ILLEGITIMATE CROSSING-OVER 275 



(iv) Secondary Structural Change. Crossing-over in structural 

 hybrids of the kind that has been discussed gives rise to recognisably 

 new forms of chromosomes. There is an exchange of dissimilar 

 segments between chromosomes with two such segments, as a result 

 of crossing-over in the region between them. We have seen how it 

 occurs in the differential segments of multiple interchange hybrids 

 and also in dislocated segments in inversion hybrids. This crossing- 

 over resembles interchange with the difference that it can occur 

 only at meiosis and will occur only at particular places. Moreover, 

 it will occur with a regular frequency. We can therefore distinguish 

 by the later results between this secondary structural change and 

 primary interchange even when its immediate consequence of 

 chiasma formation is not seen. 



The precise circumstances of secondary structural changes have 

 been determined cytologically and genetically in two experiments. 



(i) Females of Drosophila were taken having, in the first place, two 

 structural differences between different parts of their X chromosomes, 

 viz., attachment and non-attachment of Y' (Fig. 121) ; fragmenta- 

 tion into two parts, one of them attached to the fourth chromosome 

 and non-fragmentation ; and, in the second place, two gene differ- 

 ences lying between the two structural differences. They were 

 crossed to males having a single normal (i.e., non-fragmented and 

 non-attached) X chromosome with the two corresponding genes 

 recessive. The gene differences immediately adjoined the two 

 structural differences, so that, in examination of the offspring, if 

 crossing-over was seen to have taken place in an oocyte by the genetic 

 character of the individual derived from it, new chromosome types 

 having the two kinds of structure at each end of the X chromosome 

 should be found cytologically in that individual, and where no 

 genetical evidence of crossing-over was seen, one of the two 

 maternal types of X should be found. This was tested in four 

 experiments. Crossing-over with formation of new chromosome 

 types was found in 156 flies ; no crossing-over was found with no 

 new chromosome types in 203 flies ; two unexpected types were 

 found and three contrary to expectation ; these are believed to be 

 due to errors of genetical recording such as cannot be eliminated 

 entirely (Stern, 1931). 



