Evaluation of Olive Germplasm in Iran on the Basis of Morphological Traits: Assessment of Zard and Rowghani Cultivars S.M. Hosseini-Mazinani and S. Mohammadreza Samaee National Research Center for Genetic Engineering and Biotechnology P.O. Box 14155-6343 Tehran Iran H. Sadeghi Department of Horticulture Mazandaran University Sari Iran J.M. Caballero IFAPA, CIFA Almeda del Obispo Apdo. 3092, Córdoba Spain Keywords: inter- and intracultivar variation, Olea europaea, multivariate analysis Abstract Iran, located in the Eastern Mediterranean, shares many geographical characteristics and common historical roots with the countries of the Mediterranean basin, which are home to the major known cultivars of olive. Iran, for its part, harbors numerous olive cultivars, whose morphological and biological characters are yet to be characterized. This is important both from commercial and nutritional points of view. Building upon earlier work and in conjunction with a molecular genetics program for molecular characterization of Iranian olive cultivars, we undertook to characterize olive germplasm of Iran on the basis of morphological characters. To reach this goal, we started our research with the study of two important cultivars in Iran, Zard and Rowghani. This survey was performed over a one-year period. Samples were gathered from 281 ( Zard, 212 and Rowghani, 69) adult trees in eight regions with different ecological condition in the North of country (Roodbar, Manjil, Aliabad, Joodaki, Vakhman, Bahramabad, and Gilvan). Forty samples of flower, fruit, endocarp and leaf were collected from each tree. Multivariate statistical methods including cluster analysis and Principal Component Analysis were used for data analysis. Our preliminary results indicate a considerable variation within the traditionally recognized cultivars of olive in Iran. A more detailed assessment of morphological characters is in progress. INTRODUCTION It is believed that the olive tree originated in the Mediterranean in prehistoric times. Iran lies in the Eastern Mediterranean - a part of the world that has been cradle of ancient civilizations and has been implicated as a possible place of birth for the olive tree. The early history of olive tree in Iran has been shrouded in uncertainty but we know that olive was mentioned in ancient Iranian religious hymns of two thousand years ago. The history of olive implantation in the major olive-growing region of the country (Roodbar) has been documented for the past nine hundred years (Tabatabaie, 1995). Today, Iran, with an olive crop area of 35,000 hectares and an olive oil production of 3,000 tons per year, is one of the olive-growing countries of the world. The olive gene pool in this country constitutes a potentially important subset of the olive gene pool in the world. A scientifically sound knowledge of Iranian olive varieties forms the basis of further genetic studies of Iranian olive and contributes to the worldwide drive to identify and preserve the genetic variation in olive. Traditionally, several cultivars of olive have been recognized in Iran. These cultivars are grown together in the olive-growing regions of the country. The most important of these cultivars are Zard and Rowghani. Earlier, only limited morphological studies have been performed on Iranian olive cultivars. We attempt to investigate and further characterize these cultivars through statistical analysis of their Proc. XXVI IHC IVth Int. Symp. Taxonomy of Cultivated Plants Ed. C.G. Davidson and P. Trehane Acta Hort. 634, ISHS 2004 Publication supported by Can. Int. Dev. Agency (CIDA) 145
morphological traits. This is part of the first comprehensive investigation of the morphological characters of olive germplasm in Iran and it is conducted in conjunction with our ongoing research program on molecular characterization of olive germplasm by use of molecular markers. MATERIALS AND METHODS Plant Material The plant specimens were collected randomly from 212 Zard, and 69 Rowghani adult trees in eight regions with different ecological condition in the North of country (Roodbar, Manjil, Aliabad, Joodaki, Vakhman, Bahramabad. and Gilvan). Thirty-one morphological characters were measured on each individual (Table 1). Forty-organ samples from the South-facing sides of trees were characterized for each parameter according to the method prepared by the EU RESGEN CT 96/97 project, coordinated by the International Olive Oil Council (Cantini et al., 1999). Leaf, inflorescences (at the white bud stage), fruit (described when color change was complete), and