AND OF HYBRID PIGEONS. 
13 
ment is derived directly from a chromatin nucleolus. The 
chromatin nucleoli when present lose their identity apparently 
and are absorbed or built up into the spirem when the prepara¬ 
tions for division begin. Moreover, the accessory chromosome 
of McClung retains its individuality and persists throughout 
the entire period of spermatogenesis to take part finally in the 
formation of the spermatozoa. In the pigeon the filament, as 
has been stated, is cast out into the cytoplasm to apparently 
disintegrate and disappear. During the succeeding division 
numerous small black dots (Fig. 5) are seen scattered about in 
the cytoplasm. They seem to be due in part to the presence of 
this cast out chromatin and in part to small globules which, 
judging from their reaction to reagents, are of a fatty nature. 
After the chromosom.es have undergone their transforma¬ 
tion inside the nucleus and have approached the nuclear mem¬ 
brane, the centrosome which is seen lying in a clear area of the 
sphere, divides into two, one of which moves along the outer 
periphery of the nucleus to the opposite pole. The first ap¬ 
pearance of the spindle fibers is as radiations which spread 
around the nucleus from the centrosome. No further detail 
could be made out regarding the origin of the spindle. 
The nuclear membrane which has persisted in a very dis¬ 
tinct form so far disappears and the next stages show the 
chromosomes in connection with the spindle. The polar radi¬ 
ations extend only a short distance until lost to view. They 
are but poorly developed at any of the divisions. Fig. 5 shows 
a side view of spermatogonium just ready for division. The 
spindle is short and broad and the chromosomes lie in a con¬ 
fused band at the equator. The individual chromosomes are 
for the most part in the shape of short bent rods or loops, with 
the closed end of the loop turned towards the center of the 
spindle. Their form, which can be recognized only with diffi¬ 
culty, is seen to best advantage in a favorable view from one 
pole. They often split longitudinally before the spindle is 
formed and this adds to the confusion when massed. In divid¬ 
ing, if they have not already halved before coming to the 
spindle, they split lengthwise, beginning at the inner closed end. 
While moving towards the poles it is not uncommon for the free 
ends of a chromosome to fuse, thus forming a small ring or 
