The value of forest fragments for maintaining amphibian diversity in Madagascar

Biological Conservation 191 (2015) 707–715

Contents lists available at ScienceDirect

Biological Conservation journal homepage: www.elsevier.com/locate/bioc

The value of forest fragments for maintaining amphibian diversity in Madagascar Jana C. Riemann a,⁎, Serge H. Ndriantsoa b, Noromalala R. Raminosoa b, Mark-Oliver Rödel c, Julian Glos a a b c

Zoological Institute, Animal Ecology and Conservation, University of Hamburg, Martin-Luther-King Platz 3, 20146 Hamburg, Germany Département de Biologie Animale, Université d'Antananarivo, Antananarivo 101, Madagascar Museum für Naturkunde, Leibniz Institute for Evolution and Biodiversity Science, Invalidenstr. 43, 10115 Berlin, Germany

a r t i c l e

i n f o

Article history: Received 24 February 2015 Received in revised form 11 June 2015 Accepted 9 August 2015 Available online xxxx Keywords: Habitat fragmentation Rainforest Amphibian conservation Species richness Species composition Disturbance

a b s t r a c t Forest fragmentation often causes biodiversity loss, but there is no consistent pattern on species' reactions. Considering the alarming rate of deforestation in the tropics, and the fact, that large areas of protected continuous forest are limited, it becomes increasingly important to determine the biodiversity value of fragmented forests. In order to investigate fragmentation effects on rainforest frogs in Madagascar and to assess the conservation value of these fragments, we analyzed amphibian diversity in a continuous rainforest and nearby forest fragments. We hypothesized that species richness is lower in fragments compared to continuous forest, and that fragmentation leads to altered assemblage composition. We found no fragmentation effects on species richness, demonstrating that fragments may maintain local species richness comparable to continuous forest. The presence of streams was the most important factor for high species richness, independent of fragmentation status. However, we detected fragmentation effects on species composition. As expected, several species were restricted to continuous forest, but many species occurred in both forest types, and some species were only found in fragments. Rainforest amphibians in our study area were less sensitive to fragmentation than expected. Adaptations to natural disturbances like cyclones could be one reason to explain this. However, as some species exclusively occurred in continuous forest and species composition differed between continuous forest and fragments, we conclude that fragments cannot substitute continuous forest blocs, but are generally important for maintaining amphibian diversity in Madagascar, especially if they comprise streams. Forest fragments should hence be included in conservation planning. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Habitat loss and fragmentation are major threats to biodiversity (Dirzo and Raven, 2003). Forest fragmentation is a process resulting in the decrease of total habitat amount, an increasing number of smallersized and isolated habitat patches, and an increasing ratio of edge to interior habitat. Generally, forest fragments exhibit severe ecological changes including species extinctions and altered ecosystem functions (Laurance et al., 2011). However, reactions to anthropogenic habitat alterations and fragmentation can differ markedly between species, taxonomic groups and ecosystems (Fahrig, 2003; Gardner et al., 2009; Irwin et al., 2010; Laurance et al., 2011). Deforestation in the tropics proceeds at an alarming rate (Hansen et al., 2013). Protected areas are limited in area and connectedness, and their current coverage fails to protect overall global biodiversity

⁎ Corresponding author. E-mail addresses: [email protected], [email protected] (J.C. Riemann), [email protected] (S.H. Ndriantsoa), [email protected] (N.R. Raminosoa), [email protected] (M.-O. Rödel), [email protected] (J. Glos).

http://dx.doi.org/10.1016/j.biocon.2015.08.020 0006-3207/© 2015 Elsevier Ltd. All rights reserved.

including numerous threatened species (Rodrigues et al., 2004). The forest cover outside protected areas has declined markedly in tropical forests since the 1980s (DeFries et al., 2005). Considering the ongoing agricultural expansion in the tropics (Laurance et al., 2014), it can be assumed that fragmented forest will become the dominant forest type in human altered tropical landscapes in the future. To be able to maintain overall tropical biodiversity in the long-term, it is hence essential to determine the biodiversity and conservation value of human-modified landscapes (Daily, 2001; Gardner et al., 2009; Irwin et al., 2010). So far, most attention has been drawn to secondary habitats. There is increasing evidence that secondary forests and fragments may have the potential to contribute to biodiversity conservation (Barlow et al., 2007; Gillespie et al., 2012; Mendenhall et al., 2014). However, there is still a considerable lack of knowledge concerning the degree to which tropical biodiversity can persist in human-modified landscapes (Gardner et al., 2009). More than one third of extant amphibian species are currently considered threatened (Stuart et al., 2004). Various, often interacting factors have been identified as causes of this global amphibian crisis, habitat loss and alteration belonging to the most severe causes (Sodhi et al., 2008; Stuart et al., 2004). Accordingly, the majority of fragmentation

708

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

studies addressing amphibians so far revealed negative effects on their diversity (e.g., Bell and Donnelly, 2006; Cabrera-Guzmán and Reynoso, 2012; Vallan, 2000). However, increased amphibian diversity in forest fragments has also been observed (Tocher et al., 1997). Madagascar is one of the world's biodiversity hotspots with an outstanding degree of endemism (Myers et al., 2000). Amphibian species richness, including many so far undescribed candidate species, is expected to comprise up to 465 native and endemic species (Vieites et al., 2009). Madagascar's rich and unique ecosystems are threatened by high rates of deforestation and the remaining forest cover is highly fragmented (Green and Sussman, 1990; Harper et al., 2007). A recent review of species' responses to anthropogenic disturbances in Madagascar revealed that overall very little is known and responses differ even within lower taxonomic levels and between ecoregions (Irwin et al., 2010). Our study aims to contribute to a better understanding of the response of highly diverse tropical amphibian assemblages to habitat fragmentation for the implementation of future conservation strategies with special emphasis on the conservation value of forest fragments. We examined patterns of amphibian diversity in a continuous rainforest and nearby forest fragments to reveal fragmentation effects on rainforest frogs in Madagascar. A fragmented landscape where a relatively large continuous forest part that can act as control site is still present is an ideal model system to learn about the value of forest fragments as amphibian habitats. In particular, we compared species richness between forest fragments (b 20 ha) and continuous forest (non-fragmented area of Ranomafana National Park, N40,000 ha), and evaluated patterns of assemblage composition in both forest types. We investigated stream and terrestrial habitats to equally account for stream depending species (either semiaquatic or stream breeding species) and for species that are completely independent from running waters. We hypothesized for both habitat types that 1) species richness is lower in forest fragments compared to continuous forest, and that 2) fragmentation leads to altered assemblage composition in forest fragments. 2. Material and methods 2.1. Study system Field work was conducted in the Ranomafana National Park (RNP, 21°02′–21°24′S, 47°20′–47°35′E), East Madagascar, and in forest fragments located east of RNP (Fig. 1). RNP comprises 43,500 ha of continuous mid-altitude montane rainforest (500–1300 m a.s.l.) with an annual precipitation between 1700 and 4300 mm (Wright and Andriamihaja,

