Improvement of seed yield potential in tall fescue

Abstract

Modest and unreliable seed yield limits the usefulness of some forage tall fescue cultivars. In preliminary experiments, significant correlations for seed weight per panicle but not panicle number were established between spaced plants and sown rows of half-sib progeny derived from them. However, seed yield of the sown rows in the second year was not correlated with first year values. Half-sibling families of an elite breeding pool were sown in rows at three contrasting sites in autumn 1999. The following spring, panicle numbers were counted and in summer seed was harvested and weight per panicle calculated. Families varied significantly for seed yield and yield components, and family values were correlated between sites. Seedlings were raised from remnant seed of the best performing families (based on first year data at the 3 sites) and assessed pre-anthesis for panicle number. Selected plants were isolated and seed harvested. After progeny tests, three sub-sets of elite plants were isolated to produce improved selections. Plot trials sown at Lincoln (New Zealand) and at Pergamino (Argentina), showed significant seed yield improvements from the selections compared with "Grasslands Advance," a standard cultivar derived from the same breeding pool