, Revue suisse de Zoologie, 103 (4): 801-833; décembre 1996 Variation in species richness ofplants and diverse groups ofinvertebrates in three calcareous grasslands ofthe Swiss Jura mountains Bruno BAUR1, Jasmin JOSHU, Bernhard SCHMID?, Ambros HÄNGGU, Daniel BORCARD^, JosefSTARY5, Ariane PEDROLI-CHRISTEN6, G. Heinrich THOMMENL Henryk LUKA?, Hans-PeterRUSTERHOLZ*, PeterOGGIER', Stephan LEDERGERBER' & Andreas ERHARDT« 1Institut fürNatur-, Landschafts- und Umweltschutz derUniversität Basel (NLU), St. Johanns-Vorstadt 10, CH-4056 Basel, Switzerland, - Institut fürUmweltwissenschaften, UniversitätZürich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland, 3 Naturhistorisches Museum Basel, Augustinergasse 2, CH-4001 Basel, Switzerland, 4Institut de Zoologie, Université de Neuchâtel, Rue Emile-Argand 1 1 CH-2007 Neuchâtel, Switzerland, 5 Institute ofSoil Biology, Academy ofSciences ofthe Czech Republic, Na sâdkâch 7, CZ-37005 Ceské Budëjovice, Czech Republic, 6 Centre suisse de cartographie de lafaune, Musée d'histoire naturelle, Terreaux 14, CH-2000Neuchâtel, Switzerland, 7 Forschungsinstitut fürbiologischen Landbau, Bernhardsberg, CH-4104 Oberwil, Switzerland, 8 Botanisches Institut derUniversität Basel, Schönbeinstrasse 6, CH-4056 Basel, Switzerland. Variation in species richness of plants and diverse groups of inver- tebrates in three calcareous grasslands of the Swiss Jura mountains. - Species richness and abundance of vascular plants and several groups of invertebrates (spiders, oribatid mites, diplopods, grasshoppers and bush crickets, ground beetles, butterflies and terrestrial gastropods) were re- corded in three calcareous grasslands (Nenzlingen, Movelier and Vieques) in the northwestern Swiss Jura mountains. Species richness varied both between taxonomical groups and between sites (species richness ranges: 96-116 vascular plants, 60-66 spiders, 18 oribatid mites at each site, 1-7 diplopods, 10-16 grasshoppers and bush crickets, 19-21 ground beetles, 32- 46 butterflies and 15-21 terrestrial gastropods). Species overlap (number of species that occurred at all sites) was relatively large in terrestrial gas- tropods (59.1%), butterflies (56.5%), vascular plants (53.8%) and grass- hoppers (47.1%), but relatively low in oribatid mites (32.3%), spiders Manuscriptaccepted 16.01.1996. 802 BRUNO BAURETAL. (25.0%), ground beetles (18.4%) and diplopods (12.5%). Diversity expressed by the Shannon-Wiener index (H') was compared for five groups of invertebrates. Diversity was largest in spiders and ground beetles and lowest in terrestrial gastropods. Different taxonomical groups had their maximum diversity at different sites: each grassland had the highest diversity in at least one group. The three sites also varied in the abundance of different invertebrate groups. Most groups had the highest densities in Nenzlingen and the lowest densities in Vieques. All three sites contained a high proportion of species listed in the Red Data Book of Switzerland with values averaging 49.5% in grasshoppers and bush crickets, 28.9% in butterflies, 18.9% in vascular plants. 11.2% in terrestrial gastropods, and 6.7% in ground beetles. One spider species (Oxyptilapullata) and two mite species {Epilohmannia cylindrica minima and Pergalumna myrmophila) were recorded for the first time in Switzerland. Key-words: Acari - Araneae - Biodiversity - Calcareous grassland - Coleoptera - Diplopoda - Gastropoda - Lepidoptera - Orthoptera - Vascular plants. INTRODUCTION Many human activities, such as intensive agriculture, urbanisation, defores- tation, and the building of roads and railways reduce natural habitats to remnants of different sizes. In addition to the overall decrease in the area available for the orga- nisms, the fragmentation of habitats leads to a division of existing populations into isolated subpopulations of small size and to changes in habitat characteristics in the fragments (Soulé 1986, Saunders et al. 1991). It is generally assumed that habitat fragmentation contributes significantly to the local extinction of animal and plant species (Saunders etal. 1991). In the northwestern Jura mountains, unfertilized calcareous grasslands were large and continuous at the beginning ofthis century, but since the 1950's changes in agricultural practices such as the use of modern machinery, chemical fertilizers, her- bicides and pesticides, and new breeds of plants