No. 2.] COMPARATIVE CYTOLOGICAL STUDIES. 
427 
As we had concluded for the preceding species, so also in the 
present and in the species of nemerteans yet to be described, 
the nucleoli are in all probability accumulations within the 
nucleus of a substance taken up from the cytoplasm, this sub¬ 
stance being related to that which constitutes the yolk balls. 
In the least mature germinal vesicles of T. catenulatum we 
found one or two very large, lightly staining nucleoli; these stain 
in the same way and show the same structure and degree of 
refraction as do the daughter yolk balls (Figs. 107 and 116). 
Further, I have noticed in the cytoplasm small yellowish spher¬ 
ules (yolk-ball fragments) which are in every way similar to the 
smaller nucleoli, and quite frequently I have observed one or two 
of them pressed so close against the outer surface of the nuclear 
membrane as to cause a depression of the latter (Figs. 112 and 
118). In other words, it would seem that the substance of 
some of the yolk-ball fragments is taken into the nucleus and 
in the latter is re-formed into nucleoli. As long as yolk balls 
or their fragments are found within the cytoplasm lightly 
stained nucleoli of approximately the same dimensions as these 
may be seen in the nucleus. I have never seen a pore in the 
nuclear membrane through which a yolk-ball fragment could 
penetrate, though this membrane sometimes appears to be 
thinner at the point of contact with a yolk-ball fragment than 
at other points in its circumference. But in the third stage, 
when all yolk balls and their fragments have disappeared and 
the whole cytoplasm is thickly filled with their derivatives, the 
mature yolk spherules, large, faintly staining nucleoli, are no 
longer present in the nucleus, but the smallest nucleoli present 
at this time resemble in form, size, and stain, the yolk globules. 
Therefore we must conclude that the young, small nucleoli 
which first appear about the end of the third nucleolar stage 
represent mature yolk spherules, or at least that the substance 
of the two is equivalent. While the nucleoli of the first gener¬ 
ation (formed in the first stage) are commencing to degenerate, 
new nucleoli of a second generation begin to arise in the 
nucleus, and the latter, which may serve as nourishment for 
the chromatin threads, differ from the former genetically, in that 
they are not assimilated portions of yolk-ball fragments, but 
