i64 STRUCTURAL HYBRIDS 



must be a criterion of their common ancestry, that is, of their 

 homology (Montgomery and Sutton). This conclusion is borne out 

 by the usual absence of pairing in haploid plants (Ch. X). 



{b) In tetraploids, as also in triploids, pairing takes place between 

 corresponding parts of the pairing chromosomes at random, so that 

 the chromosomes often change partner. Pairing is therefore not a 

 property of the chromosome as a whole but of its constituent parts, 

 and parts of chromosomes in structural hybrids pair according to 

 their individual similarities with parts of other chromosomes. 



(c) The pairing of chromosomes is subject to the conditions of 

 meiosis being normal ; and these conditions can be modified by 

 genetic factors, the abnormal factor being recessive in the known 

 cases (Ch. X). Meiosis, as we shall see later, is one of the most 

 susceptible of all physiological processes to genetic abnormality (Ch. 

 X). Many F^ hybrids which betray no abnormality in any other 

 respect (beyond hybrid vigour) go wrong in meiosis. Where this 

 is due to errors in spindle behaviour we know it is not inherent in the 

 structural differences of the chromosomes ; it must be genotypically 

 controlled. But where it follows a failure of chromosome pairing 

 this failure may be either structurally or genotypically controlled. 



It follows that failure of pairing is not in itself evidence of dis- 

 similarity of the chromosomes concerned, although pairing, which 

 gives rise to chiasma-formation, is evidence of similarity. 



(d) Likewise, genetic conditions may determine a failure to pair 

 of particular members of the complement (as in the micro- 

 chromosomes of Hemiptera, v. Ch. X), or of certain parts of the 

 chromosomes (as in the distal parts of the chromosomes of 

 Mecostethus). Such failure is certainly not the consequence of 

 dissimilarity, because these chromosomes and parts of chromosomes 

 are inherited in the same way as the rest of the complement. 



(e) It has been seen that, while pachytene pairing is the first 

 visible condition, chiasma formation is to be regarded as a second 

 and immediate condition of the metaphase pairing of chromosomes. 

 Therefore the interpretation of chromosome pairing in hybrids at 

 metaphase must take into account the frequency and distribution 

 of chiasmata in the parent. First, pairing in a hybrid of a type 

 with localised chiasmata will be dependent on the identity of the 



