DYSCENTRIC HYBRIDITY 265 



results, and it is probably rare. This is no doubt due to the differen- 

 tial segments undergoing non-homologous pairing in continuation 

 of the pairing of the ends, which alone can therefore cross over and 

 form chiasmata. The genetical consequences of such behaviour 

 will be considered in relation to CEnothera (Ch. IX). 



(iii) Inversion Hybrids (a) Structural Types. We have so far 

 considered organisms in which the changed and imchanged segments 

 have the same linear order in relation to the centromere. These 

 may be described as eitcentric structural hybrids, and their behaviour 

 at meiosis is remarkable merely in showing multiple associations, 

 like those of polyploids, and unequal bivalents, sometimes with 

 lateral chiasmata. Structural change may, however, reverse the 

 linear order in relation to the centromere, and we then have a 

 dyscentric structural hybrid. The simplest way in which this change 

 can arise is by simple inversion of a segment. Such changes have 

 long been recognised genetically in Drosophila, and have recently 

 been identified in the salivary glands and at pachytene in maize. 

 They may also arise with translocation of a segment from one 

 chromosome to another or from one arm of a chromosome to another. 

 Thus, if a comma represents the centromere, a chromosome 

 ahcd,efgh becomes abcd,egfh by simple inversion and abfgcd,eh by 

 inversion of fg combined with internal translocation ; this last by 

 crossing-over with a normal chromosome will give ahfgcd,efgh, in 

 which the inverted segment is reduplicated. It will be noticed in 

 this last chromosome that fg is not inverted in relation to the ends, 

 but only in relation to the centromere, and inversion so defined is 

 the essential property of a dyscentric hybrid. Inversion, including 

 the centromere, it should be noted, is not inversion with respect to 

 the centromere ; thus ahcd,efgh -f- ahce^dfgh gives a eucentric 

 hybrid (D., 1936 d). 



(b) The Results of Crossing-Over. Inversion, whether including 

 the centromere or not, results in the reverse pairing of a loop at 

 pachytene (in Zea, Tulipa and Chorthippus) , and is recognised in 

 Drosophila by an exactly similar behaviour in the salivary glands 

 (Ch. V). Sometimes, however, the inversions may pair straight at 

 pachytene, non-homologous parts associating in a proportion of 

 cells by torsion and not by attraction (McClintock, 1933, on Zea^ 



