Cent. Eur. J. Biol.• 6(4) • 2011 • 639-658 DOI: 10.2478/s11535-011-0041-9

Central European Journal of Biology

Properties of shrubforest edges: a case study from South

Research Article

László Erdős1,*, Róbert Gallé1, Zoltán Bátori1, Mónika Papp2, László Körmöczi1

1Department of Ecology, University of Szeged, H-6726 Szeged, Hungary 2Institute of Botany and Nature Protection, University of West-Hungary, H-9400 Sopron, Hungary Received 03 January 2011; Accepted 17 April 2011

Abstract: Knowledge on edge properties is important from a conservation perspective. Our study was carried out in the ancient vegetation mosaic of the Villány Mts, South-Hungary. Sampling was conducted along eight transects, each running from a rock sward through a shrubforest patch into another rock sward. Unlike most studies, we identified edge position objectively, using a moving split-window analysis. Five habitat types along each transect were distinguished: north-facing rock sward interior, north-facing edge, shrubforest interior, south- facing edge, and south-facing rock sward interior. In the forty 2 m² plots, a total of 157 species were found. Species richness and Shannon- diversity of the edges was higher than those of the shrubforest interiors, but not significantly different from the rock swards. Cover did not differ significantly among habitat types. We found only a few edge-related species. No differences between differently-oriented edges were revealed. Species composition of the edges was influenced mostly by the rock sward matrix. We hypothesize that ecological conditions of the edges resemble those of the rock sward interiors. Thus, sward species can penetrate into shrubforest edges, entailing a similar composition of edges and rock swards, resulting in similar diversities. Edges might be viewed as refugia for valuable of rock swards.

Keywords: Natural edge • Edge effects • Edge diversity • Species richness • Edge-related species • Orientation • Matrix • Transect • Moving split-window • Villány Mts.

© Versita Sp. z o.o.

1. Introduction on the resource distribution between neighbouring communities. Unfortunately, there have been relatively Edges (also referred to as transitional zones, ecotones few empirical studies that measure species or ecological boundaries) are recognized as playing an richness or diversity within edges. Moreover, only a small important role in community ecology, landscape ecology fraction of those do have the advantage of delineating the and nature conservation [1]. Ecological responses to boundary objectively [13,14]. edges are a focus of considerable research interest Although the existence of edge-related (ecotonal) [2-4]. One of the central issues is biodiversity within edges. species (i.e. species that are restricted to edges or at According to the most wide-spread hypothesis [5-8], least occur significantly more frequently in edges) was species diversity is greater in edges than in either of the once widely accepted [5,11,15], the concept has been two adjoining communities. Another hypothesis [9-12] recently challanged [2,13]. Even the definition of the claims that diversity is increased only within blurred edges edge-related species is unclear in the literature [13]. whereas sharp edges are less diverse compared to the Again, more case-studies would be necessary. two neighbouring communities. However, new theories Considerable differences between the microclimatic predict that diversity of edges may be intermediate to that conditions of variously oriented edges can be measured of the adjoining communities [3,4]. A predictive model of [16-18]. Therefore, it is expected that differences in edge effects [3] suggests that positive, negative, neutral vegetation features may occur between north-facing or transitional edge responses may occur, depending and south-facing edges [19]. In the northern temperate

* E-mail: [email protected] 639 Properties of shrubforest edges: a case study from South Hungary

zones, south-facing edges often show stronger edge shrubforest is dominated by Quercus pubescens and responses [3]. However, it is likely that orientation has Fraxinus ornus, the height of which varies between a weaker effect on vegetation than on microclimate [3]. 4-6 m. Under the canopy, there are several shrubs Since two different communities meet at every (e.g. Ruscus aculeatus and Ligustrum vulgare) and edge, both of them should be taken into account. This lianas (mainly Tamus communis). Typical species means that both forest patches and the matrix must not in the herb layer include Dictamnus albus, Geum be neglected [20,21], and edges should be scrutinized urbanum, Polygonatum odoratum and Symphytum in relation to both adjoining communities. However, tuberosum ssp. nodosum [34]. The open rock sward several studies restrict their attention to the edges has a vegetation cover of 40-70%, while the remaining and forest interior areas, not considering the matrix surface is rock and bare soil. The dominant species [18,22-24]. is Festuca dalmatica. Other typical species are for Although alpine treeline ecotones have been example: Artemisia alba, Botriochloa ischaemum, extensively studied [14,25,26], other types of edges Chrysopogon gryllus and Melica ciliata. The abundance free from human influence form a neglected topic of mosses and lichens is considerable. The rock sward [19]. Therefore, a natural vegetation mosaic of karst is especially rich in protected and endangered species shrubforests and rock swards was chosen for our (e.g. Asplenium javorkeanum, Colchicum hungaricum, investigation in the Villány Mountains (South Hungary). Convolvulus cantabrica and Trigonella gladiata) [35,36]. We examined the species richness, Shannon-diversity and vegetation cover of the edges and compared the data to the interior areas of shrubforests and rock swards. Moreover, we examined whether there are differences between south-facing and north-facing edges. We also searched possible edge-related species. Finally, we made an attempt to reveal the effects of the matrix on the edges. Doing this, we were able to explain the observed diversity and species richness patterns. Although a wide variety of definitions exist [1,2,27], here we define edges as boundaries between adjacent communities [3,28]. In this article, we use the term matrix to refer to the community type covering more than 50% of the area [29].

2. Experimental Procedures 2.1 Study area Our research was carried out in the Villány Mountains (South Hungary) (Figure 1A). The bedrock consists of limestone and dolomite that is partly covered by loess [30]. Mean annual temperature is 10-10.5°C. The coldest month is January, the hottest July [31]. Mean annual precipitation is 670-690 mm [32]. Because of the east- west direction of the Villány Mountains, microclimates of the southern and of the northern slopes are quite different, with the southern slopes being considerably hotter and dryer [33]. The northern slopes are covered mainly by mesophilous forests, whereas the southern slopes are covered by xerophilous forests and grasslands. For our research, two regions of the Villány Mountains were chosen, Szársomlyó Mountain and Figure 1. The location of Hungary and (inset) of the Villány Mts Tenkes Mountain. On the southern slopes of these (A). Schematic diagram of the layout of a transect (plots outlined in bold were used for the statistical analyses) mountains, hundreds of karst shrubforests patches (B). representative z-score profile of a transect with two are scattered in a matrix of open rock sward. The karst significant edges (C).

640 L. Erdős et al.

The southern slopes of both mountains had been distance from randomized data for half-window size k grazed by sheep, but grazing ceased at least 50 years (expected mean), and SDexp,k is the standard deviation ago [37]. The very high number of relict (e.g. Colchicum of the squared Euclidean distance values from the hungaricum, Festuca dalmatica, Orobanche nana and randomized data for half-window size k. The overall Tamus communis), endemic (Dianthus pontederae and mean and standard deviation are computed from 99 Koeleria majoriflora) and relict-endemic (Vincetoxicum randomizations and for the whole transect. Random pannonicum) species [38] as well as some historical reference was made with random shift of species. The data from the 13th century [39] suggest that vegetation distribution of each species is shifted a random number can be considered natural, despite the former grazing. of units along the transect. Occurrences that are shifted Szársomlyó Mountain has been protected since 1944 beyond the end of the transect are wrapped back on to [40], while Tenkes Mountain is being slated for protection the opposite end [45,46]. Random shifts maintain the (A. Márkus, personal communication). spatial pattern of each individual species, randomizing only the interspecific patterns [47,48]. Z-scores for 2.2 Vegetation sampling each window midpoint position were averaged over Eight shrubforest patches consisting of old Q. pubescens 1 to 10 half-window sizes. Average z-scores were and F. ornus trees were selected randomly within the plotted against window midpoint position, resulting in vegetation mosaic. They were about 10-20 m in diameter, a z-score profile (Figure 1C). As the distribution of the a common size in xerophilous shrubforests throughout overall mean of the squared Euclidean distance from Hungary [34,41,42]. All patches were surrounded randomized data is very close to normal distribution by rock swards in every direction. 1 m² contiguous (L. Körmöczi unpublished data; on the basis of several sampling plots were arranged in transects running from thousands of randomisation tests), z-scores greater than a rock sward through an adjoining shrubforest patch into 1.65 are considered significant at a 5% probability level another rock sward (Figure 1B). Percentage cover of all [49]. One of the edges did not prove to be significant, species (both woody and non-woody) but it was included in the analyses since it was very less than 50 cm in height was estimated in April and in close to the significance level (z=1.536). Computations July. Since only the understory vegetation was sampled, were carried out using the statistical language R 2.10.1 our results apply to this vegetation layer exclusively. (R Development Core Team, , , http://www.R-project.org). 2.3 Identification of the edge positions We wanted to avoid subjective identification of edge 2.4 Selecting and grouping the plots positions. Therefore boundary delineation was done Plots were selected for the analyses as follows: To select objectively, applying a moving split-window analysis the edge plots, we defined edges to be two meters [43,44]. Moving split-window analysis is based on wide so that two plots were situated on the edge. One the spatial dependence of vegetation transect data. plot was on either side of the boundary line identified Dissimilarity of neighbouring areas of the transect by the moving split-window analysis (Figure 1B). This is expected to be low, but at the region of structural is undoubtedly arbitrary, but some considerations changes, i.e. at the edges, it becomes significantly support our decision on edge width. First, the moving high, so that it appears as a peak in a dissimilarity split-window analysis confirms this width, indicating profile diagram. Comparison of neighbouring areas is relatively sharp and thin edges [44]. Second, our performed in a “window” that consists in the simplest field experience seems to verify a boundary width of case of two adjacent plots of the transect. Increasing the approximately two meters. Also, similar edge widths window size results in a spatial series that can emphasize were found in the shrubforests of the Hungarian the significant discontinuities of the vegetation. Transdanubian Mountain Ranges [41]. As comparative function, we used the squared Shrubforest interior plots and rock sward interior Euclidean distance. The significance of the boundaries plots were selected from the central area of the patches, was tested with the z-score transformation according to where edge effects are expected to be absent or minimal Eq. 1, (Figure 1B). Data of the adjacent 1 m² plots were combined by averaging cover values. As a result, we

