Plant Biosystems, Vol. 143, No. 1, March 2009, pp. 1-13 Taylor & Francis Taylor & Francis Group Thermophilous deciduous forests in Southeastern Europe A. CARNI1, P. KOSIR1, B. KARADZIC2, V. MATEVSKI3, S. REDZIC4, & Z. SKVORC5 1 Institute of Biology, Scientific Research Centre of the Slovenian Academy of Sciences and Arts, Ljubljana, 2Institute for Biological Research "Sinisa Stankovic", Beograd, 3 Institute of Biology, Faculty of Natural Sciences and Mathematics, Skopje, ^Department of Biology, Faculty of Science, University of Sarajevo, Sarajevo, and 5Faculty of Forestry, University of Zagreb, Zagreb Abstract This paper deals with the numerical elaboration of the database of 1764 releves of thermophilous deciduous forests assigned by the authors to the order Quercetalia pubescentis in the northwestern part of Southeastern Europe. After elimination of releves which were dominated by mesophilous deciduous and coniferous tree species, the stratification of releves was carried out. The TWINSPAN classification revealed eight ecologically and phytogeographically interpretable groups. Additionally the analysis of Pignatti indicator values passively projected on the PCA diagram of the eight groups, and a chorological analysis of individual groups were made. The analyses revealed that all groups in general match with the traditionally accepted alliances, such as Quercion pubescenti-sessiliflorae, Aceri tatarici-Quercion, Quercion confertae, Quercion petraeae-cerris, Syringo-Carpinion, Pruno tenelle-Syrinion, Carpinion orientalis and Fraxino orni-Ostryion. Finally, a synsystematic classification of the elaborated group is proposed, and the nomenclature is harmonised with the International Code of Phytosociological Nomenclature. The results are also presented in a synoptic table together with calculation of the diagnostic species. Abbreviation: ICPN, International Code of Phytosociological Nomenclature Keywords: Chorology, ecology, fidelity, forest, Southeastern Europe, syntaxonomy, Two Way INdicator SPecies Analysis Introduction The thermophilous deciduous forests of the order Quercetalia pubescentis spread out in the warm parts of Europe. They are climazonal in warmer parts of Europe and extrazonal in the northern, cooler regions. In some parts of Europe, this type of vegetation has already been well elaborated on a regional level, for instance in Austria (Willner & Grabherr 2007), Italy (Blasi et al. 2004), Germany (Oberdorfer 1992), the Czech Republic (Knollová & Chytrý 2004), Slovakia (Roleček 2005), Hungary (Kevey & Borhidi 2005), Bulgaria (Rousakova & Tzonev 2003) and Greece (Bergmeier & Dimopou-los 2008), but in the region under study such an overview has not yet been carried out. We treated the order of thermophilous deciduous forests Quercetalia pubescentis as a part of the class Querco-Fagetea, as do the studies by most authors in the region (e.g. Horvat et al. 1974; Jovanovic et al. 1986) and as also considered elsewhere (Rivas-Martinez et al. 2001, 2002). This is justified by many mesophilous plants linking the orders of the class Querco-Fagetea together. Certain recent classifications (Mucina 1997; Rodwell et al. 2002) include this type of vegetation into a special class, that of Quercetea pubescentis. In Southeastern Europe, studies on the vegetation according to the central European Braun-Blanquet methodology (Braun-Blanquet 1964) has a long tradition. Horvat (1938), Tomazic (1939) and Rudski (1949) are among the pioneers who studied thermophilous deciduous forests. They were followed by numerous authors who are also among the authors of the tables drawn up and selected for this contribution. Since, due to objective circumstances, cooperation in this area has been rendered Correspondence: A. Carni, Institute of Biology, Scientific Research Centre of the Slovenian Academy of Sciences and Arts, Novi trg 2, P.B. 306, SI-1001 Ljubljana. Tel: + 386 1 47 06 311. Fax: + 386 1 425 99 97. Email: carni@zrc-sazu.si ISSN 1126-3504 print/ISSN 1724-5575 online © 2009 Societa Botanica Italiana DOI: 10.1080/11263500802633881 2 A. Čarni et al. difficult in the last few years, we were only recently able to form a group which prepared a survey of the vegetation of thermophilous deciduous forests in this area in compliance with the modern methodology. The aim of the research was to collect material on the thermophilous deciduous forests of the Querce-talia pubescentis in the region, and then to establish the major vegetation types resulting from the numerical analysis, and to compare them with the major types recognised in the traditional expert-based classification. The groups are presented in a synoptic table, with calculation of diagnostic species, chorological and ecological characteristics as well as their distribution patterns. Methods and materials Geographical scope We studied the thermophilous deciduous forests of the order Quercetalia pubescentis in the northwestern part of Southeastern Europe, in the territory of Slovenia, Croatia, Bosnia and Herzegovina, the former Serbia and Montenegro, and Macedonia, all states of the former Yugoslavia. The territory is a complex structure, since it comprises a part of the Pannonian basin, the coasts of the Mediterranean Sea, southern hillsides of the Alps and various mountain chains in the Balkans. The territory is classified as the Euro-Siberian region, i.e. the Apennine-Balkan and the Pannonian-Carpathian provinces, and the Mediterranean region with two provinces: the Adriatic, and the Greek/Aegean province (Rivas-Martinez et al. 2004). Object of the study The objects of our research are thermophilous deciduous forests dominated by various deciduous species such as Quercus sp. div., Ostrya carpinifiolia, Carpinus