Scientia Horticulturae 126 (2010) 138–144 Contents lists available at ScienceDirect Scientia Horticulturae journa l homepage: www.e lsev ier .com Phenot ly P.J. Terzo Agricultural Un hens, a r t i c l Article history: Received 6 No Received in re Accepted 29 Ju Keywords: Intra-populati Nei’s index Population str Phenotypic div Tomato ions, racter genet gnific netic es us (c) i appli ide d ersity drace the majority of traits, high degree of variation was observed within landrace populations. Most landrace populations had similar magnitudes of phenotypic diversity (H¯p) ranging from 0.24 to 0.52. In most of the landraces, stempubescence density, foliage density and plant growth type had the lowest phenotypic diversity within landraces (Hs). The heterogeneity of each population was mainly attributed to various traits related to fruit form. Conclusively, the in depth analysis of the phenotypic diversity revealed the 1. Introdu The redu ersicum L.) McCouch, 1 utilization drace popu (Zeven, 199 breeding sc for their ef production The effic require the 1995). The at random, ronmental f is related t mixtures of sent divers both farme ∗ Correspon E-mail add 0304-4238/$ – doi:10.1016/j. allelic and genotypic richness of the studied tomato landrace populations that can be exploited in tomato breeding programs. © 2010 Elsevier B.V. All rights reserved. ction ction of genetic variation in tomato (Solanum lycop- through domestication and breeding (Tanksley and 997) has resulted in the need for conservation and of all existing genetic resources. Heterogeneous lan- lations are among themost important genetic resources 8) and have been, and will continue to be used in plant hemes. The characterization of landraces is imperative ficient use in plant breeding efforts to improve crop (Fischbeck, 1989). ient conservation and exploitation of landraces also study of their genetic diversity structure (van Hintum, genetic diversity in any local population does not occur but is structured based on various biological and envi- actors. The high level of within landrace heterogeneity o its adaptability (Cooper et al., 2000). Landraces are phenotypes (Kiesling cited in Zeven, 1998), and repre- e and dynamic genepools that evolve over time under r and natural selection pressure (Hawtin et al., 1997). ding author. Tel.: +30 210 5294626; fax: +30 210 5294622. ress:
[email protected] (P.J. Bebeli). The role that farmers play is both direct and/or indirect as happens through farming system changes (Fruwirth cited in Zeven, 1998), or due to social reasons (Zeven, 2002). The influence of farmers on landraces can significantly change allele frequency resulting in either an increase or decrease of the heterogeneity level (Zeven, 2002). The introduction of tomato into Europe fromCentral and South- ern America at the beginning of the 16th century represented a genetic bottleneck for the cultivated tomato germplasm in Europe (Rick, 1976). The genetic heritage of the tomato was further eroded by the development of vintage and modern cultivars (Williams and St. Clair, 1993). Genetic profiles of tomato landraces are clearly dif- ferent from those of modern tomato hybrids (Carelli et al., 2006; García-Martínez et al., 2006; Terzopoulos and Bebeli, 2008). A part of the initial diversity may have been conserved in a range of lan- draces that have been cultivated for centuries and many of these are still commonly found at local markets (García-Martínez et al., 2006). Using morphological/agronomical traits, biochemical char- acteristics and molecular markers, significant levels of phenotypic and genetic diversity among tomato landraces have been observed (Carbonell-Barrachina et al., 2006; Carelli et al., 2006; Terzopoulos and Bebeli, 2008; Mazzucato et al., 2008, 2010; Goncalves et al., 2009; Labate et al., 2009; Terzopoulos et al., 2009). However, infor- mation about the variation present within each tomato landrace is see front matter © 2010 Elsevier B.V. All rights reserved. scienta.2010.06.022 ypic diversity in Greek tomato (Solanum poulos, P.J. Bebeli ∗ iversity of Athens, Department of Plant Breeding and Biometry, Iera Odos 75, 11855 At e i n f o vember 2009 vised form 15 June 2010 ne 2010 on diversity ucture ersity a b s t r a c t Landraces are heterogeneous populat different approach to population cha used per se or as a very interesting structure of tomato landraces is of si within tomato landraces and their ge characterize 34 Greek tomato landrac within and among the landraces, and Nei’s genetic diversity statistics was the landraces. The results showed a w analyzed (mean total phenotypic div meanphenotypicdiversity among lan / locate /sc ihor t i copersicum L.) landraces Greece generally have wide intra-population diversity, and require a ization than that used for the homogeneous varieties. Either ic resource in breeding programs, investigating the genetic ant importance. However, information on variation present structure is limited. The present study was designed to: (a) ing 36 morphological traits, (b) estimate phenotypic diversity dentify the traits that contributed to landrace heterogeneity. ed. Principal component analysis (PCA) was used to classify iversity present in the whole collection for most of the traits HT =0.47 with the majority of values around the mean). The s (GST)was0.21. Somevegetative traitshadGST >0.56,while for P.J. Terzopoulos, P.J. Bebeli / Scientia Horticulturae 126 (2010) 138–144 139 Table 1 Geographical data of the tomato landraces studied. Collection code Province Location Local name Latitude Longitude Altitude (m) GR008/82 Halkidiki Ormilia Milo 40◦17′ 23◦32′ 40 GR012/82 40◦17′ 