STUDIES ON INSECT SPERMATOGENESIS. 347 



the clictyosomes which offer most interesting comparisons with 

 my own observations on the Hemiptera. He failed to confirm 

 Platner's description of a very accurate splitting of each dictyo- 

 some during the division stages, but his work bears out the con- 

 clusion that some sort of fragmentation occurs like that so clearly 

 demonstrated in Euschistus. Fragmentation of the original 

 Golgi bodies or " batonettes " is also indicated by the difficulty 

 which Gatenby ('18) experienced in staining the Golgi elements 

 during the division stages. I have had the same experience with 

 Euschistus, in which the Golgi bodies stain readily with Fe- 

 hematoxylin after the proper technique but the dictyosomes fail 

 entirely of demonstration. Gatenby ('17) also noted the aggre- 

 gation of his " acroblasts " (probably dictyosomes) around the 

 centrioles during the spermatocyte divisions of many Lepidoptera 

 but how they attain this position was not indicated. 



Deineka ('12) has described the division of the Golgi ap- 

 paratus in cells of Descemet's membrane, and the general outline 

 is distinctly similar to the type of behavior in Euschistus. The 

 dictyosomes formed by fragmentation of the Golgi reticulum 

 are arranged in an equatorial belt encircling the metaphase chro- 

 mosome-plate, and thence they wander to the opposite poles of 

 the spindle, becoming in this way divided into two distinct polar 

 groups. Finally, in Euschistus, I have been able to follow the 

 process step by step, and the relation of d-ictyokinesis to the gen- 

 eral mitotic figure is completely demonstrated. 



If we try now to gain some general conception of the processes 

 by which cytoplasmic division is accomplished, we note in the 

 first place one fundamental fact applicable to all cases. The 

 cytoplasmic elements are so distributed during the early division 

 stages that the constriction of the cell wall will ultimately divide 

 tJicm in some regular mass proportion. It has been held that 

 this distribution is a matter of chance only, but I believe that the 

 facts set forth above point rather to the existence of a definite 

 mechanism for the regular arrangement of the cytoplasmic ele- 

 ments in division. Indeed, the case of Eusclustus provides con- 

 vincing evidence that the mitochondria and Golgi apparatus may 

 undergo a division accomplished in some way by activities that 

 focus in the centrioles. The evidence from other sources tends 



