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
Zoological Institute, Animal Ecology and Conservation, University of Hamburg, Martin-Luther-King Platz 3, 20146 Hamburg, Germany
b
Département de Biologie Animale, Université d'Antananarivo, Antananarivo 101, Madagascar
c
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 a b s t r a c t

Article history: Forest fragmentation often causes biodiversity loss, but there is no consistent pattern on species' reactions. Con-
Received 24 February 2015 sidering the alarming rate of deforestation in the tropics, and the fact, that large areas of protected continuous
Received in revised form 11 June 2015 forest are limited, it becomes increasingly important to determine the biodiversity value of fragmented forests.
Accepted 9 August 2015
In order to investigate fragmentation effects on rainforest frogs in Madagascar and to assess the conservation
Available online xxxx
value of these fragments, we analyzed amphibian diversity in a continuous rainforest and nearby forest frag-
Keywords:
ments. We hypothesized that species richness is lower in fragments compared to continuous forest, and that frag-
Habitat fragmentation mentation leads to altered assemblage composition. We found no fragmentation effects on species richness,
Rainforest demonstrating that fragments may maintain local species richness comparable to continuous forest. The pres-
Amphibian conservation ence of streams was the most important factor for high species richness, independent of fragmentation status.
Species richness However, we detected fragmentation effects on species composition. As expected, several species were restricted
Species composition to continuous forest, but many species occurred in both forest types, and some species were only found in frag-
Disturbance ments. 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 includ-
ed in conservation planning.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction including numerous threatened species (Rodrigues et al., 2004). The

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 smaller-
sized and isolated habitat patches, and an increasing ratio of edge to in-
terior 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: jana.riemann@uni-hamburg.de, jcriemann@gmail.com
(J.C. Riemann), nsehel2006@gmail.com (S.H. Ndriantsoa), rsnoro@yahoo.fr
(N.R. Raminosoa), mo.roedel@mfn-berlin.de (M.-O. Rödel), julian.glos@uni-hamburg.de
(J. Glos).
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 as-
sumed 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 de-
termine 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 in-
creasing 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 con-
sidered 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

http://dx.doi.org/10.1016/j.biocon.2015.08.020
0006-3207/© 2015 Elsevier Ltd. All rights reserved.
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 out-
standing 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 re-
cent 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 fragmenta-
tion 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 for-
est 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 sys-
tem 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 Na-
tional Park, N40,000 ha), and evaluated patterns of assemblage composi-
tion 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 inde-
pendent 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 as-
semblage 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 frag-
ments located east of RNP (Fig. 1). RNP comprises 43,500 ha of continu-
ous mid-altitude montane rainforest (500–1300 m a.s.l.) with an annual
precipitation between 1700 and 4300 mm (Wright and Andriamihaja,
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).
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, in-
cluding 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 am-
phibian 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 frag-
ments comprise streams, data from stream and terrestrial transects
were analyzed separately. In the following, we refer to stream and ter-
restrial 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 with-
in 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.
 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 sam-
pling 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 distur-
bances of the study system as low as possible and to ensure equal sam-
pling 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 identi-
fied by morphology in the field (according to Glaw and Vences, 2007).
Afterwards, all individuals were immediately released at point of cap-
ture, 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 ad-
dition, own reference data was recorded for several species with a Ro-
land 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 re-
cordings 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, Univer-
sity 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 com-
pared 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 detect-
ed) per transect between continuous forest and forest fragments for ter-
restrial 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 for-
est 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 composi-
tional 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 co-
ordinates 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 configura-
tions 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 spe-
cies presence/absence data with function “adonis” from R package
vegan (Oksanen et al., 2011). P-values were obtained from 999 permu-
tations. 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) represent-
ed by 1026 individuals that were detected visually, complemented by
710 J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

Table 1 fragments we found 40 species in total, 13 in terrestrial habitat and 38

Species detected per habitat type (CF: continuous forest, Ranomafana National Park; FF: in stream habitat.
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:
We recorded no effects of forest fragmentation on species richness
near threatened, EN: endangered). Names in parentheses refer to candidate species ac- as we found no differences in species richness per transect between
cording to Vieites et al. (2009). continuous forest and forest fragments in stream habitats (Wilcoxon
rank sum test, W = 20, p = 0.204, n = 17) and terrestrial habitats
Species CF FF TS
(W = 74, p = 0.183, n = 21; Fig. 2).
