OPCE-STR.vp Acta Bot. Croat. 69 (2), 237–247, 2010 CODEN: ABCRA 25 ISSN 0365–0588 A morphometric study on Scorzonera L. taxa (Asteraceae) from northeast Anatolia SERDAR MAKBUL1*, ZAFER TURKMEN2, KAMIL COSKUNCELEBI3, OSMAN BEYAZOGLU3 1 Rize University, Faculty of Arts and Sciences, Department of Biology, 53100 Rize, Turkey 2 Giresun University, Faculty of Arts and Sciences, Department of Biology, Giresun, Turkey 3 Karadeniz Technical University, Faculty of Arts and Sciences, Department of Biology, Trabzon, Turkey Phenetic traits of 39 populations belonging to 19 taxa of Scorzonera L. (Asteraceae) from north Anatolia were analyzed with the use of numerical methods. Principal component analysis (PCA) showed that pubescence and length of achenes, the shape of outer phyllaries, and average length of flowering capitula, plant pubescence, root shape and state of the plant stem are the best variables to distinguish the examined taxa. In addition, it was also found that binary are more important than quantitative characters in discrimi- nating the examined Scorzonera taxa. Numerical results based on 25 morphological char- acters were discussed and compared with traditional taxonomic treatments. Key words: Phenetics, Scorzonera, systematics, Turkey Introduction The genus Scorzonera L. (Asteraceae) numbers about 160 species, ancient Mediterra- nean by origin, belonging to the subtribe Scorzonerinae Dumort. (Lactuceae Cass., Cicho- rioideae) is widely spread in the arid regions of Eurasia and Africa (NAZAROVA 1997, BRE- MER and ANDERBERG 1994). The genus includes up to 47 species in Turkey, some of which are endemic (DURAN 2008). This genus appears simple at first sight. However, it is a taxo- nomically difficult genus with several complexes of closely related species (CHAMBERLAIN 1975). The difficulties of identifying Scorzonera derive from the genus not having been sufficiently well investigated by taxonomists, many questions thus remaining controversial (NAZAROVA 1997). In recent years Scorzonera has been the subject of caryological (NAZA- ROVA 1997, GUARDIA and BLANCA 1987), ethnobotanical (RIVERA et al. 2006, ERTUÐ 2000), chemical (ZIDORN et al. 2003, MAGIATIS et al. 2001) and phenetic (MAVRODIEV et al. 2004) studies that have improved our understanding of the systematics of this genus. ACTA BOT. CROAT. 69 (2), 2010 237 * Corresponding author, e-mail: smakbul@hotmail.com U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:38 Color profile: Disabled Composite 150 lpi at 45 degrees Nevertheless, there are still systematic problems. To date, no comprehensive numerical taxonomic study has been conducted on Scorzonera species, which are often difficult to distinguish from each other and whose infrageneric classification is only preliminary. The numerical approach is indeed a powerful tool in resolving the relationships and taxonomy of closely related species with complex variation patterns. The purpose of this study is to evaluate the extent of the variations in morphological characters and determine the taxo- nomic value of phenetic traits in classifying Scorzonera taxa as represented by 39 popula- tions distributed in NE Anatolia. Materials and methods Thirty-nine populations of nineteen Scorzonera taxa were collected from NE Anatolia and used as operational taxonomic units (OTUs) in this study. Locality information and