MATTHAI— RECENT COLONIAL ASTR^ID^ 23 



appreciated in the case of forms like Edwardsia, where the longitudinal muscles of all the 

 non-directive mesenteries are turned to one pole, but in adult coral polyps, with the 

 hexameral arrangement of the mesenterial couples and with no appreciable difference 

 in the relative size of the two grooves, it is impossible to distinguish one pole from 

 the other. Parker (110, p- 268) pointed out the same difficulty in the " diglyphic " forms 

 of Metridium. In my Group II, in which neither directive mesenteries nor directive 

 grooves are present in the adult condition, the distinction breaks down. Reliable criteria, 

 if there be any for determining one pole from the other, can be expected only from 

 a comparative study of the developmental sequence of the mesenteries. Van Beneden, 

 homologising a cerianthid larva with the segmented larva of Amphioxus or the embryo of 

 Peri'patus, substituted anterior for ventral, posterior for dorsal, the oral-disc being 

 regarded as dorsal or "face neurale." Faurot, after his exhaustive embryological studies 

 of Hexactinians, prefers to use ventral and dorsal instead of sulcus and sulculus. 



In my Group II, in which directive mesenteries and bilateral symmetry are wanting*, 

 no two grooves are distinguishable from the others, but in many polyps the grooves appear 

 (in transverse section) on the whole deeper than in Group I, owing to the folding of the 

 stomodseal wall into the inter-mesenteric chambers. G. Y. and A. F. Dixon (27 — 30), 

 Carlgren (20), McMurrich (98, 99) and Parker (110) have observed variations in the 

 number of directive grooves above and below their normal number (also total absence) 

 within the same species of certain genera, viz. Sagartia, Bunodes, Metridium, and have 

 correlated it with a corresponding variation in the number of directive couples of 

 mesenteries. Parker and McMurrich studied this variation more thoroughly in the case 

 of Metridium marginatum and Sagartia spongicola, emphasising the numerical corre- 

 spondence between directive couples and grooves. Perhaps a similar relation subsists 

 in coral polyps also, as will be seen by a comparison of Group I with Group II. 



The stomodseal ectoderm has a faintly striated border below the row of cilia. Its 

 nuclei are round, rod and spindle-shaped, and more or less homogeneously stained dark ; 

 they are massed together in each ridge below the striated border, around a deeper, some- 

 what granular region, which is almost devoid of nuclei. The nature and arrangement of 

 the nuclei give this ectoderm a characteristic appearance. Through the central proto- 

 plasmic regiion, slender filaments generally pass backwards into the mesoglsea. Nematocysts 

 are never numerous ; when present, they may belong to any of the three types. Small 

 vacuoles, as described on p. 7, are usually present in the peripheral protoplasm above 

 the layer of nuclei. 



The endoderm is, as a rule, extremely thin, and algse are comparatively rare. The 

 mesogleea may or may not be thickened at the attachments of the mesenteries ; when 

 thickened, it projects into the ridges. 



Mesenteries. In coral literature there has been a great confusion in the terminology 

 employed to denote the soft radial partitions of the polyps and the calcareous partitions of 

 the corallites, " septa " having been used for both the structures by some English and 

 German authors and " cloisons " by French authors. The name "mesenteries" ("les 

 mes^nteroides " of Lacaze Duthiers) is now well established for the radial partitions of the 



* A trace of bilateral symmetry is however left in Group II in the lateral compression of the stomodseum. 



