38i GEOGRAPHICAL SYSTEMS. 
plains of Tafilet or Sigilmessa* But if, as the 
localities render probable, the travellers directed 
their course through Fezzan, they would, on 
emerging into the desert, be considerably to the 
south of those rivers ; and to direct their course 
northwards, would have been in direct contra- 
diction to their original object. Their travelling 
westward puts Bornou out of the question ; so 
that, though the distance be somewhat great, I 
do not well see to what other river their course 
could carry them, except the Niger. 
These data being established, Herodotus, as 
usual, proceeded to complete his system by arbi- 
trary suppositions. The Nile coming from the 
west, it appeared very natural to conceive, that 
the river of the Nasamones might fall into, and 
form the main branch of that greatest of the 
streams of Africa. The opinion, in fact, has 
not been controverted by actual observation ; 
though Major Rennell's learned arguments seem 
to have left it destitute of all shadow of proba- 
bility. * It appears, however, by the testimony of 
recent travellers,! that the same belief, founded 
on very nearly the same data, is still prevalent in 
Northern Africa. 
Herodotus evidently considered Africa as sur- 
* Geographical System of Herodotus, 
t Jackson, Horneman. 
