394 
Notes . 
always remain neutral. Those, however, which are so related to the 
fibres, exert a most obvious influence on the direction of these, and 
a whole sheaf or strand of spindle-fibres may sometimes be seen 
to divert from the main direction, and to end blindly on such 
granules. 
I have been unable to demonstrate the existence of special centro- 
somes occupying definite positions at each end of the cell, and I am 
not clear as to how they might fit in with the granules. Perhaps it is 
conceivable that in some cases the individual granules might fuse and 
form one large terminal body which one might then regard as a ‘ cen- 
trosome,' and it is certain that in many cells the main portion of the 
Woodcut 2 — Fig. A. Pollen-mother cell of Liliurn Martagon in early stage of 
division : n the nucleolus. B. Later stage, showing multipolar spindle. 
achromatic spindle does stand in relation to bodies near the poles of 
the cell, but even in these cases there are almost always to be seen, in 
preparations which allow of a definite conclusion being arrived at, 
divergent strands separating from the main spindle and terminating in 
granules otherwise situated. Eventually, during the separation of the 
daughter-chromosomes, the spindle seems to become more regular, 
but the difficulties of tracing its real relationships simultaneously 
increase, and I do not as yet feel able to speak with entire confidence 
as to its further fate. 
Shortly after I had, somewhat unwillingly, arrived at the conclusion 