2003). It provides most of the remaining rainforest habitat in the Ranomafana region. The remaining forest fragments around RNP are embedded in a matrix of cultivated land (slash and burn agriculture; “tavy”), clear cut, and secondary bush and shrub vegetation. We chose nine different forest fragments that range in size between two and 16.5 ha (Fig. 1, Appendix A). Five forest fragments comprise streams, including one fragment with two streams. Aerial photographs from 1950 showed that all but two of the studied fragments were separated from continuous forest by that time already and interviews with local people revealed that all studied fragments were at least 50 years separated from RNP. The Ranomafana region corresponds to one of the centers of amphibian diversity within Madagascar with almost 120 taxa known (Glaw and Vences, 2007; Vieites et al., 2009; own unpubl. data). 2.2. Sampling design We determined species richness and composition on transects distributed along streams and in terrestrial forest parts inside RNP and forest fragments. We included stream and terrestrial habitats to equally account for stream depending species (either semiaquatic or stream breeding) and for species that reproduce independent from streams (i.e., phytotelmata, pond or terrestrial breeders). As not all studied fragments comprise streams, data from stream and terrestrial transects were analyzed separately. In the following, we refer to stream and terrestrial transects as habitat types, and continuous forest and forest fragments as forest types. We established a total of 38 independent line transects (50 × 2 m; Marsh and Haywood, 2010): 22 transects were located inside RNP (control sites) and represented continuous forest (terrestrial: 11, streams: 11), and 16 transects were spread over nine different forest fragments (terrestrial: 10, streams: 6). Following the sampling scheme of terrestrial transects (searching a band of 2 m width), stream transects included one meter riparian vegetation on each stream bank in addition to the water body. Terrestrial transects were at least 50 m apart from the next stream. We kept a minimum distance of 200 m between transects of the same habitat type, and stream transects had no direct upstream connections. Transects were geographically spread over RNP as far as possible according to accessibility and logistic constraints to control for geographic distances between fragments and transects located within RNP. Transects inside forest fragments followed the topography of the respective fragment and were initially at least 50 m apart from the next forest edge, except two stream transects that were about 25 m from the next forest edge.

Fig. 1. Schematic view of the study area and its position on Madagascar (insert). Shown are study sites (black stars) inside Ranomafana National Park (gray area; black line: park border) and studied forest fragments (dark gray areas).

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

We conducted standardized visual and acoustic transect sampling (Rödel and Ernst, 2004), visiting each transect nine times. Visual sampling was particularly useful for terrestrial leaf litter species, acoustic sampling was particularly needed for the sampling of arboreal species and species calling from hidden places, such as tree holes, phytotelmata or dense leaf litter. Since the suitability of both methods differs between species, a combination of acoustic and visual methods is recommended to detect all major ecological guilds of frogs (Rödel and Ernst, 2004; Vences et al., 2008). We may however have missed species with weak calls inhabiting tree holes, Pandanus or Ravenala trees, if those habitats where not present on or close to a transect. We refrained from using drift fences with pitfall traps since this method is the least efficient one for sampling rainforest amphibians (Rödel and Ernst, 2004; Rosa et al., 2012). Drift fences with pitfalls are useful when particularly targeting fossorial species (Rödel and Ernst, 2004), but although well investigated, fossorial amphibian species are unknown from RNP (Vieites et al., 2009). After an initial training phase from February to April 2010, transects were sampled by two teams, led by JCR and NSH, each accompanied by one assistant. All transects were sampled alternately by both teams to avoid a potential observer bias. Forest types were sampled alternately to avoid temporal effects on sampling results. Transects were sampled during day and night as in the study area nocturnal as well as diurnal species occur (Glaw and Vences, 2007). Each transect was surveyed twice (once by day, once by night) between May and June 2010, which corresponds to a drier period in the year. In 2011 all transects were visited seven times (three times by day, four times by night), from the beginning of the heavy rainy season in January until the drier season in June. Visual sampling covered all visible individuals from leaf litter up to arboreal structures in approximately 2.5 m height and in addition the aquatic habitat on stream transects. Transects were walked at constant speed (approx. 2.5 m/min). We refrained from displacing logs and rocks and from pulling apart vegetation (e.g. Pandanus leafs) to keep disturbances of the study system as low as possible and to ensure equal sampling effort in all sites. All visually detected individuals were captured, measured, sexed, marked via toe clipping (no functionally important toes, following the recommendations by Grafe et al., 2011) and identified by morphology in the field (according to Glaw and Vences, 2007). Afterwards, all individuals were immediately released at point of capture, except some individuals that were taken as vouchers (see Appendix B). Visual transect sampling was interrupted for the duration of handling time and for acoustic sampling at four fixed points (start, 12.5 m, 25 m, 37.5 m). At each point we registered all calling individuals in a 12.5 m distance straight, left and right for 5 min and identified them to species level. The size of the acoustic transect was chosen according to the ability to determine each species' advertisement call, as calls from greater distances cannot be precisely identified (Rödel and Ernst, 2004). Team members learned to distinguish the advertisement calls during the initial training phase (according to Vences et al., 2006). In addition, own reference data was recorded for several species with a Roland EDIROL R-09 recorder and a Sennheiser MKE 400 microphone (sampling frequency: 44.1 kHz, recording mode: wav 24 bit), including previously unknown calls (see e.g. Riemann et al., 2012). The call recordings are stored at the Institute of Animal Ecology and Conservation at Hamburg University. Voucher specimens were euthanized in a chlorobutanol solution, preserved in 75% ethanol, and deposited at the Museum für Naturkunde Berlin, Germany (ZMB) and the Department of Animal Biology, University of Antananarivo, Madagascar (UADBA). All toe clips were collected as tissue samples and stored in pure ethanol for DNA barcoding (or dried and stored for further analyses). Barcoding, based on a fragment of the mitochondrial 16S rRNA gene, was used to confirm identification to species level (Vences et al., 2005, 2008). We used primers 16SFrogL1 (CAT AAT CAC TTG TTC TTT AAA) and 16SFrogH1 (GAT CCA ACA TCG AGG TCG) of Vences et al. (2010) and followed the molecular protocol

709

described in Ndriantsoa et al. (2013). We determined sequences at least once for species that could be identified by a distinct morphology. Concerning highly cryptic species (e.g. taxa of the subgenera Chonomantis and Ochthomantis) and supposed undescribed taxa, barcoding was repeated for each transect sampling. Obtained sequences were compared to published data on GenBank or own reference sequences (see Appendix B for collection and GenBank accession numbers). 2.3. Data analyses We compared species richness (absolute number of species detected) per transect between continuous forest and forest fragments for terrestrial and stream habitats using Wilcoxon rank sum tests. We further compared patterns of species composition between both forest types. A reliable identification of the effects of fragmentation on species composition requires an understanding of spatial effects, i.e., the spatial structuring of assemblages. Spatial turnover occurs at all spatial scales and species assemblages can be spatially structured even without a spatial environmental signal (Beck and Khen, 2007; Ernst and Rödel, 2008). Since all transects representing continuous forest were located within the same connected forest patch, and due to land-use patterns all studied forest fragments were located east of RNP (Fig. 1), spatial effects could not be ruled out. If spatial structuring occurs, it can be difficult to disentangle fragmentation effects from spatial effects. We used Mantel tests to analyze whether differences in assemblage composition among transects are related to spatial (i.e., geographic) distance. We tested correlations between compositional differences and spatial distance among pairs of transects within each forest type, i.e., within continuous forest and among fragments, and across the whole data set. If correlations within forest types and across forest types are different, this would support the hypothesis that fragmentation effects exist and compositional differences are not caused by spatial distance alone (Ghazoul, 2002; Ramage et al., 2013). Distance matrices were calculated based on Bray–Curtis dissimilarities using species presence/absence data for compositional data and the Euclidean distance for spatial distances calculated from geographic coordinates of transects. Mantel tests were performed with function “mantel” from R package vegan (Oksanen et al., 2011) based on Pearson's product–moment correlation. P-values were obtained from 4999 permutations. We used non-metric multidimensional scaling (NMDS) to visualize and evaluate patterns of dissimilarity among transects based on their species composition. The ordination was constructed from a Bray– Curtis dissimilarity matrix using species presence/absence data. NMDS was performed with function “metaMDS” from R package vegan (Oksanen et al., 2011). This method uses random starting configurations to find a stable global solution. We then performed permutational multivariate analyses of variance (perMANOVA) (Anderson, 2001; McArdle and Anderson, 2001) to test the hypothesis of differences in species composition between continuous forest and forest fragments for terrestrial and stream habitats. This non-parametric permutation based variant of MANOVA partitions sums of squares of multivariate data equivalent to univariate ANOVA and the pseudo F statistic can be calculated directly from any distance measure (Anderson, 2001). We performed perMANOVA based on Bray–Curtis dissimilarities using species presence/absence data with function “adonis” from R package vegan (Oksanen et al., 2011). P-values were obtained from 999 permutations. All statistical analyses were performed in R (R Development Core Team, 2011). 3. Results 3.1. Species richness Transect sampling revealed a total of 57 species (Table 1) represented by 1026 individuals that were detected visually, complemented by

