THE DESICCATION OF AFRICA.- 167 



stream was eight and a half times greater than that of the 

 inland one. 



The place where the Zambesi originally rose is now occu- 

 pied by the Kebrabasa Cateracts, which, have a fall of i6d 

 feet. On the right bank stands the mountain Zakavuma and 

 on the left, Morumba, some 3,000 feet high. The river is 

 jammed between these two mountains in a channel with per- 

 pendicular sides and less than 50 yards wide; the rocks are 

 water-worn and smooth, with huge pot-holes even 100 feet 

 above low-water. We can at once recognise the relationship 

 of this gorge to that below Stanley Pool on the Congo, and 

 the Falls of Bussa on the Niger. The swell of the land in 

 which the Victoria Falls are situated runs in a north-north- 

 westerly direction and is a secondary water-parting; at one 

 time the Zambesi cut back to as far as the Batoka Gorge 

 and all the water above the Falls was restrained within the. 

 Ngami depression, with an outlet to the Makarikari, and so 

 through the Kalahari to the Orange River. This was actually 

 the arrangement of the water system that the Portuguese 

 imagined to exist, before Livingstone ascended the Zambesi 

 from the mouth of the Falls in i860; it is only a coincidence, 

 based on faulty information, since even the early missionaries 

 at Zomba could scarcely have been there before the Zambesi 

 breached the ramparts of the Batoka basalts. 



The Makarikari depression to the east of Lake Ngami is 

 a vast plain with the sides rising gradually or in terraces, 

 like in the old Lake Bonneville in Utah. In this plain the 

 salt pans lie sunk a few feet beneath the general surface and 

 are occasionally filled with water. Formerly there was a lake, 

 Kumado, in the southern corner, corresponding to the present 

 Ngami in the bed of its former extension; in Chapman's time 

 (1852) it was still a marsh, in the centre of which the Chief 

 Chapo had his town- The fact that in both depressions the 

 water accumulated in the southern ends leads me to think 

 that the natural outlet was formerly here, and not in the 

 north or east. The Gwai Poort, "the land of a thousand 

 vleys." forms a long bay in the high walls that separate the 

 depression from the Zambesi valley, but it is now partially 

 choked up with sand. On the south-west the valley of the 

 Gwai Poort is continued by the Omuramba Epukira. These 

 omirimba are shallow, grass-covered valleys, practically with- 

 out gradient; they were once river-beds but they have become 

 choked by silting or the' accumulation of wind-borne sand and 

 after rain, instead of being filled with running water they 

 become a series of vleys. A still further desiccation and all 

 connection with a river system becomes lost and the depres- 

 sions become isolated pans. The physiographical features 

 of this part of the country render it probable that the Gwai 



