COMMUNITY ECOLOGY 10(1): 75-80, 2009 1585-8553/$20.00 © Akadémiai Kiadó, Budapest DOI: 10.1556/ComEc.10.2009.1.9

The influence of habitat heterogeneity on the fine-scale pattern of an assemblage in a sand grassland

A. Torma1,2 and L. Körmöczi1

Department of Ecology, University of Szeged, Szeged, Közép fasor 52, H-6725, Hungary

Corresponding author: E-mail: [email protected]

Keywords: Epigeic true bugs, Habitat transition, Sand dune. Abstract: The influence of elevation and vegetation characteristics on the spatial pattern of an epigeic true bug assemblage was investigated along a transect in a sandy grassland of Kiskunság. A 55 m long transect through wind grooves and dune tops, perpendicular to the vegetation borders was established. Both the moving split window technique and the ordination method revealed that dune top habitat has a distinct Heteroptera assemblage. This sand dune habitat was characterized by the most abundant Heteroptera species. We did not observe a distinct true bug assemblage in the wind groove habitat. Canonical corre- spondence analysis and multiple linear regressions showed that the relative altitude had a greater effect on the distribution of true bugs than vegetation cover and plant species richness.

Nomenclature follows: Kondorossy (1999).

Abbreviations: CCA–Canonical Correspondence Analysis, MSW–Moving Split Window.

Introduction Tóthmérész 1997, 1998, Magura 2002, Horváth et al. 2002, Maelfait and De Keer 1990, Martin and Major 2001, Guido Habitat heterogeneity of grasslands derives from and Gianelle 2001, Bieringer and Zulka 2003), which help to changes of several environmental factors (microrelief, expo- understand the properties of transitions and patterns in ac- sure, soil moisture, soil organic carbon content). Human land cordance with the environmental variables. use practices (grazing, moving, fertilizing, burning) also in- This study was carried out in the Kiskunság region which fluence the heterogeneity of grasslands (for review, see Ben- is situated in the middle part of the Hungarian Great Plain. ton et al. 2003). Heterogeneous habitats consist of more or The mosaic-like landscape structure of the region consists less distinct habitat patches that are separated by boundaries. mainly of agricultural fields and forest plantations, but some If a boundary is sharp, then a well-defined zone between the patches of the original natural habitats also occur. One of the patches can be considered as ecotone with an abrupt change latter is Bócsa-Bugac sand dune area (10920 ha) belonging in the community structure (van der Maarel 1990). The tran- to the Kiskunság National Park. The Bócsa-Bugac region is sition between the stands, however, may be gradual without a typical mosaic-like complex with wind grooves and dunes sharp boundary in accordance with the gradual change of shaped by the wind. These habitats differ in microclimate background factors. In this case, it is difficult to determine conditions, vegetation and invertebrate communities (Kör- the type and characteristics of such a transition, but it would möczi et al. 1981, Györffy and Pollák 1983, Harmat 1984, be necessary to consider the structural properties and behav- Gallé et al. 1987, 1988). iour of the boundaries at a fine scale (Zalatnai et al. 2004, 2008). The Heteroptera fauna of Kiskunság is relatively well- known, numerous rare and unique species occur here due to In general, the spatial pattern of vegetation is the basis of the xeric and unique habitats (Bakonyi and Vásárhelyi 1987). distinguishing habitat patches. It has often been assumed that True bugs are an ecologically very diverse group, including the assemblages are influenced by vegetation saprophagous, zoophagous and mainly phytophagous spe- (Dennis et al. 1998). Several studies reveal the relationship cies (Dolling 1991). The richness of the Heteroptera fauna between the characteristics of arthropod assemblages and correlates strongly with total diversity (Duelli and vegetation properties; mainly the richness and composition Obrist 1998). True bugs have an important role in grassland of plant species and vegetation physiognomy are important ecosystems mainly as a primary consumer (Knowlton 1967) in this respect (e.g. Murdock et al. 1972, Dennis et al. 1998). and they are suitable ecological objects to study the effect of Other studies emphasize the impact of other background fac- grassland patchiness. tors, e.g. soil conditions or management practices (Rushton et al. 1991, Sanderson et al. 1995). Recently, several papers The aims of this study were a) to reveal the fine-scale have been published about arthropod communities related to pattern of epigeic Heteroptera assemblages; b) to detect if edge effect or grassland managements (e.g. Magura and there are any boundaries indicated by true bugs; c) to seek 76 Torma and Körmöczi correspondence between vegetation and Heteroptera assem- pitfall trapping was a suitable method for ecological investi- blage patterns in a heterogeneous sandy grassland habitat. gations in open sand grasslands. Only the adult individuals were identified at the species level. Material and methods Data analysis

