EFFECTS OF TERRA FIRME-FOREST STRUCTURE … 25 433-458.pdfcia da disponibilidade de recurso alimentar na percepção de habitat pelas corujas. Abstract. – Owls are a poorly-studied

433

ORNITOLOGIA NEOTROPICAL 25, 433–458, 2014© The Neotropical Ornithological Society

EFFECTS OF TERRA FIRME-FOREST STRUCTURE ON HABITAT USE BY OWLS (AVES: STRIGIFORMES)

IN CENTRAL BRAZILIAN AMAZONIA

Priscilla Esclarski1 & Renato Cintra2

¹Programa de Pós-Graduação em Ecologia, Instituto Nacional de Pesquisas da Amazônia (INPA), CP 478, 69067-375, Manaus, Amazonas, Brasil. E-mail: [email protected]

²Coordenação de Biodiversidade, Instituto Nacional de Pesquisas da Amazônia (INPA), CP 478, 69067-375, Manaus, Amazonas, Brasil. E-mail: [email protected]

Resumo. – Efeitos da estrutura da Floresta de Terra-Firme no uso de hábitat por corujas (Aves:Strigiformes) na Amazônia central brasileira. – As corujas, apesar de atuarem na regulação de pre-sas e controle biológico, é um grupo ainda pouco estudado quanto à distribuição e abundância, princi-palmente as espécies neotropicais. Os componentes da estrutura da floresta têm sido usados paraexplicar padrões de uso de habitat e a estrutura de comunidades de aves, porém, os estudos relacio-nando tais componentes às corujas concentram-se em espécies do hemisfério norte. O presente estudoanalisou se os componentes de estrutura da floresta influenciam o uso de habitat (ocorrência eabundância) por seis espécies de corujas em uma floresta de terra-firme na Amazônia central do Brasil.Para amostrar ocorrência e abundância foi usado o método playback, em 30 pontos distantes entre siem 1 km, nos meses de outubro e novembro de 2012. Em análise de regressão linear múltipla obtive-mos relação entre a variação na abundância de quatro espécies de corujas e componentes da estruturada floresta associados à disponibilidade de alimento; caburé-da-amazônia (Glaucidium hardyi) e distân-cia do igarapé (p = 0,023), corujinha-orelhuda (Megascops watsonii) e profundidade da serrapilheira (p =0,045), coruja-de-crista (Lophostrix cristata) e troncos mortos no chão (p = 0,042), murucututu (Pulsatrixperspicillata) e troncos mortos no chão (p = 0,009). Em relação a ocorrência, e usando regressão logís-tica múltipla, obtivemos relação somente entre a presença de murucututu e troncos mortos no chão (p =0,050). Assim, a influência dos componentes da estrutura da floresta difere de acordo com a espécie decoruja, demonstrando as diferenças interespecíficas no uso de micro-habitats, porém reflete a importân-cia da disponibilidade de recurso alimentar na percepção de habitat pelas corujas.

Abstract. – Owls are a poorly-studied avian group, despite their well-established role in prey regulationand biological control. For Neotropical species, distribution and abundance are especially poorly known.Structural components of forests have been used to explain patterns of owl habitat use and communitystructure, but such analyses have largely focused on species in the northern hemisphere. The presentstudy examines whether components of forest structure influence habitat use (occurrence and abun-dance) for six species of owls in an upland forest in central Amazonian Brazil. Between October andNovember 2012, a playback method was used to sample occurrence and abundance in 30 points, eachseparated from the next by 1 km. Multiple linear regression analysis revealed relationships between fourowl species and components of forest structure associated with food availability: Amazonian Pygmy Owl(Glaucidium hardyi) and distance to nearest stream (p = 0.023), Northern Tawny-bellied Screech Owl(Megascops watsonii) and leaf-litter depth (p = 0.045), Crested Owl (Lophostrix cristata) and dead fallentrunks on forest floor (p = 0.042), and Spectacled Owl (Pulsatrix perspicillata) and dead fallen trunks onforest floor (p = 0.009). A multiple logistic regression also revealed a significant association (p = 0.050)between the Spectacled Owl and dead fallen trunks on forest floor. The influence of the components of

434

ESCLARSKI & CINTRA

forest structure differs between the species, demonstrating interspecific differences in micro-habitat use,and reflecting the importance of food resource availability in habitat choice. Accepted 12 December2014.

Key words: Amazonian Pygmy Owl, Glaucidium hardyi, Black-Banded Owl, Strix huhula, Crested Owl,Lophostrix cristata, Mottled Owl, Strix virgata, Northern Tawny-bellied Screech Owl, Megascops watsonii,Spectacled Owl, Pulsatrix perspicillata, Amazonian Brazil, environmental heterogeneity, forest-structurecomponents, habitat use, Strigidae.

