318 KOFOID AND SWEZY. 



complement of anterior flagella is in part completed rather late (PI. 3, 

 Fig. 33) in mitosis. 



The division of the blepharoplast is difficult to detect and seems to 

 present quite a variety of appearances and to bear no constant rela- 

 tion to the progress of division in other organelles. We are quite 

 unable to find a granule at the base of each flagellum in the blepharo- 

 plast. Subdivisions of this structure seem to be concerned with 

 mitosis rather than to indicate relations to individual flagella. Owing 

 to the range in form which is presented it may be that the phenomena 

 have none of the significances which we here ascribe to the stages we 

 find. We conclude, however, that they are all referable to two steps 

 in mitosis. The first is the division of the blepharoplast into two 

 equal daughters which sooner or later migrate to the poles of the 

 elongating nucleus and place themselves at the poles of the forming 

 spindle. Each carries with it a daughter undulating membrane 

 complex (PI. 2, Figs. 20, 22) and its complement of the old and new 

 anterior flagella as above described. This division and migration is 

 always to be found as the prophase passes to the metaphase of mitosis. 



The second step is the division of each blepharoplast fPl. 2, Fig. 23) 

 into a centrosome at the apex of the spindle to which apparently no 

 flagella or extra-nuclear structures are attached and into the basal 

 granule which retains connections with the undulating membrane 

 complex, paradesmose, and the anterior flagella. We are unable to 

 trace the connections of the rhizoplast and axostyle during this period. 

 This second step, however, does not seem either to be of long duration, 

 for few instances of it have been found, nor to have a definite and fixed 

 location in the mitotic sequence. It may occur for example (PI. 1, 

 Fig. 8) in the early prophase prior to polar migration, and the blepharo- 

 plast appear to be made up of four granules or two groups of two each, 

 or it may occur in the anaphase (PI. 2, Fig. 23), and again it may 

 occur at one pole and not at the other (PI. 3, Fig. 24). These facts, 

 together with its infrequency at stages such as the anaphase when it 

 is most to be expected, lead one to doubt the universality of its occur- 

 rence. It seems, however, to be a significant process, and not an 

 irrelevant disintegration of the blepharoplast, for the parts into which 

 it subdivides have definite relations to structures of the organism, 

 and relations, moreover, which are widely recognizable among Proto- 

 zoa and other flagellated cells, for one granule persists in its location 

 at the apex of the spindle, or in the comparable posib'^n on the 

 daughter nuclei, and the other remains in the position of basal granule 

 to the whole group of flagella. 



