ALLOPOLYPLOIDY 



than in reducing the total amount of it. The reason for this is 

 mechanical. Chromosomes usually begin to pair at their ends and 

 when they reach a place where homology changes, owing to a 

 difference in the linear sequence of the two partners, they may fail 

 to fmd their new partners. They may even go on pairing by torsion 

 with the wrong partners. In either case there is a wastage of the 

 time and energy needed for pairing. There is less crossing-over; 

 fewer chiasmata are formed; these are more exclusively near the 

 ends; and some chromosomes may altogether fail to be paired at 

 metaphase. Hence, in crosses between distantly related species, there 

 is generally a reduction, and sometimes a complete absence, of 

 pairing. The structural differences between the homologous chromo- 

 somes have become so great as to prevent the formation of those 

 chiasmata whose presence would enable us to define the differences. 

 In such hybrids there can be no regular segregation. The gametes (or 

 spores) receive unbalanced assortments of chromosomes in varying 

 numbers, and sterility results. 



The sterility of a hybrid with defective pairing need not, however, 

 be final. Its case is much like that of the haploid. Separation of 

 nuclei may fail at the first anaphase in some mother cells. Gametes 

 are then formed with the unreduced and unsegregated number of 

 chromosomes. Such diploid gametes are fertile and, if they fuse, 

 they give tetraploid offspring. But tetraploids arising in this way 

 are something new. They have four sets of chromosomes, but these 

 chromosomes can associate only in pairs. They form bivalents which 

 segregate regularly. An allotetraploid, as it is called, is therefore highly 

 fertile. It is functionally a diploid, but it combines the characters 

 and the chromosomes of two diploid species ; it is amphidiploid. 

 Such is the origin of the fertile true-breeding and giant radish- 

 cabbage hybrid, Raphaiio-brassica of Karpechenko (Fig. 33). 



The allotetraploid stands at one extreme; the antotetraploid, the 

 product of doubling a homozygous diploid such as a tomato, stands 

 at the other. Between the two is a series derived from diploid 

 parents and corresponding to every degree of differentiation of 

 which diploid hybrids are capable. As a midway type we may take 

 Primula keipensis. Its parents, P. fiorihunda from the Himalayas and 

 P. verticillata from Southern Arabia, have an average of about 

 2 chiasmata in each of their nine bivalents. In the hybrid this is 



135 



