486 MEMOIRS OF THE NATIONAL ACADEMY OF SCIEN<JES. 



upper region of the apieal polyps of 2Tadivj)or<( as below. In tig'. •!. rej)rcsenting an apical polyp, 

 no radial opening-.s occur on either .side, though an uninterrupted continuit}' of the corallum for 

 such a distance appears to be somewhat unusual. The superficial canals in exsert corallites of 

 Mddrepora are mostly longitudinal in direction: 1)ut transverse connections occur, and the canals 

 may be also interrupted by skeletal growths. Decalcitied preparations show that the canal 

 system does not as a rule prolong the gastro-ccelomic cavity at the aboral end of the polyps; in 

 both Por'ites and Madi-i-pimi the polyps and their canals are abruj^ly ti-uncated. 



The internal canal system in the Porosa is morphologically basal in origin. For it is 

 established that the whole of the skeletotrophic tissue is derived from the primary basal disk, 

 and all its cvaginations and invaginations are but so n)any foldings and complications of the 

 walls of this region of the polyp. The canals of the perforate corals ma_y therefore be compared 

 M'ith the basal communicating canals of colonial Actiniaria, such as those of Pulythoa among the 

 Zoanthese. Though exerting a profound influence on the character of the corallum, the canal 

 system of the Porosa has but little morphological significance, and in any natural classification 

 of corals appears altogether unworthj^ of the importance which has been assigned it by 

 systematists. 



The fundamental ditl'crence between colonial non-perf orate and perforate corals may be 

 thus stated: In the Aporosa the gastro-ccelomic cavities of the component poh^ps of a colony are 

 in communication oidy by superficial apertures along the c-ommon line of union of the column 

 wall and base, while the Porosa have in addition a means of comumnication bj' basal anastomosing 

 canals. 



Compared with that of the Actinian polyp the gastro-ctelomic cavity of the Madrepoi-arian 

 polyjj is much more subdivided and intruded upon, both radially and peripherally, ijarticularly 

 in its lower region. In addition to the mesenterial partitions, shared in common with the 

 Actiniaria, a like number of radiating septal invaginations usually occurs, while columcilar. 

 spinous, and synapticular productions still further break it up centrally and peripherally. 

 Moreover, the calicinal wall itself is often produced upward as a peripheral, circular wall. 

 and, so far as it extends lievond the line of union of the ct)lunin wall and base, divides the 

 coelomic cavity, as well as its mesenterial partitions, into inner and outer moieties, the latter 

 constituting the cavity of the edge-zone or Randplatte. 



During expansion the upper part of living polyps is elevatinl for some distance wholh' 

 beyond the corallum. and in distinct polyps the free portion is cylindrical, in form closeh' 

 recalling an Actinian polyp (figs. -16, 48). Here the subdivisions of the gastro-ccelomic cavitj' 

 are onlv mesenteric, and are arranged periphei'ally into entoc(i>li<' and exoccelic chambers in a 

 strictly Actinian fashion. The fleshy parts of fissiparous genera likewise become extended 

 for several millimeters, the oral disk appearing as a meandering platform fringed by the zone of 

 tentacles, and the column wall on either side as a nearly vertical sinuous parapet: thera\ity is 

 divided into entoccelic and exoccelic chambers, l)ut not with any cyclic regidarit}'. 



In retracted polyps not only is the oral region withdrawn within the calice, but the 

 perithecal wall becomes more nearly apposed to that innnediately coxering the skeleton, and 

 closeh' reproduces the outer corallar form, thus largely obliterating the intervening gastric space. 



Proximallv the polypal cavity extends some distance within the corallum. or rather the 

 corallum has intruded within the polyp; and in these lower regions the cavity becomes 

 subdivided in a most complex manner by skeletal ingrowths. In addition, the skeletotrophic 

 endoderm becomes greatly thickened proximally, and further encroaches upon the chambers, 

 but apparent!}' the mesenteries always cease before the lower termination of the polypal 

 cavity is reached." 



a Nothing like the peculiar obliteration of the polypal cavity which Sclater (1886) describes in Stephanotrochus 

 has been encountered. In all cases the crelomic cavity persists as far as the proximal floor of the polyp. According 

 to Bourne (1893, p. 219), the polyp in Fungia does not desert the lower part of the calice, but remains adherent to 

 tlie basal plate. Yet even here Bourne finds that the primary and secondary mesenteries are carried upward as 

 growth proceeds, and are confined always to the upper moiety of the calice; the lower moiety consists only of 

 chambers lined with endoderm and undivided by mesenteries. 



