A TEST OF THE TALL-DWARF HYPOTHESIS IN PEARL-MILLET, PENNISETUM-GLAUCUM (L) R BR

Positive height-grain yield relationships exist for many cereals, but cannot be fully used in breeding because of lodging and harvestability problems in tall cultivars. LAW et al. (1978) proposed a ''tall-dwarf '' hypothesis for wheat, in which the positive effects of minor height genes could be exploited by selecting for them in a major dwarfing ge ne background. The applicability of this hypothesis to pearl miller wa s tested by crossing a set of dwarf S-1 progenies (from a single popul ation) which varied in height onto two male-sterile lines. Mean (by S- 1 pollinator) hybrid grain yield was closely related to mean hybrid he ight (r(2) = 0.60) over a range of mean yields of 3.0-3.9 t ha(-1) and a range of mean heights of 126-165 cm. The effect of height was expre ssed as an increase in grain number in one cross and as an increase in grain mass in the other, indicating the importance of background gene tic effects on yield-height relationships in dwarf hybrids. The concep t of ''tall-dwarfs'' appears to be applicable to pearl millet.