C lassifica ti o n. Until 1898 no attempt was made to divide the order Ceriantheae 

 into other than generic groups, but in that year van Beneden recognized two suborders in the 

 larvae he described, the Botrucnidiferae and the Acontiferae. The Acontiferae are characterized 

 by the marginal tentacles appearing in the same order as the intermesenterial chambers with 

 which they communicate, the unpaired median tentacle, however, always appearing later than 

 the next adjacent couple; certain of the mesenteries bear acontia but never botrucnidre. In the 

 Botrucnidiferae, on the other hand, the order of succession of the marginal tentacles, after the 

 appearance of the first three couples, does not correspond with that of the mesenterial chambers, 

 but they appear on either side in alternate chambers, the intervening chambers developing 

 their tentacles only later, probably simultaneously with the appearance of the median tentacle; 

 none of the mesenteries bear acontia, but a greater or less number of thetn are provided with 

 botrucnidae. These characters seem to definitely limit two groups. 



In my discussion of the arrangement of the mesenteries of the Ceriantheae (p. 9 et seq.) 

 I have shown that in the forms with which we are at present familiar two principal types are 

 recognizable. In one of these the telocnemes are formed from the second couple of protocnemes 

 and in the other they are formed from the fourth couple. This seems to be a difference of 

 considerable importance, and on the strength of it I would recognize in the Ceriantheae two 

 families, in one of which, the family Cerianthidcr, the telocnemes are the second couple from 

 the mid-ventral line, while in the other, for which the term Arachnactida may be employed, 

 the telocnemes are the fourth couple. 



We have then two sets of characters which seem to be of oreater than generic 

 importance and the question arises as to what relation, systematically, they bear to one 

 another. The question is at present a very difhcult one to answer. An examination of the 

 known larval Botrucnidiferae seems to indicate that in some members of the group the second, 

 in others the fourth protocnemes form the telocnemes, just as is the case in the Acontiferae. 

 Thus, in Calpanthula guineensis it would seem that the telocnemes were to be formed from 

 the second protocnemes, while in Hensenanthula dactylifera and H. melo it is probable that 

 they are formed from the fourth protocnemes. What may be the fmal result in Cerianthula 

 mediterranen, and H. spinifer is less certain ; in the larvae studied the second protocnemes 

 were the longest, but this condition may be corrected in older stages, the known larvae of 

 both species being in a much earlier stage of development than those of the other forms. In 

 the only adult form known to be botrucnidiferous Botryanthus benedeni (Torrey and Kleeberger, 

 1909), the telocnemes were formeel by the fourth protocnemes. 



Accepting this conclusion, if van Beneden's groups be regarded as the more inclusive, 

 both arrangements of the telocnemes will be found in each group, and, conversely, if the 

 telocnemic conditions be taken to represent the larger division, then included in each of these 

 divisions there will be both acontiferous and botrucnidiferous forms. This is not inherently 

 impossible, but it is a difhcult matter to decide which is the more primitive character. The 

 occurrence of botrucnidae is associated with an alternating development of the deuterocnemic 

 tentacles and that of acontia with their successive development. Van Beneden's groups, having 

 thus two associated characters, seem to have the claim to the higher rank. 