23◦32′ 40 GR065/93 37◦44′ 26◦59′ 60 GR069/94 35◦27′ 27◦07′ 50 GR074-076/ 38◦09′ 20◦31′ 280 GR076/92 38◦09′ 20◦31′ 280 GR076/94 35◦54′ 26◦51′ 320 GR090-093/ 38◦03′ 20◦46′ 20 GR091/92 38◦09′ 20◦30′ 20 GR092/93 37◦44′ 26◦59′ 60 GR11/03 38◦09′ 20◦30′ 20 GR14/99 37◦36′ 26◦56′ 630 GR16/99 37◦36′ 26◦56′ 630 GR193/99 35◦02′ 26◦52′ 140 GR240/99 41◦33′ 26◦13′ 63 GR242/99 41◦33′ 26◦13′ 63 GR243/99 41◦33′ 26◦13′ 63 GR25/99 37◦30′ 26◦56′ 240 GR279/99 41◦01′ 24◦13′ 60 GR300/99 41◦02′ 22◦07′ 170 GR301/99 41◦02′ 22◦07′ 170 GR5/01 38◦01′ 21◦04′ 240 GR60/99 37◦57′ 25◦50′ 130 VE023/83 41◦05′ 25◦47′ 60 VE025/83 41◦3′ 26◦32′ 30 VE026/83 40◦28′ 25◦36′ 60 VE027/83 40◦38′ 24◦42′ 240 VE029/83 41◦ 24◦14′ 50 VE030/83 40◦42′ 24◦44′ 110 VE033/83 41◦06′ 24◦04′ 80 VE035/83 ◦ ′ ◦ ′ Pastra Santorini Omospondia still limited 2008; Terzo Molecul terizing the molecular t of populati descriptors study were 36 morpho among and identify the classify the To our kno presentwit collection a drace diver 2. Materia 2.1. Plant m Plant ma landraces (T saloniki wi locally colle with differe data are giv complete b at the Expe Greece. A to blocks with 80 cm apart easu erva o Des Halkidiki Ormilia Milo Samos Vlamari Philloto Karpathos Arkassa 92 Kefallonia Pastra Kefallonia Pastra Karpathos Stes 92 Kefallonia Argostoli Milati Kefallonia Argostoli Samos Koukari Ahladi Kefallonia Argostoli Samos Pandrosos Samos Pandrosos Lasithi Lagada Lainates Evros Kyprinos Evros Kyprinos Evros Kyprinos Souvritiki Samos Marathokampos Drama Kalambaki Pella Neochori Pella Neochori Ithaki Pantochori Pantaroza Ikaria Vardata Bourneli Evros Argiani Evros Orestiada Samothraki Therma Thasos Theologos Drama Kalamonas Thasos Potamia Drama Argyroupoli Kavala Lydia Kefallonia Pastra Kyklades Thira Thessaloniki Thessaloniki (Mazzucato et al., 2008, 2010; Terzopoulos and Bebeli, poulos et al., 2009). ar markers have proved to be useful tool for charac- genetic composition of agricultural crops. Although 2.2. M Obs Tomat echniques have recently become the popular method on analysis, phenotypic traits remain indispensable for evaluating genetic variation. The objectives of this : (i) to characterize Greek tomato landraces using logical traits, (ii) to estimate the phenotypic diversity within the Greek tomato landrace populations, (iii) to traits contributed to landrace heterogeneity and (iv) to populations using principal component analysis (PCA). wledge this is the first study of phenotypic diversity hin individual tomato landraces fromawidegermplasm nd of identifying the traits that contributed to the lan- sity by analyzing a large number of plants. ls and methods aterial and experimental design terial consisted of 34 Greek tomato (S. lycopersicum L.) able 1) obtained from the Hellenic Gene Bank of Thes- th the exception of the “Santorini” landrace that was cted. The landraces came from diverse areas of Greece nt agroclimatic conditions (Fig. 1). Their geographical en in Table 1. The experiment followed a randomized lock design with three replications and was carried out rimental Field of the Agricultural University of Athens, tal of 54–60 plants per landrace were grown in three a 140 cm from row to row distance and plants spaced in-the-row. surements i exterior col pistil scar a ments inclu Fig. 1. Geogra location site is 41 02 24 17 80 38◦09′ 20◦31′ 280 36◦25′ 25◦26′ 50 40◦39′ 23◦07′ 2 rements tions on 36 morphological traits were taken based on criptors (1996). These included: (i) nine nominal mea- ncluding: leaf type, inflorescence type, style shape, fruit or, fruit flesh color, core color, fruit shape, shape of fruit nd cross-sectional shape; (ii) 10 continuous measure- ding: petal, sepal, stamen and stem internode length, phic locations of tomato landraces collected from Greece. Name of given in italics. 140 P.J. Terzopoulos, P.J. Bebeli / Scientia Horticulturae 126 (2010) 138–144 height of the first truss, fruit weight, fruit width, fruit length, core size and number of locules; and (iii) 17 ordinal measurements including: style position, easiness of fruit skin to be peeled, fruit size, fruit flesh color intensity, ribbing at calyx end, leaf attitude, fruit exterio green shou tion, fruit sh stem pubes phological i.e. leaf atti drooping on 2.3. Data an All conti were transf by dividing that formed and Gauvre each rank w and for eac phenotypic Phenoty Nei’s geneti notypic div each landra estimatedb The propor estimated f landrace ac lated and u landrace. Compari were carrie 2000) with In order to unequal nu classes wer (2008). To i diversity w was carried plants for e replications Tukey-Kram Classific component of the frequ the comput 3. Results 3.1. Charac 3.1.1. Veget The stud having ‘inde whole colle stem pubes were homo hadmixed g had ‘interm erect’ leaf a plantswith first truss, r 3.1.2. Inflorescence traits Most of the plants (62.9%) in the collection had ‘multiparous’ type inflorescence, while 54.1% had ‘inserted’ style position and 68.7% had ‘simple’ style shape. The plants with ‘slightly exserted’ ppea d’ s edia ength ion. H Trait he w inter lor a lors o 7.3% nterm 5/93” Gene it len f the Trait he e light .2% und ion h ur ty rately y (43 sed s and 4 ’. M ion a Perc ape ) and at is locu locul enot Total tota 0.65 rang or e olor t caly lative locul 2). Phen pro ange f the were y, lea e ea olor ce of .08 ( r color intensity, green trips on the fruit, intensity of lder, fruit blossom end shape, blossom end scar condi- oulder shape, width of pedicel scar, plant growth type, cence density, foliage density and skin color. Each mor- trait was evaluated using a different number of ranks, tude had three ranks a semi-erect, a horizontal and a