lable at ScienceDirect

Acta Oecologica 58 (2014) 12e21

Contents lists avai

Acta Oecologica

journal homepage: www.elsevier .com/locate/actoec

Original article

Habitat, food, and climate affecting leaf litter anuran assemblagesin an Atlantic Forest remnant

Camila Rabelo Rievers a, Maria Rita Silvério Pires a, Paula Cabral Eterovick b,*

a Programa de Pós Graduação em Ecologia de Biomas Tropicais, Universidade Federal de Ouro Preto, 35400-000 Belo Horizonte, MG, Brazilb Programa de Pós Graduação em Zoologia de Vertebrados, Pontifícia Universidade Católica de Minas Gerais, 30535-610 Belo Horizonte, MG, Brazil

a r t i c l e i n f o

Article history:Received 1 July 2013Accepted 7 April 2014Available online xxx

Keywords:AnuraLeaf litter structureInvertebrate availabilityMicroclimateSpatial distributionTemporal distribution

* Corresponding author.E-mail addresses: pceterovick@gmail.com,

C. Eterovick).

http://dx.doi.org/10.1016/j.actao.2014.04.0031146-609X/� 2014 Elsevier Masson SAS. All rights re

a b s t r a c t

Leaf litter anuran assemblages include both species that have terrestrial development and species that,during the breeding season, aggregate around bodies of water where their tadpoles develop. The re-sources used by these two groups in the leaf litter are likely to differ, as well as their sampled speciesrichness, abundance and biomass as resource availability changes. We conducted a 12-month survey ofleaf litter anuran assemblages at three forest areas in the largest Atlantic Forest remnant in the state ofMinas Gerais in southeastern Brazil. Each month we estimated, based on capture rates, anuran speciesrichness, abundance, and biomass as assemblage descriptors. We also measured variables that couldpotentially affect these descriptors in space and time: invertebrate litter fauna (abundance and richnessof taxa), leaf litter biomass, and microclimatic conditions (air humidity, air and soil temperature, soilwater content, and rainfall). We tested for differences in these variables among areas. We used generallinear models to search for the variables that best explained variation in anuran abundance (based oncapture rates) throughout the year. We analyzed species with terrestrial development (TD) and withaquatic larvae (AL) separately. We recorded 326 anurans of 15 species. Sampled anuran abundance(correlated to species richness and biomass) was explained by air humidity and/or invertebrate abun-dance for species with TD, and by soil water content or air humidity and leaf litter biomass for specieswith AL. The variability in the results of studies on leaf litter frogs that try to find variables to explainchanges in community descriptors may be due to spatial variation of resources among areas and also tothe fact that TD and AL species are frequently analyzed together, when in fact they are likely to showdifferent responses to resources present in the leaf litter habitat, reflected on capture rates.

� 2014 Elsevier Masson SAS. All rights reserved.

1. Introduction

Neotropical leaf litter anurans are mostly represented by speciesthat lay eggs on the forest floor and species that aggregate aroundpuddles for breeding (Donnelly and Crump, 1998; Haddad et al.,2008; Walting and Donnelly, 2002). The former depend on theleaf litter habitat for foraging, shelter, and breeding with their eggsbeing deposited in the leaf litter itself (e.g., Brachycephalidae,Craugastoridae, and Eleutherodactylidae). The latter live on the leaflitter and use it to forage and for shelter, but usually breed anddeposit eggs at flooded areas or puddles on the forest floor (e.g.,Leiuperidae, Leptodactylidae, Microhylidae, and Bufonidae;Donnelly and Crump,1998; Haddad et al., 2008). Other species may

eterovick@yahoo.com (P.

served.

occasionally move through the leaf litter although they are essen-tially arboreal (e.g., Hylidae; Rocha et al., 2007). From an ecologicalstandpoint, forest leaf litter is an interesting habitat to study theresources that are important for maintaining resident anuranpopulations, because resources used for both reproduction andsubsistence may be present.

Previous research suggests that the density and diversity of leaflitter frogs is mostly related to three main niche dimensions(sensu Pianka, 1973): food (e.g., productivity, availability of foodresources), space (altitude, habitat quality, habitat heterogeneity)and time (seasonality of climatic conditions). Many of thesestudies on leaf litter anuran assemblages have been conducted inBrazil, Costa Rica, and east Africa (e.g., Gardner et al., 2007;Heinen, 1992; Lieberman, 1986; Rocha et al., 2007). In Brazil,such studies took place in the Amazon (Allmon, 1991; Gascon,1996; Rojas-Ahumada et al., 2012) and in the Atlantic Forest bi-omes (Dixo and Verdade, 2006; Rocha et al., 2001; Van Sluys et al.,2007). However, none of these studies assessed variables related

Fig. 1. Map showing the location of the Parque Estadual do Rio Doce (PERD) in the state of Minas Gerais in southeastern Brazil, and the three areas sampled within the park.

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e21 13

to food, habitat, and temporal (seasonal) effects simultaneously.Such factors may be related and present synergetic effects, so amodel selection approach is important to detect which factors orcombination of factors best explain variation in leaf litter com-munity attributes such as species richness, abundance andbiomass, also reflected on capture rates. Additionally, responses tothese factors are likely to differ between species with terrestrialdevelopment and those with aquatic larvae (see Becker et al.,2007, 2010). Species with aquatic larvae must migrate tobreeding sites to reproduce, and the existence of suitable migra-tion routes, as well as specific frog traits (e.g., tolerance to avail-able habitat conditions, migration abilities), are importantdeterminants of their success (see Becker et al., 2007; Dixo andMetzger, 2010). On the other hand, for species with terrestrialdevelopment the leaf litter may assume a greater importance forreproduction rather than movement. Regardless of reproductivemode, all species may benefit from food availability in leaf litterhabitats. The importance of taking differences in life-history traits,such as reproductive modes, of species into account for planningefficient conservation strategies has already been recognized(Becker et al., 2010).

We aimed to study factors related to the three main niche di-mensions of food, space, and time concomitantly and identifywhich of these has the greatest influence on species richness,abundance and biomass (considered as community descriptors)for species with terrestrial development (TD) and species withaquatic larvae (AL) separately. Since we sampled anurans withpitfall traps, community descriptors are based on capture rates.Our main questions were: (1) Do TD frogs respond to availableresources in a different way than AL species? That is, woulddifferent variables explain community descriptors of species withTD and AL? (2) Is temporal variation in community descriptorsaffected by the same variables at forest sites within the sameforest remnant?

