218 KIYOSHI M.VSCI. 



At about the same time the cliromatoicl lody imikes its appearaaice as a 

 round bodj near the nii.clcai- wall, being deeply stained -with acid dyes and 

 iron-haematoxylin (Figs. 21, 24). The mitochondrial granules appeal- abundantly 

 in the c^-toplasm, being scattered throughout the cell body. The granules 

 are larger than those of the spermatogouifi or the eaily spermatocyte (Fig. 

 26). 



The late prophase : — The nuclei now undergo a remarkable transfor- 

 mation, characteristic in many animals, in the coui'se of which the spiremes 

 bebome more distinct and begin to shoiieu (Figs. 28, 29). Later on the 

 spiremes shorten more and more luitil they assume long and cm-ved rod- 

 shape, and at the same time their staining capacitj- begins to increases 

 rapidly (Figs. 30, 31, 136). Soon after various bivalent chromosomes appear 

 ■which are chai'acteristic in the mouse (Figs. 32, 33, 35). According to their 

 characteristic form they are divided into the following two different trj-pes, 

 namely : 1. The first type is the single ring which is formed by the imiou 

 C)f two ends of the cmTed chromosomes. 2. The second type is the bivalent 

 rod which consists of two short chromosomes. The first type can be seen 

 only in large clnomosomes, while the smaller ones belong to the second typ?. 

 It is difficult to determine accurately in what manner the rings ai'e formed. 

 But from direct observation it seems most probable that they ai'e formed not 

 by separation of the longitudinal spUtting which sometimes appears at the 

 beginning of this stage in the preparation fixed with Champy's method. 



"When the nuclear membrane begins to disintegi-ate the chromosomes 

 shorten progressively until they assume short rod-shape (Figs. 34-36). 



The number of the bivalent chromosomes is obviously half of the diploid 

 niunber. 



At this stage the XY-chromosomes ctin not lie identified. The absence 

 of these chromosomes is probably due to the fact that in this stage they are 

 united to form a bivalent chromosome. 



Owing to its small size, iu this stage, the nucleolus is not chstinguish- 

 able with ordinary dyes (iron-hsematoxylin), whereas with ArEHBACn's method 

 it can clearly be made out (Figs. 136, 137). At the beginning of this süige 

 this body begins to disintegrate, leaving a small amoimt of chromatins lehiud, 

 which is probably used up iu the formatiou of the chromosomes (Figs. 135, 



