INTRODUCTION Gloriosa superba L., a climber belonging to the family Liliaceae, is a major high-value medicinal crop. It is one of the major medicinal plants in India cultivated for its seeds which are exported to developed countries for pharmaceutical use. In India, the plant is usually found in the Himalayan foot-hills, Central India, Tamil Nadu, Andhra Pradesh and Bengal. Seeds and tubers contain valuable alkaloids, viz., colchicine and colchicoside as major constituents, and are used for treating gout and rheumatism. Due to the action of colchicoside on spindle-fibre formation during cell-division, the plant has been identified as a potential anti-cancer drug. In the Indian Systems of Medicine (ISM), the tubers are used as tonic, antiperiodic, antihelmenthic and also against snake-bite (Gupta et al, 2005). Gloriosa could be found in the wild on natural fences a decade back, but now, it has been domesticated for economic gain (as all parts of the plant find diverse uses in ISM). Though G. superba enjoys extensive natural distribution and selective cultivation, the species is now endangered due to over-exploitation of its tubers and poor Variability for qualitative and quantitative traits in Glory lily (Gloriosa superba L.) R. Chitra1, K. Rajamani and M. Jawaharlal Horticultural College and Research Institute Tamil Nadu Agricultural University, Coimbatore-641 003, India E-mail: chitra.varadharaj@gmail.com ABSTRACT Glory lily (Gloriosa superba L.) is one of the major medicinal plants of India cultivated for its seeds which are exported to developed countries for pharmaceutical use. Identifying germplasm is an important component for efficient and effective management of plant genetic resources. Variability for qualitative and quantitative traits was investigated in 18 genotypes of G. superba collected from different regions of Tamil Nadu and Andhra Pradesh. For qualitative traits, these genotypes were subjected to diversity analysis based on NBPGR descriptors. Fourteen qualitative and twenty quantitative traits of G. superba were evaluated to assess morphological variations among the genotypes collected. In qualitative traits, a large number of genotypes of the 18 clustered together, at 77% similarity in two clusters. Dendrogram constructed on the basis of twenty quantitative traits for the same set of genotypes did not reveal any clear pattern in grouping, and the genotypes were grouped into seven different clusters. Cluster analysis based on qualitative and quantitative traits revealed a different group of genotypes for each of the data-set. This clearly indicated that less variation existed between genotypes with respect to morphological traits. These easily observable morphological traits are useful tools for preliminary evaluation, because, they offer a fast and reliable approach for assessing extent of diversity in G. superba genotypes. Key words: Gloriosa superba, morphological traits, cluster analysis seed-germination. Growing demand for seeds of G. superba in the international market and wide popularity of the plant among farmers makes it necessary to induce variability and develop lines with high-yield, high colchicine content, dwarf stature and leaf-blight resistance. In this medicinal plant, knowledge of genetic variability existing within the species will greatly aid exploitation of the variability directly as cultivars and, indirectly, as its use as base material in breeding programmes. MATERIAL AND METHODS Eighteen accessions, collected from different regions of the important Gloriosa growing states viz., Tamil Nadu and Andhra Pradesh, were grown in the field in a randomized block design with three replications, at the Medicinal Plants Unit, Botanical Garden, Tamil Nadu Agricultural University, Coimbatore, during 2007 - 2008 (Table 1). Planting was done in plots with three 5 m long rows, with inter-and intra-row spacing of 150 cm and 30 cm, respectively. The plots were irrigated at weekly intervals. Recommended agronomic practices and plant protection measures were followed to 1Agricultural Research Station, TNAU, Bhavanisagar, Tamil Nadu J. Hortl. Sci. Vol. 5 (1): 38-41, 2010 39 Table 1. Details of Gloriosa superba genotypes collected in 2007 Sl.No. Germplasm collected Genotype designation 1. Nallampalayam cultivated GS 01 2. Kallimanthayam cultivated GS 02 3. Sathyamangalam wild GS 03 4. Aruppukotai wild GS 04 5. Aruppukotai cultivated GS 05 6. Kankayam cultivated GS 06 7. Kallimanthayam wild GS 07 8. Ottanchadram cultivated GS 08 9. Moolanur cultivated GS 09 10. Jeyankondam cultivated GS 10 11. Udangudi cultivated GS 11 12. Viralimalai cultivated GS 12 