428 BOTANICAL GAZETTE [November 



centrosome has divided. In some cases, however, the daughter 

 centrosomes diverge somewhat, and very rarely they may reach 

 polar positions. Although we have observed no metaphase figure 

 of the third mitosis with undoubted centrosomes at the poles, it is 

 nevertheless probable that further search would reveal such cases. 

 It is thus possible that both Shaw and Belajeff were correct in 

 their interpretations, that they were dealing with two lots of 

 material showing different behavior at this point. 



This is apparently a stage in the life history where the centro- 

 some may be seen in the act of dropping out through failure to 

 carry out its function. A new one forms at each spindle pole for 

 the same reasons that one is developed at the preceding mitosis, 

 where it has been entirely lost from the earlier phases, though as yet 

 it is impossible to determine the nature of these reasons. Since 

 the variable behavior indicates that the organ in question is in all 

 probability a disappearing one, and since there is no organ other 

 than a centrosome which we should expect to see being eliminated 

 from spermatogenous cells, the lack of continuity, if it argues at 

 all, argues for the centrosome nature of the blepharoplast rather 

 than against it. 



The conclusions reached as the result of the present investiga- 

 tion are necessarily the same as those stated in the writer's paper of 

 191 2 and cited at the beginning of this discussion. Marsilia is 

 even more convincing than Equisetum in showing the direct deriva- 

 tion of an advanced cilia-bearing organ from a functional centro- 

 some. Since there is every reason to believe that the blepharoplasts 

 of bryophytes, pteridophytes, and gymnosperms are homologous 

 structures, it follows that they are all " ontogenetically or phylo- 

 genetically centrosomes." 



Summary 



1. In the first spermatogenous mitosis there is present at each 

 spindle pole a dense region with radiations, but no centrosome. 



2. During anaphase of the second mitosis a centrosome develops 

 at each spindle pole and at telophase divides to two daughter 

 centrosomes. These only rarely develop farther; they usually 

 degenerate at once in the cytoplasm. 