t s t s

Hyperoliidae
Heterixalus alboguttatus − − + + LC 3.2. Species composition
Heterixalus betsileo − + + + LC
Mantellidae
Aglyptodactylus sp. aff. madagascariensis (A. sp. 3) − − + − LC We detected 26 species that occurred in continuous forest and forest
Boophis albilabris − + + + LC fragments, but 31 species were found in only one forest type (17 species
Boophis cf. albipunctatus − − − + LC in continuous forest and 14 species in forest fragments; Table 1). Ac-
Boophis arcanus − − − + NA
cordingly, overall about two thirds (40) of the detected species occurred
Boophis calcaratus − − − + NA
Boophis elenae − + − + DD in forest fragments. Large and medium-sized microhylids of the genera
Boophis luciae − + − − NA Platypelis and Plethodontohyla were restricted to continuous forest,
Boophis luteus − + − + LC whereas small species of Anodonthyla and Rhombophryne occurred
Boophis madagascariensis + + + + LC in both forest types or only in fragments (Table 1). The species-rich
Boophis cf. majori − − − + NT
Boophis marojezensis − + − − LC
mantellid genera Boophis, Gephyromantis and Mantidactylus all
Boophis picturatus − + − − LC contained several species that were found in both forest types as
Boophis pyrrhus − − − + LC well as species restricted to continuous forest or forest fragments.
Boophis quasiboehmei − + − + NA Two of the three Guibemantis species were found in both forest
Boophis reticulatus − + − − LC
types, the other only in continuous forest. Thus, a distinct phyloge-
Gephyromantis asper + − − − LC
Gephyromantis boulengeri + + + + LC netic signal between inhabitants of continuous forest and fragments
Gephyromantis decaryi − + − − NT could not be observed (Table 1).
Gephyromantis enki + + + + DD Most of the species found in terrestrial habitats also
Gephyromantis plicifer + + − − NT occurred in stream habitats (Table 1). Only two species in each
Gephyromantis redimitus − + − + LC
Gephyromantis sculpturatus + + + + LC
forest type were restricted to terrestrial habitat (RNP: Platypelis
Gephyromantis tschenki + + + + DD tuberifera, Gephyromantis asper; fragments: Anodonthyla moramora,
Gephyromantis ventrimaculatus − − − + LC Aglyptodactylus madagascariensis). One of these species found only in
Guibemantis liber − + − + LC terrestrial habitat in fragments also occurred in stream habitat in con-
Guibemantis cf. depressiceps − + − − LC
tinuous forest (A. moramora).
Guibemantis tornieri − + − + LC
Mantidactylus aerumnalis − − − + LC We found a significant correlation between compositional differ-
Mantidactylus alutus − + − − LC ences and spatial distance in the whole data set for stream habitats
Mantidactylus betsileanus − + − + LC (r = 0.647, p b 0.0001) and terrestrial habitats (r = 0.465, p b
Mantidactylus femoralis − + − + LC 0.0001). We further detected a significant correlation between compo-
Mantidactylus grandidieri − + − + LC
Mantidactylus majori − + − − LC
sitional differences and spatial distance within continuous forest in both
Mantidactylus melanopleura + + − + LC habitat types (stream: r = 0.43, p = 0.011; terrestrial: r = 0.577, p =
Mantidactylus sp. aff. mocquardi (M. sp. 44) − + − + LC 0.003). However, we found no correlation between compositional dif-
Mantidactylus paidroa − + − − NA ferences and spatial distance among forest fragments (stream: r =
Mantidactylus sp. aff. betsileanus “slow calls” (M. sp. 28) − + − + NA
0.279, p = 0.112; terrestrial: r = 0.003, p = 0.478), showing that as-
Mantidactylus sp. aff. biporus “Ranomafana” (M. sp. 24) − + − + NA
Mantidactylus sp. aff. charlotteae “Ambohitsara” (M. sp. 9) − + − + NA semblages in fragments were not spatially structured. These contrasting
Mantidactylus sp. aff. charlotteae “Ranomafana” (M. sp. 13) − − − + NA patterns of spatial structuring in continuous forest and forest fragments
Mantidactylus sp. aff. cowanii “small” (M. sp. 48) − + − + NA
Mantidactylus sp. aff. mocquardi “Ambatolahy” (M. sp. 47) − + − + NA
Mantidactylus sp. aff. mocquardi “Namorona” (M. sp. 64) − − − + NA
Mantidactylus sp. aff. opiparis (M. sp. 58) + + − − NA
Mantidactylus tricinctus − − − + DD
Spinomantis sp. aff. fimbriatus (S. sp. 2) − + − + NA
Microhylidae
Anodonthyla boulengeri + + + + NA
Anodonthyla moramora + + + − DD
Platypelis grandis − + − − LC
Platypelis pollicaris + + − − DD
Platypelis tuberifera + − − − LC
Plethodontohyla brevipes + + − − EN
Plethodontohyla inguinalis + + − − LC
Rhombophyrne mierya − − + + NA
Rhombophryne sp. “Ranomafana”a (Stumpffia sp. 9) − − + + 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 con-
tinuous 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, inter-
quartile 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 dis-
tance alone but forest fragmentation affected assemblage composition.
