The pooled analysis of variance for grain yield (Table 3) revealed significant differences among genotypes, environments and significant genotype × environment interaction. The genotype × environment interaction effect was significant emphasizing the influence of environment on grain yield of chickpea genotypes under study. In the current study, as depicted by the mean grain yield of 50 genotypes ranged 1296 kg/ha (G39) to 2222 kg/ha (G8) with mean yield of 1870 kg/ha (Table 2). Among the five environments tested, highest and lowest mean grain yields were obtained from environments E5 (Warangal; 2080 kg/ha) and E2 (Palem;1526 kg/ha) respectively.Genotypic performance of pigeonpea and other crops in relation to different environments has been estimated by worker like
Xing et al., (2021), Amiri et al., (2015).
The AMMI analysis of variance for grain yield (kg ha
-1) of 50chickpea genotypes evaluated across the five environments revealed that the main effects of genotypes (G) and environments (E) accounted for 22.71% and 17.86% of the total sum of squares respectively (Table 3). The G × E interaction also accounted for 59.41% of the total sum of squares clearly depicting the diversity among the genotypes and their differential reaction to environmental conditions, which facilitates selection of genotype suitable for particular environment or multiple environments.Similar findings found in the studies by
Dehghani et al., (2006),
Tolessa et al., (2013) and
Singamsetti et al., (2021) and
Sriwichai et al., (2021) where GEI accounted maximum among the genotypes of field pea, wheat, maize and winged bean respectively indicating that the differences in the response of the genotypes across the environment were substantial and the presence of G × E interaction. AMMI revealed major part of the variation in yield is due to environment indicating the diversity of the environment in the present study when the interaction was partitioned among the first four interaction principal component axis (IPCA) as they were significant in predictive assessment. All the interaction PCA were highly significant capturing 46.33%, 28.63%, 17.46% and 7.6% of the total variation in the G × E interaction sum of squares, respectively. The first two interaction PCA axes combinedly accounted for 80.0% of the G × E interaction sum of squares, endorsing that the model’s use of GGE biplot to explain variation caused by G+E+GEI across environments was effective. Similarly,
Zewdu et al., (2020), Das et al., (2010) and
Umma et al., (2013) earlier reported that first two PCAs explained the maximum GEI. Thus, the GEI of the 50 chickpea genotypes tested in five diverse environments was mostly explained by the first two principal components of genotypes and environments. AMMI revealed major part of the variation in yield is due to environment indicating the diversity of the environment in the present study. Previous reports confirmed that in most of the cases the maximum GEI could be explained through using the first two PCAs
(Fikere et al., 2014; Biswas et al., 2019). Hence, IPCA1 and IPCA2 were used for generation of AMMI1 and AMMI2 biplots.
Biplot analysis
The results of AMMI analysis further elucidated the relative contribution of first two IPCA axes to the interaction effects by plotting with genotype and environment means as depicted in Fig 1 and 2. The mean performances along with PCA1 scores for both varieties as well as environments used to construct the biplots are presented in Table 2 and 3. In the biplot, environments are designated by the letter ‘ENV’ followed by numbers 1 to 5 as suffix (Table 1, Fig 1), while genotypes represented by numbers from 1 to 50 (Table 2, Fig 1). The quadrants in the graph represent: (QI and QII) higher mean, (QIII and QIV) lower mean, (QI and QIV) +ve IPCA1 and (QII and QIII) –ve IPCA1 scores (Fig 1). When a variety and the environment have similar sign on PCA1 axis, their interaction is meant positive and if opposite, their interaction is considered negative. Therefore, a variety with near to zero PCA1 score mean that it had small interaction effect and hence was considered as stable over wide environments. On the contrary, varieties with high mean yield and large PCA scores were regarded as clearly adapted to specific environments (
Abdi and Williams, 2010;
Askari et al., 2017; Mustapha and Bakari, 2014).
Accordingly, in the present study, the chickpea genotypes, G35, G10 and G49 exhibited high yield positive IPCA 1 score and among all genotypes, G35 is being the overall best genotype. On the other hand, G8, G43 and G20 were high yielding genotypes with negative IPCA 1 scores, While IPCA 1 for G9, G24, G7 and G13 were near to zero score and hence have less interaction with the environments out of which G9 and G13 had above average yield performance.
AMMI-2 relationships among genotypes and environments
In AMMI 2 biplot (Fig 2), The biplot 2 provides on the G×E interaction only and not like AMMI 1as the AMMI biplot 1 included main effect also. It was observed from the AMMI 2 biplot analysis (Fig 2) that genotypes showing less interaction in both axes are positioned near to the origin and vice versa. So, the genotypes placed near the origin were regarded as stable compared to others. Likewise, genotypes lying away from the origin and having long spokes were considered as highly interacting types. Hence, environments E5, E1 and E3 exerted strong interaction forces while, the rest two E2 and E4 did less. In the present study, G6, G19, G20, G21, G28, G35, G5 and G3 had more responsive since they were located far from the origin,while the genotypes G39, G17, G13 and G33 were nearer to origin and hence they were less sensitive to environmental forces. In overall, G17 (2059 kg/ha) showed very less genotype × environmental interaction and high stability with high yield.
