AQ CERIANTHARIA. 



cylindrique. Disposition biseptale faisent alterner des cloisons fertiles acontiferes avec des cloisons 

 steriles privees d'aconties et portant des fils meseiiteriqiies sur leur craspedes. Musculature endoder- 

 mique a peine represente. Paroi de la colonne epaisse et consistente, a cause du developpement pris 

 par la musculature longitudinale et par la mesoglee." In his description of the type P. irncdrni Ronle 

 adds the following concerning the filaments and acontia, after mentioning that filaments (craspedes) are 

 found on all mesenteries, except the feeblest. ■'Seulement et en alternance les unes out des fils mesen- 

 teriques sur lour craspedes et les autres n'en portent point. Les cloisons ainsi privees sont fertiles, 

 denuees de fils mesenteriques, elles possedent par contre des aconties sur uue assez grande part de 

 leur extremite inferieure; leur serie debute par S3 et continue par les numeros d'ordre impaire S^ S; Sc,". 



According to this the fertile mesenteries would bear "acontia" whilst the sterile mesenteries 

 would be provided with craspedonemes. If Roule's view were correct, Pacliyceriaiitluis would thus 

 be a genus coming under Acontiferae. But this, I think, can hardly be the case, as I suspect that 

 Roule had not clearly grasped the true nature of an "acontium". My supposition is svipported too 

 by the expression of Roule's just quoted, that the acontia are found "sur une assez grande part de 

 leur (des cloisons) extremite inferieure". As far as at present known, the acontia are never distributed 

 over the mesenteries in such a way; it seems likely therefore that Roule has confounded them with 

 the craspedonemes of the region of the ciliated tracts. Padiyceriantliiis bcvcdciii might thus stand as 

 the type of a genus included in the first Ceriantharia family, that which does not possess acontia nor 

 botrncnidac. To this genus would belong P. fii/ibriatus Mc. Murrich, C. llovdii Gosse and others (see 

 tables below p. 44—47) but not C. oligopodiis, for which a new genus must be set up, unless a larval 

 genus is taken for this form, which is not likely to be so good, as the development of the species 

 and consequently the connection between the larval and the full-grown form is not quite certainly 

 known. I call this genus Araclivaritluis and take as its type A. oligopodits (Cerf.), as this species is 

 the best known. 



I now suggest therefore a different division of Ceriantharia, based on a thorough anatomical 

 study of different species. Besides the genus Aracliiiantliiis just mentioned, we now institute two 

 more new genera Botrucnidifrr and Crriantlunpsis (Types B. iioi-jegiatsw.^^^. and C.aiiirric(iuHs\' exx). 

 Very likeh- this division may also have to be modified later when a larger number of species of 

 Ceriantharia have passed under close anatomical investigation. But by starting from those features 

 which I have indicated in the diagnosis of families and genera below, it may l^e possible gradually 

 to secure a satisfactory classification of Ceriantharia. M>- suggestion for the division of Ceriantharia 

 is of the following nature. 



Fam. I Ccrianthidae. 

 Ceriantharia without "acontia" and without cnidorages and botrncnidac. 



e n u s I . Pa cJiyceriantlnts Roule. 

 Ccrianthidae whose 2nd couple of protocnemes are short, sterile and pro\ided with an extremely 

 well developed region of the cnido-glandular tract. Arrangement of metacnemes (= denterocnemes 

 Mc. Murrich) in each quartette M.B.iii.h. (1,3,2,4) more or less distinct. 



