van Beneden should draw the line which separates his Cerinula stage from the second phase 



of development immediately at th< of the formation of the third couple of mesenterii 



A clue is probably to 1»- found in his observation that all mesenteries in the Ceriantharia 



makt- their appearance dorsal to the first couple, with the exception of the third couple. This 



intern:- mmetrical order of the succession, which. after the appearance of the third 



again resumed, and it is the resumption "f tin- symmetrical succession that marks the 



inning of the second phase of development. rhe interruption, however, seems to 1" of 



s importance than such a view would give it. and further, it may bc noted that the same 



symmetrical succession and interruption may be recognized in the development ui the protocnemes 



of the Hexactiniae, luit since the deuterocnemes do not form successively in the dorsal inter- 



mesenterial chamber it is not so striking as in the Ceriantharia. The men- fact that there is 



an interruption to the continued symmetry of succession militates against its acceptance as a 



criterion, and the facts that the same sequence exists for the first four Cerianthan couples as 



the same Hexactinian couples and ihat an entirely different order of sequence supervenes 



with the fifth couple in the two groups seem to point strongly to the recognition of the identity 



of the tirst four couples in both groups. 



I have in earlier papers taken the position that not two but three stages of development 



nized in the Hexactiniae; (i a protocnemic stage in which four couples oi 



mesenti re formed, (2) a deuterocnemic stage in which two additional couples appear, one 



in each of the lateral interspaces and (3) a zygocnemic stage in which the mesenteries appear 



in pairs. The firsl stage I believe to bc common to all the Anthozoa, and to be the only 



ited in the Alcyonaria, the deuterocneme stage occurs in the Ceriantharia, Zoantheae 



and Actiniaceae, but while it is identical in the Zoanthea- and Actiniaceae it is different in the 



riantheae, that group separating from the other two at the conclusion of the first stage. 



Finally, the zygocneme stage occurs in the Zoanthea^ and Actiniaceae, but differs in the two 



■ups. which separated at its appearance. Duerden's observations indicate that the Rugosa 



also possessed three stages, the third again differing from that of the Zoantheae and Actiniaceae, 



and from van Beneden's own observations (1898) it would seem that only two stages are 



nted in the Antipatharia and that these stages are possibly identical with the corresponding 



of the Actini and Zoantheas. This last point can only be settled, however, by a 



termination of the order of sequence of the Antipatharian mesenteries. 



1 have discussed this question of the homology of the Cerianthan protocnemes at somc 



length, because it has an important hearing upon the significance of certain structural peculiarities 



which I have noted in the forms represented in the "Siboga" collection. These peculiarities do 



not in themselves involve the identity of the protocnemes of the Ceriantharia with those oi 



the Hexactiniae, but they do point strongly toward a necessity for a recognition of four couples 



nemes in the tonner group. In considering these peculiarities it will be well to review 



briefly our knowledge of the arrangement of the mesenteries in the Ceriantharia, but before 



■ is necessary to consider the terms of orientation that are to be used. 



It has been a general custom to speak of the two extremities of an Anthozoan as oral 



and al 1 listal and proximal, and to regard the edges of the plane which passes 



8 



