STUDIES ON INSECT SPERMATOGENESIS. 323 



tion the mass of threads enlarging with the general growth of 

 the cell as a whole. 



With the prophases is inaugurated a most interesting series 

 of events to which I would like to call special attention. In this 

 particular period, including the prophases of the maturation 

 divisions and also the divisions themselves, the essential phe- 

 nomena are the same in both large and small generations ; but 

 the cell pictures differ somewhat in the fact that the mitochon- 

 dria in the small cells are in general thread-like, while in the large 

 cells they are in the form of granules and rather thick rods of 

 no great length. The following account is based on the small 

 cells, brief details of the large ones being added by way of com- 

 parison. 



The early prophases of the first maturation division present 

 nothing of special interest. Montgomery's figures of the mito- 

 chondria in this stage are essentially correct, except that the 

 threads appeared to him very much thinner than is the case when 

 the Benda stain is employed. In the middle prophases a change 

 in the arrangement of the threads becomes obvious though its 

 exact nature is difficult to state. Occasionally in a favorable cell 

 it is possible to make out a number of ends of threads directed 

 toward a centriole, but at this time one would certainly never 

 guess the nature of the ultimate arrangement of the threads of 

 which these uncertain phases give the first inkling. In the final 

 prophases the rearrangement of the threads is completed and 

 the nature of the whole process becomes entirely clear. As now 

 appears, the tangle in which the threads seemed to be at the close 

 of the growth period was much more apparent than real. The 

 result of the rearrangement of the threads is best seen in sections 

 containing one or both of the centrioles which are now located at 

 opposite sides of the nucleus, thus marking out the long axis of 

 the future spindle. Hence it is convenient to anticipate the 

 spindle and speak of polar and lateral aspects of the spermato- 

 cyte, in the sense in which those terms are customarily applied to 

 metaphase figures of a dividing cell. In a lateral view of the 

 late prophase (Fig. n), the arrangement of the mitochondria is 

 very characteristic. In the cytoplasmic zone immediately sur- 

 rounding each centriole the mitochondria are entirely absent, the 



