414 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



in a form like IsophyMia, where the septal and costal spines are very variable in the amount of 

 development. The external grooves which separate the verrucal ridges correspond with the line 

 of attachment of the internal mesenteries, and are always smooth. 



Apparently there are no permanent apertures in the column wall of Madreporarian polyps, 

 such as zoophytologists are familiar with in the "Cinclides" of the Sagartids among the Actiniaria. 

 Through these latter the thread-like "Acontia,'' loaded with nematocysts, are extruded when the 

 polyp is irritated. The majority of coral polyps, however, have the power of extruding 

 prolongations of the mesenteries bearing coiled mesenterial filaments along their edge (p. 475), but 

 these can evidently perforate any portion of the superficial tissues, the disk equally with the 

 column wall. Careful examination of the body wall, before the filaments are extruded, fails to 

 reveal any apertures, and their irregular distribution, sometimes over nearly the whole external 

 surface of the polyp, would suggest that the apertures are merely temporary and may be 

 produced at any point. On PI. VIII, tig. (i-f, is represented a section through a portion of the 

 column of a polyp of Orbicella annularis through which the filamental part of a mesentery is 

 extruded. No histological modification whatever can be made out in the wall itself; the aperture 

 is a mere interruption of the layers for the passage of the mesentery and its filament. Upon the 

 polyps settling down after irritation the filaments are slowly indrawn, and ultimately no external 

 indications remain of the apertures through which they protruded. In some cases the openings 

 have been observed to remain distinct for a short time after the indrawal was completed, but the 

 injury, if such it can he regarded, was soon completely healed. 



On full expansion of the polyp the column may extend for some distance above the coral- 

 lum, and is either cylindrical, oval, or irregular in form. Proximally, where it is fixed to the 

 skeleton, it assumes the outline of the individual corallites. and hence may he circular, polygonal, 

 or irregular. In species of Sidt rastrsea and Agaricia the column appears never to he raised much 

 above the general surface of thecorallum. and in forms like Maeandriim. with incomplete polypal 

 separation, the column on both sides rises for many millimeters as a vertical expansion, with a 

 deep valley separating one polypal row from another. 



The form and position assumed by the intercalicular portion of the column wall upon retrac- 

 tion of the polyps varies greatly. In most cases the upper region of the column becomes folded 

 inwardly over the edge of the theca, while in some it is merely drawn downward. In the 

 former condition it either comes to lie inclined downward against the oblique septa {Manicina, 

 Mseandrina, etc.), or, by the action of the endodermal circular muscle, it extends horizontally, 

 terminating in a circular margin which nearly meets at the center, and thus almost covers 

 the disk below (PI. X. tig. 7-1). In Madrepora the wall becomes merely drawn within the calice 

 without any overfolding (PI. I, tig. 2); in Siderastrsea and Agaricia the column and disk 

 are simply depressed, and come to rest upon the skeleton, leaving the tentacles and mouth wholly 

 exposed (PL XXII. tig. 150). 



Variations in the position assumed by the column wall on retraction of the polyps are some- 

 times observable even in the same species. Thus the wall in Pontes clavaria may be slightly 

 folded over the disk, or. as in Sidwastrsea, it may merely come to rest upon the corallar surface, 

 the tentacles and disk remaining exposed (PI. IV, tigs. t'A and 35). 



Among the skeletonless Actiniae the column wall is usually of some thickness, so as to 

 give more or less rigidity to the body of polyp, hut in the Madreporaria, where support is 

 afforded by the skeleton, the polypal wall is nearly always a thin, delicate, often transparent 

 structure. In both groups the thickness of the wall is mainly determined by that of the middle 

 layer— the mesoglcea, as both the ectodermal and endodermal epithelia vary comparatively little. 

 By contrast with that of most anemones the mesoglcea in the column wall of corals is. as a rule. 

 little more than a mere separating lamella between the inner and outer layers, except along the 

 line of attachment of the mesenteries, where it becomes somewhat thickened in a triangular 

 manner. 



The thickness of the column wall is also partly dependent upon the state of expansion or 

 retraction of the polyp. On full distention all three layers become greatly attenuated, the 

 ectodermal and endodermal cells largely diminished in height, and the mesoglcea scarcely distin- 



