LIMITATION OF CROSSING-OVER 351 



Any crossing-over between differential segments will result in 

 reverse interchange of the kind found in (Enothera and Pisiim and 

 hence in the mutations described. 



Cytological observations on (Enothera show that the segmental 

 interchange which is responsible for regular mutation occurs by 

 crossing-over in this way. Ring-forming plants occasionally have 

 chiasmata formed interstitially between chromosomes which do not 

 associate terminally in the ring. These chiasmata remain interstitial 

 and are seen at metaphase. Evidently there are differential 

 segments in the middle of the ring-forming chromosomes which do 

 not follow the homologies of the ends (Fig. 108). Crossing-over 

 between them at the chiasmata gives new segmental interchange 

 as required by the theory of the origin of half-mutants set out above. 

 Like all secondary structural changes, being due to crossing-over, 

 it occurs at a specific place with a specific frequency and with 

 specific results. 



We now see that the existence of three kinds of segments, pairing 

 segments, interstitial segments and differential segments in large 

 rings expected a priori from simple interchange and established by 

 direct cytological study accounts for (i) the existence of complexes 

 located in the differential segments ; (ii) the mutation resulting 

 from exceptional crossing-over between the differential segments 

 (D., 1936 a) ; (iii) the occurrence of rare crossing-over between 

 genes characteristically associated with them (Renner, 1933) and 

 lying in the interstitial segments or in the pairing segments very 

 close to them. Diploid mutations not involving segmental inter- 

 change are due to this last type of crossing-over. This may take 

 one of two forms. Either particular genes in respect of which the 

 species has been heterozygous may cross over exceptionally and 

 yield a homozygous segregate. Or the genes which cross over may 

 be lethal genes and their allelomorphs. In this case one of the two 

 complexes of a heterozygote emerges in a homozygous form. Thus 

 the mutant ochracea, 7 (2), appears spontaneously from (E. grandi- 

 flora, (14), (De Vries, 1918), and the mutant lutescens after 

 hybridisation from (E. suaveolens (Renner, 1927). Both can be 

 understood as a result of exceptional crossing-over in the interstitial 

 segments, so that lethal combinations are replaced by non-lethal. 



