CKRIANTHARIA. 



73 



I have described above a Cerianthid with botrucnidae, Botrucnidifer norvcgicus. Botrucnidae 

 occur in this species on protomesenteries 2 and 3, on the first quartette of metamesenteries, and on 

 the longest mesenteries of the second. Their position is the most aboral part of the mesenteries, that 

 is to say, they occupy exactly the same position as in the larval forms described by van Beneden. 



In the first place as to the histological structure of the isolated cnidorages, it is possible that 

 in certain cases van Beneden may be right in his statement that they consist only of epithelium, 

 in certain isolated cnidorages however I have succeeded in as- 

 certaining beyond all doubt the presence of a mesogloeal por- 

 tion (figs. 7, 8, PI. IV.). The stalked botrucnid-grapes also are 

 otherwise constituted, according to my investigations, than as van 

 Beneden describes. They are not simply an epithelial formation, 

 the mesogloea enters into the composition of these 

 grape-like organs. That is, the mesogloea of the mesentery is 

 continued into the botrucnidae, ramifying like a tree and giving 

 off a branch to each botrucnid-grape (Textfig. n). The mesogloeal 

 process passes through the grape-stalk and broadens out at the 

 base of the berry forming a sort of cup or funnel. From this 

 cavity thin threads apparently proceed in the direction of the un- 

 attached side of the grape, though the course of such extremely 

 fine threads is hard to follow. In any case the mesogloeal pro- 

 cesses sustain the usual supporting cells, which cover the outside 

 of the basal portion of the grape, and within the cavity lie the 

 bases of the very large thick-walled nematocysts, which are set 

 close together and project outwards towards the periphery of the 

 grape. Between the thick-walled nematocysts lie the thin-walled 

 ones, the smallest of which are just below the free surface of the 

 supporting cells. 



It holds good therefore that the botrucnidae and cnido- 

 rages contain not merely epithelial but also mesogloeal parts. 



Van Beneden, as already said, regards both the acontia 

 and the botrucnidae as endodermal organs. I cannot endorse this 



view: everything points rather to the ectodermal origin of both acontia and botrucnidae. As I 

 have said above, they issue from the most aboral portion of the craspedion and come into immediate 

 contact with the filaments. Whilst the acontia are continuous with the craspedion the botrucnidae 

 are separated from the craspedion; the latter circumstance however is a consequence of the structure 

 of the botrucnidae, in as much as in them the filament part is broken up into a number of lobes. 

 Moreover the presence of numerous spirocysts is a point in favour of the comparison of the botrucnidae 

 to the craspediou. And the statement of Torrey and Kleeberger that the cnidorages in Botru- 

 antlius occur on craspedonemes points the same way. If the filaments are ectodermal in the first 

 place, and this van Beneden himself assumes, then the botrucnidae are ectodermal also. Only the 



Textfigure u. Botrucnidifer norvegiciis. 

 Section of the column and a mesentery 

 with botrucnidae. On some of the botru- 

 cnid-grapes only the epithelial elements 

 are sectioned. 



The Danish Ingolf-Expedition V. 3, 



10 



