Applied Vegetation Science 16 (2013) 131 -147 Geographical and ecological differentiation of Fagus forest vegetation in SE Europe A. Marinsek, U. Silc & A. Carni Keywords Balkans; Beech forest; Chorotypes; Classification; Gradients; Phytogeography; Plant traits; Syntaxonomy Abbreviations BF = beech forests; EIV = ecological indicator values; ICPN = International Code of Phytosociological Nomenclature (Weber et al. 2000) Nomenclature Flora Europaea (Tutin et al. 1964-1993); International Code of Phytosociological Nomenclature (Weber et al. 2000). Received 8 August 2011 Accepted 30 March 2012 Co-ordinating Editor: Joop Schaminee Marinsek, A. (Corresponding author, marinsek@zrc-sazu.si), Silc, U. (urban® zrc-sazu.si) & Carni, A. (carni@zrc-sazu.si): Institute of Biology, Scientific Research Centre of the Slovenian Academy of Sciences and Arts, Novi trg2, Ljubljana, 1000, Slovenia Carni, A.: University of Nova Gorica, Vipavska 13, Nova Gorica, 5000, Slovenia Abstract Questions: What is the main syntaxonomical pattern within beech forests in SE Europe? What macroecological and ecological factors distinguish these forests? Location: SE Europe: Balkan Peninsula, from the SE Alps in Slovenia, through Croatia, Bosnia and Herzegovina, Serbia, Montenegro and the Republic of Macedonia to N and NE Greece and Bulgaria, covering ca. 400 000 km2 over a length of 1000 km. Methods: With a view to differentiating beech and beech-fir forests, a data set of 5952 published and unpublished phytosociological releves were surveyed. After stratification, 997 releves remained. Cluster analysis of the data set was used to calculate diagnostic species for each cluster. Ecological indicator values (EIV) were used to estimate ecological conditions. Average EIV, altitude, latitude and longitude for releves of each cluster were plotted in a detrended correspondence analysis (DCA) diagram for ecological interpretation of clusters and relationships between clusters. Correlations between DCA releve scores and explanatory variables (EIV, portion of life forms and chorotypes, altitude, latitude and longitude) were subsequently calculated. Results: Cluster analysis divided mesophilous beech forests of SE Europe into two major clusters. Beech forests can therefore be classified into two alliances, Aremonio-Fagion and Fagion moesiacae. Further division revealed seven beech and beech-fir forest types, which we interpreted geographically and ecologically. A significant increase in the proportion of chamaephytes, hemicryptophytes and therophytes was detected along the main macroecological gradient towards the S and E. At the same time, the proportion of geophytes and phanerophytes significantly decreased in the same direction. There was also a significant increase in the proportion of Stenomediterranean, Eurymediterranean, Mediterranean-Montane, and Eurasian species, while Boreal species, as expected, decreased toward the southeast. The main differentiation of beech forests in SE Europe is due to macroecological factors (macro-climatic and historical development of vegetation), whereas local ecological factors (particularly temperature and moisture) are reflected in the differentiation of sub-alliances. Conclusions: Our study confirmed two major groups of beech forests in the research area, which could be classified into two alliances. It also revealed that there is not just an altitudinal distribution of beech forests in the SE part of the research area, but also structural and functional changes of communities as a result of the altitudinal limitation of beech forests and changed macroclimatic factors. Introduction Beech forests (BF) make up a remarkably high proportion of the European forest landscape (Bohn et al. 2004). In Central Europe, BF occupy various sites and have a wide altitudinal range, while in S Europe, within their range limit, they can only be found in humid mountain areas (Bergmeier & Dimopoulos 2001; Dierschke & Bohn 2004). Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 131 Differentiation of Fagus forests in SE Europe A. Marinšek et al. In general, differences among Fagus forests are due to broad scale (historic, phytogeographic, macroclimatic and macroecological) and regional (edaphic, meso-climatic and ecological) factors (Bergmeier & Dimopoulos 2001). Soil ecology is usually considered to be the principal factor on a regional scale (Ellenberg 1996), while on a broader scale macroecological (geography and climate) differentiation has precedence (Dierschke 1990; Dierschke & Bohn 2004). Dierschke & Bohn (2004) proposed a differentiation of European BF into nine regional, geographically based alliances, with subsequent partition towards various sub-alliances, based on the combination and gradual disappearance of several groups of plant species, due to changed ecological factors. Which factors should be considered more important for classification is an ongoing topic among syntaxonomists dealing with the classification of European Fagus forests (Soó 1964; Horvat et al. 1974; Török et al. 1989; Dierschke 1990; Dierschke & Bohn 2004). Willner (2002) and Tzonev et al. (2006) follow an approach based on ecological factors, while Dierschke (1990), Dierschke & Bohn (2004), Dzwonko & Loster (2000) and Bergmeier & Dimopoulos (2001) emphasize geographical differentiation. Various comparative studies suggest that classifications based on both ecological and geographical differentiation are generally more adequate than those that consider ecological or geographical differences alone (e.g. Dzwonko & Loster 2000). The syntaxonomy of BF of the Balkans is far less clear than that in Central and Western Europe (Dzwonko & Loster 2000). A number of different classifications exist, mostly regional (Dzwonko & Loster 2000; Bergmeier & Dimopoulos 2001; Tzonev et al. 2006; Tsiripidis et al. 2007), as well as several check lists (Stefanovič 1986; Mar-inček et al. 1992; Vukelič & Baričevič 2002; Redžič 2007; Rexhepi 2007; Trinajstič 2008; Šik & Čarni 2012), but without a synthesis over the whole area. In earlier studies, BF of the investigated area were distinguished as regional alliances of Fagus sylvatica forests. Horvat (1938) mentioned the possibility of including forests from the SE Alps to Albania and Greece in the special genetic-geographic group called Fagion sylvaticae illyricum. Horvat (1950) later classified beech forests from Macedonia and Serbia into the Fagion illyricum alliance and pointed out their different floristic composition and smaller number of Illyrian elements in comparison with BF in the N Balkans (Slovenia and Croatia). BF of SE Serbia and Bulgaria were assigned by Soó (1963, 1964) to the Fagion daci-cum alliance, although many diagnostic species of this alliance are not present in these areas. He designated them Moesian BF, with Fagus moesiaca as the differential species. Dafis (1973) used the name Fagion moesiacae to comprise Hellenic Fagus forests (today it is designated Geranio versicol-oris-Fagion), although the alliance Fagion moesiacae was considered valid in Blecic & Lakusic (1970) for Montenegro. Horvat et al. (1974) analysed some new data and described and validated the Fagion moesiacum alliance, distributed in the C and E Balkans, but they used an illegitimate name, as Fukarek (1969) had invalidly used the same name for the alliance even earlier - for acidophilous BF. They delineated the geographical range of the Fagion moesiacum alliance according to the range of the putative taxon Fagus moesiaca. This opinion was largely accepted for the region (Jovanovic et al. 1986). The illegitimately