have reduced the size of these areas and split them into small and isolated fragments. For example, between 1950 and 1985 the total area of unfertilized calcareous grassland was reduced by 78% in the Passwang region in northwestern Switzerland (Zoller et al. 1986). This overall reduction ofthe unfertilized grassland area and the isolation ofthe remnants may have led to adramatic loss ofspecies within a short period. In 1993, a multidisciplinary coordinated project was started to experimentally analyse fundamental aspects of biodiversity using dry grasslands on calcareous soils in the northwestern Swiss Jura mountains as an exemplary model (Leser 1994, Baur & Erhardt 1995). Most ofthe field work is done at three sites (Nenzlingen, Movelier and Vieques). The present paper examines species richness and abundance ofvascular plants and various groups of invertebrates (spiders, oribatid mites, diplopods, grass- hoppers and bush crickets, ground beetles, butterflies and terrestrial gastropods) in SPECIES RICHNESS INCALCAREOUS GRASSLANDS 803 these grasslands. In particular, we address the following questions: (1) Do the three grasslands differ in plant and invertebrate species richness? (2) Are there any associations between the species richnesses of different taxonomical groups? (3) Do different taxonomical groups differ in species overlap at the three sites? (4) Do the three sites differ in the overall density of particular taxonomical groups? and (5) Do the three sites differin numberand proportion ofrare and/orthreatened species? STUDY SITES AND GENERAL METHODS The three field sites are nutrient-poor, dry calcareous grasslands (belonging to the Teucrio-Mesobrometum type; Ellenberg 1988) situated near Nenzlingen (10 km S ofBasel), Movelier (5 km N ofDelémont) and Vieques (5 km E ofDelémont). The three sites are situated within 20 km ofeach other. There are local differences between the three sites. The study site in Nenz- lingen (1.5 ha in size) is situated on a SW-facing slope with an inclination of 19-22° at an altitude of 500 m a.s.l. A deciduous forest borders the investigation area to the NE. Annual precipitation averages 991 mm in Grellingen (3 km E of the study site). Soil properties and soil profiles of the investigation area are presented in Ogermann et al (1994). Until 1993, the site was grazed by cattle from May to September with a high stocking rate. The lower part of the slope was moderately fertilized by cattle dung. The investigation area in Movelier (1.3 ha) is situated on a SSE-facing slope (inclination 20-22°) at an altitude of 780 m. Half of the site is surrounded by deci- duous forest. The humus layer is relatively thick, contains some clay and is moister than in Nenzlingen and Vieques. Until 1993, the site was grazed by cattle and a moderate amount ofartificial fertilizerwas used. The investigation area in Vieques (2.0 ha) is situated on a SE-facing slope (inclination 15-27°) at an altitude of 570 m. The humus layer is thin and there are several patches of exposed bedrock (this type of habitat is lacking at the other sites). There is mixed forest at the SW-border of the area. Until 1993, the site was exposed to a low grazing pressure by cattle. In spring 1993, we initiated a field experiment to investigate possible effects of habitat fragmentation on animal and plant populations at these sites. We experi- mentally fragmented selected areas of grassland into plots of different sizes to exa- mine whether species diversities ofplants and various groups of invertebrates change after fragmentation. The fragmentation was created by mowing the vegetation around the experimental plots (see Baur & Erhardt (1995) for a figure ofthe experimental set-up). An experimental unit (= block) contains one large (4.5 x 4.5 m), one medium (1.5 x 1.5 m) and two small fragments (0.5 x 0.5 m), all of them separated by a 5-m wide strip of mown vegetation, as well as the corresponding control plots. The experiment consists of 12 blocks distributed over the three sites (five blocks in Nenzlingen, three blocks in Movelier and four blocks in Vieques). The fragmentation is maintained by regularly mowing the isolation area. This type of fragmentation is 804 BRUNO BAURETAL. reversible, but reduces dispersal ofpollen and seeds in many plant species as well as dispersal ofseveral invertebrate species. In this paper, we present data on species richness