MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 483 



Between the calicoblast layer and the actual skeleton, Bourne has found what he terms a 

 ViinHlnij meiiihrane, separating the polypal wall from the calcareous matter. This occui's 

 wherever decalcification of properly preserved material is carefully carried out, but to my mind 

 represents the remains of the colloidal matrix in which the skeleton is deposited. At the growino- 

 apex of Madrepura a contimious ground substance remains after decalcification, and tills the whole 

 of the space occupied b\' the corallum, behaving toward reagents exactly like the mesogloea 

 between the ectoderm and endoderm. It presents a striated, scale-like appearance, altogether 

 similar to that of the calcareous libers of the skeleton of 2Ladi'epora (tigs. 18, 19). The strife have 

 manifestly been produced by the calcareous skeleton laid down within a perfectly homogeneous 

 substance. This latter is evidently secreted by the calicoblasts, but only under the most 

 favorable conditions, as near the actual tip of rapidlj' growing branches, can it be found persisting 

 throughout the skeleton. Elsewhere the organic matrix has either wholly disappeared, or is so 

 insignificant as not to persist in a continuous form on decalcification. Stages in its disapi^eai-ance 

 can be easily followed in Madrepora^ in passing downward from the apex of branches, and it 

 becomes obvious that the mesogltEa-like covering on the outside of the calicoblast layer is but 

 the densest, most recent part of the matrix, which is persistent, and within which calcareous 

 fibers will be deposited. 



The skeletotrophie tissues of Siderastrsea and Madri'pxjra ^vill be described in somewhat more 

 detail. 



The skeletotrophie liiyers of Siderastrmi present certain features which distinguish them 

 from most corals (PI. XXIV, figs. 157. ItiO). The endoderm broadens but little in passing from 

 above downward; it is strongly vacuolated throughout, and contains numerous granules and a 

 few zooxanthelliv, but all traces of cellular divisions are lost. The skeletotrophie mesogtea 

 is everywhere thin, forming only a mere dividing line between the ectoderm and endoderm. 

 The skeletogenic ectoderm has practically disappeared where the mesenterial mesoglrea unites 

 with the bodv wall, and the mesoglcea has become swollen in a triangular manner, and is finely 

 striate, giving rise to desmoidal processes. The latter are found only in association with the 

 mesenteries, where these unite with the calicinal wall, or more frequentlj- where perforated by 

 synapticula (fig. 15!>). 



The calicoblast layer of Siderastraea is remarkable in that it persists as a broad, uniform 

 layer throughout the polyp, even in regions where the corallum may be assumed to be in a less 

 active condition, as along the lateral surface of the septa. Its usual appearance is represented in 

 figs. 157, 160. It is strongly vacuolated, with numerous tine granules which stain deeph', and seem 

 as if constituting a matrix in which the vacuoles are formed. The ectoderm nowhere presents 

 the characteristics of a columnar epithelium; cell limitations are indistinguishable, and larger, 

 deeply -staining bodies among the granules are probably nuclei. Where decalcification has been 

 carefully carried out, fragments of the homogeneous organic ground substance, within which the 

 skeleton is deposited, remain behind, closely adherent to the calicoblast layer. 



Sections of decalcified material of Madrepora through the growing region at the apex of 

 branches, usually show a broad, columnar calicoblast epithelium, much broader than the ciliated 

 endoderm of the canal system (PI. 11, figs. 8, 1(1). Cell limitations are more or less clearly 

 indicated, and the cells are largely vacuolated, the protoplasm being finely granular, and more 

 concentrated toward the periphery of the layer. The nuclei arc rounded and arranged mostly 

 along a middle zone, but are not as regular in position, nor as numerous, as in the colunuiar 

 cells of the endoderm. There is no indication whatever that tlie layer is more than one cell 

 thick, or that transverse cell division ever takes place. 



In the first few sections through the actual apex, where the spaces left by decalcification are 

 very narrow, the calicoblast laj'er is even a little broader than in the sections represented, and 

 the vacuolization is not so pronounced. In sections some distance from the apex, the layer 

 begins to narrow, and soon it is represented by flattened cells which are often .scarcely percep- 

 tible. This is the condition for the most part throughout the colony, and is inconsistent with 

 much functional activitv. 



