326 ROBERT H. BOWEN. 



of anaphase phenomena, and this interpretation is strengthened 

 by a study of late anaphases in which a few threads still cross 

 from one cell to the other while the majority are already severed 



(Fig. 17). . 



The arrangement of the threads with reference to the spindle 

 is soon obscured and they spread out in each second spermato- 

 cyte in preparation for the next maturation division which occurs 

 immediately. As many cytologists have described (see the ac- 

 counts of Wilson ('12) and Montgomery ('n)), during the 

 anaphases of the first maturation division the halves of each cen- 

 triole separate in a direction at right angles to the old spindle, so 

 that the main axis of the second maturation spindle is at right 

 angles to that of the first. The centrioles are indeed already in 

 position at the conclusion of the anaphase, so that the distribu- 

 tion of the threads in the daughter cells can be followed with 

 reference to a possible further influence by the centers. As a 

 matter of fact, the threads are rearranged around the centrioles 

 as before, and the second maturation division is in general a half- 

 size replica of the first, at least up to the cutting in two of the 

 "palisade." The formation of the nebenkern from the mito- 

 chondria thus distributed to the spermatid will be taken up in the 

 next section. 



In the large cell generations the facts are essentially the same, 

 but the mitochondria are comparatively short, heavy rods with 

 numerous granules intermixed. These rods are oriented around 

 the centrioles in a most striking manner. Their behavior at the 

 close of the first maturation division is of special interest, for at 

 this time the influence of the centrioles on the rods is very clearly 

 demonstrated. In fact, as the " palisade " breaks up the rods 

 can be seen diverging toward the centrioles in two groups, show- 

 ing with the greatest clearness that the movements of the mito- 

 chondria during the entire division period are very definitely 

 oriented with respect to the centrioles. 



To summarize the behavior of the mitochondria in preparation 

 for the formation of the sperm, it may therefore be said: (i) 

 that in the early spermatocytes the mitochondria form a dense, 

 nuclear " cap," (2) which is broken up during the growth period 

 and distributed in a tangle of threads throughout the cell; (3) 



