On the whole, the impression was favourable.  Old Shimbah, the Linguister at Porto da Lenha, and other natives had assured me that the Cataracts were taller than the tallest trees.  On the other hand, the plain and unadorned narrative of the “Expedition” had prepared me for a second-rate stream bubbling over a strong bed.  The river here sweeps round from the north-west, and bends with a sharp elbow first to the south-west and then to the south-east, the length of the latter reach being between four and five miles.  As far as the eye can see, the bed, which narrows from 900 to 400 and 500 yards, is broken by rocks and reefs.  A gate at the upper end pours over its lintel a clear but dwarf fall, perhaps two feet high.  The eastern staple rises at first sheer from the water’s edge to the estimated altitude of a thousand feet,—­this is the “Crocodile’s Head” which we saw on the last march, and already the thin rains are robing its rocky surface with tender green.  The strata are disposed at angles, varying from 35deg. to 45deg., and three streaks of bright trees denote Fiumaras about to be filled.  Opposite it is the “Quoin Hill,” bluff to the stream, and falling west with gradual incline.  The noise of this higher fall can hardly be heard at Nkulu, except on the stillest nights.

Below the upper gate, the bed, now narrowing to 300 yards, shows the great Yellala; the waters, after breaking into waves for a mile and a half above, rush down an inclined plane of some thirty feet in 300 yards, spuming, colliding and throwing up foam, which looks dingy white against the dull yellow-brown of the less disturbed channel—­the movement is that of waves dashing upon a pier.  The bed is broken by the Zunga chya Malemba, which some pronounced Sanga chya Malemba, an oval islet in mid-stream, whose greater diameter is disposed along the axis of the bed.  The north-western apex, raised about fifty feet above the present level of the waters, shows a little bay of pure sand, the detritus of its rocks, with a flood-mark fifteen feet high, whilst the opposite side bears a few wind-wrung trees.  The materials are gneiss and schist, banded with quartz—­Tuckey’s great masses of slate.  This is the “Terrapin” of the Nzadi.  The eastern fork, about 150 yards broad, is a mountain-torrent, coursing unobstructed down its sandy trough, and, viewed from an eminence, the waters of the mid-channel appear convex, a shallow section of a cylinder,—­it is a familiar shape well marked upon the St. Lawrence Rapids.  The western half is traversed by a reef, connecting the islets with the right bank.  During August, this branch was found almost dry; in mid-September, it was nearly full, and here the water breaks with the greatest violence.  The right bank is subtended for some hundred yards by blocks of granite and greenstone, pitted with large basins and pot-holes, delicately rounded, turned as with a lathe by the turbid waters.  The people declare that this greenstone contains copper, and Professor Smith found particles in his specimens.  The Portuguese agents, to whom the natives carefully submit everything curious, doubt the fact, as well as all reports of gold; yet there is no reason why the latter should not be found.