No. 3.] 



THE CELL-LINEAGE OF NEREIS. 



397 



nal trochoblasts gives rise to four cells, of which three remain in 

 the prototroch, and the fourth lies above it in the upper hemi- 

 sphere of the larva. The twelve prototrochal cells thus fall 

 into four groups of three each, which are at first distinctly 

 separated from each other {cf. Figs. 38, 41). The gaps soon 

 close up, however, excepting the posterior, which persists as the 

 well-known interruption in the median dorsal {posterior) line. 



The examination of a large number of embryos has led me 

 to believe that this law of division is perfectly constant. The 

 order of division varies considerably, as is shown in the figures. 

 Thus in Fig. 32 some of the posterior girdle-cells have already 

 divided, while the anterior micromeres and girdle-cells are still 

 quiescent, and only occasionally is a specimen found that shows 

 the divisions as completely as in Figs. 35, 36. The division- 

 planes, however, appear to vary only within very narrow limits. 



The following scheme shows the derivation of the proto- 

 trochal cells and brings out certain curious points in their mode 

 of origin (see also the general cytogenetic tree, p. 382). 



f 



Anterior. 



p% - ^1,1,1,2 ^ 



^1,2 



Left, 



^1^ 1-^1,1,2 /.^i, 1,2,1 =^7 



[a 1.1.1 J 2 



1,1,1,2 -pi, ^ pi - 1,1 



f ^1.1.2 Z^l. 1.2,1 =^10 g p^ = c^' 



\^^'^i^)\ p ^1.1,1,1 ^^11 I ^2 = 



^1.2 



^1, 1, 1, 2 _ ^ : 



J- ^1.1.11 



1 



1,1.1 J ^<5i.i.(/)J 



i,2,l\ <51,1,2 j 



-' Right. 



1,1,1 



1 

 \ .1.1.11 



pi = ^l,l,2,l\ ^1,1,2 J ^1,2 J 



Posterior. 



An inspection of this scheme brings out the fact that al- 

 though the prototrochal cells are bilaterally symmetrical in posi- 

 tion, they are not so in origin, sirice the brackets on the two 

 sides do not correspond. The unlettered lines lead to the cells 

 expelled from the prototroch in the horizontal divisions {cf. 

 Figs. 27, 35, 39). 



The significance of the prototroch-formation remains to be 

 seen, since nothing is known of the phenomena in other forms. 

 The number of prototrochal cells has not hitherto been deter- 

 mined with certainty in a single form. Hatschek (No. 10), 

 confirmed by Fraipont (No. 8), describes the prototroch of Po/y- 

 gordiits as consisting of numerous cells. Von Drasche (No. 5) 



