TO CERIANTIIARIA. 



tocysts which occur in great minibers, are situated betwen the two limbs of the filament, that is to say 

 they are set in the endoderni. The entodermal part of the acontium thread is slightly larger here 

 than in the species mentioned previously. 



The threads which van Jjeneden called "acontia" are thus in my view composed in the 

 main in the same way as the craspedonemes of the ciliated tract region in Ccriatithiis vicmbranaccus 

 or a craspedoneme of the cnido-glandular tract in I\ so//7ar//{s, in the respect, that is, that they are 

 composed of an ascending and a descending limb of the filament, which encloses an 

 endodermal process of the mesentery. The only difference is, that here the endodermal part 

 is reduced in size much more than in the craspedonemes alluded to, and to such an extent that the 

 entoderm in the distal portion of the acontium completely or all but disappears, so that the two limbs 

 of the filament become completely or all but contiguous (textfigure 14). 



If we may thus assimilate the "acontium" to a craspedoneme, the question arises whether the 

 principal part of the acontium, namely the filament jiart, is a differentiation of the median streak and 

 in that case of the craspedion, or whether it is to be regarded as a modification of the cnido-glan- 

 dular tract. To answer this question, the only sure data for our guidance are to be foiuid in the 

 structure and situation of the "acontia". To take first the situation of the "acontium" and of tlie cnido- 

 glandular tract, it would seem at first sight to be of no use in establishing an homology, because 

 the cnido-glandular tract of the metamesenteries of the first and second cycles as well as of proto- 

 mesenteries 2 does not occupy the same position in the different Cerianthidae. If, for instance, we 

 look at Pacliyccriaiithus innltiplicatns and Ccrianf/ius /loydii\ both of which have very long meta- 

 mesenteries, and at protomesenteries 2 of Ccriaiitlnis vicinbraiiacciis^ we find that on the mesenteries 

 in question the cnido-glaudular tract is situated not far from the stomatodaeum in C. lloydii. but in 

 the other two cases in the most aboral part of the mesentery. We find however on closer examina- 

 tion, that this difference depends on the varying length of the ciliated tract region in the different 

 species. For, if we examine the disposition of the different parts of the filament, we find that in 

 immediate proximity to the aboral border of the stomatodaeum a ciliated tract region is developed, 

 next comes a cnido-glandular tract in the aboral direction, and lastly in many species a craspedion 

 region which at its aboral end may form a craspedoneme (in C. amcricanus). In other words, the 

 cnido-glandular tract lies between the ciliated tract region and the craspedion region. Now, as on the 

 mesenteries that bear "acontia" in Arac/iiiant/iits and Arachnacfis s. str. — the ciliated tract region 

 passes immediately into the craspedion region without forming a cnido-glandular tract region, and the 

 "acontia" are situated at the aboral end of the craspedion the situation of the "acontia" tells in favour 

 of the hypothesis that they are differentiations of the craspedion. 



Secondly, the anatomical structure of the "acontia" va. Arachnanthus owA. Arachnactis alhidn leads 

 us to a similar conclusion. For, if we exclude supporting cells, the great bulk of the filament part of 

 the acontia is composed, as we have seen, of gland cells and these mucus cells, whilst nematocysts 

 are so rare as scarcely to count as a special feature at all. A longitudinal muscle layer is also de- 

 veloped, which, like the filament part itself, is ectodermal. 



The presence in the acontia of the numerous homogeneous gland cells and the extremely 

 well developed longitudinal muscles, which are very rare in the cnido-glandular tract if not entirely 



