32 



TABULATE CORALS. 



the actual visceral cavities is of endothecal origin ; and I have 

 seen nothing of the nature of a true " coenenchyma" in any of 

 the FavositidcE. 



Wherever the walls of the corallites in the Favositidce come 

 in contact with one another, a communication between the vis- 

 ceral chambers of the polypes is effected by means of the 

 "mural pores." These are, typically, very definite apertures, 

 usually arranged in some serial order ; and for the most part 

 so small and so widely apart, that they do not appreciably inter- 

 fere with the compactness and integrity of the wall as a distinct 

 structure. In the Silurian genus Cohunnopora, Nich., the pores 

 are so numerous and close-set that, though regular in form and 

 arrangement, they give the wall a cribriform aspect ; in the 

 Cretaceous Koiimckia, E. and H., the pores are equally numer- 

 ous, but apparently irregular ; while in the Devonian Arceopora, 

 Nich. and Eth. jun., the trabecular condition of the wall is 

 quite like that of the typical Perforate Corals. In these 

 forms, therefore, we have a more or less complete approxi- 

 mation to the condition of parts in Alvcopora (fig. 13), a recent 

 Madreporarian, in which tabulae are developed. It is hardly 



Fig- '3- — A, Vertical section n{ Aheopora sf>on^iosa, Dana, showing the porous walls and the 

 tabulre and (b) two calices of the same, enlarged. Recent. (After Dana.) 



necessary to add that in those members of the Favositidce in 

 which the corallites are partially disjunct, the mural pores 

 only exist in places where the tubes are In actual contact. 



The condition of the septa varies extremely in the Favo- 

 sitidce, though it is only in some few forms that these struc- 