described name Fagion illyricum was replaced (Török et al. 1989) with a new name, Aremonio-Fagion. A nomencla-tural revision of the associations classified within the alliance Aremonio-Fagion was made by Marincek et al. (1992). During our research, we were faced with difficulties because there was no unified viewpoint concerning the syntaxonomical classification of BF, and we found that these forests have been classified into various alliances. Many authors have examined BF in the C and E part of the Balkans in the last 10 years. The proposed classification of BF of SE Serbia, the Republic of Macedonia and N and C Greece by Dzwonko & Loster (2000) did not consider Fagion moesiacae and included BF in the Aremonio-Fagion alliance. Bergmeier & Dimopoulos (2001) classified Greek BF into Fagion sylvaticae. Rodwell et al. (2002) proposed that beech and fir-beech forests of the area E of the River Drina and on the Rhodope Mountains should be classified into Doronico orientalis-Fagion moesiacae. Dierschke & Bohn (2004) placed BF of this area into two alliances: those of the C Balkans (from S Serbia and Macedonia to W Bulgaria) into Doronico columnae-Fagion moesiacae and those of NE and C Greece into Doronico orientalis-Fagion moesiacae. Tzonev et al. (2006) also did not support the concept of the alliance Fagion moesiacae and pointed to the relationship of Bulgarian mesophilous and acidophilous BF to Doronico columnae-Fagenion moesiacae and thermophilous BF to Doronico orientalis-Fagenion moesiacae but they included Bulgarian BF in three different alliances (Luzulo-Fagion, Asperulo-Fagion and Cephalanthero-Fagion) that were very close to the proposal of Willner (2002) for the S-C European beech forests. The name Fagion moesiacae was not considered in the case of BF in Bosnia and Herzegovina and was replaced with a new but invalid name, Seslerio-Fagion sylvaticae (Redzic & Barudanovic 2010). Faced with a similar question to that of Bergmeier & Dimopoulos (2001), as to how Greek Fagus forest communities syntaxonomically correspond to their Balkan and European counterparts, we tried to answer the question within the framework of SE Europe. The goals of our study were: (1) to establish the main vegetation types of BF communities in SE Europe (excluding acidophilous types) and to discuss the possible syntaxonomical interpretations of the distinguished vegetation 132 Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.X © 2012 International Association for Vegetation Science A. Marinseket al. Differentiation ofFagus forests in SE Europe types; and (2) to detect the major factors that influence BF vegetation in SE Europe. Taxonomic remarks At the subspecies level of Vagus taxa, the situation in the investigated area is complicated. There are two well-defined subspecies: Vagus sylvatka subsp. sylvatica and V. sylvatica subsp. orientalis (Denk 1999) but, in 1933, Czec-zott reported Vagus moesiaca in the Balkan Peninsula as a species having intermediate morphological characters between these two subspecies. In terms of the latest and most accepted taxonomic classification, it is a species whose taxonomic status is still unclear and it has been the subject of a number of studies (Gomory et al. 1999; Magri et al. 2006; Gomory & Paule 2010). Many authors, mainly from Serbia, still distinguish it (Cvjeticanin 2003; Cvjeticanin & Novakovic 2004; Curovic et al. 2011) and many authors allow its existence with some reservations (Gomory et al. 1999; Magri et al. 2006; Brus 2010; Gomory & Paule 2010). Distinguishing V. moesiaca as a separate taxon (whatever the rank) does not seem justified in the opinion of Gomory & Paule (2010). If such a taxon was to be used for European populations originating from Balkan glacial beech refugia, then it must be reserved for populations in the very southern part of the Balkans, as proposed by Magri et al. (2006). Methods Our study comprises forests of V. sylvatica subsp. sylvatica in the SE part of Europe; from the SE Alps in Slovenia, through Croatia, Bosnia and Herzegovina, Serbia, Montenegro, Bulgaria and Macedonia to N and NE Greece (Kg- 1). In terms of the long gradient (approximately 1000 km), climatic conditions in the SE are different from those in the NW. Generally, the temperature in the S part is higher and the precipitation is lower. A sub-mediterranean mountainous climate prevails, characterized by high winter precipitation and markedly low summer precipitation (Fig. 2). Published and unpublished releves of BF made in this area were taken into consideration for the purpose of the study (Appendix SI). All the releves were made according to the Braun-Blanquet approach (Braun-Blanquet 1964) and were stored in the TURBOVEG database (Hennekens & Schaminee 2001). From the total number of collected releves (n = 5952), we selected those of BF in which V. sylvatica subsp. sylvatica had a cover value of at least two in the tree layer according to the Braun-Blanquet scale. We treated V. moesiaca as V. sylvatica subsp. sylvatica, as many authors have also done in previous studies (Dzwonko et al. 1999; Bergmeier & Dimopoulos 2001; Fig. 1. Study area on a segment of the vegetation map of SE Europe (Bohn et al. 2004) with the distribution of beech forests (dark grey) and symbols that mark positions of each beech forest vegetation type separately. Legend: squares - alliance Aremonio Fagion, circles - alliance Fagion moesiacae, a - Cluster 1, ■ - Cluster 2,0- Cluster 3, Cluster 4, O - Cluster 5, o - Cluster 6, •- Cluster 7. Symbols for clusters are placed subjectively on the map in relation to the highest frequency of occurrence. Tzonev et al. 2006). Vagus sylvatica subsp. orientalis forests were not within the focus of our research and were not taken into consideration. We also omitted releves of the alliances Luzulo-Vagion and Geranio versicolor-Vagion. The first was omitted because of the acidophilous site characteristics of BF combined in this alliance. The second, BF of the alliance Geranio versicolor-Vagion, appear in NW Greece and show a trans-Adriatic distribution pattern with considerable floristic deviation from other Vagus forest types in eastern parts of N Greece (Bergmeier & Dimopoulos 2001; Di Pietro 2009). It was first described as Vagion hellenicum (Quezel 1967). In order to avoid an unequal proportion of communities with a high number of samples (oversampling), where possible we chose a maximum ten releves from each beech forest association, defined by name. Selection was made in such a way that different authors, different publications and different locations within the area were represented (Kosir et al. 2008; Carni et al. 2009). We a posteriori georef erenced 997 releves that remained after this selection. Where the occurrence of individual species was specified for different layers, all strata were amalgamated into one layer, by default function in the JUICE 7.0 program (Tichy 2002). In order to reduce noise in the analysis, taxa occurring in six or fewer releves were omitted from the analysis (Tsiripidis et al. 2007). Taxa treated at different taxonomic levels (e.g. subspecies, variety) were aggregated to the upper level. Records of species determined to genus level were deleted from the data set. We also excluded moss and lichen species, Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 133 Differentiation of Fagus forests in SE Europe A. Marinšek et al. (a) ISO -0- LI (fflO) -O OH o a ,*■. ' t ■ 'i □ i o—a y > 0.20 are included among diagnostic species. Diagnostic species of alliances are indicated in the first column. Diagnostic spec, for alliances