386 K. MASUI; 
nucleus, a cytoplasmic body very faintly stained with iron-haematoxylin can be 
seen on careful observation (Figs. 43—45). It is represented only as an oval 
or a round mass. This body may be the idiozome, but the centrosome can 
not be found within it (Figs. 44, 45). It is impossible to assert whether the 
absence of the centrosome in the idiozome is due simply to its small size, or 
that the former has no connection with the latter. It seems to me, however, 
that the latter view is more probable. The same view is held by Macuipa (’17) 
in the spermatocyte of a grasshopper, Atractomorpha. The nuclear wall as well 
as the idiozome began to disappear, and the division of the idiozome could not 
be observed. In the horse, however, at the late prophase the idiozome is 
divided into two spheric bodies which eradually move apart from each 
other. 
In the growth stage the mitochondrial granules appear abundantly in 
the cytoplasm, being scattered throughout the-cell body (Figs. 61, 62). 
B. Tae REDUCTION DIVISION. 
The ordinary chromosomes in the first reduction division :—At the com- 
mencement of the metaphase when the chromosomes assume short rod shape, 
both the longitudinal split and the transverse constriction of the chromosomes 
become faintly visible (Figs. 63—65). In good polar view of the equatorial 
plate of the metaphase the number of chromosomes is always counted to be 
seventeen which represents the haploid number of chromosomes in the 
spermatogonia. In the horse two kinds of metaphase plates are distinguishable, 
namely, in one half of the metaphase plates eighteen chromosomes are 
represented and in the other half nineteen. The chromosomes in this stage 
show distinct and constant differences in size (Figs. 63—65). Even though 
the difference in size between neighbouring chromosomes is very little, five large 
sized, five small sized and two very small sized ones can be distinguished 
(Figs. 63—65). 
In the strongly decolourized preparations the side view of the metaphase 
of the first division shows that sixteen bivalent chromosomes are arfanged in 
the equatorial plate with their longitudinal split parallel to the axis of the 
spindle, and thus the transverse constriction of the chromosomes are at right 
angles to the sume (Fig. 67). The transverse constriction becomes more and 