Study site The parallel transects of pitfall traps were pooled for the analyses. We used the moving split window technique Investigations were carried out in a sand dune area of the (MSW) with squared Euclidean distance and the comple- Kiskunság National Park nearby Bugacpusztaháza village. ment of Renkonen similarity index as dissimilarity functions The study site was an approximately 2.4 ha area and it was to detect the discontinuities in Heteroptera assemblages and earlier a sandy pasture. The site was fenced in the 1970-s and vegetation along the transect. This is a powerful method for no management (grazing or cutting) has been applied since detecting boundaries in the case of both vegetation (Whit- then. taker 1956, van der Maarel 1976, Körmöczi 2005) and ani- The relief is characterized by sand dunes and wind mal assemblages (Magura 2002, Gallé et al. 2007). Both grooves with an average elevation difference of approxi- quantitative and presence/absence data were applied in the mately 2 m. The elevation difference between the highest and boundary detection. lowest points of the study area is 2.8 m. The soil is loose sand The significance of the peaks was tested with the Z-score on the dunes and humous sand in the grooves. Plant associa- transformation of the squared Euclidean distance values tions of the higher sand hills are calciphilous fescue steppe (Cornelius and Reynolds 1991, Hennenberg et al. 2005). (Festucetum vaginatae) and Pannonian sand puszta (Poten- tillo-Festucetum pseudovinae), and brome sward (Brometum Random reference was made with the Monte Carlo tectorum) in the very dry patches. The 1.5-2 m deep wind method: the population patterns were randomly shifted com- grooves are covered with purple moorgrass meadow pared to each other (Horváth 1998), and 1000 randomiza- (Molino-Salicetum rosmarinifoliae, Körmöczi 1983). By tions were performed. The differential profiles were drawn now , the Molinia swards are in a transitional state towards from Z-scores averaged over 1 to 20 half-window sizes for the dry grasslands due to the drop of the water table (Kör- the vegetation and over 1 to 5 half-window sizes for the true möczi et al. 2000). bugs. For further analysis, the vegetation samples were pooled in 0.5 m quadrats (N = 55) for comparison with Het- Sampling methods eroptera samples. The Mantel test with Bray-Curtis similarity was applied To investigate the habitat heterogeneity, we established to reveal the correlation between the plant and Heteroptera a 55 m long, SW-NE running transect through wind grooves assemblages (Leps et al. 2001). We calculated the percentage and dune-tops perpendicular to the supposed vegetation bor- cover of plant species and plant species richness per quadrat ders. We recorded the relative elevation along the transect. as well as the relative elevation to use them as explanatory Percentage cover of plant species was recorded in 0.25×0.25 variables. We used the same explanatory variables for further m continously placed quadrats (N = 220) along the transect analysis. in May 2003. We used pitfall traps (55 mm in diameter) with ethylene-glycol as a killing-preserve solution for sampling Canonical correspondence analysis (CCA) was used to epigeic true bugs. 110 pitfall traps were applied in two paral- explain bug species variations in samples by the variations of lel transects. Distance between the traps was 1 m and they the ecological parameters (Ter Braak 1986). Significance for worked for two weeks three times in the year 2003(May 20 explanatory variables was obtained by the permutation test – June 3, July 29 – August 8, and September 4–18). Although available in R vegan package (Oksanen et al. 2006). Number a combined sweep net and D-vac sample is an adequate tech- of permutations was 500. Two CCA procedures were carried nique to collect true bugs according to Standen (2000), we out. The first one was carried out with the whole data set, used pitfall traps for several reasons. whereas in the second procedure the species under 0.5 % fre- quency were omitted. In spite of the fact that the explained Sweep netting is not a suitable sampling method in the inertia was double in the second procedure, the significant case of such a fine scale and it samples only the plant-dwell- factors and their significance levels were the same. Hence, ing . However, in the sand dune area of the we omitted no species from analyses, and present only