The combined analysis of variance for the 107 genotypes across four environments indicated significant genotypic differences for yield and yield-attributing traits (Table 1), indicating that genotypes were genetically diverse. The environment component also showed substantial variation for all the traits under study. Thus, the substantial genotypic differences for all traits warrant further investigation.
Analysis of variability parameters
In general, the magnitude of PCV was found to be higher than the corresponding GCV for all the traits, but the differences are quite small (Table 2). This reflected their high heritability values and indicated the less influence of environment on the expression of these traits. Similar results were also obtained by
Singh et al., (2009a) and
Kumar et al., (2013). The magnitude of both GCV and PCV were high for yield per plant and number of branches per plant. This indicated the presence of wide variation for these two traits allowing further improvement by simple selection of the individual traits. The moderate GCV and PCV for clusters per plant, number of pods per plant, plant height, 100 seeds weight and peduncle length suggested the presence of considerable variation for above said characters to allow selection for individual traits. High estimate of GCV for seed yield per plant was observed by
Patel (2000);
Longnathan et al., (2001); Samad and Lavanya (2005);
Makeen et al., (2007); Jyothsna et al., (2013) and
Hemavathy et al., (2015) and for branches per pant was observed by Narasimhulu
et al. (2013);
Patel et al., (2014) and
Singh et al., (2014). But shelling percentage, days to 50% pod maturity, days to 50% flowering, number of seeds per pod, pod length and flowering period estimated to have low GCV and PCV which indicated the narrow range of variability thereby restricting for further improvement of these characters through simple selection. Similar finding had been earlier reported by
Mehandi et al., (2013) and
Patel et al., (2014).
The magnitude of broad-sense heritability was observed to be high for all the thirteen characters indicating least environmental effects on their expression. The highest heritability value was recorded for plant height, followed by number of pods per plant and days to 50% pod maturity.
Patel (2000),
Reddy et al., (2003) and
Makeen et al., (2007); Payasi (2015) and
Hemavathy et al., (2015) also reported high heritability for the traits studied in this experiment. According to
Johnson et al., (1955) the heritability values together with genetic advance would be more useful for correlating selection criteria than heritability alone. High heritability coupled with high GA, were evident for the number of branches per plant, yield per plant, pods per plant, clusters per plant and plant height. This indicated the predominance of additive gene action, implying the potential efficacy of simple selection for trait improvement. These results were agreed upon by
Shiv et al., (2017); Ramakrishan et al., (2018); Jagdhane et al., (2017) for primary branches per plant and pods per plant,
Sana et al., (2017) and
Katiyaret al. (2015) for pods per plant and seed yield per plant. High heritability and moderate genetic advance were seen for flowering period, pod length, days to 50% flowering and days to 50% pod maturity indicating these characters can also be improved by phenotypic selection.
Correlations and path coefficient analysis
Correlation analysis revealed that the association between yield and yield contributing traits was found positive and significant at both genotypic and phenotypic level for all the traits except shelling percentage (Table 3). It suggested that direct selection for these component traits could enhance yield. As a result, greater attention should be placed on these components during selection for higher yield. Similar results for yield per plant were reported by
Makeen et al., (2007) with clusters per plant, pods per plant and seeds per pod (genotypic correlation only);
Parihar et al., (2018) with days to maturity, seeds per pod and test weight.
Among the other traits, the correlation was positive between days to 50% flowering with days to 50% pod maturity, flowering period, number of branches, number of cluster, number of pods per plant, number of seeds per pod and plant height at both the genotypic and phenotypic level. The results confirmed the earlier findings of
Srivastava and Singh (2012) and
Patel et al., (2014) for 50% flowering with primary branches per plant, clusters per plant and pods per plant;
Titumeer et al. (2014) for 50% flowering with primary branches per plant, pods per plant and seed yield per plant; and
Khanpara et al., (2012) for 50% flowering with days to maturity, pods per plant and seed yield per plant. The results were also in agreement with
Kritika (2017),
Dhoot et al., (2017) for 50% flowering with days to maturity and
Rohman et al., (2003); Kritika (2017);
Dhoot et al., (2017) for 50% flowering with plant height. Days to 50% pod maturity showed a highly significant and positive correlation with flowering period, number of branches, number of clusters, number of pods per plant, number of seeds per pod and plant height at both the genotypic and phenotypic levels.
Anita et al., (2022), Pariharet al. (2018) and
Dhoot et al., (2017) observed positive correlations between days to maturity and plant height, number of pods per plant and number of branches, as well as between days to maturity and plant height by
Dhoot et al., (2017). At both the genotypic and phenotypic levels, there was a positive correlation between the flowering period and the number of branches, clusters per plant, pods per plant and plant height. The peduncle length exhibited a positive correlation with the number of branches, the number of clusters, pods per plant, seeds per pod and the plant height at both genotypic and phenotypic levels. Almost all traits, including flowering period, peduncle length, number of clusters, pods per plant, seeds per pod and plant height showed a highly significant and positive correlation with the number of branches. Similar results were reported by
Patel et al., (2014) and
Titumeer et al., (2014). However, correlation of number of branches with the shelling percentage was negative. The correlation of number of cluster per plant was highly significant and positive with number of pod per plant, number of seed per pod, plant height and pod length at both the level. These results were in agreement with the findings of
Hemavathy et al., (2015), Anand et al., (2016) and
Parihar et al., (2018). Number of pods per plant correlated positively with seeds per pod, plant height and pod length. Pod length correlated positively with plant height at both levels. Positive associations existed between number of seeds per pod, pod length and plant height. These results are agreement with
Biradar et al., (2007); Titumeer et al. (2014) and
Makeen et al., (2007) for pod length with seed per pod and
Alom et al., (2014), Khajudpar and Tantasawat (2011);
Singh et al., (2014) for pod length with seed per pod and 100 seeds weight. Since, the correlation was quite complex among the yield attributing traits, it emphasized the importance of path analysis.
Path coefficient analysis based on genotypic correlation is presented in Table 4. The path coefficient analysis revealed that number of pods per plant had the highest positive direct effect on yield per plant (0.771), followed by 100 seeds weight (0.522) and seeds per pod (0.240); indicating their effect for positive correlation with the yield. Most of the traits showed positive direct effect and negative direct- as well as negative indirect-effects were negligible. The indirect effect of number of clusters, number of branches and plant height were positive and high via number of pods per plant. Similarly, days to 50% maturity, peduncle length and number of seed per pod recorded moderate positive indirect effect via number of pods per plant. This might be the reason for positive correlations of yield with all these traits. These positive direct effects of different traits in mungbean were supported by previous studies of
Khajudpar and Tantasawat (2011);
Srivastava and Singh (2012) for pods per plant, seeds per pod, 100 seeds weight,
Ambachew et al., (2015) for 100 seeds weight and pods per plant and
Singh et al. (2009a) for plant height. The results showed that the characters which were most important for correlation studies also proved important by path analysis. Thus, it can be suggested that correlation and path analysis study should be consider together for rapid gain for final improvement in seed yield.
In this study, the lower value of the residual effect (0.012) showed that the observed component traits could explain all of the variation in yield per plant. It revealed that the variables studied in the present investigation explained 99% of the variability in the yield and other attributes besides the traits have been studied to contribute to seed yield per plant
Srivastava and Singh (2012).
