Mapping patterns of ferns species richness through the use of herbarium data

This paper aims to analyse the spatial patterns of sampling effort and species richness of pteridophyte in a well-investigated region as Tuscany, Italy, by using data stored from a geodatabase storing information on the specimens preserved in the main herbaria of the region. A total of 6,905 records about pteridophyte specimens were extracted from the geodatabase, and 5,638 of such specimens were studied through the use of spatial statistical techniques. The data about the sampling effort and species richness were analysed in relation to topographical variables to assess any significant relationship. Specimen-based rarefaction techniques were used to compare areas with different number of detected species. The analysis of the sampling effort data showed a nonhomogeneous distribution of herbarium data, with some areas being intensively sampled and others being almost unsampled. Thus, the geographical distribution of specimens was extremely clustered. The comparison across geographical areas through specimen-based rarefaction curves showed great differences in species richness and sampling completeness. The analysis of the residuals of species–area relationships evidenced that the distance to water bodies was the only significant topographical variable in controlling species diversity.     

A numerical analysis of the mesoscale distribution patterns of vascular plants in the subarctic Kevo Nature Reserve, northern Finland

Different numerical techniques were used to detect and describe the major ecological-biogeographical patterns of vascular plant distributions at the meso-scale level in a subarctic region in Finland. The distribution patterns of 231 native taxa in 362 1 km² grid squares of the Kevo Nature Reserve were analysed by two-way indicator species analysis and detrended correspondence analysis, and were subsequently related to twenty-eight geographical, topographical, geological, and vegetational variables using simple discriminant functions and canonical correspondence analysis with associated Monte Carlo permutation tests.
The floristic variation detected reflects variations in environmental factors operative at the regional and local scales. No major broad-scale coherent geographical patterns were detected; instead, the spatial distribution of the grids with a similar floristic composition shows a scattered distribution. All the numerical techniques reveal a major gradient from alpine areas to lowland sites with rivers and rocky outcrops, and the most important explanatory variables for predicting the main floristic variation are all associated with altitude. The floristic patterns represented by the second ordination gradient mainly correlate with the abundance of mires. Partial ordinations indicate that both the geographical and geological variables explain relatively little of the species distributional patterns.
Although the meso-scale approach reveals much about the plant-environment relationships in the study area, the floristic variation appears to be determined mainly by fine-scale factors. In the most heterogeneous grids, the grid size used fails to detect accurately the ecological patterns of the species present.     

Relative importance of abiotic and land use factors in explaining variation in woody vegetation in a French rural landscape

Abstract. In a rural landscape, scale vegetation patterns of woody species are controlled by both abiotic and land use factors. The woody species composition in 126 sample plots was analysed and land use factors and some abiotic parameters were quantified using land register data. The relative importance of land use and abiotic factors was differentiated using a partial Canonical Correspondence Analysis (CCA); the influence of land use in neighbouring areas was explored by changing the scale of land use sampling. The woody species composition appears to be controlled equally by land use and abiotic factors. The fraction of floristic variation in tree species composition explained by land use variables was 33.2 % in 1980 and 30.8 % in 1992, while abiotic variables accounted for 31.2 %. Part of the 17 % of the floristic variance explained is related to the surrounding land use. Thus, when the land use of the sampled plots and the surrounding land use are considered simultaneously, up to 36.9 % of the species variation may be explained. Partial CCA enabled us to quantify the respective proportion of floristic variance which could be explained by land use (36.9 %), abiotic variables alone (20.2 %), shared variance (12.0 %) and unexplained variance (31.2 %). Our results indicated that a delayed effect of variation in land use on plant populations may exist. This delay may result either from population characteristics or from inadequate land use assessment. This study indicates the need for simultaneously examining land use and abiotic patterns in ecological studies, as many Mediterranean-type ecosystems have been shaped by these patterns.     

The influence of environmental factors on the spatial distribution of saxicolous lichens in a Norwegian coastal community

Abstract. The effects of vegetation cover, radiation, micro‐habitat variables and maritime influence on the floristic composition of a saxicolous community in Vingen, western Norway were studied. Particular emphasis is put on the local distribution of Fuscidea cyathoides, Ochrolechia tartarea, Ophioparma ventosa and Pertusaria corallina. Very little of the variation in the lichen community composition is directly related to measured micro‐environmental variables but variance partitioning shows that vegetation cover explains more of the floristic variation than radiation, maritime influence and microhabitat variables. Logistic regression analyses nevertheless indicate that the micro‐environment influences the spatial distribution of the four species. The high fraction of unexplained floristic variation, 91%, is suggested to result from (1) lack of fit of data to the response model; (2) some influential environmental variables that have not been recorded; (3) local historical factors that affect present day distribution and/or (4) apparent randomness in colonization. The results also agree with the view that the four lichen species in this study are able to co‐exist in the long‐term because of different spatial distributions resulting from different strategies with respect to ecology, dispersion and interaction.     

