STUDIES ON. INSECT SPERM ATOGENESIS. 63 



series. The patterns arc merely optical cross-sections of a platc- 

 li'ork. It is difficult to be sure of this plate-work when the ne- 

 benkern is viewed " in the round," but I have tried to represent a 

 small, tangential slice of it in Fig. 4. The plate is more or less 

 wrinkled and apparently perforated in places, which irregulari- 

 ties seem to correspond with the points of indentation, folding, or 

 union, of the various plates, as can be explained by a comparative 

 study of Figs. 4, 5 and 6. In the Hemiptera, therefore, a spireme 

 seems not to occur at this stage, and the patterns are due not to 

 a tangled thread-work but to a more or less regular plate-work. 



It is necessary now to retrace our steps to the early stages in the 

 formation of the nebenkern, and attempt an explanation of these 

 results from a broadly comparative standpoint. It will be re- 

 called that the peculiar pattern in the lepidopteran nebenkern 

 owes its origin to the arrangement of the chromophilic and the 

 chromophobic constituents of the original granular mitochondria. 

 The question then arises: are the patterns just described in the 

 Hemiptera comparable in any way to those in the Lepidoptera? 

 I believe that all the facts point to an affirmative answer to this 

 question. 



It will be recalled that from a comparison of various morpho- 

 logical types of mitochondria the assumption seemed to be war- 

 ranted that they are all alike composed of two fundamental and 

 distinct substances a chromophilic material forming the outer 

 covering of each chondriosome unit, and an internal chromophobic 

 material. Assuming this structure to be present in the thread-like 

 type of mitochondria, the transformations of the hemipteran 

 spermatid are readily explicable. During the early stages in the 

 formation of the nebenkern. the presence of the two mitochon- 

 drial materials can not be discerned any more than was possible 

 in the auxocyte period. It is, therefore, impossible to say exactly 

 how the fusion of the threads takes place, as Gatenby has done in 

 the case of the vesicular mitochondria. It seems probable, how- 

 ever, that, as in the latter case there is a process of running to- 

 gether going on within the chondriosome mass, which -finally 

 results in the production of sufficiently large groupings of the two 

 materials to allow of a visible separation. This condition is at- 