a distribution map of the OTUs are given in table 1 and figure 1. Vouchers are deposited in the Herbarium of Karadeniz Technical University, Department of Biology (KTUB). Twenty-five morphological characters were assessed: 11 related to vegetative structure, and 14 to floral structures (Tab. 2). Trait selection was based on floras and our own obser- vations. A total of 234 individuals with flowers were scored for all these characters and the averages of all the individuals from each species were combined to yield the basic data ma- trix. The raw data matrix of 25 variables as columns and 39 objects (populations) as rows were used for numerical analysis (Tab. 3). Identifications were made according to Flora of Turkey (CHAMBERLAIN 1975), Flora of Europe (CHATER 1976), Flora Iranica (RECHINGER 1977) and Flora of USSR (LIPSCHIZ 1964) and assigned to nineteen taxa as follows: Scorzonera cana var. jacquiniana , S. cana 238 ACTA BOT. CROAT. 69 (2), 2010 MAKBUL S., TURKMEN Z., COSKUNCELEBI K., BEYAZOGLU O. Fig. 1. Distribution map of the investigated taxa in NE Anatolia. Population numbers are specified in table 1. U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:40 Color profile: Disabled Composite 150 lpi at 45 degrees ACTA BOT. CROAT. 69 (2), 2010 239 MORPHOMETRIC STUDY ON SCORZONERA Tab. 1. Collection data of the examined specimens. No Taxa Locality Altitude (m) Collection Numbers 1 Scorzonera cana (C.A. Mey.) Hoffm. var. jacquiniana (W. Koch) Chamb. Trabzon: Araklý-Daðbaºý 1850 Makbul 054 2 Rize: Ovit 2400 Makbul 028 3 Giresun: Alucra 1490 Makbul 078 4 S. cana (C.A. Meyer) Hoffm. var. cana Rize: Cimil 2300 Makbul 057 5 Rize: Ovit 2400 Makbul 030 6 Giresun: Eðribel Geçidi 2350 Makbul 095 7 Erzurum: Ýspir-Moryayla 2450 Makbul 093 8 S. cana (C.A. Meyer) Hoffm. var. alpina (Boiss.) Chamb. Rize: Ovit 2400 Makbul 029 9 Artvin: Ardanuç 640 Makbul 076 10 S. eriophora DC. Gümüºhane: Tersun Daðý 2040 Makbul 050 11 Gümüºhne: Mogoldas Daðý 1650 Makbul 044 12 S. armeniaca (Boiss. et Huet.) Boiss. Bayburt: Bayburt Kalesi 1650 Makbul 059 13 Erzurum: Aºkale 2000 Makbul 048 14 Artvin: Yusufeli-Ýspir road 40. km 815 Makbul 073 15 S. parviflora Jacq. Erzurum: Aºkale 1630 Makbul 088 16 S. mollis Bieb. ssp. mollis Gümüºhane: Tersun Daðý 2040 Makbul 051 17 Giresun: Findikbeli Geçidi 1730 Makbul 080 18 S. insica DC. Bayburt: Kop Daðý 2150 Makbul 085 19 S. laciniata L. ssp. laciniata Bayburt: Kop Daðý 2100 Makbul 045 20 Bayburt: Çerçi Köyü 1700 Makbul 070 21 Artvin: Yusufeli-Yokuºlu Köyü 815 Makbul 074 22 S. inaequiscapa Boiss. Bayburt 1500 Makbul 068 23 Giresun: Alucra-ªiran, 15. km 1670 Makbul 079 24 S. sericea DC. Bayburt: Kop Daðý 2450 Makbul 089 25 S. tomentosa L. Bayburt: Kop Daðý 2160 Makbul 010 26 Giresun: Alucra 1400 Makbul 012 27 S. mollis Bieb. ssp. szowitzii (DC) Chamb. Gümüºhane: Kelkit 1635 Makbul 039 28 Gümüºhane: Tersun Daðý 2000 Makbul 064 29 Erzurum. Aºkale 1640 Makbul 090 30 S. sosnowskyi Lipschitz. Erzurum: Ýspir-Moryayla 2400 Makbul 091 31 Bayburt: Kop Daðý 2150 Makbul 086 32 S. suberosa C. Koch Erzincan: Erzincan-Kelkit 20. km 1750 Makbul 041 33 Bayburt: Çerçi Köyü 1700 Makbul 069 34 S. cinerea Boiss. Bayburt: Kop Daðý 2150 Makbul 087 35 S. pseudolanata Grossh. Bayburt 1500 Makbul 072 36 Bayburt: Köse 1650 Makbul 040 37 S. seidlitzii Boiss. Artvin: ªavºat, Sahara Mezrasý 2150 Makbul 022 38 Artvin: Yalnizçam Daðlarý 