A. F. Verrill — Comparisons of Coral Faunw. 185 



The interesting relations between the West Indian and the Brazil- 

 ian coral fauna? will be discussed in the next chapter. 



At present, it is impossible to make a satisfactory general cata- 

 logue of the West Indian corals. This is due mainly to the large 

 number of forms badly determined and too imperfectly described by 

 Duch. and Michelotti, especially those of the genera Mussa (Iso- 

 phyllia), Agaricia, Porites, etc. All of these and others need careful 

 revision, with large collections and comparisons of types. 



3.— Characteristics of the Brazilian Coral Fauna,. 



The known Brazilian reef corals are but few in number, though 

 the reefs have now been well explored, but they are of unusual inter- 

 est, partly because they constitute a very special coral fauna, and 

 partly because several of them present remarkably generalized or 

 archaic characters, combining within a single species characters 

 which ordinarily characterize two or more distinct genera. 



This would seem to indicate that this fauna is a small surviving 

 remnant of an ancient coral fauna that has mostly disappeared. 

 Possibly this fauna may date back to the early Tertiary period. 

 Certain of the still existing species may have been the ancestral 

 species from which some of the modern West Indian reef-corals may 

 have been derived, by evolution, under more tropical conditions. 



Among those of special interest in this way, I may mention, espe- 

 cially, Mussa, Braziliensis and M. tenuisepta, both of which present 

 characters of Mussa (group Symphyllia), Isophyllia, and Favia, so 

 that they might be placed about equally well in either of these three 

 groups, while they also closely resemble Acanthastrcea. 



Mussa Harttii occurs in the form of a typical Mussa, with free 

 dichotomous branches, and in the form of a Symphyllia, with abun- 

 dant exotheca, uniting the corallites completely together. But all 

 sorts of intermediate states also occur. Thus it serves to compel the 

 union of these genera. 



Mmandra conferta Ver., originally referred by me to Favia, is 

 almost exactly intermediate between the two genera in all its charac- 

 ters, and therefore shows their close relations and common origin. 



It is specifically allied to F gravida, which lives with it, and also 

 to F. fragum and Mmandra Agassizii, as well as M. clivosa, of the 

 West Indies. Indeed these five species form- a nearly continuous 

 series of closely related forms, ranging from typical Favia to typical 

 Mmandra. 



