1920] SHARP— SPERMATOGENESIS 263 



the form of a short lumpy rod or series of granules, and at no time 

 does the blepharoplast have the form of a long slender thread free 

 from the nucleus as in Woodburn's figure. The writer, therefore, 

 is inclined to attribute the disagreement for the most part to actual 



differences in the material studied rather than to differences in 

 interpretation. 



The phenomenon of fragmentation is probably the most inter- 

 esting feature of the blepharoplast of Blasia. In all previous 

 accounts of bryophyte spermatogenesis, including those of Ikeno 

 (6) on Marchantia, Wilson (io) on Pellia, Polytrickutn, and Atri- 

 chum, Woodburn (n, 12, 13) on several liverworts and Mnium, 

 Miss Black (3) on Riccia, and Allen (2) on Polytrichnm, the 

 blepharoplast is reported to elongate without breaking up into 

 smaller portions. Allen (2) states that "while the possibility of a 

 somewhat similar occurrence [fragmentation] is suggested by the 

 rather knotty appearance of the blepharoplast of Polytrichum when 

 it begins to elongate, there is no time when it is visibly resolved into 

 smaller bodies.' ' In Blasia, therefore, we have the only known 

 instance in bryophytes of such a fragmentation of the blepharo- 

 plast as occurs in Equisetum, Marsilia, and the cycads. 



Although fragmentation is in general a characteristic of the 

 blepharoplasts of the cycads, and only occasionally found in 

 pteridophytes (Equisetum and Marsilia), it is now evident that it 

 may occur in forms lower in the scale. Moreover, it is seen that 

 it is not, as might be supposed, merely a means by which large 

 blepharoplasts become transformed, for the blepharoplasts of 

 Equisetum and Marsilia, and especially those of Blasia, are very 

 small. Although the details of the process of fragmentation differ 

 in the various cases (by simple fission in Blasia and by vacuoliza- 

 tion in the other forms), it is scarcely to be doubted that the phe- 

 nomenon is a result of similar causes in all. In attempting to 

 find a possible historical reason for it, one is struck by the resem- 

 blance between the fission of the blepharoplast in Blasia (fig. io) 

 and the division of an ordinary centrosome before mitosis. If the 

 blepharoplast actually represents a centrosome, as the writer (8) 

 believes the evidence indicates, it is at least possible that its fre- 

 quent fragmentation, in spite of the fact that in the more advanced 



