254 c. ISHIKAWA : 



upon itself in such a way as to push and raise the dividing 

 centrospheres, together with the nuclei, above the general sur- 

 face. I think it is owing to this power of the spindle fibres to 

 stretch out after the division of the nuclei that the flagellum is 

 formed (Fig. 10 /'), since this is actually much longer than the 

 spindle fibres. 



The effect of the elongation of the central spindle-fibres is 

 also to bo seen in dividing individuals. Such are shown in Fig. 

 4 Pt. XT, as well as in Figs. 35 and 36 Pt. XIV ('94, c), which 

 all represent the median portion of the spindle as bulging out 

 on both sides owino; undoubtedlv to the counteraction exerted 

 on the spindle-fibres by the pressure of the cytoplasm upon the 

 poles of the spindle, an explanation already given by R. Hert- 

 wiG ('95), in the case of a similar phenomenon observed by him 

 in the division of the micronucleus of Paramœcium. To the same 

 cause also is to be attributed the formation of a small diagonal 

 figure given in Fig. 39 of my previous paper and there des- 

 cribed as the rest of the swollen portion of the spindle after 

 complete separation of the nuclei ('94, c). 



Now, the point that seeins to be of great interest is, that 

 in all the cases of Noctiluca observed by me, the outbulging 

 of the central portion of the spindle takes place only 

 in the spindle of dividing individuals, while in that 

 of spore-forming individuals, a bending of the same 

 portion is observed. 



This difference in the behaviour of the central spindle- 

 fibres is perhaps to be explained by the f\ict, that in spore- 

 formation the dividing nuclei are raised above the surface of Ûiq 

 body, while in ordinary division they lie within the body. The 

 point in question becomes all the more interesting when we 



