3IEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 485 



GASTRO-CCELOMIC CAVITY. 



The term gastro-ccjelomic is applied to the whole of the iiitenial, endoderm-litied cavity of 

 coral polyps, including aii}"^ outgrowths or continuations which it may possess. In some ways 

 the designation is preferable to the terms gastro-vaseular cavity or ccelenteron, generally 

 employed for the polypal cavity in Anthozoa. For from the considerations of van Beneden 

 (ISyl), and E. B. Wilson (ISS-t), there seems some evidence to support the view that the space but 

 incompletely inclosed by the mesenterial filaments is the morphological equivalent of the gastric 

 cavity, or enteron, of the higher Metazoa. while the remainder of the internal cavity, partitioned 

 by the mesenteries, is the morphological representative of the c«?lom of the Enterocuela. 



Among colonial corals the gastro-cceloraic cavities of all the polyps in actual union with one 

 another are in commimication. and the nutrient fluid can pass from one to the other. This is a 

 persistence of the conditions consequent upon asexual development, whether this takes place by 

 gemmation or by lissiparity. In gemmation new polyps arise whollj' or in part from the column 

 wall of other polyps, and the internal cavities of the two are common for a time. Thus the 

 developing bud of Mudrepora, shown in the series of figures on PI. Ill, arises altogether from 

 the ccenosarcal wall of the colony, and its ccelomic cavity during the primary stages is repi-e- 

 sented by one of the superficial canals of the colony. In the develoi:)ing polyp of Solenastrxa, 

 represented in section in PI. XII, fig. 87, a distinct partition wall, lined with endoderm on 

 both sides, now partly sejiarates the bud from the parent; but interruptions occur at more or 

 less regular intervals, which permit of a circulation between the two cavities. 



The mode of communication of the various polj'pal cavities in a colony varies somewhat in 

 different forms. In genera like Porites, Siderastrsea, and Agaricia, in which the poly^DS are 

 separated from one another merely along a common calicinal wall, intermesenterial apertures 

 i-emain along the line of union, while the polyps are partitioned mesenteriallj'. In Sidermtrxa 

 septal partitions also occur, at any rate during retraction; for peripherally the column wall 

 comes to rest directly upon the septal covering, so that only a very narrow space is left on each 

 side between the mesenterial and the sept;d wall (PI. XXIII, fig. 156). The channels of commu- 

 nication of four adjacent polyps of Porites over the thecal edge are represented on PI. Ill, fig. 31, 

 taken from a section through the superficial region of a colony in which the polyps were in a 

 retracted condition. The fragments of the corallum seen are the slightly exsert septa, and the 

 canals pass over and around them. 



Adjacent polyps of Orhicella and Solenastrxa are also placed in communication intermesen- 

 terially at the superficial line of union of the polyps. During retraction the apertures are not 

 connected directly with the main cavity, but through the intermediation of its perithecal 

 prolongations. The same method of superficial intermesenterial communication holds for the 

 contiguous rows in the genera reproducing by incomplete discal fission, e. g., Maiandrina 

 (PI. XX, fig. 138), Colpophyllla^ Manichia, hophyUia. In these, however, the polyps which 

 are united in the same discal system have no independent cavity, the one continuous chamber 

 is shared in common. The poh'ps do not attain true individualit\-; they can best be understood 

 (IS so many mesenterial and stomodff'al systems within a general cavitj'. 



The numerous polyps constituting a colony of Madrcpura are likewise in communication by 

 means of the superficial canals, which are continuous over the thecal edge with«the main gastric 

 cavit}' (PI. I, fig. 2); but in the great group to which Madrepora and also Porites belong — the 

 Porosa — there is another and more conjplicated system of comnmnication than that afl'orded by 

 the superficial canals. Anastomosing radial canals are given off by the basal (skeletotrophic) 

 part of the body wall, in such a way that they appear as if penetrating the corallum which 

 separates one polyp from another, and thus place the different ccelentera in union; in Jfadrepara 

 they further come into communication with the superficial canals. 



The radial canals are given (jff very closely in both Por'ttca and Mndrepora^ and are disposed 

 both mesenterially and intermesenterially, without any apparent regularity. As many as five or 

 six may be seen in a single transverse section (PI. I, figs. 3-6). Thej' are not so numerous in the 



