46 GILBERT 0. BOURNE. 



(1) to any marked tendency to departure from a radial sym- 

 metry or from the normal Actinian symmetry as shown in the 

 arrangement of the mesenteries ; (2) To the presence or 

 absence of a "Randplatte." I am not disposed to attach any 

 importance to exocoelic or entocoelic septa for classificatory 

 purposes, since it appears that in one and the same colony 

 entocoelic septa only, or both exocoelic and entocoelic may be 

 found (Madrepora). 



With regard to the symmetry. Owing to the presence of 

 "directives/' Actinia and many Madreporaria show a well- 

 marked bilateral symmetry, more marked in the case of many 

 Madreporaria by an irregularity in the arrangement of the 

 septa laterally, whereas the regularity is maintained at the 

 ends of the chief axis. But in Mussa, Euphyllia, and Lopho- 

 helia, the radial symmetry is perfect ; and this is probably a 

 more primitive condition of things. In Madrepora aspera 

 there are signs of increased differentiation along the long axis, 

 since there are no filaments on mesenteries 3, 5, 8, 10 (Fowler's 

 numeration), and in M. Durvillei there is a strongly marked 

 bilateral symmetry. This symmetry is even more marked in 

 Seriatopora and Pocillopora, and it can scarcely be doubted 

 that the latter forms are allied to the Madreporinse, although 

 the one family is aporose the other perforate. 



Recent researches have made it doubtful whether any sharp 

 distinction can be drawn between the two groups. Von Koch 

 has shown that the septa are formed from the basal ectoderm, 

 and that the theca is formed by fusion of the peripheral ends 

 of the septa in Astroides, a perforate Madreporarian. A study 

 of the adult anatomy of Astrsea, Mussa, Lophohelia, Eu- 

 phyllia, Fungia, and others demand a similar explanation. 

 Fowler has shown that a well-developed " Randplatte " with 

 extrathecal mesenteric elements exists in Rhodopsammia, a 

 perforate. Where a ccenenchyme with its correlative coeno- 

 sarc is present the extrathecal parts of mesenteries are not 

 present. The foregoing part of this discussion has shown that 

 a common coenosarc is due to nothing more than a persistent 

 connection between the " Randplatten " of adjacent polyps 