e. The ordinal traits had two to five ranks. alysis nuous traits were divided into three classes and thus ormed to ordinal. The transformation was carried out the range of continuous trait into three equal parts three discrete ranks (Bechere et al., 1996; Pagnano au, 2000). For the characterization, the frequency of ithin each trait was calculated for the whole collection h landrace separately. The data was used to calculate diversities. pic variation among landraces was calculated using cdiversity statistics (Nei, 1973). For each trait, total phe- ersity of the collection (HT), phenotypic diversitywithin ce (Hs) and its average across all landraces (H¯s) were y applyingNei’s genetic diversity index (He) (Nei, 1973). tion of phenotypic diversity among landraces (GST) was or each trait. Themeanphenotypic diversitywithin each ross all traits (referred as H¯p for clarity) was also calcu- sed for the estimation of heterogeneity of individual sonsof the H¯pof all landraces and thoseof H¯sof all traits d out using Tukey’s mean comparison method (Kuehl, the statistical software JMP-7 (SAS Institute Inc., 2007). compare the phenotypic diversity values of traits with mber of ranks, transformations on the basis of three e carried out following the method of Terzopoulos et al. dentify the traits with the highest value of phenotypic ithin landraces, a Monte Carlo sampling (Weir, 1990) out by creating 100 samples of 60 randomly chosen ach landrace and the resultant samples were used as for the Tukey’s mean comparison method to obtain er HSD (Honestly Significant Difference). ation of landraces was performed with the principal analysis (PCA) (Sneath and Sokal, 1973) on a matrix encies of ranks for all 36 traits. PCA was carried out on er program NTSYS-pc (Rohlf, 1998). terization of Greek tomato landraces collection ative traits ied Greek tomato landraces could be characterized as terminate’ plant growth type (72.4% of the plants in the ction), ‘standard’ leaf type (83.8%) and ‘intermediate’ cence density (90.9%). Although, most of the landraces genous for ‘indeterminate’ growth type, some of them rowth types at various ratios.Most of the plants (69.8%) ediate’ foliage density and60%of the plants had a ‘semi- ttitude. The collection also had 78.1% and 68.7% of the ‘short’ internode length and ‘intermediate’ height of the espectively. style a exserte ‘interm petal l collect 3.1.3. In t with ‘ core co rior co ‘red’ (1 with ‘i “GR06 fruits. The fru 67.3% o 3.1.4. In t with ‘s and 24 also fo collect The fo ‘mode quentl depres varied ‘narrow condit shape. scar sh (15.1% skin th ‘small’ ‘large’ 3.2. Ph 3.2.1. The 0.08 to had HT higher flesh c bing a had re ber of (Table 3.2.2. The (GST) r Most o values densit sity. Th flesh c presen GST ≤0 red at a frequency of 16.1% and only 3.5% had ‘highly tyles. More than half of the plants (57%) were of an te’ petal length. The sepal length was similar to the in size and frequency of distribution throughout the alf of the collection plants had ‘long’ stamen length. s related to the fruit color and size hole collection, most of the fruits had ‘red’ flesh color mediate’ intensity, ‘yellow’ skin color, ‘intermediate’ nd ‘absence’ of green shoulder. Themost common exte- f mature fruit were ‘orange’ (34.7%), ‘orange/red’ (29%), ) and ‘orange/pink’ (15.5%). Most of the plants had fruits ediate’ (41.7%) or ‘large’ (28.2%) size. Two landraces, and “Santorini”, had a significant number of ‘small’ size rally 69.6% of the plants had fruits with low weight. gth and width appeared as ‘intermediate’ in 67.8% and plants in the collection, respectively. s related to the fruit form ntire landrace collection, 35% of the plants had fruits ly flattened’ shape, while 24.7% had ‘rounded’ fruits, had ‘flattened’ shaped ones. Other fruits shapes were in the collection. More than half of the plants in the ad fruits with ‘angular’ cross-sectional shape (55.9%). pes of calyx end ribbing had similar percentages and depressed’ fruit shoulder shape appeared more fre- %) than other shapes. Plants with ‘slightly’ or ‘strongly’ hapes had similar percentages. The pedicel scar width 8.4% were ‘wide’, 32.4% were ‘medium’ and 19.2% were ost plants (71.1%) had fruits with ‘closed’ blossom end nd 66.3% of the plants had fruits with ‘flat’ blossom end entages of the plants having the four types of the pistil were as follows: ‘dot’ (36.6%), ‘stellate’ (20.1%), ‘linear’ ‘irregular’ (28.2%). Most plants (53.9%) had fruits with ‘difficult’ to be peeled and 85.3% of them had fruits with le number, while 12.5% had ‘intermediate’ and 2.2% had e number. ypic diversity of tomato landrace collection phenotypic diversity l phenotypic diversity (HT) for each trait ranged from with an average of 0.47 (Table 2). The majority of traits ing from 0.39 to 0.59. Some of the traits had HT value qual to 0.61, namely easiness of fruit skin to be peeled, intensity, fruit shoulder shape, pistil scar shape, rib- x end and width of pedicel scar. The following traits ly low HT values (≤0.26): color of core, leaf type, num- es, skin color of ripe fruit and stem pubescence density otypic diversity among landraces portion of phenotypic variation among the landraces d from 0.04 to 0.76 with an average of 0.21 (Table 2). traits had GST value ranged from 0.11 to 0.29. The GST higher or equal to 0.57 for the following traits: foliage f attitude, plant growth type and stem pubescence den- siness of fruit skin to be peeled, exterior color intensity, intensity, green shoulder intensity, number of locules, green trips on the fruit and stem internode length had Table 2). P.J. Terzopoulos, P.J. Bebeli / Scientia Horticulturae 126 (2010) 