We found significant differences in species composition be-
tween 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 reason-
able 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 separat-
ed 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 fragmen-
tation 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 impor-
tant aspect of habitat fragmentation. It is well proven that patch size
affects amphibian species richness as smaller fragments usually con-
tain 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 com-
position 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 dis-
similarity index using species presence/absence data (stress = 0.146). Distances between
transects in the two-dimensional NMDS plot represent dissimilarities in species composi-
tion. 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 indicat-
ed 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 hos-
tile 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 popula-
tions 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 for-
est 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 ter-
restrial 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 distribu-
tion 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, in-
creased wind speed and decreased relative humidity near edges in frag-
ments 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 con-
tinuous forest and forest fragments, i.e., assemblages in continuous for-
est 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 con-
tinuous forest and forest fragments supported the hypothesis that frag-
mentation 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 hab-
itats. 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 5. Conclusion

(e.g., Boophis albilabris, Boophis madagascariensis, Mantidactylus
grandidieri, Mantidactylus sp. aff. mocquardi) were also common in for- Since forest fragmentation altered assemblage composition, forest
est fragments. Andreone (1994) hypothesized that arboreal and fragments cannot substitute continuous forest. It is well proven that
stream-dwelling species may not depend on the microclimatic condi- large continuous primary forests are indispensable for protecting and
tions on the forest floor and may adapt to disturbed environments. retaining tropical biodiversity and ecosystem functions (Barlow et al.,
Indeed, many arboreal (genera Boophis, Guibemantis) and stream- 2007; Laurance et al., 2011). However, remaining forests are under
dwelling (genus Mantidactylus) mantellid species found in our enormous pressure all over the tropics due to ongoing deforestation
study seemed to be tolerant against fragmentation. However, some (Hansen et al., 2013) and habitat conversion for agriculture (Laurance
of these species were also restricted to continuous forest. On the et al., 2014). Therefore, the maintenance or enhancement of overall
other hand, several leaf litter or low vegetation associated habitat amount is an important consideration for conservation planning
Gephyromantis species also occurred in fragments. In brief, it is not (Villard and Metzger, 2014). We conclude that forest fragments are im-
clear which factors and traits predispose for fragmentation sensitivity. portant refuges of amphibian diversity in Madagascar, especially if they
It is important to point out that we did not only detect species that comprise streams. Even small forest fragments can be vital comple-
were restricted to continuous forest or occurred in both forest types, ments to a network of larger protected areas. Accordingly, forest frag-
but also species that were only found in fragments. It is known from ments should be included to a greater extent in conservation planning.
other tropical regions, that forest fragments are often invaded by open It should be noted that forest fragments are often also the last avail-
habitat species (e.g., Gascon et al., 1999; Hillers et al., 2008). This able forest parts for local populations, who highly depend on natural
explained higher species richness in fragments compared to con- forest resources (Brown et al., 2011; Ferraro, 2002). Hence, strategies
tinuous forest in the study of Tocher et al. (1997) and might explain for sustainable natural resource use are urgently needed to successfully
equal species richness in our study. However, only two species protect biodiversity in remaining forest fragments and to ensure their
(genus Heterixalus) that were exclusively found in fragments are usual- long-term existence.
ly associated with open habitat (Glaw and Vences, 2007). The eastern
rainforest belt of Madagascar was originally completely forested Acknowledgments
(Green and Sussman, 1990) and thus the majority of species in our
study region should be typical forest species. Nevertheless, frequent We thank E. Larney and the team of Centre ValBio research station
cyclone disturbances influence forest structure (de Gouvenain and for their logistic support, and especially our research assistants A. Telo
Silander, 2003; Birkinshaw and Randrianjanahary, 2007), microclimate and J. Solo for their help in the field. This study would not have been
(Turton and Siegenthaler, 2004), and tree biology (Ganzhorn, 1995), possible without the permission of several forest owners and villages
and thus may have also favored amphibian adaptations to disturbed to work in their forest fragments. We thank M. Vences, A. Strauss, and
habitats. Consequently, species adapted to natural disturbances may M. Kondermann for helping with species identification via DNA
have a higher ability to cope with anthropogenic disturbances. This barcoding. We are grateful to Madagascar National Parks and the
could be one potential reason for the resistance of many Malagasy Ministère de l'Environnement, de l'Ecologie, de la Mer et des Forêts
rainforest amphibians to fragmentation, i.e. about two third of our re- for research and export permits (017/033/10/MEF/SG/DGF/DCD.SAP/
corded species occurred in forest fragments. System specific factors SLRSE, 003/004/11/MEF/SG/DGF/DCD.SAP/SCB, 045/047/12//MEF/SG/
like cyclones could thus explain differences in organism's reactions to DGF/DCD.SAP/SCB, 115 N-EA06/MG10, 072/079 N-EA06/MG11, 044
anthropogenic disturbances between study regions. N-EA04/MG12), and to the Deutsche Forschungsgemeinschaft for
Species that were tolerant against fragmentation or even exclusively funding (grant GL 665/1-1 and RO 3064/2-1). Two anonymous
occurred in fragments most likely adapted best or may be even special- reviewers provided valuable and constructive criticism to a previous
ized on disturbed habitat and are therefore able to exist in forest frag- version of the manuscript, this is gratefully acknowledged.
