Helen Dean King 357 
the time that mitosis is completed by the division of the chromosomes 
through the middle of the dumbbells (Figs. 41, 42, 44, 45, 48), there 
seems no possibility of avoiding the conclusion that the first maturation 
mitosis in the spermatocytes of Bufo is a transverse or reduction division 
in the Weismannian sense, separating chromosomes that were united 
end to end in synapsis. 
In some of the spermatocytes the spireme formed after the synizesis 
stage does not produce dumbbell-shaped chromosomes, but breaks into rec- 
tangular blocks (Fig. 35). These rectangular shaped chromosomes also 
maintain their definite shape up to the metaphase, and are never divided 
either for the first or for the second maturation mitosis before this 
time. Owing to this persistence in the shape of the chromosomes there 
can be no doubt but that here too the first division is a reduction division ; 
as in the anaphase the chromosomes are nearly square and not long and 
narrow as would be the case if they were divided longitudinally (Fig. 52). 
An examination of Figs. 30-36 will show that before the separation 
of the chromosomes there is always a considerable variation in the size 
of the segments in the same nucleus. This difference in the size of the 
chromosomes is maintained throughout all of the subsequent stages of 
mitosis and shows very clearly in the equatorial plate of the first matura- 
tion spindle shown in Fig. 43. All twelve of the chromosomes are 
present in this case and two of them are considerably larger than the 
others. In the early prophase of the division of the primary spermato- 
gonium shown in Figs. 2 and 3, there are four of the chromosomes that 
are noticeably larger than the rest. It seems very probable that these 
chromosomes maintain their individuality throughout all of the sperm- 
atogonial divisions and are united two and two in synapsis to reappear 
in the prophase of the first maturation mitosis as the two bivalent 
chromosomes that are readily distinguished from the others by their 
size. If this is true for four of the spermatogonial chromosomes it is, 
of course, true for all. Thus the twelve chromosomes shown in Fig. 43 
are bivalent structures formed by an end to end union of homologous 
chromosomes which, as Montgomery (39) has shown, are to be traced 
back through the divisions of the secondary spermatogonia to the primary 
spermatogonia. 
The difference in the size of the chromosomes on the first maturation 
spindle is strikingly illustrated in Fig. 48 where one of the largest of 
the chromosomes lies next to one of the smallest, the mass of the former 
being more than twice that of the latter. In the metaphase and early 
anaphase of the second maturation division (Figs. 56, 57), one can also 
