A vast depression, stretching from south-south-west to north-north-east and of an average breadth of some 30 miles, separated Erdi from Ennedi; it was the depression I heard spoken of earlier as a prolongation of that of the Bahr El Ghazal, through which Lake Chad once poured its waters into the lakes of Toro and Djourab, and consequently that by which the basins of the Chad and the Nile might in ancient times have entered into communication. That being so, I took the utmost care in examining the region and determining the altitudes. The lowest point was found about 30 kilometres from Dourdouro. Its altitude was 1750 feet, or 1000 feet higher than that of Bokalia at the north-eastern extremity of the Djourab. The slope was therefore from north-east to south-west, as was confirmed by the shape of the ground and the general direction of the valleys running into that depression, and I was able to conclude that if an ancient river once flowed in the bottom of that broad valley, which is hardly likely, it ran, not towards the Nile, but towards the lowlands of the Chad. By this evidence, one of the most important items of my geographical programme was fully elucidated: the basin of Lake Chad constitutes in the centre of Africa a closed basin which has never been connected with the basin of the Nile. The lake zone, now dried up, consisting of Kanem, the lowlands of Lake Chad, and Borkou, was once the outlet for the affluents of Lake Chad and for many great rivers coming down from the mountain mass of Ennedi, Erdi, and Tibesti. Its outline at successive periods—an outline in all probability very irregular—might be indicated by the hypsometric curves 270—260—250 metres, adopting for the Lake Chad of to-day the altitude of 240 metres. Its extent at that period must have been comparable with that of the Caspian Sea at the present day, and its greatest depth some hundred metres. In the evening of the second day’s march, when we were drawing near the foothills of Ennedi, we had not yet found any well, and our tiny store of water was used up. But spying in the west a notable gap in the line of hills, I thought we should be likely to find a water-point there, and profited by the coolness of the night to try to reach it. At dawn we came out on a fine river, dried up, where we got a little water by digging holes in the sand. By good luck our guide, Sougou, recognized that we had reached Oued Mourdi, where he had come by another route some six months earlier; thanks to which discovery, after a little search we were able to bivouac beside the well of Diona.