164 
OEOFFlv’EY SMFJ'H. 
tliiit eacli cliroinosoine is bivalent^ and represents two 
cln’oinosoines joined together in syintpsis, since about sixteen 
cliroinosoines can be counted in tlie earlier spevmatogonial 
divisions. On tlie completion of the first maturation division 
it seems tliat the second division is immediately begun. 
In the second division, instead of eight chromosomes 
a])pearing on the equatorial plate, four large chromosomes 
appear of the same size as in the preceding division. There 
can be very little doubt that these chromosomes are again 
bivalent, and that consequently a second synapsis of some 
kind normally takes place in bii'ds. This double synapsis Avas 
described for the pigeon by Guyer in 1900, and it has been 
observed by him since in several other species of birds (3 
and 4). Presumably this second reduction division is only 
a pseudo-reduction, and does not involve another halving 
of the chromosomes. After the second maturation division is 
completed, the resulting cells pass into a resting phase (figs. 
6 c and 7 c) as spermatids. The transformation of these 
spermatids into spermatozoa takes place in the usual Avay by 
the elongation of the cell-body Avith its nucleus, the chromatin 
in the nucleus becoming draAvn out into a spirally tAvisted 
thread (figs. 7 d and 8). In transverse sections, under a high 
poAver, the ripening spermatids have the appearance shown in 
figs. 6cZ and 7 d, Avhere the chromatin thread is seen to lie up 
against the nuclear membrane. 
To compare this normal process Avith Avhat occurs in the 
hybrids, reference must be made to the figures 9, 10 and II, 
PI. 8, and these figures should be compared with figs. 6 , 7 and 
8. The mitoses in the hybrid cells all represent the first 
maturation division, and it Avill be seen at once that in the 
first place the chromosomes are not arranged regularly on the 
mitotic spindle; secondly, that they are of very unequal and 
irregular sizes and that they are more numerous than in the 
normal spermatocytes. It is evident that a normal synapsis 
to form eight similar bivalent chromosomes does not occur 
in the hybrids; the separate chromosomes do not come 
together and fuse in an orderly Avay, but are scattered irregu- 
