70 JOURNAL OF THE 



points except at that which corresponds to the future 

 tail end of the embryo. By this means a certain por- 

 tion of the blastoderm edoe becomes turned in on each 

 side of the median line in the posterior reg-ion, the two 

 portions running forwards side by side to the point 

 already mentioned, where no centrifugal growth occurs. 

 These two portions fuse and form the primitive streak, 

 which thus at first extends to the very edge of the 

 blastoderm. Now, however, centrifugal grow^th be- 

 gins at the posterior pole of the blastoderm, and the 

 primitive streak gradually takes up its well known 

 position at a distance from the edge. 



In opposition to this view Oscar Hertwig, Rabl, and 

 others claim that the blastoderm edge is not a part of 

 thegastrula mouth, but is a peculiarity of certain meso- 

 blastic ova, and that the blastopore is represented ex- 

 clusivel}^ by a structure known as the sickle plus the 

 primitive streak. This doctrine is based on the belief that 

 an ingrowth or invag'ination of cells takes place onh^ in 

 the region of the sickle and streak, and not round the 

 edge of the blastoderm. In a paper on the develop- 

 ment of teleost fish I have already attempted a criti- 

 cism of this view^ and will onl}^ add that it is to my 

 own mind in direct contradiction with the admirable 

 account given by Duval of the formation of the primi- 

 tive streak. On the other hand it receives support from 

 the discoveries of Kupffer on the reptilian embryo, and 

 from Roller's description of the way the streak is 

 formed in the bird embryo. 



According to Roller's account^ which is adopted by 



2. The Einbrvology of the Sea Bass. Bulletin U. S. Fish Commis- 

 sion. Washing-ton. 1891, pp. 268-271. 



3. Beitrdg-e zur Kenntniss des Hiihnerkeims in Beginne der Bebrii- 

 tung. SB. der Konig. Akad. d. Wiss Wien. 1879. — Untersuchungen 

 iiber die Blatterbildung in Htihnerei. Archiv fiir Mikros, Anat. 

 Bd. XX. 1881. 