2. Materials and methods

2.1. Study site

The study was conducted at the Parque Estadual do Rio Doce(PERD), in the southwestern portion of the state of Minas Gerais,Brazil (19�4801800�19�2902400S, 42�2801800�48�3803000W, DatumWGS 84; Fig. 1). PERD is the largest and best-preserved continuousremnant of Atlantic Forest in Minas Gerais, and encompasses35,976 ha of forest in various successional stages (Seeliger et al.,2002). The vegetation is tropical, semideciduous and seasonal,with lianas and epiphytes. The park shelters more than 56 inde-pendent lakes, being the largest lake complex in Latin America (IEF,2008). The terrain is hilly, with altitudes that vary from 200 to500 m above sea level (IEF, 2008) and the soils are acidic with lownatural fertility (Metzker, 2007; Teixeira et al., 1989). Mean annualtemperatures are 20e22 �C, and annual rainfall averages1250�1500 mm, with rains concentrated in the warm season(October to March). The dry, cold season extends from April toSeptember (IEF, 2008). A preliminary inventory of anurans found 38species in PERD, most of which have broad geographic distributions(Feio et al., 1998).

We sampled leaf litter anurans in three forest areas (Preta,Central, and Aníbal) considered as replicates within the park(Fig. 1), and chosen to represent the heterogeneity of the forest inPERD. Preta is covered by primary forest on a flat terrain (meandeclivity ¼ �2.6�). Trees reach about 30 m and show a highabundance of epiphytes. The low declivity and proximity to theDoce River and several lakes promote high humidity in this area(Metzker, 2007). In Preta, traps (described ahead) were positioned370 m from the closest body of water. Central is located on a hillyterrain (declivity can reach�22�, with amean of�19.6�). The forestis in earlier successional stages than at Preta and with shorter treesreaching up to only 25 m. The lower vegetation strata are denser

Fig. 2. Sampling design employed in each sampling area, showing the sets of pitfalltraps for anurans, pitfall traps for invertebrates, and leaf litter collectors set around a10,000 m2 quadrat.

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e2114

than at Preta, with abundance of herbs and lianas (Metzker, 2007).In Central, traps were positioned 350 m from the closest body ofwater. Aníbal is covered by secondary forest and is the steepest area(declivities can reach �28�, with a mean of �21.3�). Trees reach upto 15�20 m, and the lowest strata are densely covered by herbs,bambus, and lianas, but small openings are common (Metzker,2007). In this area, traps were positioned 140 m from the closestbody of water. Preta is located 5.4 km from Central and 6.8 km fromAníbal, and Central is located 12.3 km from Aníbal (Fig. 1). Weaimed to place all sets of pitfall traps as far as possible from bodiesof water to avoid capturing anurans en route to a breeding site.Distances varied (140e370 m) because of pond distribution at eacharea and feasibility of access.

2.2. Sampling procedures

We performed anuran sampling with 25-l pitfall traps (Cechinand Martins, 2000; Corn, 1994), placed in sets of four in a Y-shaped arrangement and connected by 10 m long and 50 cm highplastic fences. We made an attempt to sample with quadrats(with both 4 and 5 m sides) as a complementary method, and wealso conducted some occasional visual searches in an attempt toregister other species that might not fall in the traps. Althoughthe species abundance rank was the same recorded with pitfalltraps and quadrats, quadrats provided a very small sample sizethat would not be enough for the statistic analyses we intendedto perform (R. Laia et al., unpublished data). Visual searches didnot reveal leaf litter anuran species that were not sampled by thetraps. Sampling with pitfall traps is obviously influenced byanuran movement, so our community descriptors (species rich-ness, abundance and biomass) must be interpreted as a combi-nation of number of frogs present at the area and theirmovement patterns. Notwithstanding, pitfall traps are consid-ered an efficient sampling method to estimate relative abun-dance of species (Gardner et al., 2007; Heyer et al., 1994;Magurran, 1988), and it has been shown to work well for leaflitter fauna (Cechin and Martins, 2000; Corn, 1994; Liebermanand Dock, 1982).

We installed eight sets of traps in each sampling area with twosets on each side of a 10,000m2 quadrat for a total of 96 pitfall traps(32 per area; see Fig. 2). We opened and checked pitfall traps forfour days each month for a total of 48 sampling days from April2007 to March 2008. Due to logistical reasons traps could not bechecked daily so they were left open for 48 h a month with a smallvolume (about 0.6 l) of formalin (5%, to minimize smell and avoidrepelling the frogs) placed in the traps to preserve specimens.Sampling bias caused by the removal of individuals from the pop-ulation could potentially mask a seasonal effect. However, we ex-pected more captures during the rainy season and we started ourstudy at the onset of the dry season. Thus, any detection of a sea-sonal effect of more frogs during the rainy season at all would be aconservative underestimate. We correlated number of monthssampled with number of individuals captured per month for eachspecies. We expected a negative correlation if sampling wasreducing the likeliness of captures for a given species. We found nosignificant results (results not shown).

We stored captured specimens in 70% ethanol, weighed them ona digital scale to the closest 0.001 g and deposited them at theherpetological collection of the Departamento de Zoologia of theUniversidade Federal de Minas Gerais (DZUFMG). We identifiedspecimens based on herpetological museum collections (DZUFMGand MCNAM e Amphibian Collection of the Museu de CiênciasNaturais da Pontifícia Universidade Católica de Minas Gerais) andspecies descriptions. We followed the taxonomic classificationadopted by Frost (2013).

To sample the availability of prey items to anurans in the leaflitter we sampled invertebrates with pitfall traps of about 1 l (8 cmdiameter) made from recycled plastic bottles. We installed oneinvertebrate pitfall trap next to each anuran pitfall trap set, 2 maway from the central anuran trap (Fig. 2), for a total of 24 inver-tebrate pitfall traps (8 per area). These traps remained openwhenever their adjacent anuran trap was open. We filled thesetraps with water and dish soap, and preserved collected specimensin ethanol 70%. We identified arthropods to the taxonomic rank oforder with the aid of a stereomicroscope (Opton 10�) and identi-fication keys (Borror et al., 1992; ITIS, 2000). Because invertebrateswere not identified to species, invertebrate richness describesrichness of higher taxa (mostly at the rank of order).

To evaluate litterfall, we used square collectors made of a 2 mmmesh netting with area of 1 m2 and depth of 23 cm, positioned80 cm from the ground on ametal pole. We placed 12 collectors perarea, 3 at each side of the 10,000m2 quadrat, separated by 20m andlocated 5 m from the pitfall trap sets. We put all the material storedin the collectors during eachmonth in plastic bags and kept them inrefrigerators (�4 �C) until processing. Processing consisted of dry-ing the material in paper bags at 65 �C in a plant drier (Orion 515 e

Fanem) and then weighing it to the nearest 0.1 g on a digital scale(Marte, precision 0.001 g).