2200 Makbul 025 39 S. latifolia (Fish. et Mey.) DC. Bayburt: Kop Daðý 2160 Makbul 094 U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:40 Color profile: Disabled Composite 150 lpi at 45 degrees var. cana, S. cana var. alpina, S. eriophora, S. armeniaca, S. parviflora, S. mollis ssp. mollis, S. insica, S. laciniata ssp. laciniata, S. inaequiscapa, S. sericea, S. tomentosa, S. mollis ssp. szowitzii, S. sosnowskyi, S. suberosa, S. cinerea, S. pseudolanata, S. latifolia, S. seidlitzii. Two multivariate analyses were performed using SYN-TAX PC 5.0 (PODANI, 1993): cluster analysis (CA) and principal components analysis (PCA). For the CA, a pair-wise matrix of resemblance values was calculated from the standardized data matrix, using Gower’s coefficient as the coefficient of resemblance that is designed for mixed data sets (SNEATH and SOKAL 1973). A dendrogram was generated by the unweighted pair-group method using arithmetic averages (UPGMA). For PCA, the standardized data were used to create a covariance matrix, and three eigenvectors were extracted. 240 ACTA BOT. CROAT. 69 (2), 2010 MAKBUL S., TURKMEN Z., COSKUNCELEBI K., BEYAZOGLU O. Tab. 2. List of characters used in this study. Symbol Characters X1 Total plant high (cm) X2 State of the plant stem; 0: scape, 1: caulescent X3 Root; 0: cylindrical, 1: tuberose X4 Residue leaves at the base of the stem; 0: absent, 1: present X5 Stem; 0: not branched, 1: branched X6 Leaf; 0: entire, 1: pinnatisect or pinnatifit X7 Width of leaf (mm) X8 Length of leaf (mm) X9 Lamina; 0: linear, lanceolat, 1: ovat, eliptic X10 Leaf margin; 0: smooth, 1: undulate X11 Plant; 0: densely hairy (tomentose, pallose, villous), 1: slightly hairy (pubescent and sericea) X12 Number of capitula X13 Average length of flowering capitula (mm) X14 Average length of outer phyllaries (mm) X15 Average length of inner phyllaries (mm) X16 Shape of outer phyllaries; 0: linear, lanceolat, 1: ovat, eliptic X17 Ligulae color; 0: yellow, 1: purple X18 Length of achenes (mm) X19 Pappus color; 0: white or cream, black or brownish X20 Pappus; 0: completely plumose or barbellat, 1: plumose together with barbellat X21 Lower surface of inner phyllaries; 0: hairless, 1: hairy X22 Average row number of phyllaries X23 Apex of outer phyllaries; 0: unforked, 1: forked X24 Surface of achenes; 0: smooth, 1: rough X25 Achenes; 0: glabrous, 1: hairy U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:40 Color profile: Disabled Composite 150 lpi at 45 degrees Results Dendrogram of 25 phenetic variables corresponds to 19 taxa (Fig. 2). All taxa fall into two main clusters. The first group is labeled »a« and consists of 30 populations belonging to 15 taxa. The second cluster is labeled »b« and consists of the rest of the examined popu- lations belonging to 4 taxa (Scorzonera mollis subsp. mollis, S. inaequiscapa, S. mollis subsp. szowitzii, S. suberosa). The cluster »a« is divided into small subclusters labeled »c« and »d«. PCA results using 25 characters are given in figures 3 and 4, showing the distribution of OTUs and variables on the first two components (axis). The results of principal component analysis show eigenvalues as percentages of explained variance, and cumulative percent- ages (Tab. 4). PC-1 and PC-2 had the highest score for the eighteen of the twenty five traits (Tab. 4). These variables are X3, X10, X13, X14, X15, X16, X17, X18, X25 for PC-1 and X4, X6, X9, X11, X12, X19, X20, X23, X24 for PC-2. PC-3 received the highest score only for the remaining seven variables used in this study. Eleven of the twenty-five traits ana- lyzed in the present study explained more than 50% of the total variation. The highest vari- ation rates observed among the twenty-five characters and accounting for the first three ACTA BOT. CROAT. 