INTRODUCTION Soybean (Glycine max L. Merrill) is a high-nutritional oilseed crop grown in India and around the world. It is a member of the Papilionaceae family and has the chromosome number 2n=40. It is known as the "miracle bean" or "golden bean" because of its properties such as protein content (40-42%) and 20% edible oil, in addition to minerals and vitamins. Soybean protein contains all of the essential amino acids, along with cardio-friendly oil that meets 30% of the world's vegetable oil requirements (Khan et al., 2022). Therefore, soybean could be regarded as an ideal food crop for the people of poor and developing countries as it contains high quality protein and reasonable quantity of oil as a source of energy (Kumar et al., 2018). The first and most important step in any crop improvement programme is genetic diversity analysis. There are several important applications for genetic diversity among genotypes in crop improvement. Estimates of genetic divergence provide the extent of diversity existed within the available germplasm and moreover, evaluation of genetic diversity is important to know the sources of genes for a particular trait (Meena et al., 2017). This diversity analysis information can be used to classify germplasm for cultivar identification, assist in parent selection for hybridization, and reduce the number of genotypes required to sample a wide range of genetic variability. A genetically diverse parent is required to increase the likelihood of selecting better seggregants for various characters (Adsul and Monpara 2014). Multivariate analysis, such as Mahalanobis D2 statistics, is extensively used in genetic divergence research findings to group genotypes so that more diverse genotypes are grouped into the most distant clusters. It is also useful in determining the relative contribution of each trait to total divergence (Swar et al., 2021). MATERIAL AND METHODS The present investigation was carried out at the research farm, Agricultural Research Station (Adilabad) and PJTS Agricultural University, Hyderabad during 2021. The experimental material comprised of 55 soybean germplasm lines with five checks viz., JS 93-05, JS 335, KDS-753, AISb-50 and Basara. Basara is the local check of Telangana. All the entries were evaluated in Randomised Complete Block Design (RCBD) in three replications by keeping inter and intra-row spacing of 45 and 10 cm respectively. The observations were recorded on whole plot basis for days to 1st flowering, days to 50 % flowering, days to 1st pod initiation, days to maturity, biomass and germination % whereas, data were recorded for plant height (cm), number of nodes on main stem of plant, number of branches plant-1, number of clusters plant-1, number of pods plant-1, 100 seed weight (g), seed yield plant-1 (g) and harvest index (%) traits based on five randomly selected plants of each entry per replication and protein content (%) and oil content (%) were estimated as per protocols of AOAC (1990). The recorded data was subjected to analysis of variance and Mahalanobis D2 statistics were used for genetic divergence analysis. Analysis of variance for Randomised Complete Block Design (RCBD) was performed as per the method suggested by Panse and sukhatme (1978). The genotypes were clustered by using Tocher's method. The intra- and inter-cluster distances were calculated and were used to describe the genotype relationship with the help of the formula proposed by Singh and Chaudhary (1977). The dissimilarity coefficient between genotypes was organised into a reasonable hierarchical system and estimated using the D value proposed by Sneathe and Sokal (1973). All the above mentioned analyses were performed using WINDOWSTAT software. RESULTS AND DISCUSSION A. Grouping of Accessions into Various Clusters The Tocher's method was used to group 60 soybean germplasm lines based on D2 values. A total of 14 clusters were formed, within which two clusters contained multiple genotypes and the remaining 12 clusters contained only one genotype, indicating that genotypes are highly diverse in nature. The lines viz., KDS 1175, Z-5, PS 1682, Z-P1, Z-3, DLSb 3, KDS 1169, JS 20-03, PS 1675, Asb-62, KDS 1187, Z-16 belongs to clusters III, IV, V, VI, VII, VIII, IX, X, XI, XII, XIII and cluster XIV respectively which resembles that their distinct genotypic composition from the remaining genotypes. Cluster I was found to be the largest comprising of 32 genotypes followed by cluster II with 16 genotypes (Table 1, Fig. 