the Kiskunság, epigeic true bugs are more abundant (Harmat results of the first procedure. Moreover, rare species may be 1984, Kerényi 1997); moreover, Spungis (2005) found a good indicators for ecological quality of habitat (Lyons et al. higher species richness of epigeic true bugs then plant-dwell- 1995) or species richness (Chase et al. 2000), therefore ex- ing true bugs at a coastal dune area. Suction sampling collects clusion of these species may result in losing valuable infor- from both the vegetation and ground surface (Samu mation (Cao et al. 2001). and Sárospataki 1995), but it needs a great effort to sort the material collected from open sandy grassland because of the To reveal the relationship between the bug assemblages great amount of sand picked up during the sampling process. and habitat heteromorphy we used multiple regression analy- Finally, both Harmat (1984) and Kerényi (1997) found that sis. The structure of Heteroptera assemblages was described Fine scale pattern of true bugs 77 by the species richness and abundance given as the average lous species were situated mainly in the grooves forming rec- number of species per trap and the total number of individu- ognizable vegetation patches. The boundaries detected along als per trap, respectively. About 75% of the total number of the transect were rather wide transitional zones with the av- individuals belonged to the three most abundant species erage width of 3m. Midpoints of the boundaries revealed by therefore we computed further regressions to reveal the im- the MSW-analysis were at 12 m, 20.5 m and 43.5 m. pact of factors on these species. The number of individuals A total of 911 adult individuals belonging to 41 Heterop- of these species was transformed by square root transforma- tera species were collected. Eleven species occurred with tion, which resulted the best fit of linear regression. only one specimen in the traps and about 75 % of the indi- Sperman’s coefficient of rank correlation was used to an- viduals belonged to the most abundant species, Emblethis alyse correlations between the distribution of the mono- ciliatus (49.40%), Pionosomus opacellus (17.12%) and Pi- phagous species Piesma kochiae and its host plant. esma kochiae (6.37%). Analyses were carried out with R (R Development Core With the MSW analysis of the quantitative data, we Team 2007) and SPSS 11.5 for WINDOWS software pack- detected two significant peaks (Fig. 1c) that delineated the ages. Heteroptera assemblage of the dune-top habitat. This was the place of the most abundant occurrence of the dominant spe- Results cies. Significant Z-score peaks at 37 m and 40 m using quan- titative data emphasize the large abundance of the dominant The vegetation of the transect consisted of 52 species, 19 Heteroptera species. Analysis of binary data, however, of which were with frequency of over 10%. The most fre- showed peaks at 12, 17, 22 and 27 m, the first two of which quent species were the xerotolerant grasses and cryptogam were significant with Renkonen index (Fig. 1a), and the sec- species on higher reliefs, but longleaf (Falcaria vulgaris), ond two with squared Euclidean distance (Fig. 1b), indicat- which was the most frequent species in the transect, occurred ing considerable differences in species composition and spe- also in the grooves. Longleaf indicates a certain degraded cies frequency, respectively (Körmöczi 2005). Full state of the grassland due to the long dry period. Most of the coincidence of the Z-score peaks of true bugs and vegetation species showed a strongly aggregated spatial distribution, was observed only at one point (12 m) where both the vege- thus the arrangement of the dominant species marked the tation and the Heteroptera assemblage showed a sharp characteristic segments of the transect, which are identified change. as community types of the sandy grassland (Zalatnai et al. 2008). The Mantel test showed no significant correlation be- tween the Heteroptera and plant assemblages (r = 0.183; Based on the composition of vegetation, four patches p=0.0589). were separated by a narrower peak (at 12 m) and two wider transitional zones (at 19-21 m and 42-45 m; see Fig. 1d). This The aim of the correspondence analysis was to reveal the pattern corresponded to the relief profile, since the mesophy- variation in Heteroptera samples induced by vegetation cover, plant richness and relative altitude. The ordination re-

