and S.
mossam bicus.
The examples discussed below show the relationship between past geol-
ogical-hydrological events and the natural distribution of tilapias.
DISTRIBUTION OF ZILLII, S. GALILAEUS AND S. NILOTICUS
These three 'soudanian' species have a very wide common range (Senegal,
Niger, Chad, Nile, Jordan, Lake Turkana (Rudolf), Lake Albert) resulting
from former interconnections of the Chad and Nile basins.
S. niloticus, originating from the upper Nile in Uganda evidently moved
southwards, colonizing all the western Rift lakes down to Lake Tanganyika.
It also colonized central and western Africa, via the Chad and Niger basins.
Its expansion is still taking place; it has not yet reached some of the tribu-
taries of the upper Niger and it is rare in the coastal rivers of western Africa.
Conversely, T. zillii and S. galilaeus have spread eastwards towards the Nile
and the first Rift lakes.
It seems likely that the Chad-Nile connection came into existence later
than the southward push of S. niloticus along the Rift lakes, and later than
the separation of Lake Albert from Lake Edward, which could explain why
S. niloticus occurs in the latter lake though T. zillii and S. galilaeus did not
reach it. T. zillii and S. galilaeus were present in the Ubangi and Uele Rivers,
now tributaries of the Zaire river, before these two rivers were captured by
the Zaire. This capture must have taken place before the Chad-Nile con-
nection since S. niloticus is absent from the Ubangi-Uele system. The Ituri
River (Zdlre, central Africa) is a particular case: T. zillii evidently colonized
it in an eastward direction, but not S. galilaeus, which found itself competing
against S. niloticus already established. Before its integration with the
Zaire basin, the Ituri wadnked with Lake Albert, which was populated with
S. niloticus. In the great ZaTre basin, the expansion of T. zillii was prevented
by the presence of T. tholloni.
The S. galilaeus coming from the north did not manage to invade the Zaire
basin, which seems to be an environment hostile to microphagous cichlids,