STUDIES ON INSECT SPERMATOGENESIS. 32! 



sperms are of equal value in fertilization. I have nevertheless 

 found these size differences of great practical value in cytological 

 studies, for one has at his command on the one hand a complete 

 series of stages in miniature, and on the other a selection of the 

 same stages reproduced on a larger scale. In the description 

 which follows I shall refer to the dimegalous cells simply as 

 " large " and " small," since the unusually small generation pre- 

 sents no points of special interest. 



II. Stages Preparatory to Sperm Formation. 



The history of the nucleus from the spermatogonial cell 

 through the close of the second maturation division has been so 

 minutely studied in the Hemiptera (see more particularly Mont- 

 gomery ('n) and Wilson ('12)) that further description would 

 be superfluous. I will accordingly proceed at once to a consid- 

 eration of the formed elements of the cytoplasm with which I 

 have been more especially interested. 



A brief reference must first be made to the centrioles, since a 

 knowledge of their behavior is essential to a proper understand- 

 ing of other cytoplasmic phenomena. At the close of the final 

 spermatogonial division the cells are so small and crowded that 

 the centrioles can not be identified. They may first be made out 

 with certainty in the spermatocytes during the growth period 

 and, when once relocated, can easily be followed through to the 

 close of the second maturation division. The point to be noted 

 is that when the centrioles reappear they are already located on 

 opposite sides of the nucleus in position for the formation of 

 the first maturation-spindle. Furthermore, the nuclear mem- 

 brane bulges out and the cell wall is drawn in opposite each 

 centriole (Fig. n), so that the centrioles touch (or nearly touch) 

 the wall of both nucleus and cell, the cytoplasm in the neighbor- 

 hood of the centrioles being thus greatly restricted. A section 

 through both centers accordingly gives a very characteristic pic- 

 ture (see Montgomery ('n), Fig. 63). 



Of the remaining cytoplasmic constituents, the mitochondria 

 are the best known and have already been described in Euschis- 

 tus by Montgomery ('n). His account was, however, based on 

 faulty material and I have found it necessary to make a reex- 