z=(di,k-dexp,k)/SDexp,k (Eq. 1) had 2 m² plots in each of five habitats: north-facing rock sward interior, north-facing edge, shrubforest interior,

where z is the z-score, di,k is the squared Euclidean south-facing edge, south-facing rock sward interior distance at position i if the half-window size is k, (Figure 1B). Altogether there were forty 2 m² plots dexp,k is the overall mean of the squared Euclidean distributed in five habitats and along eight transects.

641 Properties of shrubforest edges: a case study from South Hungary

In the fidelity measurements, the two slope aspects si,j=∑min(pk,i, pk,j) (Eq. 2) were pooled. Thus we had 16 rock sward plots and 16

edge plots but only 8 shrubforest interior plots in these where si,j is the Renkonen similarity index, pk,i is the

analyses. Our preliminary results indicated that unequal relative cover of species i in plot k, and pk,j is the relative plot number had adverse effects on significance. cover of species j in plot k. We drew box-plots using the Therefore, two additional 1 m² plots were selected SPSS package (version 17.0) for Windows (SPSS Inc., from the shrubforest interiors along each transect and Chicago, USA). combined into 2 m² plots. Thus 48 plots were used for In a complementary analysis, we searched for the fidelity calculations. species that avoid one of the habitats. To do this we merged the neighbouring habitats of forest and edge 2.5 Statistical analyses and determined diagnostic species that avoided the To assess the effect of habitat type (i.e. north-facing rock rock sward using JUICE 7.0 [52]. Diagnostic species of sward interior, north-facing edge, shrubforest interior, the merged groups forest and edge indicate that these south-facing edge, south-facing rock sward interior) species can not enter the rock sward habitat. Similarly, on the species richness, diversity and total cover of the rock sward and edge habitats were merged and the vegetation, we performed a series of generalized diagnostic species were determined against the linear models with nested random effects. We modeled shrubforest. species richness with a Poission error distribution, and diversity and total cover with a Gaussian error distribution. We treated transect as a random effect 3. Results and habitat type as a fixed effect in the models [50]. We also performed post-hoc multiple comparisons. 3.1 Species richness, Shannon-diversity and We used the lme4 (http://lme4.r-forge.r-project.org/) total cover of the edges and multcomp [51] packages in the R framework We found a total of 157 species in the 40 plots. 84 (R Development Core Team, Vienna, Austria, http:// species occured in the shrubforest interior areas, 113 www.R-project.org). We applied Q-Q plots and scale- species within north-facing edges, 91 within south- location plots to test the assumptions of the linear facing edges, 104 species in the north-exposed rock mixed models. Outlier samples were excluded from swards and 105 species in the south-exposed rock further analysis. swards (Supplemental Tables 1-5). To identify possible ecotonal species, we compiled a According to the generalized linear model, habitat list of species that occured exclusively in edge plots. We type had significant effect on the species richness also identified diagnostic species, that is, species that (χ2=42.91, P<0.001), Shannon-diversity (χ2=28.483, are concentrated in one of the three habitats (ignoring P<0.001), and a marginally significant effect on total slope aspect). For this purpose, fidelity values were cover (χ2= 9.6058, P= 0.04762). calculated using JUICE 7.0 [52]. The phi-coefficient The post-hoc comparisons showed that species was computed, since it is an appropriate measure of richness in the shrubforest interior was significantly lower fidelity [53]. Species with positive phi-coefficients were than that of the other habitat types (Figure 2A, Table 1). considered significant diagnostic species. Fisher’s No significant differences in species richness between exact test was carried out to exclude non-significant the other four habitat types were found. Shannon- diagnostic species. As a result, we got four groups diversity of the shrubforest interior was significantly of species: diagnostic species of the shrubforests, lower than that of the other habitat types, while we found diagnostic species of the edges, diagnostic species of no significant differences between Shannon-diversity of the rock swards and non-diagnostic species. the other four habitat types (Figure 2B, Table 1). The To study the effects of the surrounding matrix on the post-hoc comparisons did not show significant edges, we calculated the proportion of the diagnostic differences between cover values of the habitat types species of the shrubforests and of the rock swards that (Figure 2C, Table 1). occurred within each edge. In this way we were able to determine which of the two neighbouring communities 3.2 Edge-related species has a greater influence on the edge using the Mann- We found only a few potential edge-related species. Whitney test in Past 1.99 [54]. We also examined the We found 14 species which did not occur outside floristic similarity between each individual edge and of the edges (Achillea collina, Aconitum anthora, the neighbouring communities using the Renkonen Cerastium semidecandrum, Euonymus verrucosus, similarity index given by Eq. 2. Festuca valesiaca, Lactuca sp., Leontodon hispidus,

642 L. Erdős et al.

3.3 Effects of the edge orientation According to the statistical analyses, the north- and south-facing edges did not differ significantly in species richness, Shannon-diversity or total cover (Table 1). 3.4 Effects of the matrix on the species composition of the edges Edges were influenced mostly by rock swards. Within edges, the proportion of the rock sward diagnostic species was 19.81% on average, while proportion of the shrubforest diagnostic species was only 2.92% on average (Mann-Whitney U=13, P<0.0001). Greater influence of rock swards on edges was supported by the Renkonen similarity index as well: similarity was significantly higher between edges and rock swards than between edges and shrubforests (U=61, P<0.01). We found 29 diagnostic species when rock sward and edge were merged. These species live in the rock sward and edge habitats, but cannot enter the shrubforest interior. In contrast, there are only five diagnostic species of the forest and edge combined. This smaller group of species is typical of shrubforests and edges, but cannot enter the rock sward habitat.

4. Discussion

Edges are key components of landscapes, so their study is essential in landscape ecology and conservation biology [3]. In this article, we studied edges of an ancient shrubforest-rock sward mosaic in the Villány Mts (South Hungary). We examined how edge properties (species richness, diversity, vegetation cover, edge-related species, effects of the exposition and influences of the matrix), relate to the neighbouring communities. Most of the earlier studies on edge properties determined edge position subjectively [18,41,55-57]. In our study, we Figure 2. Species richness (A), Shannon-diversity (B) and total cover (C) of the five habitat types. Boxes delimit the determined edge location objectively, using the moving first and third quartiles, and horizontal lines in the boxes split-window analysis. Moreover, unlike many other indicate the median. Whiskers include all data values that fall within 1.5 multiples of the interquartile range, while studies [22,24,56-59], we examined natural edges. dots represent outliers. Both species richness and Shannon-diversity of the edge was higher than that of the forest interior (Table 1). Pseudolysimachion spicatum, Rosa arvensis, Salvia Higher species richness or diversity of edges compared sp., Sonchus asper, Sorbus domestica, Valerianella to forest interiors has been observed in several studies locusta and Veronica chamaedrys). In addition, all [18,22,60,61], although different patterns also exist of them were quite rare, occuring only in one or two [59,62,63]. It was suggested [23] that edge diversity plots, which means that their appearance in edges may only be higher than that of the forest interior if the may be random. edge is old (i.e. not newly-created) or stable for a longer We identified 7 shrubforest diagnostic species, period. This statement is in good agreement with the 3 edge diagnostic species and 19 rock sward results of this study of old, intact edges. diagnostic species (Table 2). The extremely small We found that both the species richness and the number of the edge diagnostic species is particularly diversity of edges were intermediate to those of the conspicuous. neighbouring communities, although neither measure

643 Properties of shrubforest edges: a case study from South Hungary

Species richness

Shrubforest Edge N Edge S Rock sward N Rock sward S Shrubforest — z=4.747 z=4.059 z=5.022 z=5.971 Edge N *** — z=-0.712 z=0.287 z=1.286 Edge S *** NS — z=0.999 z=1.996 Rock sward N *** NS NS — z=0.999 Rock sward S *** NS NS NS —

