No. 3.] THE PROTOZOA AND METAZOA. ^2)7 



D. The Mechanism of Mitosis in Noctiluca. 



The peculiar type of mitosis in N'octiliica may perhaps 

 throw some light on the mechanism of mitosis in general. 

 Here the entire absence of astral rays passing from the sphere 

 into the surrounding cytoplasm excludes all contractility 

 hypotheses in so far as they are based upon active physio- 

 logical contractility of these fibers (Van Beneden, Boveri, 

 Flemming, Reinke, etc.), and the same remark applies to 

 Heidenhain's view that mitosis is affected through elastic 

 tension of the rays. Furthermore, there is no morphological 

 evidence to show that division is affected by contraction of 

 the mantle-fibers (Hermann), for, as shown above, these do 

 not shorten and thicken, but remain the same throughout 

 mitosis. The possibility that their substance is taken up by 

 absorption into the sphere, as Wilson has suggested for Toxo- 

 pticustcs ('96), is removed by reason of two facts, viz., the 

 daughter-spheres do not enlarge in the anaphase (spore mitosis) 

 and the mantle-fibers can always be traced to distinct points 

 in the sphere, i.e., to the centrosomes. It must be, therefore, 

 that the separation of the daughter-chromosomes is caused 

 either by an active divergence of the spheres, or by growth of 

 the central-spindle fibers and the consequent passive separa- 

 tion of the spheres (Driiner). The former is improbable 

 because of the entire absence of the antipodal cones or astral 

 rays, leaving the second as the only mechanical hypothesis 

 which agrees with the facts. 



My conception of the process is, then, as follows : the central- 

 spindle lying within the ring of chromosomes is advantageously 

 placed for exerting the necessary dividing force. The nuclear 

 membrane disappears and mantle-fibers connect the ends of 

 the chromosomes with centrosomes in the spheres. The cen- 

 tral-spindle elongates, causing separation of the spheres ; the 

 mantle-fibers, remaining firm, move with the spheres, dragging 



('94), Lavdowsky ('94), Carnoy and Lebrun ('97), and others who regard the 

 nucleolus as the seat of the centrosome. On the contrary, it is a body which 

 much more resembles the so-called " nucleolus-centrosome" described by Bal- 

 biani in Spirochona, and should be, I think, placed with the latter structure as 

 one of the primitive forms of the sphere. See appendix, p. 49. 



