THE EMBRYOLOGY OF THE OPOSSUM 43 



size before they eventually become noticeably smaller than 

 the retarded cells at the opposite pole. 



However, even with this correction the attempt to relate 

 the early polarity to the later polarity meets further diffi- 

 culties. If, for instance, the difference in rate of cell division 

 at the two poles of the cleavage stages be considered to per- 

 sist long enough to be responsible for the thickness of the 

 'embryonic area' in contrast to the thinness of the 'tropho- 

 blastic area,' this means that the cells of the 'embryonic area,' 

 which have divided more slowly according to the theory, and 

 belong to an earlier cell generation, must correspondingly be 

 fewer in number. Hartman's own cell counts show this not 

 to be the case. 



Hartman's counts in four blastocysts of what I call stage 11 

 show an average of 93.5 cells in the 'embryonic area' and 

 ninety-six cells in the 'trophoblastic area' ('19, p. 62), which 

 is practically identical. The average number of endodermal 

 cells in these same vesicles is thirty-three. If these come 

 from the ' embryonic area ' as he believes, then instead of fewer 

 cells in this area there are at least 30% more than in the 

 'trophoblast.' Or if his theory be modified so as to have them 

 come from the 'trophoblastic area,' their excess would not 

 be enough to indicate a difference of even 1-cell generation. 

 The only possible way to reconcile these figures with the 

 theory of a persistent difference in rate of cell division at the 

 two poles is to assume that the trophoblastic pole is constantly 

 contributing cells to the embryonic pole (just the opposite 

 of what Heuser and Streeter, '29, assume to be the case in 

 the pig). 



One other point which must be mentioned is that though 

 so much depends upon this assumption, the cells of the 

 'embryonic area' have never really been shown to be larger 

 than the cells of the 'trophoblast.' In cross sectional view they 

 are definitely thicker (at least in some stages), which means 

 that their lateromedial dimension is greater. On the other 

 hand, in surface views of whole mounts they appear definitely 

 smaller, as their two surface dimensions are less. This has 

 been shown in the figures by Selenka, Hartman, and myself. 



