84 MADREPORARIA. 



The presence or absence of tabulfe has probably no independent taxonomic value, as it 

 is apparently invariable in the massive types of growth in which the calicles have to 

 lengthen very rapidly. The same phenomenon is found in glomerate Turbinarians. 



THE VARIATIONS IN THE C(ENENCHYMA. 



Dana suggested that the genus Astrccopora might be divided into two groups — perhaps 

 even genera — one having the cce.nenchyma highly echinulate, and the other more porous and 

 granular. Milne-Edwards adopted this suggested division. But, as we have seen above, 

 these changes are but variations upon that structure of the ccenenchyma which is typical of 

 the genus and one of its most marked features. Explanate forms tend to have the horizontal 

 element more pronounced than the vertical costal pillars, resulting in the smoother porous 

 form of ccenenchyma, while massive forms tend to have the vertical element more pronounced 

 than the horizontal synapticular element, and are thus more likely to be echinulate. There 

 are, however, other variations in the texture of the ccenenchyma, which appear to have a 

 taxonomic value : the echinulae themselves vary considerably in form and appearance. The 

 texture may be very coarse or delicate, loose or compact. Again, the typical structure may 

 be obscured, the ccenenchyma being simply reticular, or solid, or granular. 



ON THE CLASSIFICATION OP THE SPECIMENS BASED 

 UPON THESE VARIATIONS. 



On the general subject of dividing the members composing a genus of corals into species, 

 I refer the reader to what I have said under the similar heading in the Introduction to the 

 Catalogue of Turbinarians.* In the present case, we have only a small number of specimens 

 to deal with, and when those which are obviously specifically identical, coming from the same 

 locality, have been separated, the small number remaining can easily be described singly as 

 new types, a process which is hardly practicable in a large collection in which the variations 

 shade off into one another by an infinite number of gradations. 



As in my classification of the Turbinarians, I have again selected the methods of growth 

 as the leading feature in dividing the specimens. It is, however, not easy to ascertain, in all 

 cases, to which type of growth certain specimens belong. In such cases, I have mentioned 

 this doubt. Beyond this difficulty, the smallness of the Collection allows of its being divided 

 into separate types, without any sense of excessive multiplication of new species. 



With regard to the existing types, the usual difficulty in identifying them with actual 

 specimens was, in this case, reduced to a minimum. There were only five recorded species. 

 This number was reduced to four, as soon as I convinced myself that A. jndvinaria and 



* Vide su^a, p. 19. 



