OPCE-STR.vp Acta Bot. Croat. 71 (1), 31–50, 2012 CODEN: ABCRA 25 ISSN 0365-0588 eISSN 1847-8476 Classification of mesic grasslands and their transitions of South Transdanubia (Hungary) ATTILA LENGYEL1,*, DRAGICA PURGER2, JÁNOS CSIKY3 1 Department of Plant Systematics, Ecology and Theoretical Biology, Eötvös Loránd University, Pázmány P. s. 1/C, H-1117 Budapest, Hungary 2 South Transdanubian Environment Protection Institute, Köztársaság tér 7, H-7623 Pécs, Hungary 3 Department of Plant Taxonomy and Geobotany, University of Pécs, Ifjúság u. 6, H-7624 Pécs, Hungary Abstract – Relevés from meadows and pastures of South Transdanubia (Hungary) are evaluated by clustering and ordination methods. The relevé selection focused on the Arrhenatheretalia order but its transitions towards other types were also included. The groups of relevés are delimited and described according to differential, dominant and con- stant species. Ecological conditions of the groups were compared using indicator values. Nine groups were distinguished, four of them belonging strictly to the order Arrhenathe- retalia. Each alliance of Arrhenatheretalia presented in the study area (Cynosurion, Arrhe- natherion) was represented by two groups. Groups from these two alliances are separated along a light gradient, while groups of the same alliance differ in nutrient values. Within Cynosurion, the nutrient-poor group cannot be identified unambiguously as any syntaxa previously known from Hungary. The nutrient-rich Cynosurion meadows are similar to Lolio–Cynosuretum, however, they show a stronger relationship with wet meadows. Within Arrhenatherion, Pastinaco–Arrhenatheretum is recognised as a hay meadow of nutrient-rich soils. The other meadow type is similar to Filipendulo–Arrhenatheretum, thus raising syntaxonomical problems. There are transitional groups towards semi-dry and wet meadows, one dynamic phase and one outlier group among the other five clusters. Keywords: vegetation, meadows, pastures, indicators, phytosociology, syntaxonomy Introduction Mesophilous grasslands are among the most species-rich herbaceous vegetation types in Hungary. Their presence is linked with the traditional agricultural system but with the in- tensification of farming, the area of these valuable habitats is decreasing. It is an urgent ACTA BOT. CROAT. 71 (1), 2012 31 * Corresponding author, e-mail: lengyelat@caesar.elte.hu Copyright® 2012 by Acta Botanica Croatica, the Faculty of Science, University of Zagreb. All rights reserved. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:05 Color profile: Disabled Composite 150 lpi at 45 degrees challenge to reveal and document the diversity of mesic grasslands in order to help their conservation. Despite their conservation importance, mesophilous meadows are a neglected category in the Hungarian vegetation classification. According to the current syntaxonomy (BORHIDI 2003), mesophilous meadows are included into the Arrhenatheretalia order. This order is represented by three alliances in Hungary: Arrhenatherion comprising lowland and colline hay meadows, Cynosurion covering lowland to montane pastures and Phyteumo–Trisetion comprising montane hay meadows; the last one is certainly not present in South Trans- danubia. Previously, works aimed at synthesizing the variability of the Arrhenatheretalia order (SOÓ 1971, 1973; BORHIDI 2003) were based on data collected unevenly from local or subregional studies (e.g. JUHÁSZ-NAGY 1959; MÁTHÉ and KOVÁCS 1960; JEANPLONG 1960; KOVÁCS 1994, 2002; LÁJER 2002) without the application of more objective methods. There have been a few papers dealing with the meadows of the southern part of Transdanubia. However, they either only focus on the description of the stands of a restricted area or meadow type (DÉNES 1997, LÁJER 2002) or their sampling and analytical methodology is unclear (HORVÁT 1962, 1972). There are more comparative publications about other meadow types which have transi- tions towards mesic meadows. A classification of semi-dry grasslands has recently been prepared by ILLYÉS et al. (2009) either exclusively for Hungary, or together with other Cen- tral European countries as well (ILLYÉS et al. 2007). The classification of wet meadows of Hungary has been clarified by BOTTA-DUKÁT et al. (2005). In a few neighbouring countries, the syntaxonomy of mesophilous meadows has been investigated more thoroughly. The mesic meadows of Austria have been classified by MUCINA et al. (1993) and ELLMAUER (1994). In Slovakia, a classification has been proposed by UHLIAROVÁ et al. (2007) and JANI[OVÁ et al. (2007) and recently refined by focusing on the Western Carpathians by ROZBROJOVÁ et al. (2010). In Romania meadow communities are reviewed by COLDEA et al. (1991) and SANDA et al. (1999). Short descriptions of Croatian meadow associations are presented in TRINAJSTI] (2008), while recent broad-scale classification studies have been published by STAN^I] (2000, 2008). There are papers containing reviews of previously sur- veyed meadows and pastures in Serbia (e.g. KOJI] et al. 2004, BLA@EN^I] et al. 2005), yet they have not been evaluated nor have their results been synthesised. Some brief analyses are available for Slovenia (e.g. ^ARNI 2001, ZELNIK 2007, ZELNIK and ^ARNI 2008). Our aim is to prepare the classification of the meadows belonging to the Arrhenathe- retalia order in South Transdanubia and to relate our results to the existing syntaxonomical system. We intend both to differentiate among associations of this order and to delimit them from types of other orders. Materials and methods Study area The study area is situated in the southern part of Hungary (Fig. 1). The major, northern part of the area belongs to the South Transdanubian geographic district, while the southern lowlands (e.g. Drava Plain) are parts of the Great Hungarian Plain (DÖVÉNYI 2010). For simplicity, hereafter we use the term 'South Transdanubia' for the entire study area, includ- ing its plain subregion. The southern border of the examined area is the Drava River and the 32 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:06 Color profile: Disabled Composite 150 lpi at 45 degrees national border. The other limits are defined according to the distribution of mesophilous meadows (MOLNÁR et al. 2008), which approximately match the 700 mm isohyet in the east and in the north, and are delimited by the dominance of sand substratum in the west (Belsõ-Somogy). At the northern part of the study range there are the Mecsek Mts, which consist of several kinds of bedrock (mainly Jurassic and Triassic limestone). This insular mountain