We obtained climatic data from the Instituto de PesquisasEspaciais (INPE), collected by a meteorological station installedinside the park. Climatic data included soil temperature and watercontent, rainfall, air temperature and relative humidity. For rainfallwe used data from the sampling period (4 days) plus the threeprevious days because rainfall distribution was not uniformthroughout each month. For the other data we used monthly meanvalues, which showed highly significant correlations with thevalues for the same 7-day period (results not shown). The meteo-rological station is located 15.7 km from Aníbal, 14 km from Pretaand 13 km from Central. Although meteorological data can varyspatially, we considered the likelihood of this to be minimal among

Table 1Number of individuals recorded and biomass for anuran species at the three studied areas in the Parque Estadual do Rio Doce (PERD): Preta, Central, and Aníbal. Mean biomassrefers to individual means.*Species with terrestrial development.

Species Preta Central Aníbal Total Total biomass (g) Mean biomass (g)

Brachycephalidae*Ischnocnema parva (Girard, 1853) 19 2 0 21 10.9 0.51*Ischnocnema surda (Canedo et al., 2010)a 7 1 2 10 8.5 0.84

BufonidaeRhinella pombali (Baldissera, Caramaschi

and Haddad, 2004)16 1 0 17 229.5 13.49

Craugastoridae*Haddadus binotatus (Spix, 1824) 17 18 3 38 127.8 3.36

Eleutherodactylidae*Adelophryne sp. n.b 0 2 0 2 0.5 0.26

LeptodactylidaeLeptodactylus marmoratus (Steindachner, 1867) 0 1 0 1 0.7 0.70Leptodactylus latrans (Steffen, 1815) 7 0 3 10 41.1 4.57Physalaemus cuvieri (Fitzinger, 1826) 0 1 2 3 4.5 1.49Physalaemus sp. (cf. crombiei)c 2 156 3 161 213.2 1.32

MicrohylidaeChiasmocleis schubarti Bokermann, 1952 15 6 2 23 35.1 1.52Stereocyclops incrassatus (Cope, 1870) 32 1 4 37 208.4 5.63

Total 115 189 18 323 880.2 3.42

a Ischnocnema surda was recently described as occurring in the municipalities of Ouro Preto, Rio Piracicaba, and Santa Bárbara (Canedo et al., 2010).b Adelophryne sp. n. resembles a new species found in the municipality of Catas Altas (C. R. Rievers personal observation).c Physalaemus sp. (cf. crombiei), a species from the P. signifer group, resembles Physalaemus crombiei Heyer andWolf, 1989, which is only known from its type locality in the

municipality of Santa Tereza in the state of Espírito Santo (Heyer andWolf, 1989). The call of the species found at PERD is very similar to that of P. crombiei, but more studies areneeded to confirm the species identity.

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e21 15

the sampled areas. To confirm this prediction, we measured airtemperature and humidity in the field every month just prior tosampling and near each pitfall trap set using a thermohygrometer(Kestrel) positioned just above the leaf litter. We then compared allmeasurements taken for temperature and humidity among thethree areas with analyses of variance mixed models in the softwareSystat 12 (Wilkinson, 2007), with temperature and humidity asdependent variables, areas as fixed effect and months as randomeffect.

Fig. 3. Rarefaction curves based on Jackknife I species richness estimator for leaf litteranurans in the three areas sampled in the Parque Estadual do Rio Doce (PERD) insoutheastern Brazil.

2.3. Statistical analyses

We estimated species richness for each area (Preta, Central andAníbal) using Jackknife I of the software EstimateS (Colwell, 2006)to evaluate how effective our sampling effort was at recording allspecies present. We compared the three areas regarding: (1)invertebrate abundance (2) invertebrate richness of taxa, and (3)leaf litter biomass using KruskaleWallis tests. We also comparedtemporal distribution of these variables for the three areasthroughout the year using circular correlations in the softwareBioEstat 5.0 (Ayres et al., 2005) to see if monthly variation wasequivalent among areas and, using Mann Whitney tests, betweendry and wet seasons to check for seasonal variation. We aimed touse these results to aid to our understanding of anuran abundancevariation among areas and throughout the year.

We conducted general linear models (GLM) in the softwareSystat 12 (Wilkinson, 2007) to select the best independent vari-ables to explain temporal variability in the dependent variables ofanuran abundance, richness, and biomass. As independent vari-ables, we considered invertebrate abundance and richness of taxa,and leaf litter biomass, air and soil temperatures, rainfall, amount ofwater in the soil, and relative air humidity. Before we conductedmodel selection using Akaike criteria (AIC; Quinn and Keough,

2002), we tested each variable for normality and log transformedsome of them to achieve normality (invertebrate abundance andabundance of TD and AL species at Central, biomass of TD species atPreta). We also conducted Pearson or Spearman correlations (fornormal and non-normal data, respectively) using the softwareBioEstat 5.0 (Ayres et al., 2005). For highly and significantly corre-lated variables, one was chosen to include in the model selectionunder the assumption that either would provide the same infor-mation. We used Bonferroni correction to reduce the probability ofType I errors (false positives; Rice, 1989). We conducted this wholeprocedure for Preta and Central separately, once considering justTD species and once considering just AL species. Aníbal was notanalyzed due to small sample size.

Table 2Temporal distribution of leaf litter anurans from April 2007 to March 2008 at the Parque Estadual do Rio Doce (PERD), including monthly occurrence (number of individuals),species richness, abundance and biomass, invertebrate abundance and number of taxa, dry leaf litter biomass, and climatic variables. Climatic variables refer to mean monthlyvalues, rainfall represents total for seven days (three days before and four during sampling) obtained from the meteorological station. The other variables are presented astotals obtained for the three areas together.

Species Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar

BrachycephalidaeIschnocnema parva 1 4 9 3 3 1Ischnocnema surda 1 1 1 2 4 1

BufonidaeRhinella pombali 7 1 4 5

CraugastoridaeHaddadus binotatus 1 2 4 7 10 7 3 1 3

EleutherodactylidaeAdelophryne sp. n. 1 1

LeptodactylidaeLeptodactylus marmoratus 1Leptodactylus latrans 6 1 1 1 1Physalaemus cuvieri 2 1Physalaemus sp. (cf. crombiei) 4 5 6 6 17 37 9 9 64 1 3

MicrohylidaeChiasmocleis schubarti 1 1 20 1Stereocyclops incrassatus 1 2 1 32 1

Anuran species richness 5 5 4 4 6 5 3 3 9 3 4 4Anuran abundance 14 9 10 12 32 58 19 14 133 3 9 10Anuran biomass (g) 36.2 10.7 13.9 17.0 40.5 65.2 37.8 19.8 395.0 18.7 66.6 158.8Leaf litter invertebrate

abundance (numberof taxa)

670 (9) 671 (10) 408 (10) 215 (10) 265 (11) 299 (10) 688 (10) 1683 (10) 2813 (11) 1394 (10) 2939 (10) 1970 (10)

