Applying statistics for nonsequential yield component analysis - Information

In the paper, an application of the methodology for analyzing yield as affected by its components that develop at the same ontogenetic level is discussed; it may also be applied to any model in which several traits developing non-sequentially affect their product. The methodology is called "nonsequential yield component analysis". Two applications are presented; the proposed approach is compared with path analysis, commonly applied for yield component analysis, and Piepho's approach. In one example, grain yield of two cultivars of winter triticale (×Triticosecale Wittmack), Fidelio and Bogo, as affected by harvest index and biomass yield, was studied; the results of the three approaches compared were similar. However, in the study of nitrogen uptake as affected by nitrogen-uptake efficiency and crop nitrogen supply (the problem for which yield component analysis may be applied because of the underlying multiplicative model), the results of the methods were different. For both cultivars, path analysis and Piepho's method showed a much stronger influence of crop nitrogen supply on nitrogen uptake than that of nitrogen-uptake efficiency. In the proposed method, however, both these component traits influenced nitrogen uptake to a similar extent. The proposed method is shown to be methodologically better than the other approaches, and to provide different results in some situations. Hence, the new method should be used for nonsequential yield component analysis although it is suggested that path analysis may be applied if the determination coefficient of the linear model is close to 100%.