Shannon-diversity

Shrubforest Edge N Edge S Rock sward N Rock sward S Shrubforest — z=3.036 z=4.214 z=4.827 z=5.531 Edge N * — z=1.079 z=1.692 z=2.495 Edge S *** NS — z=0.634 z=1.498 Rock sward N *** NS NS — z=0.885 Rock sward S *** NS NS NS —

Cover

Shrubforest Edge N Edge S Rock sward N Rock sward S Shrubforest — z=2.436 z=1.199 z=2.416 z=2.522 Edge N NS — z=-1.236 z=-0.020 z=0.086 Edge S NS NS — z=1.217 z=1.322 Rock sward N NS NS NS — z=0.106 Rock sward S NS NS NS NS —

Table 1. Results of the pairwise post-hoc comparisons of the five habitat types (* P<0.05, ** P<0.01, *** P<0.001, NS: non-significant).

was significantly different from the rock sward interiors diversity due to the effects of the shadow. The fact (Figure 2, Table 1). Therefore, the shrubforest edges that species richness and diversity of edges are very under study may be termed noninteractive [4] or similar to those of rock swards may indicate that edge transitional [3] from the point of view of diversity. conditions and resources are similar to those of the rock A transitional edge response is expected if the two sward interior. adjoining habitats are rather different and if there are no Vegetation cover is expected to change along the resources concentrated along the edge [3]. Intermediate interior-edge-exterior gradient [19,41,56,68]. However, species richness of edges has been reported from we found that total cover of the understory did not differ several edge types [13,14,64]. significantly among habitat types (Table1 ). The probable Tilman & Pacala [65] stated that diversity is lowest in explanation for this is that both the shrubforest and the high-nutrient habitats due to shading by tall and plants rock sward understory communities have generally low that cover extensive areas. Mészáros et al. [66] found cover values: the former due to the shading of the trees that diversity is low under the dense canopy of a forest and shrubs, the latter due to the open rock surfaces. because a few shade-tolerant dominant species exclude The edges may have low cover values as a result of the most of the other species. According to Auclair & Goff combination of the above factors. [67], the diversity of mature dry habitats exceeds that of Two alternative definitions of edge-related (i.e. mature mesic habitats. These suggestions are in good ecotonal) species have been suggested [13]. According agreement with our field experience: soil of shrubforest to the first definition, edge-related species are those patches is much deeper than that of the rock swards, that are found only within edges. There have been resulting in more moisture and greater nutrient supplies studies that have revealed the existence of species (L. Erdős, unpublished data). This in turn entails higher confined exclusively to edge zones [13,57,59,68]. We and taller vegetation and as a consequence, reduced found 14 such species. However, since all of them were

644 L. Erdős et al.

species phi Significance It is widely accepted that in the northern hemisphere, south-facing edges exhibit greater edge influence than Shrubforest north-facing edges [3,19]. South-facing and north- Alliaria petiolata 0.341 * facing edges may differ in structure [69], frequency Fragaria moschata 0.395 * and species richness of native or invasive species Geum urbanum 0.459 ** [55,70-72], plant vitality [73], species composition [68] Polygonatum odoratum 0.289 * and species distributions [22]. However, in the present Quercus pubescens 0.275 * study, we found no significant differences between the Rosa canina 0.398 * species richness, diversity or vegetation cover of north- Tamus communis 0.480 *** facing and south-facing edges (Table 1). We conclude that characteristics of the shrub and canopy layers Edge should also be scrutinized. More comparative studies Elymus hispidus 0.312 ** with similar communities in the region are needed. Geranium rotundifolium 0.236 * Effects of the matrix on the forest patches are Taraxacum erythrospermum 0.370 ** increasingly recognized [21,74]. In the Villány Mts, the rock sward matrix considerably influences the Rock sward species composition of shrubforest edges. Burton [70] Acinos arvensis 0.252 * emphasized that there are two types of edge effects. The Arabidopsis thaliana 0.567 *** forest affects the neighbouring segment of the sward, Botriochloa ischaemum 0.338 ** while the sward affects the outer zone of the forest. Cerastium brachypetalum 0.312 ** The latter type seems to be much more important in the Chrysopogon gryllus 0.458 ** case of the vegetation mosaic of the Villány Mts. Similar Euphorbia cyparissias 0.375 ** results were reported by Dierschke [68], who found Helianthemum ovatum 0.454 *** that several species of the nonforest habitat penetrate Linaria genistifolia 0.315 * into the edge, resulting in a floristic similarity between Medicago minima 0.405 ** the edge and the open habitat. Interestingly, other Melica ciliata 0.312 ** studies have found that edges are more similar to forest interiors than to the neighbouring non-forest community Myosotis stricta 0.275 * [24,41,61]. In contrast with our results, in the shrubforest- Orobanche picridis 0.332 * rock sward mosaic of the Hungarian Transdanubian Sanguisorba minor 0.355 ** Mountain Ranges, edges are more similar to the forest Saxifraga tridactylites 0.529 *** interior [41]. The most likely explanation of this is that Sedum sexangulare 0.459 ** in the Transdanubian Mountain Ranges, edges possess Stachys recta 0.355 ** their own zone of shrubs with a height of 0.5-1 m. Under Trifolium campestre 0.316 * the dense canopy of the shrub layer, conditions are Verbascum lychnitis 0.315 * similar to the forest interior. However, this outer zone Viola arvensis 0.289 ** (also called „Saum”) is lacking in the Villány Mts, so microclimatic conditions characteristic of the rock sward Table 2. Diagnostic species of the shrubforests, edges and rock can penetrate into edges. swards (* P<0.05, ** P<0.01, *** P<0.001). This also casts light on the possible mechanisms underlying the observed diversity and species richness quite rare, we guess that none of them are true edge- pattern. We suggest that in the studied shrubforest-rock related species since it is highly probable that they sward mosaic, ecological factors within edges resemble occured within the edges by chance alone. According those within rock sward interiors. As a consequence, to the second definition, edge-related species occur rock sward species can encroach into shrubforest significantly more frequently in edges than elsewhere. edges, resulting in a similar composition of edges and Edge-related species that fit this definition have been rock swards. This in turn leads to similar diversities. reported in some previous works [41,64], while no Although both shrubforests and rock swards are significant edge-related species were found by others valuable communities, rock swards undoubtedly support [13]. Three such species were identified by our analyses many more protected, relict and endemic species. Our (Table 2). They may be considered edge-related, but we finding that shrubforest edges provide habitats for have to emphasize that the results must be treated with several rock sward species may be important from a care and more studies are needed on this topic. nature conservation perspective: if rock swards suffer

645 Properties of shrubforest edges: a case study from South Hungary

degradation events, shrubforest edges can serve in field studies and data analysis, as well as tothe as refugia for threatened plant species, making later anonymous reviewers for their useful and constructive re-colonization possible. comments on an earlier version of the paper. We are also thankful to the Inspectorate for Environment, Nature and Water for allowing us to carry out these Acknowledgements studies in the strictly protected nature reserve of the Villány Mts. We thank the technical support of the We are indepted to Ágnes Méri, Gábor Lőrinczi, Department of Plant and Geobotany of the István Maák and Tamás Morschhauser for their help University of Pécs.

