Multivariate Analysis for Selection of High Yielding and Early Genotypes in Pigeonpea [Cajanus cajan (L.) Millsp.] for North Western Plain Zone of India

Pigeonpea [Cajanus cajan (L.) Millsp.], is often cross pollinated crop propagated through seeds. It is mostly cultivated in the tropical semi-arid regions of the world and belongs to the family fabaceae. Being India’s second most significant pulse crop after gram or chickpea the crop holds immense potential in contributing to the Indian food basket. Seeds are important sources of dietary proteins (21%), lipids (2.3%) and carbohydrates (67%) (Sodavadiya et al., 2009) along with vitamins like riboflavin, niacin, thiamine and minerals like calcium, magnesium, zinc, copper and iron (Talari and Shakappa 2018). Apart from dietary supplements the crop also finds other uses as feed, fodder and fuel. Being a legume the deep tap root system of the crop enables fixation of atmospheric nitrogen thereby enhancing the soil fertility (Rao et al., 2016). Biometrical analysis of a population through genotypic and phenotypic coefficient of variances, correlation study, path coefficient analyses, heritability and genetic advance explicates the underlying variability existing in a population empowering plant breeders to identify the reliable yield attributing characters for which selection would be effective (Singh et al., 2019). This further helps in selecting potentially better performing lines from the population on the basis of the reliable yield and yield attributing characters as identified from the biometrical analyses (Pushpavalli et al., 2017 and Sharma et al., 2021).

Pigeonpea being a long duration pulse crop takes 160-180 days to mature and hence finds low acceptance in different cropping systems in India. Hence identification of early maturing short duration genotypes, along with high seed yielding traits are desirable for crop intensification programmes. In this context, the current investigation highlights the study of different genetic variability parameters in pigeonpea genotypes evaluated in small scale trial to identify superior high yielding and early maturing pigeonpea lines which could be utilized in further crop improvement fitting in cropping systems suitable for the North West Plain zones of India.

The trial was conducted during Kharif 2020-21 at the crop research field of Pulses Section, Department of Genetics and Plant Breeding, CCS Haryana Agricultural University, Hisar. Experimental design followed was randomized complete block design with three replications. The experimental materials consisted of thirty genotypes along with three check varieties viz. Manak, Pusa 992 and PAU 881 (Table 1).



The observations regarding days to flowering, days to maturity, number of branches per plant, number of pods per plant, 100 seed weight, plant height and yield were recorded from five randomly collected plants in each replication for each genotype. The planting system consisted of four rows per genotype. The row to row and plant to plant spacing was 45 ×10 cm. The mean data obtained was subjected to analysis of variance as laid down by Panse and Sukhatme (1989). The estimation of genotypic and phenotypic correlation coefficients from the phenotypic and genotypic components of variances were carried out as per the methodologies put forward by Fisher (1954) and Al-Jibouri et al. (1958). Similarly, broad-sense heritability, genetic advance and genetic advance as percentage of mean, genotypic and phenotypic coefficient of variation (GCV and PCV) were also analyzed (Lush 1940; Burton 1952; Allard 1960 and Johnson et al., 1955). Path co-efficient analysis was worked out to estimate the direct and indirect effects of different traits on the yield (Dewey and Lu, 1959). Cluster analysis was carried out as per Tocher’s method (Rao, 1952). Group distances based on multiple characters was carried out as per Mahalanobis (1936) utilizing the D² statistic. Data analyses were carried out in INDOSTAT software (https://www.indostat.org). Principle component analysis was carried out using PAST4.02 software package (Hammer et al., 2001).

Mean performance of advance pigeonpea genotypes is depicted in Table 1 and analysis of variance was done for the seven different quantitative characters. The variation among the thirty-three genotypes for the seven quantitative characters was highly significant (p<0.01) which justified subsequent biometrical analysis. All the seven different quantitative morphological and phenological characters significant variability was evident across the genotypes under study. Such divergence also justifies the variation in the population for the characters which could be utilized for crop improvement. Sufficient variability in pigeonpea has been in a number of published findings (Aswini et al., 2021; Fousiya et al., 2021; Sharma et al., 2021).
 
Genotypic coefficient of variation, phenotypic coefficient of variation, heritability and genetic advance
 
The estimates of heritability, genetic advance and the coefficients of variation enable prediction and assessment of gain under selection (Sahu et al., 2015). In the current investigation highest values for GCV and PCV were found in case of seed yield per plant (16.67, 17.40) followed by pods per plant (14.76, 16.95). Similar findings were reported by Patel et al., (2021) in seed yield and pods per plant having high GCV and PCV. For all the characters, values of PCV exceed the values of GCV slightly except for the trait branches per plant where sufficient difference between GCV and PCV is observed (Table 2).



This shows that the effect of environment in the manifestation of that character is high and plant selection on basis of this character would not be effective. For fruitful selection high heritability is desirable with high genetic advance (GA). The broad sense heritability for almost all the characters were high viz. plant height (PH) (90.8%), days to flowering (DF) (96.8%), days to maturity (DM) (90.8%), 100 seed weight (HSW) (92.3%) and seed yield per plant (SYPP) (91.7%). Correspondingly, high values of GA and genetic advance as percent of mean (GAM) was noticed for the characters PPP and SYPP. This is implicative of underlying additive gene action for the traits. This also means that the characters are fixable and hence selection for such traits would be effective. The character DM showed high heritability but low GAM (Table 2). This implies underlying non-additive gene action and hence scope for heterosis breeding. The lowest heritability was noticed in case of branches per plant (BPP) (40.3%), whereas pods per plant (PPP) depicted moderate heritability (75.8). Similar results for high heritability for PH, DF, HSW were found in the published results of Saroj et al., (2013); Vanniarajan et al., (2021).
 