Dry leaf litter biomass (g) 172.0 139.2 97.0 226.2 205.0 315.9 368.0 438.6 83.6 212.0 268.4 260.3Rainfall (mm) 21.5 0.5 0.0 0.0 0.0 1.75 0.0 27.0 29.5 4.0 0.0 27.0Mean air temperature (�C) 22.61 18.62 18.25 17.42 21.31 21.55 22.82 23.47 24.96 23.86 25.26 24.51Humidity (%) 85.16 88.00 82.12 83.98 75.18 70.92 64.14 76.83 79.16 78.23 76.25 84.43Mean soil temperature (�C) 26.46 23.88 22.81 22.31 25.78 27.02 29.62 27.70 31.24 28.21 31.76 28.39Soil water content (%) 0.38 0.27 0.28 0.26 0.25 0.24 0.25 0.35 0.37 0.29 0.30 0.41

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e2116

3. Results

3.1. Composition and structure of leaf litter anuran assemblages atPERD

During the whole study 326 anurans representing 15 speciesfrom seven families (Table 1) were sampled with pitfall traps. Frogsof the family Hylidaewere excluded from our analyses because thesefrogs are arboreal and were probably captured in pitfall traps whilemoving through the leaf litter (see Giaretta et al., 1999; Rocha et al.,2007). From the remaining individuals, 115 (35.7%) were sampled atPreta, 189 (58.7%) at Central, and only 18 (5.6%) at Aníbal. Thenumber of species recorded was larger for Central (n ¼ 10), followedby Preta (n ¼ 8) and Aníbal (n ¼ 7). Estimated species richness(mean � standard deviation) was 8.91 � 0.91 for Preta, 13.58 � 2.12for Central, and 7.91�0.91 for Aníbal. The rarefaction curves becamestable for Preta and Aníbal, but not for Central (Fig. 3).

Most anurans were collected during the rainy season (58%), butfrom those collected during the dry season almost half (43%) werecollected in September, just prior to the onset of the rainy season.The number of species recorded per month varied from three tonine (Table 2).

3.2. Distribution of invertebrates in the leaf litter at PERD

Aníbal was the area with the highest invertebrate abundance(6336 individuals), followed by Central (5435 individuals) and

Preta (2244 individuals). The three areas were significantlydifferent regarding invertebrate abundance throughout the year(KruskaleWallis test statistic ¼ 7.166, p ¼ 0.028, df ¼ 2) as well asinvertebrate richness of taxa (KruskaleWallis test statistic ¼ 7.185,p ¼ 0.028, df ¼ 2). Invertebrate abundance was higher during therainy season (U < 0.0005, p ¼ 0.004, df ¼ 1) but invertebraterichness was nearly constant throughout the year (U ¼ 15.5,p¼ 0.598, df¼ 1; see Table 2). In general, the taxonomic groups thatwere better represented were Collembola, Hymenoptera, Acari,Orthoptera, Coleoptera, and Diptera (Appendix 1). The areas werenot correlated regarding invertebrate abundance throughout theyear (Preta and Central: r ¼ 0.108, p ¼ 0.726, df ¼ 11; Preta andAníbal: r ¼ �0.177, p ¼ 0.563, df ¼ 11; Central and Aníbal:r ¼ �0.076, p ¼ 0.804, df ¼ 11). However, all three areas showedpeaks of abundance during the rainy season, fromOctober toMarch(Table 2).

3.3. Leaf litter characterization at sampling areas

Amounts of leaf litter varied among areas (KruskaleWallis teststatistic¼ 7.459, p¼ 0.024, df¼ 2), and were correlated throughoutthe year between Preta and Aníbal (r¼ 0.613, p¼ 0.034, df¼ 10) butnot between Central and Aníbal (r ¼ 0.080, p ¼ 0.804, df ¼ 10) orCentral and Preta (r ¼ 0.006, p ¼ 0.986, df ¼ 10). Leaf litter biomassdid not differ significantly between seasons (U ¼ 10.0, p ¼ 0.200,df ¼ 1), although its highest values were recorded on October andNovember (Table 2).

Fig. 4. Relationships among variables selected in the best models to explain temporal variation (from April 2007 to March 2008) in leaf litter anuran abundance at the ParqueEstadual do Rio Doce (PERD). Filled symbols refer to Preta (A, B), open symbols refer to Central (CeF). Circles refer to species with terrestrial development (TD), triangles refer tospecies with aquatic larvae (AL).

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e21 17

3.4. Climatic variables at PERD

Rainfall showed peaks before/during some sampling periods,mostly within the rainy season (in November, December, March,and April; Table 2). Air and soil temperatures varied during thestudy period, with a cooler period between April and September2007 and higher temperatures during the rainy season, fromOctober to March. Air humidity did not show a seasonal pattern,whereas soil water content was usually higher during the rainyseason (Table 2). Temperatures (measured in the field) variedamong months (MS ¼ 34.09, F ¼ 8.132, df ¼ 33, p ¼ 0.007) but notamong areas (MS ¼ 2.45, F ¼ 0.584, df ¼ 33, p ¼ 0.450). Humidity(measured in the field) did not vary among months (MS ¼ 16.13,F ¼ 0.250, df ¼ 33, p ¼ 0.621) neither among areas (MS ¼ 1.20,F ¼ 0.019, df ¼ 33, p ¼ 0.892).

3.5. Anuran assemblage response to prey availability, leaf littercharacteristics and climatic variables

We excluded Aníbal from the analyses of temporal variation inanuran assemblages because of its low sample size. For species withTD (Table 1), species richness, abundance and biomass were allrelated in both Preta and Central (Preta: richness vs. abundance:r ¼ 0.725, p ¼ 0.008; richness vs. biomass: r ¼ 0.725, p ¼ 0.008;abundance vs. biomass: r ¼ 0.640, p ¼ 0.025; Central: richness vs.abundance: rs¼ 0.888, p¼ 0.0001; richness vs. biomass: rs¼ 0.845,p ¼ 0.0005; abundance vs. biomass: rs ¼ 0.980, p < 0.0001; n ¼ 12pairs for all analyses). For the species with AL (the remaining spe-cies, see Table 1) species richness, abundance and biomass were allrelated at Preta (richness vs. abundance: rs ¼ 0.871, p ¼ 0.0002;richness vs. biomass: rs ¼ 0.838, p ¼ 0.0007; abundance vs.

