454 .MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



Madrepora, and as regards the protocnemes it persists tor a long time in the growth of all others. 

 The union with the stomodseum of the fifth and sixth protocnemic pairs is assumed after a time 

 }>v forms characterized by cycles of isocnemic mesenteries: but they remain incomplete in Actin- 

 ians and corals (Zoanthids, Pontes, Madrepora) characterized by anisocnemic pairs throughout. 

 Perhaps an earlier phylogenetic stage is represented by the Cerianthids and Alcvonarians. in which 

 the incomplete moieties altogether fail to appear. 



Much difference is apparent as regards the position at which the mesenteries first make their 

 appearance. The first two or three pairs seem to arise in the angle between the stomodasal wall 

 and the outer wall of the larva, and then to grow in both directions — that is, down the stomo- 

 dteal wall and the larval body wall, the latter extension being the more rapid. This is very 

 clearly shown in the figure of the living larva of Isophyllia, and in the sections through the oral 

 region (PI. XVII). While the members of the middle pair extend all the way down the 

 stomodaml wall, and nearly as far as the aboral pole of the larva, the two smaller pairs pass 

 scarcely at all down the stomodseum, and only for a short distance along the larval wall. The three 

 first mesenterial pairs in Favia probably arise in the same circumoral position, but the fourth 

 pair is first apparent on the larval wall a little below the stomodaHim (Pis. XIV, XV), and the 

 fifth and sixth pairs arise still farther down. 



Early bud polyps of Madrepora also illustrate the same relationship (PI. III). Here the 

 four primary mesenterial pairs are seen in the angle between the outer wall of the bud and the 

 stomodseal wall. They extend the whole length of the stomodseum, and for some distance along 

 the outer coenosarcal wall. On the other hand, the rudiments of the fifth and sixth pairs are 

 first seen on the outer wall, some distance removed from the oral aperture. 



Probably in most Actiniaria and Madreporaria the fourth, fifth, and sixth protocnemic pairs 

 arise independently of the stornodreum, and some way from the oral pole. When the appearance 

 of the tentacles has established the topographical regions of disk and column wall, the metacnemic 

 pairs are also found to arise somewhere on the latter, usually nearer the oral than the aboral 

 extremity. Onh T later do they grow upward, and then inwardly along the discal wall, and in 

 most cases ultimately reach the stornodaeum and extend down it. 



A marked distinction may thus be established in the place of origin of the different mesen- 

 terial pairs, the distance from the oral apertures varying with the relative age of the mesenteries; 

 the earliest pairs arise circumorally, the later pairs are some distance removed. This further 

 supports the contention that the Anthozoa are not primitively cyclical forms, but suggest an 

 ancestry in which the organs appeared bilaterally, in an antero-posterior succession. 



Several early stages, obtained in the development of bud polyps, suggest that in the asexual 

 method of increase there is the closest agreement in the order of appearance of the mesenteries 

 with that above described for the sexually produced polyps. The earliest stages have not been 

 secured, and the evidence is therefore not so complete as in the larva 1 . The youngest bud is one 

 of Cladocora arbuscula, in which eight protocnemes are already present, all united with the 

 stomodseum. Shortly below the stoniodseum only four mesenteries remain, and bear mesenterial 

 filaments; then two of these disappear, and the remaining couple are continued much farther, and 

 bear filaments almost to their termination. The musculature 1 at this stage is too weak to permit 

 of the actual arrangement in pairs being determined by means of it, but from the greater length 

 of one pair of mesenteries, and the stronger development of its filaments, it may reasonably be. 

 assumed that it represents the first pair of mesenteries, and that the order of disappearance of 

 the others indicates their successive origin. 



In other buds of Cladocora examined, all the protocnemes are already developed, the 

 Edwardsian mesenteries complete, and the fifth and sixth pairs incomplete, just as in larva? at or 

 about the time of fixation (PI. VIII, fig. ill). Buds of Solenastraea have also been secured, in 

 which only four pairs of mesenteries are complete and bear mesenterial filaments, while the two 

 incomplete pairs are without filaments, and disappear in advance of the other mesenteries (Pis. 

 XI and XII). The bud polyp of Astrangia, whose mesenterial plan is represented on p. 460, 

 indicates a somewhat later stage. In the very early bud of Madrepora, already referred to. all 



