MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 417 



better developed, and symbiotic alga; or zooxanthellffi are nearly always present. The cilation is 

 feeble, and rarely detcrniinablo in preserved material. 



The endodernial layer is of niueh the same character throughout the polyp, whether in the 

 column wall, tentacles, disk, skeletotrophie tissues, or forming the mesenterial epithelium. It 

 may vary slightly in thickness in different regions, and in the greater or less preponderance of 

 glandular cells, while in nearly all the species a remarkable modification of the skeletotrophie 

 endoderm takes place in the lower regions of the polyp. The layer here becomes much thicker 

 and loses its distinctly cellular character, appearing finely reticular. So greatly thickened does 

 the endoderm become that it often nearly obliterates the gastro-ccelomic cavity in the most 

 proximal region of the polj'p. The chief constituents — nuclei, cytoplasm, zooxanthellse, and in 

 some cases granular gland cells — are mostly accumulated in a narrow peripheral zone, the deeper 

 portion l)eing vacuolated oi- bearing only fine granules (PI. X, figs. To and 7.5). 



Zooxanthellffi occur in large numbers within the endoderm cells of all the species studied, 

 with the exception of Plu/Uangm ainericnna and Ash'aiuji.a solifar/n. The}^ are usually 

 distributed throughout the polyp, but are more numerous in the exposed tissues (coluuni wall, 

 disk, tentacles) than in the endoderm of the mesenteries and skeletotrophie tissues; they 

 even occur within the internal canals of the perforate genera Mndrepora and Porifes, but are 

 never found free or detached within the polypal cavities except in larvfe. As described on 

 page 437, the organisms are the principal cause of the coloration of many coral polyps. Large 

 oval nematocysts occur in the endoderm of Porltes and Iladrcpora, but are absent from most 

 other genera. Their numl)ers and distinctive form in the genera mentioned are such as to leave 

 no doubt that thej' are actually formed in the endoderm, not free examples injested from the 

 ectoderm. 



The circular eudodei'mal musculature of the column wall appears to ))e always present in 

 coral polyps, as in Actinian polyps, though varying much in the degree of its development; 

 as a rule it is stronger at the uppermost region of the column wall than below. Sometimes 

 the fibrils are scarcely to be found an}' where, while in other species they become strongly 

 developed distally, and give rise to a typical diffuse sphincter muscje, such as is characteristic 

 of man}' Actinite (e.g., Chri/nactis). This is seen in species of Orbicella, especially in the 

 large 0. careriiosa, but also in the smaller O. annularis (PI. VIII, fig. 65). Here, in retracted 

 polyps, the mesoglQ?a is thrown into deep folds for additional support to the musculature. The 

 mu.scle fibers lining the hollows or grooves never become separated from the superficial layer, 

 as happens in Actinians where the muscle is truly mesogkeal. In other species of corals 

 the mesogkea forms only ver}' slight folds, while again it may be perfectly smooth, indicating 

 a very weak muscular development. 



The sphincter muscle is more strongly developed in Isoj>Iii/Jlia. dlpsacea than in any other 

 species here studied. In vertical sections of the uppermost region of the colunm wall the mesogkea 

 displays one or more special thickenings which are much plaited, the whole lined with muscle fibers 

 (PI. XVII, fig. 121). The structure very closely recalls the type of sphincter described by Iladdon 

 (1898, p. 132) as occurring in the Actinian Macmdaetyla^ and there termed a ""restricted"' 

 sphincter nuiscle. It represents a stage of muscular development more complex than that 

 described as "diffuse." The plaitings appear on several axes of greater or less complexity; while 

 in the ""circumscribed" sphincter muscle of Actinian anatomy they are restricted to a single axis. 

 The amount of development of the sphincter umscle is manifestl}' dependent upon the size of 

 the polyp, the polyps of hophyllia and Orhlcella being among the largest studied. 



The action of the circular sphincter muscle is to liring about the overfolding of the distal 

 region of the column wall upon retraction of the polyps. This occurs in nearly all corals, and, 

 as alreadj' observed, it results that the column wall almost completely hides the disk and tentacles, 

 leaving a small central o])ening over the oral aperture. Circular constrictions may occur in flic 

 column wall without any retraction of the disk, in this case the action of the columnar 

 musculature is probably the same as before, but the retractor muscles of the mesenteries have 

 not come into play and di-awn downward the oral region of the polyj). 



(t. H. Fowler (ISSSu, p. 12) was the first to record the presence of an undoubted s[iliincter 