Alliance AF Fm Group No. (Clusters) 1 2 3 4 5 6 7 No. of releves 97 80 78 114 263 219 146 Average number of species 36 35 37 37 24 27 13 Average number of beech forest species 18 17 11 13 1 9 3 Montane BF (Lamio orvalae-Fagenion) AF Corydalis cava 36 - - - 3 2 - Isopyrum thalictroides 36 1 - 2 3 2 - Pahs quadri folia 72 40 14 25 7 13 - Arum maculatum 42 9 3 - 11 9 1 Galanthus nivalis 25 1 4 1 2 1 - Leucojum Vernum 19 2 - 1 - 1 - Cardamine waldsteinii 26 9 - 3 1 2 - Sambucus nigra 42 18 6 3 18 8 1 Cardamine bulbifera 80 39 15 29 40 64 19 Cardamine kitaibelli 19 5 1 - 1 1 - Lunaria rediviva 20 8 - - 2 1 - Adoxa moschatellina 20 4 1 4 1 4 - Asplenium scolopendrium 26 16 5 - 5 1 - Dryopteris filix-mas 84 72 28 46 38 58 20 AF Acer pseudoplatanus 86 69 72 61 52 20 10 Stellaria nemorum subsp. glochidisperma 8 - - - - - - Allium ursinum 22 5 3 3 2 9 - Polygonatum multiflorum 43 20 37 8 19 1 - Vicia oroboides 22 - 5 12 4 - 1 AF Lamium orvala 36 28 21 12 3 3 1 AF Ranunculus lanuginosus 18 6 - 12 1 - - Anemone ranunculoides 14 1 - 2 1 9 - Corydalis solida 9 - - - 2 2 1 Anemone nemorosa 55 42 21 44 20 32 1 AF Senecio nemorensis subsp. fuchsii 32 12 27 24 4 - - Ranunculus ficaria 9 - - - 4 1 - Ulmus glabra 32 25 21 4 14 3 1 AF Actaea spicata 43 38 24 28 7 14 2 Veratrum album 23 5 3 21 2 7 - AF Aconitum vulparia 16 5 6 10 - - - Beech-fir forests (Lamio orvalae-Fagenion) 99 AF Abies alba 25 15 43 24 11 3 Rubus fruticosus 1 35 - 3 3 1 - AF Oxalis acetosella 49 82 10 39 14 35 5 AF Rhamnus alpinus subsp. fallax 11 44 5 19 3 4 - Senecio nemorensis agg. 14 45 8 11 6 16 1 Ajuga reptans 44 8 14 21 10 2 Sambucus racemosa 2 20 3 - 5 1 - Sanicula europaea 44 71 23 26 37 36 9 Hordelymus europaeus 10 28 - 1 3 15 - Athyrium filix-femina 51 64 14 23 31 25 8 Carex sylvatica 38 51 18 11 27 3 5 AF Cardamine trifolia 22 36 8 24 2 - - AF Lonicera nigra 3 21 - 14 1 - - Dryopteris dilatata 4 19 1 5 1 5 - Polystichum aculeatum 27 49 10 32 17 19 4 Viola reichenbachiana 34 65 26 33 41 46 15 Rubus idaeus 11 38 4 24 5 23 15 AF Festuca altissima 11 22 3 15 3 1 - Adenostyles alliariae 20 - 13 1 5 1 Solanum dulcamara 3 10 1 - 2 - - Euphorbia amygdaloides 39 68 37 55 32 48 18 Myosotis sylvatica 12 20 - 4 3 9 3 136 Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.X © 2012 International Association for Vegetation Science A. Marinšek et al. Table 1. (Continued). Differentiation ofFagus forests in SE Europe Thermophilous BF (Ostryo-Fagenion) AF Fraxinus ornus 11 1 6 10 19 7 5 Cyclamen purpurascens 42 21 1 42 6 1 Tanacetum corymbosum 3 - 6 1 4 1 Carex flacca - - 8 - 2 - AF Solidago virgaurea 9 5 2 28 12 1 2 Ostrya carpinifolia 4 2 7 10 5 7 4 AF Sorbus aria 8 8 5 31 3 1 2 Melittis melissophyllum 4 - 4 8 15 2 3 Primula vulgaris 7 - 5 11 10 7 3 Clematis vitalba 16 5 7 3 15 5 1 Convallaria majalis 9 - 0 11 3 6 1 Rosa arvensis 12 - 1 - 14 5 4 Campanula trachelium 7 4 8 3 9 7 2 Peucedanum oreoselinum - - 7 1 - - Laserpitium latifolium 1 1 9 4 1 - Vincetoxicum hirundinaria - - 9 4 2 1 Sesleria autumnalis 1 1 3 4 6 2 Tamus communis 13 - 3 4 15 4 1 Acer campestre 13 - 6 1 25 5 1 Berberis vulgaris 1 - 2 - - - Sorbus torminalis - - 1 - 8 4 1 Cornus sanguinea 9 _ 22 4 3 1 Asarum europaeum subsp. caucasicum 2 - 13 1 - - AF Aposeris foetida 27 2 0 25 6 2 Euphorbia dulcis 19 2 8 2 8 - Viburnum lantana 7 - 3 4 6 4 AF Carex digitata 18 9 0 28 8 7 1 Centaurea montana - - 0 2 - - Cornus mas 3 - 3 - 13 7 1 AF Buphthalmum salicifolium - - 5 9 - - Cruciata glabra 2 1 8 1 7 1 1 Crataegus monogyna 3 2 8 7 19 5 AF Helleborus niger 10 8 8 19 - 1 Anthericum ramosum - - 9 1 - - AF Erica herbacea - - 5 10 - - AF Anemone trifolia - 1 9 15 1 1 Ligustrum vulgare 2 - 4 - 5 1 Galium sylvaticum agg. 12 5 6 28 14 4 1 Crataegus laevigata - - 0 - 3 - Melampyrum sylvaticum 2 1 4 4 2 1 Salvia glutinosa 29 34 9 15 25 11 1 Campanula persicifolia - - 4 1 2 5 2 Peucedanum austriacum 2 - 3 3 2 1 AF Hacquetia epipactis 21 9 7 11 2 - AF Lonicera xylosteum 18 18 6 21 3 12 1 Rhamnus cathartica 1 - 8 - 1 - AF Galium laevigatum 2 - 7 14 1 - Brachypodium sylvaticum 18 9 6 6 25 12 4 Acer obtusatum 3 2 1 3 8 8 2 Staphylea pinnata 8 4 4 - 1 - Viburnum opulus 2 9 2 1 - Serratula tinctoria 2 8 - 1 1 Mercurialis ovata - 9 - 2 1 2 Altimontane and subalpine BF (Saxifrago rotundifoliae-Fagenion) 1 AF Asplenium viride 1 1 - 39 - 1 AF Rubus saxatilis - 2 3 33 - 1 AF Rosa pendulina 8 10 8 51 4 4 AF Clematis alpina - - 1 26 - - AF Lonicera alpigena 29 36 6 66 2 6 1 AF Valeriana tripteris 1 1 17 37 - 1 1 Sorbus mougeotii - - - 20 1 1 AF Sorbus aucuparia 12 36 5 57 2 13 5 Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 137 Differentiation of Fagus forests in SE Europe A. Marinšek et al. Table 1. (Continued!. AF Adenostyles alpina 5 4 1 27 - - - AF Phyteuma ovatum - - 3 18 - - - Vaccinium myrtillus - 26 4 45 3 11 7 Valeriana montana - - - 17 1 1 - AF Thalictrum aquilegiifolium 2 4 1 22 1 1 - Sesleria albicans - - - 13 - - - AF Polygonatum verticillatum 20 25 - 45 4 14 1 Polystichum lonchitis 6 9 1 32 2 8 6 AF Gentiana asclepiadea 21 18 33 52 11 5 1 Hypericum umbellatum - - - 11 - - - AF Laserpitium krapfii - - 6 18 1 - - Laburnum alpinum - - 5 17 1 1 - AF Homogyne sylvestris 4 11 23 33 1 - - AF Calamagrostis varia - 9 13 26 2 - - Gymnocarpium dryopteris 4 1 - 17 - 1 - Rhododendron hirsutum - - 1 12 - - - Saxiiraga cuneifolia - - - 10 - - - Luzula sylvatica 4 18 1 42 6 25 20 Carey, ferruginea - - - 9 - - - Ranunculus platanifolius 7 5 - 19 1 2 - Salix appendiculata 1 - - 10 - - - Saxiiraga rotundifolia 12 16 - 37 2 22 8 Homogyne alpina - - - 9 - 1 - Asplenium ruta-muraria 1 1 6 18 1 1 3 Aster bellidiastrum - - - 8 - - - Thymus serpyllum - - - 9 1 - - Aquilegia vulgaris 1 - 1 11 1 - - AF Aconitum lycoctonum subsp. vulparia 3 - 3 12 - - - Calamagrostis arundinacea 1 1 18 26 1 5 10 Carex brachystachys - - - 7 - - - Melica nutans 5 1 13 23 7 3 - AF Maianthemum bifolium 5 14 10 24 3 - - Larix decidua - - - 6 - 1 - Cicerbita alpina 3 5 - 13 1 3 - Gymnocarpium robertianum 1 4 3 11 1 - - Lowland BF (TV/to tomentosae-Fagenion sylvaticae) Carpinus betulus 19 1 15 4 54 3 6 Quercus petraea 18 - 19 - 33 1 5 Circaea lutetiana 24 9 - - 29 10 1 Festuca drymeja 9 16 3 7 37 19 16 Rubus hirtus agg. 23 30 18 12 54 42 16 7/7/0 cordata 1 1 12 1 16 - - Stellaria holostea 3 - - 3 14 5 - Galeopsis tetrahit 3 - - - 11 4 - Glechoma hirsuta 9 15 - 11 25 5 1 Viola odorata - - 1 - 9 4 - Montane BF of the central and south-eastern part [Doronico columnae-Fagenion moesiacae) Fm Lapsana communis - - - 1 2 15 1 Fm Moehringia trinervia 3 2 1 2 8 26 10 Melica uniflora 5 2 18 1 22 32 4 Poa chaixii - - - 1 - 8 - Pulmonaria rubra - - - 1 1 9 1 Helleborus cyclophyllus - - - - 1 8 - Fm Potentilla micrantha 4 1 6 1 7 28 24 Fm Geum urbanum 1 1 - 3 15 21 5 Epilobium montanum 8 35 - 18 15 40 20 Campanula sparsa - - - - 1 10 5 Luzula luzulina - 2 - 1 - 8 - Fm Lathyrus laxiflorus - - - - 7 18 17 Digitalis viridiflora - - - - - 5 - Veronica officinalis - 8 - 4 8 20 10 138 Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.X © 2012 International Association for Vegetation Science A. Marinsek et al. Table 1. (Continued). Differentiation ofFagus forests in SE Europe BF of the SE part, under the influence of a Mediterranean climate {Doronico orientalis-Fagenion sylvaticae) Fm Abies borisii-regis - - - - - 9 30 Orthilia secunda - 2 1 16 2 12 35 Fm Lathyrus alpestris - - - - - 6 18 Silene multicaulis - - - - - - 10 Monotropa hypopitys 1 - - - - 1 12 Viscum album s.lat. - 1 - 1 - - 9 Fm Luzula forsten 1 - 1 - 2 12 16 Doronicum Orientale - - - - - 6 10 Species diagnostic for more than one suballiance AF Cardamine enneaphyllos 64 22 19 52 5 5 - Galium odoratum 88 72 18 22 62 78 30 Hedera helix 43 1 55 6 33 10 4 AF Mercurialis perennis 59 30 63 54 23 14 4 AF Picea abies 32 81 22 56 8 8 3 Galium rotundifolium 2 45 - 2 6 15 39 AF Prenanthes purpurea 36 69 33 63 18 32 12 AF Veronica urticifolia 2 31 12 42 5 7 1 AF Daphne mezereum 54 45 68 85 10 15 - AF Cirsium erisithales 2 9 29 37 - - - AF Hepatica nobilis 9 - 37 35 5 3 - AF Carey, alba - 2 19 18 - - - Fm Poa nemoralis 2 4 3 18 16 47 55 Other species diagnostic for alliance Aremonio-Fagion Symphytum tuberosum agg. 43 35 21 39 14 31 1 Lilium martagon 33 14 26 28 7 12 4 Omphalodes verna 10 14 10 8 1 - - Phyteuma spicatum 14 11 18 18 2 1 - Other species diagnostic for alliance Fagion moesiacae Veronica chamaedrys 10 11 3 15 17 37 36 Galium aparine 1 1 - 1 10 11 5 Physospermum cornubiense - - - - 2 10 7 Other species with high frequncy Fagus sylvatica 100 100 100 100 100 100 100 Lamiastrum galeobdolon agg. 56 54 31 24 44 46 9 Mycelis muralis 41 72 24 53 48 71 51 Acer platanoides 35 14 33 4 28 15 4 Pulmonaria officinalis 34 16 40 11 24 15 1 Asarum europaeum 32 29 37 14 28 8 1 Geranium robertianum 30 41 5 21 27 44 7 Aremonia aghmonoides 29 60 18 42 19 57 30 Heracleum sphondylium 28 5 14 18 7 4 8 Prunus avium 25 2 26 3 30 10 1 Fragaha vesca 25 46 27 38 32 29 13 Lathyrus vernus 25 16 29 16 25 4 2 Euonymus latifolius 22 8 12 9 4 8 1 Corylus avellana 22 18 27 9 29 11 3 Aegopodium podagraha 21 4 8 9 8 14 - Urtica dioica 18 2 - 3 6 13 5 Doronicum austriacum 16 8 1 8 2 8 - Ruscus hypoglossum 16 10 12 3 21 2 3 Luzula luzuloides 15 10 10 11 22 24 23 Milium effusum 15 10 - 2 3 8 1 Fraxinus excelsior 13 6 13 4 14 4 2 Daphne laureola 11 16 3 1 3 5 1 Aruncus dioicus 11 6 8 10 6 1 - Scilla bifolia 11 1 1 1 - 7 1 AF = alliance Aremonio-Fagion] Fm = alliance Fagion moesiacae] 1 = montane beech forests of Lamio orvalae-Fagenion; 2 = beech- -fir forests of Lamio orvalae-Fagenion] 3 = Ostryo-Fagenion; 4 = Saxifrago rotundifoliae-Fagenion] 5 = 7/7/0 tomentosae -Fagenion sylvaticae; 6 = Doronico columnae-Fage-nion moesiacae] 1 = Doronico orientalis-Fagenion sylvaticae. Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 139 Differentiation of Fagus forests in SE Europe A. Marinsek et al. 0.0 0-5 DCA1 Fig. 4. Detrended correspondence analysis (DCA) of seven releve clusters, with passively projected explanatory variables. The numbering of the clusters is the same as in Table 1. according to altitudinal gradient. We therefore consider Axis 1 to be the generally accepted macroecological gradient in the Balkans, which runs along the Dinaric Alps in a NW-SE direction. As expected, the correlation between Axis 1 and altitude is not significant (Table 2), because a wide altitudinal range is characteristic of the study area (Fig. 5) and BF are found along the whole altitudinal gradient in northern and southern regions. The distribution and range of altitudes among clusters is shown in Fig. 5. The correlation between DCA releve scores of Axis 1 and 2 and the mean ETV, tested with Kendall's coefficient, showed some significant differentiation along both axes (Table 2). ETV for light and temperature show a significant increase along the geographical gradient toward the SE (Axis 1), while indicators for moisture, reaction and nutrients show a significant decrease. All correlations between Axis 2 and ETV are significant. Altitude and ETV for moisture and nutrients significantly decrease, while indicator values for light, temperature, continentality and reaction significantly increase (Table 2). Along Axis 1, there is a significant increase in the proportion of chamaephytes, hemicryptophytes and therophytes toward the SE. At the same time, the proportion of geophytes and phanerophytes significantly decreases in the same direction (Table 2) and reflects changed ecological conditions. We also found correlations between Axis 1 and the proportion of chorotypes differed significantly different. Stenomediterranean, Eurymediterranean, Mediterranean-Montane, Eurasian and Montane S European species show MOO 2200 2000 1600 1600 E 1400 1100 Z < iOOO BOO 600 400 20D 0 I 12«*»-75% ~~T Neii-QnElwMaiKe I I. ITT 1 2 3 4 S B T CllHMf Fig. 5. Box-whiskers graph presents altitudes and altitudinal range for each cluster individually. an increasing trend toward the SE, while Boreal species decrease in the same direction. These results confirm our assumptions of geographically or macroecologically based differentiation of BF vegetation. The highest number of species can be observed in the BF of the N group (Fig. 6), especially in thermophilous BF (Cluster 3 - Ostryo-Fagenion) and in altimontane and subalpineBF (Cluster 4 - Saxifrago-Fagenion). Species numbers in clusters of the SE group (5, 6, 7) are in general lower. The lowest number of species and BF species is observed in BF of the SE part under the influence of a mediterranean climate (Cluster 7 - Doronico orientalis-Fage-nion sylvaticae). In general, species richness decreases along the geographical gradient. Comparing Clusters 4, 6 and 7 in Figs 5 and 6, it is obvious that altitude does not affect species richness. Both types of BF thrive at the high altitudes, but Cluster 4 contains the highest number of species and Cluster 7 the lowest. Since the analysis shows two main groups of BF in the investigated area (Fig. 3), we attempted to determine significant differences between them. We therefore compared the vegetation characteristics of the northern group (alliance Aremonio-Fagion) and the southern group (alliance Fagion moesiacae). The results of the Mann-Whitney U-test in Table 3 show statistically significant differences in the proportions of all life forms between the groups, except for hemicryptophytes and phanerophytes. Table 2 also indicates a higher proportion of geophytes and phanerophytes in the northern group, while other life forms have a higher proportion in forests of the alliance Fagion moesiacae. All ETV are significantly different between the two groups (Table 3). The table shows that BF of the alliance Fagion moesiacae thrive in conditions with higher temperatures, lower rainfall and are lower nutrients than stands of the alliance Aremonio-Fagion. 140 Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.X © 2012 International Association for Vegetation Science A. Marinsek et al. Differentiation ofFagus forests in SE Europe cd cm cm °3 SO 70 50 ■i » 40 i0 § £ 20 10 0 -10 räl ALL SPECIES ffl BF SPECIE* eiub 1 2 3 4 5 6 7 Cluster Fig. 6. Box-whiskers graph presents the average number of species and average number of BF species defined in individual clusters. Beech forest species were defined according to Willneret al. (2009). i- ü) - S Ln . .. S m >- CD o In terms of the proportion of chorotypes in the species composition of the two groups, there are significantly increased numbers of endemic, Stenomediterranean, Eurymediterranean, Eurasian, Atlantic, Montane S European and widespread species in the BF of alliance Fagion moesiacae. All of those indices highlight not only the different ecological factors in each of the two groups but also a different structure of species traits composition. Oj CM p LO m SO LO Discussion The first level of division in the dendrogram (Fig. 3) revealed the main geographic division and reflects the geographic differentiation of BF in SE Europe. The northern group (including Clusters 1-4) was classified into the alliance Aremonio-Fagion and the second, the southern group (including Clusters 5-7), into the alliance Fagion moesiacae. In the nomenclature revision of Illyrian BF of Slovenia, Croatia, SW Hungary, S Austria and NE Italy, Marincek et al. (1992) differentiated four suballiances within the Aremonio-Fagion alliance. Our analysis revealed three of them in our study: Cluster 1 as mesophilous Lamio orvalae-Fagenion, Cluster 2, which incorporates mostly fir-beech forests, according to Marincek et al. (1992) is also classified into the Lamio orvalae-Fagenion suballiance, Cluster 3 as thermophilous Ostryo-Fagenion, while Cluster 4 was altimontane Saxifrago rotundifoliae-Fagenion. Cluster 4 includes not only altimontane to subalpine BF but also lower montane ones (e.g. Arunco-Fagetum). This is confirmation of Willner (2002), who pointed out the close floris-tic relationship between Arunco-Fagetum and Anemono trifoliae-Fagetum. Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 141 Differentiation of Fagus forests in SE Europe Table 3. Mann-Whitney U-test of differences in life forms, indicator values and chorotypes between the northern group (AF: alliance Aremonio-Fagion) and southern group (Fm: alliance Fagion moesiacae) of BF vegetation types. Variable z P-level Valid N AF Valid N Fm %C -7.9606 0.0000* 324 526 %G 5.8837 0.0000* 368 611 %H -1.7326 0.0832 367 622 %N -3.8153 0.0001* 318 385 %P 1.6052 0.1084 369 628 %T -10.7364 0.0000* 132 311 Altitude -1.3653 0.1722 369 628 Light -4.3717 0.0000* 369 628 Temperature -11.8139 0.0000* 369 628 Continentality 4.1931 0.0000* 369 628 Moisture 12.8679 0.0000* 369 628 Soil reaction 14.4035 0.0000* 369 628 Nutrients 4.4918 0.0000* 369 628 End. -4.6764 0.0000* 37 28 StM -10.0349 0.0000* 170 264 EuM -8.4353 0.0000* 180 335 MeM -2.7316 0.0063 321 250 EuA -5.4226 0.0000* 369 628 Atl -5.211 0.0000* 84 80 MSE 3.8103 0.0001* 345 409 Bor 2.65 0.008 368 597 Wis -8.8218 0.0000* 293 465 End.= Endemic; StM = Stenomediterranean; EuM = Eurymediterranean; MeM = Mediterranean-Montane; EuA = Eurasian; Atl. = Atlantic; MSE = Montane south European; Bor = Boreal; WiS = widespread species. 'Significant atP < 0.001. The suballiance Epimedio-Fagenion, which is not clearly separated, is found in Cluster 3. The similarity between Os-tryo-Fagenion and Epimedio-Fagenion was already emphasized by Willner (2002), but these two suballiances are ecologically different, since Ostryo-Fagenion appears on steep southern slopes over shallow rendzinas and Epimedio-Fagenion in lowlands on deeper soils. They have many common thermophilous species. Their exact syntaxonomi-cal position therefore needs further research. The southern group of BF includes Clusters 5-7. BF of Cluster 5 thrive in the marginal regions of BF in the investigated area (Fig. 1) but over a wide geographical range. These marginal regions extend from Austria to the Black Sea region, and it is usually difficult to classify such communities on the edge of an area of distribution. In sub-optimal conditions, species of BF become weak competitors and gradually disappear. These are common features of such lowland BF types (Cluster 5). Possible reasons for a different species composition in Cluster 5 in relation to other beech forest vegetation types are also: • sub-optimal growing conditions (lower precipitation and higher temperature), A. Marinsek et al. • growing in solitary montane islands of the sub-Panno-nian region, mostly surrounded by Quercus and Carpinus forest types, • the longest distance from the nearest glacial refuge area (Magri et al. 2006; Willner et al. 2009) and possible delayed post-glacial dispersal of European BF understo-rey species. In the opinion of Magri et al. (2006), the post-glacial expansion of Fagus appears to have been limited by large plains with a continental climate and by important river valleys, such as the Hungarian plain and the lower Danube valley, while Willner et al. (2009) observed the highest species richness in areas close to potential glacial refuge areas, • all of the above. A difficulty that occurs in the classification of vegetation in Cluster 5 is that many species appear that are also characteristic of Quercus and Carpinus forests (Marincek & Cami 2000) and, in the SE part, species of Quercion frainetto (Cami et al. 2009; Kavgaci et al. 2010; Lyubenova et al. 2011). There are actually three possibilities for solving this situation: 1. Split Cluster 5 arbitrarily into a northern and southern part and include each part into the alliances of those areas; northern lowland BF (including those from Bosnia and Herzegovina, Croatia and SE Slovenia) to alliance Aremonio-Fagion, and those from Serbia and Bulgaria lowland to alliance Fagion moesiacae. This division is geographically founded. It results from a lack of beech forest species and presence of lowland and thermophilous species that unify all BF from E Slovenia to the Black Sea region into one group. 2. Include them in the species-poor C European Fagion syl-vaticae alliance, 3. Describe a new (sub)alliance of the marginal communities, extending from the Black Sea region to outcrops of the S Alps. An aspiration for a new suballiance, which would comprise BF and beech-fir forests from the Pannonian region in Croatia, already appeared when Vukelic & Baricevic (2007), while investigating beech-fir forests, pointed out the need for a new suballiance into which these forests would fit. They merely proposed a potential name, Festuco drymeiae-Fagenion sylvaticae, which could be classified into the alliance Aremonio-Fagion. In terms of our investigations, BF of the alliance Fagion moesiacae unexpectedly also extend over the whole SW edge of the Pannonian plain (Fig. 1). This finding is in agreement with Rivas-Martinez et al. (2011), who designated the border between the oceanic and continental bio-climate in the Balkan Peninsula. This confirms that optimal ecological conditions for BF are linked to an oceanic bioclimate. In any case, the alliance comprises three clusters, which represent the following groups of forests: %c -7.9606 0.0000* 324 %G 5.8837 0.0000* 368 % H -1.7326 0.0832 367 % N -3.8153 0.0001* 318 % P 1.6052 0.1084 369 %T -10.7364 0.0000* 132 Altitude -1.3653 0.1722 369 Light -4.3717 0.0000* 369 Temperature -11.8139 0.0000* 369 Continentality 4.1931 0.0000* 369 Moisture 12.8679 0.0000* 369 Soil reaction 14.4035 0.0000* 369 Nutrients 4.4918 0.0000* 369 End. -4.6764 0.0000* 37 StM -10.0349 0.0000* 170 EuM -8.4353 0.0000* 180 MeM -2.7316 0.0063 321 EuA -5.4226 0.0000* 369 Atl -5.211 0.0000* 84 MSE 3.8103 0.0001* 345 Bor 2.65 0.008 368 Wis -8.8218 0.0000* 293 142 Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.X © 2012 International Association for Vegetation Science A. Marinsek et al. 1 Lowland BF (Cluster 5), comprising beech forest communities from the lowest altitudes, mainly occurring on the sub-Pannonian plain from SE Slovenia to the N and E Bulgaria. In relation to the aforementioned three possible solutions for these forests, we suggest a new subal-liance Tilio tomentosae-Fagenion sylvaticae (typified in Appendix SI) for such lowland forests. 2 Montane BF of the continental part of the C Balkans (Cluster 6), which have been traditionally treated as the suballiance Doronico columnae-Fagenion moesiacae. Such BF, mainly from SE Serbia, the Republic of Macedonia and N Greece and high altitudes southward to E-C Greece, were already classified in the same suballiance (Dzwonko & Loster 2000; Bergmeier & Dimopoulos 2001). 