The Cytogenetics of Speciation in Asiatic Cotton.

HE cultivated cottons are divided into two groups: the diploid Asiatic species Gossypium arboreum and G. herbaceum and the New World amphidiploids G. hirsutum and G. barbadense. The taxa within each of these groups are readily crossed, and the F, in each case is vigorous and fertile. F, and subsequent generations, however, contain either depauperate types or plants that closely resemble one of the parents. HARLAND (1936) , after considerable study on the genetics of G. hirsutum and G. barbadense, concluded that the breakdown following hybridization of these species was due to the rearrangement of their coadapted modifier systems. SILOW (1944a) employed much the same argument to account for the specific differentiation between G. arboreum and G. herbaceum, and rejected any explanation based on chromosome structural differences between the two species. STEPHENS, in a 1950 review citing genetic, cytological, and plant breeding evidence, indicated that multiple gene substitution alone cannot account for the differentiation in these species. He suggested that species of Gossypium differ from one another by many chromosome structural differences too small to be easily detected by cytological methods. He further set forth three cytogenetical criteria by which cotton species could be tested for this cryptic structural differentiation: ( 1 .) Selective elimination of donor-parent genotype in interspecific backcrosses. (2.) A reduction in crossing over in interspecific crosses. (3.) Deviation from autopolyploid segregation ratios in synthetic polyploids of interspecific hybrids. STEPHENS (1949) and LEWIS and MCFARLAND (1952) in separate studies that complemented each other demonstrated selective elimination of donor-parent genes in backcrosses of G. hirsutum x barbadense hybrids. LODEN (1950) previously had shown that recombination for the red (R,)-cluster (Cl,) linkage is less frequent in the G. hirsutum X barbadense hybrid than in either parent. The present study was designed to test all three of STEPHENS’ criteria in a series of crosses involving the two Asiatic diploid species G. arboreum and G. herbaceum. The genetic marker stocks used in these crosses and the alleles they carry are listed in Table 1.