69 (2), 2010 241 MORPHOMETRIC STUDY ON SCORZONERA Fig. 2. Cluster analysis – UPGMA. The clusters discussed in the text are marked with letters. Popu- lation numbers are given in table 1. U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:41 Color profile: Disabled Composite 150 lpi at 45 degrees 242 A C T A B O T .C R O A T .69 (2),2010 M A K B U L S .,T U R K M E N Z .,C O S K U N C E L E B IK .,B E Y A Z O G L U O . Tab. 3. Raw Data Matrix using Numerical Analysis (Raw: Character, Column: Taxa). 25.83 0 0 1 0 1 3.13 10.51 0 0 1 3.33 21.60 7.50 20.16 0 1 9.33 0 0 1 1 3 1 0 25 0 0 1 1 1 2.98 12.16 0 0 1 5.16 22.50 8 22.16 0 1 10 0 0 1 1 3 1 0 24.83 0 0 1 0 1 2.42 13.33 0 0 1 1.33 24.66 6 17 0 1 9.66 0 0 1 1 3 0 0 20.83 0 0 1 0 1 1.88 8.83 0 0 1 3.16 22.16 6.83 19.16 0 1 8.50 0 0 1 0 3 1 0 11.33 0 0 1 0 1 1.80 11.33 0 0 1 2.83 22.50 5.58 16.66 0 1 9.66 0 0 1 1 3 1 0 18.66 0 0 1 0 1 2.61 12.50 0 0 1 3.83 22.16 8 23.83 0 1 11.60 0 0 1 1 3 1 0 18.66 0 0 1 1 1 2.15 10 0 0 1 6 19.16 6 18.16 0 1 9.50 0 0 1 1 3 0 0 6 0 0 1 0 1 1.10 7.25 0 0 1 1.66 16.25 5.83 14.16 0 1 8.16 0 0 1 1 3 1 0 9.33 0 0 1 0 1 1.80 7 0 0 1 3.50 13.83 6.50 18 0 1 9 0 0 1 1 3 1 0 24 1 0 0 0 0 1.33 16.60 0 0 0 3.83 22.66 13 24 0 1 11 1 1 1 1 2 0 0 33.16 1 0 0 1 0 1.85 15.33 0 0 0 2.50 26.50 12.50 31.50 0 1 6.33 1 1 1 1 2 0 0 20 1 0 1 1 1 2.56 15.66 0 0 1 5.16 15.83 4.83 16.83 0 1 8.66 1 0 1 1 3 1 0 34.16 1 0 1 1 1 3.18 13.83 0 0 1 3.16 15.66 5 21.83 0 1 10.33 1 0 0 1 3 1 0 31.16 1 1 1 1 1 2.60 13.83 0 0 1 8 18.83 4.83 16.66 0 1 8.50 1 0 1 1 3 1 0 54.66 1 0 0 0 0 11.16 27.66 0 0 1 1 24.66 7.83 23.33 1 1 7.66 0 0 1 0 3 1 0 27.33 1 1 1 1 0 0.66 13.33 0 1 1 2.66 32.16 10.16 29.50 1 1 16.66 0 0 0 0 2 0 1 21.50 1 1 1 1 0 0.45 13.66 0 1 1 4 32.50 11.16 30.66 1 1 14 0 0 1 0 2 0 1 33.50 1 0 1 0 0 3.45 17.66 1 0 1 3.16 31.33 9.83 28.33 1 1 20.16 0 0 1 0 3 0 1 51.66 1 0 1 1 1 4.50 20 0 0 1 3.33 18.16 5.16 22.33 0 1 9.16 0 0 0 1 3 1 0 48.83 1 0 1 1 1 3.50 19.16 0 0 1 3 13.50 5.33 21.16 0 1 11.33 0 0 0 1 3 0 0 49 1 0 0 1 1 3.73 10.50 0 0 1 3.66 22.33 6.16 21.16 0 1 11.33 0 0 0 0 3 1 0 12 0 1 1 0 0 0.90 8.83 0 1 1 1 26.66 7.50 27.83 1 1 15.50 0 1 1 1 3 0 1 13.33 0 1 1 0 0 0.96 9 0 1 1 1 26.66 8.83 34.50 1 1 16 0 1 1 1 3 0 1 4 0 0 1 0 0 2.66 3 0 0 1 1 14.66 5 10.83 0 1 7.33 0 0 1 1 3 0 0 49 1 0 0 1 0 2.13 8.25 1 1 0 12.66 19.66 9 16.50 0 1 10.33 0 1 0 1 2 0 0 32.16 1 0 0 1 0 1.83 8.83 1 1 0 6.16 16.83 7.16 17.16 0 1 8.66 0 1 0 1 2 0 0 13.83 0 1 1 0 0 0.48 10.50 0 1 1 3 27.66 8.83 26.33 1 1 15.83 0 1 1 1 2 0 1 12.16 0 1 1 0 0 2.41 11.50 0 1 1 1.66 24.16 9.83 24.33 1 1 16.50 0 1 1 1 2 0 1 23.66 0 1 1 0 0 2.33 13.66 0 1 1 6.50 27.50 8.16 29.33 1 1 14.16 0 1 1 1 2 0 1 66 1 0 1 1 0 2.18 10.33 1 0 1 10.66 16.83 5.16 15.66 0 1 11 1 0 1 1 2 0 0 65.33 1 0 1 1 0 1.96 11.83 1 1 1 18 15 5.83 16.83 0 1 11.16 1 0 1 1 2 0 0 10.16 0 1 1 0 0 0.58 8.66 0 1 1 1.66 22 7.50 17.16 1 0 15.66 0 1 0 1 3 0 1 8.83 0 1 1 0 0 0.96 8.33 0 1 1 2.50 25 9 20.33 1 0 16 0 1 0 0 3 0 