710

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

Table 1 Species detected per habitat type (CF: continuous forest, Ranomafana National Park; FF: forest fragments; t: terrestrial, s: stream habitats; TS: threat status according to IUCN red list criteria (IUCN, 2014), NA: not assessed, DD: data deficient, LC: least concern, NT: near threatened, EN: endangered). Names in parentheses refer to candidate species according to Vieites et al. (2009). Species

CF t

Hyperoliidae Heterixalus alboguttatus Heterixalus betsileo Mantellidae Aglyptodactylus sp. aff. madagascariensis (A. sp. 3) Boophis albilabris Boophis cf. albipunctatus Boophis arcanus Boophis calcaratus Boophis elenae Boophis luciae Boophis luteus Boophis madagascariensis Boophis cf. majori Boophis marojezensis Boophis picturatus Boophis pyrrhus Boophis quasiboehmei Boophis reticulatus Gephyromantis asper Gephyromantis boulengeri Gephyromantis decaryi Gephyromantis enki Gephyromantis plicifer Gephyromantis redimitus Gephyromantis sculpturatus Gephyromantis tschenki Gephyromantis ventrimaculatus Guibemantis liber Guibemantis cf. depressiceps Guibemantis tornieri Mantidactylus aerumnalis Mantidactylus alutus Mantidactylus betsileanus Mantidactylus femoralis Mantidactylus grandidieri Mantidactylus majori Mantidactylus melanopleura Mantidactylus sp. aff. mocquardi (M. sp. 44) Mantidactylus paidroa Mantidactylus sp. aff. betsileanus “slow calls” (M. sp. 28) Mantidactylus sp. aff. biporus “Ranomafana” (M. sp. 24) Mantidactylus sp. aff. charlotteae “Ambohitsara” (M. sp. 9) Mantidactylus sp. aff. charlotteae “Ranomafana” (M. sp. 13) Mantidactylus sp. aff. cowanii “small” (M. sp. 48) Mantidactylus sp. aff. mocquardi “Ambatolahy” (M. sp. 47) Mantidactylus sp. aff. mocquardi “Namorona” (M. sp. 64) Mantidactylus sp. aff. opiparis (M. sp. 58) Mantidactylus tricinctus Spinomantis sp. aff. fimbriatus (S. sp. 2) Microhylidae Anodonthyla boulengeri Anodonthyla moramora Platypelis grandis Platypelis pollicaris Platypelis tuberifera Plethodontohyla brevipes Plethodontohyla inguinalis Rhombophyrne mierya Rhombophryne sp. “Ranomafana”a (Stumpffia sp. 9)

FF s

t

TS

fragments we found 40 species in total, 13 in terrestrial habitat and 38 in stream habitat. We recorded no effects of forest fragmentation on species richness as we found no differences in species richness per transect between continuous forest and forest fragments in stream habitats (Wilcoxon rank sum test, W = 20, p = 0.204, n = 17) and terrestrial habitats (W = 74, p = 0.183, n = 21; Fig. 2).

s

− − + + LC − + + + LC

3.2. Species composition

− − − − − − − − + − − − − − − + + − + + − + + − − − − − − − − − − + − − − − − − − − − + − −

− + − − − + + + + − + + − + + − + + + + + + + − + + + − + + + + + + + + + + + − + + − + − +

+ + − − − − − − + − − − − − − − + − + − − + + − − − − − − − − − − − − − − − − − − − − − − −

− + + + + + − + + + − − + + − − + − + − + + + + + − + + − + + + − + + − + + + + + + + − + +

LC LC LC NA NA DD NA LC LC NT LC LC LC NA LC LC LC NT DD NT LC LC DD LC LC LC LC LC LC LC LC LC LC LC LC NA NA NA NA NA NA NA NA NA DD NA

We detected 26 species that occurred in continuous forest and forest fragments, but 31 species were found in only one forest type (17 species in continuous forest and 14 species in forest fragments; Table 1). Accordingly, overall about two thirds (40) of the detected species occurred in forest fragments. Large and medium-sized microhylids of the genera Platypelis and Plethodontohyla were restricted to continuous forest, whereas small species of Anodonthyla and Rhombophryne occurred in both forest types or only in fragments (Table 1). The species-rich mantellid genera Boophis, Gephyromantis and Mantidactylus all contained several species that were found in both forest types as well as species restricted to continuous forest or forest fragments. Two of the three Guibemantis species were found in both forest types, the other only in continuous forest. Thus, a distinct phylogenetic signal between inhabitants of continuous forest and fragments could not be observed (Table 1). Most of the species found in terrestrial habitats also occurred in stream habitats (Table 1). Only two species in each forest type were restricted to terrestrial habitat (RNP: Platypelis tuberifera, Gephyromantis asper; fragments: Anodonthyla moramora, Aglyptodactylus madagascariensis). One of these species found only in terrestrial habitat in fragments also occurred in stream habitat in continuous forest (A. moramora). We found a significant correlation between compositional differences and spatial distance in the whole data set for stream habitats (r = 0.647, p b 0.0001) and terrestrial habitats (r = 0.465, p b 0.0001). We further detected a significant correlation between compositional differences and spatial distance within continuous forest in both habitat types (stream: r = 0.43, p = 0.011; terrestrial: r = 0.577, p = 0.003). However, we found no correlation between compositional differences and spatial distance among forest fragments (stream: r = 0.279, p = 0.112; terrestrial: r = 0.003, p = 0.478), showing that assemblages in fragments were not spatially structured. These contrasting patterns of spatial structuring in continuous forest and forest fragments

+ + − + + + + − −

+ + + + − + + − −

+ + − − − − − + +

+ − − − − − − + +

NA DD LC DD LC EN LC NA NA

a RNP taxa formerly assigned to the paraphyletic genus Stumpffia, see Peloso et al. (2015).

2140 acoustic records. Generally, species richness was higher in stream habitats than in terrestrial habitats, independent of forest type. In continuous forest (RNP) a total of 43 species were detected, with 15 species found in terrestrial habitat and 41 species in stream habitat. In forest

Fig. 2. Species richness per transect for stream (s) and terrestrial (t) habitats in continuous forest (CF, gray boxes) and forest fragments (FF, white boxes). Shown are median, interquartile range (box), and location of the minimum and maximum (whiskers). Species richness was always higher in stream habitats. No differences in species richness could be detected between continuous forest and forest fragments, neither in stream (W = 20, p = 0.204, n = 17) nor in terrestrial habitats (W = 74, p = 0.183, n = 21).

