558 CELL MECHANICS 



in inversion hybrids. But unlike them (according to Rhoades and 

 McQintock, 1935) these double broken bridges form new rings again. 

 There is no evidence yet of this happening to broken bridges after 

 meiosis. In these it must be assumed undoubtedly where the bridges 

 are single that the new ends become normal ends. Their survival as 

 ends explains the occurrence of lateral chiasmata. There is, therefore, 

 some difficulty in supposing that delayed re-fusion occurs. And 

 hence there is also a difficulty in assuming a specificity in ends 

 without the additional obnoxious assumption that breakage-ends 

 change and become true ends later. 



Catcheside, on the other hand, considers that breakage-ends must 

 reunite immediately. His two assumptions of reciprocity and imme- 

 diate fusion make structural change strictly analogous to crossing- 

 over, as it is described above, and thus simplify the whole problem. 



In order to understand the apparent non-fusability and conse- 

 quent permanence of ends it is necessary to consider the mechanics 

 of breakage and reunion. The process like crossing-over demands 

 movement : the breakage ends must move towards one another 

 before they can reunite. This movement may be due to attraction 

 between the ends or to change within the broken chromosomes as in 

 the case of crossing-over. The first explanation agrees with the 

 view that the breakage-ends have a power of attraction that the 

 true ends lack. The second explanation agrees with, and indeed 

 itself explains, non-fusability of the ends, in this way : the relic 

 spirals show a state of stress of all parts of the chromosome except 

 the ends, for these are free to uncoil. Two breakage-ends are at once 

 released from this stress and will fly apart from one another. They 

 are expected to move when broken, although true ends would 

 not. 



Whatever the mechanical conditions, however, their conse- 

 quences are now clear. We have to consider the type of change 

 produced in relation to its survival at mitosis, and in growth and 

 reproduction. 



Survival at mitosis depends on whether they are efficient chromo- 

 somes, i.e. on whether they have one centromere and two ends, 

 or are inefficient, i.e. having more or fewer centromeres or ends. The 

 permanence of centromeres and ends, however, has two important 



