130 ME. H. M. BERNAED ON EECENT POEITIDjE. 



This brief liistorical sketcli shows that almot't all who have 

 studied the Madreporaria liave come to the conclusion that 

 JPorites is in some way related to the Madreporids. The reasons 

 for this conclusion may be briefly arranged under the following 

 heads : — 



1. The general similarity of their polyps, with twelve tentacles 



in a single ring. 



2. The fact that the septa are mostly in some low multiple of six. 



In Porites there are almost invariably twelve. 



3. The skeletal walls are porous, and in both tend to form reti- 



cular ccenenchyraas *. 



Although the real value of this last point, viz., the possession 

 of porous or reticular walls, has never till recently been under- 

 stood, it was nevertheless a common character in striking contratt 

 with the solid mural structures found in the Astraeidse. 



As opposed to these ccmmon characters uniting the two 

 families, we have certain difl'erences keepiug them apart. These 

 were described by Dana, who limited the i'amily to two genera, 

 Porites ?mdi Goniopora, as follows: — (1) Extraoidinary porosity 

 of the Poritid skeleton as compared with the ruore regularly 

 lamellate skeleton of the Madreporidse ; (2) the fact that, in the 

 Poritidse, the skeleton in its relation to the polyp is purely basal 

 and never rises to a deep cup ; (3) that, as it grows, the small 

 central depression of the calicle fills up, so that the stars are 

 hardly or not at all traceable thrcugh the substance of the 

 corallum, as they alw^ays are in the Madreporidse. 



On the other hand, Milne-Edwards and Haime, carried along 

 by their theory of the origin of the Madreporarian skeleton, 

 believed that the " trabecular '' character of the septa in Porites 

 was the fundamental distinction, the septa in the Madreporidse 

 being lamellate. The remaining differences above quoted from 

 Dana were no longer applicable, because several other genera 

 with the so-called *' trabecular " septa were now included. 



As it is useless to attempt to discuss these resemblances and 

 differences until we understand clearly what is meant by the 

 terms used, it is necessary to sketch the fundamental theory 



* Milne-Edwards and Haime also added that the families agreed in having 

 no tabula;, which are so ccmmon in the ^istrseids. This distinction is incorrect. 

 I have already described tabulse in Astrieopora and Titrbinaria, and find them 

 also in Gonio^pora and specially numerous in Porites. 



