188 Darling : Sex in dioecious plants 



one. This seems to be a very general phenomenon. In a very 

 few cases, however, two bodies were seen in each daughter nucleus 

 (fig. 27). Many successive stages were observed in the same 

 loculus of an anther (figs. 25-30 were taken from the same locu- 

 lus) so there can be little doubt about the general order and suc- 

 cession of these changes. 



As a general rule the secondary chromatin body to be found 

 in the one nucleus ultimately fuses or coalesces with the larger 

 one so that each daughter nucleus simulates a resting condition 

 (fig. 29). At this stage the nuclei appear very similar to the 

 nucleus of the mother cell, possessing the small non-staining 

 bodies about the nuclear wall. Occasionally the secondary 

 chromatin mass found in the one nucleus does not fuse with the 

 larger one ; this lack of fusion is probably due to the shortness of 

 time which intervenes between the first and second divisions. 



Owing to the rapidity of development and to the small size of 

 the nuclei, the formation of the chromosomes for the second divi- 

 sion could not be followed with the detail obtained for the first. 

 No spireme was noticed during these stages, but in many cases 

 the spherical chromatin masses were clearly seen coming from the 

 nucleolus, as observed in the formation of the thread for the first 

 division. It should be noted incases where the secondary chrom- 

 atin mass has not fused with the larger one, — as sometimes 

 occurred in the reconstruction of one of the daughter nuclei after 

 the first division, — that this secondary mass as well as the large 

 one gave off spherical bodies which served to build up the spireme 

 thread (fig. 31). In later stages many nuclei were found that 

 showed the bivalent chromosomes coming from the nucleolus 

 (figs. 30-32), clearly indicating the same general phenomenon as 

 occurred in the formation of the chromosomes for the first divi- 

 sion. The number of chromosomes formed in this manner could 

 not be determined with accuracy, but there is no reason to believe 

 that there is a difference in behavior from the first division, where 

 eight were formed by the segmentation of the spireme and five 

 from the nucleolus. 



In the polar view of the metaphase stage of the second division 



i 



(fig. 33), thirteen chromosomes are to be seen in each plate. 

 Owing to the small size, the division of the chromosomes could 



