The distribution of Neotropical bats in heterogeneous environments has been studied by several authors (Mendes et al. 2014; Heim et al. 2015; Jaffé et al. 2018), and has resulted in interesting findings about their spatial distribution and habitat use (Bernard & Fenton 2003; Heer et al. 2015; Dias-Silva et al. 2018). Such studies have also served as source of knowledge for making decisions about priority areas for the conservation of Neotropical bats (Fahr & Kalko 2011; Gomes et al. 2015). The structure and composition of bat assemblages is mostly determined by features of the vegetation at local and landscape scales (Medellín et al. 2008; Bobrowiec et al. 2014; Marciente et al. 2015). Variation in local and landscape habitat attributes plays an important role in structuring assemblages of bats at the regional scale by increasing beta diversity among sites (Fahr & Kalko 2011; Marciente et al. 2015).
Different species of Neotropical bats respond differently to the landscape in which they are inserted (Fahr & Kalko 2011; Barros et al. 2014; Marciente et al. 2015). For example, in a study in Mexico (Avila Cabadilla et al. 2012), variation in phyllostomid assemblages and populations was associated with variation in local and landscape habitat attributes. In this study area, the abundance of nectarivorous bats tended to be negatively associated with the mean area of dry forest patches, while the abundance of frugivorous bats was positively associated with the percentage of riparian forest (Avila-Cabadilla et al. 2012). Studies have shown that bats in altered and urban landscapes also respond differently to environmental attributes (Heim et al. 2015; Kalda et al. 2015; Treitler et al. 2016). Agroforestry systems in Costa Rica, for example, may represent a good strategy for conserving bat species richness, but have a notable effect on the structure of their assemblages (Harvey & González Villalobos 2007). The activity of insectivorous bats in a Brazilian urban landscape (Araújo & Bernard 2016) was found to be significantly more associated with areas with vegetation, indicating that green remnants were hotspots for the maintenance of bat activity in the studied area.
Vegetation classes and a multiscale spatial approach must be considered when evaluating the responses of bats to variation in landscape attributes (Avila-Cabadilla et al. 2012; Kalda et al. 2015). Some species of bats may be more sensitive to landscape changes because of specialized diets and habitat uses (Klingbeil & Willig 2009), Thus, concern has arisen regarding species that play critical roles in ecosystem functioning by maintaining the structure and processes of ecosystems (Heim et al. 2015; Meyer et al. 2015; Araújo & Bernard 2016; Treitler et al. 2016).
Given this scenario, little is known about the responses of bat faunas to the presence of many exotic species of plants immersed in a landscape of natural vegetation. Thus, the occurrence and abundance of bats in an area with such characteristics, the Botanical Garden Inhotim, were investigated. Inhotim integrates the Rede Brasileira de Jardins Botânicos (RBJB), and possesses 140 ha of about 2,100 species of exotic plants (during time of this study 2008-09) from various parts of the world, and five small artificial lakes. The botanical garden is surrounded by approximately 145 ha of seasonal semideciduous forest typical of a region of transition between the Atlantic Forest and Cerrado biomes (Ab’Saber 1977) hereafter called “native forest”. The native forest is inserted in a private protected area, the Reserva Particular do Patrimônio Natural Inhotim (RPPN Inhotim). Our aim in this work was to assess whether the bat assemblages present in the native forest and the botanical garden differ by surveying their nocturnal activity.
The Botanical Garden Inhotim of the Instituto Cultural Inhotim is located in the municipality of Brumadinho (20°7’34.15”S, 44°13’15.54”W, Fig. S1), state of Minas Gerais, Brazil. Inhotim is located in a region known as the Iron Quadrangle, which is part of the Serra do Espinhaço complex located in the transition between the Atlantic Forest and Cerrado biomes (Ab’Saber 1977). The climate of the region is tropical altitudinal (Alvares et al. 2013) with dry and cold winters, and hot and wet summers.
Four sampling points were selected in each habitat (native forest and botanical garden) and data were collected on two nights in the botanical garden and two nights in the native forest monthly between May 2008 and April 2009, for a total of 48 nights, with 24 nights in each type of habitat and six nights at each of the four sampling points. Bats were captured using six mist nets (3 m x 14 m long; 20 mm mesh) each night. Nets were installed in locations of likely bat traffic, such as water features, paths, roads and rocky outcrops. Nets were set before sunset, left open for 6 hours (18:00-24:00 h), and checked every 20 minutes. Sampling points in the native forest were at least 1 km from sampling points in the botanical garden. The distance between sampling points in the native forest was at least 600 m, while that in the botanical garden was at least 300 m.
Captured animals were identified and marked, and biological data collected (length of forearm, body mass, sex, age and reproductive status) for inclusion in a database. Bats were marked with a numbered aluminum band attached to the forearm. Because of the size and weight of the bands, individuals weighing less than 8 grams were not marked. Specimens with questionable taxonomic identification were collected and identified (bibliografia) in the laboratory and deposited in the reference collection (Table S1) of the Programa de Pós-Graduação em Biologia de Vertebrados of the Pontifícia Universidade Católica de Minas Gerais.
The bat assemblages were represented by nonmetric multidimensional scaling (NMDS) and tested using non-parametric permutational multivariate analysis of variance (PERMANOVA). The percentage contribution of each species to the dissimilarity between the two study areas was generated through analysis of percentage similarity (SIMPER). These analyses used the number of individual captures for each species in each habitat type, and a Bray-Curtis index of similarity with 9 999 randomizations. Nights without captures were not considered. A generalized linear model (GLM) was performed to determine if bat species richness differed significantly between the two habitats. Analyses were performed in R (R Core Team 2017) using the Vegan package (Oksanen 2017), with the exception of NMDS, which was carried out using the software PAST (Hammer 2017).