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

indicated that compositional differences were not caused by spatial distance alone but forest fragmentation affected assemblage composition. We found significant differences in species composition between continuous forest and forest fragments in stream habitats (perMANOVA, pseudo F [1,15] = 6.38, p = 0.001) and terrestrial habitats (pseudo F[1,19] = 11.65, p = 0.001). NMDS revealed distinct patterns of dissimilarities in species composition between continuous forest and forest fragments in both habitat types (Fig. 3). The stress value (0.146) of the final two-dimensional solution indicated a reasonable preservation of ordering relationships of the multidimensional among-transect dissimilarities. Terrestrial and stream habitats were separated along axis one, which was basically an effect of lower species richness in terrestrial compared to stream habitats in both forest types. Forest types, i.e., continuous forest and forest fragments, were separated along axis two. As species richness was equal in both forest types in stream and terrestrial habitats respectively, this separation visualizes the significant differences in species composition between continuous forest and forest fragments. 4. Discussion Ongoing habitat loss and fragmentation are threatening the globally outstanding biodiversity of Madagascar (Allnutt et al., 2008) and today's protected areas might not be sufficient to protect all extant species in the long-term. It is hence especially important to understand fragmentation effects on biodiversity and assess the diversity and conservation value of disturbed habitats (Irwin et al., 2010). We found no effects of forest fragmentation on species richness, which was unexpected since the majority of fragmentation studies so far revealed negative effects on amphibian species richness (e.g., Bell and Donnelly, 2006; Bickford et al., 2010; Cabrera-Guzmán and Reynoso, 2012; Pineda and Halffter, 2004), including former studies in Madagascar (Lehtinen and Ramanamanjato, 2006; Vallan, 2000). These former findings mainly reflect patch size effects, being an important aspect of habitat fragmentation. It is well proven that patch size affects amphibian species richness as smaller fragments usually contain less species than larger ones (e.g., Bell and Donnelly, 2006; Cabrera-Guzmán and Reynoso, 2012; Lehtinen and Ramanamanjato, 2006; Vallan, 2000). Those species-area relationships are well established and consistent through different taxa in forest fragments in the tropics (Hill et al., 2011). Tocher et al. (1997) likewise revealed that amphibian species richness decreased with decreasing fragment size, but found higher species richness in fragments than in primary

Fig. 3. Non-metric multidimensional scaling (NMDS) showing differences in species composition between transects in stream (s) and terrestrial (t) habitats in continuous forest (CF, gray triangles) and forest fragments (FF, white triangles) based on Bray–Curtis dissimilarity index using species presence/absence data (stress = 0.146). Distances between transects in the two-dimensional NMDS plot represent dissimilarities in species composition. Transects are grouped according to forest and habitat type.

711

forest samples of the same habitat amount. We found that the same habitat amount (transect area) in forest fragments harbored equal numbers of frog species than continuous forest. Our result thus indicated that effects of fragmentation per se, i.e., the breaking apart of habitat, are weaker than expected and may have only minor effects on species richness, if any, compared to habitat loss (Fahrig, 2003). Therefore, all remaining forest patches (independent of patch size) that still provide suitable habitat may have an important role in maintaining amphibian diversity and should be considered in conservation planning. The diversity of forest fragments may also highly depend on the structure and usability of the matrix, as it might act either as corridor or barrier for dispersal. Several species associated with forest habitat are able to use some matrix habitats (e.g. Gascon et al., 1999). The vicinity to continuous forest areas, that harbor potential source populations, may be also an important factor affecting species richness in forest fragments. Thus, highly isolated fragments or fragments located in a more hostile matrix preventing potential source-sink dynamics might harbor less species than the originally forested area per habitat amount. However, even small and isolated fragments can maintain populations of matrix avoiding species if they contain the necessary habitat structures for reproduction (Dixo and Metzger, 2010). Important factors for the maintenance of amphibian diversity in forest fragments are vegetation structure, structural heterogeneity and especially the availability of breeding habitats (Bickford et al., 2010; Hillers et al., 2008; Pineda and Halffter, 2004; Vallan, 2000). In our study the presence of a stream was an important factor for high species richness, independent of the surrounding forest type. Surprisingly, this also applied for species that are not directly stream depending, i.e., not stream breeding or semiaquatic. Most of the species found in terrestrial habitats also occurred in stream habitats. Some species may depend on both habitat types, e.g., species that reproduce in streams but inhabit terrestrial forest parts outside the mating season. Other species (e.g., direct developers or microhylid species that reproduce in water-filled tree holes) may occur in the vicinity of streams by chance, as their distribution may not be affected by the presence of a stream. However, even for those species that are not directly stream dependent the vicinity to a stream could be important, as microclimatic factors may be influenced by streams. Forest fragmentation can lead to higher temperatures, increased wind speed and decreased relative humidity near edges in fragments and hence changes in microclimatic conditions which affect many amphibian species (Lehtinen et al., 2003). Streams could buffer against microclimatic changes and hence forest fragments comprising streams are especially important for maintaining amphibian diversity. Although species richness was not affected by forest fragmentation, our results indicated that amphibian assemblages were influenced by fragmentation. We found different patterns of spatial structuring in continuous forest and forest fragments, i.e., assemblages in continuous forest were spatially structured but assemblages in fragments were not. Spatial structuring of species assemblages is a common pattern (Ernst and Rödel, 2008; Hillers et al., 2008) and could hence be expected in an intact system. The contrasting patterns of spatial structuring in continuous forest and forest fragments supported the hypothesis that fragmentation affected species composition and compositional differences between continuous forest and forest fragments were not caused by spatial distance alone. As expected, some species were restricted to continuous forest habitats. Large and medium-sized microhylids (genera Platypelis and Plethodontohyla) seemed to react most sensitive to fragmentation as they were never found in fragments. Small fragment sizes in our study could be one reason. Vallan (2000) reported a minimum patch size of 30 ha for microhylids in the highlands due to unfavorable microclimatic conditions in smaller fragments. However, small microhylids (genera Anodonthyla and Rhombophryne) occurred in both forest types or even predominantly in fragments. A large body size has been identified as one trait associated with high extinction risks in frogs (Lips et al., 2003; Sodhi et al., 2008). Nevertheless, this was only true for

