Assessment of morphological diversity in brinjal (Solanum melongena L.) genotypes using qualitative descriptors and the Shannon-Wiener diversity index

Introduction

Brinjal (Solanum melongena L.) is one of the most widely grown and economically important vegetable crop of the family Solanaceae, with a diploid chromosome number of 2n = 2x = 24. It is cultivated extensively across the tropical and sub-tropical regions of Asia and Africa and is often described as the 'King of vegetables' on account of its versatility and usage in different culinary purposes (Choudhary et al., 2009) [1] . India, together with the wider Indo-Burma region, is known as the primary centre of origin and diversity of the crop (Kousalya et al., 2024) [2] and this rich native variability in terms of fruit size, shape, color, growth habit, yield, pest and disease resistance, climate adaptability and quality attributes has historically formed the foundation for varietal improvement through selection and hybridization Brinjal is valued not only as an inexpensive source of dietary fibre, minerals and phenolic antioxidants, but also for its wide-ranging culinary uses and its role in the livelihoods of resource-poor farmers who grow numerous local landraces alongside a limited number of improved cultivars and hybrids. However, the increasing popularity of a small set of high-yielding hybrids has progressively narrowed the genetic base under commercial cultivation, while a very large reservoir of morphologically distinct landraces, traditional types and locally adapted genotypes continues to be maintained on-farm or in institutional germplasm collections without adequate characterization. Systematic documentation of this variability is therefore a prerequisite for its conservation and effective utilization in breeding programmes. Morphological characterization using standard, internationally accepted descriptors is the first and most accessible step in germplasm evaluation, since qualitative traits of plant growth habit, leaf, flower and fruit are largely under simple genetic control and are unaffected by most environmental fluctuations. Frequency distribution of genotypes across the classes of a qualitative descriptor can be summarized using diversity indices originally developed in information theory and later adapted for the analysis of biological and genetic variation. Among these, the Shannon-Wiener (also referred to as the Shannon-Weaver) diversity index has become one of the most widely used tools for quantifying phenotypic diversity in plant genetic resources, since it simultaneously reflects both the number of classes present (richness) and the evenness with which genotypes are distributed among them (Shannon et al., 1949) [3] . A higher value of H′ for a given trait indicates a more even spread of genotypes across a greater number of descriptor classes, whereas a low value points to the dominance of one or a few classes and hence limited exploitable variability for that character. Since the index compresses a full frequency distribution into a single number (Zeven et al., 1991) [4] , a recommendation has been followed in the present study by reporting the complete descriptor-class distribution together with the H′ values. The Shannon-Wiener index has been successfully applied to describe morphological diversity in brinjal germplasm collections of varying size and origin, including large national gene bank holdings profiled with the help of DIVA-GIS (Gangopadhyay et al., 2013) [5] and regional landrace collections from Eastern India (Tripathy et al., 2025) [6] , providing a useful basis for comparison with the present set of genotypes. Although several studies have documented quantitative genetic divergence in brinjal using Mahalanobis' D 2 statistics and principal component analysis (Anbarasi et al., 2021; Lintu et al., 2023; Zabbar et al., 2024) [7,8,9] , comparatively fewer reports combine a complete set of qualitative descriptor traits with Shannon-Wiener diversity estimation for germplasm maintained under the coastal, deltaic agro-climatic conditions of Puducherry. In view of this, the present study was undertaken with the specific objective of characterizing 45 brinjal genotypes based on twenty qualitative morphological descriptors of plant, leaf, flower and fruit characters, and of quantifying the extent and pattern of diversity for each trait using the Shannon-Wiener diversity index, to identify the most polymorphic characters and to generate baseline information useful for parental selection in future brinjal breeding programmes.

Materials and Methods

The present investigation was carried out at the Department of Horticulture, Pandit Jawaharlal Nehru College of Agriculture and Research Institute, Karaikal, Union Territory of Puducherry, India (10.9522° N, 79.7825° E), which lies in the coastal, deltaic tract of the Cauvery river basin characterized by a warm, humid tropical climate. Forty-five brinjal genotypes, comprising released varieties and local/regional collections, were raised in the field following the recommended agronomic package of practices, with observations recorded from five randomly tagged, representative plants of each genotype at the appropriate vegetative, flowering and fruiting stages. Morphological descriptors covering vegetative, floral and fruit characters were recorded for each genotype by following the descriptor states standardized for brinjal germplasm characterization by IBPGR Descriptors for eggplant (1990) [10] , except two traits (Calyx color and fruit shape), which were recorded as per Tripathy et al. (2025) [6] . Each trait was scored on the internationally used 0-9 (or 1-9) class scale appropriate to that descriptor.

