16 THOS. fl. MONTGOMERY jr., 



occurs. I have not seen the first stage of the process (which would 

 probably be marked by the presence of a centrodesmosis), but have 

 seen the stages immediately following (Figs. 23, 24). This division 

 takes place within the idiozome. Between two daughter centrosomes 

 is found a clear space of a more or less oval outline, in which no 

 cytoplasmic granules are seen. This may be regarded as the first 

 stage in the formation of the central spindle, even though at first 

 separate spindle fibres cannot be distinguished in it. Occasionally the 

 clear space between the daughter centrosomes has a clearly marked, 

 sharp boundary on either side. The two centrosomes, with their con- 

 necting central spindle, now pass out of the idiozome and wander 

 towards the nucleus, the central spindle increasing only slightly in 

 length (see the cells on the left hand of Fig. 24). Finally the small 

 central spindle lies close to the surface of the nucleus, more or less 

 paratangential to the latter. In one small spindle at this stage was 

 noticed a thread connecting the centrosomes, possibly the first appear- 

 ance of a spindle fibre (Fig. 24 c.spl), since at this late stage there 

 could not be a persistance of a primary centrodesmosis. The stages 

 of the central spindle just described are not frequently seen, and the 

 observation of the central spindle must be made in the best morning 

 light. 



The next following stage observed showed the mass of chromo- 

 somes near the centre of the cell, the nuclear membrane absent, and 

 a well formed central spindle to one side of the nuclear cavity 

 (Figs. 25, 26). At this period the chromosomes have not yet reached 

 the equator of the spindle, so that the fibres of the latter may be 

 clearly traced from centrosome to centrosome. Mantle fibres are now 

 attached to the chromosomes, and delicate pole fibres make their 

 appearance : in these mitoses the latter fibres are more delicate than 

 even the central spindle elements, while the mantle fibres are clearly 

 marked. Since the central spindle makes its first appearance in the 

 cytoplasm before the nuclear membrane has disappeared, it may 

 be concluded that it has a cytoplasmic origin; but the mode of 

 formation of the pole and mantle fibres could not be determined. 



Gradually the stage of the monaster (equatorial plate) is reached 

 (Figs. 25 — 29), all appearances pointing to the mantle fibres as the 

 mechanical agents in the movement of the chromosomes. The chromo- 

 somes become grouped closely together in the equator, though oc- 

 casionally a central opening may be detected in the chromosomal 

 plate (Figs. 33 — 36), through which then the central spindle probably 



