,Q CERIANTHARIA. 



glyph is broadest in the genus Arac/iiiaufhiis^ in that ahnost half the total number of mesenteries 

 (8—10) join on to the siphonoglyph. 



The ectoderm of the siphonoglyph contains, besides strongly ciliated supporting cells, niuctis 

 cells in great abundance. As a rule these mucus cells are most numerous at its boundary with the 

 stouiatodaeum. Sometimes as in Ccriantheopsis amcricanus the mucus cells are spread almost uniformly 

 over the whole surface of the siphonoglyph. Here and there nematocysts occur, but always in small 

 numbers. 



The hyposulcus as a rule is rather short. In Pacliycrrianflnts arstiiarii it is however rather 

 well developed (Torrey and Kleeberger 1909, fig. i). It is still longer in the ^^x\\\s Anichiiautlnts^ 

 where it is longer than the stomatodaeum itself. The ectoderm of the hyposulcus is provided with 

 fewer mucus cells than that of tlic siphonoglyph, which i.* the reason why more supporting cells 

 occur in it than in the other. In other respects its structure agrees with that of the ectoderm of 

 the siphonoglyph. In tlie genws Arachnanthus as also in the larval gewu^ ^Irac/niacf is, the hyposulcus 

 is provided along its free sides with filament differentiations (compare p. 36 and Carlgren 1912, 



P- 364, 371)- 



The heraisulci are unequally developed. Whilst they are absent in the genus Aracluiaiiflnis 



and rudimentary in Botnicnidi/rr, in the rest of the species examined by me they are more or less 

 distinctly developed. In P. solitaritts and multiplicatus and in C. Uoydii they correspond in structure 

 and in appearance. Their outline is rounded in transverse section and calls to mind a transverse 

 section of an undivided median streak, though they are not so sharply distinguished from the mesen- 

 teries as these (compare fig. 4, PI. i). Tlieir structure corresponds with that of a median streak, 

 tliat is, there are mucus cells and for the most part numerous spirocysts also. In V. iiioiibnuiacnis 

 the structure deviates to some extent from the above. Lengthwise the hemisulci are very elongated 

 flattened, and are connected with one border of the directive mesenteries. They are strongly ciliated 

 like the siphonoglyph and contain chiefly supporting cells, whilst the mucus cells, thick-walled nema- 

 tocysts and spirocysts are sparse. The hemisulci in Ceriaiitlicopsis amcricaims (fig. 4, PI. 5) occupy 

 an intermediate position between those already mentioned. On the homology of the hemisulci and 

 hyposulcus with the filaments I may refer to page 63. 



3. Arrangement of the mesenteries and the reproductive organs. 



The arrangement of the mesenteries varies a good deal in different species, though a definite 

 scheme of their arrangement ma\- always be discovered even if tlie regular arrangement is again and 

 again interfered with. In P. solifarhis especially such disturbances are very connnon and might almost 

 obscure the typical arrangement (Carlgren 1912 text-figure 4a, b) but in other species also, as in 

 P. untltiplicatus and Botrucnidifer norvegicus^ irregularities in the arrangement present themselves. 

 At any rate in P. solitarhis the irregularities in the grouping of the mesenteries are connected with 

 an irregular development of the tentacle cycles. As to the number of the primary mesenteries, the 

 protomesenterics, my views coincide completely with van Bene den's. The first three mesenteries 

 on either side of the directive plane arc in my opinion protomesenteries; the nictamesenteries there- 

 fore begin with tlic fourth coui)le. Below I give the precise grounds for this opinion of mine. 



