456 
J. Osawa 
embryo-sac, except around the nuclei. The four nuclei in the upper 
portion of the sac arrange themselves in the usual manner and form the 
egg-apparatus as well as the upper polar nucleus, while the four nuclei 
situated at the lower part of the sac form the three antipodal cells and 
the lower polar nucleus. 
The mature embryo-sac is typical in every respect. The synergids 
lie side by side and the ovum is very large, protruding a little below them. 
The cytoplasm of the synergids is very dense and stains deeply with 
haematoxylin and has usually a large vacuole at its lower end (Fig. 41). 
That of the ovum is also greatly vacuolated. The antipodal cells appear 
to be degenerated in an earlier stage, and they remain as deeply staining 
irregulär mass at the lower end of the sac. The two polar nuclei, situated 
in the central part of the sac, fuse to each other before fertilization. The 
primary endosperm-nucleus tlius produced approaches very near the 
egg-apparatus and remains there, until fertilization takes place. The 
further development of the embryo-sac after fertilization was studied 
by Schwere (1896). 
B. Taraxaoum albidum Dahlst. 
a) The development of the pollen-grain. 
In the presynaptic stage the nucleus of the pollen-mother-cell of 
T. albidum has, as usual, a large nucleolus and many delicate linin-fibres, 
on which many chromatin-granules, variable in size and in shape, are 
scattered irregularly (Fig. 42). With the growth of the nucleus they 
increase in size and in number, and the linin-fibres become more distinct 
(Fig. 43). The approach of synapsis is indicated by the gradual eon- 
traction of the reticulum towards the nucleolus (Fig. 44), until finally 
linin and chromatin-granules are fusecl into a single dense mass, as in 
T. platycarpum (Fig. 45). After a comparatively long duration of synaptic 
stage there occurs again the gradual loosening of this mass (Fig. 46), and 
the spirem becomes thicker and sliorter than before. It gradually increases 
in thickness and at the same time decreases in length, thus producing 
a dense homogeneous spirem, but I could never observe the spirem which 
has the double structure, as it was the case with T. platycarpum, though 
many sections of spirem at this stage were carefully studied. 
When the spirem has shortened and thickened there appears the 
indication of the cross-segmentation, thus transforming the spirem into 
a chain of definite number of chromosomes (Fig. 47). At an early stage 
of this process each chromosome appears to be a more or less bent rod- 
