1884.] 75 [Hyatt. 



blastema which plays an important part in succeeding stages 

 appears also to be artificial, or else an abnormal, product ; since 

 it replaces in these preparations the normal membranes of the 

 blastula and succeeding stages, which we have repeatedly seen to 

 be exclusively cellular, as generally described by other authors. 



These observations, and our own, render it probable that 

 the azygos cell and the differentiation of the endoblast and 

 ectoblast are probably common to the earliest stages of segmenta- 

 tion in most of the groups of the Carneospongiae, though ngt yet 

 observed among Keratosa. 



The stage shown in Keller's figure 12 (op. cit., pi. 18) where 

 the ovum has a large " endoderm " cell above, and six " ectoderm " 

 cells below is equivalent to the seven-celled stage in Suberites 

 and Halichondria ; the central "endoderm" cell is the azygos 

 cell, and the six " ectoderm " cells are in position and appearence 

 the precise equivalents of the six coronal cells. A difference in 

 color enabled Keller to see that this azygos cell by its division 

 gave rise to the cells of his "endoderm" or endoblast layer which 

 filled the interior and plugged up the opening or what would 

 otherwise have been an opening in the ectoblast layer. The 

 corresponding stage in Halichondria incrustans, Eastport, was 

 characterized by the farther division of the coronal cells and 

 of the azygos cell, which were not traced. The morula became 

 multicellular and a hollow amphiblastula. of one layer was 

 formed, the cells being exceedingly irregular in some specimens, 

 and quite regular in others. 



In Hal. distorta and panicea ? and Hym. caruncula, similar mo- 

 rulae were observed which appeared to be solid for periods more 

 or less prolonged and then became hollow. In Chal. limbata a 

 solid morula was present and seemed to remain solid until a late 

 multicellular stage. No distinction between the cells in respect to 

 color, or form, was observed in the outer layer, and yet the cen- 

 tre in Chal. limbata was loosely filled with smaller cells which 

 we supposed to originate from delamination as in other species. 

 These examples show that the differentiations apparent in the 

 earlier stages of the morula, were not necessarily visible in suc- 

 ceeding stages, but that they were usually succeeded in the spe- 

 cies mentioned by a form indistinguishable from an ordinary 

 archiblastula. In species with colored cells, however, the amphi 