is surrounded by hills covered by young (Pannonian, Pleistocene) sediments. South from this colline-montane landscape lies the alluvial plain of Drava River. The high- est peak of the area is 682 m in the Mecsek Mts, while the lowest point is 87 m on the Drava Plain. There is a climatic continentality gradient increasing from the northwest towards the southeast (LOVÁSZ 1977). The study range is influenced by the sub-Mediterranean climate the most strongly in Hungary. The potential vegetation of the sites where field sampling was carried out is deciduous forest, dominated by Quercus spp., Carpinus betulus and Fagus sylvatica. Some sites located along the Drava are on drained felled alluvial forests with Quercus robur, Fraxinus angustifolia and Carpinus betulus. Vegetation data Field sampling was carried out from mid-May to late-June between 2004 and 2009. In total, 211 phytosociological relevés were collected according to the Zürich-Montpellier method (DENGLER et al. 2008). Plots were located in vegetation stands corresponding with the mesic meadow categories (E1, E2, E34) of the Hungarian Habitat Guide (BÖLÖNI et al. 2003) which is also used for the recently finished country-wide habitat mapping project (MOLNÁR et al. 2007). According to this system, the category of mesic meadows is some- what wider than that of the syntaxonomic system. In terms of syntaxonomy, this allows the inclusion of Violion caninae grasslands and transitional stands towards Brometalia erecti and Molinietalia meadows, beyond Arrhenatheretalia which forms the core of this vegeta- ACTA BOT. CROAT. 71 (1), 2012 33 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA Fig. 1. The study area U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:07 Color profile: Disabled Composite 150 lpi at 45 degrees tion type. During sampling, the authors preferred stands under regular and traditional man- agement (grazing, mowing) and avoided obvious within-plot heterogeneity in factors affecting vegetation (e.g. small-scale disturbances). However, finding 'typical' stands ac- cording to the previously described associations was not aimed at since we wanted to repre- sent the total variation as much as possible. Relevés were taken in 4×4 or 5×5 m2 plots which are the recommended sizes for this vegetation type (LÁJER et al. 2008). Species cover was estimated on a percentage scale but they were log-transformed before analysis. Poorly distinguishable taxa were merged (a list is available from the authors upon request). The data set of 211 relevés was stratified geographically (KNOLLOVÁ et al. 2005) into ca. 3×3 km2 grid squares and 5 relevés were selected from each cell using the heterogeneity-con- strained random resampling method (HCR, LENGYEL et al. 2011). The number of trials in HCR resampling was 10 000 and the complement of the Ruzicka index (which is the Jaccard index generalized to abundance data, PODANI 2000) was used for calculating pair-wise dissimilarities between plots. The resampled data set consisted of 155 relevés. Data analysis Before classification, a noise elimination procedure was applied according to BOTTA- -DUKÁT et al. (2005). The data set was analysed by principal coordinates analysis (applying the Ruzicka index again) and the scores of the relevés on the significant ordination axes were used as input variables for classification. The first 19 axes were considered significant because their eigenvalues exceeded the expectations based on the broken stick model (LEGENDRE and LEGENDRE 1998). Relevé coordinates (155 relevés as objects, each with 19 scores as variables) were clas- sified based on their pair-wise Euclidean distances using Ward's agglomerative method (PODANI 2000). The optimal level of the clustering was determined by the Optimclass 1 method (TICHÝ et al. 2010) using p<10–3 threshold for characteristic species. Correspond- ingly, the dendrogram was cut at the nine-cluster level and relevé groups were described by their differential, constant and dominant species (DENGLER et al. 2008). Species with high affinity to a cluster were treated as differential species rather than truly characteristic spe- cies because in a wider comparison fidelity values may change significantly. Species fideli- ties to clusters were measured by the phi coefficient (CHYTRÝ et al. 2002) adjusted to equal group sizes (TICHÝ and CHYTRÝ 2006) and those reaching phi = 0.2 and p<10–3 significance by Fisher's exact test were considered differential. Species occurring in at least 80% of the plots of a cluster were listed among the constant species. A species was considered domi- nant if it reached 30% cover in at least 10% of the relevés of a given cluster. Unweighted mean ecological indicator values (BORHIDI 1995, HORVÁTH et al. 1995) were calculated for each plot and clusters were compared on box-and-whisker plots. Relationships of clusters were also examined by ordination. Canonical analysis of prin- cipal coordinates (ANDERSON and WILLIS 2003) with cluster memberships as nominal ex- planatory variables was applied in order to concentrate more on cluster separation. Ordina- tion results are shown on spider plots which represent ordination and classification together. Each relevé is linked to the centroid of its group on the scatter plot, thus the dia- gram resembles spiders with 'legs' starting from the centroid (the 'body') and ending in the relevés' positions. Mean indicator values of relevés were passively projected onto the dia- gram. The goodness of fit of the projection was estimated by a permutation test. 34 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:07 Color profile: Disabled Composite 150 lpi at 45 degrees All analyses were performed in R Environment (version 2.10.0, R DEVELOPMENT CORE TEAM 2009), using vegan (OKSANEN et al. 2010) and cluster (MAECHLER et al. 2005) pack- ages. The nomenclature of plant species follows KIRÁLY (2009). Results Description of groups The dendrogram of relevés is shown in figure 2. Differential, constant and dominant species are summarized in tables 1, 2 and 3 respectively. Group 1 Number of relevés: 31; mean species number per relevé: 41.0. Most relevés of Group 1 originate from moderately-grazed extensive pastures from hilltops and hillsides. In the syntaxonomic system, this type would be placed into the Cynosurion alliance within the Arrhenatheretalia order. Beyond typical elements of mesic meadows, these stands are also characterised by species tolerating nutrient-poor and dry soil (e.g. Agrostis capillaris, Anthoxanthum odoratum, Festuca pseudovina & valesiaca, F. rubra, Hypochoeris radicata). Group 2 This group consists of two outlier relevés, and has therefore not been interpreted. Group 3 Number of relevés: 24; mean species number per relevé: 37.0. Relevés classified to this group were taken in stands similar to but more disturbed than those of Group 1. The most common forms of disturbance are overgrazing, drying and soil erosion with additional leaching. Some of the stands are probably old fields which are ir- regularly or just recently mown or grazed. Accordingly, there are many weeds among the differential species of this group (e.g. Convolvulus arvensis, Trifolium arvense). We con- sider this type a dynamic state of other, more regularly managed mesic meadows, mostly of Cynosurion types. ACTA BOT. CROAT. 