BioOne.org will be down briefly for maintenance on 17 December 2024 between 18:00-22:00 Pacific Time US. We apologize for any inconvenience.
Open Access
How to translate text using browser tools
1 May 2016 Forest specialist and generalist small mammals in forest edges and hedges
Hella Schlinkert, Martin Ludwig, Péter Batáry, Andrea Holzschuh, Anikó Kovács-Hostyánszki, Teja Tscharntke, Christina Fischer
Author Affiliations +
Abstract

Agricultural intensification often leads to fragmentation of natural habitats, such as forests, and thereby negatively affects forest specialist species. However, human introduced habitats, such as hedges, may counteract negative effects of forest fragmentation and increase dispersal, particularly of forest specialists. We studied effects of habitat type (forest edge versus hedge) and hedge isolation from forests (connected versus isolated hedge) in agricultural landscapes on abundance, species richness and community composition of mice, voles and shrews in forest edges and hedges. Simultaneously to these effects of forest edge/hedge type we analysed impacts of habitat structure, namely percentage of bare ground and forest edge/hedge width, on abundance, species richness and community composition of small mammals. Total abundance and forest specialist abundance (both driven by the most abundant species Myodes glareolus, bank vole) were higher in forest edges than in hedges, while hedge isolation had no effect. In contrast, abundance of habitat generalists was higher in isolated compared to connected hedges, with no effect of habitat type (forest edge versus hedge). Species richness as well as abundance of the most abundant habitat generalist Sorex araneus (common shrew), were not affected by habitat type or hedge isolation. Decreasing percentage of bare ground and increasing forest edge/hedge width was associated with increased abundance of forest specialists, while habitat structure was unrelated to species richness or abundance of any other group. Community composition was driven by forest specialists, which exceeded habitat generalist abundance in forest edges and connected hedges, while abundances were similar to each other in isolated hedges. Our results show that small mammal forest specialists prefer forest edges as habitats over hedges, while habitat generalists are able to use unoccupied ecological niches in isolated hedges. Consequently even isolated hedges can be marginal habitats for forest specialists and habitat generalists and thereby may increase regional farmland biodiversity.

Agricultural intensification often leads to the destruction and fragmentation of natural habitats, such as forests. Thereby the resulting isolation of these natural habitats in agricultural landscapes boosts local population extinction of many organisms (Fischer and Lindenmayer 2007, Bennett and Saunders 2010). Increasing landscape heterogeneity by adding areas of natural and semi-natural habitats can counteract this negative trend by increasing habitat connectivity and additionally often contributes to the conservation of farmland biodiversity (Tewksbury et al. 2002, Bennett and Saunders 2010). Hedges are human introduced linear woody elements consisting of shrubs and/or trees in the agricultural landscape that provide various ecological functions, such as corridors for forest species or refuges for open land species (Baudry et al. 2000). Therefore, hedges function as seminatural habitat and can increase connectivity between forests in agricultural landscapes, mitigating potential negative effects of habitat fragmentation (Niemelä 2001, Bennett and Saunders 2010). However, the extent to which species use different habitat types and are affected by habitat isolation depends on species specific habitat specialisation, dispersal ability and their ability to cope with landscape modifications (Niemelä 2001, Fischer and Lindenmayer 2007).

Small mammals are of ecological importance as they represent an important link in the food chain and their occurrence may not only influence their food resources, such as seeds and invertebrates (Abt and Bock 1998), but also population densities of their predators (Aschwanden et al. 2005, Koks et al. 2007). On the other hand, small mammals, such as Microtus arvalis (common vole), are often considered as agricultural pests in central Europe, causing economic damage to crops (Jacob et al. 2014) or preying on invertebrates that perform ecosystem services, such as spiders, carabid beetles or earthworms (Abt and Bock 1998). Small mammals can also be reservoir hosts and vectors of zoonotic diseases, such as hantavirus and therefore source of infection in humans (Voutilainen et al. 2012).

Small mammal communities, and individual species e.g. Myodes glareolus (bank vole) and Apodemus sylvaticus (longtailed field mouse, wood mouse), are often positively affected by large size of forest patches, high connectivity of patches of old forest and structural heterogeneity of the forest floor (Fitzgibbon 1997, Vieira et al. 2009, Mortelliti et al. 2011). However, studies analysing the impact of forest patch isolation in agricultural landscapes on small mammals were often conducted in forests or forest patches, but not in smaller woody landscape elements such as hedges (but see Michel et al. 2006, Gelling et al. 2007, Silva and Prince 2008) and did not take species specific habitat specialisation into account (but see Yahner 1983, Bentley et al. 2000, Tattersall et al. 2002). Small mammals show distinct habitat specialisation and can be classified into forest and open land specialists and habitat generalists, each responding differently to changes in landscape complexity (Gentili et al. 2014). Forest specialists such as Apodemus flavicollis (yellow-necked mouse) and M. glareolus mainly live in forest habitats, avoiding open agricultural areas (Pollard and Relton 1970, Kotzageorgis and Mason 1996, Tattersall et al. 2002, but see Broughton et al. 2014). Therefore, forest specialists may benefit from hedges, which generally can function as habitat, dispersal corridors or stepping stones for different species (Davies and Pullin 2007). In contrast, habitat generalists, such as Sorex araneus (common shrew) and A. sylvaticus, are able to thrive in a wide range of environmental conditions and frequently occur in open agricultural landscapes (Heroldová et al. 2007, Tattersall et al. 2002, Wang and Grimm 2007). Thus, habitat generalists should benefit less from decreasing isolation of forests through hedges compared to forest specialists (Bentley et al. 2000).

