Experimental Studies on Germinal Loealization. 207 



Patella with which in the main my observations agree, though 

 the arrangement of the cells of the prototroch is somewhat 

 schematized. The embryo becomes ciliated at about eight 

 to ten hours (depending on the temperature) the first cells to 

 acquire cilia being the sixteen primary trochoblasts. For a brief 

 period the prototroch consists of only those sixteen cells, still ar- 

 ranged in four separate groups (Fig. 8). The cilia are from the 

 first arranged, not in vague patches or tufts, but in very definite 

 oblique transverse rows, which bear a marked resemblance to 

 the swimming plates of a ctenophore — indeed, it hardly seems 

 forced to compare the embryo directly at this period to a larval 

 ctenophore. At the same time, or a little later, the group of 

 small cells at the apical pole, derived mainly, if not wholly, from 

 the apical rosette, develops a tuft of flexible but non-vibratile sen- 

 sory flagella, and constitutes the apical organ. 



The ctenophore-stage is of short duration. In two or three 

 hours several cells lying in the gaps between the four groups of 

 primary trochoblasts also become ciliated and ultimately enter the 

 prototroch as secondary trochoblasts. These cells are not more 

 than half the size of the primary trochoblasts (a point of im- 

 portance in connection with the experimental results) , and at 

 first bear much smaller cilia. There are at least three and prob- 

 ably four of these trochoblasts in each quadrant (with the pos- 

 sible exception of the posterior group, in which I have only 

 certainly seen two of these cells), (Figs. 10-12), giving a total 

 of 28 to 32 cells in the prototroch, to which possibly still others 

 may be added. While I have not traced step by step the exact 

 origin of these cells, their position in the embryo leaves little 

 doubt that in each quadrant two of them are derived from the 

 first quartet (/. e.^ from derivatives of the i'- cells), and this 

 is demonstrated to be the case by the experimental evidence. The 

 position of the third cell (Fig. 10) shows almost beyond a doubt 

 that it is derived from the second quartet, /'. e., from 2^ and 

 probably from 2.^\ The experimental evidence again proves 

 that at least one, and in some cases two, trochoblasts are derived 

 from the second quartet. As may be seen in Fig. 10, a second 

 cell lies next to the one described, the position of which indicates 



