974 MARINE BIOLOGY OF THE SUDANESE RED SEA. 
The very wide distribution of the deposits among these hills, even as far 
north as Kossér (lat. 26° N.), postulates an agent the range of which was 
coterminous with the ancient Red Sea, such as, for instance, the evaporation 
of its salt water. The Bab el Mandeb Straits, between the Red Sea and 
Indian Ocean, are not only very narrow but show signs of comparatively 
recent volcanic action, so it is not at all difficult to imagine their becoming 
blocked and again opened, possibly more than once, before those final move- 
ments which we have considered. If we are right in attributing the faulting 
up of the barrier reefs to the actual opening of the Rift Valley, it is probable 
that at the time the gypsum was deposited the area was occupied by a much 
shallower sea, in which the deposition of gypsum by evaporation would 
readily take place. 
The Rawaya Salt Field——I am the more inclined to this theory by my 
acquaintance with the Rawaya Salt Field, where a bed of gypsum is in actual 
process of formation. This lake of salt (Pl. 32. fig. 2) is a fault-depression 
shut off from the sea except by subterranean leakage on the south side. 
During the winter sea-water enters this way and so renews the supply, but 
in summer the sea-level is slightly lower and scorching winds remove this, 
leaving the surface firm, dry salt. Beneath the salt, ata depth of two or three 
feet, the deposit changes into a mass of loose, white crystals of gypsum, the 
formation of which is evidently recent and still continuing, a fresh supply 
being deposited from the sea-water added during each winter. How deep 
the gypsum may extend I do not know, but if this depression is as deep as the 
adjacent ones, there may be thirty to sixty feet of recent gypsum to the three 
feet of salt. 
DoNnNGoNAB PLAIN. 
The west shore of Dongonab Bay from the point marked by an arrow (on 
Map 2, Pl. 29) to its head, a distance of about 10 miles, is favoured over the 
whole Sudan coast of the Red Sea in that it is formed of sand beaches sloping 
quickly into water two to three fathoms deep. Atabout half a mile from the 
shore lie coral-reefs and shoals, which will be described later, enclosing a 
series of harbours, shallow as stated, but unobstructed by coral, quite ideal 
for the small craft which alone navigate this coast. Dongonab Harbour is 
that one of the series which has but moderately complicated entrances, one of 
the best beaches, and is nearest to the well. 
The shore itself is extremely low, about three feet above mean water-level, 
in winter often only one foot above the sea. For from a quarter to half a 
mile inland its surface is of sand alone, perfectly level but for hillocks of loose 
sand where vegetation occurs. Further inland is the gravel ridge mentioned 
on page 270, which runs northwards from Dongonab Hill and separates the 
valley in which are the wells from the sand plain. The mass of this ridge is 