ments. 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 Appendix A. Characteristics of forest fragments. Given are size,
not sampled, but some may be indeed restricted to fragmented areas. perimeter (peri), nearest distance to continuous forest (Ranomafana
Given the fact that assemblages naturally tend to be spatially structured National Park border: dist RNP) and geographic coordinates of
(Ernst and Rödel, 2008; Hillers et al., 2008), as proven in the continuous geographical centers of fragments. Additionally, species richness
forest part in our study, species that exclusively occurred in forest (SR) per transect for terrestrial (t) and stream (s) transects is provided
fragments may also represent remnant populations. If a species' major for each fragment; ⁎ = fragments with stream; ⁎⁎ = fragment with 2
distribution area was located in parts of the original forest cover streams and 2 transects per habitat type
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 Fragment name Size Peri Dist RNP geographic SR t SR s
[ha] [km] [km] coordinates
of forest fragments becomes even more important.
In many highly deforested areas on Madagascar small forest Ambohipierenana 2.09 0.70 0.11 21°14′50.84″S, 5 –
fragments represent the last remaining habitats for amphibians 47°26′02.25″E
Ankasaopasina 2.44 0.98 1.62 21°14′10.83″S, 2 –
(Andreone et al., 2005, 2008). It remains unknown whether popula- 47°31′43.85″E
tions in forest fragments represent viable populations in the long-term Andalangina_Left* 2.45 1.06 10.50 21°17′46.06″S, 8 12
and whether matrix quality and structure allow for migration and 47°35′55.16″E
genetic exchange between populations in different fragments or frag- Imaloka Beremby 7.31 2.55 1.40 21°14′07.15″S, 4 –
47°31′21.17″E
ments and continuous forests. Since frogs have relatively short genera-
Ambolo⁎⁎ 7.49 3.37 3.77 21°15′50.67″S, 7/9 19/12
tion times and all studied fragments were separated from continuous 47°30′31.53″E
forest for at least fifty years, we assume that the observed patterns for Antenna 9.44 2.00 0.63 21°15′40.26″S, 5 –
species richness and composition in our study will persist over time. 47°26′44.61″E
However, more information about the genetic diversity of fragment Andalangina_Right* 10.88 2.63 10.73 21°17′50.57″S, 3 18
47°36′18.86″E
populations and matrix effects is urgently needed.
 J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715 713

Appendix A (continued)
(continued) Appendix B (continued)
(continued)

Fragment name Size Peri Dist RNP geographic SR t SR s Species Voucher number GenBank accession
[ha] [km] [km] coordinates
redimitus KT240669 - KT240671
Antaramanavana* 16.42 2.69 1.96 21°14′23.98″S, 5 18 Gephyromantis ZMB 81892, UADBA A 43137 - KT240429 - KT240434,
47°30′20.88″E sculpturatus UADBA A 43139, UADBA A 62087 KT240672 - KT240687
Sahadikaina* 16.52 3.53 2.54 21°14′47.98″S, 3 18 Gephyromantis ZMB 81895–ZMB 81896,
47°31″21.69″E tschenki UADBA A 43140, UADBA A
62088, UADBA A 64024
Gephyromantis ZMB 81897, UADBA A 62089 - KT240688 - KT240689
ventrimaculatus UADBA A 62090
Guibemantis liber ZMB 81901–ZMB 81906, KT240435, KT240692 -
Appendix B. Collection numbers of voucher specimens deposited UADBA A 43141 - UADBA A KT240694
at the Museum für Naturkunde Berlin, Germany (ZMB), voucher 43142, UADBA A 62091 - UADBA
specimens deposited at the Department of Animal Biology, University A 62094
of Antananarivo, Madagascar (UADBA), and GenBank accession num- Guibemantis cf.
