Pooled ANOVA
Pooled analysis of variance revealed significant differences among the environments, genotypes and genotypes × environments interactions for seed yield per plant (g). This indicated differential/non-linear response of genotypes to the environments. The environments (E) effect were the most important source of yield variation, accounted for 44.13% of total sum of squares (TSS) followed by genotypes, error and G × E interactions effects which accounted for 23.22%, 16.11% and 15.88% of TSS, respectively (Table 1).
Mean comparison
Mean seed yield under E
1 (Yp) was 6.71 g and ranged from 4.84 g (UM-302) to 9.69 g (UM-40). While, mean seed yield under E
2 (Ys) was 4.80 g and ranged from 3.53 g (UM-163) to 6.18 g (UM-60 and UM-124). Thus the data indicated that mean seed yield per plant decreased under stress and the range was wider in E
1 as compared to E
2. The genotypes UM-40, UM-51, UM-112, UM-38 and RMt-143 showed higher seed yield and genotypes UM-302, RMt-1, UM-163, UM-44 and RMt-305 showed lower seed yield in E
1. Whereas, genotypes UM-124, UM-60, UM-55, UM-28 and UM-4 recorded higher seed yield and genotypes UM-163, RMt-305, UM-47, UM-302 and UM-45 showed lower seed yield in E
2 (Table 2). The most of genotypes showed increase seed yield in which genotypes UM-40, UM-51, UM-112, UM-38 and RMt-143 showed highest seed yield in comparison to short duration variety RMt-305, respectively (Table 2).
A wide range of variation was observed among the different stress indices for all 48 genotypes of fenugreek and
Kumar et al., (2020) in mungbean also found the same findings. To evaluate drought tolerant genotypes using TOL index, higher value of TOL demonstrates more changes of genotype yield in stress and non-stress conditions and shows the susceptibility to non-stress condition.
Fernandez (1992) and
Rosielli and Hamblin (1981) stated that selection based on TOL index leads to selection of genotypes with their yields in non-stress condition are low and have lower MP. The results of this experiment showed that UM-53, RMt-1, UM-301, UM-50 and UM-4 were the most tolerant and UM-40, UM-112, UM-47, UM-100 and UM-26 were the most sensitive genotypes to the moisture stress based on TOL index. For SSI, the higher value refers to more susceptible to stress, therefore, the genotypes UM-112, UM-40, UM-47, UM-100 and UM-26 were the least tolerant genotypes and UM-53, RMt-1, UM-301, UM-50 and UM-4 were more tolerant genotypes. The SSPI resulted the same genotype ranking as TOL. Mean productivity (MP), geometric mean productivity (GMP) and stress tolerance index (STI) showed similar ranking of genotypes relative to stress tolerance
(Sofi et al., 2018).
Based on STI, the greater the difference between the yields found in normal and stress conditions, the smaller the amount of stress tolerance index and
vice versa. Thus, genotypes UM-55, UM-28, UM-56, UM-124 and UM-40 were found moisture stress tolerant with high STI and high seed yield under normal irrigation and moisture stress conditions, while genotypes UM-163, UM-302, RMt-305, UM-44 and RMt-1 displayed the lowest amount of STI and seed yield under moisture stress environment. GMP resulted the same genotype ranking as STI. For MP, the higher value refers to more tolerant to moisture stress, therefore, the genotypes UM-40, UM-55, UM-28, UM-56 and UM-51 were more tolerant whereas, the genotypes UM-302, UM-163, RMt-305, RMt-1 and UM-44 were least tolerant to moisture stress. YI can be used as a selection criterion, although it only ranks cultivars on the basis of Ys (mean seed yield in E
2). Based on YI, genotypes UM-60, UM-124, UM-55, UM-28 and UM-4 had the highest YI and Ys, hence more tolerant whereas, UM-163, RMt-305, UM-47, UM-302 and UM-45 had the lower YI and Ys. According to K1STI, the genotypes UM-40, UM-51, UM-112, UM-38 and RMt-143 were the most tolerant whereas, the genotypes UM-302, RMt-1, UM-163, UM-44 and RMt-305 were the most sensitive. According to K2STI, the genotypes UM-124, UM-60, UM-55, UM-28 and UM-4 were the most tolerant whereas, the genotypes UM-163, RMt-305, UM-47, UM-302 and UM-45 were the most sensitive (Table 2). It was concluded that MP, GMP and STI values are convenient parameters to select high yielding genotypes in both the stress and non-stress conditions (
Susmitha and Ramesh, 2020) whereas relative decrease in yield under stress, TOL, SSI and SSPI values are better indices to determine tolerance levels.
Correlation coefficient
To determine the most desirable stress tolerance index, the correlation coefficient between Yp, Ys and stress indices were calculated (Table 3). The best indices are those which have high correlation with seed yield in both E
1 and E
2 conditions and would be able to identify potential upper yielding and drought tolerant genotypes
(Talebi et al., 2007).
Seed yield under stress condition (Ys) had a weak positive association (r = 0.216) with seed yield under non-stress condition (Yp), indicating that high potential yield under optimal conditions does not necessarily result in improved yield in a moisture stress environment (and the opposite is true) because the genes controlling yield and drought tolerance are different (
Rosielle and Hamblin, 1981). Similar findings were reported by
Fernandez (1992),
Mohammadi et al., (2010), Farshadfar et al., (2013) and
Sahar et al., (2016). The seed yield under non-stress (Yp) had significant positive association with TOL (0.841), SSI (0.720), STI (0.810), MP (0.894), GMP (0.815), SSPI (0.841) and K
1STI (0.995), whereas non-significant and positive association with YI (0.214) and K
2STI (0.217). The seed yield under stress (Ys) had significant positive association with STI (0.741), MP (0.630), GMP (0.740), YI (1.000) and K
2STI (0.996), while K
1STI (0.185) exhibited non-significant and positive association. In addition, TOL (-0.347), SSI (-0.501) and SSPI (-0.346) showed significantly negative association with yield under stress (Ys). The indices STI, MP and GMP exhibited good correlation with seed yield under both the environmental conditions, therefore, selection based on MP, GMP and STI will result in the selection of genotypes with higher moisture stress tolerance and yield potential in both the environments, while TOL, SSI, YI and SSPI exhibited good correlation with seed yield under moisture stress condition. Similar findings were also reported by
Siahsar et al., (2010) in lentil, Zare (2012) and
Saeidi et al., (2013) in barley,
Singh et al., (2015) and
Mohammed and Kadhem (2017) in wheat. Thus, these indices may be used as selection criteria in breeding programme for moisture stress tolerance.
Ranking method
The estimated values of various stress tolerance indices indicated that the identification of moisture stress tolerant genotypes based on a single criterion was contradictory. Different indices introduced different or same genotypes as stress tolerant. To determine the most desirable moisture stress tolerant genotype according to the all indices, mean rank of all indices were calculated and based on this criterion the most desirable and moisture stress tolerant genotypes were identified (Table 4). The genotypes UM-124, UM-60, UM-55, UM-28 and UM-4 were identified as the most tolerant genotypes for moisture stress, while genotypes UM-163, RMt-305, UM-45, UM-302 and UM-47 as the most sensitive (Table 4). Such strategies of using different tolerance indices and ranking pattern for screening of tolerant genotypes were used by several other workers such as
Farshadfar et al., (2012), Farshadfar et al., (2014) and
Mohammed and Kadhem (2017) in wheat.