212 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
outer section (Fig. 14) all the cells are fused at their bases. In the inner section (Fig. 15) 
one of the cells (Ij is free from the periblast layer below c. p. Just a little nearer the i; 
middle line of the blastoderm the other cell (‘2) also becomes free of the periblast. The ' 
segmentation cavity, of which s. c. in Fig. 15 represents a part, is very obvious in surface ^ 
views of this stage, when the lower surface of the blastoderm is brought into focus. In | 
Fig. 3 the upper surfaces of the eight cells are closely joined, edge to edge, but at the.|j 
lower level the central ends of the cells inclose a well-marked space, s. c., the floor of . 
which is formed by the periblastic layer. Occasionally there occurs at this stage a ; 
displacement of cells, or one of the blastomeres is retarded in its cleavage, so that ^ 
there results an irregular blastoderm as shown in Fig. 4. In the Bass such irregu- i 
larities are rare. I was surprised to find how comparatively common they were in 
mackerel eggs. • | 
In Fig. 6 the fourth furrow is seen beginning. The sixteen cells formed by this ' 
furrow, after they have suffered another nuclear division, are shown in Fig. 8, PI. i 
Lxxxviii. A section through a-h of Fig. 8 is given in Fig. 16, IT. lxxxix. The 
segmentation cavity (s. c.) is now plainly established, the four central cells of Pig. 8 
being entirely free from the periblast, and the cavity even extending well under the 
peripheral cells. i 
The fifth act of cleavage is indicated by the nuclear figures in Fig. 8. The cleav- 
age plane of the corner cells {x) is meridional, but in the remaining peripheral cells the 
plane is equatorial. The tour central cells on the other band suffer a horizontal cleav- i 
age (plane parallel to the surface of the blastoderm), which can only be observed in i 
sections, Fig. 16, PI. lxxxix. By no means do all of the eggs jiass through this act 
of cleavage in such a strictly bilateral fashion. In a watch crystal of eggs, at least I 
one-half of them will be found to deviate from the type. The most common variation i 
concerns the terminal and lateral pairs of cells (w, n; m', n' ; 1, r ; V, r\ in Fig. 8). In i 
one or two of these cells the cleavage is often meridional instead of equalorial. Thus ; 
in Fig. 9, I and r' (both belonging to the lateral pairs) exhibit the variation, and in Fig. , 
10, m', belonging to a terminal pair, likewise divides meridionally. The four central I 
cells invariably divide in a horizontal i)laue, and it is rare to find a corner cell [x) i 
which varies. Occasionally, however, one is found dividing equatorially. 
Fig. 17, PI. xo, lies in the plane c-d of Fig. 8, and is from a blastoderm, in which 
this stage of cleavage is near its close. The thirty-two cells are here nearly separate. 
When they are completely established and the resting stage comes on, there is invaria- 
bly some rearrangement which partially destroys the bilaterality of the germ. Even, 
however, where the rearrangement is coupled with a previous variation, such as Figs. 
9 and 10 exhibit, the origin of the thirty-two cells can often be made out. Thus Fig. 
11 shows a resting blastoderm of this stage, in which I have given each cell the letter 
of its immediate parent, as used in Figs. 8, 9, and 10. In Fig. 12, PI. lxxxix, how- 
ever, in which the thirty-two ceils have suffered nuclear division, the displacement of 
cells has been so greaf that it is impossible to trace the original bilaterality, just as it 
was impossible to follow with certainty the movements of the cells in the living egg 
under the microscope. During the last minutes of the thirty -two cell stage, it may be 
said that all degrees in the loss of bilaterality are found, and consequently it is only in 
certain embryos that the cleavage from thirty-two into sixty-four can be accurately 
followed. 
There are certain general features in this sixth act (32 into 64) of cleavage, which 
were found in all the eggs I studied, and from which, coupled with the behavior of 
