Interrelationships between Panicle Weight, Grain Yield, and Grain Yield Components in Oat

The value of panicle weight as a selection criterion for improving grain yield in oat ( Avena sativa L.) is largely unknown. Panicle weight in oat combines two of the three primary yield components, spikelets per panicle and seed weight, into a single trait that is easy to measure. Three oat populations were evaluated in 1987 through 1989 for the effect of selection for panicle weight on grain yield. Genotypes in each population were identified as having either high, intermediate, or low panicle weight in F 5 progeny rows, and progenies were evaluated for panicle weight and grain yield in the F 6 and F 7 generations. A random group of genotypes in each population served as a control. The high panicle weight group did not differ from the random group for mean panicle weight and mean grain yield in all three populations. However, the low panicle weight group had the lowest mean panicle weight and mean grain yield in two (X4020‐15‐l/‘Ogle’ and ‘Centennial’/‘Coker 81‐32’) of the three populations. The correlation between F 5 panicle weight and F 6 and F 7 grain yield was 0.34, 0.06, and 0.56 for X4020‐15‐l/Ogle, ‘Don’/X3530‐47, and Centennial/Coker 81‐32, respectively. Panicle weight had a high positive phenotypic correlation with spikelets per panicle (0.57 to 0.90), but a high negative correlation with number of panicles per unit area (−0.54 to −0.81). Although high panicle weight was not consistently associated with high grain yield, the fact that low panicle weight identified low grain yielding genotypes in two of three populations suggests that low panicle weight would be useful in a truncation selection approach for grain yield improvement. Visual selection, however, would probably be as effective and require less resources.