1982 Volume 57 Issue 1 Pages 25-57
An electrophoretic survey of up to forty presumed isozyme loci was carried out in Asian and African cultivated rice (O. sativa and O. glaberrima) and in weedy and wild forms of African O. breviligulata on a total of 1, 948 strains. Hidden variability was checked by a test of heat stability at thirty isozyme loci in the same species.
The mean gene diversity index ("heterozygosity") was high (0.23) in O. Sativa, medium (0.14) in wild O. breviligulata and low (0.06 and 0.03) in weedy O. breviligulata and O. glaberrima.
In contrast, a maximum of seven alleles at a single locus could be distinguished in wild O. breviligulata while only three at the most were found in O. sativa and two in O. glaberrima and the weedy forms of O. breviligulata.
Calculations of genetic distances showed that the cultivated, wild and weedy African species formed a genetic group distinct from O. sativa. Multivariate analysis of the data on an individual strain basis confirmed this fact and showed in turn that most varieties of O, sativa tended to cluster in two groups which correspond to the so-called Indica and Japonica subspecies. There was however a continuous array of intermediates between the two groups.
Multivariate analysis also showed that the endemic African strains with the most genetic affinity to O. sativa were certain strains of the weedy form of O. breviligulata.
Analysis of F1 pollen sterility among 115 strains of O, sativa permitted the extraction of one small Indica and one small Japonica group of strains characterized by a high pollen sterility relationship but most strains were intermediate.
Each group had little gene diversity, with more than 80% of the loci fixed for one allele. Their isozyme patterns were remarkably complementary in that most of the various gametic associations of alleles found in the numerous intermediate strains could be explained by hybridization between varieties belonging to one and the other group. Consequently, these were assumed to represent the ancestral isozyme patterns of the Indica and Japonica subsp.
Similar genetic distances, which point to a divergence time of one to a few millions years ago were found between O. glaberrima and the ancestral types of Indica and Japonica in the three combinations.
Assuming the neutral theory of isozyme polymorphism, the data confirmed that O. glaberrima was domesticated from O. breviligulata independently of O, sativa. They also suggest that, at the origin of O. sativa, the Indica and the Japonica types were also domesticated independently. The large diversity of O. sativa would result: (i) primarily from introgressions that occurred between the ancestral cultivated types, together with the selection imposed by man, and (ii) secondarily, following the dispersion of cultivars, from the introgression of genes of wild rice in various areas. Some of the weedy forms of O. breviligulata could have originated from natural hybrids between the two cultivated species, O. sativa and O. glaberrima.