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e2118

biomass: rs ¼ 0.940, p < 0.0001; n ¼ 12 pairs), but not at Central,where just anuran abundance and biomass were related to eachother (richness vs. abundance: rs ¼ 0.225, p ¼ 0.482; richness vs.biomass: rs ¼ 0.423, p ¼ 0.171; abundance vs. biomass: rs ¼ 0.676,p ¼ 0.016; n ¼ 12 pairs). The different relationships among inde-pendent variables for species with AL at Central was probablycaused by the smaller variation in species richness among monthsso we used just anuran abundance as independent variable for allthe analyses but the analysis for TD species at Preta (in this case weused log transformed biomass because abundance data was notnormal).

Invertebrate abundance and richness of taxa (Appendix 1) werehighly correlated in both areas (Preta: rs ¼ 0.706, p ¼ 0.010, n ¼ 12pairs, Central: rs ¼ 0.887, p ¼ 0.0001, n ¼ 12 pairs), air temperaturewas correlated with soil temperature (rs¼ 0.893, p< 0.0001, n¼ 12pairs) and rainfall was correlated with soil water content(rs ¼ 0.688, p ¼ 0.013, n ¼ 12 pairs). We used invertebrate abun-dance, soil temperature and soil water content for analyses of bothareas.

The temporal variation in abundance of TD species was bestexplained by humidity alone at Preta (r ¼ 0.719, MS ¼ 81.04,p ¼ 0.008, df ¼ 10; AIC ¼ 62.18, Fig. 4A); the next best modelsincluded leaf litter biomass (AIC ¼ 62.7) and both variables (hu-midity and leaf litter biomass, AIC¼ 63.1). The temporal variation inabundance of TD species was best explained by humidity andinvertebrate abundance combined at Central (r ¼ 0.938, MS ¼ 1.89,p ¼ 0.005, df ¼ 5; AIC ¼ 8.78; Fig. 4C, E); the next best modelincluded these two variables plus soil water content (AIC ¼ 10.7).For AL species variation was best explained by soil water contentalone at Preta (r¼ 0.858, MS¼ 21.51, p¼ 0.003, df¼ 7; AIC¼ 30.15;Fig. 4B) and by leaf litter biomass and humidity combined at Central(r¼ 0.726, MS¼ 3.97, p¼ 0.034, df¼ 9; AIC¼ 35.82; Fig. 4D, F), withno other significant models.

4. Discussion

4.1. Composition and structure of the leaf litter anuran assemblagesat PERD

The number of leaf litter anuran species recorded in this studywas similar to leaf litter species richness in other Atlantic Forestsites (between eight and 16 species; Giaretta et al., 1997, 1999;Rocha et al., 2000, 2001, 2007; Van Sluys et al., 2007). Anuranabundance, on the other hand, shows a lot more variation amongstudies, which can be due to real variation among sites or samplingmethod/effort. Besides our study, several others show the oldfamily Eleutherodactylidae (now represented, in the present study,by Brachycephalidae, Craugastoridae, and Eleutherodactylidae;sensu Hedges et al., 2008) to be the most common in neotropicalleaf litter anuran communities (e.g., Fauth et al., 1989; Heinen,1992; Rocha et al., 2007).

Sampling effort was apparently not thorough enough to recordall leaf litter anuran species for at least one of the three studiedareas in PERD (Central), but still it was more efficient than quad-rats in this area (R. Laia et al., unpublished data). Rocha et al.(2001), on the other hand, found small quadrats to work well atAtlantic Forest areas in Ilha Grande, southeast Brazil. The successof any particular sampling method may vary from site to site,representing an obstacle for standardization of sampling methodson a large scale. Although our data cannot be compared to datacollected in quadrats, our results have the merit of (1) consideringa broad range of variables and niche dimensions in the sameanalysis and (2) testing their effects on capture rates of TD and ALspecies separately. Both approaches are innovative in leaf litteranuran assemblage studies.

Anuran biomass is affected by both species size and abundance.The large size of Rhinella pombali, for instance, accounted for a highbiomass at Preta, whereas the high abundance of Physalaemus sp.(cf. crombiei) was responsible for increasing the biomass at Central.Still, anuran biomass was related to species richness and diversitythroughout the year, except for AL species at Central. Themaximumvalues of anuran biomass and abundance were recorded during therainy season, especially in December, when many species arebreeding. Thus, movement of frogs to breeding sites or in search formates probably increased capture rates in December. Rainfall alsoreached a peak in this month, contributing to high water content inthe soil (see Table 2). However, the general pattern was a negativerelationship among anuran abundance and humidity or soil watercontent (see Section 4.3; Fig. 4). Biomass showed a peak in Marchdue to the high number of records of R. pombali, a heavy toad.Temporal distribution of leaf litter anurans varies from site to site,what is likely to depend on species composition and climate vari-ables, among other causes. Their abundance and biomass may behigher during the dry season as recorded in Central America(Lieberman,1986;Walting and Donnelly, 2002) and Africa (Gardneret al., 2007), or during the rainy season, as recorded in the Amazon(Allmon, 1991), Costa Rica (Heinen, 1992) and the present study, orthey may not vary considerably between seasons, as recorded atlocalities in southeastern Brazil (Giaretta et al., 1999) and in CostaRica (Scott, 1976).

4.2. Prey availability, leaf litter characteristics and climaticvariables

The dominant invertebrate groups in leaf litter of the studiedareas included Hymenoptera, Orthoptera, and Acari, as was also thecase at a site in Costa Rica (Lieberman, 1986). Other abundantgroups included Collembola, Coleoptera and Diptera. The areasstudied in PERD showed higher invertebrate abundance during therainy season, but the number of taxa recorded varied little amongmonths, remaining relatively constant throughout the year (seeTable 2). Still, the number of invertebrate taxa was correlated toinvertebrate abundance per month. Factors that may influencetemporal distribution of leaf litter invertebrate communitiesinclude predatoreprey interactions, climate, and food availability(Wolda, 1979). Responses of tropical leaf litter invertebrate faunasto climatic factors are variable, so that higher species richness orabundance may be recorded in dry or rainy seasons or seasonaldifferences may not happen at all (e.g., Guyer, 1990; Lieberman andDock, 1982; Lieberman, 1986; Wolda, 1979).

In the studied areas, leaf litter seemed to accumulate by the endof the dry season and then showed a sharp decrease in December.The onset of the rainy season (October and November) may haveincreased litterfall by washing litter down the threes. In thissequence, reduced litterfall may have resulted from the renewal oftrees with production of new leaves. Although this is not the case inthe collectors used to sample litterfall, leaf litter is likely to becarried to lower sites by the high volume of rain registered inDecember. Higher humidity during the rainy season may also havecontributed to leaf litter decomposition, increasing invertebrateabundance and decreasing leaf litter accumulation. Sin et al. (2008)showed introduced Eleutherodactylus coqui (that consumes largeamounts of leaf litter invertebrates) to increase decompositionrates in Hawaii because frog excrements and carcasses acceleratednutrient cycling.