The influence of past land-use on understory plant distributions in a near-natural deciduous forest in Denmark

In regions with a long cultural history, past land-use may strongly determine floristic composition in seemingly natural vegetation. Here we examine if past land-use affects understory plant distribution in near-natural forest in a historically and environmentally heterogeneous setting, namely the 17 ha freshwater island of Borgnr (Denmark). In 107 randomly located 100 m² plots we recorded species composition and various environmental variables. The plots were classified into three groups according to past land-use: (1) Disturbed sites close to 12^th and 20^th century ruins and associated open areas; (2) formerly grazed meadow areas; and (3) forested sites without indications of strong anthropogenic influence. Past land-use, environmental variables and species traits were used to interpret floristic gradients. Unconstrained and constrained ordinations clearly separated plots with different past land-use, and showed that the main floristic gradients on Borgø correlated with hydrological-edaphic conditions and past land-use. Partial constrained ordination using the environmental parameters as covariables showed that past land-use had a unique influence on floristic composition.     

Geohistorical and current environmental influences on floristic differentiation in the Ryukyu Archipelago, Japan

Aim Floristic differentiation in the Ryukyu Archipelago has been explained primarily by geohistory, specifically landbridge formation and vicariance at the Tokara and Kerama Gaps, two deep‐sea channels through the island arc. This ignores current environmental effects, which may also be important. We therefore tested whether the floristic differentiation pattern is explained primarily by the historical effect of the gaps as barriers, or whether a better understanding of floristic differentiation is achieved when both historical and current environmental factors are incorporated. Location Ryukyu Archipelago, Japan: an assemblage of continental islands. Methods We compiled a presence–absence matrix of 1815 plant species on 26 islands. Floristic dissimilarity distances between islands were calculated using Simpson’s similarity index and analysed using cluster analysis. We also conducted multiple regression on distance matrices (MRM) to examine the significance of the historical factors of the gaps and current environmental factors: geographical distance among islands and differences in island area and maximum elevation. Results We detected clear patterns of floristic differentiation across the gaps. Using the two gaps as explanatory variables, the MRM showed that both had significant effects on floristic dissimilarity distance. However, when geographical distance was added to the explanatory model, the Kerama Gap effect disappeared. When all five explanatory variables were used, the Tokara Gap and geographical distance had positive effects, but area difference had a negative effect. The Kerama Gap and difference in maximum elevation had no effect. Main conclusions The geographical pattern of floristic differentiation appears to indicate the influence of both gaps. However, the MRM indicates that the floristic differentiation across the Kerama Gap is no more than could be explained solely by geographical distance. Across the Tokara Gap, however, floristic differentiation is larger than geographical distance alone can explain. This additional differentiation is attributable to the effect of the historical barrier. To verify the significance of historical effects of vicariance on island biota, the confounding effects of geographical distance must be considered. The distance decay of floristic similarity and negative effect of area difference on floristic differentiation demonstrate that floristic differentiation is better understood by incorporating both historical and current environmental factors.     

Gap disturbance regime and composition in the Atlantic Montane Rain Forest: the influence of topography

In the Atlantic Montane Rain Forest of south-eastern Brazil, a field study was carried out to describe the forest disturbance regime, analyse canopy gap composition and evaluate the influence of habitat parameters on gap tree species composition. We characterized canopy gaps considering the group of variables as follows: area, type and number of tree/branch falls, topographic position, soil coverage and surrounding canopy trees. Gap composition was assessed at species level by measuring all individuals inside gaps higher than one meter. Mean gap area of the 42 canopy gaps analysed was 71.9 ± 9.0 m2 (mean ± SE). Out of the studied gaps, 35.7% were created by uprooted and by snapped trees, 16.7% by dead-standing trees and 11.9% by the fall of large branches. The disturbance regime was characterized by gap openings predominantly smaller than 150 m2 and by spatial patterning related to topography. Ridges had smaller gaps and higher proportions of gaps created by branch falls; slopes had bigger gaps generally created by uprooting events. The more abundant and frequent species were shade tolerant and the more species-rich families found inside gaps did not differ from the forest as a whole. Pioneer species were rare and restricted to medium and large size classes. The Indicator Species Analysis and the Canonical Correspondence Analysis indicated gap area, topography and the percentage of soil cover by the genera Calathea and Ctenanthe were the predominant variables correlated with woody species distribution. So, topography emerged as an important issue not only to the gap disturbance regime, but also to gap colonization. In respect to the influence of gap processes on the Atlantic Montane Rain Forest regeneration, our results support the view that canopy gap events may not be working as promoters of community wide floristic shifts.     

