264 Br. F. A. Bather — Studies in Edrioasteroidea. 



accessory cover-plates. Were it not for these accessory plates (1914, 

 p. 165, text-fig. 3) the supposition that the supra-oral cover-plates in 

 that species were suturaily united to form a solid tegmen would be 

 inconsistent with the presence of pores all round the peristome. 



The Food-grooves of Edrioaster and many other Edrioasteroidea 

 are so distinct that they call to mind those structures in Cystidea 

 Bhombifera that used to be called "recumbent arms". Those, 

 however, are produced by the proliferation of circumoral plates over 

 the outside of the theca. The resemblance is really closer to the 

 epithecal grooves of the Diploporita. Their anatomical relations, and, 

 in E. levis, their connection with the interambulacrals, show that the 

 floor-plates are part of the thecal wall, not different in origin from 

 the other thecal plates. As in the Diploporita we see to have been 

 the case, so in Edrioasteroidea we may infer that the grooves passed 

 out over the theca from the angles of the mouth, and that as they 

 became deeper and more fixed, and as the cover-plates increased in 

 size, so the subjacent plates became regularized. That is one possible 

 hypothesis, and it seems to be the one held by Dr. Foerste (1914, 

 p. 425). Another hypothesis is that the grooves passed out from the 

 angles of the mouth between the thecal plates and not over them. 

 This view is suggested by the appearances in Stromatocystis, and 

 would have the advantage of making the floor-plates a double series 

 from the outset. The disadvantage is the great weakening of the 

 peristomial region which the hypothesis implies. 



In any case a stage was reached sooner or later in which the 

 ciliated grooves, with the underlying extensions of the nervous and 

 water-vascular systems, passed over the floor-plates and under the 

 cover-plates as in Diploporita. But from the Diploporita all 

 Edrioasteroidea differ in the absence of exothecal projections 

 (brachioles) bordering the grooves or forming the ends of their 

 branches. From this point of view alone the Edrioasteroidea present 

 an ambulacral structure divergent from that of all other Pelmatozoa, 

 and this is why they have certainly better claims than the Blastoidea, 

 for instance, to rank as a separate class. In those Edrioasteroidea 

 whose skeleton, like that of Edrioaster, bears witness to the 

 presence of ampullae, this distinction is enhanced, not changed in 

 character. 



The structure of the Peristome has been described for Edrioaster 

 buchianus (1900, p. 197, PI. X, Fig. 2) and E. bigsbyi (1914, 

 pp. 164-6, text-fig. 3, PL XIV, Fig. 2). It is clear that, as 

 E. Billings noted in 1854, the gullet into which the food-grooves led 

 passed into the stomach through a firmly-built ring or mouth-frame, 

 of which the inner adcentral border turned down so as to form a rim, 

 probably for the attachment of the stomach wall. As viewed from 

 the oral aspect, after removal of all cover-plates, the chief elements 

 of this frame appear to be five perradial plates, of which the exposed 

 portions have a triangular outline. Viewed from the inside of the 

 test, there are also seen to enter into this frame five interradial 

 elements. In all the interradii except the posterior, and in all 

 perradii, these elements appear to be formed of fused portions of 

 floor-plates. In the posterior interradius there also enters into the 



