208 PROCEEDINGS OP THE AMERICAN ACADEMY. 



the other derivatives of D^-^ (D^-^ aud D»-"^ of Fig. 10, left half) are 

 in the eighth generation. 



Finally, the nearly median optical section of this egg, reproduced in 

 Figure 14, shows at a ijlance the absurdity of Van Beneden et Julin's 

 interpretation. Here the relatively large columnar cells d®-'', d®-^, 

 a®-®, and a®*^ of my interpretation,* are derived by them from the thin 

 superficial cells D®*^ and A*^'^ of the preceding stage (Fig. 13), the 

 first two together with D®-^ as a result of two successive divisions of 

 the single cell D"--^. On the other hand the small cubical cells D''^ 

 D'-^, A^*^, A'^"\ and A''*^ are taken by them to represent, undiminished 

 hy division, the voluminous columnar cells d*^"'', d^-^, a^***, a^*^ and 

 a^"^ respectively, of the preceding stage. 



It is thus seen from an examination of Van Beneden et Julin's own 

 figures, that they have reversed the dorsal and ventral relations in 

 passing from the 32-cell to the 44-cell stage, and have identified the 

 endodermal half of one with the ectodermal half of the other. This 

 fact has been overlooked by all subsequent writers on tunicate em- 

 bryology, for all have accepted as correct the results obtained by Van 

 Beneden et Julin. 



The question next arises, Is the orientation given by Van Beneden 

 et Julin for the 32-cell stage, or that given for the 44-cell stage, the 

 correct one ? A strong presumption that the latter is the case exists 

 in the close correspondence in form and size between the endoderm 

 cells of the gastrula figured by them (their Fig. lid) and the columnar 

 dorsal cells of their 44-cell stage. This presumption becomes as 

 strong as anything except direct observation of Clavelina can make 

 it, when one follows the cell lineage of Ciona through to a stage which 

 marks the beginning of gastrulation, such as I have represented in 

 Figures 11 and 12. 



In order to avoid the extension of this paper to an undesirable 

 length, I have not figured the stages intervening between the 46-cell 

 stage and gastrulation. A detailed study of these stages, however, al- 

 lows me to speak with confidence of the changes which have occurred. 



In the 32-cell stage (Figs. 3 and 4), all the cells of both hemi- 

 spheres were seen to be in the sixth generation, though, as has been 

 stated, the cells of the dorsal half had divided a little later than those 

 of the ventral half. These dorsal cells now remained quiescent, while 

 all the cells of the ventral or ectodermal half, except the small pos- 

 terior cells C®^ and D®^, divided, and thus brought about the 46-cell 



* See Explanation of Plates, Figure 14. 