138–144 141 Table 2 Values of HT , GST , the mean phenotypic diversity within populations (H¯s) for each morphological trait across all landraces and the number of occurrences of the traits with the statistically highest and lowest Hs in the landrace populations studied. Traits HT GST H¯sa Numberofoccurrence in34populations Intensity of 9 (0.4 Easiness to b 7 (0.2 Ribbing at ca 5 (0–0 Fruit cross-s 4 (0.2 Green trips o 3 (0.2 Blossom end 3 (0.1 Shape of pis 1 (0–0 Exterior colo 9 (0.3 Fruit should 9 (0.1 Width of ped 8 (0.1 Petal length 5 (0.0 Stamen leng 5 (0.1 Sepal length 5 (0–0 Style positio 4 (0.1 Fruit exterio 3 (0.2 Fruit size 3 (0.1 Fruit shape 1 (0.1 Fruit blossom 9 (0–0 Fruit length 9 (0.0 Stem interno 8 (0.1 Size of core 7 (0.0 Fruit weight 6 (0–0 Style shape 5 (0.0 Fruit width 5 (0.0 Flesh color 4 (0–0 Intensity of 2 (0–0 Inflorescenc 2 (0–0 Height of the 0 (0.0 Leaf attitude 5 (0–0 Number of l Skin color of Plant growth Leaf type Foliage dens Stem pubesc Core color HSDb Mean a Range of H b HSD (Tuke 3.2.3. Phen For each average of son method diversity w type and ste had high H¯ of fruit skin shape, gree Most of the ulations (Ta landraces’ h ribbing at c easiness to color inten landraces, r phenotypic traits (Table 3.2.4. Mean The m across all average of parison m significantly flesh color 0.61 0.05 0.5 e peeled 0.61 0.07 0.5 lyx end 0.65 0.13 0.5 ectional shape 0.60 0.10 0.5 n the fruit 0.55 0.04 0.5 scar condition 0.59 0.13 0.5 til scar 0.64 0.18 0.5 r intensity 0.51 0.04 0.4 er shape 0.62 0.19 0.4 icel scar 0.61 0.21 0.4 0.53 0.15 0.4 th 0.52 0.14 0.4 0.55 0.20 0.4 n 0.53 0.16 0.4 r color 0.53 0.15 0.4 0.53 0.16 0.4 0.50 0.21 0.4 end shape 0.48 0.20 0.3 0.51 0.23 0.3 de length 0.41 0.07 0.3 0.52 0.29 0.3 0.42 0.14 0.3 0.45 0.22 0.3 0.50 0.29 0.3 0.39 0.11 0.3 green shoulder 0.34 0.04 0.3 e type 0.45 0.29 0.3 first truss 0.36 0.15 0.3 0.58 0.57 0.2 ocules 0.26 0.08 0.24 (0–0 ripe fruit 0.23 0.22 0.21 (0–0 type 0.39 0.63 0.14 (0–0 0.19 0.29 0.14 (0–0 ity 0.49 0.76 0.12 (0–0 ence density 0.18 0.61 0.07 (0–0 0.08 0.11 0.06 (0–0 0.13 0.47 0.21 0.37 s in each morphological trait is in parenthesis. y-Kramer Honestly Significant Difference) at P≤0.05. otypic diversity within individual landraces trait studied, H¯s ranged from 0.06 to 0.59 with an 0.37 (Table 2). Based on the Tukey’s mean compari- , the traits with the significant low mean phenotypic ere color of core, foliage density, leaf type, plant growth m pubescence density (H¯s ≤ 0.14). The following traits s values (≥0.53): blossom end scar condition, easiness to be peeled, flesh color intensity, fruit cross-sectional n trips on fruits, and ribbing at calyx end (Table 2). traits had a wide range of H¯s values across the pop- ble 2). The traits that contributed the most to within eterogeneity differed from landrace to landrace. The alyx end and pistil scar shape, cross-sectional shape, be peeled, fruit blossom end scar condition and flesh sity had the highest H¯s values in 7, 7, 6, 6, 5 and 4 espectively (Table 2). Traits with the lowest value of diversitywithin individual landracesweremainlyplant 2). phenotypic diversity of each landrace ean phenotypic diversity within each landrace traits (H¯p) ranged from 0.24 to 0.52 with an 0.40 (Table 3). Based on Tukey’s mean com- ethod the majority of landraces did not differ from each other (HSD=0.17). Only two small fruited lan the most from the fi “GR240/99” (Table 3). 3.3. Landra The first 43.6% of th PCA could p groups (Fig a distinct g light fruits, of the first “GR193/99” from the r growth typ (data not s each, were “VE026/83” the remain them. Thes with larger (data not sh Highest Lowest 7–0.67) 4 – 0–0.66) 6 – .66) 7 1 3–0.67) 6 – 1–0.67) 3 1 7–0.66) 5 1 .72) 7 1 4–0.6) – – 2–0.62) – – 2–0.63) 1 – 8–0.62) – 1 2–0.57) 2 1 .61) 1 1 4–0.66) – – 4–0.62) – – –0.66) 1 – 2–0.63) – – .61) – – 8–0.6) – 1 6–0.58) – – 4–0.64) – 14 .61) – 1 3–0.54) 1 1 4–0.61) – 1 .57) 1 1 .59) 1 1 .5) – 4 4–0.49) – 1 .67) 2 15 .58) – 4 .67) 1 8 .64) 2 21 .51) – 19 .65) 1 22 .49) – 28 .56) 1 – draces, “Santorini” and “GR065/93”, which were homogeneous landraces, differed significantly ve most heterogeneous ones, namely “GR243/99”, , “GR14/99”, “GR242/99”, “GR16/99”, and “GR5/01” ces classification three axes of principal component analysis explained e total variance. Nevertheless, classification with the lace almost all Greek tomato landraces studied into six . 2). Two landraces, “Santorini” and “GR065/93”, formed roup due to that many of their plants had small and as well as short stem internode length and low height truss. A group of the landraces “GR60/99”, “GR076/94”, and “GR300/99” could also be clearly differentiated est due to many determinate and semi-determinate e plantswithheart, cylindrical andpyriform fruit shapes hown). Two big groups, having 9 and 13 landraces placed closely in the plot (Fig. 2). While the landraces and “GR11/03” stood apart and did not form a group, der of two small groups had small differences between e landraces were characterized as having many plants and heavier fruits than the landraces in other groups own). 