Studies on the Cytologv of some Species of Taraxacum. 
459 
because the special membrane-formation sometimes occiirs arround these 
missed chromosomes; but generally they seem to be produced by the 
amitotic division of some of the daughter nuclei, already mentioned. 
In the loculus of the anther, just before the flower opens, we find 
many irregulär pollen-grains. They are, firstly, very variable in size, 
and secondly, their contents are also various; some having only vacuoles, 
others plenty of contents, which stain deeply with haematoxylin, indi- 
cating that their degeneration has already taken place. 
b) The development of the embryo-sac. 
The presynaptic stage of the nucleus of the embryo-sac-mother-cell 
is similar to that of the pollen-mother-cell, as mentioned before. In the 
synaptic stage the linin-net-works and chromatin-granules contract 
gradually, separating away from the nuclear wall, until a dense ball is 
produced (Fig. 64). Düring this stage linin-threads become more distinct 
and even in denser portion their structure may be discernible as a delicate 
spirem. After a comparatively long duration, there occurs the gradnal 
loosening of these threads as usual, which is easilv recognized, because 
then the threads become much thicker and shorter (Fig. 66). The spirem 
now extends loosely in the nuclear cavity and continues to thicken and 
shorten, thus attaining the so-called “dolichonema” stage. This thickening 
process, however, does not proceed equally throughout the whole length 
of the spirem, so that it nowpresents a more or less moniliformappearance, 
the thicker regions of the spirem being connected together by the thinner 
portions (Fig. 67). When the contraction proceeds further, the difference 
of thiekness of the spirem becomes more evident and there appears the 
indication of the cross-segmentation of the spirem into a number of chro- 
mosomes (Fig. 68). At this stage or earlier, I have never observed the 
parallel arrangement of the spirem, though in both the pollen-mother- 
cell and the embryo-sac-mother-cell of T. platycarpum it was very clearlv 
observed; on the contrary it appears always as a single one as in Figs. 67 
and 68. 
After a short time there occurs the gradual breaking up of delicate 
linin-fibres, which connect the chromosomes to each other, thus pro- 
ducing a few short chains of chromosomes and each of these chains 
finally divides transversally into the definite number of chromosomes 
(Fig. 68). 
At an earlier stage of the Segmentation each chromosome appears 
to be variously bent long bodies with uneven or angular outline. They 
are distributed irregularly in the nuclear cavity and there is no pairing 
