STUDIES ON INSECT SPERMATOGENESIS. 337 



tion. Students of the mitochondria in Mollusca and Lepidoptera 

 have long supposed the chondriosomes to be composed of two 

 different substances an outer chromophilic envelope staining 

 intensely with crystal violet and related stains, and an inner, 

 medullary or " chromophobe " material which stains little or not 

 at all with all the usual reagents. By means of Cajal's Go'lgi 

 method I have succeeded (Bowen, '19) in impregnating this 

 medullary substance with the greatest clearness, the chromophilic 

 envelope being left quite colorless and transparent. Fuller treat- 

 ment of these observations lies beyond the scope of this paper, 

 but I would like to emphasize that a duplex chemical structure 

 has thus been conclusively demonstrated at least for some mito- 

 chondria. 



Finally, it remains to trace the role of the Golgi apparatus in 

 the differentiation of the sperm a subject to which a rather 

 more complete treatment may be accorded by reason oi its great 

 theoretical importance. The preceding section gave an account 

 of the Golgi apparatus up to the final anaphase of the second 

 maturation division, when the daughter chromosome plates have 

 completed their journey to the spindle poles and the dictyosomes 

 are grouped in two clusters close to the chromosome plates and 

 encircling the spindle remains (Fig. 18). I have followed out 

 the further stages in both Brochymena and Euschistus, where 

 they are essentially the same. In Brochymena the mitochondria 

 seem to condense directly into the nebenkern, while the dictyo- 

 somes, already fusing with each other to form larger masses or 

 Golgi bodies, are scattered over its whole available surface (Fig. 

 19) . They are always thus closely in contact with the nebenkern, 

 showing no tendency toward a more general distribution, and 

 until the final fusion of all the Golgi elements they retain this 

 close connection with the surface of the nebenkern. In Euschis- 

 tus, on the contrary, the Golgi bodies while remaining in the 

 vicinity of the nebenkern tend to be rather scattered. 



The fusion of the Golgi bodies continues rapidly, their number 

 constantly decreasing while their size correspondingly increases 

 (Fig. 20). Soon definite little aggregates more or less plate-like 

 or spherical in shape are formed whose periphery impregnates 

 heavily just as did the Golgi bodies in the growth period, this 



