ACTINOZOA. 259 



called " atolls " which have been described as occurring in rocks of 

 Devonian or Carboniferous age. Not only is it theoretically im- 

 probable that an " atoll " of such high antiquity as the Devonian 

 should have so far escaped destruction by denudation that its 

 original form should still be recognisable, but these supposed atolls 

 are found in areas which can be proved to have undergone extensive 

 disturbance. In such cases, therefore — as, for example, in the 

 "atolls" described by M. Dupont in the Devonian limestones of 

 Belgium — it must be borne in mind that the observed ring-like 

 arrangement of the limestones may well be explained as the result 

 of denudation acting upon a series of strata which contain coralline 

 limestones amongst their members, and which have been folded in a 

 complex manner. 



Owing to their peculiar mode of formation, it is clear that those por- 

 tions of a coral-reef which are formed by corals growing in place must 

 always be terminated by more or less abrupt boundaries ; and, for the 

 reasons above given, it is doubtful if the existing outcrops of any of the 

 known Palaeozoic limestones can be regarded as actually corresponding 

 with the edges of old reefs. Still it may be permissible to assert that 

 coral-reefs existed in Palaeozoic time, provided the term " reef" be used 

 in a wide sense. There are, namely, Palaeozoic limestones which are 

 more or less largely composed of corals which grew i?i situ, and these 

 may be looked upon as constituting formations similar to existing coral- 

 reefs, though they are mostly of small thickness, and though their present 

 outcrops must be the result of denudation rather than of original ar- 

 rangement. On the other hand, very many of the coralline limestones of 

 the Palaeozoic period have been only partially, or not at all, produced by 

 corals growing in place, but are essentially of the nature of " reef-rock," 

 being mainly composed of small corals, or of fragments of the larger 

 forms, intermingled with the debris of innumerable calcareous organisms 

 of other kinds, such as Crinoids, Brachiopods, Foraminifera, &c. The 

 oldest known coralline limestones occur in the Ordovician period ; but 

 still more extensive coralline deposits are found in the Silurian. The 

 corals of these belong principally to the Rugose and Perforate divisions 

 of the Madreporaria, but there are also various forms (such as the 

 HeliolitidcB) belonging to the Alcyonaria. Moreover, these ancient 

 coralline limestones are largely, often preponderatingly, made up of the 

 remains of Stromatoporoids, which are referable to the Hydrozoa, or 

 of Monticuliporoids, the precise zoological place of which is not ab- 

 solutely certain. In the Devonian period we meet with very extensive 

 coralline limestones which, zoologically speaking, are closely allied to 

 those of the Silurian, and in the formation of which the Stromatoporoids 

 again play a very important part. The vast coralline limestones of the 

 Carboniferous period are mostly characterised by the occurrence of 

 Rugose corals {Lithostrotion, Lonsdaleia, Cyathophyllum, &c), along 

 with Perforate types, such as Syringopora; but the Stromatoporoids 

 appear to be now wholly wanting. No true coralline limestones have 

 hitherto been recognised in the Permian rocks ; but towards the close of 

 the Triassic period, true coral-reefs were largely developed in Western 

 Europe. Still more extensive reefs were formed during Jurassic times 

 in South-western and Western Europe and in Britain, and true reefs also 



