200 
Allen— Nuclear Division in the 
later by the splitting of the thread. From some of her figures (’ 96 , 
Figs. 15-18), it seems possible that the dots she describes may be the 
chromomeres ; but others (’ 96 , Figs. 13a, 13$ ; ’ 97 , Figs. 1 a , 3#, 4, 5) make 
it plain that the ‘ chromatin dots ’ are merely the external swellings on the 
thread. The description of Farmer and Moore (’ 95 ) also seems to refer 
to these swellings. They find that, before synapsis, the much-convoluted 
linin filament ‘ becomes charged with chromatin granules which are 
especially arranged along the parallel edges of the somewhat flattened 
thread.’ After synapsis, as the thread thickens, the chromatin becomes 
more abundant in it, and its distribution along the edges is then very 
apparent. Splitting occurs in the middle line between the rows of 
chromatin material. In a more recent paper, however, Farmer and 
Moore (’ 03 ) give figures of the longitudinally split spirem which show 
a row of what appear to be the real chromomeres in each daughter-spirem. 
From my preparations it is apparent that the external dots or granules 
are not the chromatin bodies themselves, but merely the portions of these 
bodies which project beyond the linin in which they are imbedded. The 
appearance of the two rows of ‘ dots 5 is not, therefore, a preliminary to 
longitudinal splitting, since the dots of a pair are simply the external 
projections of a single chromomere. 
After the fusion of the parallel threads, and at about the time that the 
spirem first shows signs of a looser arrangement, the c sickle-shaped ’ 
nucleoles begin to lose their extremely flattened appearance and to return 
to a more or less spherical shape. All stages in this transition may be 
found. On regaining the rounded shape (Fig. 19), the nucleoles again show 
the vacuolated condition characteristic of earlier stages. The nuclei from 
this time onwards resemble those of the pre-synaptic period in the size of 
the nucleoles and in their varying number, usually two or three, sometimes 
more, appearing in each nucleus. Nucleoles may now be found in any 
part of the nucleus, but for the most part they lie either close to, or within, 
the still closely crowded coils of the spirem. The nuclear membrane again 
displays a decided affinity for stains, and from this time onwards it stands 
out in triple-stained sections as a fairly thick, dark blue line. 
While these processes are going on in the nucleus, the cell as a whole 
has been undergoing a change of shape. At a very early period, even 
before the completion of the fusion of the parallel nuclear threads, the cell 
walls connecting the pollen mother-cells into a tissue begin to dissolve, and 
the cells round up. By the time of the return of the nucleoles to their 
spherical shape, the cell walls have for the most part disappeared, and 
the cells have separated and rounded up, each surrounded only by a plasma 
membrane. In many places, however, two or more cells are directly 
in contact with one another by their plasma membranes, their outlines 
being flattened along the surfaces of contact. The size of the cavity of the 