Floristic patterns of epiphytes in the Brazilian Atlantic Forest,a biodiversity hotspot

To assess the main factors driving epiphytic angiosperm distribution throughout the Brazilian Atlantic Forest, we compiled 57 floristic surveys and analysed species composition under the influence of environmental variables, space and vegetation type using canonical correspondence analysis (CCA), similarity (Sorensen) and Mantel's tests. The indicator value index (IndVal) was used to find indicator species of each Brazilian Atlantic Forest vegetation type. Group sharpness analysis was performed in order to determine the appropriate group partition level. CCA showed a separation of the epiphytic flora reflecting temperature and rainfall gradients. Mantel's test showed that environment and space were highly correlated with floristic similarity. Cluster analysis, indicating floristic similarity, resolved five groups, mainly grouped by region. Clear differentiation of the Brazilian Atlantic Forest epiphytic flora on a north–south axis with a strong correlation with temperature and rainfall gradients was found. The role of space and environment on species composition varied according to distinct epiphytic species groups. In particular, for Bromeliaceae and Orchidaceae, the main factor associated with floristic similarity was space. Indicator species were found for all vegetation types apart from the Seasonal Semideciduous Forest that seemed to represent a subset of a more humid forest type. © 2015 The Linnean Society of London, Botanical Journal of the Linnean Society, 2015, 179 , 587–601.     

The phytogeography of European and Mediterranean heath species (Ericoideae, Ericaceae): a quantitative analysis

The geographic ranges of heath species in Europe and the Mediterranean and their relationships with environmental (climatic and ecogeographic) variables and biological features are analysed by means of multivariate methods. In particular, twinspan classifications into floristic elements and floristic regions, DCA floristic ordinations, CCA environment-constrained ordinations and CCA biology-constrained ordinations are carried out. Results of the analyses show a correspondence with conventional regionalization analyses based on broader criteria, and less correspondence with numerical analyses of other taxonomic groups at a similar scale. This lack of fit depends on the particular history and ecology of the taxonomic groups under study. A number of climatic (temperature and water stress) and geographic (coast length) variables are associated with different types of heaths according to their geographical ranges (continental, Mediterranean, Atlantic). Biological features of heaths account for a small part of the variation in range, but the association of temperate heaths with a preference for acid soils, of Mediterranean heaths with pubescence, and of Atlantic heaths with plant height is of interest. Heath species richness throughout Europe and the Mediterranean is analysed by multiple regression analyses and, apart from a strong influence of area size, a significant effect of water conditions, temperature and proximity to sea is detected. The area with the highest heath species richness is Western Mediterranean. Additional classification, ordination and multiple regression analyses of heaths in the Iberian Peninsula (Spain and Portugal, the areas with the highest heath diversity) revealed similar patterns to those found in Europe and the Mediterranean. The effect of heterogeneity of the studied units at this latter scale is removed in the Iberian analysis because of the relative homogeneity of the units considered at this scale.     

Floristic composition across a climatic gradient in a neotropical lowland forest

Abstract. This study deals with the floristic composition of lowland tropical forest in the watershed of the Panama Canal. The floristic composition of large trees in 54 forest plots was analysed with respect to environmental factors, including precipitation, geologic parent material, stand age, topography, and soils. The plots contain 824 species of trees with a diameter at breast height ≥10 cm and represent a regional flora with exceptional β‐diversity. Plot data indicate that the Panamanian forest is strongly spatially structured at the landscape scale with floristic similarity decreasing rapidly as a function of inter‐plot geographic distance, especially for distances <5 km. The ordinations and patterns of endemism across the study area indicate broad floristic associations well correlated with Holdridge life zones. The results indicate the positive aspects of life zone classification at regional scales, while simultaneously highlighting its inadequacy for finer scales of analysis and resource management. Multivariate gradient analysis techniques (Non‐metric Multidimensional Distance Scaling and Detrended Correspondence Analysis) show clear patterns of floristic variability correlated with regional precipitation trends, surficial geology, and local soil attributes. Geologic and edaphic conditions, such as acidic soils or excessively drained limestone substrates, appear to override the effects of precipitation and modify forest composition. We conclude that the Panamanian forest shows clear patterns of spatial organization along environmental gradients, predominantly precipitation. The rapid decline in floristic similarity with distance between stands also suggests a role for dispersal limitation and stochastic events.     

