REDUCTION AND SEGREGATION 247 



was the basis of the 3 : i segregation in the diploid generation. 

 These are predictions based on cytological principles and they have 

 been verified genetically (v. Wettstein, 1924 ; Andersson, 1927). 

 Tests of this kind will be described where they affect debatable 

 points, but in regard to the chromosome theory in general, they are 

 too numerous to be recapitulated here. They have been dealt with 

 elsewhere (Stern, 1928 ; Sansome and Philp, 1932). 



More important, for our present purpose, is the showing of a 

 parallelism between the rules of heredity and the rules of chromosome 

 behaviour, for this parallelism is a help in directing enquiry. Bearing 

 in mind this distinction, we will now consider the evidence as 

 affecting (i) segregation of homologous chromosomes, (ii) crossing- 

 over, and (iii) qualitative differentiation. 



2. SEGREGATION 



The pairing and separation of the chromosomes is clearly parallel 

 to genetic segregation, as pointed out by Sutton in 1902. The 

 assumption that the one determined the other made it possible 

 to predict haploid or gametic segregation, which has since been 

 abundantly demonstrated genetically. Special cytological observa- 

 tions have since made it possible to verify converse predictions in 

 various ways, as follows : — 



1. Carothers (1921) showed that when grasshoppers {Circotettix) 

 with unequal pairs of chromosomes at meiosis were bred with 

 related forms homozygous for chromosome structure, they produced 

 offspring half with one type of chromosome and half with its mate. 

 This was a combined demonstration of (a) chromosome permanence 

 and heredity, {h) meiotic reduction and mendelian segregation, and 

 (c) fertilisation and mendelian recombination. 



2. Federley (191 2, 1931) showed that PygcBra hybrids having 

 no pairing and no reduction of chromosome number at meiosis 

 {v. Ch. X) showed no segregation of parental characters. 



3. Similarly in parthenogenetic organisms where pairing fails 

 there is no segregation (cf. Ch. XI). 



4. In all allotetraploids formed by doubling of the chromosome 

 number in a hybrid there is a suppression of segregation. This 

 suppression is partial or almost complete according as pairing is 