1 52.16 1 0 0 1 0 1.41 11.83 1 0 1 5.33 19.16 6.16 13 0 1 8.66 0 0 1 1 3 0 0 8.16 0 1 0 0 0 0.81 7.25 0 1 0 3.16 12.66 4.50 11.83 0 1 6.33 1 0 1 1 2 0 0 7.50 0 1 0 0 0 0.86 6.83 0 1 0 3.33 13.33 3.50 12 0 1 5.41 1 0 1 1 2 0 0 11.83 0 0 0 0 0 2.08 7.83 0 0 0 1.83 17.66 8 14.83 0 1 5.91 0 1 1 1 2 0 0 10.16 0 0 0 0 0 1.33 5.50 0 0 0 2.50 14.16 5.50 14.16 0 1 5.66 0 1 1 1 2 0 0 75 1 0 0 1 0 2.18 11.83 1 0 0 44.16 18.16 7.50 15.50 0 1 9.50 1 1 1 1 3 0 0 U : \ A C T A B O T A N I C A \ A c t a - B o t a n 2 - 1 0 \ 2 0 0 M a k b u l . v p 1 1 . l i s t o p a d 2 0 1 0 1 3 : 2 8 : 4 1 C o l o r p r o f i l e : D i s a b l e d C o m p o s i t e 1 5 0 l p i a t 4 5 d e g r e e s ACTA BOT. CROAT. 69 (2), 2010 243 MORPHOMETRIC STUDY ON SCORZONERA Fig. 3. Principal component analysis of 39 populations projected onto the first two axes. For popu- lation number explanations see table 1. Fig. 4. Principal component analysis of 25 variables projected onto the first two axes. For variables number explanations see table 2. U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:44 Color profile: Disabled Composite 150 lpi at 45 degrees components are 92.25% (X25), 83.705% (X16) and 69.471% (X2). These characters are among the binary traits examined in this study. Only the first three components were taken into account because of their eigenvalues. The three components account together for 60.55% variation (Tab.4). The first component accounts for 27.23% variation. The second component accounts for 17.75%, the third component accounts for 15.57% and together they account for 60.55 %. Discussion There are many clustering methods and similarity coefficients in the literature as can be seen in SNEATH and SOKAL (1973). In this study, UPGMA along with Gower’s coefficient has been used for clustering the 39 OTUs (Fig. 2). In order to determine how well this dendrogram represents the underlying matrix of resemblances, the cophenetic correlation 244 ACTA BOT. CROAT. 69 (2), 2010 MAKBUL S., TURKMEN Z., COSKUNCELEBI K., BEYAZOGLU O. Tab. 4. Percentage of variance of variables accounted for by first three components. Names of variables PC-1 PC-2 PC-3 X1 15.981 7.052 61.736 X2 7.373 7.770 69.471 X3 57.119 0.730 4.527 X4 8.108 45.162 0.147 X5 13.995 4.502 34.571 X6 26.887 51.816 0.339 X7 10.890 11.774 25.331 X8 0.385 6.451 59.081 X9 3.948 30.556 17.282 X10 49.736 14.570 1.507 X11 3.927 65.503 3.842 X12 8.451 24.685 9.668 X13 56.878 3.881 16.841 X14 36.768 3.774 12.378 X15 43.064 3.743 21.436 X16 83.705 1.424 4.333 X17 17.493 0.209 2.170 X18 64.405 3.017 10.716 X19 14.161 20.058 0.687 X20 25.361 30.846 1.906 X21 0.655 0.023 8.316 X22 11.527 7.026 21.506 X23 5.888 51.826 0.362 X24 21.179 47.062 0.389 X25 92.789 0.340 0.715 Percentage of variance explained 27.23 17.75 15.57 Cumulative percentages of variance explained 27.23 44.98 60.55 U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:44 Color profile: Disabled Composite 150 lpi at 45 degrees coefficient (rcs) was also calculated. It has generally been found to vary from 0.6 to 0.95, depending on the methods producing the dendrogram and the nature of the differences among the specimens classified (SNEATH and SOKAL 1973). Our dendrogram had a cophe- netic correlation of 0.72. This means that the dendrogram provides a fairly accurate repre- sentation of the resemblances among OTUs. All the examined