712

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

microhylids in our study whereas the largest mantellid species (e.g., Boophis albilabris, Boophis madagascariensis, Mantidactylus grandidieri, Mantidactylus sp. aff. mocquardi) were also common in forest fragments. Andreone (1994) hypothesized that arboreal and stream-dwelling species may not depend on the microclimatic conditions on the forest floor and may adapt to disturbed environments. Indeed, many arboreal (genera Boophis, Guibemantis) and streamdwelling (genus Mantidactylus) mantellid species found in our study seemed to be tolerant against fragmentation. However, some of these species were also restricted to continuous forest. On the other hand, several leaf litter or low vegetation associated Gephyromantis species also occurred in fragments. In brief, it is not clear which factors and traits predispose for fragmentation sensitivity. It is important to point out that we did not only detect species that were restricted to continuous forest or occurred in both forest types, but also species that were only found in fragments. It is known from other tropical regions, that forest fragments are often invaded by open habitat species (e.g., Gascon et al., 1999; Hillers et al., 2008). This explained higher species richness in fragments compared to continuous forest in the study of Tocher et al. (1997) and might explain equal species richness in our study. However, only two species (genus Heterixalus) that were exclusively found in fragments are usually associated with open habitat (Glaw and Vences, 2007). The eastern rainforest belt of Madagascar was originally completely forested (Green and Sussman, 1990) and thus the majority of species in our study region should be typical forest species. Nevertheless, frequent cyclone disturbances influence forest structure (de Gouvenain and Silander, 2003; Birkinshaw and Randrianjanahary, 2007), microclimate (Turton and Siegenthaler, 2004), and tree biology (Ganzhorn, 1995), and thus may have also favored amphibian adaptations to disturbed habitats. Consequently, species adapted to natural disturbances may have a higher ability to cope with anthropogenic disturbances. This could be one potential reason for the resistance of many Malagasy rainforest amphibians to fragmentation, i.e. about two third of our recorded species occurred in forest fragments. System specific factors like cyclones could thus explain differences in organism's reactions to anthropogenic disturbances between study regions. Species that were tolerant against fragmentation or even exclusively occurred in fragments most likely adapted best or may be even specialized on disturbed habitat and are therefore able to exist in forest fragments. In contrast, disturbance sensitive species or species with specific habitat or microclimatic requirements that are not met in forest fragments become locally extinct. Of course, some of the species only detected in fragments may also occur in parts inside RNP that were not sampled, but some may be indeed restricted to fragmented areas. Given the fact that assemblages naturally tend to be spatially structured (Ernst and Rödel, 2008; Hillers et al., 2008), as proven in the continuous forest part in our study, species that exclusively occurred in forest fragments may also represent remnant populations. If a species' major distribution area was located in parts of the original forest cover that was mainly deforested except of one or some remnant forest patches that represent today's forest fragments, it might be even range-restricted to the forest fragment(s). In this light, the conservation of forest fragments becomes even more important. In many highly deforested areas on Madagascar small forest fragments represent the last remaining habitats for amphibians (Andreone et al., 2005, 2008). It remains unknown whether populations in forest fragments represent viable populations in the long-term and whether matrix quality and structure allow for migration and genetic exchange between populations in different fragments or fragments and continuous forests. Since frogs have relatively short generation times and all studied fragments were separated from continuous forest for at least fifty years, we assume that the observed patterns for species richness and composition in our study will persist over time. However, more information about the genetic diversity of fragment populations and matrix effects is urgently needed.

5. Conclusion Since forest fragmentation altered assemblage composition, forest fragments cannot substitute continuous forest. It is well proven that large continuous primary forests are indispensable for protecting and retaining tropical biodiversity and ecosystem functions (Barlow et al., 2007; Laurance et al., 2011). However, remaining forests are under enormous pressure all over the tropics due to ongoing deforestation (Hansen et al., 2013) and habitat conversion for agriculture (Laurance et al., 2014). Therefore, the maintenance or enhancement of overall habitat amount is an important consideration for conservation planning (Villard and Metzger, 2014). We conclude that forest fragments are important refuges of amphibian diversity in Madagascar, especially if they comprise streams. Even small forest fragments can be vital complements to a network of larger protected areas. Accordingly, forest fragments should be included to a greater extent in conservation planning. It should be noted that forest fragments are often also the last available forest parts for local populations, who highly depend on natural forest resources (Brown et al., 2011; Ferraro, 2002). Hence, strategies for sustainable natural resource use are urgently needed to successfully protect biodiversity in remaining forest fragments and to ensure their long-term existence. Acknowledgments We thank E. Larney and the team of Centre ValBio research station for their logistic support, and especially our research assistants A. Telo and J. Solo for their help in the field. This study would not have been possible without the permission of several forest owners and villages to work in their forest fragments. We thank M. Vences, A. Strauss, and M. Kondermann for helping with species identification via DNA barcoding. We are grateful to Madagascar National Parks and the Ministère de l'Environnement, de l'Ecologie, de la Mer et des Forêts for research and export permits (017/033/10/MEF/SG/DGF/DCD.SAP/ SLRSE, 003/004/11/MEF/SG/DGF/DCD.SAP/SCB, 045/047/12//MEF/SG/ DGF/DCD.SAP/SCB, 115 N-EA06/MG10, 072/079 N-EA06/MG11, 044 N-EA04/MG12), and to the Deutsche Forschungsgemeinschaft for funding (grant GL 665/1-1 and RO 3064/2-1). Two anonymous reviewers provided valuable and constructive criticism to a previous version of the manuscript, this is gratefully acknowledged.

Appendix A. Characteristics of forest fragments. Given are size, perimeter (peri), nearest distance to continuous forest (Ranomafana National Park border: dist RNP) and geographic coordinates of geographical centers of fragments. Additionally, species richness (SR) per transect for terrestrial (t) and stream (s) transects is provided for each fragment; ⁎ = fragments with stream; ⁎⁎ = fragment with 2 streams and 2 transects per habitat type

Fragment name

Size [ha]

Peri [km]

Dist RNP [km]

Ambohipierenana

2.09

0.70

0.11

Ankasaopasina

2.44

0.98

1.62

Andalangina_Left*

2.45

1.06

10.50

Imaloka Beremby

7.31

2.55

1.40

Ambolo⁎⁎

7.49

3.37

3.77

Antenna

9.44

2.00

0.63

10.88

2.63

10.73

Andalangina_Right*

geographic coordinates

SR t

SR s

21°14′50.84″S, 47°26′02.25″E 21°14′10.83″S, 47°31′43.85″E 21°17′46.06″S, 47°35′55.16″E 21°14′07.15″S, 47°31′21.17″E 21°15′50.67″S, 47°30′31.53″E 21°15′40.26″S, 47°26′44.61″E 21°17′50.57″S, 47°36′18.86″E

5

–

2

–

8

12

4

–

7/9

19/12

5

–

3

18

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715 Appendix B (continued) (continued)

Appendix A (continued) (continued) Fragment name

Size [ha]

Peri [km]

Antaramanavana*

16.42

2.69

1.96

Sahadikaina*

16.52

3.53

2.54

Dist RNP [km]

geographic coordinates

SR t

SR s

21°14′23.98″S, 47°30′20.88″E 21°14′47.98″S, 47°31″21.69″E

5

18

3

18

Appendix B. Collection numbers of voucher specimens deposited at the Museum für Naturkunde Berlin, Germany (ZMB), voucher specimens deposited at the Department of Animal Biology, University of Antananarivo, Madagascar (UADBA), and GenBank accession numbers. Names in parentheses refer to candidate species according to Vieites et al. (2009); * = RNP taxa formerly assigned to the paraphyletic genus Stumpffia, see Peloso et al. (2015); ** = species determined by their call only (determination according to Vences et al., 2006)

Species Hyperoliidae Heterixalus alboguttatus Heterixalus betsileo Mantellidae Aglyptodactylus sp. aff. madagascariensis (A. sp. 3) Boophis albilabris Boophis cf. albipunctatus** Boophis arcanus Boophis calcaratus Boophis elenae Boophis luciae

Boophis luteus Boophis madagascariensis Boophis cf. majori** Boophis marojezensis Boophis picturatus Boophis pyrrhus Boophis quasiboehmei Boophis reticulatus Gephyromantis asper Gephyromantis boulengeri Gephyromantis decaryi Gephyromantis enki

Gephyromantis plicifer Gephyromantis

713

Voucher number

Gephyromantis ventrimaculatus Guibemantis liber

Guibemantis cf. depressiceps** Guibemantis tornieri Mantidactylus aerumnalis Mantidactylus alutus Mantidactylus betsileanus Mantidactylus femoralis

ZMB 81832–ZMB 81833, UADBA A 43144, UADBA A 62097 ZMB 81834, UADBA A 43111

KT240397 - KT240403, KT240630 - KT240634

ZMB 81842, UADBA A 43117

KT240636

ZMB 77315–ZMB 77316 ZMB 81843–ZMB 81844, UADBA A 64013 ZMB 81845, UADBA A 64015 ZMB 81847–ZMB 81850, UADBA A 43119 - UADBA A 43120, UADBA A 62068, UADBA A 62070, UADBA A 62075 ZMB 81851–ZMB 81853, UADBA A 62071, UADBA A 64017 ZMB 81854–ZMB 81856, UADBA A 43122, UADBA A 62069