3 Montane BF influenced by a mediterranean climate (Cluster 7), included in the suballiance Doronico oriental-is-Fagenion sylvaticae. BF of N and C Greece were already classified by Dzwonko & Loster (2000) and Bergmeier & Dimopoulos (2001) into the suballiance Doronico orien-talis-Fagenion moesiacae (invalid name - see Appendix 1). The syntaxonomic classification of these three southern suballiances on the level of higher rank syntaxa has not yet been unified. The separation of suballiances Doronico columnae-Fagenion moesiacae and Doronico oriental-is-Fagenion sylvaticae was based on the altitudinally and the strongly geographically distinct groups of diagnostic species. Dzwonko & Loster (2000) placed them as separate suballiances of Aremonio-Fagion, while Bergmeier & Dimopoulos (2001) and Tzonev et al. (2006) classified them into Fagion sylvaticae. The reason is probably a lack of comparable data from the intermediate area between the SE part of the researched area and C European Fagion sylvaticae forests. These forests were classified to different alliances, as already proposed in the past but not validly described, according to ICPN, by Quezel (1967), Dafis (1973) and Horvat et al. (1974). Because we are taking a larger area into consideration, we suggest that both of them, together with suballiance Tilio tomentosae-Fagenion sylvaticae, should be classified into the alliance Fagion moesiacae, although Cluster 7 is rather isolated from the rest of the data set and indicates a somewhat unique floristic composition. According to our results, the number of species significantly decreases towards the southeast and BF of that area are relatively species-poor, especially those belonging to Doronico orientalis-Fagenion sylvaticae, mainly located in NE Greece, on the edge of the alliance area. The area's centre lies in SW Serbia and in W and C Bulgaria (Fig. 1). It is therefore hard to find appropriate diagnostic and differential species for the alliance. At the same time, the question is raised of the low number of species, including BF species (Willner et al. Differentiation of Fergus forests in SE Europe 2009), in that region. It is clear from Fig. 6, which shows the average number of species in different clusters, that the highest number can be found in the group of clusters belonging to Aremonio-Fagion and the lowest number of species in both cases can be found in Clusters 5 and 7. There are various reasons for the lower number of species in the SE part of the investigated area. One possible reason is that BF form the timber line in the SE (Dzwonko & Loster 2000; Bergmeier & Dimopoulos 2001), in contrast to the NW part of the researched area (SE Alps, Dinarids), where the timber line consists almost entirely of coniferous forests (Pinus mugo, Larix decidua, Picea abies). There are therefore no other forest communities above them in the SE and the rare species exchange is restricted to grassland communities above and to forests in contact with lower belts. The species pool is consequently limited. Another reason could be the acidophilous character of the substrate in the southern part (Tsiripidis et al. 2007). Ewald (2003b) claims that the pool of C European flora consists of a majority of vascular plant taxa that are restricted to very base-rich and calcareous soils and offers the hypothesis that Pleistocene range contractions caused the extinction of more acidophilous than calciphilous species, because acid soils were much rarer when refugial areas were at their minimum. Willner et al. (2009) also found that acidic BF contributed little to the number of beech forest species and that soil type diversity is a weak predictor of the number of beech forest species. They found distance to the nearest potential refuge area to be the strongest predictor of beech forest species richness. On the basis of our analysis, we found forests of Cluster 6 had the highest species richness in the SE part of the region (Fig. 6). On the basis of their and our own findings, we can corroborate one of the potential refuge areas in SW Bulgaria, proposed by Mag-ri et al. (2006). Syntaxonomy should reflect the main factors that cause the differentiation of vegetation types. Regionally defined units reflect macroclimatic patterns, while locally defined vegetation units usually reflect patterns connected with local ecological factors, such as soil properties or altitude (Knollová & Chytrý 2004). We found the influence of different climatic types in the Balkan Peninsula were very important for distinguishing these two forest types (alliances). Our study also confirmed ecological factors as the basis for the second level of division (Fig. 3), which reflect regionally defined units. We found that soil reaction was an insignificant factor for differentiation, which is in agreement with Di Pietro (2009), who claimed that the role of soil pH, which serves to distinguish basiphilous BF (Fagion s. 1., Fagetalia) and acidophillous BF (Luzulo-Fagion, Querce- Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 143 Differentiation of Fagus forests in SE Europe A. Marinšek et al. talia robori-petraeae), appears not to be applicable to the Apennines. The question of the prevailing impact of ecological as opposed to macroecological and phytogeographical gradients on vegetation, and vice versa, arises in many studies (Willner 2002; Tzonev et al. 2006; Tsiripidis et al. 2007). Unfortunately, these studies were done on the northern and southern edges of the distribution area and did not comprise the whole diversity of Balkan BF. A quick glance at the vegetation units distinguished in the past, at separate local levels, shows the delimitation of vegetation on the basis of various criteria. Marincek et al. (1992) classified BF based on ecological conditions within the framework of the geographically distinguished Aremonio-Fagion alliance. Tzonev et al. (2006) established that Bulgarian F. sylvatica communities do not show any distinct pattern of geographical differentiation but follow edaphic and local topo-climatic gradients. The reason for such results is the relatively small investigation area. Strong phytogeographical differentiation along a N-S gradient is suggested for BF in S Serbia, Macedonia and N and C Greece (Dzwonko et al. 1999; Dzwonko & Loster 2000). This is in accordance with the results of our study, although Dzwonko & Loster (2000) neglected refugia in the S Balkans and classified all BF communities (except acidophilous BF) into the alliance Aremonio-Fagion. Together with a significantly increased proportion of life forms (chamaephytes, therophytes) and chorotypes (above all Stenomediterranean, Eurymediterranean, Mediterranean-Montane species) toward the SE, the results show that beech forest communities are differentiated on the basis of (phyto)geographical and macroclimatic conditions, while ecological factors have a stronger influence on vegetation on a smaller scale. There is not only narrower altitudinal distribution of BF in the SE part of the research area, but also structural, functional and geo-ele-mental changes of BF as a result of changed macroclimatic factors and their development in the post-glacial period. Conclusions Our investigation confirmed the division of BF into two alliances in the region (Rodwell et al. 2002). Classification to lower units and the status of some forests, especially lowland BF on the edge of the area, extending from the Pannonian lowland (mainly Croatia and Serbia) to NE Bulgaria, is not very clear and needs further investigation. The findings of Tsiripidis et al. (2007) that the S Balkan Peninsula is a place in which views and hypotheses on the role of ecological and geographical factors in the differentiation of beech forest vegetation are partly complementary and partly contradictory, can be generalized for the whole of SE Europe. Acknowledgements We would like to express our thanks to the many