320 ON THE INTERNAL FORM [ch. 



it might be torn asunder into two portions, the split coinciding 

 with the locus F^. 



Suppose a number of such bodies to be scattered throughout the 

 medium. Let at first the regions ¥„, and F^ be entirely outside the 

 space where the bodies are situated: and, in making this supposition 

 we may, if we please, suppose that the loci which we are calhng 

 Fa and F^, are meanwhile situated somewhat farther from the axis 

 than in our figure, that (for instance) F^ is situated where we have 

 drawn F^, and that F^ is still farther out. The bodies then tend 

 towards the poles; but the tendency may be very small if, in 

 Fig. 98, the curve and its intersecting straight hne do not diverge 

 very far from one another beyond Fa\ in other words, if, when 



Fig. 99. 



situated in this region, the permeability of the bodies is not very 

 much in excess of that of the medium. 



Let the poles now tend to separate farther and farther from one 

 another, the strength of each pole remaining unaltered; in other 

 words, let the centrosome-foci recede from one another,- as they 

 actually do, drawing out the spindle-threads between them. The 

 loci Fa, F^ will close in to nearer relative distances from the poles. 

 In doing so, when the locus F^ crosses one of the bodies, the body 

 may be torn asunder; if the body be of elongated shape, and be 

 crossed at more points than one, the forces at work will tend to 

 exaggerate its foldings, and the tendency to rupture is greatest 

 when Fa is in some median position (Fig. 99). 



When the locus Fa has passed entirely over the body, the body 

 tends to move towards regions of weaker force; but when, in turn, 

 the locus Fi, has crossed it, then the body again moves towards 

 regions of stronger force, that is to say, towards the nearest pole. 



