156 BOTANICAL GAZETTE [FEBRUARY 
recently divided binucleate cysts were found. Their chromatin 
strands (fig. 57) still showed by their orientation the position of the 
chromosomes from which they had been derived. As stated above, 
the position of the mother nucleus is still clearly indicated by a 
starch-free area in the cytoplasm. 
No centrosomes or asters, except the pseudoaster above noted, 
were seen in connection with any of the spindles. The poles are 
very sharp, without any surrounding zone of denser cytoplasm in 
which a centrosome might have been concealed. There is no 
indication that astral bodies have any part in the formation of the 
nuclear membrane, as in Synchytrium decipiens.s While but very 
few of the critical stages were seen, it seems evident that, if there 
were any such conspicuous asters as in that plant, they would cer- 
tainly have appeared in the preparations studied. 
In the intermediate mitoses, spindle formation conforms in a 
general way to that in the primary nucleus, but the metaphases 
(figs. 58, 59) are so different that at first sight they would seem to 
be of an entirely different type. The differences, however, are not 
so great as would appear. In the smaller nuclei nearly as great an 
amount of chromatin is used in the formation of the chromosomes 
as in the larger. Their spirems are therefore much larger propor- 
tionately, and, instead of lying within the spindle, stretch nearly 
across the nuclear cavity. Sometimes such spindles show a con- 
siderable amount of chromatin which is not utilized in the formation 
of the chromosomes, but is cast out, as in the earlier divisions. 
Frequently, however, all of the chromatin goes into the spirem 
(fig. 59). The karyosome, which is so strongly developed in the 
primary nuclei, becomes gradually less and less prominent in later 
nuclei, until in the many-nucleate cyst the chromatin assumes the 
condition of a typical reticulum, although it is not finely divided, 
but remains in rather large masses which are connected by coarse 
linin strands (figs. 27, 29). In consequence of the different disposi- 
tions of the chromatin in these nuclei, the residual chromatin cast 
out during their mitosis does not take the form of large spherules, 
but is finely subdivided (fig. 58). Such a condition was also seen 
3 §, taraxaci is without karyodermatoplasts according to the recent results of 
Batty (Jahrb. Wiss. Bot. 50: r10. 1911). 
