The influence of plant densities on gene action estimates and associations in seven winter wheat parents and their F₂ progeny

The F₂ progeny from a diallel cross involving seven winter wheat parents along with the parents were grown at the Hyslop Agronomy Farm near Corvallis, Oregon to determine the influence of three plant densities on gene action estimates for yield and its primary components. The plant densities were designed to provide different levels of competitions involving solid, six and 12 inch spacing within the rows with one foot spacing between the rows. Gene action estimates were obtained by the combining ability analysis as well as narrow sense heritability estimates. In addition, path-coefficient analysis was utilized to investigate the direct and indirect associations of the primary components of yield under different stresses of competition resulting from the changes in population densities. The morphological characters measured were 1) total yield per plant, 2) kernel weight 3) number of kernels per spikelet, 4) number of spikelets per spike, 5) tiller number and 6) plant height. A small additive gene action for yield was noted in the six inch spacing while additive gene action effects could not be detected in the 12 inch and solid plantings. Yield being a complex trait seems to be affected by the environmental changes resulting from different plant densities. Consistent general combining ability estimates were observed for kernel weight and plant height in all the plant densities, indicating a small genotype-environment interaction. In the spaced plantings additive gene action estimates were obtained for tiller number and spikelets per spike while there was no evidence of additive gene effects in the solid planting. No additive gene action was noted for kernels per spikelet in the 12 inch planting while six inch and solid plantings revealed considerable genetic variability. These results would suggest that the genotypes are susceptible to environmental fluctuations for the traits tiller number, spikelets per spike and kernels per spikelet. The correlation coefficients reveal that in spaced plantings tiller number, spikelets per spike and kernels per spikelet are significantly and positively related to yield. In the solid seeding only spikelets per spike was significantly associated to yield. When the four variables were considered in terms of their associations with yield it was observed in the F₂ that in spaced plantings all the four components of yield have direct positive influence on yield. In the solid seedings however, spikelets per spike and kernels per spikelet had high positive direct effects on yield while tiller number and kernel weight showed a negative direct influence on yield. The data revealed that spikelets per spike and kernels per spikelet are the most important traits contributing towards yield. However, the results obtained with correlation coefficients indicate that a negative association existed between these two traits as well as between kernels per spikelet and tiller number suggesting the possible existence of a biological limitation between these components of yield. The results indicate that a breeding program with emphasis on the selection of plants in competitive conditions in the early generations may make the selection work more efficient. Moreover, increases in yield which considers each of the components separately or in combination of two or more would offer the most promise. By this procedure the breeder would take advantage of the large amount of additive genetic variances associated with each of the components and at the same time take into consideration any biological limitations which may exist.