orientalis, Fraxinus ornus, to mention only the most important ones. In the northern part of the region, these forests occur up to a height of 600 m where they form the extrazonal vegetation, whereas in the south they appear up to 1000 m or even higher, and form the climazonal vegetation. All releves of communities classified by their authors in the order Quercetalia pubescentis were collected from the literature. We excluded the releves whose dominant tree species (cover values 3, 4 and 5) are species of mesophilous climazonal and other forest types, above all mesophilous and coniferous ones, e.g. Abies alba, Acer platanoides, A. pseudoplatanus, Carpinus betulus, Fagus sylvatica subsp. sylvatica, F. sylvatica subsp. moesiaca, Fraxinus excelsior, Picea abies, Pinus nigra, P. sylvestris, Quercus ilex, Tilia cordata, T. platyphyllos, Ulmus glabra. The releves with an incomplete list of herb species indicated by the authors were also not included in the analyses. Methods The 1764 releves of forest vegetation of Southeastern Europe collected from the literature were entered into the TURBOVEG (Hennekens & Schaminée 2001) database. After exclusion of releves dominated by the forest species mentioned above, 1715 releves of thermophilous forests were left. The initial data set of 1715 releves was then stratified. Stratified resampling was made by phyto-sociological association as indicated by the authors. This means that up to ten releves of one association were selected in such a way that different authors, different publications and different locations were represented (Košir et al. 2008). As the associations mostly appear in a broad phytogeographical region, a special geographical stratification is not needed. After stratification 604 releves remained, originating from 82 associations. As many authors did not record mosses, we excluded them from our analysis before numerical processing. For the purpose of numerical analysis and in the synoptic table we unified the system of layer division, which differs from author to author. All sublayers of the tree layer were integrated into one, whereas for woody species, the herb and scrub layers, tree saplings and seedlings, and lianas were united into a single scrub layer. Additionally, we integrated certain plant subspecies and varieties into the level of species or aggregates. Then we carried out a classification by TWIN-SPAN (Hill 1979), run under the JUICE 6.5 programme (Tichý 2002). TWINSPAN pseudospe-cies cut levels for species abundances were set to 0— 5-25 percentage scale units. Initially, six levels of division were chosen and the minimum group size for division was set to five releves. Later on, different levels of division were accepted, resulting in eight groups of releves interpretable in terms of ecology and phytogeography. In the group distribution map only those localities where the majority of releves of individual groups is located are presented. The diagnostic species of eight individual groups were determined by calculating species' fidelity (Chytrý et al. 2002; Havlová et al. 2004) and are presented in the synoptic table (Table I). As a fidelity measure we used the phi coefficient in the JUICE programme. In these calculations, each group of releves was compared with the rest of the releves in the data set, which were taken as a single undivided group. Each of the eight groups was virtually adjusted to 1/8 of the size of the entire data set, while holding the percentage occurrences of a species within and outside a target group the same as in the Forests in Southeastern Europe 3 Table I. Synoptic table of the TWINSPAN classification. See also supplementary material available online at: http://www.informaworld. com/mpp/uploads/supplementary_files_363558_1236759373086.zip Group number 1 2 3 4 5 6 7 8 No. of releves 108 15 68 57 35 39 174 108 Proportion of geoelements in groups (%) Stenomediterranean 1.9 0 1.2 2.2 0 2.1 6 1.6 Eurimediterranean 7.6 10.2 9.4 8.9 10.3 8.2 10.5 6.7 Mediterranean-montane 1.3 0 0.6 1.7 6.6 1 3.3 12.5 Eurasian 66.9 65.3 61.4 58.7 55.9 58.8 50.6 52.9 SE European 5.7 6.1 7.6 6.7 8.1 8.2 8.1 8 Atlantic 1.9 2 0.6 1.1 0 0 2.1 1.3 Eurosiberian 10.2 10.2 11.7 10.6 9.6 11.3 9.3 12.5 Cosmopolite 2.5 4.1 3.5 2.2 1.5 8.2 4.2 1.9 Balkan 1.9 2 4.1 7.8 8.1 2.1 6 2.6 Species diagnostic for one group Group 1 Quercus petraea agg. t 97 47 58 20 8 5 23 Tilia tomentosa s 49 13 4 13 1 2 Festuca drymeja h 22 2 13 Carex pilosa h 20 12 1 Campanula persicifolia h 40 19 14 14 3 11 10 Sorbus aucuparia s 11 3 Group 2 Quercus pedunculiflora et robur t 100 Quercus pedunculiflora et robur s 93 Arum maculatum h 100 3 2 3 3 Polygonatum latifolium h 2 80 2 11 10 1 Torilis japonica h 2 73 1 21 2 Stachys germanica h 67 4 3 2 Rosa corymbifera s 60 3 2 Ulmus minor s 7 80 7 2 15 4 1 Acer tataricum t 2 67 1 5 Ulmus minor t 1 60 1 2 10 6 Rosa gallica s 47 4 2 1 Heracleum sphondylium h 47 2 1 3 Geum urbanum h 12 93 24 44 9 44 10 6 Acer tataricum s 37 87 28 12 20 8 2 1 Asparagus tenuifolius h 2 73 1 5 17 3 13 17 Prunus spinosa agg. s 4 73 31 7 21 15 10 Peucedanum alsaticum h 27 Festuca gigantea h 27 Viola alba h 5 60 4 5 14 3 16 3 Crataegus pentagyna s 33 3 3 5 Malus sylvestris s 40 18 4 2 5 Solanum nigrum h 20 Alopecurus pratensis h 20 Buglossoides purpurocaerulea h 17 80 25 28 6 21 37 7 Poa angustifolia h 27 6 9 3 Sorghum halepense h 13 Euphorbia salicifolia h 13 Physalis alkekengi h 13 Calystegia sepium h 13 Doronicum hungaricum h 13 Tamus communis s 15 60 18 16 11 31 20 15 Cornus sanguinea s 17 53 19 7 6 21 16 17 Crataegus monogyna s 59 100 68 53 49 87 57 32 Group 3 Quercus frainetto t 6 47 66 35 3 Sorbus domestica s 7 7 31 9 3 6 3 Prunella vulgaris h 2 20 26 14 3 3 2 Veronica officinalis h 