Besides habitat type and effects of isolation, habitat structure described through e.g. hedge width, height and ground cover can also influence small mammal occurrence. A high level of microhabitat complexity through high shrub diversity, high vegetation cover and low percentage of ground covered by bare-soil provide refuges from predators and are known to increase overall small mammal abundance and species richness (Gelling et al. 2007, Silva and Prince 2008). Additionally, increasing hedge width increases the amount and quality of the habitat and thereby may positively affect small mammal abundance of forest specialists as well as habitat generalists (Gelling et al. 2007, but see Bellamy et al. 2000).

In the present study we investigated the effects of habitat type and isolation on small mammal abundance, species richness and community composition, paying particular attention to the abundance of forest specialists and habitat generalists. We therefore compared three different types of structurally similar linear woody landscape elements (hereafter referred to as ‘forest edge/hedge type’): shrub dominated forest edges, sharing large parts of their borders with the forest; hedges connected to forest, sharing a small part of their borders with the forest; hedges isolated from forest by at least 300 m, with no structural connectivity to forests. Along with effects of habitat type (forest edge versus hedge) and hedges' isolation (connected versus isolated hedge), we tested the influence of habitat structure (percentage of ground covered by bare-soil and forest edge/hedge width) on small mammal occurrence.

We hypothesised that

  • 1) forest specialists are affected by habitat type and isolation, with highest abundance in forest edges, intermediate abundance in connected hedges and lowest abundance in isolated hedges, and that there is no influence of habitat type or isolation on habitat generalists;

  • 2) small mammal community composition changes with decreasing dominance of forest species from forest edges to isolated hedges;

  • 3) abundance of both forest specialists and habitat generalists increases with decreasing percentage of ground covered by bare-soil and increasing forest edge/hedge width.

  • Material and methods

    Study sites and sampling design

    The study sites were located in the surroundings of Göttingen, Lower Saxony, Germany (51°5′N, 9°9′ E; for a map see Batáry et al. 2012). We chose six forest edges, six hedges connected to forest and six hedges isolated from forest (Fig. 1, Supplementary material Appendix 1 Fig. A1). Forest edges with a distinct shrub layer distinguishable from the inner part of the forest were chosen, which were structurally similar to hedges regarding plant composition, stratification, shrub height, or width and can be essentially viewed as hedges bordering forest (Supplementary material Appendix 1 Fig. A1a; for a detailed list of hedge characteristics see Table 1). Vegetation of forest edges and hedges was dominated by Prunus spinosa (blackthorn), Crataegus spp. (hawthorn), and Rosa spp. (rose), and trees and shrubs such as Acer spp. (maple) and Cornus sanguinea (dogwood) were frequently interspersed. Forest edges and connected hedges adjoined deciduous forests dominated by Fagus sylvatica (beech), which was not masting in 2008 in Lower Saxony, the year before our study took place (Nordwestdeutsche Forstliche Versuchsanstalt 2008). Forest edges and hedges were either bordered by winter cereal fields or winter oilseed rape fields (cf. Kovács-Hostyánszki et al. 2013). Landscape composition around all 18 study sites was dominated by arable land and forest in a 500 m radius around study sites, but differed between forest edges and hedges due to their different configuration (cf. Batáry et al. 2012).

    Figure 1.

    Study design: (a) forest edge, (b) connected hedge, (c) isolated hedge.

    f01_86.jpg

    Table 1.

    Parameters describing habitat structure of forest edges, connected hedges and isolated hedges. Overall minimum and maximum values and mean values with standard errors per forest edge/hedge type are given. Differences of habitat structure between forest edges, connected hedges and isolated hedges were assessed by calculating F- and p-values from a one-way ANOVA (DF 2,15) when parameters were normally distributed or by x2-values from a Kruskal-Wallis rank sum test (DF 2) when parameters were not normally distributed.

    t01_86.gif

    To reduce spatial autocorrelation the minimum distance between study sites was 500 m. The distance between isolated hedges and the nearest forest was at least 300 m, exceeding mean distances of small mammal movements (Kozakiewicz et al. 1993, Kotzageorgis and Mason 1996, Shchipanov et al. 2011), even though in some cases Myodes glareolus are able to move larger distances, depending on sex and season (Kozakiewicz et al. 2007). Forest edge or hedge length was at least 200 m and connected hedges adjoined perpendicular to the forest.

    Parameters characterising habitat structure of forest edges (by its distinct shrub layer) and hedges (Table 1, Supplementary material Appendix 1 Table A1) were measured in August 2009 at eight sampling points spacing 25 m (Fig. 1). Percentage of hedges' ground covered by bare-soil, litter and vegetation (grasses and herbs) were estimated with 10% precision at each sampling point from both sides of the forest edge or hedge within 1 × 1 m. Height and width of the shrub layer were measured with 0.5 m precision, excluding the forest tree layer in case of forest edges. Mean values of variables describing habitat structure were calculated per study site.

