382 GEOGRAPHICAL SYSTEMS 
the name of Libya suggests the region south of 
the Atlas, much more than the depths of the in- 
terior, which were always designated by the appel- 
lation of Ethiopia. 
While it seems impossible to resist the force of 
these arguments, there are other circumstances in 
Ptolemy's description, which would lead us to look 
in Nigritia for the country which he here describes. 
The Nuba Palus is a name belonging to quite a 
different part of the continent, from that watered 
by the Adjidi. Mount Mandrus, as the western 
termination of the Niger, suggests in the strongest 
manner the territory of Manding. Panagra agrees 
almost equally with Wangara, and Caphas with Kaf- 
faba. The description, both by Ptolemy and Agathe- 
merus,* of the Niger and Gir, as fxeyisroi iror<xfjioi 9 
rivers of the greatest magnitude, and the placing 
them on a level with the Nile, agrees very ill with any 
streams which descend from the Atlas. The same 
may be said of the direction of the stream of the 
Niger east and west, while all the rivers of Sigil- 
messa run from north to south. Finally, unless 
some communication had been opened in that age 
with this part of central Africa, it seems difficult 
to discover how the Roman empire could have 
been supplied with the precious commodity of gold, 
which must have been in extensive demand, and of 
which no peculiar want seems to have been felt. 
The Periplust proves it to have been obtained 
* Geog. Greec. Minor. II. 49. 
f lb. I. 4, 10, SI. 
