STRUCTURE AM) DEVELOPMEMT OP' COELOPLAXA. 49 



As will be clear from the foregoing description, as well as from 

 the accompanying table, the divisions of the micromeres in the sub- 

 phar}-ngeal octants, i. e, the divisions of the cells arising from M, 

 generally come somewhat later than those of the corresponding micro- 

 meres in the subtentacular octants, i. e., the divisions of the cells pro- 

 duced by E — this tendency is also known in the cleavages of the egg 

 of ordinary ctenophores. In the eggs under my examination, it took 

 roughly eight hours for unsegmented eggs to develop to the 64-cell 

 stage; one cleavage required from three to five minutes to accomplish. 



For some time after the stage described last, micromeres keep 

 on multiplying by division of the previously existing micromeres, and 

 also by new buddings from macromeres. Meanwhile, E cleaves into two, 

 and the number of macromeres becomes sixteen. Then, the epibolic 

 gastrulation begins to take place. 



G astrulation. 



In the stage at the close of the segmentation stage just prior to 

 gastrulation, the macromeres are arranged in a bowl-shaped group, at 

 the margin of which on the concave side, there exist assemblages of 

 micromeres, while on the opposite side, the macromeres are exposed. 

 Then some "mesodermal cells" of Metschnikoff ('85) (fig. 15, mes) are 

 budded off from the macromeres on the convex side of the group. 

 Unfortunately I could ascertain nothing regarding the fate of those cells. 

 In the mean time, the micromeres are multiplying rapidly and gradually 

 spreading over the macromeres. While this change is going on about 

 the micromeres, the concavity of the macromere group diminishes, and 

 the gap between the micromeres on the micromere pole enlarges (fig. 

 16). Fig. 17 represents a stage during the process seen from the 

 micromere pole, where a moderately large elliptical gap exists between 

 micromeres, showing macromeres within. At the macromere pole too, 

 there is another but somewhat smaller opening. The two openings of 

 the two poles communicate with each other through a fairly wide passage 

 exi.sting between the macromeres and elongate in the transverse direction. 

 Both of the openings are closed up by multiplying micromeres with the 

 advancement of the gastrulation. The questions which pole is closed 

 first, and which pole becomes the oral pole, were discussed much among 

 previous writers on the development of the Ctenophora (cf. Chun, '80). 

 My observation in the the present form has shown clearly that, it is 