JQ413974–JQ413975 KT240637, KT240990

KT240405 - KT240406, KT240642

ZMB 81863

KT240643, KT240995

ZMB 81867

KT240644 - KT240646, KT240996 KT240407 - KT240409

ZMB 81885–ZMB 81889, UADBA A 43134 - UADBA A 43136, UADBA A 62085 - UADBA A 62086 UADBA A 43132

redimitus Gephyromantis sculpturatus Gephyromantis tschenki

GenBank accession

ZMB 81830, UADBA A 62096

ZMB 81868, UADBA A 62074 ZMB 81869–ZMB 81870, UADBA A 43125 - UADBA A 43127 ZMB 81871–ZMB 81873, UADBA A 43123 - UADBA A 43124, UADBA A 64019 ZMB 81878, UADBA A 43131, UADBA A 62081 ZMB 81879–ZMB 81883, UADBA A 62082 - UADBA A 62084 ZMB 81884

Species

KT240991 - KT240992 KT240638 - KT240639

KT240993 - KT240994

KT240997 - KT240998

KT240410 - KT240414, KT240647 KT240415 - KT240423, KT240648 - KT240659, KT240999 KT240660 KT240424, KT240661 KT240665

KT240425, KT240666 KT240668 KT240426 - KT240428,

Mantidactylus grandidieri Mantidactylus majori

Mantidactylus melanopleura Mantidactylus sp. aff. mocquardi (M. sp. 44) Mantidactylus paidroa Mantidactylus sp. aff. betsileanus “slow calls” (M. sp. 28) Mantidactylus sp. aff. biporus “Ranomafana” (M. sp. 24) Mantidactylus sp. aff. charlotteae “Ambohitsara” (M. sp. 9) Mantidactylus sp. aff. charlotteae “Ranomafana” (M. sp. 13) Mantidactylus sp. aff. cowanii “small” (M. sp. 48) Mantidactylus sp. aff. mocquardi “Ambatolahy” (M. sp. 47) Mantidactylus sp. aff. mocquardi “Namorona” (M. sp. 64)

Voucher number

GenBank accession

KT240669 - KT240671 ZMB 81892, UADBA A 43137 KT240429 - KT240434, UADBA A 43139, UADBA A 62087 KT240672 - KT240687 ZMB 81895–ZMB 81896, UADBA A 43140, UADBA A 62088, UADBA A 64024 ZMB 81897, UADBA A 62089 KT240688 - KT240689 UADBA A 62090 ZMB 81901–ZMB 81906, KT240435, KT240692 UADBA A 43141 - UADBA A KT240694 43142, UADBA A 62091 - UADBA A 62094

ZMB 81912, UADBA A 43143, UADBA A 62095 ZMB 81913

KT240437 - KT240440, KT240695 - KT240699 KT240702 - KT240703 KT240704 - KT240709

ZMB 81914, ZMB 81916, UADBA A 43147

KT240441 - KT240446, KT240710 - KT240712, KT240714 ZMB 81937–ZMB81938, KT240447 - KT240455, UADBA A 43163, UADBA A 62099 KT240715 - T240728 - UADBA A 62100, UADBA A 62114 UADBA A 43156, UADBA A 43161 KT240456 - KT240466 ZMB 81940–ZMB 81941, UADBA A 43157 - UADBA A 43158, UADBA A 64029 - UADBA A 64031 ZMB 81942–ZMB 81944 ZMB 81947–ZMB 81948

KT240467 - KT240471, KT240729 - KT240734 KT240531 - KT240538, KT240845 - KT240857

ZMB 81956–ZMB 81957, UADBA A 62103, UADBA A 62111 - UADBA A 62112 ZMB 81918–ZMB 81920, ZMB 81922, UADBA A 43149, UADBA A 62104 - UADBA A 62106

KT240472 - KT240485, KT240736 - KT240771

ZMB 81925, UADBA A 43160

KT240495 - KT240496, KT240788 - KT240792

KT240497 - KT240530, KT240700 - KT240701, KT240793 - KT240844

ZMB 81927–ZMB 81930, UADBA KT240602 - KT240613, A 43145 - UADBA A 43146, KT240913 - KT240936 UADBA A 43154, UADBA A 62110

ZMB 81931–ZMB 81932, UADBA A 43153, UADBA A 62098, UADBA A 62108 - UADBA A 62109

KT240486 - KT240494, KT240773 - KT240787

ZMB 81933, ZMB 81935–ZMB 81936, UADBA A 43162, UADBA A 62113, UADBA A 64037 - UADBA A 64039 ZMB 81949–ZMB 81951, ZMB 81953–ZMB 81954, UADBA A 64040 - UADBA A 64041

KT240571 - KT240583, KT241005 - KT241009

ZMB 81952, UADBA A 62115

KT240539 - KT240570, KT240858 - KT240896

KT240600 - KT240601, KT240909 - KT240912

714

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

(continued) Species

Voucher number

GenBank accession

Mantidactylus sp. aff. opiparis (M. sp. 58) Mantidactylus tricinctus Spinomantis sp. aff. fimbriatus (S. sp. 2) Microhylidae Anodonthyla boulengeri

ZMB 81955, UADBA A 62116

KT240584 - KT240599, KT240897 - KT240908

ZMB 81958–ZMB 81964, UADBA A 62107 ZMB 81968, UADBA A 43168 UADBA A 43169, UADBA A 62119

KT240614, KT240937 KT240943 KT240619 - KT240622, KT240948 - KT240951

ZMB 81970–ZMB 81973, UADBA A 43112 - UADBA A 43116, UADBA A 62063 - UADBA A 62064 ZMB 81976–ZMB 81977, UADBA A 62065 ZMB 81984 ZMB 81987, UADBA A 64047

KT240404, KT240635

Anodonthyla moramora Platypelis grandis Platypelis pollicaris Platypelis tuberifera Plethodontohyla brevipes Plethodontohyla inguinalis Rhombophyrne miery* Rhombophryne sp. “Ranomafana”* (Stumpffia sp. 9)

KT240615, KT240944 KT241010

ZMB 81988, UADBA A 43165 KT240616 - KT240617

ZMB 77453–ZMB 77456, UADBA-A 62120– UADBA-A 62124 ZMB 81994–ZMB 81997, ZMB 81999, UADBA A 43170, UADBA A 62125 -UADBA A 62127, UADBA A 64052

