540 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



"diffuse endodennal," '"restricted oiidoderiniil," "'constricted endodermal," " aggregated, " 

 "single or doulile me.sogla?al,'' and one or another is usually found to be characteristic of 

 Actiniarian families. 



When the polyps of the Madreporaria are taken into account, this structural feature is found 

 to be almost entirely wanting. In luost corals a circular endodermal muscle is present, ))ut it 

 rarel}' undergoes anj' increased development toward the apex of the colunui. such as can be 

 regarded as constituting a special sphincter. In the lai'gor polyps, like those of Orhiccllii and 

 I.'iojjInjUla,A slight concentration of niuscle fibers takes place, but only deserving of the title of 

 "diffuse endodermal muscle," which represents the simplest form of sphincter development 

 (PI. VIII, hg. tio). In IsophijlJiii (PI. XI. tig. 121) the mesogheal processes supporting the 

 musculature Ijecome a little more thickened and branched, and the whole structure mav perhaps 

 be regarded as having attained the next type of nuiscular complexity, that known as the 

 " restricted endodermal." 



The mesenterial nuisculature, likewise, presents no important differentiations in the coral 

 species studied. Different degrees of development of the longitudinal retractor muscle are 

 indicated by more or less deeply folded or l)ranched mesoglcjeal plaitings, but present none of 

 the variety of form met with in the Actinia'. The basilar muscle is absent, and the parieto- 

 basilar appears to be the same. 



The nature of the column wall and disk in Madreporarian polyps, likewise, affords few distin- 

 guishing features. It is in all cases devoid of the simple or complex outgrowths, suchasacrorhagi, 

 adhesive or spheroidal verrucfe, which characterize many genera and even families of Actiniaj. 

 In the more or less retracted condition, under which coral polyps generally will be studied, one 

 can merely distinguish forms with a smooth surface, as contrasted with exteriors which are 

 verrucose, the latter condition a result of the costal or septal denticulations upon which the 

 tissues come to rest. 



Among the Actiniaria, Carlgrcn (1893) has employed the presence or absence of ciliated bands 

 to the mesenterial filaments, and the presence or absence of an ectodermal musculature and 

 ganglion layer on the colunui wall, as features of diagnostic and phylogenetic importance. As 

 already shown, the mesenterial filaments of all the Madreporaria are alike in the absence of 

 true independent ciliated bands, and any nuiscular or nervous elements with which the ecto- 

 derm of the column wall may be provided do not form distinct layers. 



On the other hand, certain polypal characteristics occur within the Madreporaria which are 

 either wanting or do not assume much importance within the Actiniaria. Among these mav be 

 mentioned the resulting form of the colony due to the method of vegetative growth, the canal 

 system of perforate corals, the presence of S3'napticular perforations, the se|jtal and other invag- 

 inations of the basal disk, and the extrusion of mesenterial filaments along with the mesenteries 

 to which they are attached. All these will be found to be of systematic importance in polypal 

 studies, though not attaining the value assigned them in works concerned only with the skeleton. 



The canal system of the Porose coi'als, representing as it does merely complicated outgrowths 

 of the liasal wall, has lint little morphological significance, though modifying the corallum and 

 polypal tissues profoundly. The importance assigned the presence or absence of the canal 

 systems in the classificatory scheme of Milne Edwards and Haime has been of great utility, but 

 the character has not that fundamental value which one would desire for a primary division, and 

 can not take precedence of the mesenterial and septal arrangement in any natural system. The 

 same must be said also of the synapticula, which are the chief characteristics of the Fungacea. 

 They are merely skeletal growths connecting adjacent septa, but in the genera Sldernatrpea^ 

 Agaricia, and Fiingid they are certainly associated with peculiarities in the form of the tentacles 

 and their wide distance apart, hence their presence may be indicative of some deeper natural 

 relationship. 



I have thus ))riefly indicated the anatomical and histological features in the species of Madre- 

 poraria here studied which are available for purpo.ses of founding a more natural classification 

 than those yet proposed. It must be admitted that twenty-six species, distributed among twenty 