142 P.J. Terzopoulos, P.J. Bebeli / Scientia Horticulturae 126 (2010) 138–144 Table 3 Phenotypic diversity values (Hp) of 34 Greek tomato landraces using all morpholog- ical traits studied. Landraces Phenotypic diversity (Hp) Mean Min Max GR243/99 0.52 0 0.71 GR240/99 0.49 0 0.68 GR14/99 0.49 0.14 0.66 GR242/99 0.48 0.03 0.66 GR16/99 0.48 0.11 0.67 GR5/01 0.47 0 0.65 GR076/94 0.43 0.11 0.67 GR193/99 0.42 0 0.68 GR090-093/92 0.42 0 0.67 GR301/99 0.42 0 0.66 GR076/92 0.41 0 0.65 VE027/83 0.41 0 0.66 GR074-076/92 0.40 0 0.67 Omospondia 0.40 0 0.67 Pastra 0.40 0 0.67 GR60/99 0.40 0 0.67 GRC 091/92 0.40 0 0.68 GR25/99 0.40 0 0.65 VE026/83 0.39 0 0.69 GR069/94 0.39 0 0.67 GR11/03 0.38 0 0.66 GR279/99 0.38 0 0.64 VE029/83 0.38 0 0.66 GR300/99 0.37 0 0.67 GR008/82 0.37 0 0.63 VE035/83 VE030/83 VE033/83 VE025/83 VE023/83 GR092/93 GR012/82 Santorini GR065/93 HSDa Mean a HSD (Tuke 4. Discussi 4.1. Charac The resu and HT valu Fig. 2. Princip 36 morpholog of Greek tomato landraces. Variation has been observed among and within Greek landraces in almost every trait evaluated. This is in agreement with the suggestion by Dr. Stavropoulos, Curator of the Hellenic Gene Bank, that Greek farmers have kept part of the initial toma Mavrona e concurrent et al., 2009 observed am ated. In Greec soil and irr landraces u Greek farm with consu developme different an Grandillo e quence of t related to fl on the fruit ability for t was observ 2010). Besides developme aract caro ne c o ora ycop sens 0.37 0 0.66 0.36 0 0.7 0.36 0 0.64 0.34 0 0.6 0.34 0 0.65 0.34 0 0.7 0.34 0 0.66 The ch tion of Lycope leads t sis of l is less 0.29 0 0.66 0.24 0 0.64 0.17 0.4 y-Kramer Honestly Significant Difference) at P≤0.05. on terization of Greek landraces lts of the present study, both characterization results es (Table 2), revealed a rich diversity in the collection al Component Analysis (PCA) for 34 Greek tomato landraces based on ical traits. The six resulted groups are indicated in separate circles. lycopene b summer te large perce collection. The pres local comm transportat are many f condition o Large size o to leakage o In general, fruit quality low levels The presen aroma and heavy fruits exploited in for the rece 4.2. Structu The prop (GST) depen cific social other rando Self-pollina population and Brown ues were self-pollina phological v from Moroc the present to diversity andhelped to expand it in landraces (Traka- t al., 2002). The result of the present work was also with the findings of our previous study (Terzopoulos ) in that a significant amount of heterogeneity was ong and within the 14 Greek tomato landraces evalu- e, tomato landraces are cultivated in small fields where igation conditions are easily controlled. Farmers grow sually for their own needs for food. It seems that ers are still interested in characteristics associated mption. They give priority to the amelioration of fruit nt through introducing andmaintaining genotypeswith d more attractive fruit characteristics. According to t al. (1999), the influence on fruit size is often a conse- he influence on fruit shape. Fruit size and shape are also oral traits (Brukhin et al., 2003). The focus of farmers shape and size resulted in the presence of high vari- he fruit related traits in our collection. A similar result ed in Italian tomato landraces (Mazzucato et al., 2008, farmers, climatic conditions can also play a role in the nt of landraces. High temperatureswill affect fruit color. eristic red color of tomato fruit results from a combina- tenoid pigments, particularly lycopene, and b-carotene. ontent accounts for the fruit redness, while b-carotene nge color (Grierson and Kader, 1986). The biosynthe- ene is sensitive to the temperature, while b-carotene itive. Temperatures above 30 ◦C can strongly inhibit iosynthesis (Steven and Rick, 1986). Therefore, high mperatures in Greece might be the main reason why a ntage of orange fruit color was observed in the landrace ent collection included fruits that were appropriate for unities’ flavor demands rather than for long distance ion and long shelf life. In the landraces evaluated, there ruits with large size of blossom end scar having open r with stellate or irregular shape of pistil scar shape. f blossom end scar can reduce the value of the fruit due f juice and/or enhancing pathogen (Elkind et al., 1990). the fruit weight is negatively correlated with several traits (Lecomte et al., 2004). Large fruits tend to have of lycopene and high levels of pH (Chen et al., 1999). t collection could be a promising resource for flavor, quality variability since it contained a small number of . The variability regarding tomato quality factors can be tomato breeding schemes since consumers complain nt varieties due to lack of flavor (Causse et al., 2001). re of phenotypic variations ortion of phenotypic variation among plant landraces ds on the mating system. It is also related to the spe- and economical situation of the collection origin, or m factors (Hamrick and Godt, 1997; Bellon et al., 2003). ted species often present considerable population-to- variation in intra-population genetic diversity (Schoen , 1991). Using morphological traits, various GST val- observed in collections of self-pollinated or mainly ted species. Dje et al. (1998) found a significant mor- ariation (62.7%) among the fields of sorghum landraces co whereas 37.3% of variation was within the fields. In study, a low value of the mean GST (0.21) was observed P.J. Terzopoulos, P.J. Bebeli / Scientia Horticulturae 126 (2010) 138–144 143 among the Greek tomato landrace populations (Table 2). However, 21 of the traits studied had HT values higher or equal than 0.5, and four had values between 0.45 and 0.49 (Table 2). Furthermore, 13 of the traitswere found to have relatively ‘high’ to ‘medium’ H¯s values (higher or e the mean p Our results observed w to fruit. A lo having low (1999), Bell a constant e farmers. Sin ture of land crosses (Cle landrace va genotypes w from introd In the p had similar landrace ac landraces, “ agreement pollinated a wheat land that pheno of 87 land Labuschagn from 0.12 t out of 44 ba 4.3. Traits c Estimati sity is impo the landrac In the prese most freque