318 bulletin: museum of compakative zoology. 



mosomes results. In its essentials the process, which is a rather dif- 

 ficult one to trace, is probably as follows. The chromomeres draw 

 together in pairs (as in the cell represented in Figure 60), which finally 

 fuse so completely that the component parts can no longer be distin- 

 guished. The cell shown in Figure 60 is particularly instructive, since 

 in it six such pairs have already formed, while the remaining (twelve) 

 chromomeres are still entirely separated from eacli other. In Figure 61 

 the process is practically completed, although the double nature of each 

 of the chromosomes is still apparent, and even the connecting linin 

 strands yet persist. Simultaneously, upon the fusion of the chromo- 

 meres, the twelve resultant chromosomes arrange themselves in the 

 "equatorial plate "for the first maturation division. The chromosomes, 

 to be regarded on account of their origin and reduced numbers as 

 bivalent, are thus formed by the pairing of preexisting chromatin struct- 

 ures, in exactly the same way that they are in the case of the sperma- 

 togonia! division. In discussing the probable meaning of the pi'eseut 

 type of numerical or " pseudo " reduction of the chromosomes it is 

 essential to keep this fact clearly in mind. 



Metaphase. — As stated above, the chromosomes in the metaphase of 

 the first maturation division (Fig. 61) are only half as numerous as the 

 somatic chromosomes. The spindles at this phase are of two kinds, pre- 

 senting in side view very different aspects, a fact which for a long time 

 proved very puzzling (Figs. 64, 65) ; but careful study of many prepara- 

 tions has convinced me that the differences are apparent rather than 

 real, and are due to external causes, not, as I at first supposed, to the 

 occurrence of two types of mitosis. The apparent differences between 

 the two are shown in Figures 64 and 65. The form represented in Fig- 

 ure 64 is by far the more common, and is no doubt to be regarded as 

 the typical one. The chromosomes, here distinctly dumb-bell-shaped, 

 are elongated in the direction of the spindle axis ; they are apparently 

 dividing, not by means of a longitudinal split, but by a gradual process 

 of pulling apart. With careful focussing they can often be counted with 

 comparative accuracy, even in side views of the spindle, and there is no 

 doubt but that they are present in the reduced number, probably twelve. 

 Such cells are very numerous, — in sections, always densely packed and 

 usually occurring in groups of from five to twenty, amid spermatocytes 

 of the same generation, but in different stages of the prophase. They 

 lie near the margin of the gonad immediately beneath the peripheral 

 layer of secondary spermatocytes, spermatids, and spermatozoa. The 

 chromosomes are seldom well shown in polar views owing to their great 



