Abstract
Genotype × environment interaction (GEI) is of special interest in breeding programs to identify adaptation targets and test locations as well as to determine the most favorable genotypes. There are several nonparametric procedures used to interpret the GEI in multi-environmental trials. The purposes of this investigation were (i) to compare the effect of correction on Huehn’s nonparametric stability statistics and (ii) to use nonparametric statistics for a GEI study on lentil. Nine improved lentil genotypes and one local cultivar were grown in 5 sites during two consecutive years. Results of the nonparametric analysis demonstrated both additive and crossover GEIs. According to uncorrected nonparametric statistics, genotypes G8 and G9 were the most stable and based on corrected nonparametric statistics of Huehn, genotypes G1, G2 and G10 were the most stable. In this investigation, mean of ranks (MR) and coefficient of variation of ranks (CV) with (6)iSwere associated with high mean yield (within the dynamic concept of stability), but the other nonparametric statistics were not positively correlated with mean yield and were identified within a static concept of stability. Results also indicated that corrected nonparametric statistics were not suitable for simultaneous selection of mean yield and stability. Such an outcome could be used to delineate predictive, more rigorous recommendation strategies as well as to help define stability concepts to identify recommendations for lentil and other crops.
Highlights
Most statistical procedures assume that data follow a certain distribution, especially normal distribution
Nonparametric statistical procedures make use of nominal and ordinal scales so data are arranged in an ascending order and assigned ranks according to those observations (Bredenkamp, 1974; Spearman, 1904)
Most plant breeders prefer simultaneous selection for mean yield and stability because the selected genotypes must have high mean values coupled with stable performance
Summary
Most statistical procedures assume that data follow a certain distribution, especially normal distribution. These procedures are known as parametric statistics and estimate population parameters (such as mean and standard deviation) that delineate the underlying distribution of a dataset (Steel and Torrie, 1980). There are some other statistical procedures, which do not make tight assumptions about distribution of data and are known as distribution free or nonparametric methods (Huehn, 1979). Nonparametric statistical procedures make use of nominal and ordinal scales so data are arranged in an ascending order and assigned ranks according to those observations (Bredenkamp, 1974; Spearman, 1904). Ranking classifies observations according to their values but not to their absolute differences. Nonparametric procedures are used less often parametric procedures despite certain advantages (Kubinger, 1986)
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call