STUDIES ON INSECT SPERMATOGENESIS. 339 



The acroblast is a more or less hemispherical body, pressed 

 close to the nuclear membrane (Fig. 21), the periphery of which 

 impregnates very heavily except for the portion touching the 

 nucleus, where the substance which takes the osmic acid seems 

 to be absent. The substance within the peripheral layer blackens 

 very little and the picture as a whole recalls the similar arrange- 

 ment of substances in the Golgi bodies of the growth period. 

 The acroblast is completed at about the time the centrioles mi- 

 grate around the nucleus to their position near the nebenkern, 

 and is always located on the side of the nebenkern opposite that 

 on which the centrioles will come to lie. As the centrioles migrate 

 into their definitive position, the acroblast moves around the 

 opposite side of the nucleus to an anterior pO'Sition (Fig. 23). 

 After pausing for a time (Fig. 24) it continues on around the 

 nucleus, retracing the path over which the centrioles had trav- 

 eled, and eventually comes to rest close to the centrioles (Fig. 

 25) but, with reference to its former position, on the opposite 

 side of the nebenkern. Here it remains until the final step in 

 the formation of the acrosome is completed. 



As the acroblast slowly makes its way around the nucleus of 

 the spermatid it undergoes a process of differentiation which, 

 there is reason to believe, is of fundamental importance. The 

 first evidence of this is the appearance at the base of the acro- 

 blast that is at the surface next the nuclear membrane of a 

 transparent, spherical vesicle or bubble (sometimes there is more 

 than one) which seems to be forming out of the colorless sub- 

 stance inside the acroblast (Fig. 23). At first it is very small 

 but, growing rapidly, it gradually protrudes farther beyond the 

 border of the acroblast which is thus lifted away from the nuclear 

 membrane (Fig. 24). This transparent, structureless vesicle is 

 to be the acrosome, and by this name it may be conveniently re- 

 ferred to even at this early stage in its formation. It is always 

 in contact with the nuclear membrane while the acroblast fits 

 over it like a shell. As the acrosome continues to enlarge, the 

 acroblast seems to diminish somewhat in size, apparently at the 

 expense of the non-staining substance, and eventually comes to 

 form a mere appendage to the acrosome itself (Fig. 28). At the 

 point where the acrosome touches the nucleus, a small granule 



