Alternative model for path analysis of small-grain yield

The model most widely used for path analysis of grain yield in small grains assumes bidirectional causal pathways between yield components; however, because yield components develop sequentially, it is questionable from a biological standpoint whether a yield component could influence others that develop earlier. The purpose of this research is to describe the theory and application of an alternative model for path analysis of grain yield that takes into account the sequential development of yield components. Path analysis was performed assuming bidirectional and unidirectional causal relationships among yield components of eight spring barley (Hordeum vulgare L.) cultivars grown at Palmer and Fairbanks, AK, in each of two years. Yield and the yield components spikes per square meter, kernels per spike, and kernel weight were estimated. The unidirectional model seems to reflect operative causal pathways more realisitically, and results in simpler estimates that present a clearer picture of the relationships among grain yield and yield components. Of the three yield components, spikes per square meter resulted in the greatest positive contribution to yield, while the effect of kernels per spike on kernel weight was the most negative compensatory response among yield components.
Two models for path analysis of grain yield in barley were developed under two assumptions of yield component causal relationships. One model assumed bidirectional and the other unidirectional relationships among yield, spikes per square meter, kernels per spike and spike weight. The unidirectional model seemed to give a simpler and more realistic picture of yield and yield component interactions. Among the yield components, spikes per square meter had the most positive effect on yield, while the effect of kernels per spike on kernel weight was the most negative compensatory response.