3 25 21 1 4 Moltkia doerfleri h 10 1 Group 4 Lathyrus laxiflorus h 1 7 49 1 1 Luzula forsteri h 1 12 49 3 {continued) 4 A. Carni et al. Table I. {Continued). Group number 1 2 3 4 5 6 7 8 Trifolium pignantii h 2 3 46 3 2 Veronica chamaedrys h 33 53 34 86 11 13 26 11 Campanula sparsa h 19 Helleborus cyclophyllus h 1 26 2 1 Aremonia agrimonoides h 5 10 49 3 9 18 Silene italica h 1 9 32 5 1 Trifolium patulum h 1 19 1 Pteridium aquilinum h 1 15 47 11 15 Potentilla micrantha h 10 34 54 3 5 15 9 Ptilotemon strictus h 16 1 Malus pumila s 1 18 1 1 Lychnis coronaria h 4 22 33 5 Physospermum cornubiense h 4 33 35 46 11 5 Carex muricata h 12 1 Asphodelus albus h 14 1 1 Cruciata laevipes h 2 40 15 37 14 1 Acer obtusatum h 1 4 35 11 15 Fagus sylvatica ssp. moesiaca s 4 3 23 3 1 6 Scutellaria columnae h 1 1 16 1 Chamaespartium sagittale h 1 14 1 1 Clinopodium vulgare h 24 40 46 67 3 13 29 22 Lapsana communis h 6 13 13 25 5 1 1 Astragalus glycyphyllos h 5 27 16 26 3 2 Carex caryophyllea h 12 18 2 Hieracium murorum h 3 9 26 3 12 Galium pseudaristatum h 3 16 21 2 2 Group 5 Artemisia alba h 2 34 3 1 1 Viola tricolor s. lat. h 1 31 1 1 Melica ciliata h 34 3 6 Rosa pimpinellifolia s 23 1 3 Euphorbia epithymoides h 2 20 1 1 Delphinium fissum h 17 1 Sedum ochroleucum h 17 3 1 Teucrium montanum h 1 3 23 1 7 Lychnis viscaria h 4 14 Alyssum murale h 11 Alyssum montanum h 11 Stipa pulcherrima h 11 Seseli peucedanoides h 11 Arabis hirsuta h 3 17 1 3 Rubus idaeus s 1 4 17 1 1 Rhamnus saxatilis s 1 1 23 8 3 5 Scleranthus serpentini h 11 1 Achillea clypeolata h 11 3 Potentilla argentea h 6 17 2 Juniperus oxycedrus h 1 5 31 17 5 Dianthus petraeus h 9 Hypericum rochelii h 9 Echium russicum h 9 Allium flavum h 9 Linaria angustissima h 9 Minuartia verna ssp. collina h 9 Sesleria nitida h 9 Thymus praecox h 9 Cotoneaster integerrimus h 9 Potentilla recta h 6 1 2 20 3 1 Coronilla varia h 7 7 1 12 26 3 2 6 Sesleria rigida h 9 1 Bromus riparius h 9 1 Euphrasia species h 1 9 Aurinia saxatilis h 9 3 {continued) Forests in Southeastern Europe 5 Table I. (Continued). Group number 1 2 3 4 5 6 7 8 Group 6 Prunus mahaleb t 62 1 Parietaria officinalis h 44 1 Sedum telephium h 36 Juglans regia t 38 2 Juglans regia s 13 2 38 1 3 Clematis vitalba s 3 22 11 20 85 25 22 Prunus mahaleb s 1 3 31 69 16 19 Cornus mas t 1 1 38 5 Celtis australis t 26 1 Vitis vinifera s 1 31 2 1 Tilia cordata s 3 31 5 Celtis australis s 26 2 Corylus avellana t 4 28 1 4 Fraxinus excelsior s 3 21 Asplenium ceterach h 1 6 54 17 19 Fraxinus excelsior t 21 1 2 Geranium robertianum h 1 6 2 36 4 12 Berberis vulgaris s 1 6 38 5 18 Tilia cordata t 4 1 28 8 Acer monspessulanum s 1 5 11 49 19 19 Viburnum lantana s 15 10 2 20 64 15 46 Carpinus orientalis s 25 31 16 20 72 37 8 Asperula purpurea h 1 11 36 13 6 Calamintha sylvatica h 6 7 5 3 36 10 5 Syringa vulgaris t 13 1 Crataegus nigra s 10 Ramonda serbica h 10 Viola suavis h 1 13 1 Acer monspessulanum t 4 9 38 22 6 Acer pseudoplatanus t 1 3 6 26 1 19 Asplenium trichomanes h 6 4 2 20 44 17 33 Group 7 Quercus pubescens t 25 27 12 21 23 83 32 Clematis flammula h 10 Paliurus spina-christi s 7 3 11 Dioscorea balcanica h 2 9 Group 8 Ostrya carpinifolia s 1 6 12 9 9 69 Amelanchier ovalis s 2 2 34 Ostrya carpinifolia t 1 9 7 30 75 Cyclamen purpurascens h 3 35 Sorbus aria s 7 1 3 5 1 39 Calamagrostis varia h 23 Sorbus aria t 1 6 3 33 Peucedanum oreoselinum h 1 3 5 35 Solidago virgaurea h 1 3 3 1 29 Erica herbacea s 3 21 Cotoneaster nebrodensis s 1 2 6 8 2 28 Mercurialis ovata h 1 7 30 Buphthalmum salicifolium h 3 2 5 27 Laserpitium siler h 15 Pinus nigra t 2 16 Clematis recta h 3 16 Convallaria majalis h 4 7 1 2 3 6 31 Asplenium ruta-muraria h 3 3 22 Anthericum ramosum h 1 1 5 24 Thalictrum minus h 1 3 3 2 20 Hepatica nobilis h 1 3 2 18 Campanula pyramidalis h 1 14 Anemone trifolia h 10 Sesleria tenuifolia h 1 11 (continued) 6 A. Čarni et cd. Table I. (Continued). Group number 1 2 3 4 5 6 7 8 Sorbus umbellata h 2 11 Carex humilis h 1 26 6 26 Frangula rupestris s 9 3 7 24 Phyteuma scheuchzeri ssp. columnae h 9 Lamiastrum galeobdolon agg. h 1 3 12 Spiraea chamaedryfolia s 11 3 15 Moehringia muscosa h 1 11 Thesium bavarum h 8 Helleborus dumetorum ssp. atrorubens h 7 Rubus saxatilis s 7 Aster amellus h 7 Sesleria albicans h 7 Polygala chamaebuxus h 7 Asperula cynanchica h 1 3 14 Polypodium vulgare h 1 2 9 3 2 17 Lonicera xylosteum h 6 2 10 7 25 Species diagnostic for more than one cluster Poa nemoralis agg. h 66 19 65 23 28 9 2 Quercus petraea agg. s 54 29 60 6 3 1 23 Tilia tomentosa t 40 1 5 Quercus cerris s 18 100 71 74 6 21 20 6 Quercus cerris t 37 100 87 91 18 43 17 Quercus jrainetto s 2 47 60 25 6 2 Euonymus europaeus s 6 73 10 9 6 44 11 15 Acer campestre t 16 67 28 2 6 59 14 5 Brachypodium sylvaticum h 13 80 50 44 23 69 34 17 Festuca heterophylla h 9 47 54 31 3 9 13 Primula veris s. lat. h 1 30 31 10 8 Syringa vulgaris s 2 3 100 72 7 1 Cotinus coggygria s 16 6 69 54 20 37 Carpinus orientalis t 26 12 5 74 49 3 Sesleria autumnalis h 1 7 5 48 58 Other species with high frequency Fraxinus ornus h 69 56 39 54 56 52 82 Fraxinus ornus t 58 41 9 23 67 70 68 Cornus mas h 47 46 51 14 67 57 34 Acer campestre s 44 67 40 32 3 62 24 13 Vincetoxicum hirundinaria h 34 40 26 5 57 13 26 51 Galium mollugo agg. h 36 7 26 33 54 31 24 38 Teucrium chamaedrys h 7 13 21 33 37 26 51 49 Fragaria vesca h 32 27 37 39 31 15 34 21 Dactylis glomerata h 40 51 54 26 13 33 17 Rubus jruticosus agg. h 27 27 47 32 26 46 14 12 Euphorbia cyparissias h 13 27 32 30 54 18 22 24 Helleborus odorus h 47 44 21 9 28 45 12 Lathyrus niger h 36 47 51 33 20 13 Pyrus pyraster s 18 47 32 30 11 28 11 10 Corylus avellana s 5 22 40 20 49 17 29 Sorbus torminalis h 41 37 44 9 5 15 21 Ligustrum vulgare h 17 20 34 12 9 23 34 23 Quercus pubescens