SECONDARY STRUCTURAL CHANGE 277 



somes of new shapes would appear in the second. Such new 

 chromosomes have been found in the progenies of two triploid 

 Crepis hybrids ; C. capillaris X C. aspera (Nawaschin, 1927) 

 and C capillaris X C. tedorum, CCT (Holhngshead, 1930 b). In 

 the first, a chromosome of capillaris, in the second, one of tectorum, 

 had lost a portion of one arm. In both it is to be supposed that 

 the conditions of zygotene pairing in the triploid, where one thread 

 at any point is unpaired, had rendered possible crossing-over 

 between relatively translocated segments. Evidence that such 

 crossing-over occurs in the triploid with segregation of cross-overs 

 and non-cross-overs is found in the fact that the derived tetraploid 

 (c/. Table 26) is a weak plant unlike any tetraploid produced by 

 simple doubling. 



The attachment of the X and Y' chromosomes in Drosophila 

 resembles a secondary structural change, occurs always in the same 

 way and, like the half-mutants, with a definite frequency, once in 

 1,500 to 2,000 times (Stern, 1929 a). Since the association occurs 

 at a point where the two chromosomes are probably homologous, 

 it is possible that it arises through crossing-over {cf. D., 1931 a, 

 and Fig. 121). This conclusion has now been confirmed by Kauf- 

 mann (1934, cf. D., 1935/), who has found the complementary cross- 

 over types. 



Analogous with this is the origin of new chromosome types 

 following crossing-over in structural hybrid Drosophila whose 

 hybridity is the result of X-ray changes, as in Stern's experiment 

 described above {cf. MuUer, 1930 a). Similarly, crossing-over 

 within an inversion including the centromere will give the kind of 

 chromosome that is found in secondary trisomies in Datura and Zea. 

 Since it will survive only as an extra chromosome, its discovery 

 indicates that the change determining it is not rare. 



Secondly we have allopolyploids in which, as an exception, 

 homologous but structurally differentiated chromosomes from 

 different sets will pair. Two abnormalities of inheritance not found 

 in diploids are conceivable in such allopolyploids, and both have 

 been found. Consider three homologous but different pairs of 

 chromosomes in a hexaploid, AA, BB, CC. One of these may be 

 lost whether at mitosis or at meiosis to give ABC, AB-. In breeding 