and abundance of vascular plants and various groups of invertebrates at the beginning of the long-term experi- ment. We have tried to minimize any additional disturbances of the fragmented eco- system by using non-destructive methods whenever possible. Data were collected either within the experimental blocks (plants, grasshoppers and terrestrial gastropods), at the border ofeach experimental block (oribatid mites), and/or in the areas adjacent to or surrounding the experimental blocks (spiders, diplopods, grasshoppers, ground beetles, butterflies and terrestrial gastropods). Species determination ofgrasshoppers, butterflies and gastropods occurred in situ. Pitfall trapping was used for other taxo- nomical groups in which species identification is difficult in the field (spiders, diplo- pods, ground beetles). Details on the collecting methods are given in the corres- ponding sections dealing with the different groups oforganisms. DATA ANALYSIS Diversity has two components, the species richness measured by the total number ofspecies in an area, and the equitability of species abundances. The number of species was used as a measure of the relative richness of the grasslands, but it cannot be thought of as the absolute richness of a site, since numerous rare species may not have been recorded in this study. The diversities of different taxonomic groups were compared using the Shannon-Wiener index (H', base e) and the equi- tability or eveness (E) of distribution of individuals among the species (Lloyd & Ghelardi 1964, Krebs 1988). Species-abundance distributions were plotted to compare the proportion ofrare species (species with one individual) among the different taxonomical groups. The similarity ofthe different species assemblages were compared using the coefficient of S0rensen (based on presence/absence data) and the Renkonen index (based on abundance data; Krebs 1988, Mühlenberg 1989). VASCULAR PLANTS JasminJoshi and Bernhard Schmid Methods Plant species richness was recorded in all fragments and control plots during the growing seasons of 1993 and 1994. Data on the presence/absence ofplant species were obtained in Nenzlingen in May, June and August 1993 and in May and August 1994, in Movelier in June and August 1993 and at the beginning of June and in August 1994, and in Vieques at the beginning ofJuly and in August 1993 and in May and August 1994. In August 1994, the recording in Nenzlingen took place after mowing ofthe experimental plots. Therefore, the presence ofsome grass species may have been overlooked. 1 SPECIES RICHNESS INCALCAREOUS GRASSLANDS 805 Nomenclature ofthe vascularplants follows Binz & Heitz (1990). Endangered and threatened species are listed in Landolt (1991). Results and Discussion A total of 143 plant species were found in the three investigated calcareous grasslands (Appendix 1). The highest plant species richness was found in Movelier (116 species), where only 24 plots were investigated. In Nenzlingen we recorded 1 1 plant species (in 40 plots) and in Vieques 96 species (in 32 plots). Seventy-seven of the 143 species (53.8%) were recorded at all three sites (Table 1). Movelier also had the highest proportion of plant species of the total species pool found exclusively at one site (15.4%) compared with Nenzlingen (7.7%) and Vieques (4.9%) (Table 2). Table 1 Species richness and species overlap of diverse taxonomical groups at the three investigation sitesNenzlingen, Movelierand Vieques. Taxonomical Numberofspeciesreco:rdedin Total number Number(%) of group Nenzlingen Movelier Vieques ofspecies speciesthat recorded occurred atall (speciespool) sites Vascularplants 111 116 96 143 77 (53.8) Spiders 60 63 66 108 27 (25.0) Oribatidmites 18 18 18 31 10 (32.3) Millipedes 1 4 7 8 1 (12.5) Grasshoppers 13 16 10 17 8 (47.1) Groundbeetles 19 19 21 38 7 (18.4) Butterflies 32 46 40 46 26 (56.5) Terrestrial gastropods 21 16 15 22 13 (59.1) Table2 Number of species and percentage of the total species pool (in parentheses) that occurred exclusively atone site. Taxonomical group Nenzlingen Movelier Vieques Plants 11 (7.7) 22 (15.4) 7 (4.9) Spiders 14 (13.0) 19 (17.6) 21 (19.4) Oribatid mites 4 (12.9) 7 (22.6) 5 (16.1) Millipedes 1 (12.5) 4 (50.0) Grasshoppers 3 (17.6) Groundbeetles 10 (26.3) 6 (15.8) 9 (23.7) Butterflies 7 (15.2) Terrestrial gastropods 4 (18.2) 1 (4.5) All three sites were dominated by Bromus erectus, a typical