agg. s 16 33 9 21 11 10 40 27 Hedera helix h 30 13 7 3 56 34 22 Rosa arvensis s 39 13 25 26 3 36 6 10 Euphorbia amygdaloides h 32 25 40 33 17 10 Viola reichenbachiana h 27 21 33 23 31 20 Melica uniflora h 21 6 47 11 33 24 8 Tanacetum corymbosum h 21 33 24 14 9 17 31 Euonymus verrucosus s 20 4 4 20 41 28 27 Chamaecytisus hirsutus h 31 19 12 17 10 10 39 Melittis melissophyllum h 5 24 26 3 10 28 37 Galium sylvaticum agg. h 23 43 11 9 5 11 29 Juniperus communis h 19 15 28 3 24 33 (continued) Forests in Southeastern Europe 7 Table I. (Continued). Group number 1 2 3 4 5 6 7 8 Polygonatum odoratum h 20 6 16 14 3 22 40 Hypericum perforatum h 28 13 16 5 29 13 10 6 Lathyrus venetus h 25 28 40 18 8 Viola hirta h 5 19 42 6 5 21 20 Geranium sanguineum h 10 30 34 14 30 Dactylis polygama h 20 40 13 26 9 8 2 Rhamnus catharticus h 9 33 1 9 15 16 23 Festuca valesiaca agg. h 7 13 21 11 20 28 5 Symphytum tuberosum h 2 24 28 14 20 16 Pyrus pyraster t 6 33 26 11 10 9 2 Species values are percentage frequencies. Diagnostic species for the groups of releves (defined as those with phi >25.0) are shown, ranked by decreasing value of the phi coefficient, indicated by shading: t, tree layer; s, shrub layer; h, herb layer. Groups of releves in the table correspond to the following, traditionally accepted alliances: lyQuercionpubescenti-sessiliflorae; 2, Aceri tatarici—Quercion; 3, Quercion confertae; 4, Quercion petraeae-cerris; 5, Pruno tenellae-Syringion; 6, Syringo-Carpinion Orientalin 7, Carpinon Orientalin and 8, Fraxino orni—Ostryion carpinifoliae. original data set (Tichý & Chytrý 2006). We also calculated Fischer's exact test and gave zero fidelity value to the species with significance P < 0.001. The threshold phi value for the species to be considered as diagnostic was set at 25.0. Since the diagnostic species are calculated on the basis of a matrix of thermophilous deciduous forests they are only used for the purpose of differentiating the stands within the order Quercetalia pubescentis (Knollová & Chytrý 2004). Species that appear in at least 50% of releves of an individual group are treated as constant. For further interpretation of the ecological conditions of the relevé groups, unweighted average indicator values were used (Zelnik & Carni 2008). Average values for the eight groups were calculated using Pignatti indicator values (Pignatti et al. 2005) in the JUICE programme, were passively projected onto the Principal Components Analysis (PCA) biplot (CANOCO 4.5; ter Braak & Šmilauer 2002) in order to reveal ecological relationships among these groups and thus explain environmental gradients underlying the main ordination axes. Square-root transformed percentage frequencies were used as the input data. We also determined a chorological spectrum of groups following Pignatti et al. (2005) and Josifovic (1970-1977). The endemic Balkan species were introduced and distinguished from northern endemics, i.e. SE European species (illyric species in the broader sense). Only the species present in at least three releves within the group were taken into consideration. The chorological spectrum of individual groups of releves is presented as proportions (percentages) of each group of species in the entire species composition of the group, and shown at the top of Table I. The nomenclature is according to Flora Europaea (Tutin et al. 1964-1980), except Scleranthus serpenti-ni Beck. Fagus moesiaca has been considered as Fagus sylvatica subsp. moesica (Gomory et al. 1999). The taxon Quercus virgiliana is treated as Quercus pub-escens (Skvorc et al. 2005). Results TWINSPAN groups of releves and their interpretation The dendrogram in Figure 1 shows eight groups of releves resulting from the TWINSPAN classification of the data set that are ecologically and phytogeo-graphically interpretable. On the first level, there is a divison of forest dominated by various oak species {Quercus frainetto, Q petraea, Q cerris) on the one hand, and those dominted by Quercus pubescens, Carpinus orientalis and Ostrya carpinifolia on the other. Futher interpretation of individual groups is given below. Group 1. Constant species in the tree layer are Quercus petraeae and Fraxinus ornus. The diagnostic species are Quercus petraea, Tilia tomentosa, Festuca dymeia, Carex pilosa, Campanula persicifolia and Sorbus aucuparia, which point out the thermophilous character of these communities that appear on deeper soils. The releves from this group appear on the edge of the Panonnian basin, in the northern part of the research area (Figure 2). The majority of releves from this group are traditionally classified in the alliance Quercion pubescenti-sessiliflorae. Group 2. Constant species in the tree layer are Quercus robus, Quercus cerris, Ulmus minor and Acer campestre. The diagnostic species are the most numerous in this group, from the point of view of distribution, species composition and ecological conditions. The diagnostic species are Quercus robur, Arum maculatum, Polygonatum latifolium, Torylis japonica, Stachys germanica and many others 8 A. Carni et al. showing a higher trophic status and humidity of sites. These communities can be found in the Pannonian basin (Figure 3). The releves from this group are traditionally classified in the alliance Aceri-Quercion. Group 3. Constant tree species in the tree layer are Quercus cerris and Q. frainetto. Among diagnostic species we can find Q. frainetto, Q. cerris, Festuca heterophylla, Prunella vulgaris, Sorbus domes-tica, and Veronica officinalis, which indicate warm sites with deeper soil horizons. In this, and in groups 4, 5 and 6, more Balkan species can be 1 8 Figure 1. Dendrogram of the TWINSPAN classification of thermophilous deciduous forests