524 CELL MECHANICS 



has long been clear from the fact that it is only at this point that 

 the chromosome lies on the spindle or is " attached " to it, as the 

 phrase used to be. This view has now been verified by the study 

 of acentric chromatids produced by crossing-over in inversions and 

 by X-ray breakage. Such chromatids are entirely passive. Con- 

 gression and orientation, together with the anaphase separation of 

 chromatids and the movement of terminalisation, may therefore be 

 defined as centric reactions of the chromosomes. 



The third process is distribution. The centromeres come to lie so 

 that the bodies of their chromosomes are more or less evenly distri- 

 buted on the plate. It consists therefore in a modification of the 

 centric reaction by the body repulsion of the chromosomes. 



These three processes take place as a rule at the same time. 

 The reason for distinguishing between them is that each may fail or 

 vary independently of the other two and for special reasons. There 

 is no evidence of congression and orientation varying at mitosis 

 except for small chromosomes, which may conduct their movements 

 out of step with the large ones of the same complement — either 

 sooner or later (Upcott, unpublished). 



Distribution, on the other hand, is subject to three natural 

 sources of variation at mitosis. The first is the existence of the 

 secondary attractions which modify the simple equilibrium position 

 produced by equal repulsion. The second is the generalised body- 

 repulsion considered at prophase. This leads to a characteristic 

 difference in the distribution of large and small chromosomes. 

 Necessarily, since only the centromeres regularly lie on the spindle, 

 the widest distribution of the chromosomes is attained when the 

 longer chromosomes lie on the periphery. This always happens 

 when the bodies of the chromosomes are held in the equatorial 

 plane by the orientation of their centromeres and by their own 

 rigidity. This condition applies to meiotic bivalents with several 

 chiasmata. It does not apply however to chromosomes that bend 

 freely and may turn perpendicular to the plate, as long thin chromo- 

 somes always do at mitosis. Hence the contrast in the arrangement 

 of the chromosomes of Hosta at mitosis and meiosis (Akemine, 

 1935 ; Fig. 149, cf. Fyfe, 1936). 



The third source of variation is that which determines the hollow 



