I/O THE GERM-CELLS 



into two parts, of which the anterior gives rise to the end-knob, while 

 the posterior is transformed into a rin;^, which wanders back to its 

 final position at the posterior end of the so-called "connecting-piece." 

 From this it follows that the latter body ( VerbindimgsstUck) does not 

 correspond to the middle-piece of the salamander (here represented 

 by the small disc-shaped body at the base of the nucleus), but belongs 

 to the flagellum ]:)roper. The origin of the axial filament and end- 

 knob is, however, nearly the same in the two cases. In the guinea- 

 pig the process is somewhat more complicated and is not quite cleared 

 up bv Meves ; but the origin and fate of the ring is the same, and the 

 end-knob passes into the neck of the spermatozoon as in the rat. 

 Taken together, these observations conclusively show that in mam- 

 mals and Amphibia the end-knob is a derivative of the centrosome, 

 thus sustaining, though with some modifications, Hermann's earlier 

 conjecture ('92) as to the nature of this body; and they overturn 

 Niessing's result ('96) that the centrosome passes into the acrosome. 

 As in the salamander, the acrosome is formed from an idiozome 

 derived in the guinea-pig from the remains of the attraction-sphere 

 (Meves), while in the rat, according to Lenhossek, it is independently 

 formed in the cytoplasm without relation to the preceding mitotic 

 figure or the centrosomes. Within the sphere appears a small, deeply 

 staining body, resembling a centrosome, yet staining differently from 

 the true centrosome, which enlarges to form the acrosome, while 

 about it is formed a clear substance forming the " head-cap " (p. 139). 

 In the rat the acrosome remains small (" Spitzcnknopfchen " of Mer- 

 kel;; in the guinea-pig it becomes nearly as large as the nucleus 

 itself (Fig. 86). An interesting feature in the formation of the 

 mammalian spermatozoon is the casting off of a portion of the 

 spermatid-cytoplasm in the form of a ''cytoplasmic vesicle" or ** tail- 

 vesicle," which degenerates without further use (Fig. ?)G). This pro- 

 cess, described by Meves ('99) in the guinea-j^ig, is closely similar to 

 that which occurs in the spermatozoid-formation in ferns (p. 144). 



Resume. In reviewing the foregoing facts we find, despite many 

 variations in detail, three points of fundamental agreement, namely : 

 ( I )the origin of the sperm-nucleus from that of the spermatid ; (2) the 

 origin of a part at least of the "middle-piece" from the spermatid- 

 centrosomes; and (3) the outgrowth of the axial filament from one of 

 the spermatid-centrosomes. It is clear, however, that the term middle- 

 piece has been applied to structures of quite different morphological 

 nature, which agree only in lying behind the nucleus. Thus in the 

 salamander the inner centrosome gives rise to the main body of the 

 middle-piece ; in the rat or in man it gives rise only to the small disc- 

 shaped body lying in the "neck" in front of the so-called middle- 



