THE H.EMOFLAGELLATEg. 



239 



of the numerous divisions (in fig. 27, e represents a small 

 family, and f a large one where the parent-individual is 

 very distinct). It is important to note that the kineto- 

 nucleus changes its position during the commencing stages 

 of division and comes to lie alongside the trophonucleus, or 

 even passes to the other side of it, i. e, nearer the flagellar 

 end. 



Fig. 28. — a — d, rosette segmentation in T. lewisi ; e, daughter- 

 individual ; F, one dividing. X 1750. (After L. and M.) 



This variety of division forms a direct transition between 

 binary fission and the other characteristic method of T. lewisi, 

 viz. segmentation ; indeed, such a family as that just described 

 often greatly resembles a rosette, but is generally distin- 

 guishable therefrom by the presence of a parent-individual. 

 In rosette-formation, on the other hand, segmentation is 

 multiple and equal. ^ The body assumes an ovoid to spherical 



' McNeal (1. c.) doubts the occurrence of true equal segmentation. He 

 thinks that the " budding " process is rather concerned in all cases, i. e. that 

 a parent-individual can always be recognised. The fact, however, that equal 

 multiple longitudinal fission occurs, in which there is no sign of a parent- 

 individual (cf. fig. 25 g), at least makes it possible that the process of equal 

 segmentation occurs. The condition seen in fig. 28 B is easily derivable 

 from that in the former figure. See also next page. 



