MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 429 



two illustrate the same fact. In l><>th species studied six pairs of perfect mesenteries form the 

 first cycle, six alternating pairs make up a second cycle, and there may be twelve pairs forming 

 a third cycle. As just mentioned, however, this last cycle is rarely completed. Further, an 

 examination of the macerated skeleton shows that in very few instances is the full complement of 

 septa, viz.. 6, 6, L2, .'4. present. Usually in S. radians only a few pairs of th • third-cycle 

 mesenteries occur, the number varying with the size of the polyp, while in the larger S. siderea 

 nearly all the pairs are present, and even some membersof a fourth cycle. 



This incomplete cyclic development in the case .if the mesenteries is repeated in the last cycle 

 of entocoelic tentacles, and. the organs being widely apart, the imperfection of the cycle becomes 

 more pronounced externally. With few exceptions the hexameral plan can be traced only as far 

 as the first and second cycles. The third cycle may comprise any number of members from one 

 to twelve, while the outermost cycle of simple tentacles contains the sum of the members of all 

 the three inner cycles. Further, there is a tendency in most species of corals for the two inner 

 cycles to constitute but one cycle of twelve, in the same way that as the polyps increase in size 

 the mesenteries of the second cycle tend to unite with the stomodseum, and the first two orders 

 of septa form only one cycle. 



A considerable discal space intervening between the different tentacular cycles in Siderastrasa, 

 as compared with most other corals, it is clear that the two conditions alluded to above find their 

 outward expression in individual tentacles occurring at varying distances from the center of the 

 disk, and thus giving rise to the characteristic irregularity. In a fully developed, long-estab- 

 lished polyp, the cycles are more regular than in a young individual. Moreover, were the cycles 

 of tentacles in other coral species to be separated by such comparatively wide discal interspaces, 

 instead of being arranged closely in a narrow zone, similar cyclic irregularities would be more 

 generally noticed. 



The arrangement of the tentacles in Agaricia very closely resembles that characteristic of 

 the genus Siderastrseq, but the organs are never bifurcated, and are not distinctly stalked. They 

 remain exposed during the retracted condition of the polyp, and during ordinary retraction 

 can usually be seen as mere pointed or triangular tubercles, but when expanded they become 

 more digitiform, with an opaque white area at the apex. They are often brightly colored 

 by comparison with the rest of the polypal wall. In several colonies of the form I identify 

 as Agaricia fragilis, I was unable to determine the presence of any tentacles in the living 

 condition, even with the aid of a lens. In sections through the disk they are. however, 

 recognizable as slight, nematoevst-bearing thickenings of the ectoderm (PI. XXIV, fig. 163). 

 The organs are better developed in Agaricia agaricites. 



As in Siderastrasa, the individual tentacles are widely separated from one another, and are 

 distributed over nearly the whole discal area, one above the apparent centripetal termination of 

 each of the larger septa. No tentacles occur over the members of the smallest cycle of septa, 

 which transverse sections demonstrate as exocoelic. In this absence of exoecelic tentacles the 

 genus Agaricia is unique among the forms here studied, with the exception of the axial polyps 

 of Madrepora. 



The majority of the tentacles are arranged so as to form an inner cycle, but the number 

 composing it is variable, and the cyclic character is only approximate. Outside there are a few 

 scattered examples at different distances from the center, suggesting no cycle relationship. 

 The number in the inner cycle varies from five to nine, while the total number in any polyp may 

 be from thirteen to twenty-four. 



In Agaricia mesenterial increase appears to be in constant progress, corresponding with the 

 growth of the individual polyp, though in no regular cyclic manner. Similarly with the 

 tentacles: the inner cycle includes all the older tentacles, and outside this are the. later-formed 

 members which appear irregularly. Probably it is best to regard the tentacles as acyclic, no 

 exotentacles being developed. Counting the tentacles of many polyps gives odd numbers as 

 often as even, while in the case of species with exocoelic tentacles even numbers predominate. 

 The irregularity in the disposition of the tentacles in Agaricia should be compared with the 

 irregular arrangement of the mesenteries represented on PI. XXIV, fig. 161. 

 Vol. s— No. 7 3 