Temperature and humidity did not vary considerably amongareas, and the effects of climate are likely to be more noticeablethroughout the year. At sites in Panama temperature and humiditydid not vary noticeably and did not relate to anuran richness orabundance (Lieberman, 1986). Vargas and Bolaños (1999)

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e21 19

attributed the similarity of herpetofaunas at forest sites inColombia to their similar microclimatic conditions.

4.3. Anuran assemblage response to prey availability, leaf littercharacteristics and climatic variables

Humidity helped explain abundance of TD species in both areas,as well as AL species at Central. Humidity is expected to beimportant for both anuran movement and terrestrial breeding(Duellman and Trueb, 1994). Although this variable was notincluded in the best model to explain abundance of anurans with atadpole stage at Preta, soil water content (that related to rainfall)took its place. However, contrary to what was expected, the re-lationships were all negative (see Fig. 4). A possible explanation fora higher number of captures in drier months is that species movemore out of their breeding season (that is the wet season), thusfalling in pitfall traps more often. During the breeding period,species are likely to use smaller areas that may be either defendedterritories or favored calling sites and their vicinities. However, apeak of rainfall and abundance of anurans occurred in December.The great amount of rainfall may have triggered the reproductiveactivities of many AL species that depend on the formation ofbodies of water to breed (Duellman and Trueb, 1994). Movement ofthese species in search of breeding sites may have caused the in-crease in captures in December. Suchmovement is likely to result ina positive association between rainfall and recorded anuran speciesrichness and abundance at certain sites (e.g., Allmon, 1991; Scott,1976).

Arthropod abundance had a negative relationship with anuranabundance at Central, considering TD species. This may also be aneffect of lower anuran movement during the breeding season thatcoincides with the time when invertebrates are more abundant inthe leaf litter. Many authors have mentioned the possible influenceof prey availability on patterns of anuran abundance (e.g. Fauthet al., 1989; Walting and Donnelly, 2002), but just a few testedthis relationship (Giaretta et al., 1999; Lieberman, 1986; Liebermanand Dock, 1982; Toft, 1980), some of which found a positive rela-tionship between anuran and arthropod abundance (Giaretta et al.,1999; Lieberman, 1986; Toft, 1980).

Leaf litter biomass also showed a negative relationship withanuran abundance. Litterfall, which produces the leaf litterbiomass, peaked at the onset of the rainy season (October,November) prior to the peak in anuran capture (December), whenlitterfall was the lowest. Additionally, January had low anurancapture and high litterfall. Leaf litter biomass may help retainmoisture in the leaf litter habitat. This moisture was reduced afterheavy rains in December, which may have contributed to thenegative relationship between anuran abundance and soilmoisture.

Leaf litter biomass is expected to represent availability ofimportant microhabitats for both invertebrates and anurans, beinga good predictor of number of taxa (McArthur, 1972). Few studieshave evaluated herpetofauna distribution in relation to leaf litterbiomass at sampled areas (Giaretta et al., 1999; Scott, 1976; VanSluys et al., 2007). No significant relationship was found in theAmazon (Allmon, 1991; Rojas-Ahumada et al., 2012), whereas sig-nificant relationships were recorded at sites in southeastern Brazil(Giaretta et al., 1999; Van Sluys et al., 2007) and southeastern Asia(Scott, 1976). Leaf litter may influence arthropod communities(Heinen, 1992) that are involved in leaf litter decomposition (Houet al., 2005), altering the availability of microhabitats for anurans(Scott, 1976). Nutrient cycling resulting from leaf litter decompo-sition can also enhance productivity that may affect consumers,such as frogs, through food webs (via increase in abundance ofherbivorous arthropods; Deichmann et al., 2012).

Amphibians are known to be sensitive to vegetation structureand microclimate (Dixo and Martins, 2008; Keller et al., 2009;Whitfield and Pierce, 2005). Forest species tend to be more abun-dant in areas with dense canopy, high humidity and stable tem-perature (Urbina-Cardona and Londoño-Murcia, 2003). Lieberman(1986) interpreted seasonal variations in leaf litter amphibian dis-tribution as a response to variation in air temperature and hu-midity, or changes in leaf litter depth and biomass/productivity.Donnelly et al. (2005) assumed seasonal variation in rainfall toinfluence arthropod communities in the leaf litter and, conse-quently, prey availability for anurans, affecting their distribution,abundance, and diversity in Costa Rica. Allmon (1991), on the otherhand, found no significant relationship between leaf litter anuranassemblage attributes and habitat variables such as leaf litter depthand biomass, vegetation structure, and microclimate (rainfall andhumidity).

Our results and the available literature (see citations in theabove paragraphs) indicate that all variables likely to influenceleaf litter anuran assemblages are intertwined and anuranresponse may vary according to the species and its needs, whichmay explain why factors influencing leaf litter anuran assem-blages vary so much among localities. It is also important to keepin mind that capture rates in pitfall traps are closely linked toanuran movement through the leaf litter. The effects of anuranactivity levels on capture rates may vary among samplingmethods (e.g., pitfall traps vs. quadrats) and among species withTD or AL, and the effects of such variation on the outcomes ofdifferent studies remain to be quantified. We showed that vari-ables that influence activity levels of species with TD and AL maydiffer in some aspects between them and among close sites, butpart of this variation can be explained through variation inmovement and capture rates, which influence the outcomes ofdifferent sampling methods.

5. Conclusion

Based on previous studies on leaf litter anuran assemblages,several environmental factors have been shown to influenceanuran species distribution and activity levels, however, therelative importance of each factor may vary from site to site andthroughout the year. Such variation, combined with samplingbias, can produce even more different outcomes among studies.Our study showed an important relationship among humidityand anuran capture rates, what may be more an effect of anuranmovement patterns than abundance itself. Separating the effectsof abundance per se and activity levels on capture rates isimportant to understand the underlying mechanisms that shapeleaf litter anuran assemblages and to compare different areas.Besides, species with different needs are frequently pooled andanalyzed together, masking possible patterns. Studies on leaflitter anuran communities encompassing large geographical andtemporal scales and combining different sampling methodswould be interesting to separate local, from regional, fromtemporal variation and then identify whether there are factorsthat assume a greater importance at any of these scales.