KT240618, KT240945 KT240946 KC351191–KC351193

KT240623, KT240952

References Allnutt, T.F., Ferrier, S., Manion, G., Powell, G.V.N., Ricketts, T.H., Fisher, B.L., Harper, G.J., Irwin, M.E., Kremen, C., Labat, J.-N., Lees, D.C., Pearce, T.A., Rakotondrainibe, F., 2008. A method for quantifying biodiversity loss and its application to a 50-year record of deforestation across Madagascar. Conserv. Lett. 1, 173–181. Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecol. 26, 32–46. Andreone, F., 1994. The amphibians of Ranomafana rain forest, Madagascar — preliminary community analysis and conservation considerations. Oryx 28, 207–214. Andreone, F., Cadle, J.E., Cox, N., Glaw, F., Nussbaum, R.A., Raxworthy, C.J., Stuart, S.N., Vallan, D., Vences, M., 2005. Species review of amphibian extinction risks in Madagascar: conclusions from the Global Amphibian Assessment. Conserv. Biol. 19, 1790–1802. Andreone, F., Carpenter, A.I., Cox, N., du Preez, L., Freeman, K., Furrer, S., Garcia, G., Glaw, F., Glos, J., Knox, D., Köhler, J., Mendelson III, J.R., Mercurio, V., Mittermeier, R.A., Moore, R.D., Rabibisoa, N.H.C., Randriamahazo, H., Randrianasolo, H., Raminosoa, N.R., Ramilijaona, O.R., Raxworthy, C.J., Vallan, D., Vences, M., Vieites, D.R., Weldon, C., 2008. The challenge of conserving amphibian megadiversity in Madagascar. PLoS Biol. 6, e118. Barlow, J., Gardner, T.A., Araujo, I.S., Ávila-Pires, T.C., Bonaldo, A.B., Costa, J.E., Esposito, M.C., Ferreira, L.V., Hawes, J., Hernandez, M.I.M., Hoogmoed, M.S., Leite, R.N., Lo-ManHung, N.F., Malcolm, J.R., Martins, M.B., Mestre, L.A.M., Miranda-Santos, R., NunesGutjahr, A.L., Overal, W.L., Parry, L., Peters, S.L., Ribeiro-Junior, M.A., da Silva, M.N.F., da Silva Motta, C., Peres, C.A., 2007. Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proc. Natl. Acad. Sci. 104, 18555–18560. Beck, J., Khen, C.V., 2007. Beta-diversity of geometrid moths from northern Borneo: effects of habitat, time and space. J. Anim. Ecol. 76, 230–237. Bell, K.E., Donnelly, M.A., 2006. Influence of forest fragmentation on community structure of frogs and lizards in northeastern Costa Rica. Conserv. Biol. 20, 1750–1760. Bickford, D., Ng, T.H., Qie, L., Kudavidanage, E.P., Bradshaw, C.J.A., 2010. Forest fragment and breeding habitat characteristics explain frog diversity and abundance in Singapore. Biotropica 42, 119–125. Birkinshaw, C., Randrianjanahary, M., 2007. The effects of cyclone Hudah on the forest of Masoala Peninsula, Madagascar. Madagascar Conserv. Dev. 2, 17–20. Brown, K.A., Flynn, D.F.B., Abram, N.K., Ingram, J.C., Johnson, S.E., Wright, P., 2011. Assessing natural resource use by forest-reliant communities in Madagascar using functional diversity and functional redundancy metrics. PLoS ONE 6, e24107. Cabrera-Guzmán, E., Reynoso, V.H., 2012. Amphibian and reptile communities of rainforest fragments: minimum patch size to support high richness and abundance. Biodivers. Conserv. 21, 3243–3265. Daily, G.C., 2001. Ecological forecasts. Nature 411, 245. DeFries, R., Hansen, A., Newton, A.C., Hansen, M.C., 2005. Increasing isolation of protected areas in tropical forests over the past twenty years. Ecol. Appl. 15, 19–26.

de Gouvenain, R.C., Silander Jr., J.A., 2003. Do tropical storm regimes influence the structure of tropical lowland rain forests? Biotropica 35, 166–180. Dirzo, R., Raven, P.H., 2003. Global state of biodiversity and loss. Annu. Rev. Environ. Resour. 28, 137–167. Dixo, M., Metzger, J.P., 2010. The matrix-tolerance hypothesis: an empirical test with frogs in the Atlantic Forest. Biodivers. Conserv. 19, 3059–3071. Ernst, R., Rödel, M.-O., 2008. Patterns of community composition in two tropical tree frog assemblages: separating spatial structure and environmental effects in disturbed and undisturbed forests. J. Trop. Ecol. 24, 111–120. Fahrig, L., 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Evol. Syst. 34, 487–515. Ferraro, P.J., 2002. The local costs of establishing protected areas in low-income nations: Ranomafana National Park, Madagascar. Ecol. Econ. 43, 261–275. Ganzhorn, J.U., 1995. Cyclones over Madagascar: fate or fortune? Ambio 24, 124–125. Gardner, T.A., Barlow, J., Chazdon, R., Ewers, R.M., Harvey, C.A., Peres, C.A., Sodhi, N.S., 2009. Prospects for tropical forest biodiversity in a human-modified world. Ecol. Lett. 12, 561–582. Gascon, C., Lovejoy, T.E., Bierregaard Jr., R.O., Malcolm, J.R., Stouffer, P.C., Vasconcelos, H.L., Laurance, W.F., Zimmerman, B., Tocher, M., Borges, S., 1999. Matrix habitat and species richness in tropical forest remnants. Biol. Conserv. 91, 223–229. Ghazoul, J., 2002. Impact of logging on the richness and diversity of forest butterflies in a tropical dry forest in Thailand. Biodivers. Conserv. 11, 521–541. Gillespie, G.R., Ahmad, E., Elahan, B., Evans, A., Ancrenaz, M., Goossens, B., Scroggie, M.P., 2012. Conservation of amphibians in Borneo: relative value of secondary tropical forest and non-forest habitats. Biol. Conserv. 152, 136–144. Glaw, F., Vences, M., 2007. A Field Guide to the Amphibians and Reptiles of Madagascar. third ed. Vences & Glaw Verlag, Cologne. Grafe, T.U., Stewart, M.M., Lampert, K.P., Rödel, M.-O., 2011. Putting toe clipping into perspective: a viable method for marking anurans. J. Herpetol. 45, 28–35. Green, G.M., Sussman, R.W., 1990. Deforestation history of the eastern rain forests of Madagascar from satellite images. Science 248, 212–215. Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A., Thau, D., Stehman, S.V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A., Chini, L., Justice, C.O., Townshend, J.R.G., 2013. High-resolution global maps of 21st-century forest cover change. Science 342, 850–853. Harper, G.J., Steininger, M.K., Tucker, C.J., Juhn, D., Hawkins, F., 2007. Fifty years of deforestation and forest fragmentation in Madagascar. Environ. Conserv. 34, 325–333. Hill, J.K., Gray, M.A., Khen, C.V., Benedick, S., Tawatao, N., Hamer, K.C., 2011. Ecological impacts of tropical forest fragmentation: how consistent are patterns in species richness and nestedness? Philos. Trans. R. Soc. B 366, 3265–3276. Hillers, A., Veith, M., Rödel, M.-O., 2008. Effects of forest fragmentation and habitat degradation on West African leaf-litter frogs. Conserv. Biol. 22, 762–772. Irwin, M.T., Wright, P.C., Birkinshaw, C., Fisher, B.L., Gardner, C.J., Glos, J., Goodman, S.M., Loiselle, P., Rabeson, P., Raharison, J.-L., Raherilalao, M.J., Rakotondravony, D., Raselimanana, A., Ratsimbazafy, J., Sparks, J.S., Wilmé, L., Ganzhorn, J.U., 2010. Patterns of species change in anthropogenically disturbed forests of Madagascar. Biol. Conserv. 143, 2351–2362. IUCN, 2014. The IUCN Red List of Threatened Species, version 2014.3 URL:, http://www. iucnredlist.org (downloaded on 09 February 2015). Laurance, W.F., Camargo, J.L., Luizão, R.C., Laurance, S.G., Pimm, S.L., Bruna, E.M., Stouffer, P.C., Williamson, G.B., Benítez-Malvido, J., Vasconcelos, H.L., Houtan, K.S.V., Zartman, C.E., Boyle, S.A., Didhamm, R.K., Andrade, A., Lovejoy, T.E., 2011. The fate of Amazonian forest fragments: a 32-year investigation. Biol. Conserv. 144, 56–67. Laurance, W.F., Sayer, J., Cassman, K.G., 2014. Agricultural expansion and its impacts on tropical nature. Trends Ecol. Evol. 29, 107–116. Lehtinen, R.M., Ramanamanjato, J.-B., 2006. Effects of rainforest fragmentation and correlates of local extinction in a herpetofauna from Madagascar. Appl. Herpetol. 3, 95–110. Lehtinen, R.M., Ramanamanjato, J.-B., Raveloarison, J.G., 2003. Edge effects and extinction proneness in a herpetofauna from Madagascar. Biodivers. Conserv. 12, 1357–1370. Lips, K.R., Reeve, J.D., Witters, L.R., 2003. Ecological traits predicting amphibian population declines in Central America. Conserv. Biol. 17, 1078–1088. Marsh, D.M., Haywood, L.M.B., 2010. Area-based surveys. In: Dodd Jr., C.K. (Ed.), Amphibian Ecology and Conservation — a Handbook of Techniques. Oxford University Press, Oxford, pp. 247–262. McArdle, B.H., Anderson, M.J., 2001. Fitting multivariate models to community data: a comment on distance-based redundancy analysis. Ecology 82, 290–297. Mendenhall, C.D., Frishkoff, L.O., Santos-Barrera, G., Pacheco, J., Mesfun, E., Quijano, F.M., Ehrlich, P.R., Ceballos, G., Daily, G.C., Pringle, R.M., 2014. Countryside biogeography of Neotropical reptiles and amphibians. Ecology 95, 856–870. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853–858. Ndriantsoa, S.H., Riemann, J.C., Vences, M., Klages, J., Raminosoa, N.R., Rödel, M.-O., Glos, J., 2013. A new Stumpffia (Amphibia: Anura: Microhylidae) from the Ranomafana region, south-eastern Madagascar. Zootaxa 3636, 575–589. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O'Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2011. Vegan: community ecology package, R package version 2.0-2 URL: http://CRAN.R-project.org/package=vegan. Peloso, P.L., Frost, D.R., Richards, S.J., Rodrigues, M.T., Donnellan, S., Matsui, M., Raxworthy, C.J., Biju, S., Lemmon, E.M., Lemmon, A.R., Wheeler, W.C., 2015. The impact of anchored phylogenomics and taxon sampling on phylogenetic inference in narrowmouthed frogs (Anura, Microhylidae). Cladistics http://dx.doi.org/10.1111/cla.12118. Pineda, E., Halffter, G., 2004. Species diversity and habitat fragmentation: frogs in a tropical montane landscape in Mexico. Biol. Conserv. 117, 499–508. R Development Core Team, 2011. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria (ISBN: 3-900051-07-0, URL: http://www.R-project.org/).

