276 CHROMOSOMES IN HEREDITY: MECHANICAL 



(ii) An interchange and reduplication heterozygote in maize was 

 recognised by its having a ring of four chromosomes, one of which 

 had a small extra segment seen at prophase of meiosis. When this 

 was crossed back to a form homozygous in respect of both the 

 interchange and the reduplication the progeny could be classified 

 cytologically according to the observed occurrence of crossing-over 

 between the point of interchange and the point of reduplication. 

 They could also be classified genetically according to the occurrence 

 of crossing-over between factors known on previous evidence to 

 occur in the segment between the interchange difference and the 

 reduplication difference. The two classifications agree to the 

 extent that would be predicted from the previously known linear 

 order of the factors studied (Creighton and McClintock, 1931). 



The characteristic results by which we recognise secondary 

 structural change are the appearance of frequent new chromosomes 

 of particular types, especially fragments, and genetically the appear- 

 ance of new mutations. In CEnothera this method of mutation will 

 be described in detail (Ch. IX). 



The new fragments probably arise in one of two ways. Either 

 an interchange results from crossing-over and is unequal, so that 

 one very large and one very small new chromosome are produced. 

 Or the crossing-over is between relatively inverted segments, giving 

 a bridge which breaks and so leaves a deficient chromosome. This 

 second method will probably be effective in polyploids. 



The circumstances favouring secondary structural changes are of 

 several kinds, which require separate consideration. First, we have 

 the interspecific hybrids wh/ich are presumably structural hybrids 

 such as those that have already been considered. Thus in derivatives 

 from interspecific crosses with Avena saliva and of the hybrid 

 A. strigosa X A. barbata, Nishiyama (1933, 1934) and Philp (1934 b) 

 have found new types of small chromosome attributable to crossing- 

 over (c/. also Buxton and D., 1932, in Digitalis; Katayama, 1935, 

 in jEgilolricum ; and Ono, 1935, in Rumex). 



Hybrids that have been made between Crepis species having 

 chromosomes of different shapes, reveal the occurrence of secondary 

 structural changes. It would be expected that, crossing-over 

 having taken place between them in the first generation, chromo- 