GGE Bi-plot analysis
GGE Biplot of Environment-view for yield
Environment centered GGE biplot used to estimate the pattern of environments (Fig 3). Lines connecting the test environments to biplot origin as environment vectors were drawn to compare the association between environments. The angle cosine between two environments is used to explain the extent of correlation among them
(Dehghani et al., 2010). Environments E2 and E4 are positively correlated (an acute angle). The presence of wide obtuse angle among environments is an indication of high cross over genotype × environment interaction (
Yan and Tinker, 2006). In the present study, the environments E3 and E1 are negatively correlated (an obtuse angle).
GGE Biplot on environment for comparing environments with ideal environment
An ideal environment is representative and has the highest discriminating power (
Yan and Tinker, 2006). Differentiating capability and precise representation of testing environments are two pivotal roles of GGE biplot. The concentric circles depicted in Fig 4 and 5 aid in visualizing the length of environment vectors, which are the indicators of the discriminating ability of environments and also standard deviation within the respective environments.
Kang-Bo-Shm et al. (2015). The environments E1 and E5 are most discriminating. Accordingly, the environments E2 is the most representative whereas the environments E1 and E5 are least representative. Test environments possessing discriminating ability as well as representativeness are highly suitable for selecting adaptable genotypes. Discriminating but non representative test environment like E1 and E5 are useful for selecting adaptable genotypes.
Biplot of stability and mean performance of genotypes across average environments
The line passing through biplot origin and the average environment depicted with a single arrow constitutes the average environment axis (AEA). Projections of genotypic markers to AEA represent the mean genotype yield (Fig 6). Genotypes are ranked along the ordinate. The results revealed that the line perpendicular to the average environment axis (AEA) separated the genotypes into two groups (those above average and those below average). The genotypes with above average yield are G35, G28, G17, G49, G23, G30, G8, G36, G40, G32, G15, G20, G47. The genotypes with below average yield are G44, G11,G29, G34, G33, G18, G45, G1, G48, G12, G45, G29,G46, G39 and G38. Of all the genotypes G8 and G35 were high yielders, while G39 was the poorest yielder. The AEA ordinate is the double arrowed line that passes through the biplot origin and is perpendicular to the AEA abscissa. Greater projection onto AEA ordinate regardless of the direction means greater stability. (
Yan, 2001,
Yan et al., 2007). The genotypes G10, G23, G28, G18, G12, G9, G27 and G24 with shorter projections are stable over environments. Among these, genotypes stable and high yielding are G10, G23, G28, G17, G9 and G24. Whereas the genotypes G18, G12 and G27 were considered stable but had low seed yield. Therefore genotypes G10, G23, G17 and G28 were identified as stable and high yielding genotypes. The genotype G17 recorded the shortest projection and lies closest to the AEA ordinate hence chosen as best genotype and is at par in yield with the best checks.An ideal genotype express its full genetic potential and perform consistently across the environments.
(Kaya et al., 2006, Yan and Tinker 2006). The environment (ENV2) has been located in the closest of the axis and is thought to be best environment for assessing the stability of the genotypes. This environment is the most representative of all. The ability of genotypes to withstand the aberrant weather conditions in the representative environment are the best ones for the selection. These results are in compliance with the findings of
Akinyosoye, (2022) and
Tiwari et al., (2022) that the environment closest to the axis/concentric point is ideal for selection of stable genotypes. Environment 5 (ENV5) recorded longest projectiles from the axis implying that this is the best environment for selecting high yielding genotypes but it does not mean that genotypes chosen in these environment are stable, however genotypes selected for high yield in these environments can be utilized for breeding for high yielding varieties.
Yan and Kang (2003) reported that the distances from the biplot origin were indicative of the amount of interaction exhibited by genotypes over environments or environments over genotypes.
“Which-Won-Where” pattern analysis
Genotypes having specific adaptative ability for specific environment or group of environments were identified using “Which-Won-Where pattern analysis” and “ranking of genotypes in individual environments” using GGE Biplot tools. In comparison to AMMI, GGE bi-plot precisely identifies the best G × E interaction and also clearly indicates as to which genotype performs best in what type of environment thereby aiding in identification of mega-environments (
Yan and Kang, 2003). The studied environments were divided into three mega environments
i.e., E1, E3and E5. In mega environment E5, the winning genotype is G31, while the genotype G6 is the winner in mega environment E1, whereas, E2 and E3 are closely related and fall under the mega environment E4 and the winning genotype is G35.The other promising genotypes in the mega environment are G10, G8, G28, G40 and G22. The polygon outlook of GGE bi-plot (Fig 7) gives inputs on the best genotype(s) in each environment. The vertex genotypes (G3, G2,G46,G6 and G35) have the longest vectors, in their respective direction, which is a measure of responsiveness to environments (Fig 7). Vertex genotypes in every sector gave highest yields in the environments under that sector. The genotype with the high yield in E1, E2 and E4 is G35. The other vertex genotypes G46, G39, G38,G7andG2 are the poorest in all environments because there is no environment in their sectors. Our findings are in agreement with those of
Ebadi et al., (2010) and
Akinyosoye (2022) who reported that genotype in each environment sector could be grown at each of the locations where they demonstrated comparative advantage in seed production. Genotypes that did not fit into any of the environment sector are not appropriate for cultivation in any of the environments under study.
