238 K. mTsuKURi and c, ishikawa. 



that given in figs, la and h. When we make a careful study of the 

 hitter embryo, some such series of changes as Ave have sketched out 

 will become an absolute necessity. Our views are in the main like 

 those of Weldon and of Strahl, but we think we have filled in more 

 details. Strahl, it is true, says that the passage begins in the 

 middle of the primitive streak. We are inclined to think that in 

 his figs. 8 and 9, Taf. xiv (No. 8), he has stages in which the dif- 

 ferentiation of the ectoblast from the entoblast has not proceeded as 

 far as the dorsal lip of the blastopore. In our view, the 2nd ond 3rd 

 processes given in his account have the closest relations to each 

 other. Our hypothesis also makes what takes place in Reptilia 

 harmonise well with the development of lower forms, especially of 

 the Amphibia. 



Discussion of the Eesults of our Observations. 



In an article published as early as 1875 Balfour (No. 1, p. 208) 

 states that " Amphioxus is the Vertebrate whose mode of develop- 

 ment in its earliest stages is the simplest, and the modes of develop- 

 ment of other Vertebrates are to be looked upon as modifications of 

 this, due to the presence of food material in their ova." In the same 

 article, as well as in several subsequent publications (Nos. 2 and 3), 

 he endeavoured to work out the comparison of the vertebrate 

 development with the idea given in the above quotation for its 

 foundation. Above all, he has insisted that the mesoblast always 

 arises as paired masses, one on each side of the median line, and that 

 these two masses are to be regarded as paired diverticula of the 

 alimentary canal. Recently 0. Hertwig (No. 6), in connection wàth 

 the " Coelomtheorie " of himself and his brother, has worked out this 

 idea very completely in Amphibia, and has also shown, from the 

 investigations of other W(^rkci's, how the same idea could lie carried 