1). Similar results are observed in findings of Pawar et al. (2015), Naik et al. (2016). B. Average Inter and Intra-cluster Distances The intra-cluster D2 values ranged between zero and 113.86. Cluster II (113.86) had the greatest intra-cluster distance among the 14 clusters, followed by Cluster I (84.44), indicating the presence of some amount of genetic divergence in the genotypes belonging to these clusters. The genotypes carefully selected from these clusters may be used in future recombination breeding programmes. The remaining ten clusters had zero intra-cluster distance as they each had a single genotype (Table 2). The inter-cluster distance ranged from 56.34 to 858.60. Clusters VII and XIII had the longest and best inter-cluster distance (858.60), followed by clusters IX and XIV (807.51), clusters VII and X (794.75), clusters VII and IX (755.77), and clusters IV and X (697.77). Similarly, the relatively short inter-cluster distance was observed between clusters V and VIII (56.34), followed by clusters IV and VII (95.97), clusters III and V (97.94), clusters X and XIII (110.47) and clusters I and III (120.67) (Table 2). The greater the distance (D2) between clusters, the greater will be the genetic divergence between accessions belonging to that cluster (Ramyashree et al., 2016). Crossing between genotypes from different clusters increases the variability of the gene pool. Soybean germplasm lines from clusters VII (Z-3) and XIII (KDS 1187) were found to be the most divergent, and thus may be used in recombination breeding programmes to maximise heterosis. C. Cluster Means of the Characters A considerable difference was noticed among the cluster means for all the characters. The genotypes in cluster IX had minimum mean performance for days to 1st flowering (23.00), days to 50% flowering (29.00), days to 1st pod initiation (36.00), days to maturity (86.00), plant height (40.00 cm), and germination percentage (71.67%). The genotypes in cluster VI had a maximum mean for days to 1st flowering (44.00), days to 50% flowering (47.00), days to 1st pod initiation (66.00) and days to maturity (115.00). Cluster XIV had the highest means for branches per plant (7.80), clusters per plant (29.00) and nodes per plant (30.47), but it had the lowest mean for oil content (15.63).Cluster X had the lowest mean performance for branches per plant (3.20), clusters per plant (9.20), pods per plant (23.00), single plant yield (11.20) and biomass (388.45) and it had the highest value for 100 seed weight (15.33). For nodes per plant cluster, VIII recorded a minimum value of 10.93 and for 100 seed weight, XII (11.23) had a minimum value. Cluster VII had the highest mean for both seed germination percentage (87.00) and pods per plant (88.87). Cluster III had the lowest value for protein content (37.87). The traits like yield (25.40) and biomass (8054.37) showed maximum mean in cluster IV. For the harvest index, clusters XI (44.50) and XIII (35.07) recorded maximum and minimum mean values. Cluster XI had the highest mean for protein content (43.93), whereas cluster XIII had the highest mean for oil content (22.10) (Table 3). D. The Contribution of each Traits towards Total Divergence Each trait's contribution to total diversity is calculated based on the number of times the character appears in the first rank. The seed yield per plant appeared 560 times in the first rank, indicating the greatest contribution to total diversity. Biomass, on the other hand, appeared once in the first rank, indicating a lower contribution to total diversity. The relative contribution of each traits towards total divergence is mentioned in Table 4 and Fig. 2. The results revealed the highest contribution of seed yield per plant (31.26%) towards total diversity, followed by harvest index (21.18%), plant height (12.76%), seed germination rate (10.00%), while other traits like oil content (0.16%) were noticed as the lowest contribution towards diversity, followed by pods per plant (0.33%), biomass (0.5%) and days to 1st pod initiation (1.46%). The result suggests that the soybean accession selected for the present study are mostly divergent for seed yield per plant, harvest index, plant height and seed germination rate as these four characters contributed 75.2 % to the total diversity. The present findings are in agreement with the results obtained by Chandel et al. (2013) for harvest index and pods per plant, Manav and Arora (2018) for seed yield per plant and Kachadia et al, (2014).