Vol. 65.] NEIGHBOURHOOD OF THE VICTORIA FALLS. 397 



mouth, and it is broad for several miles upwards. The red sand 

 lies on the upper slopes of both valleys, and extends down to the 

 alluvial flat bordering both the Zambesi and the Maramba. The 

 latter river flows down the north side of the tributary valley, and 

 enters the Zambesi 2 miles above the Falls. It rises in what are 

 described as ' very rough stony hills,' 25 or 30 miles from its 

 mouth, and 500 or 600 feet above it. From May till October it is 

 almost dry to within a mile of the Zambesi : the bed of the river, 

 15 or 20 feet below the sides, holding pools of water connected by 

 rapids a few inches deep. The Nansunzu, a stream of the same 

 character, joins it about 3 miles from the Zambesi. Many sections 

 are to be seen in the banks of the Maramba showing a coarse 

 gravel containing quartzite from the neighbouring basaltic kopjes, 

 fragments of basalt from cliffs bordering on the river, and 

 sandstone and quartzite more or less waterworn, probably derived 

 from the country near the source of the river. A coarse con- 

 glomerate of rolled stones and pebbles in sand is seen sometimes 

 resting directly upon the basalt and covered by silty sand, but more 

 often overlying grey or yellowish sand at 6 or 7 feet above the 

 bed of the river, and covered by 6 or 8 feet of loamy sand and 

 soil coloured by red sand from the slopes of the valley. The 

 section figured by Mr. Lamplugh, 1 in which he found, in the gravel 

 at 5 feet from the top of the bank, an implement with the chipping 

 freshly preserved, is no longer to be recognized ; but close to it I 

 took from the base of the gravel a flake but little waterworn, of a 

 whitish quartzite, of which there are many pieces in the gravel. 

 It lay 7 feet above low-water level and 5 feet below the top of the 

 bank. An extraordinary flood in January 1908, 2 the height of 



1 Journ. Anthrop. Inst. vol. xxxvi (1906) p. 164. 



2 This flood occurred on January 29th, 1908. It rose suddenly and fell 

 quickly, being due to heavy rainfall in the hills near the sources of the river. 

 It seems to have been exceptionally high, as native huts standing well above 

 ordinary floods were swamped. In July and August flood-drift remained on 

 trees and bushes along the riverside above Livingstone, 4 and 5 miles from its 

 mouth, as much as 20 feet above the low water. An exact measurement of 

 the height of the flood was possible where the railway crosses the river a mile 

 and a quarter from the Zambesi by a bridge with a 100-foot opening. The 

 flood reached the underside of the girders, which measured 18 feet 7 inches 

 above the water-level in July, and is 2923"7 feet above sea-level. The Zambesi 

 was then low. and its level at the mouth of the Maramba was not more than 

 2903 feet above sea-level. The Maramba flood must have fallen to that, so 

 that there was a fall of 20 feet in about 2000 yards. Flood-drift remaining 

 7 or 8 feet above the July water-level more than half way down to the Zambesi 

 corroborated this estimate. Other traces of the flood remained in newly 

 exposed surfaces of the gravel-banks higher up the river, and in .sharp notches 

 4 or 5 feet deep cut in the sandy edge of the banks by the water as it poured 

 off the flooded valley. In the bed of the river shoals of waterworn stones 

 3 and 4 inches and more in diameter remained, sometimes overlain by deposits 

 of smaller gravel and sand several feet thick, presenting sections like those in 

 the banks. 



High torrential floods are said to occur every year in the Maramba, and to 

 be followed in a few days by heavy rainfall at Livingstone. On January 31st, 

 1908, two days after the flood, 2 inches of rain fell, and on February 2nd & 

 3rd 9 inches of rain were registered at Livingstone. 



Q. J. G. S. No. 259. 2 n* 



