C. CROSSLAND—-PHYSICAL DESCRIPTION OF KHOR DONGONAB. Piel 
Company, now abandoned, rising white and imposing, strangely alone in the 
midst of this yellow desert. Between us and them is the salt lake, acres of 
wet salt, like a frozen sea, which was their raison d’étre, yet another of these 
fault depressions, in this case completely shut away from the sea. (Pl. 82. 
fig. 2.) 
Formation of the Barrier Reefs.—As noted in my former paper, the barrier 
reefs are a series of large areas covered with an intricate network of reefs 
and shoals with comparatively deep lagoons between. We see from our 
examination of Rawaya how many of these deeper passages have arisen. 
Imagine Rawaya at a level say ten feet lower. The greater part of the 
land remaining above water would then be quickly cut down to form reef 
flats, the two hills remaining as islands, the canals as lagoons and passages 
between the reefs. Another series of lagoons would be formed by the 
solution and degradation of the reef flats, the outer edges of which are 
protected by growing corals and remain as long and narrow surface reefs, 
generally crescentic or incompletely circular in form. This is obviously the 
origin of the reef maze lying between Rawaya and Makawar, and, by 
inference, of the rest of the barrier system. (See fig. 1, p. 266.) 
In my previous paper it was stated that there were no hills at Salak 
(Maps 1 and 3) to make a complete correspondence between Ras Salak and 
the barrier reefs extending southwards from it, and Rawaya and its reefs. 
Since then I have had opportunities of seeing more of that neighbourhood, 
and I find that there is a line of raised coral running along the coast for 
about ten miles, which, though only ten to twenty feet above sea-level, is 
very distinctly marked off by a broad flat salt marsh separating it from other 
raised coral westwards. This latter is continuous with the shore reefs about 
the harbour of Little Salak, so that we have, on land, elevated continuations 
of both systems of reef, separated by a wide valley of fine alluvium formed 
by the filling in of the lagoon between the reefs. The sketch-map (PI. 30) 
makes the arrangement clearer. 
The shore of this salt marsh is obviously extending seawards, probably 
under the influence of sandstorms as much as, or more than, of rain floods. 
The correspondence of Ras Salak with Ras Rawaya and of the salt marsh 
with Khor Dongonab is thus complete, the former being on a smaller scale. 
The origin of the Ship Channel.—The size of this channel between the 
fringing and barrier reefs is very considerable, its average width being 
from 2 to 4 miles and depth from 15 to 200 fathoms. (Map 1, Pl. 28.) 
There is nothing to hinder large steamers travelling up it from Suakin 
to Mohamed Gul, and of course that section lying between Port Sudan and 
Suakin is so used. The traveller by native boats of 20 to 50 tons sees the 