71 (1), 2012 35 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA Fig. 2. Dendrogram of relevés U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:08 Color profile: Disabled Composite 150 lpi at 45 degrees 36 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. Tab. 1. Differential species of the groups with the values of phi coefficients. Group: 1 2 3 4 5 6 7 8 9 Centaurium erythraea 0.49 –- –- –- –- –- –- –- –- Leontodon autumnalis 0.37 –- –- –- –- –- –- –- –- Holcus lanatus 0.36 –- –- –- –- –- –- –- –- Festuca pseudov. & valesiaca 0.36 –- –- –- –- –- –- –- –- Anthoxanthum odoratum 0.32 –- –- –- –- –- –- –- –- Trifolium dubium 0.31 –- –- –- –- –- –- –- –- Agrostis capillaris 0.31 –- –- –- –- –- –- –- –- Luzula campestris 0.29 –- –- –- –- –- 0.26 –- –- Festuca rubra 0.29 –- –- –- –- –- –- –- –- Daucus carota 0.28 –- –- –- –- –- –- –- –- Cynosurus cristatus 0.28 –- –- –- –- 0.42 –- –- –- Trisetum flavescens 0.28 –- –- –- –- –- 0.27 –- –- Trifolium pratense 0.27 –- –- –- –- –- –- –- –- Picris hieracioides 0.26 –- 0.29 –- –- –- –- –- –- Lotus corniculatus 0.26 –- –- –- –- –- –- –- –- Ranunculus bulbosus 0.26 –- –- –- –- –- –- –- –- Trifolium repens 0.25 –- –- –- –- –- –- –- –- Hypochoeris radicata 0.21 –- –- –- –- –- –- –- –- Leontodon hispidus 0.20 –- –- –- –- –- –- –- –- Centaurea stoebe –- 1.00 –- –- –- –- –- –- –- Danthonia decumbens –- 0.98 –- –- –- –- –- –- –- Peucedanum oreoselinum –- 0.95 –- –- –- –- –- –- –- Sherardia arvensis –- –- 0.55 –- –- –- –- –- –- Moenchia mantica –- –- 0.49 –- –- –- –- –- –- Trifolium arvense –- –- 0.48 –- –- –- –- –- –- Odontites vernus –- –- 0.43 –- –- –- –- –- –- Trifolium campestre –- –- 0.42 –- –- –- –- –- –- Agrimonia eupatoria –- –- 0.41 –- –- –- –- –- –- Convolvulus arvensis –- –- 0.39 –- –- –- –- –- –- Geranium columbinum –- –- 0.38 –- –- –- –- –- –- Ranunculus polyanthemos –- –- –- 0.62 –- –- –- –- –- Festuca rupicola –- –- –- 0.48 –- –- –- –- –- Pimpinella saxifraga –- –- –- 0.44 –- –- –- –- –- Pastinaca sativa –- –- –- 0.44 –- –- –- –- –- Medicago falcata –- –- –- 0.41 –- –- –- –- –- Thymus glabrescens –- –- –- –- 0.93 –- –- –- –- Sanguisorba minor –- –- –- –- 0.88 –- –- –- –- Brachypodium pinnatum –- –- –- –- 0.88 –- –- –- –- Teucrium chamaedrys –- –- –- –- 0.78 –- –- –- –- Trifolium montanum –- –- –- –- 0.76 –- –- –- –- Thesium linophyllon –- –- –- –- 0.76 –- –- –- –- U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:08 Color profile: Disabled Composite 150 lpi at 45 degrees ACTA BOT. CROAT. 71 (1), 2012 37 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA Group: 1 2 3 4 5 6 7 8 9 Vicia tenuifolia –- –- –- –- 0.76 –- –- –- –- Koeleria cristata –- –- –- –- 0.76 –- –- –- –- Centaurea scabiosa –- –- –- –- 0.73 –- –- –- –- Dianthus pontederae –- –- –- –- 0.69 –- –- –- –- Asperula cynanchica –- –- –- –- 0.68 –- –- –- –- Lathyrus latifolius –- –- –- –- 0.61 –- –- –- –- Peucedanum cervaria –- –- –- –- 0.61 –- –- –- –- Anthyllis vulneraria –- –- –- –- 0.61 –- –- –- –- Onobrychis viciifolia –- –- –- –- 0.61 –- –- –- –- Salvia pratensis –- –- –- –- 0.61 –- –- –- –- Plantago media –- –- –- –- 0.59 –- –- –- –- Lolium perenne –- –- –- –- –- 0.63 –- –- –- Agrostis stolonifera –- –- –- –- –- 0.62 –- –- –- Cynodon dactylon –- –- –- –- –- 0.56 –- –- –- Bromus racemosus agg. –- –- –- –- –- 0.50 –- –- –- Carex distans –- –- –- –- –- 0.44 –- –- –- Bellis perennis –- –- –- –- –- 0.43 –- –- –- Verbena officinalis –- –- –- –- –- 0.42 –- –- –- Filipendula vulgaris –- –- –- –- –- –- 0.74 –- –- Betonica officinalis –- –- –- –- –- –- 0.57 –- –- Carex pallescens –- –- –- –- –- –- 0.49 –- –- Campanula patula –- –- –- –- –- –- 0.48 –- –- Sanguisorba officinalis –- –- –- –- –- –- 0.47 –- –- Vicia cracca –- –- –- –- –- –- 0.40 –- –- Lathyrus pratensis –- –- –- –- –- –- 0.40 –- –- Cruciata glabra –- –- –- –- –- –- 0.40 –- –- Knautia drymeia –- –- –- –- –- –- 0.40 –- –- Rumex acetosa –- –- –- –- –- –- 0.38 –- –- Saxifraga bulbifera –- –- –- –- –- –- 0.36 –- –- Colchicum autumnale –- –- –- –- –- –- 0.31 –- –- Carex hirta –- –- –- –- –- –- 0.29 –- –- Crepis biennis –- –- –- –- –- –- –- 0.28 –- Ranunculus acris –- –- –- –- –- –- –- 0.26 0.40 Ranunculus repens –- –- –- –- –- –- –- –- 0.71 Symphytum officinale –- –- –- –- –- –- –- –- 0.68 Cirsium canum –- –- –- –- –- –- –- –- 0.64 Sonchus arvensis –- –- –- –- –- –- –- –- 0.60 Lychnis flos-cuculi –- –- –- –- –- –- –- –- 0.56 Lysimachia vulgaris –- –- –- –- –- –- –- –- 0.47 Potentilla reptans –- –- –- –- –- –- –- –- 0.46 Poa trivialis –- –- –- –- –- –- –- –- 0.41 Tab. 1. – continued U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:08 Color profile: Disabled Composite 150 lpi at 45 degrees Group 4 Number of relevés: 9; mean species number per relevé: 42.6. This cluster contains transitional relevés towards semi-dry meadows, mainly from low- lands. Syntaxonomically this group should be interpreted as a cluster of transitional stands between Arrhenatheretalia and Brometalia erecti orders. Its stands are mown, grazed or overgrazed, but species-rich. Differential and dominant species are generalists and distur- bance tolerants which occur both on mesic and xeromesic meadows (e.g. Festuca rupicola, Medicago falcata, Pimpinella saxifraga, Ranunculus polyanthemos). 