in Southeastern Europe corresponding to the eight alliances. found due to the distribution pattern of the releves. The releves from this group are mainly distributed in eastern Croatia, eastern Bosnia and Serbia (Figure 3). The majority of releves from this group are traditionally classified in the alliance Quercion confertae. Group 4. Q. petraea and Q. cerris are constant tree species in this group. The diagnostic species are numerous, indicating higher altitudes, more mesic and slightly acidic stands, including some Balkan endemics, such as Campanula sprasa, Lathyrus laxiflorus, Trifolium pignatti, Veronica chamaedrys, etc.; futher diagnostic species are listed in Table I. The releves are distributed in Macedonia, southern Serbia and Montenegro (Figure 2). The majority of releves from this group are traditionally classified in the alliance Quercion petraeae-cerris. Group 5. There is no constant species in the tree layer of this group, but Syringa vulgaris, Cotinus coggygria and Fraxinus ornus are constant in the scrub layer. Beside the physiognomical differences within group 6, the releves from this group have many diagnostic species indicating open, dry habitats and the relict position of stands, such as Artemisia alba, Melica ciliata, Rosa pimpinellifolia, Viola tricolor and many others listed in Table I. The releves from this group are distributed in the canyons of central Serbia (Figure 4). The majority of releves from this group are traditionally classified in the alliance Pruno tenellae-Syringion. Figure 2. The distribution of the releves of group 1 {Quercion pubescenti-sessiliflorae •) and group 4 {Quercion petraeae-cerris A)- Forests in Southeastern Europe 9 Group 6. The constant species in the tree layer are Carpinus orientalis, Fraxinus ornus, Primus mahaleb and Acer campestre. Diagnostic species show the thermophilous character of communities and often have a Mediterranean-continental distribution pattern, they are Juglans regia, Primus mahaleb, Parietaria officinalis, Sedum telephium, and others which are further enumerated in Table I. The releves from this group are found in the continental part of the Balkan Peninsula, in eastern Serbia (Figure 4). The majority of releves from this group are traditionally classified in the alliance Syringo-Carpinion orientalis. Group 7. The constant species in the tree layer is Quercus pubescens. Diagnostic species are Quercus pubescens, Clematis flammula, Paliurus spina-christi ® ROMA Figure 3. Distribution of the releves of group 2 {Aceri tatarici-Quercion •) and group 3 (Quercion confertae A)- Figure 4. Distribution of the releves from groups 5 {Pruno tenellae-Syringion +)3 6 (Syringo-Carpinion orientalis 7 {Carpinion orientalis •) and 8 {Fraxino orni—Ostryion carpinifoliae A)- 10 A. Cami et al. and Dioscorea balcanica that share a (sub)Mediterra-nean character and grow mainly on carbonate bedrock. This group comprises more Mediterranean species. In this group, and also in groups 6 and 8, there are more species with a SE European distribution pattern. There is also a higher proportion of Balkan species due to the location of some releves in the southern part of the region. The releves from this group are found along the Adriatic Sea and in regions where the maritime influence extends deep into the continent, along the rivers Una, Neretva, Drim and Vardar (Figure 4). The majority of releves from this group are traditionally classified in the alliance Carpinion orientalis. Group 8. The constant species in the tree layer is Ostrya carpinifolia. Diagnostic species (e.g. such as Ostrya carpinifolia, Amelanchier ovalis, Cyclamen purpurascens, Sorbus aria) are enumerated in Table I; they belong to heliophilous, thermophilous communities that grow on shallow soil horizons over carbonate bedrock. In this group more Mediterranean-montane species can be found. They are distributed in the inner part of the mountain ranges along the Adriatic coast (Figure 4). The majority of releves from this group are traditionally classified in the alliance Fraxino orni-Ostryon. PCA and indicator values The PCA of the eight groups with a passive projection of Pignatti indicator values (Fig ;ure 5) shows that the o OO 1 1 1 |3 1° Moisture i 7 1, 8 ° ° ^ Nutrients ------o-------- /1 \ / 1 / i 2o / i / 1 / i 5 6 | /° q O ob 1 Soil Reaction j Continentality 1 1 -1.0 1.0 Fig ure 5. PCA of eight TWINSPAN groups and mean Pignatti indicator values plotted as supplementary variables on the ordination diagram. Eigenvalues of first two axes are 0.279 and 0.226. The groups are numbered as in Table I. Only indicator values with the highest correlations with the first two PCA axes are shown. The highest correlations with the first axis have the indicator values for Nutrients (0.8210) and Moisture (0.7993), and with the second axis for Continentality (0.7099) and Soil reaction (0.6584). division along the first axis is based on moisture and nutrients. Communities dominated by oak species (groups 1-4) thrive on the most propitious stands with deep soil horizons while groups dominated by Quercus pubescens, Carpinus orientalis and Ostrya carpinifolia (groups 5-8) occur on the shallow soils. The second axis shows the separation on the basis of continentality and soil reaction. The most continental groups are 2, 5 and 6 found in the Panonnian plain and central part of the Balkan Peninsula, and the least continental ones are those from precipitation-rich mountain ranges bordering on the eastern Adriatic coast (groups 4 and 8). The soil reaction shows the highest pH within groups 5-8 dominated by Quercus pubescens, Carpinus orientalis and Ostrya carpinifolia, thriving on shallow soils over carbonate bedrock, and the lowest in Quercus-dominated