Based on the D
2 statistics, 196 groundnut genotypes were grouped into 25 clusters (Table 1). Cluster I with 90 genotypes was the largest followed by cluster II (55) and cluster IV (21). All other clusters were mono-genotypic except cluster III with eight genotypes and Cluster XXIII with two genotypes. The magnitude of D
2 values observed in the study suggested considerable diversity in the mini core set for resistance to insect, disease and for productivity parameters. Interestingly, genotypes with different pedigree and geographic origin but with the same level of performance of the trait under study belonged to the same cluster. This may be evidenced from the belonging of the genotype ICG 13787 with bunch growth habit from West Africa and the genotype ICG 2925 with runner growth habit from South Asia to same cluster (II) might be due to having same level of resistance reaction to
Spodoptera litura (
Saleem, 2018). Earlier, no relationship between geographical distribution and genetic diversity was observed in confectionary groundnuts
(Venkateswarlu et al., 2011). Lack of relationship between genetic and geographic diversity could have arisen from genetic drifts and selection in a particular environment (
Murthy and Arunachalam, 1966). Similar clustering pattern in groundnut based on productivity traits and reaction to late leaf spot has been reported earlier by
Vijay (2015) and
Chaudhari et al. (2017). Therefore, selecting genotypes as parents in the crossing programme on the basis of genetic divergence analysis would be more rewarding than the choice made only on the basis of geographic diversity (
Bhakal and Lal, 2015;
Chaudhari et al., 2017). Conspicuously, 20 clusters (V, VI, VII, VIII, IX, X, XI, XII, XIII, XIV, XV, XVI, XVII, XVIII, XIX, XX, XXI, XXII, XXIV and XXV) were solitary indicating that these genotypes might have completely different genetic makeup from the rest of the genotypes and from each other and thus led to the formation of such monogenotypic clusters. Also, such solitary clusters generally exhibit
-superior/inferior performance for few traits due to total isolation preventing gene flow or due to intensive natural/artificial selection for diverse adaptive complexes
(Chaudhari et al., 2017; Mahesh and Hasan, 2018).
Twenty nine genotypes showing resistance to
Spodoptera litura limited their distribution among nine clusters (I, II, IV, XI, XII, XIII, XIV, XVI and XXIII) where cluster II had fifteen such genotypes followed by six in cluster IV. Distribution of resistant genotypes in different clusters indicated diversity with respect to other traits as well. For example, ICG 13787 and ICG 928 were resistant to
Spodoptera litura but they differed with respect to their performance to late leaf spot, shelling per cent and yield per plant. Five genotypes resistant to late leaf spot were found distributed in the clusters II, IV, XVIII, XIX and XXI indicating diversity among them. Interestingly, two rust resistant genotypes were distributed in two clusters (IV and XXII). Conspicuously, the genotype ICG 2381 with resistance to all the three biotic stresses (Saleem, 2018) belonged to cluster IV. The susceptible check genotypes JL 24 and TMV 2 belonged to cluster III where no resistant genotypes belonged. Therefore, diversity in distribution of resistant genotypes into different clusters facilitates utilizing them in resistance breeding programme as a diverse source of material from different genetic background.
Maximum inter-cluster distance observed between cluster XXIII and cluster VI (836.06) followed by cluster XX and cluster XV (733.80) and cluster XV and cluster XX (733.80) (Fig 4), indicated that crossing between genotypes belonging to these clusters may help in production of transgressive segregants leading to opportunity for selecting better genotypes in succeeding generations
(Zaman et al., 2010; Bhakal and Lal, 2015;
Chaudhari et al., 2017; Mahesh and Hasan, 2018). Lowest inter cluster distance between cluster VIII and cluster IX (24.68) (Fig 4) suggested existence of low genetic diversity between them. Maximum intra cluster distance among the genotypes belonging to the cluster IV (95.94) followed by cluster III (78.49), cluster II (61.09) and cluster I (42.97) (Fig 4) indicated substantial variation existing among the genotypes of such clusters. Therefore, the 21 genotypes in cluster IV having highest intra-cluster distance when compared to 90 genotypes in cluster I having low intra cluster distance can be effectively used in the breeding programme to diversify the population and also may be used to produce good recombinants from the same cluster. Earlier,
Chaudhari et al., (2017) also reported more intra cluster distance in the cluster containing less number of genotypes in groundnut. Larger inter cluster distances than intra cluster distances indicated wider diversity present among the genotypes of distant groups.
