620 ROBERT H. BOWEN 



in Call OS a mi a than in Pygaera, correlated apparently 

 with the very much smaller amount of available acrosomal 

 material. The thread-like nature of the acrosome and its 

 attachment so near the insertion of the axial filament offer 

 the possibility of a natural error in interpretation which 

 should be guarded against in the study of other sperms. If one 

 studied only the later stages of sperm formation, or observed 

 the earlier ones inaccurately, a most olnious conclusion would 

 be that the acrosome was really a forward growth from the 

 spermatid centrioles, comparable in its method of formation 

 to the tail filament. It is possible that some of the accounts 

 which have been given of the role of centrioles in acrosome 

 formation may actually be traceable to errors of this nature. 

 In any event it is clear that great care should be exercised in 

 the future in interpreting thread-like formations in the sperm 

 head, since it is now evident that topographical relationshij) 

 to the centrioles ma}' not be in the least degree indicative of 

 organic relationship. It is interesting to note in this place that 

 Goldsmitli (1919) has described a thread-like formation in the 

 sperm head of Cicindela, without, however, offering any 

 explanation of its homologies. I think it proba])le that the 

 facts made out in Callosamia ^^^ll furnish a clue to the 

 enigmatic structures described by Goldsmith. 



The Apyrexe Sperms. 



In the saturnid moths large numbers of ap\Tene sperms 

 are formed after the normal sperm formation has been largely 

 completed. The nucleus in every case is reorganized in the 

 spermatid in a normal manner, as Gatenb}' (19176) found in 

 Pier is brassicae. However, at an early stage the sper- 

 matid becomes visibly abnormal by reason of the improper 

 orientation of the nucleus and nebenkern, their relative posi- 

 tions being exactly reversed in the abnormal spermatids. 

 The result is quite striking when an entire cyst is observed, 

 the nebenkerns being adjacent to the cyst wall instead of the 

 nuclei, as is so characteristic in insect testes. 



As the nebenkern elongates the nucleus moves back along 