groups 1-4, appearing in more acid sites with deeper soil horizons. Discussion After the numerical analysis, and the calculation of diagnostic species and distribution of releves, confirmed also by the correlation with Pignatti indicator values and the chorological spectrum, it can be concluded that the groups generally correspond to the traditionally recognised alliances. General view of communities The general view of communties (Figure 1, Table I) shows that the whole data set, i.e. order Quercetalia pubescentis, can be divided into two parts (groups of alliances). One is dominated by various deciduous species of the genus Quercus (except the calciphilic Q. pubescens). The other is formed by communities dominated by Carpinus orientalis, Ostrya carpinifolia and Quercus pubescens. Such a division has already been established by Lakusic et al. (1982), who separated these two groups as orders and described a new order, Ostryo-Carpinetalia orientalis. The group of oak forests {Quercion pubescenti-sessiliflorae, Aceri-Quercion, Quercion confertae, and Quercion petraeae-cerris) appears in sites with deeper, slightly acid soil horizons. Quercus petraea is diagnostic for alliances from the (sub)montane vegetation belt {Quercion petraeae-sessiliflorae and Quercion petraeae-cerris), Quercus cerris is diagnostic for Balkan and Panonnian alliances Quercion confertae, Quercion petraeae-cerris and Aceri Quercion. Quercus frainetto is diagnostic for Balkan and Panonnian lowland alliances, i.e. Aceri-Quercion and Quercion confertae. The other group of forests dominated by Quercus pubescens, Carpinus orientalis and Ostrya carpinofolia is spread in the areas under the Mediterranean influence. In the coastal region, the most important Forests in Southeastern Europe 11 tree species is Quercus pubescens (Carpinion orientalise, that is substituted by Ostrya carpiniofolia (Fraxino orni—Ostryion) in the (sub)montane vegetation belt, while Carpinus orientalis is characteristic for the coastal (Carpinion orientalis) as well as for continental communities (Syringo-Carpinion). The diagnostic values of constant tree species also reflect the ecological amplitude of the alliances. In the syntaxonomic scheme the Pruno tenellae— Syringion is separated from the order of thermophi-lous deciduous forests of the Quercetalia pubescentis and classified in the Fraxino orni—Cotinetalia order of thermophilous deciduous scrub communities, based on physiognomic and floristic differences. Proposed syntaxonomic scheme Quercion pubescenti-sessiliflorae Br.-Bl 1932 Aceri tatarici-Quercion Zolyomi et Jakucs 1957 Quercion confertae Horvat 1954 Quercion petraeae-cerris (Lakusic et Jovanovic 1980) all. nova Syringo-Carpinion orientalis Jakucs 1959 Carpinon orientalis Horvat 1954 Fraxino orni-Ostryion carpinifoliae Tomazic 1940 Quercetalia pubescentis Klika 1933 Pruno tenellae-Syringion (Jovanovic 1979) all. nova Fraxino orni-Cotinetalia Jakucs 1961 Querco-Fagetea Br.-Bl. et Vieger in Vlieger 1937 Description of the syntaxa Quercion pubescenti-sessiliflorae is an alliance with a Central European distribution pattern and has the most southern irradiation in this region; Tilia tometosa and Acer tataricum can form the basis for a geographical differentiation of the alliance (Bar-icevic & Vukelic 2006; Baricevic et al. 2006). Aceri tatarici-Quercion is the most continental alliance and is distributed in the Pannonian plain (Jovanovic 1997); Quercion confertae is an alliance distributed in the lowlands under the influence of the continental climate (Jovanovic 1997). Quercion petraeae-cerris is found at higher altitudes in the eastern and southern part of the study area in contact with beech forests (Kojic et al. 1998, Em 1964); Syringo-Carpinion orientalis is an extrazonal Mediterranean vegetation thriving in the continental part of the central Balkans (Misic 1981, 1997). It is confirmed based also on differences in floristic composition and physiognomy that the Pruno tenellae-Syringion forms a separate syntaxon thriving on sunny, steep slopes (Misic 1981). Carpinion orientalis is found in the waste areas under the influence of the Mediterranean climate; on more humid sites along the Adriatic coast it is classified as Ostyo-Carpinenion orientalis and in drier sites in the southern part of the Balkans as Syringo-Carpinenion orientalis (Horvat et al. 1974; Poldini 1988), Fraxino omi-Ostryion is found in the inner part of the mountain chains along the Adriatic coast at higher altitudes showing some similarities to the vegetation of the Erico-Pinetea (Horvat 1959; Tomazic 1940; Wallnofer 1993). Description of new syntaxa Quercion petraeae-cerris (Lakusic et Jovanovic 1980) all. nova hoc loco Holotypus: Quercetum cerris Vukicevic 1966 holotypus hoc loco The original description is not valid according to Article 1 of the ICPN. The description of the alliance corresponds to the description of group 4; diagnostic species are indicated in Table I. Pruno tenellae-Syringion (Jovanovic 1979) all. nova hoc loco. Holotypus: Artemisio camphoratae—Amygdaletum na-nae Jovanovic 1954 holotypus hoc loco The original description is not valid according to Article 1 of the ICPN. The description of the alliance corresponds to the description of group 5; diagnostic species are indicated in Table I. Conclusions This work presents the first survey of thermophilous deciduous forests in the northwestern part of Southeastern Europe based on the numerical approach and can form the basis for futher elaboration. A revision of the nomenclature of associations is needed, since the one based on geographical and ecological epithets is still widely used. Further research is needed to analyse and to understand the ecological