74 DR. H. B. GUPPY. 
through our misconceptions of its dimensions, we have been led to 
introduce a great cause to explain a very small effect. The slightly 
raised margins can be easily explained by causes dwelt upon by 
Murray, Agassiz, and others. No movement of the earth’s crust is 
necessary for this purpose. The mode of growth of corals, the 
action of the waves, and the influence of the currents, afford 
agencies quite sufficient to produce the slightly raised margins of 
an atoll. | 
The development of the islands of an atoll into horse-shoe or 
crescentic islands, as in the instance of Keeling Atoll, or into perfect 
small atolls or atollons, as in the Maldive Group, is a subsequent 
process to be shortly explained. These small atolls and horse-shoe 
islands only assume their characteristic forms after the island has 
been thrown up by the waves. Such was the conclusion I arrived at 
concerning small atolls and crescent-shaped coral islands in the 
Solomon Islands (Proc. Roy. Soc., Hdinburgh, 1885-86, p. 900); and, 
as just stated, I have formed the same opinion concerning the islands 
of Keeling Atoll. There is, in the first place, the island from which 
“lateral extensions grow out on either side so as to ultimately form 
a horse-shoe reef,” which itself under favourable conditions may 
develop into a small atoll. In the Solomon Islands I imperfectly 
grasped the method by which these changes in form are effected. 
In Keeling Atoll I saw the process in operation, and I arrived at the 
conclusion that whenever a coral island stems a constant surface- 
current, the sand produced by the breakers on the outer edge of the 
reef will mostly be deposited by the current on each side of the 
island in the form of two literal banks or extensions, giving the 
island ultimately a horse-shoe form, with the convexity presented 
against the current. The process may be aptly compared to the 
formation of a \/-shaped ridge of sand when a stake or some other 
obstacle is placed in a river bed. The stake represents the original 
small island thrown up by the waves. The \/-shaped ridge of sand 
represents the arms of the horse-shoe island which are subsequently 
formed. The back-wash or eddy may in the river-bed join the arms 
of the \/-shaped ridge of sand. Ina similar manner a horse-shoe 
island may have a bank thrown up across the mouth, and thus a 
small atoll is formed. Such is the process, imperfectly disclosed to 
me in the Solomon Islands, that I found illustrated in all its stages 
in Keeling Atoll. In the Keeling Islands, however, it was necessary 
to satisfy myself of the reality of the agencies chiefly concerned in 
this process. For instance, I had to ascertain how and to what 