    Small mammal sampling

    Small mammal sampling was conducted twice, in May and in July 2009 during the growing season, before the crop harvest. 20 multiple-capture live-traps (Ugglan-traps, 240 × 60 × 90 mm; Grahnab, Sweden) were placed with 10 m spacing in the middle of each forest edge or hedge (Fig. 1) (cf. Silva and Prince 2008). In case of connected hedges starting point of trap lines was 5 m away from forest edges. Trapping took place for three consecutive trapping nights. Before the first trapping night traps were pre-baited for one night to increase trapping efficiency. Traps were baited with rolled oats in the evening before sunset and checked in the early morning after sunrise. To sample the small mammal community representatively without any bias to species with certain food preference, no high-energy bait for shrews was used, since a pilot study in the area showed that capture probability but also mortality of shrews increased by additionally using mealworms as bait. Trapped small mammals were identified to species and marked with permanent micro tattoos (FST Laboratory Animal Microtattoo System, green tattoo paste, sterile hypodermic needles 27 G × ½”) on the tail base to enable identification when recaptured.

    Abundance of small mammals per forest edge or hedge was calculated by excluding recaptures (cf. Fischer et al. 2011, Michel et al. 2006), as there was no difference in trappability (percentage of recaptures) among forest edges, connected and isolated hedges (one-way analysis of variance: F2,15 = 0.44, p = 0.65). Further, in case of the connected hedges, there was no difference in the number of trapped individuals among traps with different distance from the forest edge (analysis of variance: F19,15 = 1.34, p = 0.16), therefore we pooled these data. Small mammal species were classified after Niethammer and Krapp (1978, 1982, 1990) and IUCN (2013) according to their habitat specialisation. M. glareolus and Apodemus flavicollis were categorised as ‘forest specialists’ (depend on forests, forest edges and forest patches), Apodemus agrarius (striped field mouse), A. sylvaticus, S. araneus and Sorex minutus (pygmy shrew) as ‘habitat generalists’ (no distinct habitat preference) and Crocidura leucodon (bicoloured white-toothed shrew) and M. arvalis as ‘open land specialists’ (depend on open agricultural landscapes).

    Statistics

    Correlation analyses (Pearson's product-moment correlation when both variables were normally distributed and Spearman's rank correlation when at least one variable was non-normally distributed) were performed to test for multicollinearity between all measured variables that describe habitat structure. Correlated variables (threshold level: |r| >0.7; Table 2; Dormann et al. 2013) and variables that were highly skewed (vegetation cover; Table 1) were excluded from further analysis. The independent variables selected that describe habitat structure were percentage of ground covered by bare-soil and forest edge/hedge width (hereafter referred to as bare ground and width).

    Linear mixed-effects models (lme function; Pinheiro and Bates 2000) with a maximized log-likelihood implemented in the R package nlme (ver. 3.1–111; Pinheiro et al. 2013) were used to test the effects of forest edge/hedge type (forest edge versus connected hedge versus isolated hedge), habitat structure (bare ground and width) and all two-way interactions using R 3.0.2 software (< www.r-project.org>). Total small mammal abundance, species richness, abundance of each of the different habitat specialisation groups, and abundance of each of the two most abundant species (M. glareolus, S. araneus) were used as response variables. Study sites were included as random effect to model the independence of errors with respect to temporal autocorrelation (two trapping sessions within the same study site; Pinheiro and Bates 2000). To account for normal error distribution and homoscedasticity, response variables were square root transformed when necessary. Model selection and averaging was done by an information criteria approach using multimodel inference from the R package MuMIn (ver. 1.10.5; Barton 2014). From the global model, all candidate models (18 different models) were generated and ranked based on their AICc values (Akaike information criterion for small sample size corrections). Further, ΔAICc values in relation to the model with the minimum AICc, and Akaike weights, which provide a relative weight of evidence for each model and likelihood-ratio based R2-values were calculated. Models with ΔAICc ≤ 2, which provide substantial empirical evidence for the model (Burnham and Anderson 2002), were averaged and the relative variable importance, weighted averages of parameter estimates, standard errors and confidence intervals were calculated. Confidence intervals of parameter estimates, which did not include zero, were considered to influence response variables significantly (Grueber et al. 2011). Contrasts between forest edges, connected and isolated hedges were investigated by re-ordering factor levels.

    Table 2.

    Correlation coefficients of parameters describing habitat structure of hedges and forest edges. Variables highlighted in grey exceeded threshold level of |r| > 0.7. Bare ground and width were used for further analysis.

    t02_86.gif

    Table 3.

    Total small mammal abundance, species richness and abundances of captured species per trapping session per forest edge, connected and isolated hedge (2 trapping sessions, 6 study sites: n = 12 respectively). Total values and mean values with standard errors are given.

    t03_86.gif

    Small mammal community composition was assessed in relation to forest edge/hedge type (forest edge versus connected versus isolated hedge) and habitat structure (bare ground and width) by using a constrained ordination model. To decide on the ordination method, we first ran a DCA (detrended correspondence analysis, an indirect gradient analysis) to estimate the gradient length which measures the beta diversity in small mammal community composition (Leps and Smilauer 2003), for species data summed up over both trapping sessions. If the gradient length is shorter than 3.0 a linear constrained ordination method (RDA, redundancy analysis) is appropriate and if the longest gradient is larger than 4.0 the data are heterogeneous and a unimodal method (CCA, canonical correspondence analysis) should be applied (Leps and Smilauer 2003). Since the gradient length was shorter than 3.0, indicating species with a linear response, we decided to perform partial RDA (Leps and Smilauer 2003) using the R package vegan (ver. 1.17; Oksanen et al. 2011). Prior to analysis the species matrix was square root transformed. The species matrix was constrained by either forest edge/hedge type, bare ground or width, with the other two variables used as conditioning variables. The significance of constrained variables for each partial model was assessed by calculation of permutation tests based on 999 permutations. Pseudo-F-values with the corresponding p-values and ordination plots are given. To test for possible shifts of community composition among habitat specialisation groups, abundances of forest specialists and generalists were compared for forest edges, connected and isolated hedges, respectively using two-tailed Student's t-tests.