people who kindly helped us to compose the database and for their comments, in particular Lojze Marinček (Ljubljana), Željko Škvore (Zagreb), Sulejman Redžič (Sarajevo), Jugoslav Brujič (Banja Luka), Marijana Novakovič and Marko Pero-vič (Belgrade), Rosen Tzonev (Sofia) and Vlado Matevski (Skopje). Thanks are also due to Jean-Paul Theurillat (Champex-Lac) for advice concerning syntaxonomy. Wolfgang Willner and an anonymous referee contributed with many constructive comments on the manuscript. Martin Cregeen improved the English. The authors acknowledge financial support from the state budget through the Slovenian Research Agency (Projects No. LI-9737 and PI-0236). References Bergmeier, E. 1990. Wälder und Gebüsche des Niederen Olymp (Káto Olimbos, NO-Thessalien). Ein Beitrag zur systematischen und orographischen Vegetationsgliederung Griechenlands. Phytocoenoloßia 18: 161-342. Bergmeier, E. & Dimopoulos, P. 2001. Fagus sylvatica forest vegetation in Greece: syntaxonomy and gradient analysis. Journal of Vegetation Science 12: 109-126. Bergmeier, E. & Dimopoulos, P. 2008. Identifying plant communities of thermophilous deciduous forest in Greece: species composition, distribution, ecology and syntaxonomy. Plant Biosystems 142: 228-254. Blečič, V. & Lakušič, R. 1970. Der Urwald Biogradska Gora in Gebirge Bjelasica in Montenegro. Akad. Nauka Umjet. Bosne Her-ceg, Poseb. Izd. 15: 131-139 + 4tab. Bohn, U., Neuhäusl, R., Gollub, G., Hettwer, C, Neuhäuslova, Z., Raus, T.H., Schlüter, H. & Weber, H. 2004. Karte der natürlichen vegetation Europas/map of the natural vegetation of Europe. Maßstab/Scale 1: 2.500.000. Münster, DE. Braun-Blanquet, J. 1964. Pflanzensoziologie. Grundzüge der Vegetationskunde, 3rdedn. Springer, Berlin, DE. Bruelheide, H. 2000. A new measure of fidelity and its application to defining species groups. Journal of Vegetation Science 11: 167-178. Brus, R. 2010. Growing evidence for the existence of glacial refugia of European beech (Fagus sylvatica L.) in the southeastern Alps and north-western Dinaric Alps. Periodicum Biologomni 112: 239-246. Carni, A., Košir, P., Karadzic, B., Matevski, V., Redžič, S. & Škvore, Ž. 2009. Thermophilous deciduous forests in Southeastern Europe. Plant Biosystems 143: 1-13. Chytrý, M., Tichý, L., Holt, J. & Botta-Dukát, Z. 2002. Determination of diagnostic species with statistical fidelity measures. Journal of Vegetation Science 13: 79-90. Curovič, M., Steševič, D., Medarevič, M., Cvjetičanin, R., Pantič, D. & Spalevič, V. 2011. Ecological and structural characteris- 144 Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.X © 2012 International Association for Vegetation Science A. Marinsek et al. tics of monodominant montane beech forests in the national park Biogradska Gora, Montenegro. Archives of Biological Sciences 63: 429^40. Cvjeticanin, R. 2003. Fitocenoze bukve u Srbiji. Sumarstvo 55: 107-112 (In Serbian). Cvjeticanin, R. & Novakovic, M. 2004. Fitocenoloska pripadnost bukovih suma u istrazivanim sastojinama na Ozrenu-Sokobanja. Sumarstvo 65: 97-104. Czeczott, H. 1933. A study on the variability of the leaves of beeches: F. orientalis Lipsky, F. sylvatica L. and intermediate forms. Parti.. RocznikDendrologiczny 5: 45-121. Dans, S. 1973. Taxonömisis tis dasikis vlastiseos tis Ellädos. (Gliederung der Waldwegetation Griechenlands.). Epist. Epet. Geopon. Dasol. Schol. Arist. Paneptist. Thessalonikis 15: 75-88. Denk, T. 1999. The taxonomy of Fagus in western Eurasia, 1: Fagus sylvatica subsp. orientalis (= F. orientalis). 2: Fagus sylvatica subsp. sylvatica. Feddes Repertorium 110: 177-200, 381-412. Di Pietro, R. 2009. Observations on the beech woodlands of the Apennines (peninsular Italy): an intricate biogeographical and syntaxonomical issue. Lazaroa 30: 89-97. Dierschke, H. 1990. Species-rich beech woods in mesic habitats in central and Western Europe: a regional classification into suballiances. Vegetatio 87: 1-10. Dierschke, H. & Bohn, U. 2004. Eutraphente Rotbuchenwälder in Europe. Tuexenia 24: 19-56. Dzwonko, Z. & Loster, S. 2000. Syntaxonomy and phytogeo-graphical differentiation of the Fagus woods in the Soutwest Balkan Peninsula. Journal of Vegetation Science 11: 667-678. Dzwonko, Z., Loster, S., Dubiel, E. & Drenkovski, R. 1999. Syn-taxonomic analysis of beechwoods in Macedonia (former Republic of Yugoslavia). Phytocoenologia 29: 153-175. Ellenberg, H. 1996. Vegetation Mitteleuropas mit den Alpen. - 5. Aufl. Ulmer, Stuttgart, DE. Ewald, J. 2003a. The sensivity of Ellenberg indicator values to the completeness of vegetation releves. Basic and Applied Ecology 4: 507-513. Ewald, J. 2003b. The calcareous riddle: why are there so many calciphilous species in the Central European flora? Folia Geo-botanica 38: 357-366. Fukarek, P. 1969. Prilog poznavanju biljnosocioloskih odnosa suma i sibljaka nacionalnog parka Sutjeska. ANU BiH, special editionXI. Odjeljenjeprirodnih imatematickihnauka 3: 189-291. Gömöry, D. & Paule, L. 2010. Reticulate evolution patterns in western-Eurasian beeches. Botanica Helvetica 120: 63-74. Gömöry, D., Paule, L., Brus, R., Zhelev, P., Tomovic, Z. & Gracan, J. 1999. Genetic differentiation and phylogeny of beech on the Balkan peninsula. Journal of Evolutionary Biology 12: 746-754. Hennekens, S.M. & Schaminee, J.H.J. 2001. TURBOVEG, a comprehensive database management system for vegetation data. Journal of Vegetation Science 12: 589-591. Horvat, I. 1938. Biljnosocioloska istrazivanja suma u Hrvatskoj. Glasnik za sumske pokuse 6: 127-279. Horvat, I. 1950. Sumske zajednice Jugoslavije (Les associations forestieres en Yugoslavie). Institut za sumarska istrazivanja: p. 73. Differentiation of Fagus forests in SE Europe Horvat, I., Glavac, V. & Ellenberg, H. 1974. Vegetation Südosteuropas. Geobotanica selecta 4, G. Fischer, Stuttgart, DE. p. 767. Jordanov, D. (ed.). 1963-1979. Flora Republicae Popularis Bulgaricae. In Aedibus Academiae Scientiarum Bulgaricae, Serdicae (InBulgarian). Josifovic, M., (ed.). 1970-1977. Flora SR Srbije. I-IX. Srpska Akademija Nauka i Umetnosti, Beograd (In Serbian). Jovanovic, B., Lakusic, R., Rizovski, R., Trinajstic, I. & Zupancic, M. 1986. Prodromus phytocoenosum Jugoslaviae ad mappam vegetationis M 1: 200 000. Naucno vece vegetacijske karte Jugoslavije, Bribir-Ilok. p. 46. Kavgaci, A., Carni, A., Tecimen, H.B. & Ozalp, G. 2010. Diversity and ecological differentiation of oak forests in NW Thrace (Turkey). Archives of Biological Sciences 62: 705-718. Knollovä, I. & Chytry, M. 2004. Oak-hornbeam forests of the Czech Republic: geographical and ecological approaches to vegetation classification. Preslia 76: 291-311. Kosir, P., Carni, A. & Di Pietro, R. 2008. Classification and phyto-geographical differentiation of broad-leaved ravine forests in southeastern Europe. Journal of Vegetation Science 19:331-342. Lyubenova, M., Tzonev, R. & Pachedjieva, K. 2011. Syntaxonomy of Quercetea pubescentis (Oberd. 