endocarp were studied. Samples were collected from the mid-shoot portion of the current year s growth from the most representative shoots at shoulder level (approximately 1.5 m from the ground). For each tree, the average for each characterized trait was used in statistical analysis. Olive plants used in this study had been propagated vegetatively. Data Analysis Qualitative and quantitative data sets were analysed separately. Quantitative variables were standardized (mean = 0, variance = 1) for numerical analysis (Manly, 1986). Qualitative data sets were scored as the presence or absence of a characters. The Euclidean distance was used as a dissimilarity coefficient for quantitative characters. Similarity value (F) was calculated between individuals for qualitative characters using Nei and Li s (1979) and Jaccard s (1908) methods. Similarity matrices, based on different genetic distance estimators, were compared using the Mantel matrix correspondence test (Mantel, 1976). In order to quantify intra- and intercultivar relatedness, cluster analysis was performed on the total and selected characters. The methods used for these analyses were by the complete linkage method (Peeters and Martinelli, 1989) and unweighted pairgroup (UPGMA) (Benzecri, 1973). The best assessment of fit between dendrogram and original data was determined using the cophenetic correlation coefficient (Romesburg, 1984). To achieve data reduction, to clarify the relationships between two or more characters and to divided the total variance of the original characters into a limited number of uncorrelated new variables, Principal Component Analysis (PCA), (Wiley, 1981) was used. Ordination of the cultivars was performed on the first two PCA axes (Harris, 1999). Multivariate statistical analysis was performed by using NTSYS (Rohlf, 1998) and SPSS/PC (Norušis, 1999). RESULTS AND DISCUSSION High cophenetic correlation (> 0.8) obtained confirmed the efficiency of the clustering algorithms used. There was significant correlation between similarity matrices obtained using different genetic distance estimators (r = P < 0.05). We studied two important olive cultivars in Iran using morphological (quantitative and qualitative) traits and using multivariate statistical methods. This study was performed in three steps: Quantitative and qualitative characters were studied separately in cultivars of Zard and Rowghani cultivars. The complete sets of data on Rowghani cultivar is shown (Fig. 1, 2 and 3). Data on Zard cultivar is omitted in this paper due to lack of space. This survey showed that individuals from the same cultivar lie in different groups. Pattern of clustering was different with qualitative and quantitative characters (Fig. 1 and 2). However, it did not change the overall results. In other words, individuals merely moved from one group to another. Also, cluster analysis based on quantitative traits established more groups than qualitative traits, which is not surprising because 146
quantitative traits are controlled by numerous genes and are also influenced by environmental conditions and therefore exhibit more variance than qualitative characters. In order to solve these problems, especially to eliminate environmental effects, it is recommended that the quantitative characters used for grouping be chosen from data sets gathered over several years and from different places. In the second step, different cultivars from the same ecological region were inspected, but there was no coincidence between grouping patterns of individuals and their cultivars. In other words, individuals from different cultivars were placed in the same cluster (Fig. 1 and 2). In the third step, cultivars from different regions were studied. Also here, analysis based on quantitative and qualitative data sets could not place individuals in justifiable groups and did not reveal any correlation between grouping pattern and geographical distribution. In other words, individuals belonging to different regions lie in the same group. Ordination of the cultivars based on the first two PCs supported the clustering results (Fig. 3). CONCLUSION Our preliminary results indicate a considerable level of variation within the traditionally recognized olive cultivars in Iran. A more detailed assessment of morphological characters is in progress. This morphological study is being conducted in parallel with a molecular genetic program for molecular characterization of olive germplasm in Iran by use of molecular markers. The availability of molecular data along with more detailed morphological studies can help in discriminating olive cultivars of Iran. ACKNOWLEDGEMENTS We wish to thank Drs. M. Sheidai, J. Koochmeshgi and A. Mohammadi for their helpful comments on the manuscript. Literature Cited Benzecri, J.P. 1973. L analyses des donnees. Tome I. La taxonomie. Eds. Dunod, Paris. Cantini, C.A., Cimato, A. and Sani, G. 1999. Morphological evaluation of olive germplasm present in Tuscany region. Euphytica. 