No. 2.] COMPARATIVE CYTOLOGICAL STUDIES. 513 
losungsvorgange,” due to chemical changes in its substance. 
Cf. the movements described by Flemming (’ 97 ) for the ovum 
of Ascidia . 
The nucleolus has in some cases a viscid consistency (as 
described by me for Stichostemma ) and then may be irregular 
in form ; in other cases it is more fluid, and this is probably 
the case when it has regularly a spherical shape, i.e., the globu¬ 
lar form characteristic of drops of a thin liquid. Its more or 
less fluid consistency allows changes of form, division into 
particles, and fusions of neighboring nucleoli. 
The division of a nucleolus into two or more parts is a 
normal and regular phenomenon in many cells, though all 
nucleoli do not show this property. Two kinds of nucleolar 
division may be distinguished : ( 1 ) that mode by which the 
nucleolus becomes elongated and then breaks into two or more 
parts, whereby the daughter-nucleoli are usually capable of 
further division ; and ( 2 ) that mode by which the nucleolus 
fragments nearly simultaneously into a number of small gran¬ 
ules. From my own observations the former mode is evinced 
by the nucleoli of the muscle and giant gland cells of Piscicola , 
the giant cells of Doto , and the germinal spots at certain 
stages in the ovogenesis of the metanemerteans. This mode of 
division cannot be regarded as a phenomenon of nucleolar 
degeneration, since the nucleolus and its products may often 
continue to increase in size during the process of division. But 
the second mode, that by which the nucleolus breaks into a 
large number of granules, since it is particularly characteristic 
of the nucleolus in nuclear division, may be regarded as a 
process of degeneration ; the case of divisions during nuclear 
division shall be considered later. A strange mode of nucleolar 
division has been described by A. Schneider (’83). According 
to his observations on Klossia , the smaller nucleoli are portions 
of the inner substance of the larger nucleoli and wander out of 
each larger one by passing through the pore (“ canal micropy- 
laire ”) of the cortical substance of the latter; this intranucleolar 
origin of the smaller nucleoli is still open to question, since it 
was not observed in life, and since the canal micropylaire was 
observed in only one nucleolus. Marshall (’92) has described 
