376 bulletin: museum of comparative zoology. 



plete. And since Downing's figures suggest that be saw only a few of 

 the stages in the metamorphosis, it is not improbable that a division of 

 the centrosome, and migration of one of the resultant halves takes place 

 in Hydra, as it certainly does in Gonionemus. Gorich (:03 a ) states that 

 the history of the centrosome is the same in various other hydroids and 

 medusae ; and since I have myself been able to confirm his account in 

 Aurelia and Metridium, as well as in Gonionemus, it seems safe to con- 

 clude that the process observed in the latter animal is the one general 

 among coelenterates. 



b. The archoplasmic structures. — Our first accurate knowledge of these 

 structures in this group is due to Pictet ('9l), who discovered a homo- 

 geneous mass lying beside the nucleus in the spermatid of the siphono- 

 phore Halistemma. This mass he believed to be formed by the coalescence 

 of granules which "sont . . . des produits d'elimination de la derniere 

 division caryocinetique des spermatocytes." In Gleba, likewise, there is 

 the same " Nebenkern," and also in the spermatozoon a spherical middle 

 piece shown by its staining reactions to be of similar nature. Ballowitz 

 ('94) has described much the same condition in the actinian Tealia, find- 

 ing a "Nebenkern" and also one or two smaller masses lying near the 

 point of origin of the tail. These observations were extended by Eet- 

 zius (:04, :05), who found that the middle piece in several other coelen- 

 terates — e. g., Cyanea, Tubularia, Clava, Sertularia — consists of two or 

 more spherical archoplasmic masses. With this account the condition in 

 Gonionemus closely agrees. Gorich (:03 a , : 04) seems to have entirely 

 overlooked this substance in the spermatid of Aurelia and Sycandra. 

 But since I have myself demonstrated, by use of the iodine method, that 

 the archoplasm plays the same role in Aurelia that it does in Gonione- 

 mus, I attribute his failure to detect it to his dependence on haematoxy- 

 lin staining and on such plasma dyes as Bordeaux red. Taking these 

 various observations together, we can safely conclude that the middle 

 piece among coelenterates in general is an archoplasmic structure. 



Most authors who have traced the origin of the middle piece have 

 maintained that it is descended from the interzonal portion of the spindle. 

 This view is maintained especially by Wilcox ('95,'96); Erlanger ('97). 

 Paulmier ('99) ; and Baumgartner (:02). Gonionemus, however, shows a 

 divergence from this type, in that the archoplasmic middle piece is de- 

 rived, at least in part, from the remnants of the polar portion of the 

 spindle, thus closely paralleling the conditions observed by Calkins 

 ('95) in Lumbricus. 



