INSECT SPERMATOGENESIS 607 



very carefully examined, especially in serial cross-sections of 

 the nebenkern, and it is now clear that this guess was correct 

 in every particular. As in the Hemiptera the division of the 

 nebenkern is foreshadowed by the symmetrical disposition of 

 the chromophilic material (tigs. 3, 5, and 45), and the division 

 itself is accomphshed in the regions unoccupied by chromo- 

 philic material soon after elongation begins. In the region of 

 the chromophilic material itself, however, the division is not 

 (usually) completed until after the final act of dissolution, 

 a point in which the Lepidoptera agree with the Coleoptera in 

 which the splitting of the nebenkern is delayed in a somewhat 

 similar manner. In the Lepidoptera, however, there seem often 

 to be more or less local irregularities in the division process, 

 and it thus happens not infrequently that the final remnant of 

 the chromophilic material is left to complete its dissolution in 

 one of the nebenkern halves, while the division plane is com- 

 pleted (fig. 9 b). The general features of the division process 

 as outhned above are well shown in figs. 9, 12, and 47. In 

 fig. 9, which represents a nebenkern at the stage of fig. 8, 

 the division above (and below) the chromophilic substance is 

 completed (fig. 9 c), but in the region of the plate-work it is 

 still incomplete (fig. 9 a), with the exception of cases like 

 fig. 9 B already noted. In fig. 47 cross-sections of a nebenkern 

 like fig. 46 are shown. The more spun-out portions at the ends 

 of the nebenkern masses are shown in fig. 47 c, while the parts 

 nearer the middle are shown in fig. 47 b, and the region of the 

 plate-work itself in fig. 47 a. Comparison of figs. 3, 5, and 6, 

 with figs. 9 a and b, and fig. 45 with fig. 47 a, shows clearly 

 how the structure of the plate-work becomes progressively 

 simplified as the chromophilic material condenses. Just 

 before its final dissolution the plate-work is reduced to a simple 

 ovoid shell (figs. 9 a and 14), such as I have described in 

 Hemiptera and Coleoptera. (Compare also the figures of 

 Doncaster and Cannon (1920) in the louse.) In these late 

 condensation stages the plate-work, as against the thread-like 

 structure of the chromophilic material, seems to me unques- 

 tionable. As the figures show, the axial filament lies in the 



T t 2 