Figure 1. Profile diagram of Z-score values for the Heteroptera assemblage (Renkonen index (a) and binary (b) and quantitative data (c) analysed with squared Euclidean distance) averaged over 1-5 half-window width and for the vegetation averaged over 1-20 half window width (d). Horizontal broken lines indicate 5% significance levels for the separate analyses. Vertical solid lines mark peaks significant on the profile diagrams (a), (b) and (c). Profile diagrams are shifted compared to each other with their own significance levels, therefore the vertical scale is not relevant. The profile of relative elevation is marked with a thick broken line. 78 Torma and Körmöczi

Table 1. Partial CCA revealing the impact of plant species rich- sults showed that the samples that were located at dune top ness, total cover of vegetation and relative elevation on true bug banded together, while the samples of lower relief were scat- assamblage. Significance of parameters was tested by a Monte- tered and overlapped each other (Fig. 2). The three con- Carlo procedure with 500 permutations. strained axes represented 9.11% of the total inertia. The first axis corresponded to elevation (r = -0.9446), and the second axis corresponded to plant species richness (r = -0.8042). Of the three explanatory variables, relative elevation and plant species richness contributed significantly to the distribution of true bugs (Table 1). The constrained axes of CCA ex- plained a low amount of the total inertia, however Oksanen Table 2. Multiple linear regression analysis revealing the rela- (2008) suggested to concentrate on the whole result instead tion between species richness and the abundance of true bugs of the proportions of inertia, because the total inertia does not and habitat properties, as well as between the abundance of dominant species and habitat properties. (Abbreviations: cov: have a clear meaning in CCA and most of the total inertia total cover of vegetation; plants: plant species richness; rel: rela- may be random noise. tive elevation) The total regression models were significant in each case (Table 2). According to the partial correlation coefficients, elevation had a significant effect on the structure of Heterop- tera assemblages and on the abundance of dominant species along the transect and the plant species richness had negative effect on the abundance of P. opacellus (Table 2). A signifi- cant correlation (r = 0.411; p = 0.002) was found between the distribution of P. kochiae and its host plant Kochia laniflora.

