304 THOS. H. MONTGOMERY, 



directly referable to the variation in the length and volume of the 

 chromosomes before the synapsis stage, providing, what is true within 

 certain bounds, that the number of granules stands in more or less 

 direct ratio to the volume of the chromosome. These counts were 

 made upon chromosomes of the earlier synapsis, while some of the 

 granules are destined to still further subdivide; in the later synapsis 

 and telophase their number would be greater, but certainly not more 

 than twice as great, and probably less than this. 



Commencing about synchronously with the above stages of de- 

 condensation of the chromatin, and continuing until after them, pro- 

 ceeds a longitudinal splitting of the granules. Occasionally this 

 splitting appears in the early anaphase before the appearance of the 

 nuclear membrane (Figs. 60, 61, Plate 19), but such splitting seems 

 to be precocious. This splitting consists in the division of each 

 granule into two, the plane of division passing through the long axis 

 of the chromosome (Figs. 63-66, 69, 71, 73, 74, 80—84, 88); the 

 plane in which two daughter granules lie coincides at first quite 

 regularly with the plane joining the two arms of the V-shaped chromo- 

 some, but later becomes more or less twisted out of this plane. 

 The process begins with a flattening and elongation of each granule 

 (Figs. 65—67, 70 — 84, 88—90), and usually begins and is completed 

 sooner at the middle than at the ends of each univalent chromosome ^). 



All these stages may be followed very clearly in JPerix>atus, and 

 result in the splitting of each and every granule (Figs. 91 — 97). The 

 split separating corresponding daughter granules becomes from now 

 on gradually wider, and is most prominent about the beginning of the 

 telophase; in the rest stage of the spermatocytes it is really widest 

 but at this stage is more difficult to determine. 



The longitudinal splitting may commence, sometimes is fully ended, 

 before the complete decondensation into granules has been effected; 

 in such cases each univalent chromosome appears as two parallel rods 



1) In consequence of the flattening and elongation of its granules 

 in one plane, the chromosome appears narrower on edge than on sur- 

 face view. In the Figs. 85 — 87 showing cross sections of the uni- 

 valent chromosomes at this stage, the sections of the chromosomes 

 appear oftener round (as I have shown them) than fiat. This optical 

 effect is due to the difficulty of seeing only one granule at a given 

 focus of the microscope, granules above and below the plane of view 

 obscuring a clear view of a single one, and thereby causing the mass 

 to appear rounded rather than llat. 