Winter Bird Assemblages in Rural and Urban Environments: A National Survey

Urban development has a marked effect on the ecological and behavioural traits of many living organisms, including birds. In this paper, we analysed differences in the numbers of wintering birds between rural and urban areas in Poland. We also analysed species richness and abundance in relation to longitude, latitude, human population size, and landscape structure. All these parameters were analysed using modern statistical techniques incorporating species detectability. We counted birds in 156 squares (0.25 km2 each) in December 2012 and again in January 2013 in locations in and around 26 urban areas across Poland (in each urban area we surveyed 3 squares and 3 squares in nearby rural areas). The influence of twelve potential environmental variables on species abundance and richness was assessed with Generalized Linear Mixed Models, Principal Components and Detrended Correspondence Analyses. Totals of 72 bird species and 89,710 individual birds were recorded in this study. On average (±SE) 13.3 ± 0.3 species and 288 ± 14 individuals were recorded in each square in each survey. A formal comparison of rural and urban areas revealed that 27 species had a significant preference; 17 to rural areas and 10 to urban areas. Moreover, overall abundance in urban areas was more than double that of rural areas. There was almost a complete separation of rural and urban bird communities. Significantly more birds and more bird species were recorded in January compared to December. We conclude that differences between rural and urban areas in terms of winter conditions and the availability of resources are reflected in different bird communities in the two environments.     

Is there environmental differentiation in the <Emphasis Type="Italic">Quercus</Emphasis>-dominated forests of west-central Mexico?

Detailed information on 38 species and 26 environmental variables was recorded from a network of 86 permanent plots across a geographical range of 10 km, in order to determine the patterns of floristic composition in Quercus-dominated forests; to elucidate environmental differentiation in such forests; and to determine whether species are partitioning their environment. To examine likely patterns of floristic composition, a data matrix expressed as relative volume + relative density was used to run non-metric multidimensional scaling. Canonical correspondence analysis extracted the environmental variation that best correlates with the observed patterns of floristic composition. Our results indicate that congeneric Quercus individuals represent the largest proportion of the species pool in the study plots. They coexist with other species having similar ecological requirements in at least three distinct floristic groups. Examination of the two largest groups and their species compositions reveals that one floristic gradient runs across the most xeric zone of the study area, and the second major floristic gradient runs across a mesic zone. The most important environmental variable explaining the observed patterns of floristic composition is altitude, although partial canonical correspondence analysis suggests that micro-habitat heterogeneity (catena position and canopy maturity) was most significant.     

Biogeographical relations of a hyperarid desert flora in eastern Egypt

The floristic composition and geographical elements of the study area were analysed resulted in 328 species representing 206 genera in 55 families. This study confirmed the record of fourteen species, mostly weeds, which can be considered as new additions to the flora of the study area. Therophytes were the dominant life form, while mono‐ and bi‐regional Saharo‐Sindian geoelements were the most represented. Ten species showed dominancy with their Q‐values ranged between 0.802 and 0.2, where Zilla spinosa and Zygophyllum coccineum were of common occurrence. Application of cluster analysis and DCA ordination techniques produced four major floristic groups (A–D) comprising seven subgroups. The correlation coefficients (r) between the different subgroups revealed high significant correlations (P = 0.01) between floristic group (B) and subgroup (C₂) and between subgroups (D₁) and (D₂). Significant correlations (P = 0.05) occurred between subgroup (D₁) and both of (A₁) and (C₂). Comparing the floristic similarities between this investigation and other relevant studies were presented and discussed. On the other hand, the low similarity index between the study area and Sinai may be attributed to the geographical position of both deserts where Sinai desert is part of the Irano‐Turanian region, while the Eastern Desert is a part of the Saharo‐Sindian region.     

