MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 453 



Transverse sections through the discal cone reveal live pairs of complete mesenteries, without any 

 indications of others. The musculature is yet too feebly developed to allow of the paired charac- 

 ter of tiie mesenteries being established l)y this means alone, but the larger interspaces and the 

 examination of lower sections prove that the enumeration added is correct. In addition to the 

 four Edwardsian mesenteries, the liilateral pair between the first and second pairs is complete, 

 while the pair between the first and third, to be seen in the sections below, is still incomplete. 

 In sections through the actual column wall, below the stomodisal region, the six pairs of protoc- 

 nemes are equal in radial length, and six alternating pairs of metacnenies occur, the sulcar mem- 

 bers a little in advance of the sulcular. In this instance the lifth pair is again complete before 

 the sixth, and the first cycle of metacnemes is established before all the protocnemes are united 

 with the stomod:eum. In the bud of AsfniiH/Hi. represented on p. 4r(i(t, fig. 8'^ the fifth mesentery 

 on the right side is completed before the one on the left side. 



In the section of the larva of Favia fragum^ represented in fig. 11-t, the pair between the 

 first and second Edwardsian pairs is already represented on the larval wall, some distance below 

 the stomodfeum. ))y small mesogheal enlargements, but no trace of any such enlargement yet 

 occurs between pairs I and III, where the sixth pair will be situated {cf. fig. 116). 



From these examples there can be no doubt that though the fifth and sixth pairs may at times 

 appear simultaneously, j^et at other times an interval occurs. The pair of mesenteries between 

 the first and second protocnemic pairs is to be regarded as the fifth in the sequence, and the 

 pair between the first and third as the sixth or last in the development of the protocnemes. The 

 enumeration of these two pairs in H. V. Wilson's figures of the mesenteries in an attached larva 

 of Manicina (PI. V, fig. 39) should therefore be reversed. 



In all the instances 3'et referred to, the fifth and sixth mesenterial pairs remain free from the 

 stomodfeum, and in numerous cases it has been found that this condition is retained for a very 

 lengthened period. In young pol3^ps of S. radians completion was not attained within the course 

 of four months, though the second cycle of mesenteries had appeared in the meantime. Likewise 

 in the young polyp of F. fraijaai. with four pairs of metacnemes, the\- are still free (PI. XIV, 

 fig. 109). All coral larva? appear to settle at or about the Edwards ia-fita,g(^ . and the septa then 

 begin to make their appearance. 



No important resting stage, in the appearance of the protocnemes, seems to be indicated in 

 any of the investigations yet conducted on the Madreporaria, though the comparatively strong 

 development of the first pair must not be overlooked. With this possible exception, the develo^:)- 

 mentfrom the first to the sixth pair progresses with uninterrupted regularity, and the same ma}' 

 be said of the further growth of the first four pairs, as concerns their union with the stomodfeum, 

 but a prolonged interval separates the further development of the fifth and sixth pairs. Lacaze- 

 Duthiers and others have endeavored to establish several resting stages in the appearance of the 

 six pairs of primary mesenteries of the Actiniaria, but the Madreporaria afford little support for 

 such. 



The incompletion of the fifth ai\d sixth j^rotocnemic pairs is j^ermanent in certain Zoantharia, 

 and therefore this condition can not necessarily be looked upon as a developmental resting stage 

 in the .sense of Lacaze-Duthiers. Wherever these pairs appear in the Actinian family Edwardsidie, 

 they remain free from the stomoda?um, and the same relationshij) holds for Gonactinia, etc. 

 Throughout the Zoantheie the fifth pair remains incomplete, and the sixth pair becomes complete 

 only in macrotypic members of the group. In the West Indian A/'2'>tas>a aiunda/a, (hes.), I have 

 found the pairs to remain free for the most part, although the second, third, and fourth orders of 

 mesenteries were fully developed. They remain permanentlj' free in Jfadrejjora and Porltes, even 

 when an increase beyond the iisual twelve takes place. Numerous other instances may be cited, 

 all tending to show that some significance attaches to the incompletion of the fifth and sixth 

 protocnemic pairs, as compared with the completion of the other four biliitcral i)airs. It is not 

 merely a lagging l)ehind in growth due to their later appearance. 



My studies lead me to believe that the earliest corals and Actinians were characterized by 

 anisocnemic pairs (excepting the directives), as compared with the isoenemic pairs of later corals 

 and Actinians. To-dav. the former condition is retained in the Zoanthids, and in Pai^ites and 



