Genetic Variability and Interrelationships among Seed Yield and Yield Components in Switchgrass.

Information on the genetic variability of seed yield and yield components and interrelationships among these traits in switchgrass, Panicum virgatum L., are limited. Accordingly, we conducted replicated experiments at Chickasha and Perkins, OK, in 1998 involving 11 lowland type switchgrass populations to assess genetic variation for seed yield and yield components and quantify interrelationships among these traits. Significant (P ≤ 0.01) variation existed among the 11 populations over locations for percent seed set and 100-seed weight. Seed yield and the yield components panicle number per plant, spikelet number per panicle, and seed number per panicle had significant (P ≤ 0.05) population × location interactions, therefore, data for these traits were analyzed separately for each location revealing significant (P ≤ 0.01) differences among populations at both locations for each of these traits. Phenotypic correlation coefficients among seed yield and yield components were computed using over-location data. Seed yield per plant had significant positive correlations with seed number per panicle (r = 0.87**), internode length (r = 0.73**), 100-seed weight (r = 0.67*), and percent seed set (r = 0.62*). Seed yield per plant also had a significant positive correlation with harvest index (r = 0.87**), while the correlation between seed yield and biomass yield was non-significant. Path coefficient analyses revealed that seed number per panicle had the greatest positive direct effect on seed yield. Results indicate the presence of adequate genetic variability within the populations studied for breeding improvement of seed yield. Selection for increased seed number per panicle would be the most effective means of indirectly increasing seed yield. Imaging software is a feasible means of rapidly counting seeds per panicle.