1948) Doing Kraft, 1955 in Bulgaria. Comptes rendus de l'Academie bulgare des Sciences 64: 565-580. Magri, D., Vendramin, G.G., Comps, B., Dupanloup, I., Geburek, T., Gomory, D., Latalowa, M., Litt, T., Paule, L., Roure, J.M., Tantau, I., van der Knaap, W.O., Petit, R.J. & de Beaulieu, J. L. 2006. A new scenario for the Quaternary history of European beech populations: palaeobotanical evidence and genetic consequences.NewP/jyto/o^/jt 171: 199-221. Marincek, L. & Carni, A. 2000. Die Unterverbände der Hainbuchenwälder des Verbandes Erythronio-Carpinion betuli (Horvat 1938) Marincek in Wallnöfer, Mucina et Grass 1993. Scopolia 45: 1-20. Marincek, L., Mucina, L., Zupancic, M., Poldini, L., Dakskobler, I. & Accetto, M. ("1992" [recte 1993]). Nomenklatorische Revision der illyrischen Buchenwälder. Studia Geobotanica 12: 121-135. Pignatti, S., Menegoni, P. & Pietrosanti, S. 2005. Bioindicazione attraverso le piante vascolari. Valori di indicazione secondo Ellenberg (Ziegerwerte) per le specie della Flora dJtalia. Braun-Blanquetia 39: 1-97. Quezel, P. 1967. A propos de quelques hetraies de Macedoine grecque. Bulletin de la Societe Botanique de France 114: 200-210. Raunkiaer, C. 1934. The life forms of plants and Statistical geography. Claredon, Oxford, UK. Redzic, S. 2007. Syntaxonomic diversity as an indicator of ecology diversity - case study Vranica Mts in the Central Bosna. Biologia 62: 173-184. Redzic, S. & Barudanovic, S. 2010. The pattern of diversity of forest vegetation of the Crvanj Mountain in the Herzegovina (West Balkan Peninsula). Sumarski list 5-6: 261-274. Rexhepi, F. 2007. The vegetation of Kosovo. University of Prishtina, Faculty of Natural Sciences, Prishtine, p. 137. Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 145 Differentiation of Fagus forests in SE Europe Rivas-Martínez, S., Rivas Sáenz, S. & Penas, A. 2011. Worldwide bioclimatic classification system. Global Botany 1: 1-634 + 4 maps. Rodwell, J.S., Schaminée, J.H.J., 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-LNV 2002/054: p. 168. Šilc, U. & Čami, A. 2012. Conspectus of vegetation syntaxa in Slovenia. Hacquetia 11 (inpress). DOI: 10.2478/vl0028-012-0006-1. Sile, U., Vrbničanin, S., Božič, D., Cami, A. & Dajič Stevanovič, Z. 2009. Weed vegetation in the north-western Balkans: diversity and species composition. Weed Research 49: 602-612. Soó, R. 1963. Bulgarische Pflanzengesellschaften II. Annales Universitatis Scientarium Budapestinensis. Section Biology 6: 175-186. Soó, R. 1964. Die regionalen Fagton-Vabände und Gesellschaften Südosteuropas. Ada Agronomica Academiae Scientiarum Hungaricae 1: 1-104. Stefanovič, V. 1986. Fitocenologija sa pregledom šumskih fitocenoza Jugoslavije, IGRO Svjetlost, OOUR Zavoda za udžbenike Sarajevo, p. 269 (In Serbian). Tichý, L. 2002. JUICE, software for vegetation classification. Journal of Vegetation Sdence 13: 451^53. Tichý, L. & Chytrý, M. 2006. Statistical determination of diagnostic species for site groups of unequal size. Journal of Vegetation Science 17: 809-818. Török, K., Podani, J. & Borhidi, A. 1989. Numerical revision of the Fagion illyricum alliance. Vegetatio 81:169-180. Trinajstič, I. 2008. Biljne zajednice Republike Hrvatske. Plant communities of Croatia. Akademija šumarskih znanosti, Zagreb, p. 179 (In Croatian). A. Marinsek et al. Tsiripidis, I., Bergmeier, E. & Dimopoulos, P. 2007. Geographical and ecological differentiation in Greek Fagus forest vegetation. Journal of Vegetation Science 18: 743-750. Tutin, T.G., Heywood, V.H., Burges, N.A., Moore, D.M., Valentine, D.H., Walters, S.M. & Webb, D.A. (eds.) 1964-1993. Flora Europaea, Vols 2-5 and Vol. 1, 2nd edn. Cambridge University Press, Cambridge, UK. Tzonev, R., Dimitrov, M., Chytry, M., Roussakova, V., Dimova, D., Gussev, C, Pavlov, D., Vulchev, V., Vitkova, A., Gogoushev, G., Nikolov, I., Borisova, D. & Ganeva, A. 2006. Beech forest communities in Bulgaria. Phytocoenologta 36: 247-279. Vukelic, J. & Baricevic, D. 2002. Novije fitocenoloske spoznaje o bukovim sumama u Hrvatskoj. Recent phytocoenological perceptions of beech forests in Croatia. Sumarskl list 9-10: 439^57. Vukelic, J. & Baricevic, D. 2007. Nomenklaturno-sintaksonom-sko odreenje panonskih bukovo-jelovih suma (Abteti-Fage-tum "pannonicum") u Hrvatskoj. Nomenclatural-syntaxonomic determination of pannonian beech-fir forests (Abieti-Fagetum "pannonicum") in Croatia. Sumarskl list 9-10: 407-429 (In Croatian). Weber, H.E., Moravec, J. & Theurillat, J.-P. 2000. International code of phytosociological nomenclature. 3rd edition. Journal of Vegetation Science 11: 739-768. Willner, W. 2002. Syntaxonomische Revision der sudmitteleu-ropaischen Buchenwalder. Phytocoenologta 32: 337-453. Willner, W., Di Pietro, R. & Bergmeier, E. 2009. Phytogeographi-cal evidence for post-glacial dispersal limitation of European beech forest species. Ecography 32: 1011-1018. Appendix 1: Nomenclature of Fagus sylvatica forest syntaxa in Southeast Europe Querco-Fagetea Braun-Blanquet et Vlieger in Vlieger 1937 Fagetalia sylvaticae Walas 1933 Aremonio-Fagion Török, Podani et Borhidi ex Marincek, Mucina, Zupanäc, Poldini, Dakskobler et Accetto 1993 Lamio orvalae-Fagenion Borhidi ex Marincek, Mucina, Zupanäc, Poldini, Dakskobler et Accetto 1993 Saxifrago-Fagenion Marincek, Mucina, Zupanäc, Poldini, Dakskobler et Accetto 1993 Ostryo-Fagenion Borhidi ex Soo 1964 (incl. Epimedio-Fagenion Marincek et al. 1993) Fagion moesiacae Bleäc et Lakusic 1970 (Syn.: Fagion moesiacum Fukarek 1969 nom. inval. [art. 2b], Fagion moesiacum Horvat, Glavac et Ellenberg 1974 nom. illeg. [art. 34], Fagion moesiacae Török, Podani et Borhidi 1989 nom. illeg. [art 31], Doronico orientalis-Fagion moesiacae Dierschke 1998 nom. inval. [art. 2b], Doronico columnae-Fagion moesiacae (Dzwonko, Loster, Dubiel et Drenkovski 1999) Dierschke 2004 [syntax, syn.]) Doronico orientalis-Fagenion sylvaticae Marinsek, Carni et Silc suball. nov. (Syn.: Doronico orientalis-Fagenion moesiacae Raus 1977 nom. inval. [art. 1], Doronico orientalis-Fagenion moesiacae Raus 1980 nom. inval. [art. 2b], Doronico orientalis-Fagenion moesiacae Raus ex Raus 1980 nom. inval. [art. 2b], Doronico orientalis-Fagenion moesiacae Raus ex Bergmeier 1990 nom. inval. [art. 5], Doronico orientalis-Fagenion moesiacae Raus ex Bergmeier et Dimopoulos 2001 nom. inval. [art. 3f]) Applied Vegetation Science 146 Doi: 10.1111/J.1654-109X.2012.01203.X© 2012 International Association for Vegetation Science A. Marinšek et al. Differentiation ofFagus forests in SE Europe Doronico columnae-Fagenion moesiacae Dzwonko, Loster, Dubiel et Drenkovski 1999 Tilio tomentosae-Fagenion sylvaticae Marinšek, Čarni et Šilc suball. nov. Newsyntaxa and lectotypifications Fagion moesiacae Blečič et Lakušič 1970: Lectotypus hoc loco: Elymo-Fagetum moesiacae Blečič et Lakušič 1970 (Nomenclatural type, lectotypus: relevé no. 2, table 4 in Blečič et Lakušič (1970)) Doronico orientalis-Fagenion sylvaticae Raus ex Marinšek, Čarni et Šilc suball. nov. hoc loco: Nomenclature type -holotypus: Lathyro alpestris-Fagetum sylvaticae Bergmeier 1990 Diagnostic species: Abies borisii-regis, Lathyrus alpestris, Luzula forsterii, Silene multicaulis, Monotropa hypopitys, Doronicum orientale. Ecological circumstances: thermophilous beech forests of continental part of the southern Balkan, nder the influence of the Mediterranean climate. Tilio tomentosae-Fagenion sylvaticae Marinšek, Čami et Šilc suball. nov. hoc loco: Nomenclature type -holotypus: Tilio tomentosae-Fagetum sylvaticae Tzonev, Dimitrov, Chytrý, Roussakova, Dimova, Gussev, Pavlov, Vulchev, Vítkova, Gogoushev, Nikolov, Borisova et Ganeva 2006. Diagnostic species: Tilia cordata, Glechoma hirsuta and Tilia tomentosa *. *taxon Tilia tomentosa has phi-value 17.5 but its geographical distribution corresponds to the areal of the the suballiance. Therefore we decided to select it as a diagnostic species. Ecological circumstances: lowland thermophilous beech forests under continental climate. Supporting Information Additional supporting information may be found in the online version of this article: Appendix SI. Bibliography of relevé data Set. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. Applied Vegetation Science Doi: 10.1111/J.1654-109X.2012.01203.x © 2012 International Association for Vegetation Science 147