The mean values of each of the 11 traits studied for all the genotypes in 25 clusters presented in Table 3 revealed that the genotype ICG 1519 belonging to solitary cluster XXIV had the highest mean 41.05 for damage due to
Spodoptera litura while the genotypes ICG 5051 and ICG 76 also belonging to solitary clusters XIV and XII had the least damage of 5.35 and 6.85, respectively. In case of late leaf spot, lowest cluster mean was observed in solitary clusters XVIII and XIX (2.5) with the genotypes ICG 12625, an accession from South America and ICG11426 an accession from South Asia, respectively. With respect to rust, lowest cluster mean was observed in solitary clusters XVII and cluster XXII (3.5). ICG 2381 (cluster IV), a land race from Brazil was found resistant to late leaf spot, rust and
Spodoptera litura (Saleem, 2018). Earlier, this genotype was reported to be resistant to rust and
Aspergilus flavus with good oil quality (
Upadhayaya et al. 2014) but low yielding (
Saleem, 2018). Therefore, diversity in distribution of resistant genotypes into different clusters facilitates utilizing those genotypes in resistance breeding programme as a diverse source of material. The susceptible genotypes JL 24 and TMV 2 have been associated with cluster III.
Highest mean number of pods per plant was found in solitary clusters XVII (ICG 87157 -28.55) followed by cluster XXIII (ICG 4538 and ICG 11855 -26.05) and lowest mean was observed in cluster X (ICGV 5286 -9.45). ICG 4538 and ICG 11855 also exhibited resistance to
Spodoptera litura. So, these genotypes can be utilised in resistance breeding to get
Spodoptera litura resistance along with desirable productivity parameters. With respect to shelling percentage, highest mean was observed in cluster XIX (74.70%) followed by cluster XXI (73.45%). However, highest mean hundred seed weight was observed in cluster XXII (ICG 13723
-55.20 g) followed by cluster XVI (ICG 513
-54.40 g). Among these, ICG 13723 showed resistance to rust and ICG 513 showed resistance to
Spodoptera litura. ICG 875 belonging to cluster XXI produced the highest mean yield per plant (30.55 g) followed by ICG 76 (25.20 g) of cluster XII. Interestingly, both these two genotypes exhibited resistance to late leaf spot and
Spodoptera litura. Hence, these genotypes may be tested under artificial epiphytotic conditions to confirm their resistance before they can be tested over locations to confirm their suitability for release as high yielding cultivars.
Among all the morphological traits studied, number of primary branches per plant (48.41%) contributed most towards divergence followed by hundred seed weight (13.13%) and reaction to
Spodoptera litura (11.33%) (Table 3). Days to fifty per cent flowering (0.03%), shelling per cent (0.59%), reaction to late leaf spot (1.3%) and rust (0.52%) contributed very less to the divergence.The present findings were in conformity with
Chaudhari et al., (2017) but in contrast to those of
Bakal and Lal (2015),
Mahesh and Hasan (2018) and
Raza et al. (2018) who opined that productivity parameters like haulm yield, shelling per cent and yield per plant contributed more to the divergence of the population while number of branches exhibited no contribution towards the divergence. In the present experiment, mini core comprised genotypes from 6 botanical varieties and each of them have different branching pattern and have different mean number of branches. Further, some of such varieties are bunch and runner types with varying number of branches which might be one of the reasons for more contribution of number of branches towards the divergence. Generally, the traits rendering maximum contribution towards divergence are given importance for selection of genotypes in breeding programme. Reaction to
Spodoptera litura significantly contributed to the diversity of the germplasm. This might have been due to different genotypes in the mini core set had different level of resistance reaction to
Spodoptera litura. Cluster means together with information on the traits that contributed maximum towards divergence would help in selection of parents for hybridization. Therefore, the genotypes ICG 76 and ICG 4412 from diverse origin having resistance to
Spodoptera litura along with higher hundred seed weight and yield per plant can be used as potential donors in the groundnut breeding programme. Multiple stress resistant (
Spodoptera litura, late leaf spot and rust) genotype ICG 2381 possessing higher number of pods and hundred seed weight but low yield per plant (9.3g) can be utilised in the backcross breeding programme as a donor to transfer the multiple biotic stress resistance to agronomically superior cultivars.