PLATE XXXIII. 



Fig. 3. Ctenolabrus 32-cell stage in division. [Fig. 3 seen from above, Fig. -1 from below] four 



(19) central floor-cells dotted, i. e. outlined with dotted lines [in Fig. 3]. (In Fig. 3] every nucleus is 

 and in state of division. Comparing this with Fig. 4 of the same age nearly, we see that the 



Fig. 4. cap consists of seventeen marginal cells and fifteen central cells. One of the marginal cells, in 



(20) cases where the cleavage is more symmetrical, as in Fig. 4, takes its place among the central 

 cells, thus making sixteen in the margin and sixteen in center. This cell is pc of the right side, 

 which is here plainly in the margin, but its inner face is much narrower than the outer, showing 

 at least a tendency to occupy a position above the marginal cells, i. e. lapping over the mar- 

 ginal cells as in Fig. 4. Looking at the central cells as seen in Fig. 4, we can say that they 

 are arranged in two double rows, crossing at right angles. The double row ( = axial row) 

 lying in axis of future embryo is two cells deep; the transverse row is two cells deep at the 

 center, but one cell deep at the extremities. 



Notice that the pair of pre-cenlral cells (pc) lap over the marginal cells farther than the pair 

 of relro-centml cells (re). The tendency of one of the pre-central cells to take a marginal posi- 

 tion seems to be quite significant, for I find that the majority of caps of this age have seventeen 

 marginal cells: four on each side, four behind, and five in front. In most cases the extra 

 marginal cell in front makes the front side more convex than the hind side. This convexity 

 is frequent in the 16-cell stage, but more frequent in the 32-cell stage. 



In both figures it is the posterior left central (a) cells that abut against the diagonally opposite 

 cells (b). I do not think this feature is universal. 



Of the central cells, four (dotted [Fig. 3, pale letters] a-d) are completely in the floor, only 

 one of them reaching to the upper side (dotted c) and four (a-d) are completely above. On 

 the right and left, between the marginal cells and the four central ones are two intermediate 

 cells (i), which extend from floor to roof, forming part of the upper and lower surface; this 

 appears to be uniform in all cases that I have studied. [In Fig. 4 the lettering is reversed, a-d 

 are floor-cells; pale a-d, root-cells). 



The pre-central and retro-central cells always lap, or lie completely above other cells; in 

 F^ig. 4 r c of the right side is completely shut out from the floor, while r c of the left side shows a 

 small part of its surface (x) in the floor. The same is true of pc of the left side. The retro-cen- 

 tral cells take part in forming the floor more often than the pre-central. In Fig. 3, the retro- 

 central cells both show considerable siu-face in the floor. In perfectly symmetrical caps, we 

 count six cells from side to side, both above and below; while from before to hind side we count 

 five above and four or five below. 



Vacuoles: — -No intercellular vacuoles could be seen in [the material used for] Figs. 3 and 4, 

 but I have other preparations of the 32-cell stage in which they show beautifully. 



The dimensions are usually much wider than long. One is much wider than in Fig. 3. 



Nuclei of four lateral marginal cells appear to divide indifferently, tangentially, or radially, 

 those before and behind almost invariably radially. 



Nuclei: — As to the nuclei, in Fig. 3, the nuclear figures of the marginal cells have a radial 

 direction. In Fig. 4, they are radial before and behind, but tangential on the sides with one 

 exception on the left side (no. 3). The nuclear figures of the intermediate cells are vertical or 

 nearly so. When the nuclear figures of the marginal cells are radial, as in Fig. 3, the inner 

 poles lie higher than the outer poles, hence I have shaded the inner poles more heavily. Even 

 when tangential I think one pole may often be higher, perhaps generally so, than the other, 

 so that in any case the division of the marginal cells results in making a ring of cells lying above 

 and overlapping more or less the extreme marginal cells. 



Cleavage-cavity: — The limits of the cleavage-cavity (6c) are shown feebly in Fig. 4. It is 

 not easily traced in succeeding stages and perhaps may be said to vanish. 



Thus a comparison of Figs. 3 and 4 shows that the direction of the nuclear figures may vai'y 

 considerably and still not obliterate the general plan. 