38 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. Tab. 2. Constant species of the groups with percentage frequency values. Group: 1 2 3 4 5 6 7 8 9 Anthoxanthum odoratum 97 100 –- –- –- –- 83 –- –- Trifolium pratense 97 –- –- –- –- –- –- –- 85 Holcus lanatus 94 100 –- –- –- –- –- –- –- Lotus corniculatus 94 –- –- –- –- –- –- –- –- Plantago lanceolata 94 100 83 89 –- –- –- –- –- Trisetum flavescens 94 –- –- –- –- –- 93 –- –- Achillea millefolium agg. 90 100 88 100 80 92 –- 93 85 Galium verum 90 –- 92 100 –- 83 100 83 85 Poa pratensis agg. 87 –- 100 –- 80 83 83 93 100 Daucus carota 84 –- –- –- –- –- –- –- –- Veronica chamaedrys 81 –- –- –- –- –- 90 83 –- Agrostis capillaris –- 100 –- –- –- –- –- –- –- Ambrosia artemisiifolia –- 100 –- –- –- –- –- –- –- Centaurea stoebe –- 100 –- –- –- –- –- –- –- Danthonia decumbens –- 100 –- –- –- –- –- –- –- Helictotrichon pubescens –- 100 –- –- –- –- –- –- –- Peucedanum oreoselinum –- 100 –- –- –- –- –- –- –- Festuca rupicola –- –- –- 100 80 –- –- –- –- Pimpinella saxifraga –- –- –- 89 –- –- –- –- –- Brachypodium pinnatum –- –- –- –- 100 –- –- –- –- Plantago media –- –- –- –- 100 –- –- –- –- Salvia pratensis –- –- –- –- 100 –- –- –- –- Thymus glabrescens –- –- –- –- 100 –- –- –- –- Trifolium montanum –- –- –- –- 100 –- –- –- –- Dianthus pontederae –- –- –- –- 80 –- –- –- –- Sanguisorba minor –- –- –- –- 80 –- –- –- –- Teucrium chamaedrys –- –- –- –- 80 –- –- –- –- Alopecurus pratensis –- –- –- –- –- –- 86 83 92 Betonica officinalis –- –- –- –- –- –- 86 –- –- Lathyrus pratensis –- –- –- –- –- –- 83 –- –- Ranunculus acris –- –- –- –- –- –- –- 80 100 Potentilla reptans –- –- –- –- –- –- –- –- 100 U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:08 Color profile: Disabled Composite 150 lpi at 45 degrees Group 5 Number of relevés: 5; mean species number per relevé: 32.4. Group 5 is a semi-dry meadow group of the Brometalia order present in the Mecsek Mts. Stands are grazed or mown, and harbour far less species than the previous, transitional group. This cluster is characterised by dry grassland species (e.g. Asperula cynanchica, Dianthus pontederae, Teucrium chamaedrys, Vicia tenuifolia). The syntaxonomic position of Groups 4 and 5 should be discussed in more detail in an analysis focusing on both semi-dry and mesic meadows. Group 6 Number of relevés: 12; mean species number per relevé: 30.5. Relevés of Group 6 originate from intensively grazed lowland pastures with nutri- ent-rich and well-watered soils. These plots have the lowest species numbers. The differen- tial species of this group (e.g. Agrostis stolonifera agg., Cynosurus cristatus, Lolium ACTA BOT. CROAT. 71 (1), 2012 39 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA Tab. 3. Dominant species of the groups with percentage frequency values. Group: 1 2 3 4 5 6 7 8 9 Lotus corniculatus 23 —- —- 11 —- —- —- —- —- Trifolium pratense 19 —- —- —- —- —- —- —- 23 Agrostis capillaris 16 50 —- —- —- —- —- —- —- Trisetum flavescens 16 —- —- —- —- —- —- —- —- Festuca rubra 10 —- —- —- —- —- —- 10 —- Holcus lanatus 10 —- —- —- —- —- 17 —- 15 Leontodon hispidus 10 —- —- —- —- —- —- —- —- Danthonia decumbens —- 50 —- —- —- —- —- —- —- Helictotrichon pubescens —- 50 —- —- —- —- 10 —- —- Peucedanum oreoselinum —- 50 —- —- —- —- —- —- —- Festuca rupicola —- —- 25 22 40 —- —- —- —- Poa pratensis agg. —- —- 17 —- —- —- —- 23 15 Arrhenatherum elatius —- —- —- 11 —- —- —- 23 —- Brachypodium pinnatum —- —- —- 11 40 —- —- —- —- Cynodon dactylon —- —- —- 11 —- —- —- —- —- Festuca pratensis —- —- —- 11 —- —- —- 17 69 Medicago falcata —- —- —- 11 —- —- —- —- —- Securigera varia —- —- —- 11 —- —- —- —- —- Teucrium chamaedrys —- —- —- —- 40 —- —- —- —- Bromus erectus —- —- —- —- 20 —- —- —- —- Vicia tenuifolia —- —- —- —- 20 —- —- —- —- Trifolium repens —- —- —- —- —- 42 —- —- —- Bromus racemosus agg. —- —- —- —- —- 17 —- —- —- Carex distans —- —- —- —- —- 17 —- —- —- Alopecurus pratensis —- —- —- —- —- —- 10 17 —- Centaurea jacea —- —- —- —- —- —- 10 —- —- U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:08 Color profile: Disabled Composite 150 lpi at 45 degrees perenne) tolerate grazing and trampling well but demand nutrient-rich and semi-humid or humid soil. Most of the sites are located on the Drava Plain where they probably originated from the overgrazing and drying of wet meadows. In the syntaxonomical system this group belongs to the Cynosurion alliance. Group 7 Number of relevés: 29; mean number of species per relevé: 40.7. Group 7 contains special, transitional meadows that belong to the Arrhenatherion alli- ance of the Arrhenatheretalia order. These stands are characterized by the co-existence of species that tolerate temporary drought and flood as well (e.g. Betonica officinalis, Filipendula vulgaris, Helictotrichon pubescens) and moisture-demanding species (e.g. Alopecurus pratensis, Carex pallescens, Sanguisorba officinalis). Besides grassland spe- cies generally characterising the mesophilous meadows (e.g. Campanula patula, Rumex acetosa, Saxifraga bulbifera, Trisetum flavescens), plants of forest edges are also present (e.g. Cruciata glabra, Knautia drymeia). This variety of species characters is explained by these meadows usually being situated on packed soils with a fluctuating water supply, be- cause of which the moist and dry periods alternate over the year. These meadows resemble drying Molinion coeruleae stands, however, the latter is a moister grassland type. More- over, many of these meadows are both grazed and mown, which also contributes to their di- versity. These are ones of the most species-rich meadows of the region. This type is distri- buted in lowlands and hills. Group 8 Number of relevés: 30; mean number of species per relevé: 34.6. This cluster consists of relevés from regularly mown, productive, tall grass stands. These meadows grow on nutrient-rich and steadily humid soils from the lowlands to the Mecsek Mts. They are 'typical' oat grass meadows, which harbour many generalist grass- land species but have very few differential species (e.g. Crepis biennis, Ranunculus acris). In the syntaxonomical system, this group corresponds to the Pastinaco–Arrhenatheretum association, which is distributed widely in other Central European countries as well. This group has transitions to almost all the others. From our preliminary analysis, we found that this group could have been delimited differently, which would slightly alter its syntaxo- nomic interpretation, depending on how transitional stands are classified. Group 9 Number of relevés: 13; mean number of species per relevé: 33. Relevés classified into Group 9 originate from the most humid sites, and should be con- sidered as members of Molinietalia meadows. These stands are located in lowlands and valleys and they can be flooded in spring, which is not usual for mesic (Arrhenatheretalia) but typical for wet (Molinietalia) grasslands. Differential species are wet meadow plants (e.g. Cirsium canum, Poa trivialis, Ranunculus repens, Symphytum officinale). Since this group is at the periphery of the scope of our analysis, it is not discussed in more detail here. Indicator values and ordination Box-and-whisker plots reveal the strong outlier character of Group 5 because it consi- derably differs from all other groups regarding all