Acknowledgments

A Research Productivity grant (301077/2010-0) was provided toP. C. Eterovick by CNPq. Permits were provided by IEF (136/06) andIbama (323/06-NUFAS-MG e 02015.007378/06-02). We would liketo thank the INPE (Instituto Nacional de Pesquisas Espaciais) forproviding climatic data from the meteorological station at PERD(Parque Estadual do Rio Doce), to the whole team of PERD forimportant logistical support, to L. B. Nascimento, F. S. F. Leite and

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e2120

Carlos Alberto Cruz for help in some specimens identification, andseveral biologists for help during fieldwork: Ivan Cordeiro, RobertaCarmo, Luiza Passos, Sara Araújo, Adriane Silva, Karina Lima, Lud-mila Faria, Leonardo Lobato, Alexandre Bahia, Arquimedes Ferreira,Michael Lindemann, Francisco Oliveira, Leonardo Salim, Janaina,Thiago Metzker, André Hirsch, Rafaela Rodrigues, Rafael Martins,Rodolfo Raimundo, Mariana Drummond, and Milena Wachlevski.We are also thankful to M. Van Sluys and R. N. Feio for helpfulcomments in an earlier version of this manuscript and to E. R. Wildfor revision of the English and suggestions.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.actao.2014.04.003.

References

Allmon, W.D., 1991. A plot study of forest floor litter frogs, Central Amazon, Brazil.J. Trop. Ecol. 7, 503e522.

Ayres, M., Ayres Jr., M., Ayres, D.L., Santos, A.S., 2005. Bioestat 5.0. AplicaçõesEstatísticas nas Áreas das Ciências Biológicas e Médicas. Sociedade CivilMamirauá/MCT-CNPq, Brasília, Brasil.

Becker, C.G., Fonseca, C.R., Haddad, C.F.B., Batista, R.F., Prado, P.I., 2007. Habitat splitand the global decline of amphibians. Science 318, 1775e1777.

Becker, C.G., Loyola, R.D., Haddad, C.F.B., Zamudio, K.R., 2010. Integrating specieslife-history traits and patterns of deforestation in amphibian conservationplanning. Divers. Distrib. 16, 10e19.

Borror, D.J., Triplehorn, C.A., Johnson, N.F., 1992. An Introduction to the Study ofInsects, sixth ed. Harcourt Brace College Publishers, Philadelphia, Pennsylvania.

Canedo, C., Pimenta, B.V.S., Leite, F.S.F., Caramaschi, U., 2010. New species ofIschnocnema (Anura: Brachycephalidae) from the state of Minas Gerais, south-eastern Brazil, with comments on the I. verrucosa species series. Copeia 2010,629e634.

Cechin, S.Z., Martins, M., 2000. Eficiência de armadilhas de queda (pitfall traps) emamostragens de anfíbios e répteis no Brasil. Rev. Bras. Zool. 17, 729e740.

Colwell, R.K., 2006. EstimateS, Version 8.0: Statistical Estimation of Species Rich-ness and Shared Species from Samples (Software and user’s guide) (13 October2011). Available at: http://viceroy.eeb.uconn.edu/EstimateS.

Corn, P.S., 1994. Standard techniques for inventory and monitoring e straight-linedrift fences and pitfalls traps. In: Heyer, W.R., Donnelly, M.A.,McDiarmid, R.W., Hayek, L.C., Foster, M.S. (Eds.), Measuring and MonitoringBiological Diversity: Standard Methods for Amphibians. Smithsonian InstitutionPress, Washington, DC, USA, pp. 118e124.

Deichmann, J.L., Toft, C.A., Deichmann, P.M., Lima, A.P., Williamson, G.B., 2012.Neotropical primary productivity affects biomass of the leaf-litter herpetofau-nal assemblage. J. Trop. Ecol. 28, 427e435.

Dixo, M., Martins, M., 2008. Are leaf-litter frogs and lizards affected by edge effectsdue to forest fragmentation in Brazilian Atlantic forest? J. Trop. Ecol. 24, 551e554.

Dixo, M., Metzger, J.P., 2010. The matrix-tolerance hypothesis: an empirical testwith frogs in the Atlantic forest. Biodivers. Conserv. 19, 3059e3071.

Dixo, M., Verdade, V.K., 2006. Herpetofauna de serrapilheira da reserva florestal deMorro Grande, Cotia, São Paulo. Biota Neotrop. 6, 1e20.

Donnelly, M.A., Crump, M.L., 1998. Potential effects of climate change on twoneotropical amphibian assemblages. Clim. Change 39, 541e561.

Donnelly, M.A., Chen, M.H., Watkins, G.G., 2005. Sampling amphibians and rep-tiles in the Iwokrama forest ecosystem. Proc. Acad. Nat. Sci. Phila. 154, 55e69.

Duellman, W.E., Trueb, L., 1994. Biology of Amphibians. McGraw-Hill, New York.Fauth, J.E., Crother, B.I., Slowinski, J.B., 1989. Elevational patterns of species richness,

evenness, and abundance of Costa Rican leaf-litter herpetofauna. Biotropica 21,178e185.

Feio, R.N., Braga, U.M.L., Wiederhecker, H., Santos, P.S., 1998. Anfíbios do ParqueEstadual do Rio Doce (Minas Gerais). Universidade Federal de Viçosa and IEF,Viçosa, Minas Gerais, Brasil.

Frost, D.R., 2013. Amphibian Species of the World: an Online Reference. AmericanMuseum of Natural History, New York, USA. Version 5.6 (9 January, 2013).Electronic database accessible at. http://research.amnh.org/vz/herpetology/amphibia/index.html (accessed on 05.06.13.).

Gardner, T.A., Fitzherbert, E.B., Drewes, R.C., Howell, K.M., Caro, T., 2007. Spatial andtemporal patterns of abundance and diversity of an east African leaf litteramphibian fauna. Biotropica 39, 105e113.

Gascon, C., 1996. Amphibian litter fauna and river barriers in flooded and non-flooded Amazonian rain forest. Biotropica 28, 136e140.

Giaretta, A.A., Sawaya, R.J., Machado, G., Araújo, M.S., Facure, K.G., Medeiros, H.F.,Nunes, R., 1997. Diversity and abundance of litter frogs in altitudinal sites atSerra do Japi, Southeastern Brazil. Rev. Bras. Zool. 14, 341e346.

Giaretta, A.A., Facure, K.G., Sawaya, R.J., Meyer, J.H.M., Chemim, N., 1999. Di-versity and abundance of litter frogs in a montane forest of southeasternBrazil: seasonal and altitudinal changes. Biotropica 31, 669e674.

Guyer, C., 1990. The herpetofauna of La Selva, Costa Rica. In: Gentry, A.H. (Ed.), FourNeotropical Rainforests. Yale University Press, New Haven, Connecticut, USA,pp. 371e385.

Haddad, C.F.B., Toledo, L.F., Prado, C.P.A., 2008. Anfíbios da Mata Atlântica. EditoraBiotrópica, São Paulo, Brazil.

Hedges, S.B., Duellman, W.E., Heinicke, M.P., 2008. New world direct-developingfrogs (Anura: Terrarana): molecular phylogeny, biogeography and conserva-tion. Zootaxa 1373, 1e182.

