MEMOIRS OF THE NATIONAL ACADEMY OE SCIENCES. 455 



the protocnemes are present, but their relative sizes are in conforiiiit,y with those of the Inids of 

 Chtdoc/ira and Soli'/iasfra'a, and may indicate a like successive origin. 



Summarizing, we find: (1) That the twelve protocnemes arise as six bilateral pairs in a defi- 

 nite sequence, which is probably the same throughout the Madreporaria, and conforms with that 

 characteristic of most Actiniaria. (2) The first two or three pairs arise at the angle l)etween the 

 stomoda>um and the larval wall, while the later mesenteries first appear on the column wall, some 

 distance from the oral aperture. (3) Two pairs of directives are always present, formed from 

 the third and fourth pairs of the mesenterial sequence. (4) The first four pairs unite with the 

 stomodicum in the order of their appearance {Edwardsia-stnge^, and a long interval elapses 

 before the fifth and sixth pairs become complete; the fifth pair may develop somewhat in advance 

 of the sixth pair. In some cases the fifth and sixth pairs are permanently incomplete. (5) The 

 development of the protocnemes in asexualh^ produced buds is in close agreement with that of 

 sexually produced polyps. 



FIRST CYCLE OF METACNEMES. 



While much attention has been given to the order of appearance of the six pairs of protoc- 

 nemes in the Madreporaria and Actiniaria, comparatively few observations have been recorded 

 with regard to the order of development of the pairs of metacnemes. For the Madreporaria, 

 the establishment of the latter becomes a matter of great importance, seeing that upon it is 

 dependent the order of appearance of the septa, a question already nmch discussed by students 

 of the hard part of corals, but with varying results. 



The transition from the protocnemic to the metacnemic stage of Anthozoan development is 

 one of the greatest morphological significance. Lacaze-Duthiers (1872) was the first to realize 

 this in the Actinia\ and in his resume of the development oiActinhi equina (p. 3(32) he writes: 



" Le nombre, la grandeur, la position et la sjmimetrie des pai'ties ne sontpas determines par 

 les memes lois a toutes les epoques. Ainsi la loi qui preside a la multiplication des parties depuis 

 Torigine jusqu' au nomljre douze, n'est pas la meme que celle qui r%it la multiplication apres 

 que ce chiffre est atteint." 



As regards sexually produced coral polyps, no previous accounts of the actual order of appear- 

 ance followed by the metacnemes are available. The embryological oliservations of Lacaze- 

 Duthiers, von Koch, and Wilson, so far as concerns the polyps themselves, practicallj^ cease with 

 the protocnemic stage. Wilson describes the mesenterial condition in young polyps of Manicina, 

 but gives no account of the sequence according to which the .stages have been reached. 



I have been fortunate in rearing young polyps of Siderastrsea radians as far as the 

 completion of the first cycle of metacnemes, and the various stages in the appearance of the 

 latter have been obtained. Full details will l)e publislicd later, but the diagrammatic figs. 6 {a-c) 

 indicate the actual results. The polyps at fixation contained onlj' the six pairs of protocnemes, 

 as usual, four pairs complete and two pairs incomplete. The fact that the protocnemes retain 

 this proportional development enables the dorsal and ventral, or sulcular and sulcar aspects, to be 

 determined. The polyps remained thus for about a month, the first and second cycles of tentacles 

 appearing in the meantime; then, in the largest specimens, a mesenterial pair was observed 

 within the dorsal exoccele on each side, situated toward the aboral region of the column. A few 

 daj-s afterwards, a similar mesenterial pair appeared within each of the middle exocceles, the 

 dorsal pairs at the same time extending higher up the column. Later, a mesenterial pair was 

 formed within each of the ventral exocceles; so that six new isocnemic pairs were now present, 

 diminishing in vertical and radial length from the dorsal to the ventral side of the polyp, 

 according to their order of appearance. The fifth and sixth protocnemic pairs remained 

 incomplete throughout. 



The six unilateral pairs of mesenteries, of three different sizes, continued their growth j^ri 

 passu with that of the polyps as a whole, and when the latter were about three months old became 

 nearly equal in size, constituting a distinct second cycle. In time, the mesenteries, growing both 

 upward and downward, extended the full IcMigth of the column and partly acro.ss the disk, but in 