indicator values (Figs. 3, 4, 5). Other 40 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:09 Color profile: Disabled Composite 150 lpi at 45 degrees groups show gradual differences in temperature (Fig. 3a), soil reaction (Fig. 4a) and continentality (Fig. 5b) values because the positions of the medians and the proportions of overlaps resemble random patterns. In contrast, groups are aggregated into 'levels' along the gradients of moisture, nutrient and light values. Groups on the same level have similar medians and highly overlapping interquartile ranges, while the levels differ significantly from each other. In moisture values (Fig. 3b), between the two extremes (Group 5 and Group 9, dry and wet, respectively) a drier level consisted of Groups 1, 3 and 4, while a moister level is formed of Groups 6, 7 and 8. There are two levels along the nutrient gradi- ent as well (Fig. 4b). More nutrient-rich clusters are Groups 6, 8 and 9, while Groups 1, 3, 4 and 7 have on average lower nutrient status. Two levels are differentiated according to light availability values (Fig. 5a). Groups 1, 3, 4 and 6 are characterised by more light, and ACTA BOT. CROAT. 71 (1), 2012 41 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA Fig. 3. Box-and-whisker plots of mean temperature (a) and moisture (b) indicator values of clus- ters. TB – Borhidi's temperature values, WB – Borhidi's moisture values. Fig. 4. Box-and-whisker plots of mean soil reaction (a) and nutrient status (b) indicator values of clusters. RB – Borhidi's soil reaction values, NB – Borhidi's nutrient demand values. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:12 Color profile: Disabled Composite 150 lpi at 45 degrees Groups 7, 8 and 9 with less. This pattern corresponds with their differences in management because relevés of Groups 1, 3, 4, 5 and 6 originate mostly from grazed sites but the others from mown meadows. The relative positions of the groups can be seen on the ordination diagram (Fig. 6). The first canonical axis explains 23.6% of the within-group variation, which is 5.0% of the total variation. The second axis of canonical analysis of principal coordinates accounts for 20.9% of the canonical variation, which is 4.5% of the total variation of species data. Indicator val- 42 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. Fig. 5. Box-and-whisker plots of mean light availability (a) and continentality (b) indicator values of clusters. LB – Borhidi's light demand values, CB – Borhidi's climatic continentality Fig. 6. Canonical analysis of principal coordinates ordination diagram with spider plot of the clus- ters and with passively projected mean indicator values U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees ues such as: temperature indicator value, moisture, nutrient status and light availability are projected onto the scatterplot with high (p<0.001) continentality indicator value, with weaker but still significant (p<0.05) fit, while soil reaction could not be fitted to the first two axes (Tab. 4). Moisture and nutrient status correlate with both ordination axes, temper- ature indicator value and light availability with only the first axis, continentality indicator value with only the second axis. The ordination diagram supports the importance of mois- ture, nutrient status and light availability gradients in the species composition. Discussion Ours is the first attempt on the classification of Hungarian mesophilous meadows using multivariate methods, even if our work focuses on a minor part of the country. Therefore, our syntaxonomic results generate rather new questions and hypotheses about the current syntaxonomic system. From our classification four clusters emerged as important types certainly belonging to the Arrhenatheretalia order. Two of these four clusters belong to the Cynosurion alliance (Group 1, Group 6) and the other two to Arrhenatherion (Group 7, Group 8). These alliances are separated along the first axis of the canonical principal coordinate ordination that corresponds also with light availability (Fig. 6), with groups of Cynosurion getting higher mean light availability val- ues. This is in line with the differences in their management (i.e. grazing for Cynosurion and mowing for Arrhenatherion), and accordingly Borhidi's light values seem to differenti- ate well between grazed and mown communities. Within both alliances, clusters differ in nutrient status (Figs. 4b, 6). These patterns were more or less incorporated previously in the syntaxonomic concept but species composi- tions of the groups were not in accordance with the association descriptions in every case. In the Hungarian syntaxonomic system (BORHIDI 2003), mesophilous pastures on nutri- ent-poor soils are included in the Festuco commutatae–Cynosuretum distributed in colline and montane areas, which contains species associated with the montane climate in Hun- gary (Carlina acaulis, Rhinanthus alectorolophus, Rh. wagneri, Danthonia decumbens, Alchemilla spp.). However, even the highest altitudes of our study area do not reach the montane belt and most of the montane characteristic species of the above association are not present in the regional species pool. Thus, pastures with low nutrient availability in ACTA BOT. CROAT. 71 (1), 2012 43 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA Tab. 4. Pearson’s correlation of indicator values with the ordination axes and the goodness of fit of their passive projections. TB – temperature, WB – moisture, RB – soil reaction, NB – nutri- ent status, LB – light availability, CB – continentality indicator value Correlation with the axes Fit CAP 1 CAP 2 R-squared p TB –0.996 0.095 0.209 <0.001 WB 0.744 0.669 0.544 <0.001 RB –0.167 0.986 0.006 n.s. NB 0.450 0.893 0.548 <0.001 LB –0.997 –0.072 0.533 <0.001 CB 0.093 –0.996 0.059 <0.05 U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees South Transdanubia did not correspond with Festuco commutatae–Cynosuretum; however, this can be explained by the lack of data from this region in the past, which prevented the Hungarian description (BORHIDI 2003) of this association being extended to the southern stands. As borne out by ROZBROJOVÁ et al. (2010), extensive pastures have species compo- sitions similar to low production hayfields. There is a low productive hay meadow associa- tion in Hungary in the Arrhenatherion alliance, Anthyllido–Festucetum rubrae, which is distributed through the whole country in colline and montane ranges according to BORHIDI (2003). However, we could not recognise this association based on our data. It seems time to revise these two