MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 459 



A wide distinction in tiieir niaiiuer of appearance thus separates the members of tlie second 

 order of mesenteries from those of the first order. The primary mesenteries appeal- in hilatei^al 

 pairs, in a succession which is first toward one aspect and tlien toward the other aspect of 

 the polyp, and so on, and only later do they constitute unilateral pairs, in which the musculature 

 is on the faces turned toward each other. With the exception of the directive pairs, the two 

 members of each unilateral pair arise at different times, the dorso-lateral pairs being constituted 

 of mesenteries II and V, and the ventro-lateral pairs of mesenteries I and VI, in the protocnemic 

 sequence; and for a long period the lateral pairs are anisocnemic. The secondary mesenteries also 

 arise in a bilateral manner, but are in uitllateml {hocnem ie) jxii I's from thehHjiiuving, and in any 

 polyp the\' are formed in only one succession, from the dorsal to the ventral aspect, alternating 

 with the primary pairs, and situated within the primary exocceles. In mature polyps the secondary 

 mesenteries are all equal, except perhaps in their vertical extent, and are arranged around the 

 polypal wall with perfect hexameral radial symmetry, all traces of their bilateral succession 

 being lost. 



Where coi-al polj'ps attain consideraiile size, as in Orhieella careniona and PhyUangla 

 ainei'icana, the members of the second order of mesenteries often become united with the 

 stomodffium. In doing so they follow the same antero-posterior succession as that characteristic 

 of their order of appearance in the young polyp (see fig. 9A, p. i64). 



SECOND CYCLE OF METACNEMES. 



The order of appearance of the second cycle of metacnemes, or third order of mesenteries, 

 may now be considered. Tliese, wlien complete, consist of twelve equal pairs, a pair within 

 eacli of the exocQ?les between the protocnemes and the first-cycle metacnemes. The succession 

 has not been followed upon any coral polyp reared directly from the larva, but sufficient evidence 

 is forthcoming from the asexually produced polyps of several species to indicate in a general 

 waj' the manner in wliich it is carried out. 



Mature polyps of Cladoc<»^a arhuscula and Astrangia solitaria usually contain a certain num- 

 ber of second-order metacnemes, but apparently never the full complement of twelve pairs. It 

 is therefore possit)le to obtain from these cei-tain intermediate stages in the establishment of the 

 cycle. PL VI, fig. 49, represents a section through the stomoda?al region of a polyp of Gladocora 

 with sixteen pairs of mesenteries. Of these the six complete pairs are jirotocnemes, the six alter- 

 nating pairs are first-cycle metacnemes, and the four pairs remaining are the only representa- 

 tives of the second-cycle metacnemes. The latter are but feeblj^ developed, and without 

 mesenterial filaments. The fact of greatest importance, in comiection with the four new pairs 

 of mesenteries, is their restriction to only one exocoele within each of the six primar\' systems; 

 they are not developed in both the exocceles within the two ventral sj'stems, as considerations of 

 symmetry would suggest. It will be also obsei-ved that in each case thej- occur within the 

 exoccele on the dorsal asjaect of the second-order mesenteries {ef\ p. 4.58). 



The polyps of Cladocora arhuscula very rarelj^ pass beyond the stage with- sixteen mesen- 

 teries, whicli corresponds externally with thirty-two tentacles. For the further mesenterial 

 sequence therefore other species will be employed. 



All the members of an isolated group of eight separate pol_yps of Astrangia, solitaria were 

 decalcified and sectionized, and the stage reached in the mesenterial development of each is 

 diagrammaticalh' represented in figs. 8 {a-g). Camera lueida drawings of a transverse section 

 from two difl:erent individuals are also given on Pis. V and VI, figs. 43, 47. The seven 

 diagrammatic figures reveal that no two polyps in the group were alike in their mesenterial 

 arrangement, so that the series may be taken as affording a fairly complete representation of 

 the order of mesenterial development generally followed in this species. 



Fig. 8« is taken fi-om the smallest of the poh'ps. In this instance three members of the 

 protocnemes are still incomplete, and only four pairs of metacnemes have yet appeared, situated 

 within the dorsal and the middle primary exocceles. The sequence of the first-cycle metacnemes 

 is evidently similar to that of the polyps of Siderastrxa above described (p. 45(!). In the next 



