STUDIES ON INSECT SPERMATOGENESIS. 341 



the contrast between the accurate, meristic division of the nu- 

 cleus and the supposedly haphazard, mass division of the cyto- 

 plasm as one of the striking features of karyokinesis. In part, 

 no doubt, we have been led to an easy acceptance of this view by 

 what appear to be the necessary implications of the modern 

 chromosomal theory of heredity, which, by its emphasis on the 

 exactness of chromosome distribution, perhaps predisposes us to 

 assign a secondary importance to the division of the specific com- 

 ponents of the cytoplasm. Such a standpoint is unwarranted, for 

 while it may be improbable that any mechanism will ever bfe 

 made out for an exact division of the cytoplasm comparable to 

 that of the nucleus, it does not necessarily follow that cyto- 

 plasmic division can be reduced to the basis of a mere mass divi- 

 sion subject only to the laws of chance. Nevertheless, at the 

 present time cytoplasmic division seems often to be treated as if 

 it were a haphazard process, the essential phenomena of which 

 are the same whether cell division be mitotic or amitotic, with of 

 course an implied exception in the case of the centrioles and the 

 achromatic division figure. 



It seems to me, however, that the facts observed in the Hemip- 

 tera point rather to a conception of cytoplasmic division as a 

 precisely ordered process. That the cytoplasm may indeed un- 

 dergo division of a remarkably regular character has long been 

 recognized in the special case of the segmenting egg ; and in the 

 so-called " determinate " eggs particularly of molluscs and as- 

 cidians the experimental studies of recent years have revealed a 

 very definite order in the distribution of cytoplasmic factors of 

 differentiation which presumably are traceable to corresponding 

 ' formative stuffs." For accomplishing such complicated cyto- 

 plasmic divisions some special mechanism seems necessary, but 

 as to its nature very little has been made out. In the division of 

 tissue cells, however, appeal can not be made, as in the case of 

 the egg, to direct experiment, and we must here rely on morpho- 

 logical studies alone. An approach to the problem may there- 

 fore be made by a comparison of the quantitative relations 

 existing between the products. Accordingly the problem of 

 cytoplasmic division may be stated for present purposes in the 

 form of a double question: (i) Is the division of the cytoplasm 