association names and descriptions (Festuco commutatae–Cynosure- tum and Anthyllido–Festucetum rubrae) based on statistical analyses, limit their applica- tion to narrower geographic areas and find appropriate names for the others. Moreover, if uniqueness and distinctness of narrow-range associations are not supported enough, it is advisable to adopt names from the syntaxonomic systems of neighbouring countries in or- der to standardise the nomenclature of syntaxa over the countries (KNOLLOVÁ et al. 2006). Based on archive data and available literature of Central European countries (HÁJKOVÁ et al. 2007, JANISOVÁ et al. 2007), Festuco–Cynosuretum, Anthoxantho–Agrostietum and Festuco–Agrostietum seem equally applicable names for southern Hungarian extensive pastures. The group of nutrient-rich Cynosurion pastures (Group 6) resembles the Lolio–Cyno- suretum association that is already reported from Hungary. According to BORHIDI (2003) this association occurs in colline and montane ranges. However, HÁJKOVÁ et al. (2007) ex- tended its altitude range to the lowlands in the Czech Republic. In our study area, similar stands are the most common in the Drava Plain, but rarer in the higher altitudes, contrary to BORHIDI (2003). Lowland stands are not only situated at lower altitudes but also in southern geographic position affected more by sub-Mediterranean climate and they harbour a few of the sub-Mediterranean species that are rare or sparse in the study region (e.g. Hordeum secalinum). Therefore, we think an urgent task to explore the relationship between the nu- trient-rich Cynosurion pastures of South Transdanubia and the Bromo–Cynosuretum stands of Croatia. It should also be further investigated how they are delimited from trampled and grazed wet meadows. This problem is also recognised by ZUIDHOFF et al. (1995) who sug- gested applying the name Junco–Cynosuretum Sougnez 1957 to relevés that they classified to a transitional cluster between Calthion and Cynosurion (mainly Lolio–Cynosuretum). Noteworthily, their analysis is based mostly on Western European relevés with dispropor- tionately few Central European data. We identified Group 8 as a Pastinaco–Arrhenatheretum elatioris association that is a common meadow community in Central Europe and also included in the Hungarian sys- tem. Nevertheless, there are some other associations characterised by Arrhenatherum elatius which are difficult to distinguish from Pastinaco–Arrhenatheretum. In Hungary, such an association is Alopecuro–Arrhenatheretum which is a mountainous type described from the northern part of the country (MÁTHÉ and KOVÁCS 1960). In Croatia, Ononi- do–Arrhenatheretum is distributed along the Drava and Mura rivers and is considered a moister and less productive meadow than Pastinaco–Arrhenatheretum (ILIJANI] and [EGU- LJA 1983, TRINAJSTI] 2008). Though, these associations are not clearly delimited from Pastinaco–Arrhenatheretum (STAN^I] 2008) and the differences are probably so slight that the applied hierarchical classification methods are not able to distinguish them as separate clusters of our current data set. Here we suggest defining these associations by supervised 44 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees classification tools (BRUELHEIDE 2000) and/or calculating similarity measures (TICHÝ 2005, VAN TONGEREN et al. 2008) between reliable reference plot data and the unidentified relevés, and using the resulting assignments for mapping their distribution. We expect that a few relevés would be identified as any of these rarer associations but the majority of Group 8 should be assigned to Pastinaco–Arrhenatheretum. It is also worth mentioning that Pastinaco–Arrhenatheretum is characterised by generalist meadow species, which are common on drying wet meadows and other dynamic types, thus the species composition of Pastinaco–Arrhenatheretum can easily appear on various sites. It is another question whether differences among the above narrow associations (Alopecuro–Arrhenatheretum and Ononido–Arrhenatheretum) and the core type (Pastinaco–Arrhenatheretum) are big- ger than the differences among dynamic phases. The low productive part of Arrhenatherion alliance had previously been covered by two associations: Anthoxantho–Festucetum rupicolae from lowlands and Anthyllido–Fes- tucetum rubrae from colline and montane ranges (BORHIDI 2003). Anthoxantho–Festuce- tum rupicolae is originally described from this area (Drava Plain) by DÉNES (1997) and some of our relevés are clearly identifiable with the published data. Anthyllido–Festucetum rubrae is considered common in deciduous forest ranges (BORHIDI 2003), but we could not recognise it in our data. Instead, our Group 7 seemed the most similar to Filipendulo– –Arrhenatheretum described originally from Austria (HUNDT and HÜBL 1983) and detected also in Croatia (TRINAJSTI] 2002). Similar stands are reported from the Bakonyalja region in Transdanubia (BAUER et al. 2001), but no phytosociological relevés are published. The presence of this association in Hungary should be validated by analysing our data together with relevés of the above countries. If this community is accepted as present in Hungary, the position of Anthoxantho–Festucetum rupicolae should also be revised because Fili- pendulo–Arrhenatheretum would be an association with a wide distribution and many vari- ants, and Anthoxantho–Festucetum rupicolae may be only its dry variant with subregional distribution. In Austria three subassociations of Filipendulo–Arrhenatheretum are de- scribed in HUNDT and HÜBL (1983): typical, wet (with Cirsium oleraceum) and dry (with Centaurea scabiosa), although the dry subassociation differs from Anthoxantho–Festu- cetum rupicolae floristically. Recently, an ecologically similar association was described (ZELNIK 2007) with the name Triseto–Centaureetum macroptili. This community repre- sents transitions between wet and mesic grasslands and it occurs on elevated part of alluviae where the water table fluctuates, as in the locations of Group 7. However, we are not aware of any occurrences of its dominant species, Centaurea jacea subsp. macroptilon, in the studied region. Moreover, Triseto–Centaureetum harbour some acidophilous species that are also very rare or absent in South Transdanubia (e.g. Nardus stricta, Festuca filiformis). We also identified a dynamic phase of several types but mainly of nutrient-poor pas- tures. Such clusters are usually not interpreted syntaxonomically. However, these degraded stands show continuous transition to other, more 'stable' types and their representation in an unsupervised classification is also a question of sampling decisions. Three of our clusters are identified as belonging to or representing transitions to orders other than Arrhenatheretalia. Aware that the transitions between Arrhenatheretalia and Brometalia or Arrhenatheretalia and Molinietalia meadows are continuous and they can only be delimited quite arbitrarily, we are careful with the syntaxonomic interpretation of ACTA BOT. CROAT. 