Correlation analysis
 
Correlation study elucidates the direction and magnitude of association which reveal the relative contributions of the characters to yield and thereby facilitate the criteria for the selection of high yielding genotypes (Sharma et al., 2022). In the present study it was seen that the values of genotypic correlations were higher than phenotypic correlations. This means that environment does influence the character association negatively. Although the value reduction for the characters DF and HSW is low which justifies their degree of character stability and association with SYPP. Such results are in correspondence to the findings of Almeida et al., (2010) and Verma et al., (2018) who also testified the character correlations of DF and HSW with yield. Table 3 represents the values of correlation coefficients for the different characters with seed yield.



Positive and significant correlations were found for SYPP with DF (0.300) and HSW (0.365). This indicates that earliness to flowering would contribute to increase in seed yield. As the traits show high heritability along with genetic advance so these characters provide reliable factors for effective selection. Similarly, higher test weight of seeds indicates bold seeds which ensures better crop establishment. Negative correlations were seen in case BPP to SYPP at both genotypic and phenotypic levels. This is implicative environmental effects for this character expression is more and hence the character should be avoided for improved plant selections. Ranjani et al., (2018) also reported the negative correlation of BPP with seed yield.
 
Path coefficient analysis
 
Path coefficient analysis is carried out which divides the correlation into direct and indirect effects quantifying the relative importance of each character (Khan et al., 2016). With yield per plot being considered as the dependent variable and the other six characters as the independent variables path coefficient analysis was carried out. At genotypic level highest positive direct effect was observed in case of PPP (1.35) followed by DF (1.22) and HSW (0.34) (Fig 1).



Similar results of significant high positive direct effects of PPP on SYPP were mentioned by Bishnoi et al., (2019); Hemavathy et al., (2019); Devi et al., (2020). The residual effect of 0.191 pointed out that the independent characters played significant role in expression of the dependent character. PPP showed high positive indirect effect on seed yield through BPP. Similarly, DF exhibited positive indirect effect on DM. Although negative direct effect on seed yield was seen in case of BPP (-1.975), PH (-1.089) and DM (-0.469). The findings are in concurrence to the reports of Sharma et al., (2021) wherein it was emphasized that selection for yield enhancement could be made through direct selection for DF and indirect selection for BPP. The results are suggestive that the above morphological i.e. HSW and PPP and phenological characters i.e. DF are effective indicators for direct selection of high yielding genotypes in pigeonpea.
 
Principal component analysis
 
Principal component analysis (PCA) study was carried out to decipher the underlying factors contributing mostly to the variability in the population. The population under the present study is highly variable and hence the PCA analysis has revealed that 65.8% of variation could be explained by the first three principal components. Trait biplot (Fig 2) revealed that for SYPP had high correlation with HSW and DF. First principal component displayed high contribution to divergence for the characters DF, PH, HSW and SYPP. Correspondingly the genotypes AH 20-10 and AH 20-23 explicated the variability for SYPP and DF. Hence hybridization between these genotypes for the concerned characters could be rewarding.



PC2 substantiated the divergence for the characters BPP, PPP and DM. Such results also found in the published works of Hussain et al., (2021) wherein PC2 showed positive correlation with PPP and BPP. Thus the above characters were identified to be the key components of variability in the population and henceforth could be forwarded for in breeding programmes as significant factors of selection.
 
Cluster analysis
 
Mahalonobis D² method is an effective tool enabling the classification of a population into clusters based upon their genetic divergence. Genotypes with diverse genetic architecture provide for increased chances of genetic recombination manifesting hybrid vigour. Cluster analysis of the population segregated the thirty-three genotypes under study into eight clusters indicating sufficient divergence in the population. The clustering pattern displayed that cluster I and II have the most number of genotypes (11), whereas cluster III, VI and VIII have only one genotype each (Table 4).



The genotypes in cluster VI, VII and VIII were found to be outstanding in terms genetic distance. Upon correlation with the yield values as shown in Table 1 genotypes AH 20-10 and AH20-23 from clusters VI and VIII showed highest yield compared to the checks. Hence, considering the genetic heterogeneity these clusters could be well utilized in hybridization programmes to develop superior recombinants. Among the recent investigations cluster analysis to decipher the diversity of pigeonpea populations were carried out by Hemavathy et al., (2019) and Sandeep et al., (2020).
 

Multivariate analyses provide idea about the racial likelihood of a population. The present investigation highlights character assessment, its association with yield and identification of superior genotypes in pigeonpea evaluated in small scale trial. Significant variability among the genotypes was observed. There is positive association between the seed yield per plot with DF and HSW. Similarly, path coefficient analysis revealed that PPP, DF and HSW had significant positive and direct effect on seed yield per plot at both genotypic and phenotypic levels. Hence significantly the characters HSW, DF and PPP turn out to be prominent traits aiding improved plant selection. Direct selection for the above traits would be suitable for yield improvement. This is suggestive that earliness in flowering along with bold seeded in pigeonpea contributes to yield enhancement. The population exhibited sufficient divergence as evident from the principal component analysis revealing three significant components which explained 65.8% of the total variation and cluster analysis which classified the 33 genotypes into eight different clusters. High inter cluster distance between Cluster VI, VII and VIII is reflective of potential possibilities of hybridization between these clusters for fruitful crop improvement. The genotypes AH 20-10 and AH 20-23 present from the respective clusters were identified as the sufficiently productive with respect to DF, PPP and SYPP. Thus the results are further suggestive of a correlated response of earliness to increased seed yield. Henceforth future crop breeding programmes for pigeonpea improvement should target early duration and high yielding genotypes as highlighted in the present investigation. The identified genotypes could thus be further evaluated in advanced breeding trials in further generations.

None