Multivalent Formation and Genetic Segregation in Some Allopolyploid Gossypium Hybrids.

gation of parental characters in allopolyploid Aquilegia showed that flower color segregated while other characters did not. This segregation was assumed to result from heterogenetic pairing of chromosomes since about one to two quadrivalents per cell were observed. Segregation in an allopolyploid of Bromus ciliatus x B. laevipes also reflected a high amount of multivalent pairing ( JAKOB 1952). However, segregation ratios may not always depend on multivalent formation since DAWSON (1941) observed almost exclusively bivalent associations in Lotus corniculatus, yet obtained tetrasomic 5: 1 ratios for the inheritance of cyanogenesis. In allopolyploids of Elymus glaucus x Sitanion spp., STEBBINS and VAARAMA (1954) found only a small amount of segregation for parental characters that differentiated the species, yet the mean number of multivalents ranged from four to five multivalents per sporocyte. In explanation, the genome formula was assumed to be A,A,A,A,E, E,E,E, with the multivalents involving chiefly the A, genomes; the small amount of segregation was ascribed to rarity of pairing of the E genomes which carried the factors distinguishing the species. A correlation of cytological behavior with genetic segregation was attempted by FAGERLIND (1937) in allotetraploid Galium Mollugo x uerum. He established certain formulae which related interand intraspecific bivalent and multivalent syndesis to the amount of genetic segregation in self ed allopolyploids. In cotton breeding work it is often desirable to transfer certain genes from wild species to cultivated cottons by means of interspecific crosses. The problem, however, is not simple since the cultivated New World cottons (n=26) are allopolyploids while the wild cottons are diploids (n=13). In synthetic amphiploids from these species the amount of segregation should be related to the extent of homogenetic and heterogenetic pairing ( WADDINGTON 1939). If pairing is completely homogenetic, only bivalents are formed from homologues and there would be no segregation of genes differentiating the parental species; consequently, transfer of genes would not be possible. However, if pairing is heterogenetic, segregation THE in only a few polyploid hybrids. A study by SKALINSKA in 1935 on the segre-