grass species of extensively managed calcareous grasslands. Sanguisorba minor, Ranunculus bulbosus and Hieracium pilosella also occurred frequently at each site. Two of the species 806 BRUNO BAURETAL. which were found at all three sites {Trifolium ochroleucon and Veronica prostrata) are listed as vulnerable in the Red Data Book of Switzerland (Landolt 1991). Furthermore, Gentiana cruciata, found exclusively in Vieques, is also considered as vulnerable in Switzerland. Altogether, seven of the 143 species (4.9%) recorded belong to the endangered species, 19 (13.3%) to the vulnerable species and one (0.7%) to the rare plant species in the northern Jura mountains (Landolt 1991). SPIDERS (Arachnida, Araneae) Ambros Hänggi Methods Pitfall traps were used to collect ground-living arthropods (Araneae, Diplo- poda and Coleoptera). The traps were white plastic jars, 7 cm deep x 7 cm in diameter, containing about 50 ml formalin (4%) with detergent. The traps were pro- tected against rain by grey plastic roofs (measuring 18 x 18 cm) that were fixed horizontally about 10 cm above ground. Specimens were removed and formalin replaced at intervals of two weeks from 5 May to 12 November 1994. At each site, nine traps were placed in groups of three (within-group distance 1 m) in three different grassland subtypes. m At each site, the grassland subtypes considered were 50 - 100 apart; they may reflect the heterogeneity of the three grasslands. In Nenzlingen the grassland subtypes were (1) a moderately fertilized patch ofthe pasture, (2) an unfertilized area (typical Mesobrometum), and (3) a patch dominated by bracken (Pteridium aquilinum); in Movelier (1) a moderately fertilized patch of the pasture, (2) an unfertilized area (typical Mesobrometum), and (3) a part of the slope structured by cow paths; in Vieques (1) an unfertilized area of the pasture with exposed bedrock (karst rocks), (2) an unfertilized area (typical Mesobrometum), and (3) a patch ofthe pasture (5 x 10 m) partly covered with bramble {Rubiis spp.). Spiders were determined by Theo Blick, Hummeltal (Germany). Nomenclature follows Maurer & Hänggi (1990). All spiders collected are deposited in the Natur- historisches Museum Basel. Results and Discussion A total of 2701 adult spiders belonging to 108 species were collected at the three study sites (Appendix 2). The sites were similar in species richness (60 species in Nenzlingen, 63 in Movelierand 66 in Vieques), but differed in species composition (Table 1, Appendix 2). Only 27 of the 108 species (25.0%) were found at all three sites (Table 1). Twenty-one (19.4%) of the total 108 species were exclusively found in Vieques, indicating an extraordinary spider community in this grassland (Table 2). The highest similarity was found between the spider communities in Nenzlingen and Movelier, whereas the spider community in Vieques was quite different from those of the other sites (Table 3). Diversity expressed by the Shannon-Wiener index (H') was SPECIES RICHNESS IN CALCAREOUS GRASSLANDS 807 Table3 Comparison of similarity in different groups of organisms at the three field sites. Ne Nenzlingen, Mo=Movelier, and Vi =Vieques. S0renser similarity index Renkoner similarity index i i Ne-Mo Ne-Vi Mo-Vi Ne-Mo Ne-Vi Mo-Vi Vascularplants 1993 0.79 0.82 0.74 _ _ _ 1994 0.79 0.82 0.76 - - - Spiders 0.59 0.59 0.54 0.57 0.35 0.41 Oribatidmites 0.56 0.67 0.50 0.61 0.40 0.64 Millipedes 0.40 0.25 0.55 0.91 0.73 0.76 Grasshoppers 1993 0.82 0.78 0.80 0.61 0.63 0.73 1994 0.91 0.74 0.76 0.60 0.53 0.60 Groundbeetles 0.47 0.40 0.55 0.31 0.21 0.24 Butterflies 1993 0.78 0.90 0.88 - - - 1994 0.79 0.87 0.92 - - - Gastropods 0.81 0.83 0.84 0.84 0.83 0.80 highest in Vieques, whereas Movelier and Nenzlingen showed slightly lower H- values (Table4). Relating the number of individuals caught (Appendix 2) to the number of trapping days (1719 trapping days per site), the activity density of spiders was lower at the three sites (Table 5) than has been reported from similar habitats (Inntal: Thaler 1985; Tessin: Hänggi 1992). The reason for this discrepancy might be that the collection period lasted only six months in the present study (animals were caught during the period ofvegetation growth). Several spider species collected are interesting from a faunistic point of view. Oxyptilapallata was found forthe first time in Switzerland (two males were collected in Vieques; one