236 Edmund B. Wilson. 



The Isolated ^^-blastomere first divides equally (Fig. loi), 

 then unequally and dexiotropically, so as to form two slightly 

 smaller micromeres, displaced towards the left (Fig. 102). Up 

 to this point the embryo remains strictly a half of the correspond- 

 ing 8-cell stage. At the succeeding division the differences be- 

 tween the open and closed types become apparent. In the former 

 case, as shown in Figs. 103-104, the divisions may occur nearly 

 typically, though frequently the cells become more or less dis- 

 placed. In the second case the cells shift during the division, so 

 as to fit accurately together; and, as is clearly shown in Figs. 105, 

 106, the two trochoblasts (shaded) may thus come to lie on op- 

 posite sides of the embryo, as is also the case with the two cells 

 of the second quartet. I have not followed out in full the later 

 cleavage of these larvae, which are very puzzling in both cases, 

 owing to either the initial or subsequent shiftings. But so much 

 is certain, that from the open type may arise an open blastula 

 (Fig- 108), while the closed type remains closed; and the effects 

 are clearly shown in the resulting larvae. Fig. 107 shows a closed 

 32/2-cell stage, with the polar body in position. This embryo is 

 diflicult to analyze in detail, but very clearly shows two rosette- 

 cells above, with the corresponding i^- cells, and on each side 

 are two cells that doubtless represent the daughter-trochoblasts. 

 The eight cells of the lower hemisphere are more difficult to 

 identify, and the cell-connections shown in the figure are only in- 

 ferred. 



The isolated M blastomere first divides unequally, forming a 

 micromere above, a macromere below (Fig. 109) ; and this is 

 followed by a leiotropic division identical with that occurring 

 in a single quadrant of a whole embryo (Fig. no). The 16/4- 

 cell stage then divides dexiotropically, producing a 32/4-cell stage 

 that may pretty accurately correspond with a single quadrant of 

 the normal 32-cell stage. Fig. in shows this stage of the same 

 individual shown in iio; this differs from a single quadrant of a 

 whole 32-cell stage only in the fact that 2^ has extended upwards 

 somewhat, so as to separate i^- from i"". Fig. 27 shows a 32/4- 

 cell stage that has separated somewhat (the cells are shown ex- 

 actly as they lay). Every cell is of correct proportion and po- 