71 (1), 2012 45 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees transitional groups. These clusters are hard to characterise because they often include stands from various transitional pathways between broader syntaxa. For example, Group 4 comprises a broad variety of transitions between mesic and semi-dry meadows, its diag- nostic species thus being the generalists, which have in common only their moisture prefer- ence. Heterogeneity of this group is also supported by Festuca rupicola being the only spe- cies that dominates more than one stand. There are three associations considered transitional between semi-dry and mesic meadows by BORHIDI (2003): Anthoxantho–Festucetum rupicolae Dénes 1997 in the Arrhenatherion alliance, Onobrychido viciaefoliae–Brome- tum erecti Müller 1966 and Carlino acaulis–Brometum Oberdorfer 1957 in the Bromion erecti, however, the presence of the latter two is not validated by statistical analysis in a re- cent review (ILLYÉS et al. 2009). Literatures of neighbouring countries also consider Ranunculo bulbosi–Arrhenatheretum Ellmauer in Mucina et al. 1993 and Filipendulo vulgaris–Arrhenatheretum Hundt et Hübl 1983 of Arrhenatherion as such intermediate types. There are several associations interpretable as transitions between wet and mesic meadows as well (e.g. Alopecuro–Arrhenatheretum (Máthé et Kovács 1960) Soó 1971, Cirsio cani–Festucetum pratensis Májovsky et Ru`i~ková 1975, Arrhenathero–Molini- etum arundinaceae Lájer 2002, Holcetum lanati Issler 1934). In our opinion, the status of these intermediate associations should be investigated in an analysis comprising many relevés from both orders they are related to. Acknowledgements We are grateful to János BODÓ, András PINTÉR, István Zsolt TÓTH and László WÁGNER for their help in the field work. We thank Prof. János PODANI and the two anonymous refer- ees for their valuable comments on the manuscript. References ANDERSON, M. J., WILLIS, T. J., 2003: Canonical analysis of principal coordinates: a useful method of constrained ordination for ecology. Ecology 84, 511–526. BAUER, N., BALOGH, L., KENYERES, Z., 2001: Vegetation and conservational problems of the fens at Tapolcafõ and Attya (Pápa-Bakonyalja; Region of the Bakony Mountains, Hungary) (in Hungarian). Botanikai Közlemények 88, 71–94. BLA@EN^I], J., RAN|ELOVI], V., BUTORAC, B., VUKOJI^I], S., ZLATKOVI], B., @UKOVEC, D., ]ALI], I., PAVI]EVI], D., LAKU[I], D., 2005: Habitat classification of Serbia (In Ser- bian). Institut za Botaniku i Botani~ka Ba{ta Jevremovac, Belgrade. BÖLÖNI, J., KUN, A., MOLNÁR, Z. s., 2003: Habitat guide (In Hungarian). MTA ÖBKI, Vácrátót. BORHIDI, A., 1995: Social behaviour types, the naturalness and relative ecological indicator values of the higher plants in the Hungarian flora. Acta Botanica Hungarica 39, 97–181. BORHIDI, A., 2003: Plant associations of Hungary (In Hungarian). Akadémiai Kiadó, Buda- pest. BOTTA-DUKÁT, Z., CHYTRÝ, M., HÁJKOVÁ, P., HAVLOVÁ, M., 2005: Vegetation of lowland wet meadows along a climatic continentality gradient in Central Europe. Preslia 77, 89–111. 46 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees BRUELHEIDE, H., 2000: A new measure of fidelity and its application to defining species groups. Journal of Vegetation Science 11, 167–178. ^ARNI, A., 2001: Vegetation of cultivated grasslands in the Gori~ko region (NE Slovenia). Acta Biologica Slovenica 44, 13–27. CHYTRÝ, M., TICHÝ, L., HOLT, J., BOTTA-DUKÁT, Z., 2002: Determination of diagnostic spe- cies with statistical fidelity measures. Journal of Vegetation Science 13, 79–90. COLDEA, G., 1991: Prodrome des associations vegetales des Carpates du sud-est (Carpates Roumaines). Documenta Phytosociologica 13, 317–539. DÉNES, A., 1997: Occurrence of Anthoxantho–Festucetum rupicolae association on the meadows of the Drava Plain (In Hungarian). Kitaibelia 2, 274–275. DENGLER, J., CHYTRÝ, M., EWALD, J., 2008: Phytosociology. In: JÖRGENSEN, S. E., FATH, B. D., Encyclopedia of ecology 4. General ecology. Elsevier, Oxford. DÖVÉNYI, Z., 2010: Magyaroszág kistájainak katasztere. MTA Földrajztudományi Kuta- tóintézet, Budapest. ELLMAUER, T., 1994: Syntaxonomie der Frischwiesen (Molinio–Arrhenatheretea p.p.) in Österreich. Tuexenia 14, 151–168. HÁJKOVÁ, P., HÁJEK, M., BLA@KOVÁ, D., KU^ERA, T., CHYTRÝ, M., ØEZNÍ^KOVÁ, M., [UMBEROVÁ, K., ^ERNY, T., NOVÁK, J., SIMONOVÁ, D., 2007: Meadows and mesic pas- tures. In: CHYTRÝ, M. (ed.), Vegetation of the Czech Republic. 1. Grassland and heathland vegetation (In Czech). Academia, Praha. HORVÁT, A. O., 1962: Meadows in the vicinity of the Mecsek Mts. (In Hungarian) Annals of the Janus Pannonius Museum 1960, 53–67. HORVÁT, A. O., 1972. Die Vegetation des Mecsekgebirges und seiner Umgebung. Aka- démiai Kiadó, Budapest. HORVÁTH, F., DOBOLYI, Z. K., MORSCHHAUSER, T., LÕKÖS, L., KARAS, L., SZERDAHELYI, T., 1995: Hungarian flora database 1.2. List of taxa and relevant attributes (In Hungarian). MTA ÖBKI, Vácrátót. HUNDT, R., HÜBL, E., 1983: Pflanzensoziologische, pflanzengeographische und landes- kulturelle Aspekte des Filipendulo-Arrhenatheretum im Wiener Wald. Tuexenia 3, 331–342. ILIJANI], Lj., [EGULJA, N., 1983: Phytozönologische und ökologische Untersuchungen der Glatthaferwiesen in der Podravina (Nordkroatien). Acta Botanica Croatica 42, 63–82. ILLYÉS, E., BAUER, N., BOTTA-DUKÁT, Z., 2009: Classification of semi-dry grassland vege- tation in Hungary. Preslia 81, 239–260. ILLYÉS, E., CHYTRÝ, M., BOTTA-DUKÁT, Z., JANDT, U., [KODOVÁ, I., JANI[OVÁ, M., WILL - NER, W., HÁJEK, O., 2007: Semi-dry grasslands along a climatic gradient across Central Europe: Vegetation classification with validation. Journal of Vegetation Science 18, 835–846. JANI[OVÁ, M., MICHÁLKOVÁ, D., [KODOVÁ, I., UHLIAROVÁ, E., ZALIBEROVÁ, M., 2007: Cynosurion cristati Tüxen 1947. In: JANI[OVÁ, M. (ed.), Grassland vegetation of Slovak Republic: electronic expert system for identification of syntaxa (In Slovakian), 110–118. Botanický ústav SAV, Bratislava. ACTA BOT. CROAT. 