102 DuGGAR : Development of the Pollen Grain 



and the few large chromatic granules are replaced by a more 

 chromatic ribbon, and by man}- more \'ery evident granules, giv- 

 ing to the ribbon a noduhite appearance, as in Fig. 35. Moreover, 

 the ribbon is coiled and looped in definite curves ; and it may 

 be readily seen that there arc no anastomoses of the general 

 thread. 



The process of chromosome formation seems to follow the 

 general method outlined for the first division of the pollen mother- 

 cell, except that here no such definite loops or rings have been 

 observed. Radiating from the nuclear . membrane in the direction 

 of the axis, especially, are kinoplasmic threads. The disappear- 

 ance of the nuclear membrane initiates the formation of a loose 

 spindle in every way equivalent to the one first formed in the 

 pollen mother-cell. The chromosomes are again irregularly scat- 

 tered in the region of the nuclear plate, and the characteristic 

 heterotypic division of this plant is unmistakable (Fig. 38). In 

 this case, however, there is no large space free from trophoplasm 

 immediately surrounding the spindle. I have also been able to 

 estimate the number of chromosomes in this division, and it cor- 

 responds to the reduced number of the male archcsporium, about 

 twenty fiv^e.- 



After the formation of the daughter nuclei, the chromosomes 



may be identified for a time (Fig. 39), but as no later telophase 

 stages were found, I have no notes concerning the reappearance of 

 , the nucleolus, or the early stages in the formation of the second 

 spindle. When the second spindle is complete (Fig. 40), a glistening 

 cell wall separates the two daughter-cells, there is no indication of 

 the former spindle fibers, and the c}'toplasm is contracted or re- 

 pelled from the newly-formed wall. The chromosomes are 

 arranged at the nuclear plate in a definite plane, and the charac- 

 teristic homotypic division is e\'idcnt. The}- separate longitudinally, 

 and the bacilloidal daughter segments pass to the poles on a defi- 

 nite alignment, as in the corresponding division in the microsporic 

 development. 



There is every reason to believe that these di\'isions are truly 

 homologous with the two divisions in the pollen mother-cell, and 

 that here we have the reducing divisions preceding the formation 

 of the female sexual nucleus. Of the monocotyledons yet pub- 