J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715 Ramage, B.S., Sheil, D., Salim, H.M.W., Fletcher, C., Mustafa, N.-Z.A., Luruthusamay, J.C., Harrison, R.D., Butod, E., Dzulkiply, A.D., Kassim, A.R., Potts, M.D., 2013. Pseudoreplication in tropical forests and the resulting effects on biodiversity conservation. Conserv. Biol. 27, 364–372. Riemann, J.C., Ndriantsoa, S.H., Vences, M., Rödel, M.-O., Glos, J., 2012. Advertisement call and male morphology of the Malagasy treefrog Boophis arcanus from the Ranomafana region, south-eastern Madagascar. Zootaxa 3250, 66–68. Rödel, M.-O., Ernst, R., 2004. Measuring and monitoring amphibian diversity in tropical forests. I. An evaluation of methods with recommendations for standardization. Ecotropica 10, 1–14. Rodrigues, A.S.L., Andelman, S.J., Bakarr, M.I., Boitani, L., Brooks, T.M., Cowling, R.M., Fishpool, L.D.C., da Fonseca, G.A.B., Gaston, K.J., Hoffmann, M., Long, J.S., Marquet, P.A., Pilgrim, J.D., Pressey, R.L., Schipper, J., Sechrest, W., Stuart, S.N., Underhill, L.G., Waller, R.W., Watts, M.E.J., Yan, X., 2004. Effectiveness of the global protected area network in representing species diversity. Nature 428, 640–643. Rosa, G.M., Andreone, F., Crottini, A., Hauswaldt, J.S., Noel, J., Rabibisoa, N.H., Randriambahiniarime, M.O., Rebelo, R., Raxworthy, C.J., 2012. The amphibians of the relict Betampona low-elevation rainforest, eastern Madagascar: an application of the integrative taxonomy approach to biodiversity assessments. Biodivers. Conserv. 21, 1531–1559. Sodhi, N.S., Bickford, D., Diesmos, A.C., Lee, T.M., Koh, L.P., Brook, B.W., Sekercioglu, C.H., Bradshaw, C.J.A., 2008. Measuring the meltdown: drivers of global amphibian extinction and decline. PLoS ONE 3, e1636. Stuart, S.N., Chanson, J.S., Cox, N.A., Young, B.E., Rodrigues, A.S.L., Fischman, D.L., Waller, R.W., 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306, 1783–1786. Tocher, M.D., Gascon, C., Zimmerman, B.L., 1997. Fragmentation effects on a central Amazonian frog community: a ten-year study. In: Laurance, W.F., Bierregaard Jr., R.O.

715

(Eds.), Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities. University of Chicago Press, Chicago, pp. 124–137. Turton, S.M., Siegenthaler, D.T., 2004. Immediate impacts of a severe tropical cyclone on the microclimate of a rain-forest canopy in north-east Australia. J. Trop. Ecol. 20, 583–586. Vallan, D., 2000. Influence of forest fragmentation on amphibian diversity in the nature reserve of Ambohitantely, highland Madagascar. Biol. Conserv. 96, 31–43. Vences, M., Chiari, Y., Teschke, M., Randrianiaina, R.-D., Raharivololoniaina, L., Bora, P., Vieites, D.R., Glaw, F., 2008. Which frogs are out there? A preliminary evaluation of survey techniques and identification reliability of Malagasy amphibians. In: Andreone, F. (Ed.), A Conservation Strategy for the Amphibians of MadagascarMonografie XLV. Museo Regionale di Scienze Naturali, Torino, pp. 233–252. Vences, M., Glaw, F., Márquez, R., 2006. The Calls of the Frogs of Madagascar, 3 Audio CDs and Booklet. Alosa-Fonozoo, Barcelona. Vences, M., Köhler, J., Pabijan, M., Glaw, F., 2010. Two syntopic and microendemic new frogs of the genus Blommersia from the east coast of Madagascar. Afr. J. Herpetol. 59, 133–156. Vences, M., Thomas, M., van der Meijden, A., Chiari, Y., Vieites, D.R., 2005. Comparative performance of the 16S rRNA gene in DNA barcoding of amphibians. Front. Zool. 2, 5. Vieites, D.R., Wollenberg, K.C., Andreone, F., Köhler, J., Glaw, F., Vences, M., 2009. Vast underestimation of Madagascar's biodiversity evidenced by an integrative amphibian inventory. Proc. Natl. Acad. Sci. 106, 8267–8272. Villard, M.-A., Metzger, J.P., 2014. Beyond the fragmentation debate: a conceptual model to predict when habitat configuration really matters. J. Appl. Ecol. 51, 309–318. Wright, P.C., Andriamihaja, B., 2003. The conservation value of long-term research: a case study from the Parc National de Ranomafana. In: Goodman, S.M., Benstead, J.P. (Eds.), The Natural History of Madagascar. University of Chicago Press, Chicago, pp. 1485–1488.