in Lilium Mart agon: II. Spermatogenesis. 197 
pearance of the nuclear membrane, partial solution of the 
nucleolus, and contraction of the chromatic thread. The 
structure which we have hitherto called the chromatic ball is 
seen to consist of the chromatic threads closely coiled round 
a mass of semi-fluid matter, which must be chiefly of nucleolar 
origin. As we have seen, the colouring in preparations stained 
with methyl-green and acid fuchsin suggests that there is also 
some amount of chromatin dissolved in this mass. The 
chromatic threads in nuclei which have just entered on 
synapsis — for instance, those at one end of a loculus which 
shows resting-nuclei at the other end — resemble that drawn in 
Fig. 1 a. Such threads are irregularly thickened, and the 
double row of dots can only be made out here and there. 
The cytoplasm which surrounds the nucleus often shows 
fibrillar structure at this time, and in the absence of a proper 
membrane the nuclear cavity is outlined by a denser web of 
cytoplasmic fibrils (Fig. 3). 
The structure I have just described is characteristic of the 
earlier period of synapsis. Later on — in such a nucleus, for 
example, as that drawn in Fig. 3 a — the nuclear membrane 
begins to reappear, nucleoli are slowly reconstructed out of 
the amorphous nucleolar mass, and the coils of the chromatic 
thread become looser. The thread itself is now of uniform 
breadth, and is bordered by two regular rows of dots (Fig. 3 b). 
These changes proceed slowly until the chromatic coils 
occupy nearly the whole of the nuclear cavity, which is once 
more bounded by a well-marked membrane. The nucleoli 
are more or less spherical and of definite outline. No trace 
of an amorphous mass remains. The spirem-stage is reached 
when the chromatic coils, which can now be seen to form 
a continuous ribbon, occupy the whole of the nuclear cavity. 
The structure of the pollen-mother-cell nucleus in the 
spirem-condition is perfectly clear. It is a good deal smaller 
than the embryo-sac nucleus at the corresponding stage 
(cf. Fig. 4 with Fig. 14 in Part I), and the coils of ribbon are 
more closely packed in the nuclear cavity. The ribbon itself 
looks broader than the ribbon of the embryo-sac nucleus ; but 