and phytogeographical circumstances within this type of vegetation in the context of a wider region, even on the European scale. For the region under study connecting Central Europe and the eastern Balkans, it is important to meet the common European standards. In recent years, there have appeared some publications dealing with the forest vegetation of Southeastern Europe (e.g. Bergmeier & Dimopoulus 2001; Dring et al. 2002; Bergmeier et al. 2004; Chasapis et al. 2004; Roussakova & Dimitrov 2005; Tzonev et al. 2006) that provide a basis for further research and synthesis. New challenges are also being faced with regard to current research projects that have revealed common features on both sides of the Adriatic Sea, on the 12 A. Čami et al. peninsulas of Southern Europe, i.e. the Balkans and the Apennines. They have a similar geographical position, and development of flora and vegetation. In both regions, there are beech forests of the alliance Aremonio-Fagion (Biondi et al. 2002; Blasi et al. 2005), and in their southern parts thermophilous beech forests of the alliance Geranio versicoloris-Fagion (Bergmeier & Dimopoulus 2001; Bergmeier et al. 2004; Di Pietro et al. 2004; Rosati et al. 2005). At the same time we can find the same suballiances in noble hardwood forests of the Tilio-Acerion (Kosir et al. 2008) and Carpinus betulus forests of the Erythronio-Carpinion (Biondi et al. 2002; Marincek & Carni 2000). Within the order Quercetalia pubescentis there are suballiances of the Carpinion orientalis on both sides of the Adriatic Sea, and a certain degree of similarity also applies to the alliances Quercion confertae and Quercion petraeae-cerris, whereas the alliance Quercion pubescenti-sessiliflorae appears in the northern parts of both regions (Pignatti & Pignatti-Wikus 1987; Poldini 1988; Blasi et al. 2001, 2004; Ubaldi 2003). These studies show that a common European work on these topics is needed. Acknowledgements We would like to express our thanks for technical assistance to Barbara Sustar and Iztok Sajko. We would like to thank Lubomir Tichy for help provided during the calculation of the Pignatti bioindicator values and Romeo di Pietro for his encouragement. The preparation of this article was co-financed by various projects of the Slovenian Research Agency (to AC and PK). References Baricevic D, Vukelic J. 2006. Flora of the order Quercetalia pubescentis Br.-Bl. (1931) 1932 in the forest vegetation of the Pozega hill area (NE Croatia). Acta Bot Croat 65(1): 67-81. Baricevic D, Vukelic J, Pernar N, Baksic, Sango M. 2006. Association Lathyro-Quercetum petraeae I. Horvat (1938) 1958 in the Pozega hill area and its comparison with other distribution areas in Croatia. Period Biol 108(6): 683-692. Bergmeier E, Dimopoulos P. 2001. Fagus sylvatica forest vegetation in Greece: Syntaxonomy and gradient analysis. J Veg Sei 12: 109-126. Bergmeier E, Dimopoulos P. 2008. Identifying plant communities of thermophilous deciduous forest in Greece: Species composition, distribution, ecology and syntaxonomy. Plant Biosyst 142: 228-254. Bergmeier E, Dimopoulos P, Theodoropoulos K, Eleftheriadou E. 2004. Zonale sommergrüne Laubwälder der südlichen Balk-anhalbinsel-eine Übersicht. Tuexenia 24: 89-111. Biondi E, Casavecchia S, Pinzi M, Allegrezza M, Baldoni M. 2002. The syntaxonomy of the mesophilous woods of the Central and Northern Apennines (Italy). Fitosociologia 39(2): 71-93. Blasi C, Di Pietro R, Filesi L. 2004. Syntaxonomical revision of Quercetalia pubesceti-petraeae in the Italian Peninsula. Fitosociologia 41(1): 87-164. Blasi C, Di Pietro R, Filesi L, Fortini P. 2001. Syntaxonomy, chorology and syndynamics of Carpinus orientalis communities in Central Italy. Phytocoenologia 31(1): 33-62. Blasi C, Fortini P, Grossi G, Presti G. 2005. Faggete e cerrete mesofile nell'Alto Molise. Fitosociologia 42(2): 67-81. Braun-Blanquet J. 1964. Pflanzensoziologie. Grundzüge der Vegetationskunde. 3rd ed. Wien: Springer. Chasapis M, Karagiannakidou V, Theodoropoulus K. 2004. Phytosociological research of Quercus coeeifera L. pseudoma-quis on Mount Chortiatis, northern Greece. Isr J Plant Sei 52: 357-381. Chytrý M, Tichý L, Holt J, Botta-Dukát Z. 2002. Determination of diagnostic species with statistical fidelity measures. J Veg Sei 13: 79-90. Di Pietro R, Izco J, Blasi C. 2004. Contribution to the nomenclatural knowledge of Fagus sylvatica woodlands of southern Italy. Plant Biosyst 138(1): 27-36. Dring J, Hoda P, Mersinllari M, Mullaj A, Pignatti S, Rodwell J. 2002. Plant communities of Albania - a preliminary overview. Ann Bot 2: 7-30. Em H. 1964. Cerovi šumski zaednici kako oddělen visočinski pojac po nekoi makedonski planini. God zbor Zemjodel šum fak 17: 256-261. Gömöry D, Paule L, Brus R, Zhelev P, Tomovic Z, Gračan J. 1999. Genetic differentiation and phylogeny of beech on the Balkan penninsula. J Evol Biol 12: 746-754. Havlová M., Chytrý M., Tichý L. 2004. Diversity of hay meadows in the Czech Republic: Major types and environmental gradients. Phytocoenologia 34(4): 551-567. Hennekens SM, Schaminée JHJ. 2001. TURBOVEG, a comprehensive data base management system for vegetation data. J Veg Sei 12(4): 589-591. Hill MO. 1979. TWINSPAN - a Fortran program for arranging multivariate data in an ordered two way table by classification of the individuals and the attributes. Department of Ecology and Systematics, Cornell University, Ithaca, NY. Horvat I. 1938. Biljnosociološka istraživanja šuma u Hrvatskoj. Glasnikza šumske pokuse 6(48): 279 p. Horvat I. 1959. Sistematski odnosi termofilnih hrastovih i borovih šuma jugoistočne Evropě. Biološki glasnik 12: 1-40. Horvat I, Glavač V, Ellenberg H, 1974. Vegetation Südosteuropas. Geobotanica selecta 10. Stuttgart: Gustav Fischer Verlag. Josifovič M. 1970-1977. Flora SR Srbije 1-9. Beograd: Srpska akademija nauka i umetnosti. Jovanovic B. 1997. Razred evrosibirskih listopadnih šuma. Querco- Fagetea Br.