References

[1] Yarrow M.M., Marín V.H., Toward conceptual test from a Nothofagus treeline in southern New cohesiveness: a historical analysis of the theory Zealand, Community Ecol., 2003, 4, 35-42 and utility of ecological boundaries and transition [15] Zólyomi B., Coenotone, ecotone and their role in zones, Ecosystems, 2007, 10, 462-476 preserving relic species, Acta Bot. Hung., 1987, 33, [2] Kark S., van Rensburg B.J., Ecotones: marginal 3-18 or central areas of transition?, Isr. J. Ecol. Evol., [16] Young A., Mitchell N., Microclimate and vegetation 2006, 52, 29-53 edge effects in a fragmented Podocarp-broadleaf [3] Ries L., Fletcher R.J. Jr., Battin J., Sisk forest in New Zealand, Biol. Conserv., 1994, 67, T.D., Ecological responses to habitat edges: 63-72 mechanisms, models, and variability explained, [17] Chen J., Franklin J.F., Spies T.A., Growing-season Annu. Rev. Ecol. Evol. Syst., 2004, 35, 491-522 microclimatic gradients from clearcut edges into [4] Strayer D.L, Power M.E., Fagan W.F., Pickett old-growth Douglas-fir forests, Ecol. Appl., 1995, 5, S.T.A., Belnap J., A classification of ecological 74-86 boundaries, BioScience, 2003, 33, 723-729 [18] Gehlhausen S.M., Schwartz M.W., Augspurger [5] Odum E.P., Fundamentals of Ecology, 3rd Ed., W.B. C.K., Vegetation and microclimatic edge effects Saunders, Philadelphia, 1971 in two mixed-mesophytic forest fragments, Plant [6] Pianka E.R., Evolutionary ecology, 3rd Ed., Harper Ecol., 2000, 147, 21-35 and Row, New York, 1983 [19] Harper K.A., MacDonald S.E., Burton P.J., Chen [7] Chiras D.D., Environmental science: action for a J., Brosofske K.D., Saunders S.C., et al., Edge sustainable future, 3rd Ed., Benjamin/Cummings, influence on forest structure and composition in Redwood City, 1991 fragmented landscapes, Conserv. Biol., 2005, 19, [8] Risser P.G., The Status of the Science Examining 768-782 Ecotones, BioScience, 1995, 45, 318-325 [20] Jules E.S., Shahani P., A broader ecological context [9] Brown J.H., Gibson A.C., Biogeography, C.V. to habitat fragmentation: Why matrix habitat is Mosby, St. Louis, 1983 more important than we thought, J. Veg. Sci., 2003, [10] van Leeuwen C.G., A relation theoretical approach 14, 459-464 to pattern and process in vegetation, Wentia, 1966, [21] Kupfer J.A., Malanson G.P., Franklin S.B., Not 15, 25-46 seeing the ocean for the islands: the mediating [11] van der Maarel E., On the establishment of plant influence of matrix-based processes on forest community boundaries, Ber. Deut. Bot. Ges., 1976, fragmentation effects, Global Ecol. Biogeogr., 89, 415-443 2006, 15, 8-20 [12] van der Maarel E., Ecotones and ecoclines are [22] Matlack G.R., Vegetation dynamics of the forest different, J. Veg. Sci., 1990, 1, 135-138 edge – trends in space and successional time, J. [13] Lloyd K.M., McQueen A.A.M., Lee B.J., Wilson Ecol., 1994, 82, 113-123 R.C.B., Walker S., Wilson J.B., Evidence on [23] Harper K.A., Macdonald S.E., Structure and ecotone concepts from switch, environmental and composition of edges next to regenerating clear- anthropogenic ecotones, J. Veg. Sci., 2000, 11, cuts in mixed-wood boreal forest, J. Veg. Sci., 903-910 2002, 13, 535-546 [14] Stowe C.J., Kissling W.D., Ohlemüller R., Wilson [24] Santos A.M.M., Santos B.A., Are the vegetation J.B., Are ecotone properties scale-dependent? A structure and composition of the shrubby Caatinga

646 L. Erdős et al.

free from edge influence?, Acta Bot. Bras., 2008, dalmatica (Hack.) Richt. in Hungary and its 22, 1077-1084 statistical comparison with ssp. pseudodalmatica [25] Armand A.D., Sharp and Gradual Mountain (Kraj.) Soó, Ann. Univ. Budapest, S. Biol., 1964, 7, Timberlines as a Result of Species Interactions, 143-156, (in German) In: Hansen A.J., Di Castri F., (Eds.), Landscape [36] Dénes A., A Villányi-hegység Chrysopogono- Boundaries: Consequences for Biotic Diversity and Festucion dalmaticae társulásai (Chrysopogono- Ecological Flows, Springer-Verlag, New York, 1992 Festucion dalmaticae associations of the Villány [26] Camarero J.J., Gutiérrez E., Fortin M.-J., Spatial Mountains), In: Csontos P., (Ed.), Sziklagyepek patterns of plant richness across treeline ecotones szünbotanikai kutatása (Ecological studies on in the Pyrenees reveal different locations for rock swards), Scientia Kiadó, Budapest, 1998, (in richness and tree cover boundaries, Global Ecol. Hungarian) Biogeogr., 2006, 15, 182-191 [37] Horvát A.O., A mecseki tájak erdei növénytársulásai [27] Hufkens K., Scheunders P., Ceulemans R., (Forest associations of the Mecsek Mountains), Ecotones in vegetation ecology: methodologies Janus Pannonius Múz. Évk., 1963, 8, 33-53, (in and definitions revisited, Ecol. Res., 2009, 24, 977- Hungarian) 986 [38] Dénes A., A Villányi-hegység flóra- és [28] Harris L.D., Edge effects and conservation of biotic vegetációkutatásának története, eredményeinek diversity, Conserv. Biol., 1988, 2, 330-332 összefoglalása, különös tekintettel a védett és [29] Forman R.T.T., Land mosaics: the ecology of ritka fajok előfordulására (History of research landscapes and regions, Cambridge University on flora and vegetation in the Villány Mountains, Press, Cambridge, 1995 a summary of results, with special regard to the [30] Lovász Gy., A Baranyai-dombság, a Mecsek occurence of rare and protected species), In: és a Villányi-hegység (Baranya Hills, Mecsek Uherkovich Á., (Ed.), A Villányi-hegység botanikai Mountains and Villány Mountains), In: Ádám L., és zoológiai alapfelmérése (The flora, vegetation Marosi S., Szilárd J., (Eds.), A Dunántúli-dombság and fauna of the Villány Mountains), Baranya (Dél-Dunántúl) (Transdanubian Hills (South megyei Múzeumok Igazgatósága, Pécs, 2000, (in Transdanubia)), Akadémiai Kiadó, Budapest, Hungarian) 1981, (in Hungarian) [39] Szita L., Nagyharsány, Száz magyar falu [31] Szilárd J., Éghajlati adottságok (Climatic könyvesháza Kht., Budapest, 2002, (in Hungarian) conditions), In: Ádám L., Marosi S., Szilárd J., [40] Reuter C., Baranya megye: Szársomlyó (Baranya (Eds.), A Dunántúli-dombság (Dél-Dunántúl) county: Szársomlyó Mountain), In: Kopasz M., (Transdanubian Hills (South Transdanubia)), (Ed.), Védett természeti értékeink (Protected Akadémiai Kiadó, Budapest, 1981, (in Hungarian) natural values in Hungary), Mezőgazdasági Kiadó, [32] Ambrózy P., Kozma F., Éghajlat (Climate), In: Marosi Budapest, 1976, (in Hungarian) S., Somogyi S., (Eds.), Magyarország kistájainak [41] Jakucs P., Dynamische Verbindung der Wälder katasztere II (Cadaster of the basic landscape und Rasen (Dynamic connections of forests and units of Hungary II), MTA Földrajztudományi Kutató grasslands), Akadémiai Kiadó, Budapest, 1972, (in Intézet, Budapest, 1990, (in Hungarian) German) [33] Horvát A.O., Papp L., A nagyharsányi Szársomlyón [42] Borhidi A., Magyarország növénytársulásai (Plant végzett mikroklímamérések eredményei (Results associations of Hungary), Akadémiai kiadó, of the microclimate measurements made on Budapest, 2003, (in Hungarian) Szársomlyó Mountain near Nagyharsány), [43] Ludwig J.A., Cornelius J.M., Locating Janus Pannonius Múz. Évk., 1964, 9, 43-51, (in discontinuities along ecological gradients, Hungarian) Ecology, 1987, 68, 448-450 [34] Dénes A., A Mecsek és a Villányi-hegység [44] Körmöczi L., On the sensitivity and significance karsztbokorerdői (Karst shrubforests of the test of biotic boundary detection, Community Ecol., Mecsek Mountains and of the Villány Mountains), 2005, 6, 75-81 Janus Pannonius Múz. Évk., 1995, 39, 5-31, (in [45] Palmer M.W., van der Maarel E., Variance in Hungarian) species richness, species association, and niche [35] Simon T., Entdeckung und Zönologie der Festuca limitation, Oikos, 1995, 73, 203-213 dalmatica (Hack.) Richt. in Ungarn und ihr [46] Horváth A., INFOTHEM program: new possibilities statistischer Vergleich mit ssp. pseudodalmatica of spatial series analysis based on information (Kraj.) Soó (Discovery and coenology of Festuca theory methods, Tiscia, 1998, 31, 71-84