not the sam these traits lowest valu within indiv were often stem pubes had the low tively (Tabl farmers’ pro tivated in sm controlled, improveme result in low low values (i.e. determ type). There tion pressu High hetero willingness needs, e.g. t farmers wh shape for d flattened an (informatio the fruit sha landrace. Si tative of an (2010) obse the same plant growth type, 20% of them had plants with flat fruit shape, 56% had round fruit shape and 24% had ovoid one. Within each landrace, the observed variability of fruit traits might increase by the continuous involvement of genotypes with different and ttrac the e co y of f ndra PCA ed in ed o high of int . In o ted s izing esul e bi e lan prev tion card tudy ) su n bet toma ntary how pop and tom hich rs, w of ge 2008 t stu ces h y. clus ek to that adde tion rm, s tra omm ructu resu d far lity re is tion wled gram oject ces - rain iewi qual to 0.45) and almost all traits had the big range of henotypic diversity within each landrace (Hs) (Table 2). suggested that the major part of phenotypic variation as within the landraces, especially in the traits related wGST is not somethingunusual, as oneof the reasons for GST could be the seed exchange among farmers. Dje et al. on et al. (2003), and Badstue et al. (2007) had observed xchange of local and exotic seed and technology among ce a landrace (or a farmer’s local variety) may be a mix- races adapted from old or modern varieties and their veland et al., 2000), the high values observed within riations could be due to the continuous involvement of ith different and more attractive fruit characteristics uction of new material by farmer’s practices. resent work, we observed that most of the landraces values of the mean phenotypic diversity within each ross all traits (H¯p) except for those two small fruited Santorini” and “GR065/93” (Table 3). This result was in with the studies of other researchers concerning self- nd/or mainly self-pollinated species. Studying durum races from Syria, van Hintum and Elings (1991) found typic diversity ranged from 0.22 to 0.68, while 68 out races had the values from 0.35 to 0.55. Assefa and e (2004) observed that phenotypic diversity ranged o 0.57 in barley landraces from Ethiopia. Twenty-six rley landraces had values from 0.35 to 0.55. ontribute the most to landrace heterogeneity ng the magnitude of inter- and intra-population diver- rtant.Understandingwhich traits contribute themost to e heterogeneity and thus adaptability is also invaluable. nt study, the most heterogeneous traits that occurred ntly within individual landraces of the collection were e for the majority of landraces. However, in most cases were related to fruit form (Table 2). Traits with the e of phenotypic diversity that occurred most frequently idual landraces were mainly vegetative traits and they the same for the majority of landraces. For example, cence density, foliage density and plant growth type est Hs values in 28, 22 and 21 of the landraces, respec- e 2). The above findings could be explained by Greek duction trends. Since Greek tomato landraces are cul- all fieldswheremany environmental factors are easily farmers play a major role in selection and landrace nt. Strong selection pressure on a trait will generally heterogeneity (Bernardo, 2002). In the present study, of heterogeneity were observed in vegetative traits inate plant growth type or indeterminate plant growth fore, it seems that the determinative factor for selec- re was the cultivation system that Greek farmers used. geneity of fruit traits probably resulted from farmers’ of using the same landrace to cover different consumer omatoes for salad and tomatoes for juice. For example, o grow “Santorini” used to maintain three types of fruit ifferent needs: rounded fruits for juice and preserves, d slightly flattened fruits for tomatoes dried in the sun n from our exploration in Thira Island). As the result, pewas themost heterogeneous traitwithin “Santorini” milarly, in the study of 25 tomato accessions represen- Italian landrace “A pera Abruzzese”, Mazzucato et al. rved that, although all of them produced plants with morea bining practic geneit 4.4. La The classifi explain to the levels lection evalua by util vious r into on pare th of our correla the Jac vious s p>0.05 relatio of the pleme traits s similar traits Greek gin, w marke sisted Bebeli, presen landra geneit 5. Con Gre ation value- popula fruit fo geneou were c istic st be the sure an variabi structu in addi Ackno Pro The pr resour tional T for rev tive fruit characteristics throughseedsexchanges. Com- introduction of new material with farmer’s selection uld result in the derivation of higher value of hetero- ruit traits compared to other traits. ces classification analysis showed that the majority of landraces were to two groups (Fig. 2). However, the first three axes nly 43.6% of the total variance. This could be attributed levels of intra-population heterogeneity and the low er-population variability observed in the landrace col- ur previous study (Terzopoulos and Bebeli, 2008), we ome of the tomato landraces from the present work inter-simple sequence repeat (ISSR) markers. Our pre- t showed that the landrace populations were grouped g sub-cluster (Terzopoulos and Bebeli, 2008). To com- drace classification result of the present studywith that ious one, the Mantel test was performed between the matrix of the PCA analysis from the present study and similarity matrix derived from ISSR data of our pre- (Terzopoulos and Bebeli, 2008). The result (r=0.007, ggested that there was not statistically significant cor- ween the two matrices. Nonetheless, the