71 (1), 2012 47 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees JEANPLONG, J., 1960: Übersichliche Schilderung der am oberen Raablauf in Ungarn gelegenen Wiesen (In Hungarian). Botanikai Közlemények 48, 101–105. JUHÁSZ-NAGY, P., 1959: Meadow and pasture associations of the Bereg Plain (In Hungar- ian). Acta Universitatis Debreceniensis 4, 195–228. KIRÁLY, G. (ed.), 2009: New Hungarian herbal. The vascular plants of Hungary. Identifica- tion key (In Hungarian). Aggteleki Nemzeti Park Igazgatóság, Jósvafõ. KNOLLOVÁ, I., CHYTRÝ, M., TICHÝ, L., HÁJEK, O., 2005: Stratified resampling of phyto- sociological databases: some strategies for obtaining more representative data sets for classification studies. Journal of Vegetation Science 16, 479–486. KNOLLOVÁ, I., CHYTRÝ, M., TICHÝ, L., HÁJEK, O. 2006: Local ranges of phytosociological associations: are they reflected in numerical classification? Biologia (Bratislava) 61, 71–77. KOJI], M., MRFAT-VUKELI], S., DAJI], Z., \OR|EVI]-MILO[EVI], S., 2004: Meadows and pastures of Serbia (In Serbian). Institut za istra`ivanja u poljoprivredi Srbije, Belgrade. KOVÁCS, J. A., 1994: Meadow associations of the Kõszeg Mts. and Kõszeg-hegyalja. In: BARTHA, D. (ed.), Vegetation of the Kõszeg Mts. (In Hungarian), 147–174. Sopron– Kõszeg. KOVÁCS, J. A., 2002: Meadows vegetation of Õrség Landscape protection area (In Hungar- ian). Kanitzia 10, 137–174. LÁJER, K., 2002: Floristical and coenological studies on meadows of the Somogy county valley of the river Drava (In Hungarian). Kitaibelia 7, 187–205. LÁJER, K., BOTTA-DUKÁT, Z., CSIKY, J., HORVÁTH, F., SZMORAD, F., BAGI, I., DOBOLYI, K., HAHN, I., KOVÁCS, J. A., RÉDEI T., 2008: Hungarian phytosociological database (CO- ENODATREF): sampling methodology, nomenclature and its actual stage. Annali di Botanica 7, 197–201. LEGENDRE, P., LEGENDRE L., 1998: Numerical ecology. Elsevier, Amsterdam. LENGYEL, A., CHYTRÝ, M., TICHÝ L., 2011: Heterogeneity-constrained resampling of phytosociological databases. Journal of Vegetation Science 22, 175–183. LOVÁSZ, GY. (ed.), 1977: Natural geography of Baranya County (In Hungarian). Baranya Megyei Levéltár, Pécs. MAECHLER, M., ROUSSEEUW, P., STRUYF, A., HUBERT M., 2005: Cluster analysis basics and extensions; unpublished. MÁTHÉ, I., KOVÁCS M., 1960: Vegetationsstudien im Mátragebirge. Acta Botanica Aca- demiae Scientiarum Hungariae 6, 343–382. MOLNÁR, ZS., BARTHA, S., SEREGÉLYES, T., ILLYÉS, E., BOTTA-DUKÁT, Z., TÍMÁR, G., HORVÁTH, F., RÉVÉSZ, A., KUN, A., BÖLÖNI, J., BÍRÓ, M., BODONCZI, L., DEÁK, J. Á., FOGARASI, P., HORVÁTH, A., ISÉPY, I., KARAS, L., KECSKÉS, F., MOLNÁR, CS., ORT- MANN-NÉ AJKAI, A., RÉV, SZ., 2007: A grid-based, satellite-image supported, multi-at- tributed vegetation mapping method (MÉTA). Folia Geobotanica 42, 225–247. MOLNÁR, ZS., BÍRÓ, M., BÖLÖNI, J., HORVÁTH, F., 2008: Distribution of the (semi-) natural habitats in Hungary. I. Marshes and grasslands. Acta Botanica Hungarica 50, 59–105. MUCINA, L., GRABHERR, G., ELLMAUER, T. (eds.), 1993: Die Pflanzengesellschaften Öster- reichs. Teil I. Anthropogene Vegetation. Gustav Fischer Verlag, Jena/Stuttgart/New York. 48 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees OKSANEN, J., BLANCHET, F. G., KINDT, R., LEGENDRE, P., O'HARA, R. G., SIMPSON, G. L., SÓLYMOS, P., STEVENS, M. H. H., WAGNER, H., 2010: vegan: Community Ecology Pack- age. R package version 1.17–0. http://cran.r-project.org/package=vegan (access: Feb- ruary 9, 2012) PODANI, J. 2000: Introduction to the exploration of multivariate biological data. Backhuys Publishers, Leiden. R DEVELOPMENT CORE TEAM, 2009: R: A language and environment for statistical comput- ing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org (access February 9, 2012) ROZBROJOVÁ, Z., HÁJEK, M., HÁJEK, O., 2010: Vegetation diversity of mesic meadows and pastures in the West Carpathians. Preslia 87, 307–332. SANDA, V., POPESCU, A., ARCUº, M., 1999: Critical revision of plant communities in Roma- nia (In Romanian). Tilia Press International, Constanþa. SOÓ, R., 1971: Aufzählung der Assoziationen der ungarischen Vegetation nach den neu- eren zönosystematisch-nomeklatorischen Ergebnissen. Acta Botanica Academiae Scien- tiarum Hungariae 17, 127–127. SOÓ, R., 1973: Synopsis systematico-geobotanica florae vegetationisque Hungariae V. (In Hungarian) Akadémiai Kiadó, Budapest. STAN^I], Z. 2000: Grasslands of class Molinio-Arrhenatheretea in northeastern Croatia (In Croatian). PhD Thesis, University of Zagreb, Zagreb. STAN^I], Z., 2008: Classification of mesic and wet grasslands in northwest Croatia. Biologia (Bratislava) 63, 1089–1103. TICHÝ, L., CHYTRÝ, M., 2006: Statistical determination of diagnostic species for site groups of unequal size. Journal of Vegetation Science 17, 809–818. TICHÝ, L. 2005: New similarity indices for the assignment of relevés to the vegetation units of an existing phytosociological classification. Plant Ecology 179, 67–72. TICHÝ, L., CHYTRÝ, M., HÁJEK, M., TALBOT, S. S., BOTTA-DUKÁT, Z. 2010: OptimClass: Using species-to-cluster fidelity to determine the optimal partition in classification of ecological communities. Journal of Vegetation Science 21, 287–299. TRINAJSTI], I. 2002: Meadows of false oat-grass and dropworth – ass. Filipendulo vulgaris – Arrhenatheretum Hundt et Hübl. (Arrhenatherion) in Croatia (In Croatian). Agro- nomski Glasnik 2002, 125–136. TRINAJSTI], I. 2008: Plant communities of Republic of Croatia (In Croatian) Akademija {umarskih znanosti, Zagreb. UHLIAROVÁ, E., JANI[OVÁ, M., UJHÁZY, K., 2007: Arrhenatherion elatioris Luquet 1926. In: JANI[OVÁ M. (ed.), Grassland vegetation of Slovak Republic: electronic expert sys- tem for identification of syntaxa (In Slovakian), 91–109. Botanický ústav SAV, Brati- slava. VAN TONGEREN, O., GREMMEN, N., HENNEKENS, S., 2008: Assignment of relevés to pre-de- fined classes by supervised clustering of plant communities using a new composite in- dex. Journal of Vegetation Science 19, 525–536. ZELNIK, I., 2007: New grassland association Triseto-Centaureetum macroptili ass. nova in Slovenia. Hacquetia 6, 77–89. ACTA BOT. CROAT. 71 (1), 2012 49 MESIC GRASSLANDS OF SOUTH TRANSDANUBIA U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees ZELNIK, I., ^ARNI, A., 2008: Distribution of plant communities, ecological strategy types and diversity along a moisture gradient. Community Ecology 9, 1–9. ZUIDHOFF, A. C., RODWELL, J. S., SCHAMINÉE, J. H. J., 1995: The Cynosurion cristati Tx. 1947 of central, southern and western Europe: a tentative overview, based on the analy- sis of individual relevés. Annali di Botanica 53, 25–47. 50 ACTA BOT. CROAT. 71 (1), 2012 LENGYEL A., PURGER D., CSIKY J. U:\ACTA BOTANICA\Acta-Botan 1-12\464 Lengyel et al.vp 26. o ujak 2012 10:25:14 Color profile: Disabled Composite 150 lpi at 45 degrees