26o CHROMOSOMES IN HEREDITY: MECHANICAL 



stress has developed between the partner chromosomes such as 

 could account for the special properties of distribution foimd in 

 chiasmata. The observed development of relational coiling, such 

 as is found at pachytene and diplotene in Chorthippus or Fritillaria, 

 is evidently compatible with no other assumption than that of 

 crossing-over at chiasmata (D., 1935 c, 1936 h). 



5. CROSSING-OVER IN STRUCTURAL HYBRIDS 



(i) Deficiency Hybrids. It is a matter of observation that the two 

 chromatids between the centromere and the first chiasma pass to 

 the same pole at the first anaphase. On the other hand, owing to 

 the exchange of partners at the chiasma, paired chromatids are 

 separated on the other side of the chiasma. Since apart from move- 

 ment of the chiasma, the paired chromatids are sister chromatids 

 from the same parent chromosome, the first division is reductional 

 for the part of the chromosome proximal to the first chiasma and 

 equational for the part immediately distal to it. 



The results of this principle are seen in the different types of 

 behaviour found in " unequal bivalents." These bivalents are 

 usually the result of pairing of two chromosomes, one of which has 

 lost an end segment (Table 20) . If no chiasma and no crossing-over 

 occur between the place where the difference shows itself and the 

 centromere, then the first division is reductional in respect of the 

 difference. If a single chiasma occurs in this segment, the first 

 division is equational in respect of the difference. These types of 

 behaviour are characteristic of particular unequal bivalents for the 

 simple reason that they depend on the structure of the particular 

 bivalents. The difference in some cases lies next to the centromere, 

 and no crossing-over can ever occur between them ; such bivalents 

 always divide reductionally at the first division in respect of the 

 difference. The difference in other cases lies at the opposite end of 

 the chromosome from the centromere ; they always form one 

 chiasma, and the first division is always equational. It is naturally 

 to be expected from what we know of the variable positions of 

 chiasmata that certain unequal bivalents should be found some- 

 times with crossing-over between the centromere and the inequality 



