484 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



In the most actively growing regions of the colony, where the calicoblast layer forms a 

 broad columnar epithelium, the mesoglcea appears as a uniformly thin layer, entirely free from 

 any connection with the corallum (figs. 8,16); but in other areas, when decalcification has been 

 carried out slowly, the middle layer displays numerous desmoidal processes on its skeletal surface 

 (figs. T. 17). The most perfect forms of desmoidal processes are conical or pyramidal, the base 

 projecting outward; usually the}' are seen in longitudinal section, but occasionally in transverse 

 section. Toward their free edge the processes stain much more deeply than elsewhere, and are 

 very finely striate, the strife being practically parallel and of equal length. The actual edge of 

 the desmoidal processes is usually jagged, as if torn from some attachment. 



Elsewhere the mesoglceal lamella exhibits smaller, less regular elevations, which are similarly 

 deeply stained and striate. When a process is cut through transversely, either radiating stria? 

 are presented at all the levels, or a punctate appearance is revealed. Nuclei are scattered about 

 the origin of the processes, but are never found within it. The processes are undoubted 

 outgrowths of the thin mesoglceal lamella, the actual continuation being readily observed; but 

 the deeply -staining character toward the free edge indicates that some structural alteration has 

 taken place. 



Madrepora offers special advantages for a study <>f the nature and relations of the skeletal 

 matrix, as, owing to the rapid growth taking place at the apex of the branches, the admixture of 

 ground substance and calcareous deposit is here better preserved than in many corals. In apical 

 polyps, which have been hardened in chromic acid, and slowly decalcified by means of weak acetic 

 acid, the organic matrix of the skeleton is obtained m situ. It presents exactly the appearance 

 of a striated mesoglceal substance, and behaves toward reagents in the same way, and is wholly 

 devoid of nuclei. The fine stria? are arranged in a scale-like manner (fig. IS). The matrix is 

 best developed at the terminal region of the corallites, and is densest toward the polypal surface, 

 gradually thinning toward what would be the primary center of calcification within the branch. 

 In most preparations it is shrunk from the calicoblast layer. 



Looking at fragments of the corallum under a low magnification (about 60) the surface has 

 the appearance of minute imbricating scales. They are somewhat polygonal in outline, and the 

 free edge is directed obliquely forward, the whole indicating a spiral arrangement, most 

 pronounced in the smaller branches. Under a higher magnification, however, each of the 

 apparent scales is seen to be but the exposed terminal portion of an obliquely arranged fiber, 

 proceeding from the deeper regions of the corallum, and the superficial scale-like appearance 

 is produced by the overlapping of the numerous fibers. This is also seen in sections of the 

 skeleton starting from the surface, but the distinction between one fiber and another practically 

 disappears a little below the surface. The free edge of each fiber is often slightly jagged, not 

 rounded and smooth, as if torn from some other connection. Viewed bj* transmitted light, each 

 fiber has an indistinct, delicate, fibrillar appearance, with alternating lighter and darker trans- 

 verse bands, somewhat recalling a striped muscle fiber (fig. lit)." The matrix left upon decalcifi- 

 cation is thus closely repeated in the microscopic characteristics of the newly formed skeleton. 



In studying the relations of the corallum to the soft tissues, sections have been made 

 through the apical region of branches in which decalcification has but partly proceeded. It is 

 found that the middle portion or center of calcification is the first to disappear by the action of 

 tlie acid, the periphery, where most organic ground substance occurs, being last. 



"I >]•. Ogilvie (1896, p. 217) also gives a figure of the enlarged calcareous fibers of Madrepora, terminating in 

 what she considers to be calicoblast scales. According to the view there presented, the scales are calcified calico- 

 blasts, but as shown above the polypal tissues afford no support for such a conclusion. The organic matrix referred 

 to by Miss Ogilvie is proved to be a homogeneous, jelly-like substance secreted by the calicoblast layer, within 

 which the calcareous fibrillse are laid down in a scale-like manner. 