Discussion

Multivariate data on Heteroptera assemblage were col- lected in a transect parallel with vegetation sampling to re- veal their spatial pattern and the patchiness of habitats. The results suggest that MSW analysis is a successful method to detect boundaries on such a fine scale and shows similar re- sults to ordination. Both methods reveal that the dune top habitat has a distinct Heteroptera assemblage. Beyond the top, at the southern slope, there is a transitional zone charac- terised by the dissolving of the dune top assemblage. MSW analysis, however, showed significant changes of species composition also at the margins of wind groove, but we did not find a distinct Heteroptera assemblage in the wind groove habitat, contrary to Harmat (1984). We did not find a consis- tent species composition of the samples situated at the wind groove; the samples varied strongly. Moreover, the abun- dance and species richness were rather low in these samples. According to Harmat (1984), five abundant species were as- sociated with wind groove habitats from which, Rhyparo- chromus pini and Xanthochilus quadratus also occurred in our sample, but their abundance was rather low. We did not observe Dimorphopterus spinolai, which was the dominant species in wind grooves, according to both Harmat (1984) and Kerényi (1997). Thus, the wind groove was not charac- terized by a distinct bug assemblage associated with this habitat type. Most bug species, particularly the dominant ones, were observed in the dune top habitat where the species composition was similar to what was ascertained by Harmat Figure 2. Canonical Correspondence Analysis (CCA) based on (1984). each observed species. On the scatterplot of site scores, the sampling sites belonging to the same relative elevation segment We hypothesized that at the point where a vegetation are marked with the same symbol. The width of a relative eleva- boundary was detected, a similar boundary should be ex- tion segment is 0.5 m. (Abbreviations: cov: total cover of vege- tation; plants: plant species richness; rel: relative elevation). pected in phytophagous insect assemblages, too. Despite the fact that the order Heteroptera includes mainly herbivorous Fine scale pattern of true bugs 79 species which are associated with the vegetation (Zurbrügg References and Frank 2006) and several studies reported the relation be- tween epigeic bugs and vegetation properties, e.g. vegetation Bakonyi, G. and T. Vásárhelyi. 1987. The Heteroptera fauna of the height and density (Bröring and Wiegleb 2005), or Kiskunság National Park. In: S. Mahunka (ed.) The Fauna of the coenological structure (Gallé et al. 2007), we did not find Kiskunság National Park. Vol. II. Akadémiai Kiadó, Budapest. clear accordance in vegetation and Heteroptera boundaries. pp. 85-108. Moreover, the Mantel test showed no correlation between the Benton, T.G., J.A. Vickery. and J.D. Wilson. 2003. Farmland biodi- species composition of the vegetation and Heteroptera as- versity: is habitat heterogeneity the key? Trends in Ecology and semblages. Evolution 18:182-188. Bieringer, G. and K.P. Zulka. 2003. Shading out species richness: Di Giulio et al. (2001) found that the bug fauna is af- edge effect of a pine plantation on the Orthoptera (Tettigoniidae fected more strongly by the texture and the microclimatic and Acrididae) assemblage of an adjacent dry grassland. Biodi- conditions of the grasslands than by plant species richness. ver. Conserv. 12:1481–1495. They suggested that in the case of zoophagous and Bröring, U. and G. Wiegleb. 2005. Soil zoology II: Colonization, polyphagous herbivore species the plant species richness distribution, and abundance of terrestrial Heteroptera in open might play a less important role than the textural diversity of landscapes of former brown coal mining areas. Ecol. Engineer- the habitat. ing 24:135–147. Cao, Y., D.P. Larsen and R. St.-J. Thorne. 2001. Rare species in In spite of the low amount of the explained inertia in multivariate analysis for bio-assessment: some considerations. CCA, we found that elevation, which primarily determines J. N. Am. Benthol. Soc. 20:144-153. the microclimate, had a greater effect than vegetation vari- Chase, M.K., W.B. Kristan, A.J. Lynam, M.V. Price and J.T. Roten- ables (plant species richness or the cover of vegetation) on berry. 