210 PROCEEDINGS OF THE AMERICAN ACADEMY. 



gastrula is formed with a dorsal blastopore by a process " intermediate 

 between epibole and invagination " ; and because of the second, the 

 blastopore closes from the sides and anterior end, so that it ultimately 

 becomes pear-shaped and lies in the posterior half of the embryo. 



The part which particular cells take in forming the organs of the 

 larva is reserved for discussion in a later paper, as that part of my 

 work is still incomplete. One thing may, however, be stated posi- 

 tively. The entire endoderm and mesoderm are not derived as stated 

 by Van Beneden et Julin from the twelve cells of the 44-cell stage 

 a«-5, a«-«, a'^-^ a*'-'*, d*'-^ d«-^ b«•^ b«•^ b«•^ b^-^, c«•^ and g^-\ 



Conclusions. 



1. The conflicting statements of Van Beneden et Julin on the one 

 hand, and Seeliger on the other, regarding the segmentation of the 

 ascidian egg, are explained by the detection of a fundamental error 

 in the work of each. 



Seeliger has determined the dorsal and ventral sides of the egg 

 correctly, but reversed the anterior and posterior ends in all his figures 

 of the early stages. Van Beneden et Julin have determined correctly 

 the anterior and posterior ends, but have reversed dorsal and ventral 

 in all stages previous to the 44-cell stage. These errors are clearly 

 indicated by a critical examination of the authors' own figures, and a 

 careful study of the cell lineage of Ciona leaves no doubt of their 

 existence. 



2. The correction of Van Beneden et Julin's unfortunate change 

 of orientation necessitates fundamental changes in their conclusions. 



(a) It is stated by them that the four smaller cells of the 8-cell 

 stage give rise to ectoderm only, while the four larger cells produce 

 both endoderm and ectoderm. On the contrary, neither the four 

 smaller cells nor the four larger ones produce ectoderm exclusively ; 

 but it is the four larger, not the four smaller ones, which give rise 

 to the greater portion, perhaps all, of the ectoderm. 



(b) I cannot at present assent to the statement of Van Beneden 

 et Julin that the separation of the germ layers is complete at the 

 44-cell stao-e. 



3. In Ciona, from the sixth to the eighth generations at least, cell 

 multiplication is more rapid in the anterior and lateral portions of the 

 egg, and this fact is an miportant element in determining the shape 

 and position of the gastrula mouth. 



Cambridge, May 9, 1894. 