647 Properties of shrubforest edges: a case study from South Hungary

[47] Bartha S., Kertész M., The importance of species richness: An example for woody plants, neutral-models in detecting interspecific spatial Forest Ecol. Manag., 2010, 259, 266-274 associations from ‚trainsect’ data, Tiscia, 1998, 31, [62] Baez S., Balslev H., Edge effects on palm diversity 85-98 in rain forest fragments in western Ecuador, [48] Horváth A., Makrai L., Variability of spatial Biodivers. Conserv., 2007, 16, 2201-2211 dependence within a pioneer plant community, [63] Łuczaj Ł., Sadowska B., Edge effects of different Tiscia, 2000, 32, 19-29 groups of organisms: vascular plants, bryophyte and [49] Zalatnai M., Körmöczi L., Tóth T., Community fungi species richness across a forest-grassland boundaries and edaphic factors in saline-sodic border, Folia Geobot. Phytotax., 1997, 32, 343-353 grassland communities along an elevation gradient, [64] Walker S., Wilson J.B., Steel J.B., Rapson G.L., Tiscia, 2007, 36, 7-15 Smith B., King W. McG., et al., Properties of [50] Crawley, M.J., The R Book, John Wiley, New York, ecotones: Evidence from five ecotones objectively 2007 determined from a coastal vegetation gradient, J. [51] Hothorn T., Bretz F., Westfall P., Simultaneous Veg. Sci., 2003, 14, 579-590 Inference in General Parametric Models, [65] Tilman D., Pacala S., The Maintenance of Species Biometrical J., 2008, 50, 346-363 Richness in Plant Communities, In: Ricklefs R.E., [52] Tichý L., JUICE, software for vegetation Schluter D., (Eds.), Species Diversity in Ecological classification, J. Veg. Sci., 2002, 13, 451-453 Communities, University of Chicago Press, [53] Tichý L., Chytrý M., Statistical determination of Chicago, 1993 diagnostic species for site groups of unequal size, [66] Mészáros I., Jakucs P., Précsényi P., Diversity J. Veg. Sci. 2006, 17, 809-818 and niche changes of shrub species within forest [54] Hammer Ř., Harper D.A.T., Ryan P.D., PAST: margin, Acta Bot. Hung., 1981, 27, 421-437 Paleontological Statistics Software Package for [67] Auclair A.N., Goff F.G., Diversity relations of upland Education and Data Analysis, Palaeontol. Electron., forests in the western Great Lakes area, Am. Nat., 2001, http://palaeo-electronica.org/2001_1/past/ 1971, 105, 499-528 issue1_01.htm [68] Dierschke H., Saumgesellschaften im Vegetations- und [55] Brothers T.S., Spingarn A., Forest fragmentation Standortsgefälle an Waldrändern (Edge communities and alien plant invasion of Central Indiana old- at vegetation and habitat gradients in forest edges), growth forests, Conserv. Biol., 1992, 6, 91-100 Script. Geobot., 1974, 6, 1-246, (in German) [56] Oosterhoorn M., Kappelle M., Vegetation structure [69] Papp M., A csáfordjánosfai tölgy-kőris-szil ligeterdő and composition along an interior-edge-exterior szegélyének fa- és cserjefaj összetétele, valamint gradient in a Costa Rican montane cloud forest, szerkezeti jellemzői (Tree and shrub composition Forest Ecol. Manag., 2000, 126, 291-307 and structural characteristics of the edge of an oak- [57] Dutoit T., Buisson E., Gerbaud E., Roche P., Tatoni elm-ash woodland near Csáfordjánosfa), Kitaibelia, T., The status of transitions between cultivated 2008, 13, 185, (in Hungarian) fields and their boundaries: ecotones, ecoclines or [70] Burton J.P., Effects of clearcut edges on trees in edge effects?, Acta Oecol., 2007, 31, 127-136 the Sub-boreal spruce zone of Northwest-Central [58] Oliveira M.A., Grillo A.S., Tabarelli M., Forest edges British Columbia, Silva Fenn., 2002, 36, 329-352 in the Brazilian Atlantic forest: drastic changes in [71] Fraver S., Vegetation responses along edge-to- tree species assemblages, Oryx, 2004, 38, 389- interior gradients in the mixed hardwood forests of 394 the Roanoke River Basin, North Carolina, Conserv. [59] Günter S., Weber M., Erreis R., Aguirre N., Influence Biol., 1994, 8, 822-832 of distance to forest edges on natural regeneration [72] Kivistö L., Kuusinen M., Edge effects on the of abandoned pastures: a case study in the tropical epiphytic lichen flora of Picea abies in middle mountain rain forest of Southern Ecuador, Eur. J. boreal Finland, Lychenologist, 2000, 32, 387-398 Forest Res., 2007, 126, 67-75 [73] Hylander K., Aspect modifies the magnitude of [60] De Casenave J.L, Pelotto J.P., Protomastro J., edge effects on bryophyte growth in boreal forests, Edge-interior differences in vegetation structure J. Appl. Ecol., 2005, 42, 518-525 and composition in a Chaco semi-arid forest, [74] Nascimento H.E.M., Andrade A.C.S., Camargo Argentina, Forest Ecol. Manag., 1995, 72, 61-69 J.L.C., Laurance W.F., Laurance S.G., Ribeiro [61] Gonzalez M., Ladet S., Deconchat M., Cabanettes J.E.L., Effects of the surrounding matrix on tree A., Alard D., Balent G., Relative contribution of recruitment in Amazonian Forest Fragments, edge and interior zones to patch size effect on Conserv. Biol., 2006, 20, 853-860

648 L. Erdős et al.

Appendix

Transect serial number 1 2 3 4 5 6 7 8

Acer platanoides - 0.05 ------Acinos arvensis - - - - - 1.75 - - Ajuga genevensis - 0.25 ------Alliaria petiolata - 0.05 - 1.25 - - 0.75 7.75 Allium flavum - - - - 0.05 0.05 0.1 - Allium sphaerocephalon - - 0.05 - - - - - Anthriscus cerefolium - - - - 0.05 - 0.05 - Arenaria serpyllifolia - - 0.1 - 0.05 - - - Arrhenatherum elatius ------2.5 - Artemisia alba - 0.25 ------Arum maculatum - - - - 0.05 - - - Asplenium javorkeanum 0.5 - 0.25 - 1 0.25 - - Brachypodium rupestre ------0.05 Bromus squarrosus - 0.05 ------Bromus sterilis - - - - 0.05 - - - Buglossoides purpureo-coerulea - 3 ------Calepina irregularis 4 1.05 ------Campanula bononiensis ------0.55 - Campanula rapunculus ------0.3 - Carex michelii - - - - 0.05 - - - Cerastium brachypetalum - 0.05 0.05 - 0.1 0.05 - - Cleistogenes serotina - - - - 1.55 1.05 - - Colchicum hungaricum - - - - 0.05 - - - Convolvulus cantabrica - - 0.5 - - - - - Corydalis solida ------1.05 0.5 Crataegus monogyna ------0.25 - Dianthus giganteiformis - - - - - 1.75 - - Dictamnus albus - - - - 9.5 7.5 - - Elymus hispidus 0.05 0.55 4 - - 2 - - Eryngium campestre - - - - - 2.05 - - Euphorbia cyparissias - 0.05 ------Euphorbia epithymoides ------0.25 - Fallopia convolvulus 0.05 0.05 0.1 0.25 0.3 0.05 0.3 0.75 Festuca dalmatica - 0.05 - - 0.05 3.75 - - Festuca rupicola ------4 - Fragaria moschata 0.55 - - - 0.5 - - - Fragaria vesca ------6.5 - Fragaria viridis 0.8 - - - 1.3 - - - Fraxinus ornus 2.5 0.05 - 0.05 0.05 3.55 - 1.25 Fumaria officinalis - - 0.05 - 0.1 - - - Gagea pratensis ------0.1 - Galium aparine 0.75 0.1 2.75 - 0.05 0.1 0.1 0.1 Galium lucidum - 1.5 0.5 - 3.25 - - - Geranium columbinum - - - - - 0.1 - - Geranium rotundifolium - 0.1 0.55 1.5 0.05 - - - Geranium sanguineum ------0.55 - Geum urbanum - 0.75 - - - - 0.1 - Glechoma hirsuta ------0.05 - Hypericum perforatum - 0.1 - 0.05 - 0.05 - - Lactuca viminea - 0.5 - - 0.1 - - -

Supplementary Table 1. Species cover data of the 2 m² shrubforest plots.

649 Properties of shrubforest edges: a case study from South Hungary

Transect serial number 1 2 3 4 5 6 7 8

Lamium amplexicaule - - 0.1 - 0.05 - - - Lamium purpureum 0.1 - - 0.05 0.1 - 0.05 - Lathyrus sphaericus 0.05 0.05 0.1 - 0.3 0.1 - - Ligustrum vulgare - - - 2.5 - - - - Medicago minima - - 0.55 - - - - - Melica ciliata - 0.25 2 - - 0.1 - - Mercurialis perennis ------2.05 - Myosotis stricta - - 0.05 - - 0.05 - - Orlaya grandiflora 0.05 3.75 7.5 - 0.1 1 - - Ornithogalum umbellatum ------0.75 - Papaver confine - - 0.05 - - - - - Pisum elatius - - 2.75 - - - - - Poa pratensis agg. - 1.75 - - 2.25 - 0.5 0.05 Polygonatum odoratum 0.25 2.75 - - - 0.75 - 0.05 Quercus pubescens 2.5 0.5 37 - - - - 1 Rosa canina 0.05 0.1 ------Ruscus aculeatus ------30 65 Sanguisorba minor - - 0.3 - - - - - Securigera varia - 0.5 ------Stachys recta - - 0.25 - - - - - Stellaria media - 0.05 - - 1 - 0.25 - Symphytum tuberosum 0.5 ------Tamus communis 11 5.75 - 0.75 0.75 - - - Taraxacum erythrospermum - 0.05 - - - 0.75 0.05 - Teucrium chamaedrys 3.5 12 9 - - 1 3 - Thlaspi perfoliatum - 0.1 - - - 0.05 0.05 - Thymus praecox - - - - - 1 - - Tragopogon dubius - - - - 0.25 0.25 - - Veronica austriaca - 0.3 ------Veronica hederifolia - - 0.05 0.3 0.05 - 0.3 - Vicia angustifolia 0.05 0.05 2.25 - 0.3 0.1 0.05 - Vincetoxicum officinale ------3 - Viola arvensis - 0.1 0.75 - 0.3 0.05 - - Viola odorata 0.25 ------0.05

continuedSupplementary Table 1. Species cover data of the 2 m² shrubforest plots.