The frequency of genotypes falling in each descriptor class was computed for every trait and the percentage distribution was worked out. Phenotypic diversity for each was quantified using the Shannon-Wiener diversity index (H′), calculated from the class-frequency distribution as follows (Shannon et al., 1949) [3] :

Where, P i is the proportion of the total 45 genotypes falling in the i th descriptor class of a given trait, and s is the number of descriptor classes actually represented (i.e., classes with a non-zero frequency). The maximum possible value of the index for a trait, H max , is given by ln(s) and the normalized Shannon diversity index, H′n = H′/H max , was additionally computed to allow direct comparison of relative diversity across traits differing in their number of descriptor classes; H′n ranges from 0 (complete uniformity) to 1 (genotypes distributed with perfect evenness across all observed classes) (Zeven et al., 1991) [4] . Agglomerative hierarchical cluster analysis was performed based on the Gower dissimilarity matrix using Ward's minimum-variance method (Ward, 1963; implemented as ward.D 2 , which operates directly on the dissimilarity matrix rather than on squared Euclidean distances). Basic statistical analysis, such as frequency distribution, was carried out in Microsoft Excel and the Shannon-Wiener diversity index and cluster analysis were carried out using R software version 4.4.2.

Results and Discussion

The 45 brinjal genotypes evaluated in the present study displayed a wide range of phenotypic variation for the majority of the 20 qualitative descriptors recorded, confirming the presence of substantial exploitable morphological diversity in the germplasm. The complete frequency distribution of genotypes across descriptor classes, together with the trait-wise Shannon-Wiener diversity index (H′), its theoretical maximum (H max ) and the normalized index (H′n), is presented in Table 1.

Vegetative and floral traits

Among the vegetative descriptors, plant growth habit showed a predominance of the intermediate class (53.33% of genotypes), followed by the upright (37.78%) and prostrate (8.89%) classes, giving a moderately high H′ of 0.918 (H′n = 0.836), indicating that although one class was numerically dominant, all three growth-habit classes were reasonably well represented. Plant height was overwhelmingly skewed towards the tall class (64.44%), with the remainder intermediate (35.56%) and no genotypes in the very short, short or very tall classes; the resulting H′ of 0.651 was nevertheless associated with the highest normalized diversity recorded among all vegetative traits (H′n = 0.939), because only two classes were actually present and the genotypes were divided between them in a near-even 64:36 ratio.

Fruit and seed traits

Fruit-related descriptors consistently returned the highest Shannon diversity values in the present study. Fruit colour at commercial ripeness was the single most polymorphic trait overall (H′ = 1.606, six of the nine possible descriptor classes represented, H′n = 0.896), with genotypes distributed among green (26.67%), milky white (11.11%), lilac grey (8.89%), purple (33.33%), purple-black (15.56%) and black (4.44%) fruited types, and none scored as deep yellow, fire red or scarlet red. Fruit length-to-breadth ratio was the second most diverse character (H′ = 1.398, five classes, H′n = 0.869), reflecting the co-existence of round/as-long-as-broad, slightly longer, and markedly elongated fruit types in the genotype set, while fruit shape (H′ = 1.325, four classes: long 24.44%, round 24.44%, oblong 37.78%, oval 13.33%; H′n = 0.956) recorded the highest normalized diversity of any trait studied, indicating an almost even spread of genotypes across all four observed fruit-shape classes. Seed colour (H′ = 1.016, H′n = 0.925) and fruit cross-section (H′ = 0.958, H′n = 0.872) also showed relatively high diversity, whereas fruit colour distribution (H′ = 0.556, uniform fruits predominating in 75.56% of genotypes) and, most notably, fruit curvature (H′ = 0.393, H′n = 0.567, with 86.67% of genotypes bearing straight fruits and only 13.33% slightly curved) were the least diverse of the fruit descriptors. Fruit apex shape, with genotypes distributed among protruded (11.11%), rounded (33.33%) and depressed (55.56%) classes, recorded intermediate diversity (H′ = 0.937, H′n = 0.853).

