Hyatt.] 90 [March 5, 



have for convenience sake designated the cavity of a colony 

 of Eudorina or Volvox) is represented at this stage only by 

 the tube connecting these two apertures, but in the next stages 

 the growth of cells by transverse divi>ion gradually separates 

 these two layers and establishes the hollow sphere of the true 

 amphiblastula, the segmentation cavity or blastocoel of Huxley. 

 The differentiation of the layers does not disappear. It, however, 

 would undoubtedly be difficult to distinguish the two layers in 

 Schultze's amphiblastula (Zeitschr. Wissen. Zool., suppl., vol. xxv, 

 pi. 20, f. 15), if it were not for the differences of color in the basal 

 or ectoblastic cells. The upper or endoblastic hemisphere is very 

 large and is formed not only by transverse division of the primi- 

 tive basals, but also from transverse division first of the lower 

 parts, and then of the upper portions of the primitive apicals. 

 The upper aperture becomes closed in this stage, but the one 

 through the ectoblast remains open longer, surrounded by the eight 

 cells, or points of the primitive, dark-colored, granular basals. 

 The cells of the endoblast in the next stage acquire cylindri- 

 cal forms with collars, and tlagella, and the basal cells remaining 

 still unciliated divide rapidly, forming the ectoblast, and grow 

 out into a hemispherical outline. The embryonic amphiblas- 

 tula in its most advanced stage of growth is thus formed and is 

 commonly supposed to be typical for sponges. We have, how- 

 ever, given our reasons for considering older stages, or the cinc- 

 toplanula, as more distinctly typical. 



The mode of formation of the ectoblast and endoblast, partly 

 by the apical and partly by basal cells, is important, since it ena- 

 bles us to see that Schultze was right in designating the cells of 

 the diploplacula as apicals and basals, instead of employing the 

 usual term of endoblast or ectoblast, or their equivalents. It is 

 evident that these early stages are not the equivalents of later 

 and apparently similar cellular transformations, and that the con- 

 tinuity of the azygos and apical cells with true endoblastic cells, 

 as in Keller's Chalin. fertilis for example, is really due to fusion 

 between an earlier and a normally later and distinct stage. The 

 diploplacula indicates an ancestor with the two layers less differ- 

 entiated than generally supposed, and certainly in a more primi- 

 tive condition than is shown by an amphimorula in which the 

 endoblast and ectoblast are separable. This form probably had 



