AND RELATED GENERA, SCOTTISH NATIONAL ANTARCTIC EXPEDITION. 65 



though this is not absolutely necessary. The arrangement of mesenteries in Minyas 

 may be explained quite simply through an enlargement of the endoccels. In any case 

 I consider the similarity in the Endoccelactis and Minyas arrangement as due to con- 

 vergence, a view which I am now able to further confirm, as Minyas, i.e. the species 

 described by me in 1895, and a closely related species, probably M. olivacea, later 

 examined by me, are stichodactyline Actiniaria, which are nearly allied to the family 

 Aurelianidse (the genera Aureliana and Actinoporus). 



With regard to the position of the family Endocoelactidse, I pointed out in 1897 

 that it must be placed fairly low in the system of Actiniaria, a view that has also been 

 taken up by M'Murrich. This is indicated not only by the absence of the sphincter 

 and the presence of spirocysts (thin-walled nematocysts) in the ectoderm of the body- 

 wall and oesophagus, but also by the absence of true differentiated basilar muscles. 

 Thus, the Endoccelactids must be Actiniaria, though they are not developed in the same 

 way as the elongated genera provided with physa (e.g. Edwardsidse, Halcampidse). 

 From a theoretical point of view we must assume the occurrence of forms which con- 

 stitute a link between the Protactininse and the Athenaria among the Actininse, i.e. 

 we must take for granted the occurrence of original Actiniaria, which by the retention 

 of the original body-shape with flat base (thus without development of a physa) have 

 lost the longitudinal muscles in the body-wall, but, on the other hand, have not yet 

 developed true basilar muscles; in the same way as I pointed out (1900, p. 57) that 

 the family Discosomidse forms a link between the Protostichodactylinse and Stichodo- 

 dactylinse. Among the Actininse type similar conditions would then prevail with 

 regard to the family Endocoelactis, if my supposition that this family has no longi- 

 tudinal muscles in the body-wall proves to be correct. Should it be the case, on the 

 other hand, that M'Murrich is right in saying that such longitudinal muscles occur 

 in Halcurias pilotus, the family Endocoelactidse must be referred to the Protactininse. 

 In this case the thickenings of the basal parts of the ectodermal cells in the body-wall 

 may be considered as rudimentary epithelial muscles, a view, however, I do not hold, 

 and a question that can only be answered by means of good material of maceration. 

 For practical reasons it would possibly be advisable in future to combine the family 

 Endocoelactidse with the Protactininse, and the Discosomidse with the Protostichodac- 

 tylinse, a grouping which I already, in 1900, p. 57 (77), pointed out as possible with 

 regard to the Discosomidse. 



In my opinion, the family Endocoelactidse must thus belong to the lowest Actininse, 

 or possibly to the more differentiated Protactininse. Probably an intimate relation to 

 any other Actininse family does not exist. 



Before concluding the account of the relations of the family Endocoelactidse to 

 other Actiniaria, we might just point out that these variations are of importance for 

 the study of the other Anthozoa. As already known, the skeleton-forming Madreporaria 

 show similar variations from the usual symmetry, as the Actiniaria, as 8- and 10-rayed, 

 radial or more bilateral forms are found even there. As we have seen that such a 



TRANS. ROY. SOC. EDIN., VOL. L. PART I. (NO. 4). 9 