classification to landraces with morphological traits could be com- to that of using ISSR data. The study of morphological ed that the majority of Greek tomato landraces have ulation structures, i.e. having homogeneous vegetative heterogeneous fruit traits. This suggested that some ato landraces could have had the same genetic ori- is in agreement with our previous study. Using ISSR e observed that seven of the landraces analyzed con- netically closely related morphotypes (Terzopoulos and ). These findings, in addition to the other results of the dy, showed that the appropriate strategy for efficient andling should focus on their intra-population hetero- ion mato landraces showed remarkable phenotypic vari- is promising for their use in the manipulation of d traits in tomato improvement. Landraces’ intra- heterogeneity was mainly due to the traits related to such as inflorescence and fruit traits. The most homo- its within each landrace were vegetative traits, which on for most of the landrace populations. The character- re of landrace heterogeneity in the present study could lt of combination of cultivation system selection pres- mers’ willingness of exchanging seeds and maintaining in fruit traits. The knowledge of landrace phenotypic essential for its utilization in future breeding schemes to its use in sustainable agriculture. gements “Environment-Pythagoras I” supported this work. is co-funded by European Social Fund and national “Operational Programme for Education and Initial Voca- ing”. “Education” II. The authors thankDr. J.P. Gustafson ng the paper, Dr. Konstantinos Aliferis, Susan Coward, 144 P.J. Terzopoulos, P.J. Bebeli / Scientia Horticulturae 126 (2010) 138–144 Christopher Sue, Georgia Paziotou, George Apergis and Konstanti- nos Triantafillou for their contribution. References Assefa, A., Labuschagne, M.T., 2004. Phenotypic diversity in barley (Hordeum vulgare L.) landraces from north Shewa in Ethiopia. Biodivers. Conserv. 13, 1441–1451. Badstue, L.B., Bellon, M.R., Berthaud, J., Ramirez, A., Flores, D., Juarez, X., 2007. The dynamics of farmers’ maize seed supply practices in the Central Valleys of Oax- aca. Mexico World Dev. 35, 1579–1593. Bechere, E., Belay, D., Mitiku, D., Merker, A., 1996. Phenotypic diversity of tetraploid wheat landraces from northern and north-central regions of Ethiopia. Hereditas 124, 165–172. Bellon, M.R., Berthaud, J., Smale, M., Aguirre, J.A., Taba, S., Aragon, F., Diaz, J., Castro, H., 2003. Participatory landrace selection for on farm conservation: an exam- ple from the Central Valleys of Oaxaca. Mexico Genet. Resour. Crop Evol. 50, 401–416. Bernardo, R., 2002. Breeding for Quantitative Traits in Plants. Stemma Press, Wood- burn, MN. Brukhin, V., Hernould,M., Gonzalez, N., Chevalier, C.,Mouras, A., 2003. Flower devel- opment schedule in tomato Lycopersicon esculentum cv. sweet cherry. Sex Plant Reprod. 15, 311–320. Carbonell-Bar pounds by tomatoes. Carelli, B.P., Ge among Bra revealed b Causse, M., Sa of organol attributes. Chen, F.Q., Fo of QTLs tum× Lyco species. M Cleveland, D.A ing farmer Cooper, H.D., S an overvie Genetic Ba Dje, Y., Ater, M allozyme Resour. Cr Dje, Y., Forcio tion genet allozyme a Elkind, Y., Galp action of t Fischbeck, G., O.H., Mars Cambridge García-Martín ation of am for tomato tional cult Goncalves, L.S Heirloom logical, ag Genet. Mo Grandillo, S., K ral variatio Grierson, D., K J. (Eds.), T 241–280. Hamrick, J.L., 37, 26–30. Hawtin, G., Iwanaga,M., Hodgkin, T., 1997. Genetic resources in breeding for adapta- tion. In: Tigerstedt, P.M.A. (Ed.), Adaptation in Plant Breeding. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 277–288. Kuehl, R.O., 2000. Design of Experiments: Statistical Principles of Research Design and Analysis. Duxbury, Thomson Learning, Pacific Grove, CA, USA. Labate, J.A., Robertson, L.D., Baldo, A.M., 2009. Multilocus sequence data reveal extensive departures from equilibrium in domesticated tomato (Solanum lycop- ersicum L.). Heredity 103, 257–267. Lecomte, L., Saliba-Colombani, V., Gomez-Jimenez, M.C., Duffe, P., Buret, M., Gausse, M., 2004. Fine mapping of QTLs of chromosome 2 affecting the fruit architecture and composition of tomato. Mol. Breeding 13, 1–4. Mazzucato, A., Papa, R., Bitocchi, E., Mosconi, P., Nanni, L., Negri, V., Picarella, M.E., Siligato, F., Soressi, G.P., Tiranti, B., Veronesi, F., 2008. Genetic diversity, struc- ture and marker-trait associations in a collection of Italian tomato (Solanum lycopersicum L.) landraces. Theor. Appl. Genet. 116, 657–669. Mazzucato, A., Ficcadenti, N., Caioni, M., Mosconi, P., Piccinini, P., Sanampudi, V.R.R., Sestili, S., Ferrari, V., 2010. Genetic diversity and distinctiveness in tomato (Solanum lycopersicum L.) landraces: the Italian case study of ‘A pera Abruzzese’. Sci. Hortic. 125, 55–62. Nei, M., 1973. Analysis of gene diversity in subdivided populations. Proc. Natl. Acad. Sci. U.S.A. 70, 3321–3323. Pagnano, M., Gauvreau, K., 2000. Principles of Biostatistics. Duxbury, CA. Rick, C.M., 1976. Tomato Lycopersicon esculentum (Solanaceae). In: Simmonds, N.W. (Ed.), Evolution of Crop Plants. Longman Group, London, UK, pp. 268–273. .J., 1998. NTSYS-Numerical Taxonomy and Multivariate Analysis System. er Pub itute I D.J., Br effect . Acad P.H.A., Franci M.A., ds.), T 09. , S.D. tic po ulos, k tom ulos, P ity in ulos, P ulation Descr avron . Des . In: P ding, tum, T in cro . (Eds hester tum, T of Syr hytica S., 199 , MA. s, C., cted b morp lentum .C., 19 127– .C., 20 retica garde rachina, A.A., Agusti, A., Ruiz, J.J., 2006. Analysis of flavor volatile com- dynamic headspace in traditional and hybrid cultivars of Spanish Eur. Food Res. Technol. 