MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 539 



1865) and Kuiith (1869-70), and i.s regarded by Neuniayr (1889) as a characteristic of the 

 highest morphological significance. The bilateralit}' in many forms is only clearly developed in 

 the early stages, but the external surface of corallites often exhibits pinnate streaks or i-idges 

 which correspond with the internal septa, and these deoionstrate conclusively the bilateral manner 

 in which the septa have arisen, e. g., Sfreptelmnia, Zajihrenfis. 



In the Cj'clocnemaria the mesenteries are as a rule arranged in two or more regular cycles, 

 while in the Entocnemaria they are in one cycle only, with merely alternately large and 

 small members. In the absence of any knowledge of the soft parts of a coral we maj- assume 

 that in general the septa of the former group will be pol^'cyclic, while those of the latter will be 

 monocyclic or at most dicyclic. Where asexual repi'oduction by incomplete discal hssion prevails, 

 there is a tendency toward the dicyclic condition, but these forms can be readily distinguished 

 from the Paheozoic types with merely large and small alternating septa. 



With regard to the further emplo5'ment of the mesenteries and septa of Madreporaria for 

 taxonomic purposes, the wide differences in the ai'rangement and nature of these organs in 

 gemmiferous and tissiparous genera may now be considered. In mature polyps of the former it 

 has been found that a cyclical regularity prevails, and two pairs of directives are alwa3's 

 present, but in the latter the introduction and continuance of fission carries with it marked 

 changes, not only in the mesenteries, but also in the septa and tentacles. No other polypal 

 characteristic seems to exert such a profound influence upon the nature of the compound coral 

 as a whole. But by no means can the results of fission be regarded as of such fundamental 

 significance as those distinguishing the Entocnemaria and Cyclocnemaria. The young polyps of 

 both gemmiferous and fissiparous genera are built upon exactly the same plan, and it is only with 

 the advent of vegetative reproduction that they liecome divergent. 



Fissiparity would appear to be a condition which may arise in any group of corals, and its 

 occurrence does not necessarily indicate any natural relationship among the forms in which it 

 prevails. In anj' classiticatory scheme it can probabl}' be regarded as onlj' of subfamily impor- 

 tance, which is practically the position assigned it bj^ Duncan." In this case the divisions, in 

 whatever families the}- occur, may be defined as follows: 



Geminantes. — Asexual reproduction takes place hy gemmation, and each polyp represents a 

 distinct individual. The tentacles, mesenteries, and septa are arranged in alternating cycles, and 

 two pairs of directive mesenteries are present in each polvp- 



FtssqxuxDdes. — Asexual reproduction takes place by stomodiwal fission, without the production 

 of morphologically complete polj'ps. The tentacles, mesenteries, and septa, after fission is estab- 

 lished, are not arranged in regular alternating cycles, and no new directive mesenteries arise. 



The arrangement and form of the tentacles in the Madreporaria can not attain that 

 systematic value which they possess in the Actiniaria. In the latter the origin of one or more 

 tentacles from a single mesenteric chamber is a character of nuich importance, and afl:'ords a 

 means of dividing the Hexactinite into the two suborders Actininte and Sticliodactylina?. In the 

 corals no instance of the stichodactylinous condition has been met with, and an}' other tentacular 

 characteristic so far disclosed seems worthy of onl^^ generic, rarely of family, recognition. With 

 the exception of the bifurcated entoc(£lic tentacles in the single geiuis Slderaxtrira, the organs 

 are invariably simple in corals. As a rule they are arranged in close, alternating, entacmreous 

 cycles, but in the Fungidic the cycles are distant and tend to lose their regularity of disposition. 

 The prevalence of the knobbed or swollen tentacular apex in corals is noteworthy, considering 

 how rarely it occurs in the Actiniaria (e. g., Corynactis, Riconlea, CornUhnorphnx). Tentacular 

 introversion is probabl}' very general throughout the Madreporaria, but rare among the 

 Actiniaria. 



The sphincter muscle is another structure which the Herwigs first brought into prominence 

 as an aid in the classification of Actinian polyps. It occurs toward the apex of the colunui wall 

 in nearly all anemones, and, next to the arrangement of the mesenteries and tentacles, occupies an 

 important place in all Actinological .studies. Various types of sphincter are recognized, such as 



« For later results, see foot-note, p. 541. 



