138 STRUCTURAL HYBRIDS 



have a recognisable effect in reducing pachytene pairing and 

 chiasma formation, (ii) A reduction in the frequency of chiasmata 

 with or without occasional failure of pairing of the chromosomes. 

 This failure always shows a curve of variation similar to that of 

 chiasma-formation. (iii) A reduction of chiasma-frequency to such 

 a point that pairing fails entirely or almost entirely. 



(iv) Complex Hybrids. Another class of hybrids occurs in 

 which the differences can be partly defined. Amongst these are 

 the complex ring-forming and sex heterozygotes which form 

 permanent self-reproducing types. Interchange, translocation and 

 deficiency can often be inferred in them, but together with such 

 changes others occur that can be located but not defined {v. Ch. XI). 



(v) Polyploid Hybrids (species or true-breeding varieties). 

 The zygote is derived from the union of identical polyploid gametes, 

 each of which is derived ultimately from the union of dissimilar 

 haploid gametes in an ancestor. These hybrids are effective^ 

 true-breeding when they have an even number of sets owing to the 

 pairing and segregation of identical chromosomes derived from 

 opposite immediate parents. In this case they are functionally 

 diploid. But they are also liable to jdeld segregates owing to the 

 occasional pairing of dissimilar chromosomes derived from their 

 ultimate diploid parents (v. Ch. VII). 



(vi) Numerical-Structural Hybrids. All types of cytologically 

 observable hybridity may occur in the same organism, e.g., in 

 triploid (Enothera (Catcheside, 1931 a), Triticum (Kihara and Nishi- 

 yama, 1930 ; Mather, 1935), and Avena (Nishiyama, 1929), and in 

 trisomic Datura (Belling, 1927 h). These will be considered in 

 relation to structural hybrids. 



(vii) Mendelian Hybrids [sensit strido). Where a difference, 

 which might be empirically described as a mendelian difference or 

 group of differences, can be described directly in chromosome terms 

 (as in the case of the trabant of Matthiola or the properties of 

 ring-formation in various genera), it is no longer convenient to 

 speak of it in mendelian terms any more than it is convenient to 

 speak of " mutation " where the change can be defined as 

 translocation or inversion. It therefore follows that mendelian 

 differences for the cytologist are those which show segregation 



