C. CROSSLAND—PHYSICAL D®SCRIPTION OF KHOR DONGONAB. 279 
been wetted by the sea. In the morning the dew drips from the roof like 
rain. The winter is generally dry and pleasant, bul. there are intervals of 
strong damp south-east winds, which carry salt moisture over everything near 
the sea, and other manifestations of dampness are as unpleasant as described 
for the late summer. 
Cementation of the Sand.—Now for the effect of these aiternations of 
extremes upon the sand of the plain. The new drift is of a pleasant light 
yellow colour, dry and loose like the sand above high-tide mark of an 
English beach. Ina few days it becomes darker coloured at the base, and 
finally this tinge spreads over the whole except perhaps the uppermost part, 
and the sand grains begin to cohere slightly. During a spell of damp weather 
the drift has drawn up moisture by capillary action from the damp plain, and 
with it salt *. Dry weather following, this water is evaporated, to be replaced 
by more drawn up from the plain (ultimately from the sea), so that in time a 
considerable quantity of salt is mixed with the sand. Rain removes the more 
soluble, leaving the less soluble constituents of sea-salt as a cement, which, in 
the case of the coarser deposits, converts loose sand into stone. The level of 
the plain is thus that at which capillary action can deposit enough binding 
material to counteract the abrasion of the wind and the sand it carries. Near 
the sea the level is temporarily raised a little above this by the dunes 
collected by Suaeda, but within this band, the vegetation being killed off by 
the consolidation of the sand, the hillocks are swept away and the level reduced 
so that the hard almost: perfectly level surface above described is produced. 
After every storm sand is deposited on the slope of the beach, thus extending 
the plain a minute distance seawards. This process would go on if no 
vegetation were present, but the Suaeda assists greatly by the increase of 
height it gives to the slope and by holding the sand together while the 
cementing process is begun. 
The longest axes of these hillocks are, as has been noticed in other sand 
formations, not in the direction of the prevailing winds, fairly strong though 
these are, but in the direction of the sand-bearing gales, 7. e., not N.H. to S. WW, 
but N.N.W. or W. to S.S.E. or E. 
Coral Reefs—In places patches of coral reef appear in the sand, but as 
these project only a few inches above the surface it is evident that they were 
cut down to sea-level after the first main movement of elevation in these 
reefs, attaining their present position by one of the minor ones which followed 
it, as described in my paper of 1907. They were then surrounded by the sand 
before the sea totally removed them. The neighbouring reefs which bound 
the lagoons seawards are almost certainly of the same age, but though areas 
* Fine salt spray is blown on to the plain during high winds, but the darkening of drifts 
from the base upwards is evidence of the importance of capillary action. 
