ON THE EMBRYOLOGY OF THE TELEOSTS. 201 



the progress of the invagination the segmentation cavity is encroached upon by the 

 lower layer cells, its floor becomes covered with cells, some arising from the hypoblast 

 while others apparently originate from the free yolk nuclei, and the cavity is shortly 

 obliterated. 1 



I admit that I am in doubt as to the part played by the intermediary layer and its 

 resulting cells. As before mentioned a portion of the cells apparently enter into the 

 floor of the segmentation cavity and are subsequently either embraced in the hypoblast of 

 invagination or are crowded by it into a mesoblastic position. This however accounts for 

 but a small proportion of the cells of the intermediary layer, and it seems to me probable 

 that the hypoblast of invagination forms only the dorsal wall of the alimentary tract while 

 the intermediary layer furnishes the ventral portion. This seems to be in full accord with 

 the formation of the alimentary tract in other forms (e. g. Batrachia) where the ventral 

 portion of the hypoblast is formed by yolk cells. 



Regarding the origin of the hypoblast in the Teleosts there seems to be a diversity of 

 opinion. Henneguy ('80 p. 402-3) describes the invagination in the eggs of the perch and 

 trout, the blastoderm being inflected at its margin and a line or fissure separating the " 

 sensorial [our lower layer cells] from the inflected portion. So far we agree with him. 

 He however states that the epidermal layer is not inflected. In this he agrees with two 

 of the figures of His ('75 pi. n, figs. 2 and 3) but not with fig. 1. Our observations were that 

 the epidermal layer of the epiblast alone is inflected. Balfour ('81, 57) says that the 

 yolk cells form the hypoblast in the smaller Teleost eggs but that in the larger as in those 

 of Elasmobranchs only a portion of the hypoblast has such an origin. We should con- 

 sider our eggs as small. KupfFer with a doubt regards the cells as forming the hypoblast. 

 Professor Van Beneden in his researches on the eggs of an unknown Teleost ('78) arrives at 

 widely different conclusions regarding the origin of the germ layers from those we have 

 formed and we cannot reconcile his results with our observations. We have seen step by 

 step, minute by minute, the progress of the invagination and it scarcely seems possible 

 that any error of observation on this point can have crept in, especially as we witnessed 

 the process many times. Yet Van Beneden totally denies that in his Teleost any invagina- 

 tion takes place. It would seem to me that he is wrong from the very start. On p. 52, 

 he considers the egg before the appearance of (our) two segmentation spheres as follows : 

 " Directly after fecundation the egg of the osseous fish divides into two very unequal cells, 

 very dissimilar, differing in constitution and significance ; the one is a germ which seg- 

 ments and from which the blastodisc is derived ; the other is formed by the deutoplasmic 

 globe * * *. This cell is the origin of the endodermic layer of the future 

 embryo." To all of this I must express an emphatic dissent. The aggregation of the 

 protoplasm at one pole of the egg and of the deutoplasm at the other cannot in any way 

 be considered as a segmentation, nor can the deutoplasmic portion be considered as a cell. 

 No one would think of regarding "a centrolecithal egg, that of a Crustacean for example, 

 as composed of two cells or the central portion as of a cellular character, yet the homology 

 between the two eggs is easily shown. On the same page he explictly says that the germi- 

 nal portion is the homologue of the ectoderm and the vitelline of the entoderm, a view 



1 Subsequent studies lead me to believe that this statement to this cavity, though I must say that there appear many diffi- 

 is an error and seem to confirm the idea of Ryder with regard culties in connection therewith. 



MEMOIRS BOST. SOC. NAT. HIST. VOL. III. 26 



