Pollen Mother-cells of Lilhim canadense. 201 
anther-sac has been increased to allow room for the rounding up of the cells, 
partly by the collapse of the bounding layer of tapetal cells, partly by the 
increase in size of the anther as a whole. In the fluid filling the sac, 
in which the cells now float, there are numerous drops of a substance which 
stains bright red in the triple stain, and deep black in the haematoxylin. 
There now begins to be formed about each mother-cell a new independent 
cell wall, which, while it is still very thin, is often demonstrable in slightly 
plasmolyzed cells. There is considerable variation in the thickness of the 
new wails of different cells at the same stage of development. At the close 
of synapsis the wall is often still very thin (Fig. 20 ), but in some instances 
it has attained a considerable thickness (Fig. 21 ) ; in the latter case its 
thickness is usually unequal on different sides of the same cell. 
The series of stages which I have described under the general term 
‘synapsis’ (Moore, ’95 a) is marked by the continuous presence of an 
aggregation of the chromatic materials within the nucleus, at first approxi- 
mately central, but soon becoming eccentric. This period is an extremely 
long one, as is evidenced by the large proportion of anthers in material 
fixed from day to day through a considerable period, the cells of which 
show synaptic figures. It seems to me certain that these stages in the lily 
extend over a period of some days at least, very probably of a week 
or more. The peculiar aggregation of the nuclear thread at this period 
seems to have been first observed by Tschistiakoff (’75). He gives (in his 
Fig. VI) a very good representation of a synaptic nucleus in the f pollen 
grain ’ of Cupressus. The figure is plainly that of a pollen mother-cell. 
Synaptic figures were observed in various pollen mother-cells by Tangl (’82), 
Strasburger (’82), Heuser (’84), and Guignard (’84), all of whom, however, 
thought the peculiar appearance due to the action of reagents. Tangl and 
Strasburger observed flattened nucleoles at the same period, and the latter 
interpreted them as secretion products, the real nucleole having, as he 
thought, disappeared at an earlier stage. Strasburger also noted the slight 
affinity of the nuclear membrane at this period for stains. E. Overton (’91) 
observed a similar stage in the prophases of the first division in the embryo- 
sac of Lilium. Van Beneden and Julin (’84) found that the chromatin 
strand in the nucleus of the primary spermatocyte of Ascaris megalocephala 
originates from a nucleolus-like mass ; a similar aggregation was found 
by Kultschitzky (’88), O. Hertwig (’90), and Brauer (’93). Brauer’s descrip- 
tion makes it plain that this mass is of the same nature as the synaptic 
aggregation in plant nuclei. Hertwig also observed flattened nucleoles at 
this or an immediately preceding stage. It is worth noting in this con- 
nexion that Winiwarter (’00), whose description of the processes occurring 
during synapsis is similar to mine, finds that in the primary oocyte of the 
rabbit the nuclear membrane is very indistinct during synapsis. 
These and numerous other observations seem to show that the peculiar 