-Bl. et Vlieg. 1937. In: Sarič M, editor. Vegetacija Srbije II, šumske zajednice 1. Beograd: Srpska akademija nauka i umetnosti. pp. 1-106. Jovanovic B, Lakušič R, Rizovski R, Trinjastič I, Zupančič M. 1986. Prodromus phytocoenosum Jugoslaviae. Naučno vece vegetacijske karte Jugoslavije, Bribir-Ilok. Kevey B, Borhidi A. 2005. The acidophilous forets of the Mecsek and their relationship with Balkan-Pannonian acidophilous forests. Acta Bot Hung 47: 273-368. Knollová I, Chytrý M. 2004. Oak-hornbeam forests of the Czech Republic: Geographical and ecological approaches to vegetation classification. Preslia 76: 291-311. Kojič M, Popovic R, Karadzic B. 1998. Sintaksonomski pregled vegetacije Srbije. Beograd: Institut za biološka istraživanja "Siniša Stankovič". Košir P, Carni A, Di Pietro R. 2008. Classification and phytogeographical differentiation of broad-leaved ravine forests in southeastern Europe. J Veg Sei 19: 331-342. Forests in Southeastern Europe 13 Lakušič R, Pavlovic D, Redžič S. 1982. Horološko-ekološka i floristická diferenciacija šuma i šikara sa bjelograbicem (Carpinus orientalis Mill.) i crnim grabom (Ostrya carpinifolia Scop.) na prostoru Jugoslavije. Glas Rep zav zaš Pri Prirod muze 15: 103-116. Marinček L, Carni A. 2000. Die Unterverbände der Hainbuchenwälder des Verbandes Erythronio-Carpinion betuli (Horvat 1938) Marinček in Wallnöfer, Mucina et Grass 1993. Scopolia 45: 1-20. Mišič V. 1981. Šumska vegetacija klisura i kanjona istočne Srbije. Beograd: Institut za biološka istraživanja "Siniša Stankovic". Mišič V. 1997. Red medunčevih šuma Quercetalia pubescentis Br.- Bl. (1931)1932. Sveza šuma grabiča sa jorgovanom Syringo- Carpinion orientalis Jakucs 1959. In: Sarič M, editor. Vegetacija Srbije II, šumske zajednice 1. Beograd: Srpska akademija nauka i umetnosti. pp. 311-347. Mucina L. 1997. Conspectus of classes of European vegetation. Folia Geobot Phytotax 32: 117-172. Oberdorfer E. 1992. Süddeutsche Pflanzengesellschaften. Wälder und Gebüsche. Jena: Gustav Fischer. Pignatti S, Menegoni P, Pietrosanti S. 2005. Valori di bioindica- zione delle piante vascolari della flora d'Italia. Braun-Blanque- tia 39: 1-97. Pignatti S, Pignatti-Wikus E. 1987. Le conosi a cerro e frainetto della penisola e della Sicilia. Not Soc Ital Fitosoc 23: 107-124. Poldini L. 1988. Übersicht des Verbandes Ostryo-Carpinion orientalis (Quercetalia pubescentis) in SO-Europa. Phytocoeno-logia 16(1): 125-143. Rivas-Martinez S, Diaz TE, Fernández-Gonzáles F, Izco J, Loidi J, Lousä M, Penas A. 2002. Vascular plant communities of Spain and Portugal. Addenda to the syntaxonomical checklist of 2001. Itin Geobot 15: 5-435. Rivas-Martinez S, Fernández-Gonzáles F, Loidi J, Lousä M, Penas A. 2001. Syntaxonomical checklist of vascular plant communities of Spain and Portugal to association level. Itin Geobot 14: 5-341. Rivas-Martinez S, Penas A, Diaz TE. 2004. Biogeographic map of Europe. Cartographic Service, University of Leon, Spain. Rodwell JS, Schaminée JHJ, Mucina L, Pignatti S, Dring J, Moss D. 2002. The diversity of European vegetation. An overview of phytosociological alliances and their relationship to EUNIS habitats. Wageningen, NL. Report EC-LNW nr. 2002/ 054. Roleček J. 2005. Vegetation types of dry-mesic oak forests in Slovakia. Preslia 77: 241-261. Rosati L, Di Pietro R, Blasi C. 2005. La vegetazione forestale delle regione Temperata del "Flysch del Cilento" (Italia meridio- nale). Fitosociologia 42(2): 33-65. Rousakova V, Tzonev R. 2003. Syntaxonomy of the oak forests in the Pleven district (Danube plain in Bulgaria). Fitosociologia 40(1): 23-31. Roussakova V, Dimitrov M. 2005. Caractéristiques phytosociolo- giques des forěts á Picea abies et Abies alba dans la montagne de Rila (Bulgarie). Acta Bot Gall 152: 563-571. Rudski I. 1949. Tipovi liščarskih šuma jugoistočnog dela Sumadije. Prirodnjački muzej srpske zemlje Posebno izdanje 25. Škvore Z, Franjič J, Idžojtič M. 2005. Population structure of Quercus pubescens Willd. (Fagaceae) in Croatia according to morphology of leaves. Acta Bot Hung 47: 193-206. Ter Braak JFC, Šmilauer P. 2002. CANOCO Reference Manual and CanoDraw for Windows User's Guide. Software for Canonical Community Ordination (version 4.5). Microcomputer Power. Ithaca, NY. Tichý L. 2002. JUICE, software for vegetation classification. J Veg Sei 13: 451-453. Tichý L, Chytrý M. 2006. Statistical determination of diagnostic species for site groups of unequal size. J Veg Sei 17: 809-818. Tomažič G. 1939. Splošen pregled vegetacije iz razreda Querco- Fagetales v Sloveniji. Zbor Prirod druš 1: 43-49. Tomažič G. 1940. Asociacije borovih gozdov v Sloveniji. I. Bazifilni borovi gozdovi. Razprazu 1: 77-120. Tutin TG, Heywood VH, Burges NA, Moore DM, Valentine DH, Walters SM, Webb DA, editors. 1964-1980. Flora Europaea. Volumes 1-5. Cambridge: Cambridge University Press. Tzonev R, Dimitrov M, Chytrý M, Roussakova V, Dimova D, Gussev C, Pavlov D, Vulchev V, Vítkova A, Gogoushev G, et al. 2006. Beech forest communities in Bulgaria. Phytocoe- nologia 36: 247-279. Ubaldi D. 2003. La vegetazione boschiva d'Italia. Manuale di Fitosociologia forestale. Bologna: Clueb. Wallnöfer S. 1993. Erico-Pinetea. In: Mucina L, Grabherr G, Wallnöfer S, editors. Die Pflanzengesellschaften Österreichs. Teil III. Wälder und Gebüsche. Jena: Gustav Fischer. Willner W, Grabherr G. 2007. Die Wälder und Gebüsche Österreichs. Spektrum, Akademischer Verlag. Zelnik I, Carni A. 2008. Wet meadows of the alliance Molinion Koch 1926 and their environmental gradients in Slovenia. Biologia 63(2): 187-196.