_. CERIANTHARIA. 



74 



stalk of the botrucnidae will be entodermal. It is difficult however to draw a hard and fast line be- 

 tween ectoderm and endoderm. And probably some cells at the base of the botrucnid-berry on the 

 outside of the cup are endodermal. 



Lastly as regards the homology of the botrucnidae, my view at the present time, as I have 

 said before, coincides with van Beneden's, that they are homologous with the acoiitia, 

 amongst other reasons because both in van Beneden's larval forms and in Botrucnidifer norvegicus 

 they form like the acontia the most aboral part of the craspedion region. It is true that the botru- 

 cnidae of Botruanthus benedeni appear to have a different position on the mesenteries, but as it is 

 very hard to determine the precise distribution of these botrucnidae, of which I may add, judging 

 solely by the account of Torrey and Kleeberger, we have no very exact description, and as such 

 distribution is totally at variance with that in the remaining Ceriantharia provided with botrucnidae, 

 I am unable at present to pay much heed to these investigations. Nor is it beyond the bounds of 

 possibility that the botrucnidae of these writers are not identical with the organs so named by van 

 Beneden. In their inner structure it is important to notice that the botrucnidae occupy an inter- 

 mediate position between the acontia and the craspedonemes of the craspedion. It is true that they 

 are quite different in shape, for instead of presenting the thread-like appearance of the acontia and 

 the craspedonemes, they are split up into a large number of grape-like organs. But such a circum- 

 stance need not preclude the essential homology of these organs, and the less so as an interruption 

 of the filaments may occur also in other Anthozoa. In Scyiophomis antarcticus for instance and in 

 some other soleless Actiniarian forms I have found a discontinuity of the ciliated tracts, and in Calli- 

 actis (Sagartia) parasitica the filaments on the least developed mesenteries shew interruptions in 

 several places in the ciliated tract region. Now if we imagine that a strongly coiled filament part 

 has parted in a large number of places and that each section so formed has rolled itself up, we shall 

 have an organ corresponding in form to a botrucnid. As to their anatomical structure, they have 

 reduced the number of mucus cells, and augmented the spirocysts and the thick-walled nematocysts, 

 but without developing any granular gland cells. 



Bringing into relation all the morphological data given above, I think that we are bound, at 

 least until further light is thrown upon the matter, to admit the homology of the botrucnidae of Bo- 

 trucnidifer and of van Beneden's larval forms with the acontia of the genera Arachnanthus and 

 Arachnactis and with the craspedonemes of the craspedion region in Cerianthcopsis amcricanus. 



Before I conclude, it may be convenient to give a short summary of the different kinds of 

 craspedonemes found in Ceriantharia, with some diagrams. 



As we have seen, craspedonemes may be developed as well or rather more from the ciliated 

 tract region of the filaments as from the cnido-glandular tract and from the craspedion region, though 

 in a more or less modified form. All these craspedonemes, - I disregard the botrucnidae are 

 flat or show themselves in transverse section as slightly rounded processes, in the middle of which 

 is an entodermal and a mesogloeal axis, and whose free edge is covered, for the most part at least, 

 with a filament part. There is thus found as a rule a from the oral side ascending limb of the fila- 

 ment and a descending limb. Only in the most aboral craspedoneme in the terminal portion of the 

 filament in P. maua does the descending limb appear to be absent. A transverse section of the eras- 



