MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 497 



simplified by regarding the latter as stomodajal budding; but, as already stated, the step 

 involves an important morphologit-al distinction. Although the stages in division of the mouth 

 or stomodanun have not been actually observed, the results to be detailed below prove conclusively 

 that in fission the stomodajal wall is actually divided into two equal or unequal parts, and that 

 the complete mesenteries inserted on each part go along with it, and help to form the new or 

 daughter polyp. The plane of fission is entoccelic, and usually at right angles to the directive 

 plane and longer oral axis; hence, only one pair of directives is retained by each of the two 

 primary daughter polyps (p. 505). Were fission to proceed no further in all probability the 

 luesenteries in their later growth would assume the hexameral plan, and the polyps would only 

 differ from a larval or liud polyp in having but one pair of directives. In most instances, 

 however, the daughter fission polyps are again subjected to fission, so that they never attain a 

 truly regular cyclic character. 



The process of polvpal gemmation and fissiparity, as revealed by individual species, will 

 be briefly described. Madixpora, Solcnastnea., and Cladocora will serve as examples of the 

 former, and McDilclna and Favia as illustrations of the latter phenomenon. 



BUDDING IN MADREPORA. 



Madrcpom is a favorable foi-m on which to study extratentacular gemmation, in a region in 

 which there is no perithecal continuation of the mesenteries (coenosarc). The early stages are 

 reproduced on PL III, figs. 22-;i7, taken from longitudinal sections of a very young bud, a little 

 below the apex of a branch. 



All the sections represented are from the left side of the median axis of the bud, but the 

 sections on the other side exhibit the same details. The right end is upper in relation to the axis 

 of the branch on which the bud was situated, and the left end is lower. Fig. 22 is from the median 

 dorso-ventral plane passing through the stomoda?um and the axial entocreles. The polyp is yet 

 scarcely raised above the general surface of the coenosai'c; the ridges above and below (right 

 and left in figure) probably indicate the commencement of the axial entoccelic tentatcles, and the 

 included depi'ession the central part of the oral disk. Compared with that of the colony 

 generallj', no histological difference is yet presented by the outer ectoderm. Communication 

 between the exterior and the superficial canal sj'stem has just been definitely established, the 

 mesogloea of the ccenosarc passing directly into that of the stomodasal wall. The stomodfeal 

 walls hang freely within a superficial longitudinal canal, differing in no important respect fi'om the 

 others around; but as the sections are taken in a longitudinal direction, the canals appear much 

 longer than in the case of transverse sections (PI. I, figs. 2-6). The endoderm of the canal has 

 undergone a marked alteration from that lining the canals and gastric cavitj^ of the polyps. It 

 is broader, more stronglj' ciliated, non-vacuolated, and zooxanthellie are practically absent, though 

 present in the surrounding canals; long, narrow, supjiorting cells, with abundant protoplasmic 

 contents, are the chief constituents. As best shown in figs. 23 and 24. the endoderm of the canal 

 becomes thinner and more normal toward the periphery- of the chamber; the inner and outer 

 layers — ectoderm and endoderm — of the stpmodseal walls are histologically alike. 



The stomoda?al wall for a few sections bej'ond that represented in fig. 22 appears as a 

 projection from the superficial wall of the colony, hanging freely within tiie canal; the periphery 

 of the projection exhibits four vertical mesoglceal strands, connected with a lower transverse 

 strand. Later, as shown in fig. 28, three central cavities appear and separate the ridges into 

 four distinct components. The lower transverse connecting strand is the horizontal continuation 

 of the stomo.dt^al wall, and tlii^ vertical strands represent the mesenteries, not yet separated from 

 one another. The stomodanil wall is continued, as it were, along the free edges of the mesenteries, 

 as often happens in adult polyps. In the next section, fig. 24, the upper and lower (right 

 and left) projections have become free, l>nt the two inner are still united l>j' the stomodseal 

 prolongation. 



The two inner mesenteries afterwards, fig. 25, become free, and now the uppermost of 

 the four is united with the boundaiy huer of the canal, and in the later sections ceases to exist. 

 Immediately below the uppermost mesentery is a slight ])rojection of the cojnosarcal endoderm 



