490 EIGENMANN. [Vol. V. 



yolk. In contrast with the former, the substance of these cells 

 is denser ; that is, stains deeper than that of the normal cells. 

 The contours are also much more regular and rounded than 

 of the blastoderm cells (Fig. n). They resemble in all respects 

 but their multinuclear condition the primitive sex-cells in older 

 eggs. They contrast very strikingly with the smaller, more 

 refringent, periblast nuclei, so we cannot possibly be dealing 

 with that structure — unless, indeed, this should be an atavistic 

 condition ; but the distinct contour and similarity between the 

 nuclei of these and the blastoderm cells speak very decidedly 

 against such a conclusion. The larger one contains three nuclei, 

 and measures 23 X 20 /x ; the smaller, i8X20/i. I have found 

 several similar cells in another egg of this stage, from another 

 ovary. 



Only one cell of all those examined in later stages approaches 

 these. It is a large uni-nucleated cell lying in the yolk just 

 below the edge of the entoderm in a larva with about ten proto- 

 vertebrse. It is stained very dark, and contains yolk granules. 

 Its position so near the oviferous mesoderm may permit the 

 supposition that it has dropped from its normal position. It 

 measures 18 X 22 fi, with a nucleus of 6 /jl. 



Another segmenting egg, the cells of which have become 

 separated by reagents, is probably still more instructive. It 

 contains one very large cell 47 fi in diameter, with a central 

 denser mass about 22 jju in diameter. I am inclined to think 

 that this is a sex-cell which has come to a state of rest still 

 earlier than usual. Two or four of such cells would give rise to 

 all the sex-cells observed in the subsequent stages. If this cell 

 really is a sex-cell, then the substance of the germinal epithe- 

 lium is segregated as early as the fifth or sixth segmentation. 

 This seems to me to be highly probable. 



Several figures (Fig. 6) would seem to indicate that one of 

 the larger cells of an early stage divides and gives rise to the 

 groups of smaller cells in a later stage. This can scarcely be 

 the case, since the number of cells in the earlier and later 

 stages are about equal, unless a number of the earlier cells 

 atrophy or are resorbed. The loss of four cells, two in the 

 gill region, and two in the region of the fifth body somite, is 

 probable, but even with the addition of these, the number of 

 cells in the last stage examined does not exceed the average 