Regional variation in kwongan in the central wheatbelt of south-western Australia

This study examined the floristic variation at 20 sites of sclerophyllous shrubland (kwongan) occurring on lateritic soils in the central wheatbelt of south-western Australia. The results show that this kwongan is as floristically rich and varied as the kwongan in the northern and southern sandplain areas of the south-west. Six sequential groups of sites can be identified along the rainfall gradient present over the study area; 29 species are selected as indicators of the floristic groups. The distribution of many of the remaining 426 species suggests a continuum of floristic change along the rainfall gradient. One hundred and forty-five species were only recorded once and thus contributed to the high species richness of the sites and the low level of floristic uniformity between the stands. There were significant trends in the proportions of lifeforms with respect to the rainfall gradient. Evidence of the response of the plants to an obvious environmental gradient and the increased understanding of this complex vegetation type will provide conservation managers with some rationale for setting priorities for reserve acquisition and management.     

The combined role of topography and overstorey tree composition in promoting edaphic and floristic variation in a Mediterranean forest

This study aimed to address which factors, other than topography, contribute to the floristic variation of forested slopes. The natural forest studied is located in the Sudoeste Alentejano e Costa Vicentina Nature Park (southwestern Portugal). We sampled topographic, edaphic, floristic and community structure variables along three bottom–top hillside transects. Multivariate analyses of soil variables (by PCA), and of woody species composition and floristic–environmental relationships (by CCA and pCCA) were performed. Environmental–floristic trends strongly associated with the elevation gradient were identified. At lower altitudes, the lowest species richness, the highest soil fertility, and the tallest and most dense (least available light) canopy occurred. The spatial variation in woody species composition and abundance was closely associated with Zn availability in the soil and litter groundcover, but these varibles had significant spatial structure in the studied forest. The non-spatially structured species variance was better predicted by soil NO₃^– and NH₄⁺. The spatial variation of species data not shared with environmental variables was also calculated. We suggest that the influence of topographic gradient on the variation of edaphic variables and on the distribution and abundance of woody species was mediated by overstorey tree composition. Locally dominant tree species, in particular Quercus faginea and Quercus suber, may function as ecosystem engineers promoting environmental changes (i.e., Zn availability in soil, litter accumulation and light availability) that influence overall floristic variation.     

Modelling floristic species richness on a regional scale: a case study in Switzerland

In this paper a multivariate linear regression model is proposed for predicting and mapping regional species richness in areas below the timberline according to environmental variables. The data used in setting up the model were derived from a floristic inventory. Using a stepwise regression technique, five environmental variables were found to explain 48.9% of the variability in the total number of plant species: namely temperature range, proximity to a big river or lake, threshold of minimum annual precipitation, amount of calcareous rock outcrops and number of soil types. A considerable part of the unexplained variability is thought to have been influenced by variations in the quality of the botanical inventory. These results show the importance of systematic floristic sampling in addition to conventional inventories when using floristic data as a basis in nature conservation. Nevertheless it is still possible to interpret the resulting diversity patterns ecologically. Regional species richness in Switzerland appears to be a function of: (i) environmental heterogeneity; (ii) threshold values of minimum precipitation; and (iii) presence of calcareous rock outcrops. According to similar studies, environmental heterogeneity was the strongest determinant of total species richness. In contrast to some studies, high productivity decreased the number of species. Furthermore, the implications of this work for climate change scenarios are discussed.     

Distributional modelling and parsimony analysis of endemicity of Senecio in the Mediterranean-type climate area of Central Chile

Aim Floristic blocks and areas of endemism resulting from a parsimony analysis of endemicity (PAE) using raw floristic data versus data generated from distributional modelling for 130 species in the genus Senecio Tourn. ex L. distributed in the Mediterranean‐type climate area of Central Chile were compared, and the results were used to identify conservation priorities for the flora of the region. Location Central Chile, between 30° and 38° S. Methods Using herbarium records, a species × area matrix consisting of presence/absence data was constructed from a 0.5° × 0.5° grid. Distributional modelling techniques incorporating vegetation formations, elevation and the contagion index were used to interpolate floristic composition of poorly known areas. Parsimony analysis of endemicity was used to identify floristic blocks and areas of endemism. Results Using the number of most parsimonious trees as an index, distributional modelling greatly optimized the results of the PAE analysis. Three floristic blocks and four areas of endemism were suggested based on the PAE results using potential distribution data not incorporating the contagion index, while four blocks and two areas of endemism were suggested from the PAE results using potential distribution data incorporating the contagion index. Floristic blocks for the northern coast, southern Andes, and northern/central Andes were found, with some blocks showing divisions within them representing distinct geographic subunits. Major breaks between and within floristic blocks were identified at 32.5°–33° S and 34.5°–35° S. Main conclusions The floristic blocks identified with the distributional modelling and PAE correspond well to results from some previous studies and support hypothesized biogeographic divisions within Central Chile. The results were similar to those obtained from parallel analysis of the entire tree flora of Central Chile. The vegetative formation‐based distributional modelling produced robust and reproducible results when used along with PAE, especially when the contagion index was incorporated, and is a useful technique for area classification. The results demonstrate the utility of Senecio as an indicator genus for biogeography and conservation in southern South America.     

