Selection of Genotypes in Mungbean Based on Pattern of Responses Using Parametric and Non-Parametric Approaches

Forty cultivars, including four testers, of mungbean (Vigna radiata L. Wilczek) were grown in four environments. Yield data were used to select the most desirable lines with respect to their pattern of responses. Selection was based on parametric and non-parametric statistics. Ten lines had a CV 20% whereas 28 lines had non-significant squared standard deviation (S²i). The line ML-1 had the lowest value of S²i and the lowest CV, indicating consistent performance in all environments. Deviation from regression was used to test the adequacy of the regression model, and the model was found adequate. Two lines, ML-370 and MUM-3, were found suitable for commercial use. Two other lines MG-46 and ML-384, were similar in response to all four testers, but were not found stable as judged by S²i and CV. To clarify confusion arising from the stability statistics, a quantitative mathematical (parametric) characterization and a qualitative description (non-parametric) are recommended.