CROSSING-OVER 555 



know intercalary, may include the centromere or not. When it 

 includes the centromere it is the same as internal interchange, that 

 is, interchange within the chromosome. Translocation consists in 

 the removal of an intercalary segment from one position to another. 

 It may be intra-radial or extra-radial ; it may be internal to the 

 chromosome or fraternal, that is, between two homologues, or 

 external, that is between two unrelated chromosomes ; it may be 

 eucentric or dyscentric. Further, translocation may be symmetrical 

 with respect to the centromere or not, in which case the piece removed 

 contains the centromere and so gives rise to a dicentric and an 

 acentric product. Interchange is internal, fraternal or external, 

 but it may be balanced or unbalanced and more or less unequal with 

 regard to the sizes of segments exchanged. 



These differences of spatial relationship have important genetical 

 results, as we saw in considering structural hybridity and secondary 

 structural change. By crossing-over all intercalary changes give 

 reduplications and deficiencies, and all dyscentric changes give 

 dicentric and acentric products, and these, whether arising in this 

 way or directly, cause breakage and loss. 



Thus both the spatial and genetic properties of a structural change 

 condition its survival. Acentric, dicentric and ring chromosomes 

 suffer mitotic elimination ; dyscentric changes, whether inversion or 

 translocation, suffer meiotic elimination ; duplications and deficien- 

 cies suffer cellular elimination, that is, death, either immediate or 

 delayed, of the whole cells containing the changed chromosomes. 

 These last two do not depend on the purely spatial properties of 

 the change but on its genetic consequences. In considering any 

 analysis of structural changes we must therefore bear in mind the 

 degree, kind and time of elimination that the changes may have 

 suffered before they were identified. 



The spatial relationships of structural change are interesting, 

 however, from another point of view. They can be considered in 

 relation to the number of breakages and number of reunions of 

 breakage-ends required to produce the observed result (Haldane, 

 unpub.). Stadler (1932) has pointed out that all viable X-ray 

 changes are apparently due to the reunion of breakage-ends. This 

 seems to be true except in the formation of branched chromosomes. 