depressiceps**
bers. Names in parentheses refer to candidate species according to Guibemantis ZMB 81912, UADBA A 43143, KT240437 - KT240440,
Vieites et al. (2009); * = RNP taxa formerly assigned to the tornieri UADBA A 62095 KT240695 - KT240699
paraphyletic genus Stumpffia, see Peloso et al. (2015); ** = species de- Mantidactylus ZMB 81913 KT240702 - KT240703
termined by their call only (determination according to Vences et al., aerumnalis
Mantidactylus KT240704 - KT240709
2006)
alutus
Mantidactylus ZMB 81914, ZMB 81916, KT240441 - KT240446,
betsileanus UADBA A 43147 KT240710 - KT240712,
Species Voucher number GenBank accession KT240714
Mantidactylus ZMB 81937–ZMB81938, KT240447 - KT240455,
Hyperoliidae
femoralis UADBA A 43163, UADBA A 62099 KT240715 - T240728
Heterixalus ZMB 81830, UADBA A 62096
- UADBA A 62100, UADBA A
alboguttatus
62114
Heterixalus ZMB 81832–ZMB
Mantidactylus UADBA A 43156, UADBA A 43161 KT240456 - KT240466
betsileo 81833, UADBA A 43144, UADBA
grandidieri
A 62097
Mantidactylus ZMB 81940–ZMB 81941,
Mantellidae
majori UADBA A 43157 - UADBA A
Aglyptodactylus ZMB 81834, UADBA A 43111 KT240397 - KT240403,
43158, UADBA A 64029 - UADBA
sp. aff. KT240630 - KT240634
A 64031
madagascariensis
Mantidactylus ZMB 81942–ZMB 81944 KT240467 - KT240471,
(A. sp. 3)
melanopleura KT240729 - KT240734
Boophis albilabris ZMB 81842, UADBA A 43117 KT240636
Mantidactylus sp. ZMB 81947–ZMB 81948 KT240531 - KT240538,
Boophis cf.
aff. mocquardi KT240845 - KT240857
albipunctatus**
(M. sp. 44)
Boophis arcanus ZMB 77315–ZMB 77316 JQ413974–JQ413975
Mantidactylus ZMB 81956–ZMB 81957, KT240472 - KT240485,
Boophis ZMB 81843–ZMB 81844, KT240637, KT240990
paidroa UADBA A 62103, UADBA A 62111 KT240736 - KT240771
calcaratus UADBA A 64013
- UADBA A 62112
Boophis elenae ZMB 81845, UADBA A 64015 KT240991 - KT240992
Mantidactylus sp. ZMB 81918–ZMB 81920, KT240497 - KT240530,
Boophis luciae ZMB 81847–ZMB 81850, KT240638 - KT240639
aff. ZMB 81922, UADBA A 43149, KT240700 - KT240701,
UADBA A 43119 - UADBA A
betsileanus “slow UADBA A 62104 - UADBA A KT240793 - KT240844
43120, UADBA A 62068, UADBA
calls” 62106
A 62070, UADBA A 62075
(M. sp. 28)
Boophis luteus ZMB 81851–ZMB 81853, KT240993 - KT240994
Mantidactylus sp. ZMB 81925, UADBA A 43160 KT240495 - KT240496,
UADBA A 62071, UADBA A 64017
aff. KT240788 - KT240792
Boophis ZMB 81854–ZMB 81856, KT240405 - KT240406,
biporus
madagascariensis UADBA A 43122, UADBA A 62069 KT240642
“Ranomafana”
Boophis cf.
(M. sp. 24)
majori**
Mantidactylus sp. ZMB 81927–ZMB 81930, UADBA KT240602 - KT240613,
Boophis ZMB 81863 KT240643, KT240995
aff. A 43145 - UADBA A 43146, KT240913 - KT240936
marojezensis
charlotteae UADBA A 43154, UADBA A 62110
Boophis ZMB 81867 KT240644 - KT240646,
“Ambohitsara”
picturatus KT240996
(M. sp. 9)
Boophis pyrrhus ZMB 81868, UADBA A 62074 KT240407 - KT240409
Mantidactylus sp. ZMB 81931–ZMB 81932, KT240486 - KT240494,
Boophis ZMB 81869–ZMB 81870,
aff. UADBA A 43153, UADBA A KT240773 - KT240787
quasiboehmei UADBA A 43125 - UADBA A
charlotteae 62098, UADBA A 62108 - UADBA
43127
“Ranomafana” A 62109
Boophis ZMB 81871–ZMB 81873, KT240997 - KT240998
(M. sp. 13)
reticulatus UADBA A 43123 - UADBA A
Mantidactylus sp. ZMB 81933, ZMB KT240571 - KT240583,
43124, UADBA A 64019
aff. 81935–ZMB 81936, KT241005 - KT241009
Gephyromantis ZMB 81878, UADBA A 43131, KT240410 - KT240414,
cowanii “small” UADBA A 43162, UADBA A