OTUs fall into two major clusters at 88.2% dissimilarity levels (Fig. 2). One, labeled »a«, consists of populations that belong to scapigerous and caulescent taxa, the other, labeled »b«, includes all the remaining OTUs representing entirely scapi- gerous taxa. When the dendrogram is carefully examined, it can be seen that cluster »a« consisting of most of the examined OTUs belongs to fifteen taxa, but cluster »b« consisting only of nine OTUs belongs to four taxa. There are no significant differences in the anatomi- cal peculiarities among the taxa in group »b« (LIPSCHIZ 1964, KAMELIN and TAGEV 1986). Cluster »a« splits into two small clusters (c, d). Cluster »c« consists of OTUs corre- sponding to scapigerous, subscapigerous and caulescent taxa. They are morphologically and anatomically related to each other. All OTUs in this group have a typically cylindrical root system, but the representatives of S. parviflora (OTU 15) and S. insica (OTU 18) form a small group because of the unforked stem and the pubescence. The representatives of S. armeniaca (OTUs 12–14) and S. laciniata ssp. laciniata (OTUs 19–21) are close to each other because of the similar morphological traits as indicated by CHAMBERLAIN (1975). These views are correlated with our results obtained from UPGMA. Additionally, our find- ings also support the results based on caryological data as indicated by NAZAROVA (1997). Although OTU 24 (S. sericea) display some different morphologic properties and grow in different habitats from the other examined populations in cluster »c«, it occurs in cluster »c«, together with the other OTUs because of the shared characters such as cylindrical root, pinnatisect leaf, similar achene and flower peculiarities. Although all representatives of S. cana (OTUs 1–7) are uniform at the specific level, they are not separated at the subspecific level. S. cana is said to be close to S. laciniata ssp. laciniata (CHAMBERLAIN 1975), but our results from UPGMA do not support this view and show that S. cana is more closely related to S. armeniaca (Fig. 2). This situation might be explained as follows; S. laciniata prefers dry habitats and is characterized by an entire leaf and S. cana prefers humid habitats and is characterized by a compound leaf. Cluster »d« includes OTUs characterized by densely hairy, generally caulescent and a few lanate scapigerous taxa (S. pseudolanata, S. seidlitzii). Although the representatives of S. pseudolanata (OTUs 35–36) and S. seidlitzii (OTUs 37–38) are morphologically similar to the representatives of cluster »b«, these scapigerous OTUs that are densely lanate and hairy occur in cluster »d«. The major group of the cluster »d« includes only caulescent to- mentose taxa. But S. latifolia (OTU 39) characterized by numerous capitula (more than 20) and S. cinerea (OTU 34) characterized by globrous achenes form a small group in cluster »d«. S. tomentosa is closed to S. latifolia (CHAMBERLAIN 1975). Our results do not support this idea; those from UPGMA show that S. latifolia is more closely related to S. cinerea and S. sosnowskyi than S. tomentosa. OTUs characterized by scapigerous or semi-caulescent, tuberose-root, glabrous or pu- bescent and entire linear lamina with undulate margins were clustered in group »b«. It seems that the representative of S. mollis ssp. mollis forms a small sub-group in cluster »b«. ACTA BOT. CROAT. 