109:173-181. Harris, S.A. 1999. RAPDs in systematics- a useful methodology? p.211-228. In: P.M. Hollingsworth, R.M. Bateman and R.J. Gornall (eds.), Molecular Systematics and Plant Evolution, Taylor & Francis, London. Jaccard, P. 1908. Nouvelles recherches sur la distribution floral. Bull. Soc. Vaud. Sci. Nat. 44:223-270. Manly, B.F.J. 1986. Multivariate statistical methods: a primer. Chapman and Hall, London. Mantel, N. 1976. A personal perspective on statistical techniques for quasi-experiments. p.103-129. In: D.B. Owen (ed.), On the History of Statistics and Probability, Marcel Dekker, New York. Nei, M. and Li, W. 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. (USA) 76:5269-5273. Norušis, M.J. 1999. SPSS/PC advanced statistics. Ver.10.0. SPSS Inc., Chicago. Peeters, J.P and Martinelli, J.A. 1989. Hierarchical cluster analysis as a tool to manage variation in germplasm collection. Theor. Appl. Genet. 78:42-48. Rohlf, F.J. 1998. NTSYS-pc (Numerical Taxonomy and Multivariate Analysis System). Ver. 2. Exeter Software, New York. Romesburg, H.C. 1984. Cluster analysis for researchers. Lifetime Learning Publications, Belmont, California. Tabatabaie, M. 1995. Olive and its oil (in Persian), Ministry of Agriculture, Tehran. Wiley, E.O. 1981. Phylogenetics: The theory and practice of phylogenetics and systematics. John Wiley, New York. 147
Tables Table 1. List of morphological characters used for the multivariate analysis of olive germplasm in Iran. Characters A) Leaf characters: 1) Blade length (mm) 2) Blade width (mm) 3) Blade length/width 4) Longitudinal curvature of blade B) Inflorescence characters: 1) Inflorescence length 2) Number of flowers C) Fruit characters: 1) Fruit weight (g) 2) Fruit length (mm) 3) Fruit width (mm) 4) Fruit length/width 5) Symmetry in position A 6) Situation of maximum transverse diameter in position B 7) Apex in position A 8) Base in position A 9) Nipple 10) Presence of lenticels 11) Size of lenticels 12) Colour at full maturity D) Endocarp (stone) characters: 1) Stone weight (g) 2) Stone length (mm) 3) Stone width (mm) 4) Stone length/width 5) Symmetry in position A 6) Symmetry in position B 7) Situation of maximum transverse diameter in position B 8) Apex in position A 9) Base in position A 10) Surface in position B 11) Number of grooves 12) Distribution of grooves 13) Termination of the apex in position A 148
Figurese ROWGHANI1 ROWGHANI28 ROWGHANI2 ROWGHANI33 ROWGHANI37 ROWGHANI5 ROWGHANI31 ROWGHANI32 ROWGHANI3 ROWGHANI27 ROWGHANI4 ROWGHANI24 ROWGHANI30 ROWGHANI35 ROWGHANI36 ROWGHANI25 ROWGHANI17 ROWGHANI34 ROWGHANI21 ROWGHANI23 ROWGHANI29 ROWGHANI19 ROWGHANI6 ROWGHANI9 ROWGHANI13 ROWGHANI16 ROWGHANI7 ROWGHANI15 ROWGHANI8 ROWGHANI10 ROWGHANI11 ROWGHANI14 ROWGHANI12 ROWGHANI18 ROWGHANI26 ROWGHANI20 ROWGHANI22 0.64 2.30 3.96 5.62 7.28 Fig. 1. Dendrogram generated by clustering UPGMA analysis based on Euclidean distance computed from pairwise comparisons of quantitative characters between 37 individuals of Rowghani cultivar. 149
0.51 0.63 0.75 0.88 1.00 ROWGHANI1 ROWGHANI2 ROWGHANI5 ROWGHANI27 ROWGHANI19 ROWGHANI6 ROWGHANI9 ROWGHANI12 ROWGHANI14 ROWGHANI15 ROWGHANI16 ROWGHANI18 ROWGHANI20 ROWGHANI26 ROWGHANI28 ROWGHANI30 ROWGHANI32 ROWGHANI33 ROWGHANI36 ROWGHANI35 ROWGHANI4 ROWGHANI29 ROWGHANI3 ROWGHANI24 ROWGHANI7 ROWGHANI17 ROWGHANI8 ROWGHANI22 ROWGHANI23 ROWGHANI25 ROWGHANI13 ROWGHANI21 ROWGHANI34 ROWGHANI37 ROWGHANI10 ROWGHANI31 ROWGHANI11 Fig. 2. Dendrogram generated by clustering UPGMA analysis of 1-F values (based on Jaccard s index) computed from pairwise comparisons of qualitative characters between 37 individuals of Rowhgani cultivar. 150
PC1 0.94 0.84 R8 R22 R23 R25 0.75 R21 R34 R37 R24 R7 R17 R13 R3 R4 R29 R1 R5 R27 R20 R26 R28 R30 R32 R33 R36 R19 R35 R6 R9 R12 R14 R15 R16 R18 0.66 R31 R0 0.57 R11-0.56-0.34-0.13 0.09 0.31 PC2 (a) PC1 2 1 0-1 R31 R6 R2 R28R33 R32 R1 R12 R14 R10 R8 R16 R18 R7 R11 R15 R9 R6 R13 R3 R17R34 R26 R25 R24 R30 R4 R35 R37 R27 R36 R21 R19 R23 R29-2 R22-3 -3-2 -1 R20 0 1 2 PC2 (b) Fig. 3. Principal Component Analysis of the pairwise difference matrix of qualitative (a) and quantitative (b) for 37 individuals of Rowghani cultivar. 151