EARLY STAGES OP DEVELOPMENT OF THE MOUSE. 65 



sufficiently emphasise the distinction between epiblast and 

 proximal trophoblast, a point to which he attaches consider- 

 able theoretical importance, and to which I shall have to 

 return later. 



In attempting to account for Robinson's error I am inclined 

 to believe that he had some difficulty in identifying the tro- 

 phoblast in a later stage (such as that which I have repre- 

 sented in fig. 8 ; compare Selenka, fig. 25, and Duval, fig. 90), 

 in which it is enormously extended and thinned out (Reichert's 

 membrane) and not always easy to distinguish from the 

 submucosa of the uterus ; and that his endeavour to give an 

 explanation of this has led him to a misinterpretation of the 

 earlier stages. In the first place he has assumed that the 

 unattached blastocyst, as it moves down the uterus to the 

 position it is permanently to occupy, does not rotate upon 

 itself in every possible direction (as it most certainly does), 

 and hence, with the help of a slight difference in staining, has 

 identified that end which he happened to find anti-mesometric 

 in a free blastocyst, and which is in reality the embryonic 

 knob, with the end which is distal, or anti-mesometric in a 

 fixed stage, that is to say the yolk-sac, and has therefore 

 labelled this end hypoblast (see his figs. 3, 4, 5; fig. A). 



In the second place he has described an " invagination " of 

 the " roof " of the " segmentation '' cavity (his words in this 

 connection are somewhat inconsistent) from two sections only, 

 figured in his figs. 6 and 7. The first of these, I think, must 

 have passed through a fold in the trophoblast, such as are of 

 frequent occurrence in material fixed with picro-sulphuric 

 acid ; this fold would be what he calls hypoblast ; what he 

 has described as the epiblast I believe to be really so, or 

 rather the embryonic knob, but I feel sure that the line by 

 which he makes the central cavity of this epiblast communi- 

 cate with the exterior is an artifact. Fig. 7, I think, must be 

 tangential (it is the tenth of fourteen and cut obliquely !), the 

 appearance of the spaces (which he regards as separate, 

 though he advances no evidence in support of this, and as 

 segmentation and yolk-sac cavities respectively^ being due to 



VOL. 43, FART 1. NEW SERIES, iS 



