Performance of new released cultivars in multi-environment trials are analyzed by various parametric and nonparametric methods for exploring stabile cultivars via biometricians. For a nonparametric estimation of stability in test environments, a new nonparametric statistic as NCV (nonparametric coefficient of variation) has been introduced which is based upon the ranks of the cultivars in each test environment. This new statistic use median as a nonparametric central tendency, and nonparametric index of statistical dispersion as inter-percentile range. The NCV nonparametric stability statistic which presented here is similar to the concept of environmental coefficient of variation which was previously proposed for detection of the stable cultivars in multi-environment trials. Our research showed that the most stable cultivars based on the lowest values of this nonparametric statistic, had the highest mean yield among studied genotypes. Plotting of mean yield versus NCV verified the above results and showed that the highest mean yielding cultivar is identified as the most stable one. This nonparametric statistic would be useful for simultaneous selection for yield and stability, so this model can provide some flexible tool for biometricians.
