GOO ROBKRT ir. nowKN 



and are finally separat(id into two equal masses by the constric- 

 tion of the cell wall. The vesicles then draw away from the 

 region of the mid-body and, gradually rounding up, regain their 

 former shape. This is shown particularly in the second matura- 

 tion division, at the close of which the nebenkern is constructed 

 (fig. 43). 



The ' aero blasts ' of Gatenby are obviously the representa- 

 tives of the (rolgi apparatus, as he has also indicated in a later 

 publication, I wish only to call attention to the intimate struc- 

 ture of the individual (lolgi l)odies, since on this point I am not 

 entirely in agreement with (jatenby's account. According to 

 my observations each Golgi body is made up on a plan essentially 

 similar to that which I have described in Hemiptera (Bowen, 

 lO'iO), except that they are somewhat smaller, and hence less 

 easy to analyse. Each Golgi rodlet is accompanied by a small, 

 plate-like mass of material, which stains relatively less than the 

 Golgi substance itself, and which, as I have shown elsewhere, is 

 to be looked upon as a portion of the fragmented idiosome. The 

 vesicular portion which Gatenby sometimes finds is obviously 

 the equivalent of this idiosomic i)ortion of each Golgi body ; 

 but I ha^ e never happened to see in my preparations any case 

 in which it presented such a vesicular appearance. 



In the maturation divisions the collection of the Golgi 

 material around the spindle poles has been correctly described 

 by Gatenby, but I cannot at present corroborate his statement 

 that the ' acroblasts ' are sorted out entire. The possibility 

 of their undergoing more or less fragmentation, such as seems 

 to occur so extensively in the Hemiptt^ra (Bowmen, 1920), ought, 

 I think, to receive much more thorough study before a final 

 decision is reached. In any event we seem to be agreed that 

 the Golgi bodies are ultimately present in the spermatids in 

 substantially the same form as in the spermatocj'tes. 



Concerning the centrioles I have nothing new to contribute, 

 except that Cook's (1910) statement concerning the loss of the 

 tail filament in the first spermatocyte division of Callosamia 

 is apparently incorrect, and was based in all probability on 

 faulty technique. 



