Evaluation of grain yield stability of rainfed lentil genotypes by parametric and non-parametric methods

Lentil as a legume crop is one of the most important food crops in developing countries (Karimizadeh & Mohammadi, 2010). In most cases, the interaction between environment and genotype occurs, complicating selection for improved yield among genotypes (Sabaghpour et al., 2004). A cultivar or genotype is considered to be more adaptive or stable if it has a high mean yield but a low degree of fluctuation in yielding ability when grown in diverse environments (Finlay & Wilkinson, 1963). The response of different genotypes in different environments and thus the evaluation of genotype interaction in the environment are of particular importance to researchers in plant genetics and breeding, which can help plant breeders to evaluate genotypes more accurately and select the best one. The purpose of this study was to identify and introduce superior genotypes in terms of yield and yield stability among the lines obtained from preliminary yield test.

Sixteen advanced lentil genotypes along with the control genotypes i.e. Kimia and Gachsaran selected from the advanced yield trial of the 2012-13 cropping year were used as planting material in Gachsaran, Khorramabad and Ilam areas for three years (2013-2016) in a randomized complete block design with three replications. Analysis of variance was performed separately in each environment and then Bartlett test was used to evaluate the homogeneity of experimental errors. Then the combined analysis of variance was performed on seed yield. Stability analysis were performed using environmental variance (S2i), coefficient of variation (CVi), Shukla's variance (2i), Wrick equivalence (Wi), Plaisted statistic ( ), Plaisted and Peterson statistic ( ) and superiority index (Pi) and nonparametric methods, , , , TOP and mean of rank.

Simple analysis of variance showed genetic differences among the genotypes. The combined analysis of variance was performed after Bartlett test, which confirmed variance homogeneity of experimental errors (χ2 = 9.5; P = 0.33). The combined analysis of variance indicated the significant effects of genotype, year, location and interactions of year × location, genotype × location and genotype × year × location. The mean seed yield of genotypes showed that out of 18 studied genotypes, seven genotypes produced higher yields than the average yield of genotypes in the all environments (1566.39 kg.ha-1), so that the highest seed yield were seen in the genotypes 15 and 16, followed by genotypes 8, 12, 11, 5 and 2. Based on environmental variance (S2i), the genotypes 3, 7, 6 and 13 and based on the coefficient of environmental variation (CVi), the genotypes 3, 7, 6 and 15 were identified as stable genotypes. Plaisted and Plaisted and Peterson methods identified the genotypes 4, 9, 2, 10 and 3 as stable genotypes. Wrick equivalence and Shukla variance also introduced the genotypes 4, 9, 2, 10, 3 and 12 as stable genotypes. The Lin and Binns superiority index identified the genotypes 16, 8, 15, 12 and 11 as the most stable genotypes. The genotypes 4, 2, 3, 10 and 9 were the most stable genotypes based on nonparametric index. Based on the index, the genotypes 1, 2, 3, 4, 6, 7, 9, 10 and 18 and based on statistics, the genotypes 3, 4, 10, 9, 1, 7, 2 and 6 were stable genotypes. Based on Fox nonparametric index, the genotypes 16, 11, 2, 8, 12, 13 and 14 were stable genotypes. The genotypes 12, 2, 9, 16, 11, 8, 4 and 3 were more stable based on the total Kong rank. The principal component analysis to evaluate the relationship between seed yield and stability indices showed that seed yield had the highest correlation with MID, TOP and PI. Therefore, these three indices can be used as the best indices to identify superior genotypes in terms of seed yield and stability.

In general, the genotypes 2, 5, 8, 11, 12, 13, 15 and 16 gave higher average yield or equal to Gachsaran control seed yield and were also stable. The genotype 16 and 11 were the most stable genotypes based on MID, TOP and PI and also had the highest seed yield and could be candidates to be released as new cultivars.