478 JEAN REDMAN OLIVER 



is no trace of a tail filament. It soon appears, however (fig. 

 10, a.f.) and may be readily found in cases in which the centro- 

 somes have migrated inward to their contact with the nuclear 

 membrane, as in figures 11 and 12. In nearly every instance 

 the tail filament projects out freely beyond the boundary of the 

 cell, and even in such an early stage as that shown in figure 10, 

 it has already reached a considerable length. The appearance 

 and growth of the tail filament must be extremely rapid, for it 

 is very difficult to find any early stages. The migration of the 

 centrosomes toward the nucleus must be quite rapid also, as 

 intermediate stages between those shown in figures 10 and 11 

 are very rare. No indication of any prolongation extending from 

 the nucleus toward the centrosomes, such as described by Meves 

 ('99), and by Duesberg ('08), could be found, and the migration 

 seems to be entirely an active one on the part of the centrosomes, 

 so far as such a visible participation of the nucleus is concerned. 

 After reaching the nucleus the anterior centrosome becomes 

 closely pressed up against the membrane (figs. 11 and 12) and 

 at the end of the first period it is fused with the membrane, 

 often almost disappearing from view in the chromatin incrusted 

 wall. In this fusion the anterior centrosome is lengthened in 

 a direction at right angles to the tail filament. In many instances 

 it appears as if the centrosome had penetrated the membrane 

 and was situated on its inner surface as in figures 13 and 14. 



The second period: from the appearance of the caudal tube to the 

 migration of the distal half of the posterior centrosome {the annu- 

 lus) along the tail filament: figures 16 to 29 , 



In the original division of Meves ('99) the second period extends 

 from the origin of the 'Manschette' up to its total disappearance 

 from the cell. As will be shown in the following pages the 

 *Manschette' does not disappear in the developing spermatid of 

 the fur seal, but persists and takes an important place in the 

 structure of the adult sperm. For this reason we are compelled 

 to seek another phenomenon which, if possible, occurs constantly 

 at the time of the disappearance of the caudal tube in other 