222, 536–542. rald, L.T.S., Grazziotin, F.G., Echeverrigaray, S., 2006. Genetic diversity zilian cultivars and landraces of tomato Lycopersicon esculentum Mill. y RAPD markers. Genet. Resour. Crop Evol. 53, 385–400. limba-Colombani, V., Lesschaeve, I., Buret, M., 2001. Genetic analysis eptic quality in fresh market tomato. 2. Mapping QTLs for sensory Theor. Appl. Genet. 102, 273–283. olad, M.R., Hyman, S., St Clair, D.A., Beelaman, R.B., 1999. Mapping for lycopene and other fruit traits in a Lycopersicon esculen- persicon pimpillenifolium cross and comparison of QTLs across tomato ol. Breeding 5, 283–299. ., Soleri, D., Smith, S.E., 2000. A biological framework for understand- s plant breeding. Econ. Bot. 54, 377–394. pillane, C., Hodgkin, T., 2000. Broadening the genetic base of crops: w. In: Cooper, H.D., Spillane, C., Hodgkin, T. (Eds.), Broadening the se of Crop Production. CABI Publishing, New York, NY, USA, pp. 1–23. ., Lefebvre, C., Vekemans, X., 1998. Patterns of morphological and variation in sorghum landraces of Northwestern Morocco. Genet. op Evol. 45, 541–548. li, D., Ater, M., Lefebvre, C., Vekemans, X., 1999. Assessing popula- ic structure of sorghum landraces fromNorth-WesternMorocco using nd microsatellite markers. Theor. Appl. Genet. 99, 157–163. er, O.B., Scott, J.W., Kedar, N., 1990. Genotype by environment inter- omato blossom-end scar size. Euphytica 50, 91–95. 1989. Evaluation of cereals in Europe. In: Brown, A.H.D., Frankel, hall, D.R., Williams, J.T. (Eds.), The Use of Plant Genetic Resources. University Press, Cambridge, Great Britain, pp. 197–211. ez, S., Andreani, L., Garcia-Gusano, M., Geuna, F., Ruiz, J.J., 2006. Evalu- plified fragment length polymorphism and simple sequence repeats germplasm fingerprinting: utility for grouping closely related tradi- ivars. Genome 49, 648–656. .A., Rodrigues, R., Do Amaral Jr., A.T., Karasawa, M., Sudre, C.P., 2009. tomato gene bank: assessing genetic divergence based on morpho- ronomic and molecular data using a Ward-modified location model. l. Res. 8, 364–374. u, H.M., Tanksley, S.D., 1999. Identifying the loci responsible for natu- n in fruit size and shape in tomato. Theor. Appl. Genet. 99, 978–987. ader, A.A., 1986. Fruit ripening and quality. In: Atherton, J.G., Rudich, he Tomato Crop. Chapman and Hall Ltd, London, Great Britain, pp. Godt, M.J.W., 1997. Allozyme diversity in cultivated crops. Crop Sci. Rohlf, F Exet SAS Inst Schoen, and Natl Sneath, San Steven, J. (E 35–1 Tanksley gene Terzopo Gree Terzopo gene Terzopo pop Tomato Traka-M 2002 toes Bree van Hin sity E.A.V Chic van Hin sity Eup Weir, B. land William dete poly escu Zeven, A 104, Zeven, A theo and l., New York. nc., 2007. JMP/Sales Department. Cary, NC, USA. own, A.H.D., 1991. Intraspecific variation in population gene diversity ive population size correlates with the mating system in plants. Proc. . Sci. U.S.A. 88, 4494–4497. Sokal, R.R., 1973. Numerical Taxonomy. W.H. Freeman and Company, sco. Rick, C.M., 1986. Genetics and breeding. In: Atherton, J.G., Rudich, he Tomato Crop. Chapman and Hall Ltd, London, Great Britain, pp. , McCouch, S.R., 1997. Seed banks and molecular maps: unlocking tential from the wild. Science 277, 1063–1066. P.J., Bebeli, P.J., 2008. DNA and morphological diversity of selected ato (Solanum lycopersicum L.) landraces. Sci. Hortic. 116, 354–361. .J., Kaltsikes, P.J., Bebeli, P.J., 2008. Determining the sources of hetero- Greek faba bean local populations. Field Crop Res. 105, 124–130. .J.,Walters, S.A., Bebeli, P.J., 2009. EvaluationofGreek tomato landrace s for heterogeneity of horticultural traits. Eur. J. Hortic. Sci. 74, 24–29. iptors, 1996. IPGRI. Tomato Descriptors, Rome, Italy. a, K., Tasios, V., Palatos, G., Koutsos, Th., Stavropoulos, N., Mellidis, V., cription, evaluation and use of non-watering type landraces of toma- roceeding of the 9th Greek Scientific Society of Genetics and Plant Thessaloniki, Greece (in Greek). h.J.L., 1995. Hierarchical approaches to the analysis of genetic diver- p plants. In: Hodgkin, T., Brown, A.H.D., van Hintum, Th.J.L., Morales, .), Core Collections of Plant Genetic Resources. John Wiley and Sons, , UK, pp. 23–34. h.J.L., Elings, A., 1991. Assessment of glutenin and phenotypic diver- ian durum wheat landraces in relation to their geographical regions. 55, 209–215. 0. Genetic Data Analysis. Sinauer Associates, Inc. Publishers, Sunder- St. Clair, D., 1993. Phenetic relationships and levels of variability y restriction fragment length polymorphism and random amplified hic DNA analysis of cultivated and wild accessions of Lycopersicon . Genome 36, 619–630. 98. Landraces: a review of definitions and classifications. Euphytica 139. 02. Traditional maintenance breeding of landraces: 2. Practical and l considerations on maintenance of variation of landraces by farmers ners. Euphytica 123, 147–158. Phenotypic diversity in Greek tomato (Solanum lycopersicum L.) landraces Introduction Materials and methods Plant material and experimental design Measurements Data analysis Results Characterization of Greek tomato landraces collection Vegetative traits Inflorescence traits Traits related to the fruit color and size Traits related to the fruit form Phenotypic diversity of tomato landrace collection Total phenotypic diversity Phenotypic diversity among landraces Phenotypic diversity within individual landraces Mean phenotypic diversity of each landrace Landraces classification Discussion Characterization of Greek landraces Structure of phenotypic variations Traits contribute the most to landrace heterogeneity Landraces classification Conclusion Acknowledgements References