asper UADBA A 62081 KT240647
(M. sp. 48) 62113, UADBA A 64037 - UADBA
Gephyromantis ZMB 81879–ZMB 81883, KT240415 - KT240423,
A 64039
boulengeri UADBA A 62082 - UADBA A KT240648 - KT240659,
Mantidactylus sp. ZMB 81949–ZMB 81951, KT240539 - KT240570,
62084 KT240999
aff. ZMB 81953–ZMB 81954, KT240858 - KT240896
Gephyromantis ZMB 81884 KT240660
mocquardi UADBA A 64040 - UADBA A
decaryi
“Ambatolahy” 64041
Gephyromantis ZMB 81885–ZMB 81889, KT240424, KT240661 -
(M. sp. 47)
enki UADBA A 43134 - UADBA A KT240665
Mantidactylus sp. ZMB 81952, UADBA A 62115 KT240600 - KT240601,
43136, UADBA A 62085 - UADBA
aff. KT240909 - KT240912
A 62086
mocquardi
Gephyromantis UADBA A 43132 KT240425, KT240666 -
“Namorona”
plicifer KT240668
(M. sp. 64)
Gephyromantis KT240426 - KT240428,
714 J.C. Riemann et al. / Biological Conservation 191 (2015) 707–715

(continued) de Gouvenain, R.C., Silander Jr., J.A., 2003. Do tropical storm regimes influence the struc-
ture of tropical lowland rain forests? Biotropica 35, 166–180.
Species Voucher number GenBank accession Dirzo, R., Raven, P.H., 2003. Global state of biodiversity and loss. Annu. Rev. Environ.
Resour. 28, 137–167.
Mantidactylus sp. ZMB 81955, UADBA A 62116 KT240584 - KT240599, Dixo, M., Metzger, J.P., 2010. The matrix-tolerance hypothesis: an empirical test with
aff. KT240897 - KT240908 frogs in the Atlantic Forest. Biodivers. Conserv. 19, 3059–3071.
opiparis (M. Ernst, R., Rödel, M.-O., 2008. Patterns of community composition in two tropical tree frog
sp. 58) assemblages: separating spatial structure and environmental effects in disturbed and
Mantidactylus ZMB 81958–ZMB 81964, KT240614, KT240937 - undisturbed forests. J. Trop. Ecol. 24, 111–120.
tricinctus UADBA A 62107 KT240943 Fahrig, L., 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Evol.
Spinomantis sp. ZMB 81968, UADBA A 43168 - KT240619 - KT240622, Syst. 34, 487–515.
aff. UADBA A 43169, UADBA A 62119 KT240948 - KT240951 Ferraro, P.J., 2002. The local costs of establishing protected areas in low-income nations:
fimbriatus (S. Ranomafana National Park, Madagascar. Ecol. Econ. 43, 261–275.
sp. 2) 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.,
Microhylidae
2009. Prospects for tropical forest biodiversity in a human-modified world. Ecol.
Anodonthyla ZMB 81970–ZMB 81973, KT240404, KT240635
Lett. 12, 561–582.
boulengeri UADBA A 43112 - UADBA A
Gascon, C., Lovejoy, T.E., Bierregaard Jr., R.O., Malcolm, J.R., Stouffer, P.C., Vasconcelos, H.L.,
43116, UADBA A 62063 - UADBA Laurance, W.F., Zimmerman, B., Tocher, M., Borges, S., 1999. Matrix habitat and spe-
A 62064 cies richness in tropical forest remnants. Biol. Conserv. 91, 223–229.
Anodonthyla ZMB 81976–ZMB 81977, Ghazoul, J., 2002. Impact of logging on the richness and diversity of forest butterflies in a
moramora UADBA A 62065 tropical dry forest in Thailand. Biodivers. Conserv. 11, 521–541.
Platypelis grandis ZMB 81984 KT240615, KT240944 Gillespie, G.R., Ahmad, E., Elahan, B., Evans, A., Ancrenaz, M., Goossens, B., Scroggie, M.P.,
Platypelis ZMB 81987, UADBA A 64047 KT241010 2012. Conservation of amphibians in Borneo: relative value of secondary tropical for-
pollicaris est and non-forest habitats. Biol. Conserv. 152, 136–144.
Platypelis ZMB 81988, UADBA A 43165 Glaw, F., Vences, M., 2007. A Field Guide to the Amphibians and Reptiles of Madagascar.
tuberifera third ed. Vences & Glaw Verlag, Cologne.