INTRODUCTION

In terms of habitat ecology, heterogeneity isdefined as the degree of discontinuity withinthe environmental conditions of a landscape(Morrison 1998). These conditions may varyaccording to the composition and structure ofvegetation or according to the flow of energyand other resources essential for a givenorganism. In recent years, based on the con-cepts of habitat and niche, environmentalheterogeneity has been widely used to ex-plain patterns of habitat use and communitystructure (Day 2000, Gaston 2000, Allen &Gillooly 2006). In the sense adopted in thispaper, habitat includes where the organismlives, as well as the conditions for the survivalof the organism, while niche is a description ofthe habitat conditions that meet the minimumrequirements of a given species, so that thereproduction rate is equal to or greater thanthe mortality rate under the impact of envi-ronmental conditions on the population(Chase & Leibold 2003).

Knowledge of how organisms are distrib-uted is an essential prerequisite for effectiveinference of which evolutionary and ecologi-cal factors determine patterns of habitatoccupation and residence (Gayne & Balda1994, Ricklefs 2004, Graham et al. 2006). Theability to use different features of the environ-ment varies greatly between organisms, butmost are able to track changes in the featuresof suitable habitat (Enfjäll & Leimar 2009).The perception of ideal habitat depends on avariety of factors, including morphologicalcharacteristics of the species (Srugley & Chai

1990, Hughes et al. 2007), social structure(Yaber & Rabenold 2002, Le Galliard et al.2005), and life history (Levin 1984) as well asenvironmental factors, such as climate (Best etal. 2007, Hughes et al. 2007), landscape struc-ture (Kuch & Idelberger 2005), and abun-dance of conspecific individuals at a specificsite (Fletcher 2006, 2007).

Environmental heterogeneity influencesdirectly or indirectly the spatial distribution,richness, and composition of an area’s avi-fauna, since individual choices will be influ-enced by such factors as forest compositionand species diversity, vegetation density andstructure, and the status of local ecologicalsuccession (Hildén 1965, Orians 1969, Wiens1969, Whitacre et al. 1990, Robinson 1994,Thiollay 1996, Amaral 2007).

The loss of habitat elements also has adirect effect on birds. Some structural com-ponents of the forest, such as tree density orthe abundance of fallen logs, furnish localizedmicro-habitats for foraging and are known tobe prime factors for site selection by nestingbirds (Rodewald & Yahner 2000, Slaght et al.2013). Despite this, the effect of spatial varia-tion in forest-structure components on habi-tat choice by birds has been little studied(Enfjäll & Leimar 2009), even though suchfactors are important for an improved under-standing of how interactions between organ-isms operate, and for the organization andstructure of populations and ecological com-munities (Begon et al. 1986). Such changesmay impact both the survival of individualbirds and the maintenance of communitystructure. However, because of species-spe-

435

FOREST-STRUCTURE EFFECTS ON OWLS

cific effects neither cause nor effect will beuniform within such a community (Orians1969, Wiens 1969, Cintra & Naka 2012).

Several studies in central Amazonia haverelated forest-structure components to avianoccurrence and abundance (e.g., Borges et al.2004, Barros & Cintra 2009, Cintra et al.2006). Rice et al. (1983) suggested that habitatselection by birds may vary seasonally accord-ing to the availability of food, and that theleaf-litter layer may also be sufficiently impor-tant in habitat choice as to influence the den-sity of ground-foraging by birds-of-prey.

Patterns of distribution and abundance ofnocturnal birds are generally related to envi-ronmental components, such as forest age,availability of cavities, edge distance, and thespatial and temporal abundance of food.According to Kavanagh et al. (1995), environ-mental generalists were more specific towardstheir prey while environmental specialistswere more general in relation to their prey.According to Martin (1998) birds have “pref-erences for micro-habitats,” which reflects theselection of a place to stay and use. Such vari-ation can be highly species-specific so that,even when the same substrate is chosen fornesting, the characteristics of the adjacentvegetation may well be different for each birdspecies.

Because tropical bird communities arerich, complex, and heterogeneous, factorsdetermining community structure are likely tobe diffuse in the way they operate (Cintra &Naka 2012). Consequently, for many speciesof rainforest birds, little is known about theirbiology, including key factors in determiningpresence and abundance, such as habitatrequirements and social structure, among oth-ers (Thiollay 2002).