Heinen, J.T., 1992. Comparisons of the leaf litter herpetofauna in abandoned cacaoplantations and primary rain forest in Costa Rica: some implications for faunalrestoration. Biotropica 24, 431e439.

Heyer, W.R., Wolf, A.J., 1989. Physalaemus crombiei (Amphibia: Leptodactylidae), anew frog species from Espírito Santo, Brazil, with comments on the P. signifergroup. Proc. Biol. Soc. Wash. 102, 500e506.

Heyer, W.R., Donnely, M.A., McDiarmid, R.W., Hayek, L.C., Foster, M.S. (Eds.),1994. Measuring and Monitoring Biological Diversity e Standard Methodsfor Amphibians. Smithsonian Institution Press, Washington and London.

Hou, P.C.L., Zou, X., Huang, C.Y., Chien, H.J., 2005. Plant litter decomposition influ-enced by soil animals and disturbance in a subtropical rainforest of Taiwan.Pedobiologia 49, 539e547.

Instituto Estadual de Florestas (IEF), 2008. Plano de Manejo do Parque Estadual doRio Doce. Available at: http://www.ief.mg.gov.br/component/content/306?task¼view (accessed on 15.04.08.).

Integrated Taxonomic Information System (ITIS), 2000. Available at: http://www.itis.gov/ (accessed on 03.05.08.).

Keller, A., Rödel, M.-O., Linsenmair, K.E., Grafe, T.U., 2009. The importance ofenvironment heterogeneity for species diversity and assemblage structure inBornean stream frogs. J. Anim. Ecol. 78, 305e314.

Lieberman, S., 1986. Ecology of the leaf litter herpetofauna of a neotropical rainforest: La Selva, Costa Rica. Acta Zool. Mex., Nueva Ser. 15, 1e72.

Lieberman, S., Dock, C.F., 1982. Analysis of the leaf-litter arthropod fauna of alowland tropical evergreen forest site (La Selva, Costa Rica). Rev. Biol. Trop. 30,27e34.

Magurran, A.E., 1988. Ecological Diversity and its Measurement. Princeton Univer-sity Press, Princeton.

McArthur, R.H., 1972. Geographical Ecology. Patterns in the Distribution of Species.Harpes and Row, New York.

Metzker, T.I.F., 2007. Estudos sobre a Dinâmica e os Estoques de Carbono da MataAtlântica do Parque Estadual do Rio Doce, Minas Gerais, Brasil (Master thesis).Universidade Federal de Minas Gerais, Minas Gerais, Brasil.

Pianka, E.R., 1973. The structure of lizard communities. Annu. Rev. Ecol. Syst. 4, 53e74.

Quinn, G.P., Keough, M.J., 2002. Experimental Design and Data Analysis for Bi-ologists. Cambridge University Press, Cambridge.

Rice, W.R., 1989. Analyzing tables of statistical tests. Evolution 43, 223e225.Rocha, C.F.D., Van Sluys, M., Alves, M.A.S., Bergallo, H.G., Vrcibradic, D., 2000. Ac-

tivity of leaf-litter frogs: when should frogs be sampled. J. Herpetol. 34, 285e287.

Rocha, C.F.D., Van Sluys, M., Alves, M.A.S., Bergallo, H.G., Vrcibradic, D., 2001. Esti-mates of forest floor litter frog communities: a comparison of two methods.Aust. Ecol. 26, 14e21.

Rocha, C.F.D., Vrcibradic, D., Kiefer, M.C., Almeida-Gomes, M., Borges-Junior, V.N.T.,Carneiro, P.C.F., Marra, R.V., Almeida-Santos, P., Siqueira, C.C., Goyannes-Araújo, P., Fernandes, C.G.A., Rubião, E.C.N., Van Sluys, M., 2007. A survey of theleaf-litter frog assembly from an Atlantic forest area (Reserva Ecológica deGuapiaçu) in Rio de Janeiro State, Brazil, with an estimate of frog densities. Trop.Zool. 20, 99e108.

Rojas-Ahumada, D.P., Landeiro, V.L., Menin, M., 2012. Role of environmental andspatial processes in structuring anuran communities across a tropical rainforest. Aust. Ecol. 37, 865e873.

Scott Jr., N.J., 1976. The abundance and diversity of the herpetofaunas of tropicalforest litter. Biotropica 8, 41e58.

Sin, H., Beard, K.H., Pitt, W.C., 2008. An invasive frog, Eleutherodactylus coqui, in-creases new leaf production and leaf litter decomposition rates throughnutrient cycling in Hawaii. Biol. Invasions 10, 335e345.

Seeliger, U., Cordazzo, C., Barbosa, F.A.R., 2002. Os Sites e o Programa Brasileiro dePesquisas Ecológicas de Longa Duração. Gráfica e Editora O Lutador, BeloHorizonte, Brazil.

Teixeira, J.L., Barros, N.F., Costa, L.M., 1989. Biomassa e conteúdo de nutrientes deduas espécies de eucalipto em diferentes ambientes do Rio Doce e MG. Rev.Árvore 1, 34e50.

Toft, C.A., 1980. Feeding ecology of thirteen syntopic species of anurans in a sea-sonal tropical environment. Oecologia 45, 131e141.

Urbina-Cardona, J.N., Londoño-Murcia, M.C., 2003. Distribuición de la comunidadde herpetofauna asociada a cuatro áreas con diferente grado de perturbaciónem la Isla Gorgona, Pacífico Colombiano. Rev. Acad. Colomb. Cienc. 27, 105e113.

Van Sluys, M., Vrcibradic, D., Alves, M.A.S., Bergallo, H.G., Rocha, C.F.D., 2007.Ecological parameters of the leaf-litter frog community of an Atlantic rain-forest area at Ilha Grande, Rio de Janeiro State, Brazil. Aust. Ecol. 32, 254e260.

C.R. Rievers et al. / Acta Oecologica 58 (2014) 12e21 21

Vargas, F., Bolaños, M.E., 1999. Anfíbios y reptiles presentes em hábitats perturbadosde selva lluviosa tropical en el Bajo Anchicayá Colombiano. Rev. Acad. Colomb.Cienc. Exactas Fis. Nat. 23, 499e511.

Walting, J.I., Donnelly, M.A., 2002. Seasonal patterns of reproduction and abundanceof leaf litter frogs in a Central American rainforest. J. Zool., Lond. 258, 269e276.

Whitfield, S.M., Pierce, M.S.F., 2005. Tree buttress microhabitat use by a neotropicalleaf-litter herpetofauna. J. Herpetol. 39, 192e198.

Wilkinson, L., 2007. SYSTAT: the System for Statistics. SYSTAT Inc, Evanston, IL.Wolda, H., 1979. Abundance and diversity of Homoptera in the canopy of a tropical

forest. Ecol. Entomol. 4, 181e190.