516 Edmund B. Wilson. 



condense into the compact form appears to vary considerably, 

 for they cannot always be distinguished until the later growth- 

 period, and it should be noted that during the pale period the 

 nuclei often show a variable number of smaller deeply-staining 

 granules. I believe, however, that there can be no doubt as to 

 the nature of the two larger bodies on account of their great con- 

 stancy, their size, and the completeness of the series that connects 

 the earlier with the later conditions (such as is shown in Fig. I, <:), 

 where no doubt of their nature can exist. The persistence of the 

 larger chromosome-nucleolus ("accessory") throughout all these 

 stages without any considerable change renders it manifestly 

 impossible that it should give rise to the ra-chromosome bivalent, 

 either directly as assumed by Paulmier and Montgomery, or by 

 division into two univalents that subsequently conjugate, as 

 described by Gross in Syromastes. 



In the early prophases the larger chromosomes resume their 

 staining capacity and condense into characteristic cross-forms 

 (Fig. i, <:), and finally into compact quadripartite tetrads or bipar- 

 tite bodies. At this time the heterotropic chromosome assumes 

 a dumbbell or quadripartite shape, and the w-chromosomes, 

 which are still quite separate and may even lie on opposite sides 

 of the nucleus, also frequently become bipartite. The nucleus now 

 contains, accordingly, eight separate chromatin-elements, one more 

 than the number of bivalents in the first mitosis, as is also the case 

 in Archimerus and Anasa, as described beyond. As the spindle 

 forms the two microchromosomes lose their bipartite shape, 

 approach each other, and in the stage just preceding the metaphase 

 finally conjugate to form the small bivalent chromosome at the 

 center of the group. Without fusing, the two halves are then 

 immediately separated, the division always taking place more 

 rapidly than in the case of the larger chromosomes (Fig. I, /). 



It is clear to demonstration accordingly, that in Alydus the small 

 central bivalent does not arise from the large chromosome-nucleolus 

 of the growth-period, but is formed by the late conjugation of two 

 separate microchromosomes that have no genetic connection with 

 that body. The same fact is shown no less clearly in Archimerus 

 calcarator (which shows eight chromosomes in the first mitosis), 

 where the m-chromosomes, and the corresponding bivalent, are 

 of extraordinary minuteness and are so much smaller than the acces- 

 sory that they could not possibly be confused with the latter (Fig. 3). 