13. Pudukottai cultivated GS 13 14. Andhra cultivated – I GS 14 15. Andhra wild GS 15 16. Z-Melur cultivated GS 16 17. Poondurai wild GS 17 18. Andhra cultivated -II GS 18 Wild : Sourced from natural habitat Cultivated : Sourced from farmers’ fields in the respective location Fig 1. Dendrogram of Gloriosa superba genotypes for qualitative traits using UPGMA based on Jaccard’s coefficient ensure a normal, healthy crop. Diversity, in terms of morphological variations among the genotypes collected, was documented as per Singh et al (2003). Morphological characters identified in the descriptor for Gloriosa superba (Saravanan and Buvaneswaran, 2003) such as habit, stem- branching, tuber shape, leaf arrangement, lamina margin, lamina colour, flower shape and biotic-stress susceptibility, were used for morphological characterization. Observations were made on five randomly-selected plants of each genotype for qualitative and quantitative traits. Qualitative traits depicting an array of characters were converted into binary characters (Sneath and Sokal, 1973) based on variations present in each trait. The presence or absence of a phenotype was given the score of 1 and 0, respectively. Quantitative data generated on various traits were standardized to zero mean and a unit variance. Sequential Agglomerative Hierarchical Non-overlapping (SAHN) clustering was performed on Squared Euclidean distance matrix using dice coefficient for quantitative and binary data, respectively, using Unweighted Pair Group Method with Arithmetic Averages (UPGMA). Analysis of data was done using NTSYSpc Version 2.02 (Rohlf, 1994). RESULTS AND DISCUSSION All the accessions collected were climbers, as reported earlier by Le Roux and Robbertse (1994). All other characters, however, showed great variation among accessions. For instance, leaf-shape variation included ovate, lanceolate and linear. Based on tuber-shape, of the 18 accessions studied, 10 had L-shape and 8 had V-shape. Leaf arrangement was mostly opposite, but alternate arrangement was also seen. Leaf lamina colour in Gloriosa was predominantly pale-green or dark-green. However, a few accessions showed dark-green lamina with pale-green streaks. As for leaf-blight susceptibility, most accessions showed no visible sign of susceptibility, while and others exhibited either low or high susceptibility (Table 2). Cluster analysis grouped 18 genotypes in two major clusters where similarity coefficients ranged from 50.5 to 100 (Fig 1). The maximum number of genotypes figured in cluster I, having 10 genotypes with a high degree of similarity (64-100%). This clearly indicated that less variation existed between genotypes with respect to morphological traits. Minimal variation between genotypes for morphological characters can be attributed to low divergence due to a greater gene flow between populations (eg., pollination, new introductions etc.) (Doulaty Baneh et al, 2007). Additionally twenty phenotypic quantitative characters were evaluated to assist diversity comparison. Per cent variation for individual traits varied from 0.05 (number of leaves per plant) to 13.55 (hundred fresh-seed weight) (Table 3). Of the twenty traits observed, hundred fresh-seed weight showed highest variation, ranging from 5.88 g (GS 17) to 12.49 g (GS 15). Pod girth, dry-seed yield per plant and dry-seed recovery also showed considerable amount of variation. A minimum of 125.26 and maximum of 621.18 numbers of leaves per plant were observed in GS J. Hortl. Sci. Vol. 5 (1): 38-41, 2010 Variability in glory lily 40 17 and GS 15, respectively. The number of pods per plant was maximum in GS 15 (46.17), and minimum in GS 17 (4.33). Dry-seed yield per plant was 4 g and 97.78 g in GS 17 and GS 15, respectively. In the present study, variation in mean performance of the genotypes may be due to interaction between environment and genotype as reported by Chandran (1987). Comparatively small difference between characters indicated that variability was primarily Table 2. Phenotype variants observed for various qualitative traits across 18 Gloriosa superba genotypes Trait Score Phenotype No. of genotypes Plant habit 1 Annual stem 18 with perennial rootstock 2 Biennial - 3 Perennial - Mode of reproduction 1 Asexual 18 2 Sexual - 3 Asexual & sexual - Plant growth habit 1 Climbing 9 2 Erect 9 Stem branching 1 Profuse 8 2 Sparse 10 Tuber shape 1 ‘V’ shape 8 2 ‘L’ shape 10 Stem pubescence 1 Glabrous 18 2 Sparse - 3 Medium - 4 Dense - Leaf arrangement 1 Alternate 4 2 Opposite 14 Lamina margin 1 Ovate 5 2 Lanceolate 10 3 Linear 3 Leaf base 1 Obtuse 18 Apex of upper leaf 1 With tendril 18 2 Without tendril - Lamina colour 1 Pale green 9 2 Dark green 6 3 Dark green 3 with pale green streaks Flower colour 1 Yellow and Red - 2 Yellow - 3 Red 18 4 White - Flower shape 1 Linear 10 2 Narrow lanceolate 8 Biotic-stress 1 Very low or no 11 susceptibility visible sign of (Leaf blight) susceptibility 3 Low 5 5 Intermediate - 7 High 2 9 Very high - due to genotypic differences, and scope for selection based on these components would be much greater in G. superba. Dendrogram construction based on twenty quantitative traits for 18 genotypes showed the genotypes to be grouped into seven different clusters (Fig 2). Cluster I had the maximum number of genotypes (five genotypes). Cluster to cluster composition revealed that clusters I, III, IV, V and VI comprised cultivated genotypes of G. superba from Tamil Nadu. Cluster II and VII included wild genotypes from Tamil Nadu. On the contrary, clusters V and VI were composed of cultivated genotypes types from Andhra and Tamil Nadu. The distribution pattern of genotypes of diverse origin in a single cluster indicates that genetic diversity observed within G. superba was not related to geographic origin. Differences noted in plant characters probably occurred over time, with free movement of plant material from location to location, and due to spontaneous induced mutations over time. Thus, for crop improvement Gloriosa plants should be selected on the basis of quantified degree of divergence as opposed to geographic origin of the genotype. These easily-observable morphological traits are a useful tool for preliminary evaluation, because, they offer a fast and reliable approach for assessing extent of diversity in G. superba genotypes. REFERENCES Chandran,V.J. 1987. Studies on the evaluation of intervariental crosses of Tomato (Lycopersicum esculentum Mill.) in F 2 and F 3 generations. M.Sc (Hort.) Thesis, Tamil Nadu Agricultural University, Coimbatore, India Doulaty Banch, H., Grassi, F., Mohammadi, A., Nazemieh, A., De Mattia, F., Imazio, S. and Labra, M. 2007. The use of AFLP and morphological markers to study Iranian grapevine germplasm to avoid genetic Fig 2. Dendrogram of Gloriosa superba genotypes for quantitative traits using UPGMA based on Squared Euclidean Distance of standardized data mean J. Hortl. Sci. Vol. 5 (1): 38-41, 2010 Chitra et al 41 erosion. J. Hortl. Sci. & Biotech., 82:745-752 Gupta, L.M., Rana, R.C., Raina R. and Meenakshi Gupta. 2005. Colchicine content in Gloriosa superba L. J. Res., (SKUAST-J), 4:238-241 Le Roux, L.G. and Robbertse, P.J. 1994. Tuber ontogeny, morphology and vegetative reproduction of Gloriosa superba L. South African J. Bot., 60:321-324 Rohlf, F.J., 1994. NTSYS-PC: Numerical taxonomy and multivariate analysis system. Version 2.2. State University of New York, Stony Brook, New York Table 3. Mean, range, coefficient of variation and standard deviation for vegetative / reproductive quantitative traits in Glory lily Traits Mean Range CV (%) SEd CD (0.05) Plant height 136.78 54.49 - 180.93 0.63 0.49 0.97 Stem girth 0.67 0.49 - 0.90 1.60 0.006 0.012 No. of leaves per plant 373.09 125.26 - 621.18 0.05 0.11 0.22 No. of branches per plant 9.68 3.33 - 18.17 4.75 0.26 0.52 Days to 50% flowering 29.02 27.00 - 32.50 1.20 0.20 0.39 No. of flowers per plant 26.81 7.33 - 57.83 2.03 0.26 0.52 No. of pods per plant 19.50 4.33 - 46.17 4.11 0.44 0.89 Pod setting percentage 69.61 57.22 - 79.80 3.87 1.51 3.00 Pod length 6.40 4.60 - 8.07 1.50 0.05 0.10 Pod girth 7.50 4.14 - 8.58 10.57 0.45 0.90 Number of seeds per pod 51.46 30.65 - 77.83 3.85 1.22 2.43 Fresh pod weight 6.44 3.70 – 8.57 1.74 0.06 0.12 Fresh seed weight per pod 5.28 2.69 - 7.84 2.64 0.07 0.15 Fresh pod yield per plant 138.86 16.47 - 400.16 3.66 2.79 5.53 Fresh seed yield per plant 117.01 11.95 - 362.34 4.39 2.80 5.54 Fresh seed recovery 80.91 71.65 - 90.41 2.94 1.31 2.60 Dry seed recovery 29.78 22.71 - 35.41 5.75 0.96 1.90 100 fresh seed weight 9.21 5.88 - 12.49 13.55 0.50 0.99 100 dry seed weight 2.65 2.14 - 3.34 0.75 0.01 0.02 Dry seed yield per plant 33.47 4.00 - 97.78 7.55 1.30 2.58 Saravanan, S. and Buvaneswaran, C. 2003. Gloriosa superba L. cultivation in Tamil Nadu: A socio- economic analysis. Adv. Pl. Sci., 16:23-28 Singh, B.M., Mahajan, R.K., Umesh Srivastava and Pareek, S.K. 2003. Minimal Descriptors of Agri-Horticultural Crops. Part IV: Medicinal and Aromatic Plants. National Bureau of Plant Genetic Resources, Pusa Campus, New Delhi, pp 59-64 Sneath, P.H.A. and Sokal, R.R. 1973. Numerical taxonomy. W.H. Freeman (ed.), San Fransisco (MS Received 13 April 2009, Revised 26 February 2010) J. Hortl. Sci. Vol. 5 (1): 38-41, 2010 Variability in glory lily