69 (2), 2010 245 MORPHOMETRIC STUDY ON SCORZONERA U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:44 Color profile: Disabled Composite 150 lpi at 45 degrees The reason for this can be explained with the different phenetic traits such as the forked and semi-caulescent stem. The other OTUs in cluster »b« have very similar morphological characters. The representatives of S. suberosa (OTUs 32–33) with lilac to purple flowers are separated from the other OTUs (Fig. 2). This delimitation is coincided with the caryological and ITS results of MAVRODIEV (2004). The ranks of subgenus and section have been used in Scorzonera by several scientists (LIPSCHIZ 1964, KAMELIN and TAGEV 1986, NAZAROVA 1997). Scorzonera cana, S. armeniaca and S. laciniata ssp. laciniata treated in subg. Podospermum (DC.) Lipschiz, S. suberosa in subg. Pseudupodospermum (Lipschiz et Krash) Lipschiz, S. seidlitzii in subgen. Scorzonera sect. Pulvinares Lipschiz and S. latifolia in subg. Scorzonera sect. Nervosae Lipschiz (LIPSCHIZ 1964). All the OTUs of these listed taxa are grouped as indicated by LIPSCHIZ (1964) and our results from UPGMA support the use of such ranks within the genus (Fig. 2). The position of the 39 examined populations on the first two components of the PCA (Fig. 3) shows that all populations examined in this study are split into three distinct clus- ters. These clusters are almost associated with the results from UPGMA except for the rep- resentatives of S. insica. This situation could be explained by the differences in the root sys- tems and the shapes of lamina, which are the most important characters, explaining most of the variation among the examined taxa. This might be caused by the insufficient specimens or population examined in this study. There is no other contradiction between PCA and UPGMA with respect to the distribution of OTUs. The first two extracted components explain 44.98% of the total variation among the ex- amined populations (Fig. 3). The three components account together for 60.55% of the variation among populations (Tab. 4). The first component emphasizes pubescence and length of achenes, shape of outer phyllaries, average length of flowering capitula, root shape and leaf margin and the second component with apex of outer phyllaries, hair types of plant and shape of lamina. The third component emphasizes length of the leaf, total plant height and state of the plant stem. However, weights are all under 50% and almost equal. In summary, the state of residue leaves at the base of the stem, pubescence and length of achenes, shape of outer phyllaries, average length of flowering capitula, root shape and leaf margin are the most important characters separating the Scorzonera species. These charac- ters are the basic ones used in most floras for separating these species (CHATER 1976, CHAMBERLAIN 1975). The present study is a preliminary step in the analysis of morphological characters by means of numerical analysis. The results basically agree with the traditional taxonomic treatments of Scorzonera. Although qualitative variables such as shape of outer phyllaries, achenes and root shape explain most of the total variation, binary characters seem to be more important than quantitative ones in separating Scorzonera taxa. Acknowledgements The authors would like to express their thanks to Dr. Ahmet Duran from the Education Faculty, Selçuk University (Konya-Turkey) for kindly helping on the identification of sam- ples. Thanks to the Scientific and Technological Research Council of Turkey for financial support. 