MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 425 



drawn within the mesenterial chambers of the polyp, while the tentacles remain directed 



normally outward. The introverted disk in von Heider's figures presents in transverse and 

 longitudinal sections much the same appearance as the introverted tentacles of Pontes in figs. :'>o 

 and 40. that is. the ectoderm is internal and the endoderm external. In many instances of 

 strongly retracted polyps the tentacles are found greatly depressed or introverted as integral 

 parts of the disk. Discal infolding is noticed more fully on p. 434. 



From all these examples it is manifest that the phenomenon of tentacular introversion in 

 both expanded and retracted polyps, and of disappearance in the discal wall of fully expanded 

 polyps, are very general among corals. They probably serve to explain the statements of some 

 of the older observers that tentacles arc wanting in certain species of corals. 



(i. von Koch (1890, p. 399) has found in the contracted polyps of the Alcyonarian, Rhizoaa n'm 

 rosea, that, in addition to the infolding of the disk and upper part of the column, the tentacles 

 undergo invagination, but only for about half their length; the proximal half still preserves the 

 normal relationship of outer ectoderm and inner endoderm. This is undoubtedly similar to the 

 process described above, only the introversion is not continued to the extreme limit, as in Porites. 

 In the living expanded polyps of the coral Astroides cafo/cularis, von Koch has also observed that 

 the terminal part of a tentacle is often drawn inwardly toward the basal part, and again pushed 

 out. the movements somewhat resembling the drawing in and pushing out of a telescope tube, 

 and continuing for some time. 



An explanation of tentacular introversion does not seem readily forthcoming, for beyond 

 the usual ectodermal longitudinal and endodermal circular fibers no special musculature is dis- 

 coverable whereby the movements may be produced; further, a decided individuality is exhibited 

 by the various members comprised in the cycles. One may surmise as a cause a difference in 

 the hydrostatic pressure between the internal cavity and the exterior, owing to variations in the 

 circulation of the nutrient fluid within the colony. But this would not account for the fact that 

 the polyp itself may remain fully expanded, and only certain of the tentacles be invaginated, 

 while the others remain extruded. 



The disappearance of the tentacular walls in the discal tissues seems more easy of explana- 

 tion. Structurally the tentacles in the Madreporaria are rarely the important differentiated 

 discal outgrowths which they have become in most Actiniaria, and when the polyps attain their 

 full expansion it can readily be understood how the tentacular walls may become involved in the 

 discal expansion, and lose the distinctness of their walls, the thick apex only remaining to indi- 

 cate their former presence. The tentacles are originally outgrowths of the disk, and can again 

 become part of it, the thickened apical knob remaining as the only evidence of a special differen- 

 tiation. 



Histology. -Histologically the walls of the tentacles present few characteristics which do not 

 occur in the column wall or disk. Such peculiar features as they display have reference to 

 their function as stinging organs. Transverse or longitudinal sections of most species exhibit 

 marked inequalities in the thickness of the ectoderm, the broader regions representing special 

 uematocyst areas. The thickenings correspond with the more opaque areas on the tentacular 

 walls in the living condition, and are best seen in sections made from tentacles in the expanded 

 condition, as in the retracted examples the wider nematocyst regions tend to overlap the inter- 

 vening narrow areas (PI. VI, tig. 50; PI. X, tig. 75). The largest battery of stinging cells is 

 at the apex, and here the outermost zone is constituted almost wholly of uematoblasts. By 

 focussing with a high power around the free edge of a nematocyst area, triangular or thread-like 

 cnidocils can usually be discerned, especially in the living tentacle, and cilia may be present over 

 the whole tentacular surface (PI. II. fig. L0). 



The nematocysts in the tentacles are mainly of the long, narrow, thin-walled form, with the 

 spiral thread closely coiled (PI. XVII. fig. 124<z). Other thin-walled forms — small and oval, or 

 large and oval with a loose spiral thread— may occasionally occur, but are never so oharacteristic 

 as the former. In the deeper parts of the ectodermal layer, brightly staining, apparently homo- 

 geneous bodies are generally seen, which represent nematocysts in various stages of development. 

 At first they are irregularly arranged at almost every angle with the surface, but as they reach 



