The hierarchical clustering method performed by Ward’s minimum variance on Euclidian distance is presented with the distribution of 185 mungbean germplasm in different clusters and dendrogram in Table 3 and Fig 1. The accessions in this observation were clustered into 7 clusters based on 25 morphological and 11 agronomic characters. The maximum number of accessions was observed in Cluster I with 60 accessions followed by Cluster II and Cluster III consists of 42 and 38 accessions, respectively. Cluster IVand VIcomprised18 accessions each whereas each of Cluster V and VII involved 5 accessions. The finding of this research was very closely related with many investigations of
Bhattacharya and Vijaylaxmi, 2005,
Singh et al., 2009, Yimram et al., 2009, Gokulakrishnan et al., 2012, Pandiyan et al., 2012, Divyaramakrishnan and Savithramma., 2014,
NB et al., 2015, Sarkar and Kundagrami, 2016,
Jain et al., 2017 and
Sen and De, 2017 who described that 10 clusters of 25 genotypes in mungbean, 11 cluster of 80 genotypes in mungbean, 6 clusters of 340 genotypes in mungbean, 6 clusters of 30 genoypes in mungbean, 8 clusters of 646 genotypes in Green gram, 8 clusters of 374 genotypes in mungbean, 13 clusters of 100 accessions in mungbean, 5 clusters of 23 genotypes in mungbean, 9 clusters of 24 genotypes in soybean and 6 cluster in 30 mungbean genotypes. This finding indicates the significant exiting of genetic diversity among the tested mungbean accessions and provides a good chance for the selection of parents for the improvement program.
The cluster mean and standard deviation of agronomic characters in different clusters have been presented in Table 4.
Among the different clusters, cluster VI showed late-flowering accessions followed by cluster VII whereas the remaining 5 clusters revealed the early flowering accessions, respectively. Similarly, days to maturity showed the same trend as days to flowering among different clusters. The result showed that the plant height at flowering time and maturity had the highest range in the cluster VI and VII and the lowest range in the remaining 5 clusters (I, II, III, IV and V), respectively.
The highestnumber of cluster was observed in cluster II. The cluster VII, VI and II provided the highest number of pod bearing branches per plant. Cluster VII and II recorded with good information due to the presence of the maximum number of pod per plant.The highest range of pod length was classified in Cluster IV, V, I and III and whereas the lowest range of pod length was observed in cluster II, VI and VII, respectively. The average 100 seed weight with above 4.0 g was observed in cluster IV, V, I and III and the average 100 seed weight with 3.19 and below was found in clusters II, VI and VII. Among the different clusters, the maximum yield per plant existed in cluster II followed by cluster IV and the minimum yield per plant was recorded in cluster VI. This result was very closely related with the research of
Yimram et al., 2009, Divyaramakrishnan and Savithramma, 2014,
NB et al., 2015 and
Jain et al., 2017 and indicated that mean performance of agronomic characters was very prominently diverse among different clusters. Therefore, the information of this result will be very useful with some interested characters for future plant breeding program and improvement of a new variety.
The results of descriptive statistics for 11 agronomic characters such as minimum, maximum, standard deviation and coefficient of variation were shown in Table 5. The highest coefficient variation was observed in pod per plant (33.29%) followed by pod bearing branches per plant (31.50%), yield per plant (28.57%), 100 seed weight (25.79) and plant height at flowering (25.49%). Moderate coefficient variation was found in plant height at maturity (18.85 %), number of cluster per plant (14.40%), days to 50% flowering (13.42%) and pod length (12.73%). Days to maturity and seed per pod which are represented as low variability among tested mungbean germplasm with the coefficient of variation of 9.52% and 6.25%. Many research findings were most similar with this finding by
Yimram et al., 2009, Abna et al., 2012, Sarkar 2014,
Singh et al., 2014, Shyamalee et al., 2016, Hapsari et al., 2018, etc. This research showed significant diversity among the tested germplasm especially for yield and its related characters such as pod per plant, pod bearing branches and number of cluster per plant.
Comparison of eigenvectors and eigenvalues for the first three principal components among 185 mungbean accessions was presented in Table 6 and scree plot constructed for 11 agronomic characters of 11 principal components was shown in Fig 2. The values of principal component which are larger than unity (>1) were found in the first three principal components (PC1, PC2 and PC3) with the eigenvalue of 5.16, 2.31 and 1.12, respectively. The analysis based on standardized data revealed that the first three principal components accounted for 78.06 % of the total variability of agronomic characters. Among the first important three principal components, high proportion of total variance (46.91%) was recorded in PC1 and other two principal components PC2 and PC3 were observed with 20.98% and 10.26% of the total variance, respectively (Table 6 and Fig 2).
Among the study of agronomic characters, days to 50% flowering, days to maturity, plant height at flowering, plant height at maturity, number of pod bearing branches per plant and 100 seed weight were more related with the first principal component (PC1) with the eigenvector of 0.34, 0.37, 0.34, 0.37, 0.33 and 0.36. The characters which contributed more relatively with PC2 were the number of cluster per plant, pod per plant and yield per plant (0.35, 0.60 and 0.48). The third principal component (PC3) was generally associated with seed per pod and pod length with the eigenvector of 0.51 and 0.37, respectively. The coefficient of the appropriate vectors was considered by the criterion of
Raji, 2002 who described that the criterion vector was greater than 0.3 which provides large effect to be considered important while the characters which have a lesser value than 0.3 were considered as non-important effects on overall variation. Principal component analysis is beneficial for the breeders to conduct specific breeding programs according to good information about the groups where certain traits are more important. The results of this study are more related with the characterization of germplasm by
Yimram et al., 2009, Abna et al., 2012, Divyaramakrishnan and Savithramma, 2014,
Singh et al., 2014, Sarkar and Kundagrami, 2016 and
Hapsari et al., 2018 who described that five, three, four, three, five and five principal components with 68.9%, 84.04 %, 79.00%, 63.87%, 73.22%, 71.11% and 76.00% of total variance in mungbean.