Transect serial number 1 2 3 4 5 6 7 8

Achillea collina ------0.25 Acinos arvensis 0.05 0.05 - 0.05 - 0.3 - - Aconitum anthora ------0.55 - Ajuga genevensis 1.75 1.25 0.75 - - - - - Alliaria petiolata ------0.1 2.55 Allium flavum - - 0.05 - 0.5 0.05 0.55 0.05 Allium sphaerocephalon 0.05 0.1 0.1 - - - - - Alyssum alyssoides - - 0.05 0.05 0.05 0.1 - - Anthericum ramosum ------7.5 Anthriscus cerefolium - - - - 1 - - - Arabidopsis thaliana - - 0.05 - - - - - Arenaria serpyllifolia - 0.05 0.1 0.05 0.05 0.1 - - Arrhenatherum elatius ------6 0.5

Supplementary Table 2. Species cover data of the 2 m² north-facing edge plots.

650 L. Erdős et al.

Transect serial number 1 2 3 4 5 6 7 8

Artemisia alba 17 ------Asperula cynanchica ------0.25 Asplenium javorkeanum - - 0.25 0.05 - 0.05 - - Botriochloa ischaemum 2 - 3.75 - 0.55 8 - - Brachypodium rupestre ------11.5 Bromus squarrosus - 0.1 0.05 - - 0.05 - - Buglossoides arvensis - 0.1 - 1.3 - - - - Buglossoides purpureo-coerulea - 0.75 - 0.5 - - - - Calepina irregularis 9.5 15 ------Campanula bononiensis ------0.25 1.25 Campanula rapunculus ------0.3 0.5 Carex michelii ------0.05 Centaurea bibersteinii - 0.3 - - 3.5 1.25 - - Cerastium brachypetalum 0.1 0.05 0.1 - 0.05 0.55 - - Cerastium semidecandrum - - - - 0.05 - - - Cleistogenes serotina 2.5 0.05 - - 3 25 - - Colchicum hungaricum - - - - - 1.25 - - Convolvulus cantabrica - - 1.3 0.25 1.05 2.5 - - Corydalis solida - - 0.05 - - - 0.1 0.3 Cruciata pedemontana - - - - 0.05 0.05 - - Crupina vulgaris - - - - - 0.5 - - Dianthus giganteiformis - - 1.25 - - 1.55 - - Dianthus pontederae ------0.05 - Dictamnus albus - - - - - 0.5 - - Elymus hispidus 4 0.25 2.75 3.5 15 6.5 0.05 1 Euonymus verrucosus - - - 1 - - 0.05 - Euphorbia cyparissias 0.75 0.05 0.05 - - - 0.05 3.25 Fallopia convolvulus 0.05 0.1 - 0.05 0.05 - - 0.05 Festuca dalmatica - 0.5 10 - - 4.25 - - Festuca rupicola ------13 1.5 Festuca valesiaca ------0.25 Fragaria vesca ------0.05 - Fraxinus ornus - 2 - 4 - 1.5 1.5 - Fumaria officinalis - 0.1 0.05 0.05 0.05 0.05 - - Gagea pratensis ------0.05 - Galium aparine 0.1 - 0.3 1.75 1.75 0.1 - 0.5 Galium lucidum - - 0.8 - - - 2.25 - Geranium columbinum - - - - - 0.05 - - Geranium molle - - - - 0.05 - - - Geranium rotundifolium 0.05 1.5 0.5 4 0.05 0.05 - - Helianthemum ovatum - - - 0.05 - - - 0.05 Helleborus odorus ------3.5 1.55 Hypericum perforatum 0.05 2.55 0.05 - 0.05 0.3 - - variegata - - 0.5 0.05 - - - - Koeleria cristata ------1.5 0.75 Lactuca sp. 0.25 ------Lactuca viminea - - - - - 0.05 - - Lamium amplexicaule - 0.05 0.1 - 0.05 0.05 - - Lamium purpureum 0.1 0.05 - 0.05 0.05 - 0.05 - Lathyrus sphaericus 0.3 0.5 0.1 - 0.05 0.1 - - Linaria genistifolia - - - - 0.05 - - - Medicago minima - 0.5 0.55 0.05 - - - -

continuedSupplementary Table 2. Species cover data of the 2 m² north-facing edge plots.

651 Properties of shrubforest edges: a case study from South Hungary

Transect serial number 1 2 3 4 5 6 7 8

Melica ciliata 0.25 1.5 3 1 1.55 1 - - Muscari comosum 0.25 - - - 0.05 - - - Myosotis stricta - 0.05 0.05 - 0.5 0.1 - - Orlaya grandiflora 1 8.5 16 8 0.55 2.55 - 0.5 Ornithogalum umbellatum ------0.25 - Papaver confine - - - - 0.05 0.1 - - Phleum phleoides 4 - 0.05 2.05 - - - - Pisum elatius - 0.05 1.75 0.1 - - - - Poa pratensis agg. 0.25 - 0.5 - 0.05 - 0.1 2.55 Polygonatum odoratum - 3.75 - 1.55 - - - 8 Potentilla arenaria - - 2.25 - 13 - 0.05 0.5 Pseudolysimachion spicatum - 0.25 ------Quercus pubescens - - 0.5 - - - - - Ranunculus illyricus ------0.05 0.1 Rosa arvensis ------0.75 Rosa canina - - - - 0.05 - - - Ruscus aculeatus ------46.5 - Salvia sp. ------3 - Sanguisorba minor 0.05 1 1 0.5 - 0.05 - 0.25 Saxifraga tridactylites - 0.05 - - 0.25 - - - Sedum telephium - 0.05 - - - - - 0.3 Silene otites - - - - - 1.5 - - Sonchus asper - - 0.05 - - - - - Sorbus domestica ------0.5 - Stachys recta - 1 2.25 - 0.05 - - - Stellaria media - - 0.05 - - - - - Stipa pulcherrima - - 0.05 - - - - - Symphytum tuberosum ------0.1 Tamus communis 7 5.75 ------Taraxacum erythrospermum - 0.3 0.75 - 1.5 0.05 0.05 0.05 Teucrium chamaedrys 3.05 17 6 8 - 1 1.3 6 Thlaspi perfoliatum - - - 0.05 - 0.05 - - Thymus praecox - - - - 0.25 3 0.05 0.5 Thymus serpyllum - 0.5 ------Tilia tomentosa - - - - 28.5 - - 0.05 Tragopogon dubius - 0.05 - - 0.05 - - - Trifolium arvense - - - - 0.8 - - - Trifolium campestre - - 0.05 - - - - - Valerianella coronata - - 0.05 - - 0.05 - - Valerianella locusta - - 0.05 - - - - - Veronica chamaedrys ------0.25 - Veronica hederifolia 0.05 - 0.05 0.05 0.1 - - - Veronica praecox - - - - 0.05 0.05 - - Vicia angustifolia 0.05 1 0.3 0.8 0.1 0.05 0.05 - Vincetoxicum officinale ------0.05 - Viola arvensis 0.1 0.3 0.3 0.1 0.05 0.1 - - Viola hirta ------0.5 1.25 Viola odorata ------1

continuedSupplementary Table 2. Species cover data of the 2 m² north-facing edge plots.

652 L. Erdős et al.

Transect serial number 1 2 3 4 5 6 7 8

Acinos arvensis 0.05 0.05 0.1 - 0.05 1.5 - - Ajuga genevensis 0.5 ------Alliaria petiolata ------5.5 9.05 Allium flavum - 0.05 0.05 - 0.05 0.1 - 0.25 Allium sphaerocephalon 0.1 0.3 0.1 - - - - - Alyssum alyssoides - - 0.8 0.05 0.05 0.3 - - Anthericum ramosum ------10 Anthriscus cerefolium ------0.1 - Arabis auriculata - - 0.25 - - 0.1 - - Arenaria serpyllifolia - 0.05 0.5 - 0.3 0.05 - - Arrhenatherum elatius ------0.05 - Artemisia alba 1 1 ------Asplenium javorkeanum - 0.5 0.3 0.55 - 0.3 - - Botriochloa ischaemum - 1 - 0.05 7.75 8 - 0.5 Bromus squarrosus - - 0.05 - 0.05 0.05 - - Bromus sterilis 0.05 - - - - - 1.05 - Buglossoides arvensis 0.05 - - 0.25 - 0.05 - - Calepina irregularis 9 1.5 ------Carex michelii ------1.5 Centaurea bibersteinii - 0.55 - - 10 1 - - Cerastium brachypetalum 0.05 0.5 0.3 0.3 1.5 0.55 - 0.05 Chrysopogon gryllus - 2.05 ------Cleistogenes serotina 1.5 4.5 - - 4.5 7 - - Convolvulus cantabrica 0.05 - 0.05 - 0.8 2.25 - 0.75 Corydalis solida ------0.3 0.5 Cruciata pedemontana - - - - 0.05 - - - Crupina vulgaris - - - - - 0.05 - - Dianthus giganteiformis - 0.25 1.75 - - 2.5 - - Dictamnus albus - - - 4 2.55 0.3 - - Echium vulgare ------2 Elymus hispidus 2.5 - 1.5 4 1 11 0.5 0.8 Eryngium campestre - - - - 0.25 - - - Euphorbia cyparissias 0.05 0.1 0.1 2.75 - - - 3 Euphorbia helioscopia - - - - - 0.05 - - Fallopia convolvulus 0.1 0.1 0.05 0.1 - 0.1 5.25 1.5 Festuca dalmatica 2 5.75 5.75 - - - - 5.25 Festuca rupicola ------1.5 - Fragaria vesca ------0.5 Fragaria viridis - - - - 0.5 - - 1 Fraxinus ornus 0.05 15 - - 12.5 - 0.1 0.1 Fumaria officinalis 0.25 0.05 0.25 0.05 0.25 0.05 - 0.05 Galium aparine 0.3 0.05 0.25 0.1 0.05 - 0.1 6.5 Galium lucidum ------0.05 Geranium columbinum - - - - 0.3 0.1 - - Geranium rotundifolium 0.05 0.3 0.3 2.5 0.1 0.05 2 0.05 Geranium sanguineum - - 1.25 - - - 0.1 0.05 Helianthemum ovatum - 1.75 - - 0.5 0.05 - 0.05 Hypericum perforatum 0.05 0.5 0.5 0.05 0.8 1 - 0.05 Iris variegata - - - 0.05 - - - - Koeleria cristata ------0.05 5.75 Lamium amplexicaule - 0.05 0.5 1.25 0.1 1.05 - - Lamium purpureum 0.05 - - - - - 2.25 -