One-hundred and ninety-one bats were captured belonging to 15 species of three families (Table 1).
The total sampling effort for the 48 nights of sampling was of 72,576 m 2/h, with 36,288 m 2/h in each habitat (native forest and botanical garden). A total of 81 individuals belonging to 12 species where captured in the botanical garden, whereas 110 individuals belonging to 14 species were captured in the native forest. There was only one recapture (Sturnira lilium), which occurred in the garden. Species richness did not differ significantly between the habitats (GLM, F=0.35, p=0.55).
Eleven species were recorded in both areas: Anoura caudifer, Artibeus lituratus, Carollia brevicauda, C. perspicillata, Desmodus rotundus, Eptesicus brasiliensis, Glossophaga soricina, Molossus molossus, Myotis nigricans, Platyrrhinus lineatus and S. lilium.
The composition of the bat fauna of the two habitats differed by only four species – Chrotopterus auritus, Phyllostomus discolor and Lasiurus blossevillii were recorded only in the native forest, and Molossops temminckii was recorded only in the botanical garden.
Despite the limited difference in bat species composition between the two habitats, the structure of the assemblages were noticeably different, as shown in Fig. 1, with only a slight overlap between the sampled points. The difference in the structure of assemblages was statistically significant (PERMANOVA, F=2.53, p=0.01). The total dissimilarity calculated by the SIMPER analysis (Fig. 2) was 81.56% , with S. lilium (20.99%) and C. perspicillata (19.01%) making the greatest contribution. Other species with notable contributions were P. lineatus (10.7%) and A.caudifer (8.15%), as well as the aerial insectivores E.brasiliensis (8.02%) and M. molossus (7.89%).
The bat species sampled represented 19% of the bat species known to occur in the state of Minas Gerais (Tavares et al. 2010) and 50% of the species recorded for the metropolitan region of Belo Horizonte (De Knegt et al. 2005; Bruno et al. 2011; Talamoni et al. 2013). The bat richness recorded in this study is similar to that reported by other studies conducted in the vicinity (Falcão et al. 2003; Gomes et al. 2015). The similarity in bat species richness and composition between the sampled habitats is indicative of two scenarios: (i) both habitats can support a similar number of species (Presley & Willig 2010); and (ii) the botanical garden is not acting as a barrier to the occurrence of most species of bats in the region. However, specific methodologies are needed to test these scenarios, such as radio telemetry, the search for roosts and the assessment of the availability of resources (Bernard & Fenton 2003; Ramos Pereira et al. 2010; Medellin et al. 2017).
Nonetheless, based on the capture data for each area, the assemblages of the forest and botanical garden have distinct structures . The exclusive occurrence of C. auritus and P. discolor in areas of native forest may indicate that these environments offer better roost conditions, such as caves and hollows of large trees (Kwiecinski 2006; Witt & Fábian 2010), and may provide a greater variety of food, since they are carnivorous-omnivorous animals (Bonato et al. 2004; Kwiecinski 2006). The species L. blossevillii is an aerial insectivorous bat with echolocation pulses that are adapted to forest environments (Schnitzler & Kalko 2001; Arias-Aguilar et al. 2018), thus favoring foraging in environments with more obstacles such as branches, lianas and trees that are characteristic of the studied native forest. On the other hand, although M. temminckii is an aerial insectivore, like L. blossevillii, it uses pulses of echolocation with greater plasticity, and thus successfully forages in both open areas, such as are present in the botanical garden, and in forest areas (Denzinger & Schnitzler 2013; López-Baucells et al. 2016).
The results presented here identify two main points responsible for the observed dissimilarity in PERMANOVA analysis between the assemblages of the two habitats: (i) the predominance of C. perspicillata (native forest 38, botanical garden 8) in areas of native forest, and S. lilium (native forest 17, botanical garden 25) and P. lineatus (native forest 4, botanical garden 12) in the botanical garden; and (ii) the greater capture of insectivores in the garden area (native forest 10, botanical garden 22). The reason for the differences in capture rate between C. perspicillata and S. lilium is not clear, as the availability of food resources and roosts in both areas was not evaluated. However, some hypotheses may be suggested. Studies conducted in southeastern and southern Brazil found C. perspicillata to have high abundances in forest areas and smaller abundances in modified areas such as urban and pasture areas (Gazarini & Pedro 2013; Muylaert et al. 2014). On the other hand, S. lilium and P. lineatus, although also abundant in forest areas, also occupy anthropic regions, such as regenerating areas and plantations (Souza et al. 2006; Nascimento et al. 2013). In addition, P. lineatus is known to use large palm leaves as roosts, as well as concrete buildings (Mendes et al. 2014), both of which occur in the studied botanical garden.
The greater insectivore capture rate in the garden was expected because insectivorous bats concentrate their feeding activity in places close to water bodies, mainly due to the availability of insects (Hintze et al. 2016; Dias-Silva et al. 2018). Thus, the presence of artificial lakes in the botanical garden were expected to favor the capture of these bats. However, the activity patterns of insectivorous bats are best investigated using bioacoustical methods (O’Farrell et al. 1999; Macswiney G et al. 2008).
It is clear that the establishment of the garden of exotic plants and the construction of artificial lakes played a crucial role in structuring the local bat assemblage, or at least promoted distinct patterns during the nocturnal activities of the bat species present. The results of this study demonstrate that bats respond differently to variation in available habitats even on a small spatial scale.
Materiales suplementarios
Supplement 1Table S1. Specimens deposited in the reference collection of the Programa de Pós-Graduação em Biologia de Vertebrados, PUC Minas.Fig. S1. Study area, where: Green markers represent sampling sites in the native forest and blue markers represent sampling sites in the botanical garden.