Hierarchical cluster analysis (Ward's method on Gower's dissimilarity)

Agglomerative hierarchical clustering of the 45 genotypes on their Gower dissimilarity matrix, using Ward's minimum-variance method, produced a dendrogram (Fig 4) that grouped the genotypes into four major clusters at a dissimilarity height of approximately 0.9-1.0, well above the height of most of the internal branching within each cluster.

Cluster III was the largest (17 genotypes), followed by Cluster II (13 genotypes), Cluster IV (9 genotypes) and Cluster I (6 genotypes; Table 2). Cross-tabulation of cluster membership against the original descriptor states (Table 3) showed that the four clusters were morphologically interpretable rather than arbitrary statistical groupings. Cluster I was uniformly tall (100%), predominantly upright in growth habit (83.3%) with a white corolla (83.3%) and was the only cluster fixed for green fruit colour at commercial ripeness (100%) combined with a markedly elongated fruit (several times as long as broad in 66.7% of genotypes) and slightly curved fruit shape (83.3%) -a combination consistent with the long, green-fruited genotypes (e.g., Brinjal Green

Overall pattern of diversity and comparison with earlier reports

Averaged across all 20 descriptors, the mean H′ was 0.884 and the mean normalized diversity index (H′n) was 0.789, indicating that, on the whole, the 45 genotypes evaluated were distributed evenly across the descriptor classes actually expressed in this germplasm set, with diversity being concentrated disproportionately in fruit-related traits. This pattern -higher Shannon diversity for fruit colour, fruit shape and fruit length-to-breadth ratio relative to vegetative and floral traits -is consistent with several earlier reports on brinjal. Working with 48 landraces from Odisha, Tripathy et al. (2025) [6] reported an overall mean Shannon-Wiener index of 0.733 with the greatest variation ob-served for fruit length-to-breadth ratio, leaf blade colour, fruit shape and petiole colour, while a large-scale characterization of 622 indigenous and exotic NBPGR accessions using DIVA-GIS reported Shannon diversity values ranging from as low as 0.01 for fruit position and 0.20 for fruit curvature up to 1.51 for fruit colour, with fruit shape (1.30) and fruit length-to-breadth ratio (1.40) again ranking among the most diverse traits (Gangopadhyay et al., 2013) [5] . The concordance between the ranking of highly diverse traits in the present study and in these two independent germplasm sets -despite differences in genotype origin, sample size and geography -suggests that fruit colour, fruit shape and fruit elongation are traits that have been subjected to comparatively less selection pressure, or have diversified more readily under farmer and breeder selection for market preference, than characters such as fruit curvature and calyx colour, which remain comparatively conserved across brinjal germplasm. A morphological survey specifically conducted on brinjal germplasm from the coastal Karaikal region reported similarly wide variation for fruit and plant traits among locally available genotypes (Bharathi et al., 2024) [11] , supporting the observation that the coastal, deltaic tract in which the present genotypes were evaluated genuinely diverse brinjal material rather than a narrow, closely related set of types. Likewise, multivariate divergence studies employing Mahalanobis' D 2 statistics on other brinjal collections have consistently identified fruit size, fruit shape and fruit colour as major contributors to overall genetic divergence (Anbarasi et [7,12,9] , and divergence analysis including wild and semi-wild relatives has shown that fruit and flowering characters differentiate cultivated types most sharply (Lintu et al., 2023) [8] , further corroborating the trait-wise pattern of diversity recorded here.

The comparatively low diversity recorded for calyx colour, fruit curvature, leaf blade length and leaf blade tip angle indicates that these characters are close to fixation in the present genotype set, with one descriptor class accounting for more than 70-85% of genotypes in each case; such traits are of limited direct value for distinguishing genotypes or for use as morphological markers in this particular collection, although they may still be diagnostic when this material is compared against germplasm from other agro-ecological regions. In contrast, the pronounced diversity for fruit colour, fruit shape, fruit length-to-breadth ratio and corolla colour identifies these as the descriptors of greatest immediate value for distinguishing among the 45 genotypes and, more importantly, for guiding the choice of morphologically divergent parents in future hybridization programmes aimed at combining desirable fruit-quality attributes, since genetic divergence for fruit traits translates directly into scope for heterosis and transgressive segregation for yield-and quality-related characters in the F 1 and subsequent generations (Anbarasi et al., 2021; Tripathy et al., 2025) [7,6] .