2000. Single species as indicators of species richness and true bugs. The regression analysis confirmed the result of composition in California coastal sage scrub birds and small CCA, indicating a strong relationship between species rich- mammals. Conserv. Biol. 14:474-487. ness and abundance of true bugs and elevation as well as be- Cornelius, J.M. and J.F. Reynolds. 1991. On determining the statis- tween the abundance of dominant species and elevation. In tical significance of discontinuities within ordered ecological the case of P. opacellus and E. ciliatus, this relationship may data. Ecology 72:2057-2070. be explained by their habitat choice. Both species preferred Dennis, P., M.R. Young and J. Gordon. 1998. Distribution and abun- the dune top, therefore their distribution related to elevation dance of small and arachnids in relation to structural het- strongly. According to the observations of Harmat (1984), P. erogeneity of grazed, indigenous grassland. Ecol. Entomol. kochiae was not associated with dune top habitat and it was 23:253-264. not an abundant species. The relationship between elevation Di Giulio, M., P.J. Edwards and E. Meister. 2001. Enhancing insect and the abundance of this monophagous species may be ex- diversity in agricultural grasslands: the roles of management and plained by the distribution of its host plant. P. kochiae corre- landscape structure. J.App. Ecol. 38:310-319. lated strongly with its host plant, Kochia laniflora which was Dolling, W.R. 1991. The . Oxford University Press, Ox- observed with high density at the top and the southern slope ford. of the dune. Duelli, P. and M. Obrist. 1998. In search of the best correlates for local organismal biodiversity in cultivated areas. Biodiver. Con- It seems that elevation and microrelief determine the serv. 7:297-309. Heteroptera assemblage structure in a sand dune ecosystem. This relation would be more convincing if there was a dis- Gallé, L., Gy. Györffy, E. Hornung and L. Körmöczi. 1988. Indica- tion of environmental heteromorphy and habitat fragmentation tinct assemblage in the wind groove, however numerous spe- by invertebrate communities in grasslands. In: J. Bohac and V. cies were represented by few individuals or occurred in few Ruzicka (eds.), Proc. Vth Int. Conf. Bioindicatores Deteriorisa- traps and they increased the noise of data (Voigt et al. 2003). tionis Regionis, Inst. Landscape Ecology CAS, Ceské The lack of a distinct bug assemblage is presumably due to Budéjovice. pp. 167-170. the relatively small size of wind grooves as already suggested Gallé, L., Gy. Györffy, L. Körmöczi, G. D-Szõnyi and B. Harmath. by Gallé et al (1987). However, Kerényi (1997) found, that 1987. Különbözõ közösségtípusok élõhely heterogenitás in- the bug assemblage in wind grooves depended more strongly dikációja homokpusztai gyepen. Környezet- és Természet- on climatic conditions than the bug assemblage at dune tops. védelmi Kutatások 6:230-271. A dry period could cause a decline in species richness and Gallé, R., A. Torma and L. Körmöczi. 2007. Epigeic invertebrate abundance in wind grooves. The long-term dry period ob- assemblages (Aranae, Heteroptera) of natural forest edges – Pre- served in the Kiskunság may cause intrinsic changes in as- liminary results. In: K. Tajovský, J. Schlaghamerrský and V. semblage structure of true bugs in this manner, but further Pi l (eds.), Contributions to Soil Zoology in Central Europe II. Èeské Budìjovice. pp. studies are needed to reveal the real impact of the dry period. 47-52. Guido, M. and D. Gianelle. 2001. Distribution patterns of four or- Acknowledgements: The authors would like to thank R. thoptera species in relation to microhabitat heterogenity in an Gallé for his technical help. The present study was supported ecotonal area. Acta Oecol. 22:175-185. by NKFP 6/013/2005 grant and the Regional University Györffy, Gy. and T. Pollák. 1983. Habitat specialization of leafhop- Knowledge Centre for Environment and Nanotechnology, per community living in a sandy soil grassland. Acta Biol. University of Szeged. Szeged. 29:153-158. 80 Torma and Körmöczi