Supplementary Table 3. Species cover data of the 2 m² south-facing edge plots.

653 Properties of shrubforest edges: a case study from South Hungary

Transect serial number 1 2 3 4 5 6 7 8

Lathyrus sphaericus - - - - 0.8 0.05 - - Leontodon hispidus - - 0.05 - - - - - Ligustrum vulgare ------15 Linaria genistifolia - - - - - 1.3 - - Medicago minima - - 11 - 0.25 - - - Melica ciliata - 1.5 7 4 - 1.5 - 0.05 Myosotis stricta - 0.1 0.05 0.1 1.55 0.3 - - Orlaya grandiflora 3 5.75 8.5 10 9 8.25 2.5 0.5 Orobanche picridis - - - 0.1 - - - - Papaver confine - - 0.1 0.05 0.05 - - - Phleum phleoides - - 0.05 - - - - - Pisum elatius - - 4 - - - - - Poa pratensis agg. - - - - 0.75 - 0.75 2.75 Polygonatum odoratum - - - 0.05 - - - - Potentilla arenaria - - 3 1.75 - - 0.25 0.5 Quercus pubescens 1.5 - 0.05 - - 20 - 0.05 Ranunculus illyricus ------0.05 0.25 Sanguisorba minor - 2 1 0.75 0.05 0.75 - 0.05 Saxifraga tridactylites - 0.05 0.1 - - - - - Securigera varia - 0.05 - 2 - - - - Sisymbrium orientale - 0.05 ------Stachys recta - 1 - - - - - 3 Symphytum tuberosum 1.5 ------Taraxacum erythrospermum 1 0.55 - - 0.3 0.25 0.75 - Teucrium chamaedrys 2.5 0.25 - 11.5 3 - 0.8 3 Thlaspi perfoliatum - 0.05 - - 0.1 - 0.1 0.05 Thymus praecox - - - - 1 1.75 1 0.05 Thymus serpyllum - 3 ------Tilia tomentosa ------0.05 - Tragopogon dubius - 0.1 0.05 - 0.55 0.75 - - Trifolium arvense - - - - 0.5 - - - Valerianella coronata - - 0.25 - 0.05 0.1 - - Verbascum lychnitis - - 0.8 - 0.05 0.05 - - Veronica hederifolia - - 0.05 0.1 - - 0.5 - Veronica persica - - 0.05 - - - - - Veronica praecox - - - - - 0.05 - - Vicia angustifolia 0.05 - 0.05 - 0.05 0.75 0.05 0.1 Vicia hirsuta ------0.05 - Viola arvensis 0.05 0.1 0.05 0.1 0.1 0.1 0.3 0.1

continuedSupplementary Table 3. Species cover data of the 2 m² south-facing edge plots.

Transect serial number 1 2 3 4 5 6 7 8

Acinos arvensis - 0.3 - 0.3 - 0.1 - 0.05 Ajuga genevensis 0.75 0.55 - 1 - - - - Alliaria petiolata - - 0.25 - - - - 0.05 Allium flavum - - 0.1 - 0.55 0.05 0.5 0.1 Allium sphaerocephalon 0.05 0.3 0.1 0.55 - - - - Alyssum alyssoides - 0.05 - 0.5 - 1.55 - - Anthericum ramosum ------1.5 2.5 Anthriscus cerefolium - - 0.05 - - - - - Arabidopsis thaliana - - - 0.1 - 0.1 - -

Supplementary Table 4. Species cover data of the 2 m² north-facing rock sward plots.

654 L. Erdős et al.

Transect serial number 1 2 3 4 5 6 7 8

Arabis turrita ------0.05 Arenaria serpyllifolia - 0.1 - 0.1 - 0.1 - - Arrhenatherum elatius ------1.25 Artemisia alba 3.5 ------Asplenium javorkeanum 1.05 0.25 0.05 0.55 - - - - Berteroa incana - - - - 10 - - - Botriochloa ischaemum 5.5 - - 0.05 10 12.5 - 1.75 Bromus squarrosus - 0.25 - 0.1 - 0.05 - - Bromus sterilis - - 0.05 - - - - 0.75 Buglossoides arvensis - 0.5 - 2.25 - - - - Calepina irregularis 3 7 - - 0.05 0.3 - - Carex michelii ------0.05 0.05 Centaurea bibersteinii - 0.1 - - 0.5 10 - - Cerastium brachypetalum 0.05 0.25 - 0.3 1.25 2 0.3 0.3 Chrysopogon gryllus - - - - 18 - - - Cleistogenes serotina 1.25 3.5 - 0.05 0.05 14 - - Convolvulus cantabrica 0.25 3.5 - 0.5 - 2 - - Corydalis solida ------0.3 0.3 Cruciata pedemontana - - - - 0.75 0.1 - - Crupina vulgaris - - - - - 0.3 - - Dianthus giganteiformis - 1.5 - 0.75 0.05 - - - Dianthus pontederae ------1.75 - Dictamnus albus - - - - - 4 - - Elymus hispidus 1 - 0.55 3 5 3.75 3 6 Erodium cicutarium - 3.5 ------Eryngium campestre - - - - - 0.25 - - Euphorbia cyparissias 1 0.05 - 0.05 - 0.05 0.55 1.05 Euphorbia epithymoides ------0.05 Fallopia convolvulus 0.1 0.05 0.05 - 0.75 - 0.1 0.05 Festuca dalmatica - 1 1 - - - - - Festuca rupicola - - 0.05 - 1.55 - 8 6.5 Fraxinus ornus - 3 5.5 - - - 2 2 Fumaria officinalis 0.1 0.75 0.3 0.3 0.1 0.3 - - Gagea pratensis ------0.05 - Galium aparine 0.3 0.05 1.25 0.5 - 0.05 - 0.1 Galium lucidum - - 0.5 - - - 2 0.8 Geranium columbinum - - - - - 0.1 - - Geranium molle - - - - 1 - - - Geranium pusillum - - - - 0.1 - - - Geranium rotundifolium 0.05 0.55 0.25 1 0.3 0.05 - 0.25 Geranium sanguineum - - 3 - - - 0.05 0.3 Helianthemum ovatum - 1 1.5 4.05 - 0.25 0.05 - Hesperis tristis - - - - - 0.05 - - Hieracium bauhini ------0.1 - Hypericum perforatum 0.05 0.05 - - 0.05 0.55 - - Iris variegata - - - - 3.5 - 0.75 - Koeleria cristata ------0.8 2 Lamium amplexicaule - 0.1 0.05 0.1 0.1 0.1 - - Lamium purpureum 0.1 - - - 0.1 - - 0.05 Lathyrus sphaericus 0.5 0.25 - - 0.1 1.75 0.05 - Linaria genistifolia - - - - - 0.1 - 0.05 Medicago minima - 5 - 14.5 0.25 1.25 - - Melica ciliata 1 1.05 1.75 0.05 1.55 1.55 2.05 3

continuedSupplementary Table 4. Species cover data of the 2 m² north-facing rock sward plots.