Conclusion

The present study demonstrates that considerable and unevenly distributed morphological diversity exists among the 45 brinjal genotypes evaluated at PAJANCOA & RI, Karaikal, with the Shannon-Wiener diversity index ranging from 0.393 for fruit curvature to 1.606 for fruit colour at commercial ripeness across the 20 qualitative descriptors examined (mean H′ = 0.884; mean H′n = 0.789). Fruit-related traits -fruit colour, fruit shape, fruit length-to-breadth ratio and fruit cross-section -together with corolla colour, contributed disproportionately to the overall diversity recorded, while calyx colour, fruit curvature and certain leaf descriptors were comparatively uniform across the genotype set. Ward's hierarchical clustering of Gower's dissimilarity matrix grouped the 45 brinjal genotypes into four morphologically divergent clusters distinguished mainly by fruit colour, fruit shape and plant stature. Cluster analysis separated a distinctive group of tall, green, long-fruited genotypes (Cluster I) from the remaining, more variable purple-and oblong-fruited genotypes. These findings confirm that the germplasm assembled for this study, comprising released varieties, landraces and breeding lines, constitutes a genetically diverse working collection that can be profitably exploited for the selection of morphologically divergent parents in brinjal hybridization programmes, for the construction of representative core or working collections, and for descriptor-based documentation and varietal identification. Fruit colour and fruit shape-related descriptors, in particular, merit priority attention in future characterization and breeding efforts targeting fruit-quality improvement in brinjal and crosses between Cluster I and Cluster II genotypes, in particular, they are recommended as a priority for future hybridization work aimed at maximizing heterosis and transgressive segregation for fruit traits.

References

1. Choudhary B, Gaur K. ISAAA Brief No. 2009;38:1-102.
2. Kousalya R, Praneetha S, Vethamoni PI, Ravichandran V, Iyanar KNS. Unveiling the genetic variability, character association and principal component analysis for yield and yield contributing traits in Brinjal (Solanum melongena L.) genotypes. Indian J Agric Sci. 2024;94(10):1039-1044.
3. Shannon CE, Weaver W. The mathematical theory of communication. University of Illinois Press; 1949.
4. Zeven AC, Hennink S. The interpretation of Nei and ShannonWeaver within population variation indices. Euphytica. 1991;51(3):235-240.
5. Gangopadhyay KK, Pandey S, Tiwari SK, Panwar NS, Meena BL, Dutta M. Diversity analysis in eggplant germplasm in India using DIVA-GIS approach. Indian J Hortic. 2013;70(4):519-525.
6. Tripathy B, Tripathy P, Jyothsna J, Sahu GS, Dash SK, Badu M, Mahapatra SK, Rout GR. Exploring heterosis, dominance effect, and genetic control in brinjal (Solanum melongena L.) landraces. Sci Rep. 2025;15(1):23032.
7. Anbarasi D, Haripriya K. Genetic divergence in brinjal genotypes for growth and yield parameters. Electron. J. Plant Breed. 2021;12(4):1408-1412.
8. Lintu P, Namboodiri RV. Genetic diversity of brinjal (Solanum melongena L.) and its wild relatives. Electron. J. Plant Breed. 2023;14(1):137-147.
9. Zabbar MA, Islam MS, Rasel M, Rahman MM, Robin AHK. Morphological characterization and genetic diversity analysis of yield and yield contributing parameters in brinjal (Solanum melongena L.) genotypes. SAARC J Agric. 2024;22(1):59-71.
10. IBPGR. Descriptor for Eggplant/Descripteurs pour I'Aubergine. International Board for Plant Resources, Rome; 1990.
11. Bharathi KI, Shanthi A, Manikandan M. Assessment of genetic diversity of brinjal (Solanum melongena L.) in the coastal region of Karaikal, India. J Sci Res Rep. 2024;30(6):407-411.
12. Younas Z, Naseer S, Kazmi A, Ali A, Wahab A, Sultana T, Shoukat I, Hameed A, Afzal M, Mashwani ZR, Rahimi M. Assessment of diversity among important brinjal (Solanum melongena) cultivars using morphological markers. J Food Qual. 2022;2022(1):4255554.