936 The American Naturalist. [October, 
In such situations the primary condition of the reefs would be that of 
barrier reefs (though differing from most barrier reefs in notsloping off 
rapidly into very deep water). The reefs of Florida are probably ex- 
amples of this sort. Moreover, it was long since recognized by Cham- 
isso, that, if the summit of a submarine volcano (or a shoal of any 
origin whatever) should be within the hundred-foot limit, the coral 
formations thereon would naturally assume the atoll form—a ring of 
reef encircling a lagoon—as the result simply of the more luxuriant 
growth of the corals on the outside than in the middle of the reef 
grounds. Murray, in his theory of atolls, in addition to the more 
luxuriant growth of corals on the outside, calls in the solvent action 
of the sea-water in the middle of the reef grounds in explanation of 
the lagoon,—an action whose importance he seems greatly to overrate. 
A strong objection to Chamisso’s theory of the origin of atolls has 
been found in the amount of coincidence which it requires. All atolls 
must represent submarine volcanoes (since no other shoals are likely to 
occur in mid-ocean), rising to within a hundred feet of the sea-level. 
The occurrence of so many independent volcanoes attaining so nearly 
the same altitude appears improbable. Murray meets this argument 
with the suggestion that forces are in action which tend to reduce all 
oceanic volcanoes to a uniform altitude, since rain, streams, and ocean 
waves tend to degrade to sea-level all peaks that rise above that level, 
while deposits of skeletons of pelagic life tend to raise the level of 
shoals which are yet too deep for coral growth. An island of any height 
could of course be leveled in time by the combined effects of subzrial 
and marine denudation ; and, as it is not unlikely that volcanic islands 
may have been formed in various geological periods, it is conceivable 
that scores or even hundreds of oceanic volcanoes, originally of var- 
ious heights, might all exist to-day in the condition of shoals. But it 
is certain that wave-action could never degrade an island much below 
the water-level. The formation of atolls with deep lagoons, on the 
basis of volcanic cones truncated by wave-action, would seem impossi- 
ble without subsidence. The depth of the lagoons, which is generally 
considerable in the larger atolls, and which sometimes amounts to more 
than three times the extreme depth of coral growth, seems, in fact, a 
conclusive argument for the subsidence theory. In commenting on the 
theory of Murray, which proposes to account for the foundation ot 
coral islands by the accumulation of the remains of pelagic life, and 
to account for deep lagoons by the solvent action of the water, Dar- 
win is reported to have said to a friend, not long before his death, that 
he could not understand how the water could possess so great solvent 
