Abstract
SUMMARY: Fifty-nine full-sib progenies produced in a North Carolina Model I design were assessed fortotal yield (sum of three harvests) under four treatments differing in plot size and density. The treatments were: swards of 0.25 m2, spaced plants at 0.6 x 0.6 m and single and double rowswith 0.3 m betweenrows and 0.15 m between plants, with ten plants per row. Estimates of plot heritabilities ranged from 0.24 for spaced plants to 0.59 for double rows. Heritability in the sward was 0.35. Additive genetic correlations between the sward and the other treatments were surprisingly high: about + 0.86 for the spaced plants and between +0.8 and +0.9 for the row plots. The implications of these estimates for the relative selection efficiencies of the non-sward treatments are discussed.