between 5 and 18 May 1994, the other between 14 September and 1 October 1994). This species is known to occur in extremely dry grasslands and Table4 Shannon-Wiener diversity index (H') and eveness (E; in parentheses) in various invertebrate groups at the three investigation sites. Where the sampling effort was unequal (oribatid mites, grasshoppers and gastropods), the values have been calculated for equal subsamples of three blocks (seeMethods). Taxonomical group Nenzlingen Movelier Vieques Spiders 2.96 (0.72) 3.02 (0.73) 3.18 (0.76) Oribatidmites 2.06 (0.72) 1.58 (0.55) 1.81 (0.64) Grasshoppers (1993) 1.72 (0.81) 1.80 (0.75) 1.73 (0.86) (1994) 1.41 (0.70) 1.77 (0.78) 1.76 (0.85) Groundbeetles 2.50 (0.85) 2.43 (0.83) 2.55 (0.84) Terrestrial gastropods 1.25 (0.51) 1.26 (0.62) 1.46 (0.70) 808 BRUNO BAUR ETAL. Table5 Relativedensities ofvarious invertebrategroups atthethree investigation sites. Taxonomical group Nenzlingen Movelier Vieques Spiders 0.61 0.49 0.48 (individuals/trappingday) Oribatid mites 18700 19'200 5'600 (individuals/m2 ) Grasshoppers (1993) 43.8 66.2 45.2 (individuals/block) (1994) 47.9 59.1 42.8 Groundbeetles 0.09 0.04 0.08 (individuals/trappingday) Terrestrial gastropods 0.83 0.18 0.13 (individuals/trap) habitats with rocks, stones or sand in southern and eastern Europe (Bauchhenss 1990). Talavera (.-Euophrys) inopinata was collected for the second time in Switzer- land (two males in Vieques between 15 and 29 June 1994 and one female in Movelier between 17 and 24 July 1994). This species was described by Wunderlich (1993) based on specimens collected in similar habitats in the region of the present study (Jura mountains between Delémont and Porrentruy; see Hänggi 1993). A further taxon, Pardosa sp. (not listed in Maurer & Hänggi 1990), has not yet been described, but may belong to the species group of Pardosa lugubris s.l. (Töpfer-Hofmann & von Helversen (1990) named this taxon Pardosa saltans (nomen nudum!)). Most probably this undescribed taxon occurs more frequently in Switzerland than P. lugubris s.str., as indicated by museum specimens (most ofthem wrongly determined as P. lugubris). Pardosa sp. was found in small numbers at all three sites. Six further species can be regarded as rare (recorded fewer than five times in Switzerland; cf. Maurer & Hänggi 1990): Bathyphantesparvulus, Centromerita con- cinna, Tegenaria pietà, Eperigone trilobata, Mioxena blanda and Dipoena prona. A single specimen of each of the first three species was collected in this study. Infor- mation on the ecology of these species is summarized in Hänggi et al. (1995). Eperigone trilobata is widespread in northern America and was most probably intro- duced to Europe some 20 years ago (Dumpert & Platen 1985). Eperigone trilobata can now frequently be found in dry, unfertilized grasslands in central Europe (Hänggi 1990, Renner 1992). Dipoenaprona (three males collected in Vieques between 24July and 5 August 1994) is a widespread species, but is rarely found in Europe (Miller 1967). Mioxena blanda (one male caught in Nenzlingen between 13 and 27 October 1994) is widely distributed but rare in northern and western Europe (Roberts 1985). No Red Data List has been compiled for spiders in Switzerland. However, 27 of the 108 (25.0%) species found in the present study have narrow niches (ste- noecious; Maurer & Hänggi 1990). Most ofthese spiders seem to be thermophilous, SPECIES RICHNESS IN CALCAREOUS GRASSLANDS 809 preferring dry, open habitats. These types of habitat have dramatically declined in Switzerland. Consequently, these 27 species can be regarded as potentially endan- gered mainly due to habitat destruction. Vieques harbours a particularly threatened spider fauna since 11 ofthe 21 species recorded exclusively at this site belong to the group ofpotentially endangered species. ORIBATID MITES (Acari, Oribatei) Daniel Borcard &JosefStary Methods Cylindrical soil samples, 5 cm in diameter and 8 cm in depth, were taken at the margins of the control areas of each experimental block, using the soil augers described by Bieri etal. (1978). The samples were divided into two parts (0-4 cm and 4-8 cm). The soil fauna was extracted by means of a MacFadyen-derived extractor (Borcard 1986). Each site was sampled three times, on 10 March, 19 July and 14 November 1994. Four samples were taken in each block on each occasion. In all, 144 samples weretaken. Results