    Results

    We recorded 805 individuals (representing the first capture) of 11 species from 2160 trapping nights (1335 captures including recaptures; for total abundance, species richness and abundance of the different species within the forest edges/hedges: Table 3, Supplementary material Appendix 1 Table A1). Trap mortality, calculated as the number of animals that unintentionally died in traps, divided by the total number of captures, was 10% for all species, with 1% mortality rate among rodents and 59% among shrews. Individuals from Neomys fodiens (Eurasian water shrew), Arvicola amphibius (European water vole), and Glis glis (edible dormouse) were excluded from analyses as these captures (four individuals in sum) were not representative due to unsuitable trapping method for this species or specialisation on aquatic habitat structures. We were not able to apply linear mixed effects models for open land specialists, as their abundance was highly right skewed, with only one or two individuals occurring in five forest edges and connected hedges and 26 individuals occurring in one isolated hedge.

    Table 4.

    Effects of forest edge/hedge type (forest edge versus connected hedge versus isolated hedge) and habitat structure (bare ground and width) on small mammal abundance, species richness and abundance of forest specialists and habitat generalists (for composition of forest specialists and habitat generalists see Table 3). Results of multi-model averaging from linear mixed effects models with ΔAICc≤2 are given (parameter estimates with standard error, confidence intervals and levels of significance denoted with * = p ≤ 0.05 and ** = p < 0.01 ). Variables indicated by ‘-’ were not part of the averaged model. LWLE: linear woody landscape element (F: forest edge, C: connected hedge, I: Isolated hedge); %BG: bare ground, W: edge/hedge width, “:” indicates two-way interaction.

    t04_86.gif

    Total abundance, abundance of forest specialists, as well as abundance of Myodes glareolus were higher in forest edges than in hedges, while isolation of hedges had no effect (forest edge > connected hedge = isolated hedge; Table 4, Fig. 2a, c, e). In contrast, abundance of habitat generalists was higher in isolated hedges than in connected hedges, while habitat type (forest edge versus hedge) had no effect (Table 4, Fig. 2d). For total species richness and abundance of Sorex araneus we did not find an influence of forest edge/hedge type (Table 4, Fig. 2b, f). Abundance of forest specialists was positively associated with decreasing percentage of bare ground and increasing forest edge/hedge width (Table 4). Species richness or abundance of any other group or species were not influenced by habitat structure (bare ground and width; Table 4; for the model section table containing all models with ΔAICc ≤ 2 see Supplementary material Appendix 1 Table A2).

    Small mammal community composition was affected by forest edge/hedge type (variation explained = 28.6 %, pseudo-F2,13 = 3.14, p < 0.01; Fig. 3a) and width (variation explained = 12.5%, pseudo-F2,13 = 2.76, p < 0.05; Fig. 3b), but not by bare ground (variation explained = 10.4%, pseudo-F1,13 = 2.29, p = 0.09). Myodes glareolus dominated the small mammal community in forest edges. The other species contributed similarly to the small mammal communities of forest edges and hedges. Myodes glareolus and Apodemus flavicollis were predominantly present in small mammal communities in wide forest edges or hedges.

    These results were also underlined by tests for possible shifts of community composition among habitat specialisation groups. We found a higher abundance of forest specialists compared to habitat generalists in forest edges (t = 3.58, p < 0.05) and connected hedges (t = 2.68, p < 0.05). Isolated hedges harboured similar numbers of forest specialist and habitat generalist individuals (t = -0.12, p = 0.91).

    Discussion

    Habitat specialisation is a key factor influencing species occurrence in different habitat types, shown for e.g. birds (Batáry et al. 2012) or arthropods (Petit and Usher 1998, Fischer et al. 2013). Studies have shown that small mammal forest specialists avoid the open agricultural matrix and are consistently found in forests and, in accordance to our results, in other woody landscape elements such as forest edges and hedges (Pollard and Relton 1970, Kotzageorgis and Mason 1996, Bentley et al. 2000). Even though our results show effects of habitat type (forest edge versus hedge) on abundance of forest specialists, we could not show an effect of isolation (connected versus isolated hedges). Studies on small mammals in forests have shown negative effects of increasing isolation when focusing on single species, such as Myodes glareolus and Apodemus sylvaticus (Kozakiewicz 1985, Paillat and Butet 1996, Fitzgibbon 1997, but see Szacki 1987). This is in accordance to our data, as forest edges in our study were linear woody landscape elements sharing half of their borders with the forest while connected and isolated hedges were less or not connected to forests. As habitat connectivity can influence immigration and population growth of forest specialists, with higher immigration, as well as growth rates in well-connected habitats than in isolated habitats, (Fahrig and Merriam 1985, Kozakiewicz 1985, Guivier et al. 2011), one would expect lower abundances of forest specialists in isolated than in connected hedges. In contrast, emigration from isolated habitats may be lower due to high matrix-patch contrast, influencing annual population densities positively (Fitzgibbon 1997). Here, distances of at least 300 m to the nearest forest seem to be still sufficient for forest specialists to reach hedges without structural connectivity to their main habitat. This is also supported by the population density ratio of forest specialists and habitat generalists: forest specialists showed overall higher population densities than habitat generalists in forest edges and connected hedges while they were similar in isolated hedges. Further, not only woody structures but also marginal habitats, such as road verges (Bellamy et al. 2000, Broughton et al. 2014), as well as other factors such as habitat configuration in terms of habitat linearity (Tattersall et al. 2002) seem to be important for the occurrence of habitat specialists possibly leading to non-significant effects of isolation in our study.

