Nuclear Division in the Hepaticae. 495 
directions, and its general character recalls the framework of 
radiating filaments in a spider s web. The nucleolus is situated 
in the middle of these radiating threads, and the appearance 
is such as to strongly suggest an intimate connexion between 
the two structures. As the linin becomes more pronounced 
in character, the centrospheres are seen on the periphery of the 
nucleus at opposite sides. They are difficult to distinguish at 
this stage, on account of the feebleness of the radiations, later 
on they are however readily recognized. 
The nucleolus now becomes very much vacuolated, and no 
longer stains with its former intensity, but as the nuclear sap 
remains quite unstainable, and the only structure which at 
this point suddenly begins to absorb dyes greedily is the 
linin, it seems not unreasonable to conclude that something 
has passed out of the nucleolus to the threadwork. At any 
rate one is otherwise driven to suppose that the material 
which leaves the nucleolus, and thereby causes its vacuolated 
appearance, alters its reactions to stains entirely as it diffuses 
out into the nuclear sap. After the vacuolation just spoken 
of the nucleolus commonly fragments, and finally disappears 
at the formation of the achromatic spindle. 
Meanwhile changes have gone on in the linin filament. 
Certain portions of it, which are relatively somewhat elon- 
gated, have become thickened, and now greedily absorb the 
ordinary nuclear stains. These thickened portions are the 
rudiments of the young chromosomes, and they certainly owe 
their present features to the increase of chromatin within the 
linin substratum. It is not easy to assign an origin to this 
chromatin which thus appears in these localized spots. 
The behaviour of the nucleolus already alluded to strongly 
suggests that it may have some share in the process, but I do 
not think it at all probable that it is alone concerned. 
Professor Strasburger 1 has suggested that the chromatin 
originates from the cytoplasm, in part at any rate, and I am 
inclined to attach importance to this possibility. Perhaps 
both nucleolus and cytoplasm may each furnish their respective 
1 Karyokinetische Problem e 5 Pringsheims Jahrb., Bd. XXVIII. 
Mm2 
