BIGELOW: NUCLEAR CYCLE OF GONIONEMUS MUKBACIIII. 377 



The origin of the acrosome has never been traced in any coelenterate, 

 although this structure has heen observed in various members of the 

 group by Pictet ('91) ; Ballowitz ('94) ; Aders (:03) ; Retzius (:04, :05) ; 

 Gorich (:03 a ,:04); and others. Its staining reactions, however, show 

 clearly that it also is of archoplasmic nature, as is now so generally con- 

 sidered the case among the higher Metazoa. (See Meves, '97, '98, '99, : 00 ; 

 Wilcox, '96; McGregor, '99; Paulmier, '99 ; Tonniges, :02 ; Blackman, 

 ■05.) 



c. Tlie axial filament. — In Gonionemus the evidence is very strong 

 that the axial hiament of the tail is an outgrowth of the substance of 

 the centrosome : a result indicated, though less conclusively, by the ob- 

 servations of Gorich (:03 a , :04) on Sycandra and Anrelia. Of especial 

 interest in this connection is the occurrence in Gonionemus of giant and 

 multiple spermatids of a type closely resembling those described by 

 Paulmier ('99) in Anasa. In the latter case, according to Paulmier, a 

 separate axial filament arises in connection with each centrosome, al- 

 though there is but a single mass of archoplasm. This has seemed to E. B. 

 "Wilson ( : 00) conclusive evidence that the filaments are actual outgrowths 

 of the centrosome. And this is entirely corroborated by Broman (:02), 

 who has found that in abnormal spermatids of mammals and elasmo- 

 branchs one filament is developed in connection with every centrosome, 

 no matter how many nuclei or spheres there may be in the cell. Still 

 further support is lent to this view by the discovery of abortive filaments 

 in connection with the centrosomes in the spermatocytes of elasmo- 

 branchs (Moore, '95), Lepidoptera (Meves, '97, :00; Henneguy, '98) ; and 

 Gonionemus (pp. 322) ; while Meves (:00) has shown that in Pygaera 

 these filaments — of which there are two in the secondary spermatocyte, 

 one attached to each arm of the V-shaped centrosome — may actually 

 persist to form the tail filaments of the spermatozoa. 



It is probable, then, that a genetic relationship between centrosomes 

 and axial filaments is very general, as has already been pointed out by 

 Meves (:02), and Korschelt und Heider (:02). But such a relationship 

 is not universal, for Tonniges (:02), P. et M. Bouin (:03) and Blackman 

 (:05) have shown that the filament in myriapods is formed from the 

 archoplasm, independent of the centrosome ; a conclusion which is 

 certainly true for the "pseudoaxial filaments" described by Blackman 

 (: 05), for these cannot possibly have any connection with the centrosome. 



