512 Proceedings of the Royal Society 
secreted by the animals which can still live in it; the lagoon thus 
becomes widened and deepened.* 
On the other hand a vigorous growth and secretion of lime takes 
place on the outer margins of the reef ; and when the water outside 
becomes too deep for reef-forming corals to live, these still build 
seawards on a talus made up of their own debris : — the whole 
atoll expands somewhat after the manner of a Fairy Eing. 
It is not necessary to call in disseverment of large atolls in order 
to explain the appearances presented in the Great M'aldiva group 
of atolls, f The coral fields rising from very many parts of these 
extensive submarine banks form atolls. The marginal atolls have 
from the first the advantage of a better supply of food. They 
elongate in the direction of the margin of the bank where the water 
is shallower than to seaward. Many of these marginal atolls have 
coalesced, and as this growth and coalescence have continued, a large 
part of the food-supply has been cut off from the small atolls situated 
towards the interior of the bank. Ultimately a large atoll like 
Suadiva atoll would be formed. The atolls in the interior would 
be perhaps wholly removed in solution, and the atoll-like character 
of small marginal but now coalesced atolls would be wholly or 
partially lost by the destruction of their inner sides.J A study 
of the charts shows all the stages in this mode of development. 
In the case of the Lakadivh, Caroline, and Chagos archipelagos we 
have submarine banks at various stages of growth towards the sur- 
face, some too deep for reef-forming species of coral, others with 
coral plantations, but all submerged several fathoms, and scattered 
amongst these some of the oldest and most completely-formed atolls 
and coral islands. It is most difficult to conceive how these sub- 
* Complete little Serpula-atolls, with lagoons from 3 to 50 feet in diameter, 
and formed in this way without subsidence, were numerous along the shores of 
Bermuda. 
t Mr Darwin’s application of his theory to this group — where the dissever- 
ment of large atolls is called in, and a destructive power attributed to oceanic 
currents, which it is very unlikely they can ever possess — has often been con- 
sidered unsatisfactory. 
1 “In speaking of Bow Island, Belcher mentions the fact that several of 
its points had undergone material change, or were no longer the same when 
visited after the lapse of fourteen years. These remarks refer particularly to 
islets situated within the lagoon. I could myself quote many instances of the 
same description,” — “Wilkes’ Exploring Expedition,” vol. iv. p. 271. 