and Discussion A total of 4079 adult oribatid mites belonging to 31 different species were captured at the three sites (Appendix 3). At each site, 18 mite species were found. However, species number is often influenced by sampling effort. To account for the lower number of blocks sampled in Movelier (n = 3), we computed the expected number of species at the two other sites, assuming that only three blocks had been sampled. The average ofthe total number of species ofeach possible combination of three blocks was 17.5 in Nenzlingen and 17.0inVieques. Eight of the 31 mite species (25.8%) were found at all three sites (Table 1, Appendix 3). In contrast, four species (12.9%) were exclusively found in Nenzlingen, seven species (22.6%) exclusively in Movelier and five species (16.1%) exclusively inVieques (Table 2). The differences in species composition between seasons were small, accoun- ting for 2.7% of the total variation as indicated by a canonical correspondence ana- lysis (CCA; ter Braak 1986). Consequently, data from the three sampling occasions were pooled for each site for further analysis. Moreover, unless noted otherwise, all mites collected (depth 0-8 cm) were considered, although the majority of the mites were found in the upper soil layer (between and 4 cm depth; Nenzlingen 84.2%, Movelier71.8% andVieques 65.6%). The among-site differences in species composition were significant as indi- cated by a CCA including all 144 samples (Monte Carlo test, 999 permutations, p = 0.001). These differences accounted for 11.4% ofthe total variation. In this analysis, Vieques represented the most heterogeneous sampling site. Overall the species richness was low when compared with otherbiotopes (such as forests), but normal for 810 BRUNOBAURETAL. open grassland (Bachelier 1978). In comparison with the two other sites, Vieques again represented an extraordinary site: the local richness (as measured by the average number of species per sample) was very low (about halfthat ofNenzlingen), but this was compensated by the overall heterogeneity ofthe site. With estimated values of slightly below 20'000 individuals/m2 the overall . densities of the mite communities in Nenzlingen (18700 individuals/m2) and Movelier (19'200 individuals/m-) corresponded well to those described in the literature for similar habitats (for a review see Bachelier 1978). In contrast, Vieques had an extremely low mite density (5'600 individuals/m2). This can be partly explained by the dryness as well as by the scarce vegetation and extremely shallow soil atthis site. Two species, Epilohmannia cylindrica minima (in Movelier) and Pergalumna myrmophila (in Vieques), wererecorded forthe first time in Switzerland. Sofar, these species have been considered as specialists that occur in the Mediterranean subregion and Pannonian district. With the exception of Minuntiiozetes semirufus, all species captured at only one of the three sites had low densities. The most important components of the investigated oribatid mite communities include semicosmopolitan species (nine species, i.e. 29% of the total species number) and holarctic species (10 species; 32%). Most of these species are eurytopic with a large ecological amplitude (e.g. OppieUa nova, Scheloribates laevigatas, Xylobates capucinus and Tectocepheus sarekensis), and are usually dominant in early stages ofsecondary succession or indi- cate non-stabilized biotopes affected by human activity. Ten species (32%) are found only in Europe. MILLIPEDES (Diplopoda, Myriapoda) Ariane Pedroli-Christen Methods Pitfall traps were used to collect millipedes as described in the spider section (see above). Nomenclature follows Pedroli-Christen (1993). Results and Discussion A total of 154 adult diplopods belonging to eight species were collected (Appendix 4). Seven species were found in Vieques, four in Movelier and one in Nenzlingen. Cylindroiulus caeruleocinctus dominated the diplopod communities at all three sites (100% in Nenzlingen, 91% in Movelier and 73% in Vieques). This species is characteristic for open and exposed habitats such as calcareous grasslands. It can occasionally occur along forest edges, but never in the forest. The remaining species (Glomeris spp. and Tachypodoiulus niger) are associated with woody habitats, butcan also occur in stands of bramble (in Vieques). None of the diplopod species recorded in the present study is rare.