    Figure 2.

    Effects of forest edge/hedge type on small mammals. (a) total abundance, (b) species richness, (c) abundance of forest specialists, (d) abundance of habitat generalists, (e) abundance of M. glareolus and (f) abundance of S. araneus. Means ± SE per trapping session per forest edge, connected and isolated hedge (2 trapping sessions, 6 study sites: n = 12, respectively) are given. Factor levels containing different letters indicate significant differences at p < 0.05, which were determined by re-ordering factor levels between forest edges, connected hedges and isolated hedges. Non-significant results from averaged linear mixed-effects models are indicated by n.s.

    f02_86.jpg

    The higher abundance of forest specialists in forest edges than in hedges, was mainly driven by M. glareolus, which was the most abundant, predominant forest specialist and in all but one study site (in one of the isolated hedges). This can also affect the ecological functions provided through small mammals. For example, raptor abundances such as the common buzzard Buteo buteo decrease with decreasing prey availability and fragmentation (Butet et al. 2010). Further, the risk of disease emergence is potentially higher in forest edges compared to isolated hedges through higher abundances of M. glareolus and therefore higher number of reservoir hosts for e.g. Puumala hantavirus (cf. Voutilainen et al. 2012). In general, our results indicate the importance of hedges as landscape elements enhancing habitat availability for forest specialists and their population survival in fragmented landscapes. However, more clarity and robustness would be obtained with higher sample size over full population cycles and more years (also including masting years, potentially leading to population outbreaks of M. glareolus; Imholt et al. 2015), as well as by choosing isolated hedges along a gradient of isolation from forest, and studying small mammals' dispersal among hedges with different degrees of isolation.

    Figure 3.

    Effects of forest edge/hedge type on small mammal community composition. (a) RDA ordination diagram of small mammal communities in forest edges, connected hedges and isolated hedges shown with minimum convex polygons. (b) RDA ordination diagram for small mammal communities in forest edges and hedges with different widths (arrow). Small mammal species: Mg = M. glareolus, Af= A flavicollis, As = A. sylvaticus, Aa = A. agrarius, Sa = S. araneus, Sm = S. minutus, Ma = M. arvalis, C1 = C. leucodon.

    f03_86.jpg

    Habitat generalists were affected by isolation (higher abundances in isolated versus connected hedges) but not by habitat type (forest edge versus hedge), probably because they frequently use various landscape elements such as forests, grasslands or agricultural fields as habitat or as food source (Pollard and Relton 1970, Bentley et al. 2000, Ouin et al. 2000, Wang and Grimm 2007). This pattern may be more pronounced after the crop harvest when fields provide little food and cover and generalist species such as A. sylvaticus individuals have to retreat into hedges (Fitzgibbon 1997, Tattersall et al. 2001). However, the most abundant generalist species Sorex araneus did not show any response to habitat type or to isolation. As shrews are insectivorous species (Wang and Grimm 2007) and therefore exploit different food sources compared to the omnivorous M. glareolus which just feeds on 8 ± 2% high energetic invertebrates (Butet and Delettre 2011), there seems to be a higher niche differentiation resulting in a complete overlap of both species in our study sites. Even if our observed shrew mortality rate of 59% is in line with the literature (Eccard and Klemme 2013, Shonfield et al. 2013) it may has influenced overall shrew abundances blurring our results. Therefore, from a methodological and welfare point of view other methods need to be used such as shorter trapping intervals to reduce shrew mortality (Stromgren and Sullivan 2014).

    Species richness did not respond to habitat type or isolation of hedges. However previous studies showed negative effects of increasing forest isolation, due to a decrease of forest specialist species and constant numbers of generalist species (Pardini et al. 2005, Michel et al. 2006). In our study the lack of response of species richness to forest edge/hedge type can be explained by generally very low numbers of species (cf. Gentili et al. 2014), as well as by differences in small mammal community composition among forest edges and hedges. Forest specialists (M. glareolus and A. flavicollis) occur in all forest edges, but in just 80% of the hedges (connected and isolated hedges), while open land species (M. arvalis and C. leucodon) occasionally retreat into hedges and could be found in 20% of the hedges compared to 8% of forest edges, resulting in similar species richness. Further, small mammal communities of connected hedges were more similar to small mammal communities of forest edges than to those of isolated hedges, emphasizing an effect of isolation of hedges on small mammal communities (Bentley et al. 2000).

    In our study decreasing percentage of bare ground was positively associated with the abundance of forest specialists, but had no influence on any other group. This is in line with other studies showing small mammals' preferences for habitats with low amount of bare ground and dense cover of herbaceous vegetation (Mazurkiewicz 1994, Ouin et al. 2000, Silva and Prince 2008), mostly interpreted as antipredatory behaviour by small mammals. Thereby not only dense cover of herbaceous vegetation and litter, but also dense shrub layers of hedges and forest edges may protect small mammals from predators (Jedrzejewska and Jedrzejewski 1990, Longland and Price 1991, Weber 2008). Further, increasing forest edge/hedge width was positively associated with the abundance of forest specialists and changed the community composition (cf. Gelling et al. 2007, Michel et al. 2007). Increasing hedge width also increases the available habitat for forest species (Gelling et al. 2007), while habitat generalists such as S. araneus, A. agrarius and A. sylvaticus also use a variety of habitats within agricultural landscapes (Pollard and Relton 1970, Ouin et al. 2000, Wang and Grimm 2007), being less sensitive to decreasing availability of forest edge or hedge habitats. Since we merely used the naturally occurring variability of percentage of bare ground and forest edge/ hedge width as explanatory variables for small mammal occurrence, the variance within sites seems to be rather low to affect abundance of habitat generalists (but see Butet et al. 2006, Silva and Prince 2008).