Harmat, B. 1984. Szünbiológiai vizsgálatok homokpusztai gyep Het- Murdock, W.W., F.C. Evans. and C.H. Peterson. 1972. Diversity and eroptera-közösségein. PhD thesis. p. 93. pattern in plants and insects. Ecology 53:819-829. Hennenberg, K.J., D. Goetze, L. Kouamé, B. Orthmann and S. Oksanen, J., R. Kindt, P. Legendre and R.B. O’Hara. 2006. Vegan: Porembski. 2005. Border and ecotone detection by vegetation Community Ecology Package version 1.8-3. http://cran.r-pro- composition along forest-savanna transects in Ivory Coast. J. ject.org. Veg. Sci. 16:301-310. Oksanen, J. 2008. Multivariate analysis of ecological communities Horváth, A. 1998. INFOTHEM program: new possibilities of spatial in R: vegan tutorial. http://cran.r-project.org. series analysis based on information theory methods. Tiscia R Development Core Team 2007. R: A language and environment 31:71-84. for statistical computing. R Foundation for Statistical Comput- Horváth, R., T. Magura, G. Péter and B. Tóthmérész. 2002. Edge ing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R- effect on weevils and spiders. Web Ecol. 3:43-47. project.org. Kerényi, Z. 1997. Poloskaközösség szerkezetváltozásai disztur- Rushton, S.P., M. Luff and M.D. Eyre. 1991. Habitat characteristics bációs hatásokra homokpusztai gyepen. MSc Thesis. p. 39. of grassland Pterostichus species (Coleoptera: Carabidae). Ecol. Entomol. 16:91-104. Knowlton, G. F. 1967. Grass bugs: A serious range problem in 1966. Utah Acad. Sci. Arts and Letters 43: 20-21. Samu, F. and M. Sárospataki. 1995. Design and use of a hand-hold suction sampler, and its comparison with sweep net and pitfall Kondorossy, E. 1999. Checklist of the Hungarian bug fauna (Het- trap sampling. Folia Entomol. Hung. 56:195-203. eroptera). Folia Entomol. Hung. 60:125-152. Sanderson, R.A., S.P. Rushton, A.J. Cherrill and J.P. Byrne. 1995. Körmöczi, L. 1983. Correlations between the zonation of sandy Soil, vegetation and space: an analysis of their effects on the grasslands and the physico-chemical condition of their soil in invertebrate communities of a moorland in north-east England. Bugac. Acta Biol. Szeged. 29:117-127. J. App. Ecol. 32:506-518. Körmöczi, L., L. Gallé, Gy. Györffy and K. Margóczi. 2000. Succes- Standen, V. 2000. The adequacy of collecting techniques for estimat- sional dynamics of sand dune plant and invertebrate communi- ing species richness of grassland invertebrates. J. App. Ecol. ties: the role of stress and disturbances. In: K. Lajtha and K. 37:884-893. Vanderbilt (eds), Coopoeration in Long Term Ecological Re- Spungis, V. 2005. Fauna and ecology of bugs (Hemiptera, Heterop- search in Central and Eastern Europe, Oregon State University, Corvallis, USA. pp. 77-83. tera) in the coastal gray dunes in Latvia. Latvijas Entomologs 42:95-105. Körmöczi, L. 2005. On the sensitivity and significance test of vege- Ter Braak, C.J. 1986. Canonical correspondence analysis: a new tation boundary detection. Community Ecol. 6:75-81. eigenvector technique for multivariate direct gradient analysis. Körmöczi, L., Gy. Bodrogközy and I. Horváth. 1981. Investigation Ecology 67:1167–1179. of biological production and bioclimate of sandy grassland in van der Maarel, E. 1976. On the establishment of plant community Bugac. Acta Biol. Szeged. 27:55-69. boundaries. Berichte der Deutschen Botanischen Gesellschaft Leps, J., V. Novotny and Y. Basset. 2001. Habitat and succesional 89:415-443. status of plants in relation to the communities of their leaf-chew- van der Maarel, E. 1990. Ecotones and ecoclines are different. Jour- ing herbivores in Papua New Guinea. J. Ecol. 89:186-199. nal of Vegetation Science 1:135-138. Lyons, J., S. Navarro-Perez, P.A. Cochran, C.E. Santana and M. Voigt, W., J. Perner, A.J. Davis, T. Eggers, J. Schumacher, R. Bähr- Guzman-Arroyo. 1995. Index of biotic integrity based on fish mann, B. Fabian, W. Heinrich, G. Köhler, D. Lichter, R. Mar- assamblages for the conservation of streams and rivers in west- staller and F. Sander. 2003. Trophic levels are differentially cenral Mexico. Conserv. Biol. 9:569-584. sensitive to climate. Ecology 84:2444-2453. Maelfait, J. P. and R. De Keer. 1990. The border zone of an inten- Whittaker, R. H. 1956. Vegetation of the Great Smoky Mountains. sively grazed pasture as a corridor for spiders (Araneae). Biol. Ecol. Monog. 26:1-80. Conserv. 54:223-238. Zalatnai, M. and L. Körmöczi. 2004. Fine-scale pattern of the bound- Magura, T. and B. Tóthmérész. 1997. Testing edge effect on carabid ary zones in alkaline grassland communities. Community Ecol. assemblages in an oak-hornbeam forest. Acta Zool. Acad. Sci. 5:235-246. Hung. 43:303-312. Zalatnai, M., L. Körmöczi and T. Tóth. 2008. Community bounda- Magura, T. and B. Tóthmérész. 1998. Edge effect on carabids in an ries and edaphic factors in saline-sodic grassland communities oak-hornbeam forest at the Aggtelek National Park (Hungary). along an elevation gradient. Tiscia 36:7-15. Acta Phytopathol. Entomol. Hung. 33:379-387. Zurbrügg, C. and T. Frank. 2006. Factors influencing bug diversity Magura, T. 2002. Carabids and forest edge: spatial pattern and edge (Insecta: Heteroptera) in semi-natural habitats. Biodiver. Con- effect. Forest Ecol. Manage. 157:23-37. serv. 15:275-294. Martin, T.J. and R.E. Major. 2001. Changes in wolf spider (Araneae) Received September 11, 2008 assemblages across woodland–pasture boundaries in the central Revised December 19, 2008 wheat-belt of New South Wales, Australia. Austral. Ecol. Accepted March 4, 2009 26:264-274.