2 4 



CERIANTHARIA. 



longitudinal ridges in the aboral part are, as in C. lloydi-i, supported by mesogloeal processes, which 

 is not the case in the oral part. The structure of the ridges agrees that of C. lloydii. Here very 

 numerous spirocysts occur. Thick -walled nematocysts of a length of 24 31 ft are not infrequent. The 

 ectoderm of the siphonoglyph is prominent and contains specially numeroiis, closely packed homogeneous 

 gland cells (mucus cells). 



The ectoderm of the hemisulci consists in the oral part chiefly of strongly ciliated supporting 

 cells; outwards from the pair of directive mesenteries the homogeneous gland cells become more 

 numerous, and a number of thin-walled nematocysts also occur; aborally the gland part gets predo- 

 minant till at last it gradually disappears. The extension of the hemisulci upon the mesentery is 

 indicated here also by the strong ciliation (Fig. 4, pi. 5). The hemisulci do not form a filament-like 

 part, as in P. multiplicatus, btit usually line only one side of the mesentery. A ciliated tract on the 

 outside is not developed, contrasting on this point with the genus Arachnanthus. The longitudinal 

 musculature of the hemisulci is faint, with the exception of the part which lies below the cnido- 

 glandular cell portion, where it is strong. 



The arrangement of the mesenteries is given in detail by Me. Murrich (1910). In the main 

 my observations agree with this investigator's, though to some degree the details vary. Not only the 

 regions of the ciliated tracts but also those of the cnido-glandular tracts lie in all mesenteries com- 

 paratively close to the lower border of the stomatodaeum, so that the craspedion region on the longer 

 mesenteries, is specially long. 



Protomesenteries i: The directive mesenteries are comparatively short, without ciliated tracts. 

 For the extension of the hemisulci on these mesenteries, see above. 



Protomesenteries 2 are long and reach nearly to the aboral pole of the animal. The ciliated 

 tract region is straight at the beginning, but lower down forms first short then longer, closely set 

 craspedonemes. The region of the cnido-glandular tract is prolonged and forms a winding band (fig. 5 

 pi. 5). The region of the craspedia is very long and takes up the greatest part of the free border of 

 the mesentery. In the most aboral part it forms a threadlike process (fig. 6, pi. 5). They are fertile. 



Protomesenteries 3 are short though distinctly longer than the metamesenteries of the 3rd and 

 4th cycles. The ciliated tract region with its craspedonemes ends higher up than on protomesenteries 

 2. The region of the cnido-glandular tract is more strongly developed than in those mesenteries. 

 As usual with the Cerianthidae they are sterile. 



The metamesenteries are arranged on the formula mBMb, that is, in every quartette there is 

 first a mesentery of the 2nd cycle, then one of the 3rd, next one of the ist and lastly one of the 4th. 

 As Me. Murrich considers that the metamesenteries begin with the 5th pair of mesenteries, counting 

 from the directive chamber, the mesenterial arrangement is according to his formula BMbni. A de- 

 viation from this rule would occur, if we may judge from Me. Mur rich's figure (IV 1910) of the 

 arrangement of the mesenteries, in the ist quartette where B takes 's place, and conversely. As far 

 as I can tell, however, from specimens examined, this is not the case, for this group also shews the 

 typical arrangement. It is characteristic of this species, and of the whole genus perhaps, that the 

 metameseuteries of the ist cycle increase in length towards the multiplication chamber, as Me. Mur- 

 rich has shewn and I can confirm. Still in several places there occurs with a certain regularity a 



