422 BOTANICAL GAZETTE [November 



separated from each other by the sterile cells cut off by the walls 

 numbered 4, it seems reasonable to hold with Belajeff that the 

 microspore develops two symmetrically placed antheridia separated 

 by wall 2. The prothallial cell degenerates during the later an- 



theridial divisions. 



The foregoing account agrees in its essential features with that 

 given by Belajeff (2) for M . elata, but in that form the develop- 

 ment in the two hemispheres is not so symmetrical. In a very 

 few cases we have seen wall 3 in the upper hemisphere strike the 

 spore wall rather than wall 2, which results in a condition more 



like that in M. elata. 



spermatogenesis 



primary 



permatids. In these mitoses 



chromosomes behave as in ordinarv vegetative mitoses 



view 



description. Attention will therefore be directed wholly to the 



centrosomes and associated structures. 



rmatogenous mitosis. — The primary spermatogenous cell 

 rig which can be called a centrosome. The cytoplasm 



P 



many 



em 



without 



he spindle forms 

 At late prophase 



me 



during anaphase these regions appear denser (fig. 14)- A little 

 later long and very distinct radiations develop about each pole; 

 at their focus there is a dense and finely granular appearance 

 (fig. 15), but no distinct body is formed. At telophase these polar 

 achromatic structures disappear; in fig. 16 their last remnants may 

 still be seen. 



Second spermatogenous mitosis. — In the prophases the second 

 mitosis is similar to the first. The spindle at first ends indefinitely 

 (fig. 17), but during early anaphase it rapidly becomes pointed. 

 Long radiations develop as in the preceding mitosis, and at each 

 pole a very minute and intensely staining granule appears (fig. 18). 

 This is the centrosome. On account of its extremely small size it 

 is practically impossible to make out the exact manner of its origin; 



