INTRODUCTION Pigeonpea (Cajanus cajan (L.) Millsp.) is a photosensitive, highly nutritious tropical grain legume crop belongs to the family Fabaceae with chromosome number (2n=22). Introduction of short duration pigeonoea varieties have been developed to accommodate in diverse cropping systems but couldn’t have comparable yield as long duration ones. Hence, breeding of varieties with high yield and early maturing is the need of the hour. Genetic diversity studies are a prerequisite for any crop improvement program. Hybridization and selection are two crucial steps in the process of plant breeding. Thus, study of many morphological characters in germplasm is important for assessment of the differences among populations in order to identify population differences and breeding potential. Plant breeders frequently measure a huge number of characteristics, some of which may not have enough discriminatory power for evaluation, characterization and management germplasm (Maji and Shaibu 2012). In such case, principal component analysis (PCA) may be used to reveal patterns and eliminate redundancy in data sets (Mahalingam et al., 2020) as morphological and physiological variations routinely occur in crop species. PCA simplifies a large data series into smaller set of variables without losing any vital information of original data set by seeking for groups that have very strong inter-correlation with per cent variation explained for each component towards the total variability (Das et al., 2017).. Biplot Analysis is an exploratory graph method used in statistic with two-dimensional chart that represents of multidimensional dataset. The research aim was planned to estimate the diversity through classification of genotype through Principal Component Analysis (PCA). MATERIALS AND METHODS Sixty eight pigeonpea genotypes along with five checks (Table 1) were evaluated on the basis of ten agro-morphological traits at Department of Pulses, Tamil Nadu Agricultural University, Coimbatore. The genotypes were raised in Augmented Block Design II (ABD II) during rabi, 21. Ten quantitative traits used for diversity analysis viz., days to fifty per cent flowering and days to maturity, plant height (cm), number of primary branches per plant, pod length (cm), number of pods per plant, number of seeds per pod, pod weight per plant (g) and hundred seed weight (g) and single plant yield (g). Principal Component Analysis (PCA) was performed using the mean data of pigeonpea accessions. The analysis was carried out using R software package. RESULTS AND DISCUSSION Analysis of Variance (ANOVA) performed for ten quantitative traits revealed presence of significant variation among genotypes and can be subjected to further analysis (Table 2). Eigen vectors and eigen values. The Principal Component Analysis (PCA) resulted in the formation of three principal components with eigen value greater than 1. The eigen value for PC1, PC2 and PC3 was 3.41, 2.24 and 1.21, respectively. Other principal components (PC4 to PC10) had discriminatory power. The first three principal components which accounted for 68.72 per cent of total variation which was depicted in the scree plot (Fig. 1) and Table 3. Hamid et al, (2011) reported 76.2% of total variance during evaluation of hundred germplasm lines for various morphological and agronomic traits. Component matrix showing latent vectors associated with first three principal components was given in the Table 4. Maximum amount of variation was contributed by PC1 (34.16%) to the total variance followed by second principal component (PC2) (22.44%) and third principal component (PC3) (12.10%). The first principal component (PC1) contributed maximum variation to the total variation showed positive loadings for all the studied traits. Similar findings were reported by Rekha et al. (2013); Tharageshwari and Hemavathy (2020) using pigeonpea germplasm accessions. The eigen vector value > 0.4 may be considered as large loadings. The trait with eigen vector values greater than 0.4 could be effectively utilized for selection which contributes greater variation to the principal component. Loadings with maximum value was observed in traits viz., number of pods per plant (0.973) followed by single plant yield (0.905), pod weight (0.87), number of branches per plant (0.229) and number of seeds per pod (0.207). The results were found accordance with findings of Hemavathy et al. (2017), where PC1 contributed maximum to total variance with positive correlation with all traits except single plant yield. The second principal component (PC2) showed 22.44 per cent of variation which exhibited positive loading for all traits except number of seeds per pod (-0.027). The traits viz., days to maturity (0.984), days to 50 % flowering (0.98), plant height (0.337) and number of branches per plant (0.108) which showed large positive loading and provide maximum contribution to percent of variance. Similar findings were reported by Yohane et al. (2020) for traits viz., plant height, number of primary branches, number of pods per plant, hundred seed weight and grain yield per plant. The third principal component (PC3) contributed 12.10 per cent of total variation. It showed positive loadings for traits viz., number of seeds per pod (0.974), single plant yield (0.233), pod weight (0.185), plant height (0.058), pod length (0.052) and number of branches per plant (0.027). Conversely, negative loading was present in traits viz., days to fifty per cent flowering (-0.012), days to maturity (-0.019), number of pods per plant (-0.001) and hundred seed weight (-0.04) which depicted that these traits showed negative contribution to the genetic divergence. The results were supported by the findings of Zavinon et al. (2019) where, the traits days to 50 per cent flowering and days to maturity showed negative contribution to divergence. PCA Biplot analysis. PCA biplot was indicated with two-dimensional chart that represents multidimensional dataset (Fig. 3). The correlation between the variables can be detected by measuring the angle between two variables. If the angle between two variables was 90 degrees, it was represented that there was no correlation between two variables. Positive correlation was detected if the angle was less than 90 degree and if the angle was more than 90 degrees, it was found that negative correlation between two variables. Vector line length was used for representation of variance level of the variable. The distance between genotypes and the variable were used to detect the contribution of variable to the genotype. Closer the distance higher the value and vice versa for farther distance. Considering the above conditions, PCA biplot was concluded that the traits days to maturity and days to fifty per cent flowering had weak positive correlation with hundred seed weight, zero correlation with pods per plant, single plant yield and pod weight. Negative correlation was observed between number of seeds per pod and the traits days to maturity and fifty per cent flowering. High positive correlation was observed among the traits viz., days to maturity, days to fifty per cent flowering, plant height, pod length and number of branches per plant. The traits viz., days to maturity, days to fifty per cent flowering, single plant yield (g), number of pods per plant and pod weight showed maximum variance as these variables were represented to have maximum vector length. Contrarily, the traits pod length, hundred seed weight and number of seeds per pod showed minimum variance with shorter vector length. The results were in concurrence with the findings of Yohane (2020) while analysing phenotypic divergence of eighty one pigeonpea genotypes for both quantitative and qualitative traits. In the PCA biplot, the best performing genotypes were concentrated in the right quadrant of the biplot. The genotype AL 2184 which observed to have minimum value for all the variables was plotted in the negative left quadrant. Likewise, the genotypes CO(Rg) 7, DPP 3-81 and ICP 9808 which showed maximum performance of all the variables plotted on the right quadrant.