231 
diterranean and the ocean, and at the east are bound- 
ed by the sandy deserts, on the one side reaching to the 
Atlantic ocean, and on the other to the Mediterranean, 
forms a real island, without being in any apparent connec- 
tion with the other mountains of Africa. 
All what we know of the deserts of sand which sur- 
round the chain of the Atlas to the east and south sides P 
proves that they are not like those of Tartary, compo- 
sed of the humus depauperatus of Linneus, that is to 
say, of an earth, which, in consequence of perpetual til- 
lage and production, is become exhausted and depri- 
ved of those organic particles which are necessary to 
vegetation. 
One may easily judge of the deserts which lie to the 
south of the Atlas, by those which I have seen to its 
north and west; I observed in these latter, nothing but 
large beds of glutinous clay, which I consider as a vol- 
canic sub-marine production; plains of moving sand, 
entirely composed of a flinty dust of quartz and felds- 
path mixed with finely pulverized shells; and banks of 
a recent chalky marl, evidently formed by the glutina- 
tion of the sand, or pulverized animal substances. 
I have in fact not found in these deserts any complete 
remains of sea-animals; but my situation prevented me 
from making assiduous researches, and it is possible 
that such remains, if they exist, are only to be found 
at a great distance to the south or west of the Atlas, as 
the fury of the waves pulverizes every thing that in these 
parts is raised to the surface of the sea. 
The shock of the waves is so great, that even at 
times of the most perfect calm, and without any pre- 