    Conclusions

    We found that small mammal forest specialists were more abundant in forest edges than in hedges, while abundance of habitat generalists was higher in isolated compared to connected hedges. However, not only forest edges, but also hedges, harboured a relatively high abundance of forest specialists compared to habitat generalists, probably due to their higher dependency on woody landscape elements in agricultural areas. Isolation of hedges had no significant effect on abundance of forest specialists showing that habitat suitability of hedges without structural connection to the forest was equivalent to connected ones. Thus small mammal forest specialists may profit from even isolated hedges by providing habitat area and refuges in fragmented agricultural areas. However, to make general predictions on the impact of isolation of hedges on generalist small mammal species in agricultural landscapes, also potential positive effects of hedges as refuges after the crop harvest have to be studied. In sum, well connected as well as isolated hedges in agricultural landscapes are marginal habitats for small mammals, independently whether they are forest specialists, habitat generalists or even open land specialists. Thereby the maintenance and creation of hedges in agricultural landscapes may be an effective tool to increase small mammal diversity and thus overall farmland biodiversity.

    Acknowledgements

    We thank Kristy Udy for linguistic improvements, Christoph Scherber for statistical advice, Sebastian Hänke for help during field work, Birgit Jauker for her contribution to the sampling design, Susanne Schiele for assistance in all kind of problems and the research group of Stefan Halle (Jena) for providing traps. This work was supported by the German Research Foundation (DFG) and the Technische Universität München within the funding programme Open Access Publishing. PB was funded by the German Research Foundation (DFG BA 4438/1-1).

    References

    1.

    Abt, K. F. and Bock, W. F. 1998. Seasonal variations of diet composition in farmland field mice Apodemus spp. and bank voles Clethrionomys glareolus. — Acta Theriol. 43: 379–389. Google Scholar

    2.

    Aschwanden, J. et al. 2005. Are ecological compensation areas attractive hunting sites for common kestrels (Falco tinnunculus) and long-eared owls (Asio otus)? — J. Ornithol. 146: 279–286. Google Scholar

    3.

    Barton, K. 2014. MuMIn: multi-model inference. R package ver. 1.10.5. — <  http://CRAN.R-project.org/package=MuMIn >. Google Scholar

    4.

    Batáry, P. et al. 2012. Contrasting effect of isolation of hedges from forests on farmland vs woodland birds. — Commun. Ecol. 13: 155–161. Google Scholar

    5.

    Baudry, J. et al. 2000. Hedgerows: an international perspective on their origin, function and management. — J. Environ. Manage. 60: 7–22. Google Scholar

    6.

    Bellamy, P. E. et al. 2000. Road verges as habitat for small mammals in Britain. — Mamm. Rev. 30: 131–139. Google Scholar

    7.

    Bennett, A. F. and Saunders, D. A. 2010. Habitat fragmentation and landscape change. — In: Sodhi, N. S. and Ehrlich, P. R. (eds), Conservation biology for all. Oxford Univ. Press, pp. 88–106. Google Scholar

    8.

    Bentley, J. M. et al. 2000. Effects of fragmentation of Araucarian vine forest on small mammal communities. — Conserv. Biol. 14: 1075–1087. Google Scholar

    9.

    Broughton, R. K. et al. 2014. Agri-environment scheme enhances small mammal diversity and abundance at the farm-scale. — Agric. Ecosyst. Environ. 192: 122–129. Google Scholar

    10.

    Burnham, K. P. and Anderson, D. R. 2002. Model selection and multimodel inference — a practical information-theoretic approach. — Springer. Google Scholar

    11.

    Butet, A. and Delettre, Y. R. 2011. Diet differentiation between European arvicoline and murine rodents. — Acta Theriol. 56: 297–304. Google Scholar

    12.

    Butet, A. et al. 2006. Factors driving small rodents assemblages from field boundaries in agricultural landscapes of western France. — Landscape Ecol. 21: 449–461. Google Scholar

    13.

    Butet, A. et al. 2010. Responses of common buzzard (Buteo buteo) and Eurasian kestrel (Falco tinnunculus) to land use changes in agricultural landscapes of western France. — Agric. Ecosyst. Environ. 138: 152–159. Google Scholar

    14.

    Davies, Z. G. and Pullin, A. S. 2007. Are hedgerows effective corridors between fragments of woodland habitat? An evidencebased approach. — Landscape Ecol. 22: 333–351. Google Scholar

    15.

    Dormann, C. F. et al. 2013. Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. — Ecography 36: 27–46. Google Scholar

    16.

    Eccard, J. A. and Klemme, I. 2013. Reducing mortality of shrews in rodent live trapping — a method increasing live-trap selectivity with shrew exits. — Ann. Zool Fenn. 50: 371–376. Google Scholar

    17.

    Fahrig, L. and Merriam, G. 1985. Habitat patch connectivity and population survival. — Ecology 66: 1762–1768. Google Scholar

    18.

