190 ROBERT H. BOWEN. 



with the plastidome of root-tip cells, viz., their orientation at 

 the spindle poles during cell division, as observed by Allen. 

 With the spindle itself, however, I can not find that they have 

 anything to do. I have followed all stages in the division of 

 rings like those of Fig. 9, preparatory to cell division, but I 

 have not been able as yet to verify Allen's account of their 

 fragmentation in the older generations of androgonia. The mode 

 of their distribution to, or their morphology in, the androcytes 

 has not yet been certainly made out. In general it may be 

 said that the plastidome of moss androgonia shows an essential 

 similarity to that of the ordinary meristem cells of spermato- 

 phyte root-tips. 



It remains now to follow the history of these various com- 

 ponents in the androcytes as they gradually transform into 

 sperm cells. The history of the osmiophilic platelets (called 

 spherosomes in an earlier report (Bowen, '266)) will first be 

 indicated. In the early androcytes, the scattered osmiophilic 

 platelets gather gradually into an irregular, elongated heap 

 lying along one side of the nucleus. It was my first belief, 

 borne out by hematoxylin preparations, that this was the 

 limosphere of authors, gradually forming by fusion of the 

 platelets. It is clear now that the whole course of events is 

 more complicated, and that the limosphere is by no means so 

 simple a structure as has been supposed. Actually the osmio- 

 philic platelets are thus assembled in connection with the 

 formation of a new structure to which it will hereafter be con- 

 venient to limit the term "limosphere," regardless of the fact 

 that as originally used by Wilson ('n) it probably involved the 

 platelets proper to some extent at least. The limosphere appears 

 first amid the platelets as an elongate body, often duplex in a 

 later stage, but finally rounding up to form a conspicuous clear 

 sphere within which is a large darkly staining granule (Figs. 

 ii to 14). This is the typical limosphere of Allen's account, 

 its earlier stages having been overlooked for the most part by 

 previous workers. During the later stages in its development, 

 the osmiophilic platelets show a less intimate relation to the 

 limosphere (compare Figs, n and 14), and subsequently the 

 relation becomes very loose, the platelets being even more 