246 ACTA BOT. CROAT. 69 (2), 2010 MAKBUL S., TURKMEN Z., COSKUNCELEBI K., BEYAZOGLU O. U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 13. listopad 2010 12:02:49 Color profile: Disabled Composite 150 lpi at 45 degrees References BREMER, K., ANDERBERG, A. A., 1994: Asteraceae: Cladistics and classification. Timbers Press, Portland. CHAMBERLAIN, D. F., 1975: Scorzonera L. In: DAVIS, P. H. (ed), Flora of Turkey and the east Aegean islands. Edinburgh University Press, Edinburgh. CHATER, A. O., 1976: Scorzonera. In: TUTIN, T. G, HEYWOOD, V. H, BURGES, N. A, VALENTINE, D. H, WALTERS, S. M, WEBB, D. A. (eds.), Flora Europaea. Cambridge University Press, Cambridge. DURAN, A., 2008: Rediscovery of the poorly known Scorzonera argyria and its relation- ships in Turkey. Biologia 63, 1078–1084. ERTUÐ, F., 2000: An ethnobotanical study of Central Anatolia (Turkey). Economic Botany 54, 155–182. GUARDIA, C. D., BLANCA, G., 1987: Karyology of the Scorzonera (Compositae) species from the Iberian Peninsula. Plant Systematic and Evolution 156, 29–42. KAMELIN, R. V., TAGEV, I. U., 1986: Survey of the species of the genus Scorzonera (Asteraceae). Botanical Journal 71, 1972–1982. LIPSCHIZ, S. J., 1964: Scorzonera. In: SHISHIN, B. K. (ed.), Flora of the USSR, 29 (in Rus- sian). Academy of Science of the USSR, Moscow, Leningrad. MAGIATIS, P., MITAKU, S., SKALTSOUNIS, A., TILLEQUIN, F., 2001: 1-Oxo-2-hydroxy-1,2- -dihydroacronycine: a useful synthon in the acronycine series for the introduction of amino substituents at 6-position and for the conversion into isopropylfuroacridones. Chemical and Pharmological Bulletin 49, 1304–1307. MAVRODIEV, E. V., EDWARDS, C. E., ALBACH, D. C., GITZENDANNER, A., SOLTIS, P. S., SOLTIS, D. E., 2004: Phylogenetic relationships in subtribe Scorzonerinae (Asteraceae: Cichorioideae: Cichorieae) based on ITS seguence data. Taxon 53, 699–712. NAZAROVA, E. A., 1997: Karyosystematic investigation of the genus Scorzonera L. s.l. (Lactuceae, Asteraceae). Caryologia 50, 239–261. SNEATH, P. H. A., SOKAL, R. R., 1973: Numerical taxonomy: The principles and practice of numerical classification. W. H. Freeman and Company, San Francisco. PODANI, J., 1993: Multivariate data analysis in ecology and systematic, A methodological guide to Syn-Tax 5.0 Package. SPB Academic Publishing, Netherlands. RECHINGER, K. H., 1977: Scorzonera L. In: Rechinger, K. H. (ed.), Flora Iranica, 122, 16–79. Graz Akademische Druck und Verlagsanstalt. RIVERA, D., OBÓN, C., HEINRICH, M., INOCENCIO, C., VERDE, A., FAJARDO, J., 2006: Gathered Mediterranean food plants–ethnobotanical investigations and historical de- velopment. Forum of Nutrition 59, 18–74. ZIDORN, C., ELLMERER, E. P., STURM, S. STUPPNER, H., 2003: Trylobibenzyls E and F from Scorzonera humilis and distribution of caffeic acid derivatives, lignans and tyrolo- bibenzyls in European taxa of the subtribe Scorzonerinae (Lactuceae, Asteraceae). Phytochemistry 63, 61–67. ACTA BOT. CROAT. 69 (2), 2010 247 MORPHOMETRIC STUDY ON SCORZONERA U:\ACTA BOTANICA\Acta-Botan 2-10\200 Makbul.vp 11. listopad 2010 13:28:44 Color profile: Disabled Composite 150 lpi at 45 degrees