Evaluation of floristic diversity in urban areas as a basis for habitat management

Questions: How can floristic diversity be evaluated in conser‐vation plans to identify sites of highest interest for biodiversity? What are the mechanisms influencing the distribution of species in human‐dominated environments? What are the best criteria to identify sites where active urban management is most likely to enhance floristic diversity? Location: The Hauts‐de‐Seine district bordering Paris, France. Methods: We described the floristic diversity in one of the most urbanized French districts through the inventory of ca. 1000 sites located in 23 habitats. We built a new index of floristic interest (IFI), integrating information on richness, indigeneity, typicality and rarity of species, to identify sites and habitats of highest interest for conservation. Finally, we explored the relationship between site IFI and land use patterns (LUP). Results: We observed a total of 626 vascular plant species. Habitats with highest IFI were typically situated in seminatural environments or environments with moderate human impact. We also showed that neighbouring (urban) structures had a significant influence on the floristic interest of sites: for example, the presence of collective dwellings around a site had a strong negative impact on IFI. Conclusions: Our approach can be used to optimize management in urban zones; we illustrate such possibilities by defining a ‘Site Potential Value’, which was then compared with the observed IFI, to identify areas (e.g. river banks) where better management could improve the district's biodiversity.     

Habitat selection by small mammals in a Victorian heathland

Habitat preferences of five species of small mammals were studied on a 6.6 ha trapping grid in heathland on an undulating, deep, sandy podzol at Cranbourne, Victoria. The 120 sites were sorted into groups using a polythetic, agglomerative, non-hierarchical clustering procedure with (i) floristic and (ii) structural data. The dispersion of 4051 trap captures over 28 months was studied in relation to these groupings. Rattus lutreolus showed no preference for any of the structural groups, but good differentiation was obtained with the floristic groups. Dispersion of R. lutreolus was related to a sedge-food index, and seasonal change in R. lutreolus dispersion was related to change in rainfall. The preference of R. rattus for areas of wet heath of high structural complexity was best revealed using the structural classification. The results for the other species tended to favour the floristic rather than structural groups. Pseudomys novaehollandiae and Isoodon obesulus preferred dry heath of a younger successional stage. Mus musculus captures showed a preference for the dry heath generally. This preference was most pronounced in spring when the population was declining. The wet community species (R. lutreolus, R. rattus) tended to be food specialists and habitat generalists and the dry community species P. novaehollandiae, I. obesulus, M. musculus), exhibited converse traits.     

Plant composition and diversity in a semi-natural Mediterranean island landscape: the importance of environmental factors

Abstract

Islands crucially contribute to the Mediterranean Basin’s high floristic diversity, which, however, is at risk facing climate and land-use changes. Besides the identification of highly diverse areas, the knowledge about factors favouring diversity is of great importance. We analysed plant species diversity and composition related to environmental factors over varied vegetation units on a former Italian prison island in the northwest of Sardinia. Due to a long history of land use with grazing and later abandonment the nowadays protected island features a semi-natural landscape and can serve as an example for strongly anthropogenic altered insular ecosystems. Floristic composition, soil properties, microclimate and ungulate abundance were assessed. Relationships of vegetation composition and diversity with abiotic variables were examined by Canonical Correspondence Analysis, which indicated the importance of air temperature, soil moisture, slope gradient and C/N ratio for floristic differentiation. Most important abiotic factors for plant species richness were relative air humidity and soil moisture, while floristic diversity was mainly determined by air temperature and pH. Furthermore, observation data pointed to an adverse influence of ungulate abundance for plant species diversity. Regarding nature conservation, grazing intensity thus must be critically taken into account, especially for sensitive vegetation units like the coastal garrigue.