655 Properties of shrubforest edges: a case study from South Hungary

Transect serial number 1 2 3 4 5 6 7 8

Minuartia verna ------0.25 - Muscari comosum 0.75 ------Myosotis stricta 0.05 0.3 0.05 0.1 1.25 3 - - Orlaya grandiflora 2.5 6 10 20 3.75 15 25 22.5 Ornithogalum umbellatum ------0.05 - Orobanche picridis - - - 0.05 - - - 0.25 Papaver confine - - 0.05 0.1 - 0.1 - - Phleum phleoides - - 1.5 2.25 - - - - Pisum elatius - - 1.5 0.8 - - - - Poa pratensis agg. ------1.75 0.05 Polygonatum odoratum 1.5 - - 0.75 - - - - Potentilla arenaria - - - 4.75 - - 7.5 3.5 Ranunculus illyricus ------0.3 0.05 Sanguisorba minor 1.5 0.05 0.75 0.5 - 1.75 0.1 0.5 Saxifraga tridactylites - 0.05 0.05 0.05 - - 0.05 - Securigera varia - - 8 - - - - - Sedum sexangulare ------0.3 0.05 Sedum telephium - - - - 1.25 - 0.1 1.05 Sisymbrium orientale - - - 1 - - - - Stachys recta - 2 - 0.05 1.05 - 1.25 1 Stellaria media ------1.5 Tamus communis - 0.5 ------Taraxacum erythrospermum ------0.3 Teucrium chamaedrys 0.05 10 - 6 - - 1.5 5.75 Thlaspi perfoliatum - - - - - 0.05 0.1 0.5 Thymus odoratissimus - - - - 0.25 - - - Thymus praecox - - - - - 0.25 4 0.55 Thymus serpyllum - 3.5 - 1.5 - - - - Tilia tomentosa ------0.05 - Tragopogon dubius - 0.05 - - 0.1 0.25 - - Trifolium arvense - - - - 14.5 0.75 0.1 - Trifolium campestre - - - - 2.25 0.25 6.5 - Valerianella coronata - 0.05 - - - 0.3 - - Valerianella sp. ------0.05 Verbascum lychnitis 2.5 - - 0.3 0.05 - - - Veronica hederifolia ------0.1 0.1 Veronica persica - 0.05 ------Veronica praecox - - - - 0.05 0.1 - - Vicia angustifolia 0.1 0.8 0.25 0.5 0.3 0.05 - - Vicia hirsuta ------0.1 Vicia lathyroides ------0.05 - Viola arvensis 0.1 0.1 0.05 0.3 0.3 0.1 0.05 0.1

continuedSupplementary Table 4. Species cover data of the 2 m² north-facing rock sward plots.

656 L. Erdős et al.

Transect serial number 1 2 3 4 5 6 7 8

Acinos arvensis 0.1 0.1 0.05 0.1 0.1 0.3 - 0.3 Ajuga genevensis 0.55 ------1.25 Allium flavum - 0.1 0.05 - 0.05 0.05 0.1 0.25 Allium sphaerocephalon 0.5 0.75 0.3 1.5 - - - - Alyssum alyssoides 0.05 0.25 0.75 0.3 1.75 0.05 - - Anthericum ramosum ------8 - Arabidopsis thaliana - 0.05 0.1 0.1 0.1 0.1 - 0.05 Arabis auriculata - - - - - 0.1 - 0.05 Arabis glabra ------0.25 - Arenaria serpyllifolia 0.1 0.1 0.8 0.75 0.1 0.1 - 2.55 Arrhenatherum elatius ------1.75 - Artemisia alba 0.25 ------Asperula cynanchica ------0.05 - Asplenium javorkeanum 0.5 0.25 0.3 2.5 - 0.05 - - Botriochloa ischaemum 0.05 - - 7.25 19 55 1.25 1.75 Bromus squarrosus - 0.05 0.05 - 0.05 0.1 - - Buglossoides arvensis 1.5 0.05 - 0.3 - - - - Calepina irregularis 3.25 5 - - 0.05 - - - Campanula bononiensis ------1.05 - Campanula rapunculus ------0.05 - Carex michelii ------1 - Centaurea bibersteinii 0.1 0.3 - - 12.5 0.1 - - Cerastium brachypetalum 0.3 0.3 0.1 1 1.75 5 0.1 0.3 Chrysopogon gryllus 3 15.5 6 - - 0.25 - 2.75 Cleistogenes serotina 2.5 2.75 - 0.25 8.5 0.05 - - Colchicum hungaricum - - - - 1 0.05 - - Convolvulus cantabrica 6.5 - 0.75 3.5 2.5 2.25 - 1 Cruciata pedemontana - - - - 0.1 0.1 - - Crupina vulgaris - - - - 0.5 - - - Dianthus giganteiformis 1.5 2 0.75 2 - - - - Dianthus pontederae ------0.5 - Dictamnus albus - - - - 2 - - 0.05 Echium vulgare ------5 Elymus hispidus 2.5 - 1 - 3.25 - 0.5 0.55 Erodium cicutarium 1 ------Euphorbia cyparissias 0.3 1.25 0.1 0.55 0.1 4.05 3.5 0.1 Euphorbia helioscopia - - - - - 0.5 - - Fallopia convolvulus 0.05 - - - 0.05 0.05 - 0.55 Festuca dalmatica 2.75 0.3 7.5 1.5 0.5 - - 9.5 Festuca rupicola ------11.5 - Fragaria vesca ------0.8 - Fragaria viridis ------0.5 Fraxinus ornus - 1.5 - - - - 3 1 Fumaria officinalis 0.05 0.05 - 0.05 0.1 0.1 - - Galium aparine 0.05 - - - - 0.05 0.1 - Galium lucidum ------0.05 1.25 Geranium columbinum - - - - 0.1 0.8 - - Geranium rotundifolium 0.05 0.1 0.05 0.1 - 0.8 - - Geranium sanguineum ------0.1 Helianthemum canum - 0.05 ------Helianthemum ovatum 3 4 2.5 3 - 1.3 1 0.75 Helleborus odorus ------0.05 -

Supplementary Table 5. Species cover data of the 2 m² south-facing rock sward plots.

657 Properties of shrubforest edges: a case study from South Hungary

Transect serial number 1 2 3 4 5 6 7 8

Hesperis tristis - - - - 0.1 - - - Hieracium bauhini ------0.3 Hypericum perforatum 0.25 - 0.3 0.05 0.1 0.1 - 0.05 Koeleria cristata ------0.55 2.5 Lactuca serriola - - - - - 0.05 - - Lactuca viminea ------0.05 Lamium amplexicaule 0.1 0.05 0.3 0.1 0.1 0.1 - - Lamium purpureum - 0.05 ------Lathyrus sphaericus 0.05 - 0.1 - 0.75 - - - Linaria genistifolia - - 0.05 0.05 0.05 - 0.05 - Medicago minima 3 3.75 6 1.75 0.55 0.05 - 0.05 Melica ciliata 0.75 2.75 1.05 - 1.55 1.5 0.1 0.55 Melica uniflora ------0.75 Muscari racmosum ------0.05 Myosotis stricta 0.1 0.5 0.5 0.3 1.05 1 - - Orlaya grandiflora 6.5 9.5 5.75 14.5 7.5 19 16.5 2 Orobanche picridis - - - - - 0.05 - 0.3 Papaver confine - - - - 0.05 - - - Pisum elatius - - 0.75 - 0.05 - - - Poa pratensis agg. ------0.5 0.05 Potentilla arenaria - 0.5 3 1.05 - - 1.25 13 Ranunculus illyricus ------0.05 0.05 Sanguisorba minor 0.25 2.25 1 1 2.25 1.25 2.75 1 Saxifraga tridactylites 0.1 0.5 0.3 0.1 0.05 1 - 0.25 Securigera varia - - - 0.05 - - - - Sedum acre - - - 0.05 - 1.5 - - Sedum sexangulare ------0.05 5 Sedum telephium ------0.05 - Sideritis montana ------0.25 Silene otites - 0.05 0.5 0.05 - - - - Sisymbrium orientale - - - - - 1.05 - - Stachys recta 0.5 - 2 0.25 - - 0.55 2 Stipa pulcherrima - 1.25 - 2.25 - - - - Taraxacum erythrospermum ------0.25 0.05 Teucrium chamaedrys 8.25 - - 2.25 - - 6 1.3 Thlaspi perfoliatum - - - 0.05 0.1 0.05 0.1 0.1 Thymus praecox - - - - 0.25 0.5 5.75 0.1 Thymus serpyllum 0.05 - - 0.05 - - - - Tilia tomentosa ------0.05 - Tragopogon dubius 0.1 - - - - 0.05 - 0.25 Trifolium arvense - - - - 0.5 - - - Trifolium campestre - - - - 0.05 - 0.05 - Valerianella coronata 0.05 - 0.25 - 0.3 0.3 - - Verbascum lychnitis - - 0.55 - 0.05 1.5 - - Veronica hederifolia - - - - 0.05 - - 0.05 Veronica persica 0.1 0.05 ------Veronica polita - 0.05 ------Veronica praecox - - - 0.05 0.1 0.05 - - Vicia angustifolia 0.3 - - - - 0.1 0.05 0.1 Vicia hirsuta ------0.05 - Vincetoxicum officinale ------2.75 - Viola arvensis 0.1 0.3 0.1 0.3 0.1 0.1 0.05 0.05 Viola hirta ------0.05 -

continuedSupplementary Table 5. Species cover data of the 2 m² south-facing rock sward plots.

658