38 Fraser.— Contributions to the Cytology of 
the definitive nucleus of the ascus (Fig. 18). The nuclei lie against each 
other, either in the ascus or in the cell from which it originates, and appear 
simply to flow together ; the two nucleoli are visible for a time, then they 
also fuse. The two spiremes mingle and cannot be distinguished after the 
first stages of fusion. 
The synapsis or second contraction now sets in (Fig. 20). The 
chromatin filament thickens, the longitudinal split is more or less obliterated, 
and the whole thread, except a few loops which run out from the main mass 
to the periphery, becomes aggregated towards one side of the nucleus. The 
synaptic stage apparently persists for some time, then, as the contraction 
loosens, the loops running out from the central mass become more obvious, 
especially as shown in transverse section of the ascus (Fig. 21), and the 
longitudinal fission is once more apparent (Fig. 22). 
The spireme next breaks up into its constituent loops, each of these 
forming a bivalent chromosome (Fig. 23), in the limbs of which the longi- 
tudinal fission can still be distinguished. The limbs may be twisted on 
each other, or united at both ends, forming a ring-shaped figure, or they 
may diverge very considerably (Fig. 24). Most of the forms described for 
the heterotype chromosomes of Phanerogams have been observed both at 
this and at later stages. All the chromosomes are not necessarily formed 
from loops, but their limbs are always derived from different portions of the 
spireme. 
The chromosomes now shorten and thicken, and the longitudinal split 
is, for the most part, obliterated, though it may still be distinguished in 
favourable cases. During this process the chromosomes become arranged 
about the periphery of the nucleus, appearing to undergo a mutual repulsion. 
As the contraction of the chromosomes proceeds, it becomes possible to 
count them ; they appear, as first stated by Guillermond ( 24 ) in 1904, to be 
sixteen in number (Fig. 25). 
During synapsis the nucleolus becomes closely pressed against the 
nuclear membrane, and assumes a characteristic sickle shape (Fig. 22). 
When the synaptic contraction loosens it remains close to the wall, and 
either retains its irregular shape or becomes once more rounded. By the 
time the chromosomes are fully formed, it is seen to be vacuolate in 
structure, but it persists, though diminished in size, till the late telophases 
of the division (Fig. 33). 
Spindle formation follows closely the process described by Harper ( 28 ) 
for Erysiphe , but the fibres are very delicate. The first stage to be 
recognized with certainty shows the two discoid centrosomes lying rather 
close together, with a separate cone of fibres radiating from each (Fig. 26). 
The centrosomes move apart, a spindle is established between them 
(Fig. 27) ; the radiating fibres become attached to the chromosomes, and 
have the appearance of drawing these on to the spindle (Fig. 28). A little 
