546 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



Ill the same way the, hirge anterior teutaele of radial pol^'ps is nearly colorless, as is also 

 the case with the small tentacle on each side of it, and less so the next tentacle. In a partly 

 retracted condition the tentacles are often darker than any other region of the polyp, the 

 endoderm heino- thickened and its zooxanthelhe crowded; when fully inflated the color of the 

 polyps as a whole becomes paler. The absence of color toward the margins of the colonies prob- 

 ably indicates tliat these are regions of rapid growth, the polypal growth being in advance of 

 that of the multiplication of zooxanthelhe. 



In no case have an}' traces of ectodermal coloration been found. This uniformity of color 

 of the West Indian Madrepoi-a is in marked contrast with the vivid and varied tints described 

 by Saville Kent for most of the species of the genus occurring on the Australian Barrier Reef 

 (1893). 



Anatomy and lihUiJogy. — The column wall is everywhere very narrow, in sections measuring 

 0.08 mm. across. The ectoderm is constituted largely of unicellular oval gland cells with clear 

 contents. Sometimes the contents of the gland cells stain slightly, and in sections the mucus 

 can often be seen extruded, forming an irregular outer covering to the ectodermal cells. Small 

 nematocysts of two kinds occur somewhat sparsely — a long, thin-walled form in which the spiral 

 thread is distinctly visible, and a smaller thick-walled oval variety in which the internal thread 

 is scarcely recognizable. At the apex of growing branches the ectoderm cells are often much 

 longer than elsewhere, measuring 0.07 mm. (figs. 7-9). 



The mesoglcea is throughout a very thin supporting liuuella, but wherever it attains much 

 thickness it is found to be clear, transparent, and homogeneous, without connective-tissue cells. 



The endoderm varies somewhat in character, according as zooxanthellre are present or absent. 

 At the tips of branches, which in the living condition are colorless, the endoderm of the column 

 wall is a very narrow layer, the cells liut little vacuolated, and the nuclei comparatively large and 

 somewhat regularly arranged (fig. 8) ; but where the symbiotic algse occur the layer is broader and 

 the cells more vacuolated. All gradations can be traced between the total absence of the alg.e 

 and their presence to such an extent as to constitute nearly the whole layer. In radial sections 

 of the column wall are seen the cut ends of delicate endoderraal muscle fibrils, arranged in 

 a circular manner, and forming a very thin muscular layer, which extends the whole length 

 of the wall, and is continuous with the circular endodermal musculature of the tentacles. 

 This diffuse endodermal musculature probably acts as a sphincter during the retraction of the 

 polyps; there is, howevei-, no concentration of the muscle fibrils on mes;)gl(eal jilaitings. such 

 as can be regarded as forming a special sphincter muscle; the mesoglujal surface remains smooth 

 throughout. 



Histologically the outer covering of the skeleton differs in no essential respects from that of 

 the column wall proper, the two being merely continuations of one another. Where the wall 

 rests upon the echinulations, continuity with the skeletotrophic tissues is established, and the 

 endoderm and mosoghea of the two pass into one another. Th(> figures show that the actual apex 

 of the echinulation is covered by its own skeletogenic ectoderm, the mesogla-u. and the outer 

 ectoderm. No muscle fibers are determinable in the coenosarcal endoderm, such as occur in the 

 free portion of the column wall. 



In sections toward the apex of axial polyps, the corallar ridges are often unprotected 

 by soft tissues. This is probably due to the very tliin walls liaving broken down during 

 decalcification, but in most carefully prepared material no remains can be found, and the very 

 broad ectoderm overlying the canals j)asses inwardly as the calicoblast l.iyer. 



The tentacles in strongly retracted radial polyps often ai)ii('ar as mere longitudinal ridges 

 of the wall of the polyps, and neither in longitudinal nor ti'ansverse sections is any part free 

 from the disk. ' In other cases, however, isolated circular sections of the tentacles are obtained, 

 showing tliat (luring retraction the organs may retain their <listiiictness as outgrowths of the disk. 

 Longitudinal sections ])resent no sharp line of separation between the upper region of the column 

 -wall and the tentacles. The transverse section through the tentacular region, represented in fig. 8, 

 exhibits the six larger and six smaller tentacles all at the .same level, and nearly filling the calicinal 

 eavitv: thev are outgrowths of both the entoccelic and exoccelic chamliers. the larger tentacles 