Plethodontohyla KT240616 - KT240617 Grafe, T.U., Stewart, M.M., Lampert, K.P., Rödel, M.-O., 2011. Putting toe clipping into
brevipes 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
Plethodontohyla KT240618, KT240945 -
Madagascar from satellite images. Science 248, 212–215.
inguinalis KT240946
Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A., Thau,
Rhombophyrne ZMB 77453–ZMB 77456, KC351191–KC351193 D., Stehman, S.V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A., Chini, L.,
miery* UADBA-A 62120– Justice, C.O., Townshend, J.R.G., 2013. High-resolution global maps of 21st-century
UADBA-A 62124 forest cover change. Science 342, 850–853.
Rhombophryne ZMB 81994–ZMB 81997, KT240623, KT240952 Harper, G.J., Steininger, M.K., Tucker, C.J., Juhn, D., Hawkins, F., 2007. Fifty years of defor-
sp. ZMB 81999, UADBA A 43170, estation and forest fragmentation in Madagascar. Environ. Conserv. 34, 325–333.
“Ranomafana”* UADBA A 62125 -UADBA A Hill, J.K., Gray, M.A., Khen, C.V., Benedick, S., Tawatao, N., Hamer, K.C., 2011. Ecological
(Stumpffia sp. 9) 62127, UADBA A 64052 impacts of tropical forest fragmentation: how consistent are patterns in species rich-
ness and nestedness? Philos. Trans. R. Soc. B 366, 3265–3276.
Hillers, A., Veith, M., Rödel, M.-O., 2008. Effects of forest fragmentation and habitat degra-
dation 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.,
References Raselimanana, A., Ratsimbazafy, J., Sparks, J.S., Wilmé, L., Ganzhorn, J.U., 2010. Pat-
Allnutt, T.F., Ferrier, S., Manion, G., Powell, G.V.N., Ricketts, T.H., Fisher, B.L., Harper, G.J., terns of species change in anthropogenically disturbed forests of Madagascar. Biol.
Irwin, M.E., Kremen, C., Labat, J.-N., Lees, D.C., Pearce, T.A., Rakotondrainibe, F., Conserv. 143, 2351–2362.
2008. A method for quantifying biodiversity loss and its application to a 50-year IUCN, 2014. The IUCN Red List of Threatened Species, version 2014.3 URL:, http://www.
record of deforestation across Madagascar. Conserv. Lett. 1, 173–181. iucnredlist.org (downloaded on 09 February 2015).
Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. Laurance, W.F., Camargo, J.L., Luizão, R.C., Laurance, S.G., Pimm, S.L., Bruna, E.M., Stouffer,
Austral Ecol. 26, 32–46. P.C., Williamson, G.B., Benítez-Malvido, J., Vasconcelos, H.L., Houtan, K.S.V., Zartman,
Andreone, F., 1994. The amphibians of Ranomafana rain forest, Madagascar — preliminary C.E., Boyle, S.A., Didhamm, R.K., Andrade, A., Lovejoy, T.E., 2011. The fate of
community analysis and conservation considerations. Oryx 28, 207–214. Amazonian forest fragments: a 32-year investigation. Biol. Conserv. 144, 56–67.
Andreone, F., Cadle, J.E., Cox, N., Glaw, F., Nussbaum, R.A., Raxworthy, C.J., Stuart, S.N., Laurance, W.F., Sayer, J., Cassman, K.G., 2014. Agricultural expansion and its impacts on
Vallan, D., Vences, M., 2005. Species review of amphibian extinction risks in tropical nature. Trends Ecol. Evol. 29, 107–116.
Madagascar: conclusions from the Global Amphibian Assessment. Conserv. Biol. 19, Lehtinen, R.M., Ramanamanjato, J.-B., 2006. Effects of rainforest fragmentation and corre-
1790–1802. lates of local extinction in a herpetofauna from Madagascar. Appl. Herpetol. 3, 95–110.
Andreone, F., Carpenter, A.I., Cox, N., du Preez, L., Freeman, K., Furrer, S., Garcia, G., Glaw, F., Lehtinen, R.M., Ramanamanjato, J.-B., Raveloarison, J.G., 2003. Edge effects and extinction
Glos, J., Knox, D., Köhler, J., Mendelson III, J.R., Mercurio, V., Mittermeier, R.A., Moore, proneness in a herpetofauna from Madagascar. Biodivers. Conserv. 12, 1357–1370.
R.D., Rabibisoa, N.H.C., Randriamahazo, H., Randrianasolo, H., Raminosoa, N.R., Lips, K.R., Reeve, J.D., Witters, L.R., 2003. Ecological traits predicting amphibian population
Ramilijaona, O.R., Raxworthy, C.J., Vallan, D., Vences, M., Vieites, D.R., Weldon, C., 2008. declines in Central America. Conserv. Biol. 17, 1078–1088.
