332 Thos. H. Montgomery. 



the blastoderm into the yolk secondarily, and with this view my con- 

 clusions agree. Then, I agree with Schimkewitsch, in contradiction 

 to the other writers, that the yolk cells do not originate the entoblast 

 or take part in the formation of the embryo; but my results differ 

 from those of Schimkewitsch in finding that no entoblast arises within 

 the cephalothorax, and in finding no good distinction between vitello- 

 cytes and phagocytes — j)erhaps because I have not studied the later 

 stages when the phagocytes are most conspicuous. There is also con- 

 fusion with regard to the two "cumuli." What I have called the 

 "anterior cumulus" evidently corresponds to the "cumulus" of Clapa- 

 rede (before segmentation of the embryo), the anterior invagination 

 of Salensky, the "white area" of Balfour, the "weisser Fleck" of 

 Lendl, and the "cumulus" of Schimkewitsch, this having a more 

 central position and forming earlier ; while my "posterior cumulus" 

 would correspond with the "posterior mound" of Salensky, the 

 "cumulus" of Balfour, the "Primitivhiigel" of Lendl, the "secondary 

 thickening" of Kishinouye, and the "weisser Fleck" of Schimke- 

 witsch. In this comparison I disagree with Schimkewitsch's conclu- 

 sion that his "weisser Fleck" arises anterior to the anterior cumulus, 

 while I agree with him that the so-called "cumulus" of Morin is, as 

 the one described by Claparede after the appearance of the segments, 

 incomparable with the gastrulation cumuli. Balbiani and Locy 

 also observed two cumuli, but since they failed to indicate the 

 boundaries of the germ disc it is difficult to understand their ac- 

 counts. My conclusion is peculiar that only the anterior cumulus 

 forms mesentoblast. 



13. Segmentation and Appendages of the Cephaloiliorax. 



The mesoblast at the termination of gastrulation is a continuous 

 single layer; its later segmentation occasions the appearance of the 

 protozonitcs (early somites). The germ disc changes from a circular 

 to an ovoid outline, then its broader end marks the region of the 

 head, and its narrower end (caudal lobe) marks the position of the 

 earlier posterior cumulus. The earliest stage seen of segmentation 

 was one with foui- ])rotozonites, which seem to appear almost if not 

 quite synchronously; these are the segments of either (1) the four 