    Fischer, C. et al. 2011. Small mammals in agricultural landscapes: opposing responses to farming practices and landscape complexity. — Biol. Conserv. 144: 1130–1136. Google Scholar

    19.

    Fischer, C. et al. 2013. The impact of hedge—forest connectivity and microhabitat conditions on spider and carabid beetle assemblages in agricultural landscapes. — J. Insect Conserv. 17: 1027–1038. Google Scholar

    20.

    Fischer, J. and Lindenmayer, D. B. 2007. Landscape modification and habitat fragmentation: a synthesis. — Global Ecol. Biogeogr. 16: 265–280. Google Scholar

    21.

    Fitzgibbon, C. D. 1997. Small mammals in farm woodlands: the effects of habitat, isolation and surrounding land-use patterns. — J. Appl. Ecol. 34: 530–539. Google Scholar

    22.

    Gelling, M. et al. 2007. Are hedgerows the route to increased farmland small mammal density? Use of hedgerows in British pastoral habitats. — Landscape Ecol. 22: 1019–1032. Google Scholar

    23.

    Gentili, S. et al. 2014. Decreased small mammals species diversity and increased population abundance along a gradient of agricultural intensification. — Hystrix 25: 39–44. Google Scholar

    24.

    Grueber, C. E. et al. 2011. Multimodel inference in ecology and evolution: challenges and solutions. — J. Evol. Biol. 24: 699–711. Google Scholar

    25.

    Guivier, E. et al. 2011. Landscape genetics highlights the role of bank vole metapopulation dynamics in the epidemiology of Puumala hantavirus. — Mol. Ecol. 20: 3569–3583. Google Scholar

    26.

    Heroldová, M. et al. 2007. Structure and diversity of small mammal communities in agriculture landscape. — Agric. Ecosyst. Environ. 120: 206–210. Google Scholar

    27.

    Imholt, C. et al. 2015. Quantifying the past and future impact of climate on outbreak patterns of bank voles (Myodes glareolus). — Pest Manage. Sci. 71: 166–172. Google Scholar

    28.

    IUCN 2013. The IUCN red list of threatened species. Ver. 2013.1. — < www.iucnredlist.org>. Accessed 26 August 2013. Google Scholar

    29.

    Jacob, J. et al. 2014. Common vole (Microtus arvalis) ecology and management: implications for risk assessment of plant protection products. — Pest Manage. Sci. 70: 869–878. Google Scholar

    30.

    Jedrzejewska, B. and Jedrzejewski, W. 1990. Antipredatory behavior of bank voles and prey choice of weasels — enclosure experiments. — Ann. Zool. Fenn. 27: 321–328. Google Scholar

    31.

    Koks, B. J. et al. 2007. Do voles make agricultural habitat attractive to Montagu's harrier Circus pygargus? — Ibis 149: 575–586. Google Scholar

    32.

    Kotzageorgis, G. C. and Mason, C. F. 1996. Range use, determined by telemetry, of yellow-necked mice (Apodemus flavicollis) in hedgerows. — J. Zool. 240: 773–777. Google Scholar

    33.

    Kovacs-Hostyanszki, A. et al. 2013. Contrasting effects of massflowering crops on bee pollination of hedge plants at different spatial and temporal scales. — Ecol. Appl. 23: 1938–1946. Google Scholar

    34.

    Kozakiewicz, M. 1985. The role of habitat isolation in formation of structure and dynamics of the bank vole population. — Acta Theriol. 30: 193–209. Google Scholar

    35.

    Kozakiewicz, M. et al. 1993. Use of space by bank voles (Clethrionomys glareolus) in a Polish farm landscape. — Landscape Ecol. 8: 19–24. Google Scholar

    36.

    Kozakiewicz, M. et al. 2007. Long-distance movements of individuals in a free-living bank vole population: an important element of male breeding strategy. — Acta Theriol. 52: 339–348. Google Scholar

    37.

    Leps, J. and Smilauer, P. 2003. Multivariate analysis of ecological data using CANOCO. — Cambridge Univ. Press. Google Scholar

    38.

    Longland, W. S. and Price, M. V. 1991. Direct observations of owls and heteromyid rodents — can predation risk explain microhabitat use? — Ecology 72: 2261–2273. Google Scholar

    39.

    Mazurkiewicz, M. 1994. Factors influencing the distribution of the bank vole in forest habitats. — Acta Theriol. 39: 113–126. Google Scholar

    40.

    Michel, N. et al. 2006. How does landscape use influence small mammal diversity, abundance and biomass in hedgerow networks of farming landscapes? — Acta Oecol. 30: 11–20. Google Scholar

    41.

    Michel, N. et al. 2007. Role of habitat and landscape in structuring small mammal assemblages in hedgerow networks of contrasted farming landscapes in Brittany, France. — Landscape Ecol. 22: 1241–1253. Google Scholar

    42.

    Mortelliti, A. et al. 2011. Independent effects of habitat loss, habitat fragmentation and structural connectivity on the distribution of two arboreal rodents. — J. Appl. Ecol. 48: 153–162. Google Scholar

    43.

    Niemelä, J. 2001. Carabid beetles (Coleoptera: Carabidae) and habitat fragmentation: a review. — Eur. J. Entomol. 98: 127–132. Google Scholar

    44.

    Niethammer, J. and Krapp, F. 1978. Handbuch der Säugetiere Europas: Nagetiere I — Rodentia I. — Akad. Verlagsgesellschaft. Google Scholar

    45.

