STUDIES ON INSECT SPERMATOGENESIS. 59 



her of small vacuoles in the periphery of the mitochondria! mass 

 (Fig. i). These are at first very indistinct and difficult to demon- 

 strate,, but they presently become clearer and form a complete 

 layer investing the periphery of the nebenkern. They seem to 

 stain little or not at all, in contradistinction to the central mass 

 which is more darkly colored. This may be due to the fact that 

 the process of differentiation is going on from without inwards, 

 or it may be due to the optical superposition of several layers of 

 vacuoles. I am rather inclined to think that the process of differ- 

 entiation is going on throughout the whole mass, but that the peri- 

 pheral portions perhaps progress more rapidly. It is a curious 

 fact that in many Bencla preparations which seem to be other- 

 wise excellent, these and the later differentiation phenomena are 

 not demonstrated at all. Further, I have not been able to get 

 any preparations (of Brocliyincna*) stained with sufficient sharp- 

 ness to show these vacuoles as they should, of course, appear on 

 the rounded surfaces of the nebenkern as one focuses up and 

 down. Zweiger ('07), however, seems to have made out this 

 stage in Forficula (see his Fig. 36), in which the whole neben- 

 kern seems made up of very small vacuoles. 



The process of differentiation by which these vacuoles are pro- 

 duced has doubtless been proceeding for some time, but is made 

 visible only when the vacuoles attain dimensions within the 

 reach of our technique. At all events, Fig. i represents only a 

 transitory condition in a continuous process which is progressing 

 toward a definite end. The successive figures which I have 

 drawn are merely selected at convenient points along the way. 

 Fig. 2 represents a stage somewhat later than the preceding figure,. 

 in which the nature of this vacuole formation has become much 

 more evident. It is clear now that the formation of non-staining 

 vacuoles is going on throughout the nebenkern mass, the outer 

 ones apparently being differentiated more rapidly than the more 

 interior ones. Thus it comes about that these vacuoles are ar- 

 ranged in rather definite layers which, in cross-section, would 

 appear as concentric rings (Fig. 2). 1 The outermost layer clears 



i This corresponds approximately to the period at which the centriole 

 migrates from its originally anterior position (Fig. i) to its definitive position 

 at the base of the sperm head (Fig. 3)- (See Bowen, '22.) 



