254 CHROMOSOMES IN HEREDITY : MECHANICAL 



of the hypothesis that is to be induced from these particular obser- 

 vations. The observations require that all chiasmata result from 

 crossing-over between two chromatids of the partner chromosomes. This 

 follows, in the light of our present knowledge of chiasmata, from 

 Janssens' assumption that the diplotene split separates the partner 

 chromosomes so that the pairs of chromatids that remain together 

 after the formation of chiasmata are sister chromatids, derived 

 from the same parent chromosome. 



It now remains for us to describe the evidence for this simplified 

 and defined chiasmatype theory and to show how it can be used in 

 finding out genetical principles from chromosome behaviour. 



4. CYTOLOGICAL EVIDENCE OF CROSSING-OVER 



(i) Kinds of Proof. The proof that crossing-over has occurred 

 between two chromatids of partner chromosomes follows on 

 accepted assumptions from the demonstration that sister chromatids 

 (from the same parent chromosome) are paired on both sides of the 

 chiasma. This demonstration must depend on our ability to make 

 a distinction between the two parent homologues. Such a distinction 

 has been made in regard to three kinds of properties : of develop- 

 ment, of function, and of form. The distinction of development was 

 the first to be made. In auto-tetraploids the chromosomes pair at 

 pachytene with exchanges of partner. If a chromosome could form 

 a single chiasma between two such exchanges of partner, crossing- 

 over must have occurred at the chiasma. Single " intercalary " 

 chiasmata of this kind are formed in tetraploids and triploids 

 (D., 1930 c, Figs. 39, 41). Similar developmental distinctions 

 depend on interlocking and relational coiling. The distinction of 

 function depends on the recognition of constant pairing properties 

 in chromosomes. It enables us to infer crossing-over from special 

 configurations in multiple interchange hybrids, and from the 

 formation of chiasmata between the chromosomes of haploids 

 (Catcheside, 1932). The distinction of form is naturally the easiest 

 to make out. It is now regularly made in many structural hybrids 

 which show not merely the occurrence of crossing-over, but also the 

 occurrence of all the expected relationships between successive 



