CERIANTHARIA. 



difference in length between the two cycles is slight. The}' are all short, shorter than protomesen- 

 teries 3 and diminish in length towards the multiplication chamber. In contrast with the metamesen- 

 teries of the ist and 2nd cycle, they have no development of the craspedia. 



The number of the metamesenteries comes to about 70 in the larger specimens. 



The filaments: The ciliated tract is specially well marked in protomesenteries 2 and the meta- 

 mesenteries of the ist and and cycles, rather more feebly in the others. The median streak contains 

 numerous spirocysts and shews especially in the metamesenteries of the ist and 2nd cycle in the 

 distal part a distinct tendency to bipartition, though by no means so pronounced as in Arachnactis 

 lobiancoi. The mesogloeal processes, on which the chief part of the ciliated tracts rest, arise however, 

 as in the majority of species for inst. as in C, membranaceus, from the middle lamella of the mesentery 

 and not from the mesogloeal fold of the median streak. In transverse section the median streak is 

 broad with 2 strong mesogloeal folds projecting towards the sides. The outline shews in the middle 

 a very deep depression and the ectoderm is rather thinner than at the sides, and in the same way 

 the gland cells and nematocysts are fewer in the middle than at the sides, and in this way also the 

 median streak shews signs of dividing into 2 tracts, and perhaps more markedly than in C. mem- 

 branaccus. In the metamesenteries of the 3rd and 4th cycles this differentiation is less advanced, so 

 that the filaments sometimes have the same appearance as in Botrucnidifer (Fig. 5, PI. 4). The longitu- 

 dinal musculature in the median streak is well developed. The ciliated tract region of the filaments 

 is straight without craspedonemes. 



The cnido-glandular tract is of the usual structure with numerous gland cells, and thick-walled 

 nematocysts broadening towards the base, of 31 34//s length. Whilst it has very wavy coils in 

 the mesenteries of the 3rd and 4th cycles and also in protomesenteries 3, it is straight or slightly 

 coiled in the metamesenteries of the ist and 2nd cycles and in protomesenteries 2 and in the case 

 of the latter is strongly suggestive of v. Beneden's acontia. Like these and tne craspedonemes 

 the mesenterial process, on which the filament rests, shews an ascending and a descending limb of 

 the filament separated by the endoderm of the mesentery. The latter is very slight at the extremity 

 of the process, so that the two limbs of the filament almost press upon each other (giving the 

 process a structure which agrees with that of the acontium as I have described it below a des- 

 cription which differs not a little from v. Beneden's but agrees in the main with Cerfontaine's), 

 higher up however towards the base of the process, this part gets distinctly larger. In fig. 4, PI. 4, which 

 -represents a transverse section taken some way from the extremity of such an acontium-like process 

 in protomesenteries 2, we see at the sides the two intersected filament-parts of the cnido-glandular tract 

 with their thick-walled nematocysts and gland cells separated by a part of the endoderm. It is noticeable 

 too that the part of the mesogloea that belongs to the filaments shews a tendency to part off from 

 the mesogloea of the endoderm, a circumstance which is of interest for the study of the structure 

 of the "acontium" in other species. Longitudinal musculature is absent in the cnido-glandular tract. 



The craspedia, that is to say the parts of the median streak that lie below (aborally) from the 

 cnido-glandular tract, are very well developed in the upper part (Fig. 6, PI. 3), and taper away more 

 and more towards the aboral pole. In their upper part are found, besides homogeneous and granular 

 gland cells, numerous spirocysts, in the lower part appear also the large thick-walled nematocysts 



The Danish Ingolt'-Kxpedition V, 3. 3 



