64 



CERIANTHARIA. 



traction of the mesenteries. In some species {C. viembranacais, P. multiplicatus, maua and C.amcricanns) 

 the ciliated tract region forms craspedonemes, which below the straight ciliated region issue one 

 after the other from the free border of the mesenter)-, and finally terminate, at least in the first three 

 species, in a more or less compressed bunch"). A typical example of this successive outgrowth of cras- 

 pedonemes is seen in protomesenteries 2, the so-called continuous mesenteries, of C. mevibranaccus^ as 

 also in P. multiplicatus in which the craspedonemes shew most ramification, and in which they occur on 

 all metamesenteries of the first and second cycles. The bunches of the craspedonemes are seen most 

 finely developed on the metamesenteries of the first cycle in C. iiu-nibraiiaccus. The structure of these 

 formations has been elucidated by O. and R. Hertwig (1879). They have shewn that the cra.spe- 

 doneme consists of a thread-like or a flattened process of the mesogloea and entoderm, over which 

 passes a descending and an ascending limb of the ciliated tract region of the filament. The ciliated 

 tract region of the filament passes from one craspedoneme to another and forms a continuous co\'ering 

 of all craspedonemes. In a transverse section therefore of the craspedoneme a ciliated tract region of 

 the filament is seen at the two poles and in the middle an entoderm part supported by the mesogloea 

 (Textfigure 12). My investigation of the structure of the craspedoneme, whether the craspedoneme 

 is simple or with bunches attached, is in complete accord with the statements of the brothers 

 Hertwig. 



Before concluding the description of the craspedonemes of the region of the ciliated tracts I 

 may mention a peculiar structure of a craspedoneme in the ciliated tract region of P. mana. Whilst 

 as a rule the craspedonemes in the ciliated tract region shew a typical structure, and externally a 

 flat ribbon-like form, on the two sides of which runs a layer of filament, there was found on several 

 metamesenteries in the most aboral jjart, a craspedoneme which in transverse section had a more 

 rounded .shape and externally suggested an "acontium". The anatomical investigation of these cras- 

 pedonemes proved that the craspedoneme consisted of a relative!}' slight endoderm part which bore 

 on one side, the oral, only a mesenterial filament of the ciliated tract region (Textfigure 13). So much was 

 quite clear, but the constitution of the extremity of the cra.spedoneme I could not determine in all 

 points with certainty, as the specimen was in a poor state of preservation. The part of the ciliated 

 tract attached to the different mesogloeal processes certainly disappears in this case, as also the meso- 

 gloeal processes, and the entoderm part is reduced in size, but apparently the portion of the ciliated 

 tracts that lies close to the median streak and bounds the inside portion of the ciliated groove is still 

 left. If, as I think i)robable, these details are correctly observed, a ciliated tract region would not 

 occtn- in its entirety in the most distal portion of the craspedoneme, but merely an undifferentiated 

 median streak with half onl>' of tlie ciliated grooves. These acontium-like craspedonemes are therefore 

 constituted — and the same is the case with the craspedonemes of the craspedia in C. aiiirricamts^ 

 as we shall see below — quite differentl)- from the acontiain the genus yir«f//«««/'/«<j and in Araclniactis^ 

 though they have the .same external appearance and correspond in situation. [Compare Carlgren (1912)]. 



The ciliated tract region is best developed on the metamesenteries of the first and second 

 cycles, and on protomesenteries 2, since protomesenteries 3 and the metamesenteries of the 3rd and 



') The bunches of the craspedonemes are not therefore van Beneden's "acontia", as this investigator considers 

 possible (189S p. 33), but doubtless M c. Murrich's "acontia" in C. americamis (1890 p. 13S, fig. I, PI 7). 