    Niethammer, J. and Krapp, F. 1982. Handbuch der Säugetiere Europas: Nagetiere II — Rodentia II. — Akad. Verlagsgesellschaft. Google Scholar

    46.

    Niethammer, J. and Krapp, F. 1990. Handbuch der Säugetiere Europas: Insektenfresser — Insectivora; Herrentiere — Primates. — AULA. Google Scholar

    47.

    Nordwestdeutsche Forstliche Versuchsanstalt 2008. Waldzustandsbericht 2008. — Printec Offset Kassel. Google Scholar

    48.

    Oksanen, J. et al. 2011. vegan: Community ecology package. R package ver. 1.17–11. — < http://CRAN.R-project.org/package=MuMIn >. Google Scholar

    49.

    Ouin, A. et al. 2000. Spatial dynamics of wood mouse (Apodemus sylvaticus) in an agricultural landscape under intensive use in the Mont Saint Michel Bay (France). — Agric. Ecosyst. Environ. 78: 159–165. Google Scholar

    50.

    Paillat, G. and Butet, A. 1996. Spatial dynamics of the bank vole (Clethrionomys glareolus) in a fragmented landscape. — Acta Oecol. 17: 553–559. Google Scholar

    51.

    Pardini, R. et al. 2005. The role of forest structure, fragment size and corridors in maintaining small mammal abundance and diversity in an Atlantic forest landscape. — Biol. Conserv. 124: 253–266. Google Scholar

    52.

    Petit, S. and Usher, M. B. 1998. Biodiversity in agricultural landscapes: the ground beetle communities of woody uncultivated habitats. — Biodivers. Conserv. 7: 1549–1561. Google Scholar

    53.

    Pinheiro, J. C. and Bates, D. M. 2000. Mixed-effects models in S and S-PLUS. — Springer. Google Scholar

    54.

    Pinheiro, J. et al. 2013. nlme: linear and nonlinear mixed effects models. — R package ver. 3.1–111. Google Scholar

    55.

    Pollard, E. and Relton, J. 1970. Hedges. V. A study of small mammals in hedges and cultivated fields. — J. Appl. Ecol. 7: 549–557. Google Scholar

    56.

    Shchipanov, N. A. et al. 2011. Two types of using of space in the resident common shrews Sorex araneus L. — Biol. Bull. 38: 92–97. Google Scholar

    57.

    Shonfield, J. et al. 2013. Reducing accidental shrew mortality associated with small-mammal livetrapping I: an inter- and intrastudy analysis. — J. Mammal. 94: 745–753. Google Scholar

    58.

    Silva, M. and Prince, M. E. 2008. The conservation value of hedgerows for small mammals in prince Edward Island, Canada. — Am. Midl. Nat. 159: 110–124. Google Scholar

    59.

    Stromgren, E. J. and Sullivan, T. P. 2014. Influence of pitfall versus Longworth livetraps, bait addition, and drift fences on trap success and mortality of shrews. — Acta Theriol. 59: 203–210. Google Scholar

    60.

    Szacki, J. 1987. Ecological corridor as a factor determining the structure and organization of a bank vole population. — Acta Theriol. 32: 31–44. Google Scholar

    61.

    Tattersall, F. H. et al. 2001. Habitat use by wood mice (Apodemus sylvaticus) in a changeable arable landscape. — J. Zool. 255: 487–494. Google Scholar

    62.

    Tattersall, F. H. et al. 2002. Is habitat linearity important for small mammal communities on farmland? — J. Appl. Ecol. 39: 643–652. Google Scholar

    63.

    Tewksbury, J. J. et al. 2002. Corridors affect plants, animals, and their interactions in fragmented landscapes. — Proc. Natl Acad. Sci. USA 99: 12923–12926. Google Scholar

    64.

    Vieira, M. V. et al. 2009. Land use vs fragment size and isolation as determinants of small mammal composition and richness in Atlantic Forest remnants. — Biol. Conserv. 142: 1191–1200. Google Scholar

    65.

    Voutilainen, L. et al. 2012. Environmental change and disease dynamics: effects of intensive forest management on Puumala hantavirus infection in boreal bank vole populations. — PloS ONE 7: e39452 Google Scholar

    66.

    Wang, M. and Grimm, V. 2007. Home range dynamics and population regulation: an individual-based model of the common shrew Sorex araneus. — Ecol. Modell. 205: 397–409. Google Scholar

    67.

    Weber, H. 2008. Gebüsche, Hecken, Krautsäume. — Ulmer. Google Scholar

    68.

    Yahner, R. H. 1983. Small mammals in farmstead shelterbelts — Habitat correlates of seasonal abundance and community atructure. — J. Wildl. Manage. 47: 74–84. Google Scholar

    Notes

    [1] Supplementary material (available online as Appendix wlb-00176 at < www.wildlifebiology.org/appendix/wlb-00176>). Appendix 1.

    © 2016 The Authors. This is an Open Access article This work is licensed under a Creative Commons AttributionNonCommercial-NoDerivatives 4.0 International License (CC-BYNC-ND) .
    Hella Schlinkert, Martin Ludwig, Péter Batáry, Andrea Holzschuh, Anikó Kovács-Hostyánszki, Teja Tscharntke, and Christina Fischer "Forest specialist and generalist small mammals in forest edges and hedges," Wildlife Biology 22(3), 86-94, (1 May 2016). https://doi.org/10.2981/wlb.00176
    Accepted: 1 November 2015; Published: 1 May 2016
    Back to Top