112 NORTH-WEST AFRICA. River," a name it no longer merits, is not a tributary of the Bahira ; it descends from the Zaghwan mountains, and its volume, increased by the water at present collected by the aqueduct of Tunis, flows southwards round a low eminence which bounds the lacustrine depression. The Tunisian Sebkhas. On the eastern shore of Tunis, the coast is skirted by numerous sebkhas, which are separated from the Mediterranean by strips of sand. But at some distance inland, depressions are also found into which fall several rivulets, whose waters quickly run dry in their saline clay beds. Such are the sebkhas which follow in succession west, south-west, and south of Susa, and which are alternately vast sheets of water and saline plains. During winter time Kairwan has often been completely cut off from the rest of Tunis by these quagmires. At the very com- mencement of the rainy season a large portion of the country is transformed into a veritable slough, leaving no other route available to the caravans except the ridges running between the hollows. The most extensive lagoon is the Sidi-el-Hani sebkha, or Lake of Kairwan, wliose surface at the period of the floods is at least 200 square miles in extent, and whose central depression, in rainy years, always retains a little water. It is completely cut off from the coastlands by the Sahel hills, whilst Lake Kclbia, not so extensive but always filled with water and even bearing boats, occasionally discharges its surplus waters into the lowlands over a ledge some 60 feet high. When the rainfall is very abundant — that is to say, on an average every eight years — the emissary called the Wed Menfes attains a coast- land lagoon, the sebkha of Jcriba, which is connected with the sea by the Ilalk-el- Mengol. Travelling at this part of the coast is rendered dangerous on account of the looseness of the soil, and till recently, before the construction of the causeway, not a winter passed without the caravans losing some of their men or animals. According to M^L de Campou and Rouire, Lake Kelbia, whose surface varies from 20,000 to 32,000 acres according to the season, forms the basin of a fluvial system as vast and even more important than that of the Mejerda, The Wed Bagla, which flows into this basin together with its tributaries the Wed Fekka, the Marguelil, and other rivers flowing »from the heights^f Central Tunis, appears on the maps recently drawn up to have a far less extensive area of drainage than the northern rivers. In several essays M. Rouire has also attempted to prove the identity of the Bagla with the river Triton of the ancient writers. But how is it po.ssible to identify with certainty a river which, according to Pliny, forms the source of the Nile, and one of whose branches is lost in the Niger ? And the lake of the same name which M. Rouire identifies with Lake Kelbia, may in fact have been that mysterious basin which different writers have sought in various places along the southern shores of the Mediterranean, Strabo placing it at Berenice, to the west of the Great Syrtis, whilst Diodorus seeks it in the vicinity of the " ocean which surrounds the world." It would assuredly be a hopeless task to endeavour to reconcile all the assertions that ancient writers have made about the