The challenge of conserving amphibian megadiversity in Madagascar. PLoS Biol. 6, e118. Marsh, D.M., Haywood, L.M.B., 2010. Area-based surveys. In: Dodd Jr., C.K. (Ed.), Amphib-
Barlow, J., Gardner, T.A., Araujo, I.S., Ávila-Pires, T.C., Bonaldo, A.B., Costa, J.E., Esposito, M.C., ian Ecology and Conservation — a Handbook of Techniques. Oxford University Press,
Ferreira, L.V., Hawes, J., Hernandez, M.I.M., Hoogmoed, M.S., Leite, R.N., Lo-Man- Oxford, pp. 247–262.
Hung, N.F., Malcolm, J.R., Martins, M.B., Mestre, L.A.M., Miranda-Santos, R., Nunes- McArdle, B.H., Anderson, M.J., 2001. Fitting multivariate models to community data: a
Gutjahr, A.L., Overal, W.L., Parry, L., Peters, S.L., Ribeiro-Junior, M.A., da Silva, M.N.F., comment on distance-based redundancy analysis. Ecology 82, 290–297.
da Silva Motta, C., Peres, C.A., 2007. Quantifying the biodiversity value of tropical pri- Mendenhall, C.D., Frishkoff, L.O., Santos-Barrera, G., Pacheco, J., Mesfun, E., Quijano, F.M.,
mary, secondary, and plantation forests. Proc. Natl. Acad. Sci. 104, 18555–18560. Ehrlich, P.R., Ceballos, G., Daily, G.C., Pringle, R.M., 2014. Countryside biogeography
Beck, J., Khen, C.V., 2007. Beta-diversity of geometrid moths from northern Borneo: effects of Neotropical reptiles and amphibians. Ecology 95, 856–870.
of habitat, time and space. J. Anim. Ecol. 76, 230–237. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Biodiver-
Bell, K.E., Donnelly, M.A., 2006. Influence of forest fragmentation on community structure sity hotspots for conservation priorities. Nature 403, 853–858.
of frogs and lizards in northeastern Costa Rica. Conserv. Biol. 20, 1750–1760. Ndriantsoa, S.H., Riemann, J.C., Vences, M., Klages, J., Raminosoa, N.R., Rödel, M.-O., Glos, J.,
Bickford, D., Ng, T.H., Qie, L., Kudavidanage, E.P., Bradshaw, C.J.A., 2010. Forest fragment 2013. A new Stumpffia (Amphibia: Anura: Microhylidae) from the Ranomafana re-
and breeding habitat characteristics explain frog diversity and abundance in gion, south-eastern Madagascar. Zootaxa 3636, 575–589.
Singapore. Biotropica 42, 119–125. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O'Hara, R.B., Simpson, G.L.,
Birkinshaw, C., Randrianjanahary, M., 2007. The effects of cyclone Hudah on the forest of Solymos, P., Stevens, M.H.H., Wagner, H., 2011. Vegan: community ecology package,
Masoala Peninsula, Madagascar. Madagascar Conserv. Dev. 2, 17–20. R package version 2.0-2 URL: http://CRAN.R-project.org/package=vegan.
Brown, K.A., Flynn, D.F.B., Abram, N.K., Ingram, J.C., Johnson, S.E., Wright, P., 2011. Peloso, P.L., Frost, D.R., Richards, S.J., Rodrigues, M.T., Donnellan, S., Matsui, M., Raxworthy,
Assessing natural resource use by forest-reliant communities in Madagascar using C.J., Biju, S., Lemmon, E.M., Lemmon, A.R., Wheeler, W.C., 2015. The impact of an-
functional diversity and functional redundancy metrics. PLoS ONE 6, e24107. chored phylogenomics and taxon sampling on phylogenetic inference in narrow-
Cabrera-Guzmán, E., Reynoso, V.H., 2012. Amphibian and reptile communities of mouthed frogs (Anura, Microhylidae). Cladistics http://dx.doi.org/10.1111/cla.12118.
rainforest fragments: minimum patch size to support high richness and abundance. Pineda, E., Halffter, G., 2004. Species diversity and habitat fragmentation: frogs in a
Biodivers. Conserv. 21, 3243–3265. tropical montane landscape in Mexico. Biol. Conserv. 117, 499–508.
Daily, G.C., 2001. Ecological forecasts. Nature 411, 245. R Development Core Team, 2011. R: a Language and Environment for Statistical Comput-
DeFries, R., Hansen, A., Newton, A.C., Hansen, M.C., 2005. Increasing isolation of protected ing. R Foundation for Statistical Computing, Vienna, Austria (ISBN: 3-900051-07-0,
areas in tropical forests over the past twenty years. Ecol. Appl. 15, 19–26. 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 conser-
vation. 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 extinc-
tion 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 Am-
azonian 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 un-
derestimation 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.