MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 475 



demonstrate how .similar, higlily specialized tissues may recur in different regions. The marked 

 dov(>iopment of the hiteral ciliated area in ilddrepora may be conceived as associated with the 

 complex circulatory system of the porose corals, but the weakness of the filaments in the allied 

 genus Parifen scarcely bears out such a suggestion. 



Mature polyps of Puritea usual!}' contain four pairs of complete mesenteries in the upper 

 region, but the dorsal directives often become free before the lower termination of the stomo- 

 d;eum is reached (tigs. 80, 41). Mesenterial tilaments, however, are found on only the three 

 remaining pairs of mesenteries, I, H, III; the free edge of pairs IV, V, VI is covered with 

 the ordinary mesenterial epithelium (PI. Ill, tig. 29). For a short distance below the stomo- 

 danun the tips of the older mesenteries are provided with a deeply-staining tissue, in no ways 

 distinguishable from the stomodseal ectoderm with which it is continuous (PI. IV, tig. 38); 

 and no shai'p boundary line here separates the filament from the rest of the endodermal epi- 

 thelium. Some distance lielow the stomodieal region, however, the filament takes on the normal 

 character, and lateral endodermal lobes may be formed on the hrst and second pairs (tig. 39), 

 though they persist for a very short vertical distance. 



The limited development of the mesenterial filaments in Porites, on only two or three of the 

 pairs of the mesenteries, is in close agreement with the results of Fowler (188S) upon polyps of 

 Seriatopom. In ^S'. suhulata Fowler found the mesenterial filaments to be well developed on only 

 one pair of mesenteries, the two marked 3 and 10 in the author's notation, and corresponding with 

 the pair marked I, I in the present paper; the mesenteries numbered 1, 5, S, Id, corresponding 

 with pairs II, III, wei-e generally devoid of any '"filamentar" thickening. The proportional 

 development of the filaments thus corresponds with the order of appearance of the mesenteries, 



EXTRUSION OF MESENTERIES AND FILAMENTS. 



In corals reproducing by gemmation the filaments are i-arely so strongly developed as in 

 tissiparous species. In the latter, certain of the filaments ai-e more important than others, and 

 t)ecome greatly folded and convoluted, attached to the free edge of the mesentery all the way 

 (PI. XXII, fig. 1-48). The mesenteries bearing such strongly developed filaments are capable of 

 partial extrusion through the walls of the polyp, either upon irritation or injury to the latter, 

 sometimes in such quantities as to nearly hide the surface of the colony. Extrusions may appear 

 at any part of the column wall or disk, as well as through the mouth. In the living polyp the 

 mesentery and filament sent out are easily distinguished one from the other; the former is 

 usuallj^ thin, colorless, and transparent, while the latter is opaque white, and disposed in irregular 

 loops and coils. The extruded mesenteries in some species are faintly green in color, perhaps 

 due to the large number of zooxanthelliB in the endodermal epithelium. In polyps preserved 

 with the mesenteries thus partly extruded, some of the filaments are also found displaced within 

 the upper polypal regions, and even within the chambers of the edge-zone. In Cladoeora the 

 filaments have been observed to enter the tentacular cavities, and occasionallv they are found in 

 the perithecal continuations of the polypal cavity. 



The phenomenon takes place most readily in fissiparous genera, upon strong irritation or 

 after rough handling of the colony, and the extrusions are more copious in these forms than in 

 genera reproducing by gemmation, but probably there are few corals in which it may not occur 

 occur to a greater or less degree. It has, however, never been observed in the numerous colonies 

 of S'idi'mstrxa kept under observation. 



Examination of the column wall and disk of the polyps, under ordinary' conditions, fails to 

 reveal any apertui'es comparable with the " Cinclides " of the Actiniaria, through which it may be 

 supposed that the hlaments can pass. The absence of cinclides. and the irregular disposition 

 of the extrusions over an\' part of the column wall and disk, make it evident that the openings are 

 merely temporary, and capable of formation at any jioint, structural continuity being again 

 established when the mesenteries are indrawn. On examining the surface of the column wall, 

 immediately oh withdrawal of the filaments, the apertures could lie observed, and have been found 

 to remain open for a short time; gradually, however, they close over, and all evidence of their 

 former presence is lost. 



