TRANSACTIONS OF THE SECTIONS. 103 



and with mountain-ranges running for the most part in parallel lines north-east and 

 south-west. The southern slopes of the Atlas chain rise from a depression which in 

 several parts, especially to the south of Tunis, is many feet below the level of the 

 Mediterranean. From this depression the Sahara is for the most part a system of end- 

 less terraces, some of which are only a few miles apart, while others are expanded into 

 plains of from 50 to 100 miles in width, and which, so far as my observations and the 

 information I could gather from native caravans and a trustworthy guide extended, in 

 an unbroken series to within three days' journey of Timbuctoo, when the traveller will 

 probably find himself on the northern watershed of the valley of the Niger. 



As we advance, on every stage is written the record of the retiring ocean, which 

 gradually, by the elevation of its southern shores, was driven back and back to the 

 northward, till the last long inlet from the Gulf of Cabes to Tuggurt was drained and 

 evaporated, leaving its traces in the salt plains, and occasional moisture of the Wed 

 R'hir and Chott el Melah — the ancient Lake Tritonis. 



There are several singular exceptions to the course of the mountain-ranges above 

 mentioned, which are generally the local causes of the oases. 



Thus at Laghouat we find several elliptical basins of diminishing size piled one on 

 another. The lowest and largest rests on the flat surface of the secondary rock, which 

 is the base of the whole system. Several great fissures, which pervade all these super- 

 imposed basins, allow the water to percolate. It then rests on the impermeable rock, 

 draining through a very thin stratum of gravel or sand into any depressions, whence 

 it is raised by artesian wells, and creates an oasis. 



From the Sebaa Rous to Laghouat, all these ranges appear to belong to the lower 

 chalk formation. Limestone predominates, and forms the ridges of the Sahari, Senalba, 

 and Djellal mountains. It is of saccharoid structure, and of a variable colour, generally 

 greyish white. In many of the plains there is sandstone, sometimes hard, and at other 

 times so soft as to yield to the pressure of the fingers. This sandstone encloses nodules 

 of flint of various colours and semi-transparent. By disaggregation these become de- 

 tached from the softer medium in which they were imbedded. As the wind sweeps 

 the sand, they form shingly beaches of pebbles, many of them of a pretty chalcedony, 

 which is exported in some quantity to Paris. 



The upper deposit of limestone is marked by regular beds of gypsum of vast extent, 

 which are found in every district of the Sahara, but never in the secondary formation 

 of the Atlas region. 



South of Laghouat, the furthest French outpost, we came upon a shallow alluvial 

 deposit of the very latest tertiary and diluvian formation. Near the mountains this is 

 often composed of rolled pebbles in a limestone matrix. On the plains it is a white 

 calcareous rock, a sort of crust, very hard at the surface, but soft and friable below, 

 where it is mixed with green or grey clay, and encloses many crystals of gypsum. 



The diluvian formation may be traced more or less distinctly, I believe, between 

 all the ranges, even as far north as the Zahrez, near Djelfa. 



I was particular] v struck by the fact that several of my fossil shells from these super- 

 ficial deposits proved specifically identical with freshwater tertiary fossils from the 

 region of the Black Sea. May not further research perhaps reveal, that at no very 

 distant geologic epoch a vast chain of freshwater lakes, similar to those of North 

 America at the present day, extended from the plateaux of the western Sahara as far 

 as the neighbourhood of the Caspian 1 



The basin of the M'zab country further still to the south supplied me only with a 

 few fossils, apparently miocene. 



In turning from the M'zab southwards to Waregla, and thence north-east towards 

 Tuggurt and the Gulf of Cabes, the geological system appears to be the same, but with 

 fewer distinct little basins, and with more extensive diluvial deposits. 



As far as we could trace them, the basins are generally horizontal up to Biskra in 

 the north, and Gufza in the east, or very slightly inclined, consisting of alternating 

 beds of greensand (?), gypsum, and clay. These beds extend almost without inter- 

 ruption, or with very slight depressions, from latitude thirty-one degrees north to 

 thirty-five degrees north, and from longitude five degrees east to nine degrees east. 



The most interesting portion of this district is the Wed R'hir, a long line of depres- 

 sion sloping from the Touareg desert, latitude thirty degrees north, and longitude five 

 degrees east (circiter), with its surface occasionally moistened by salt lakes, but with- 