The ecology of Neotropical raptors is lit-tle known, especially concerning aspects offorest structure used for the selection of habi-tats or territories (Throstom 2000). In somediurnal predators, such as Eurasian Spar-

rowhawk (Accipiter nisus) and Osprey (Pandionhaliaetus), population density is regulated bythe availability of prey and of favorable sitesfor nesting (Newton et al. 1977, Van Daele &Van Daele 1982). The situation is currentlyless certain for nocturnal raptors, especially inthe tropics, where species’ biology is oftenpoorly known, especially in relation to habitatuse and to the question which components offorest structure most directly influence occu-pation and residence (Amaral 2007, Motta-Junior & Braga 2012).

Most studies relating components of for-est structure to the occurrence of owls havebeen conducted in the northern hemisphere.Several authors have suggested that a generalpreference among owls exists for habitats inmature forests, and that differences in thestructure of the preferred habitat may varywith species’ body size (Zwank et al. 1994,Gayne & Balda 1994, Hunter et al. 1995, Man-zur et al. 1998, Peery et al. 1999, La Haye &Gutiérrez 1999). Evidence suggests thatsmaller species often use more open areas andchoose shrubs as nesting sites (McCallum &Gehlbach 1988), while larger species preferdenser canopy and cavities in trunks for nest-ing (Bull et al. 1989, Belthoff & Ritchinson1990).

Dead, broken, but still-standing trunks areoften used by nesting owls (Hershey et al.1998), and areas with a greater number ofthese will support a higher abundance of owls.Food-resource availability has also beenrelated to the occurrence and abundance ofowls, and indeed some species form pairsaccording to the abundance cycles of mainprey (McInvaille & Keith 1974, Ellinson 1980,Smith et al. 1981, Village 1982, Smith & Gil-bert 1984, Sparks et al. 1994). In addition,both the presence of fallen logs and greatdepths of leaf litter have been shown to pro-vide shelter for owls’ prey species, such asrodents, lizards, crickets, spiders, and beetles,which use leaf litter for concealment (Kiltie

436

ESCLARSKI & CINTRA

1981). Such sites can both attract and supportmore owls, and so increase their local abun-dance (Smith & Gilbert 1984).

Sometimes, similar features combine toinfluence both food and nest sites. For exam-ple, this is true for the Blakiston’s Fish Owl(Bubo blakistoni), perhaps one of the best-stud-ied owl species, which has been found to nestin valleys and near to water bodies (Spangen-berg 1965, Pukinski 1973, Surmach 1998,Slaght 2011). Nest occurrence is also relatedto tree-trunk diameter, the availability of still-standing, but dead trees, and fallen logs clog-ging waterways (the latter being associatedwith abundance of salmon, an importantfood: Slaght et al. 2013).

Survival and area-size usage in the Spot-ted Owl (Strix occidentalis caurina) increase withthe proportion of available habitat in late suc-cessional stages and with the amount of edgehabitat, but decrease with the distance fromthe closest neighbor (Schilling et al. 2013),while the area used and breeding periodexpand in more fragmented habitats. Thisaccords with the suggestion by Filloy & Bel-locq (2013) that spatial variation in the abun-dance of forest birds is mainly due tostructural components of the forest.

In the Amazon rainforest, there are fewstudies that relate the habitat structure withthe distribution, abundance, and behavior ofowls (see Willis 1977, Enriquéz-Rocha &Rangel-Salazar 2001, Borges et al. 2004,Sberze et al. 2010). A recent study of noctur-nal birds in the Brazilian Amazon analyzedhabitat use in primary and secondary forest(Sberze et al. 2010), but among the speciesanalyzed only two were owls – Crested Owland Amazonian Pygmy Owl. However, theirlevels of occurrence did not differ signifi-cantly between the two forest types.

A more recent study in the Reserva Flo-restal Adolpho Ducke (RFAD) evaluated theeffects of forest-structure components onhabitat use by six owl species (Barros & Cin-

tra 2009), describing their general spatial dis-tribution from records in 72 plots spacedfrom each other for 1 km and distributed in alarge spatial scale of 64 km2. The studyshowed very clear patterns of spatial distribu-tion and influence of environmental hetero-geneity in different species of owls (Barros &Cintra 2009). The abundance of still-standingdead trunks was used to explain the variationin density of Crested Owl and NorthernTawny-bellied Screech Owl, with the lattershowing more frequent use of areas withhigher tree abundance. The AmazonianPygmy Owl preferred forest areas nearstreams. In this study, playback was not usedas a complimentary technique to listeningrecords at point counts (Granzinolli & Motta-Junior 2010), even though combining bothmethods can increase detections, thus mini-mizing false absence records (Mackenzie et al.2002). So, even when underestimating therecords by using this technique the studyshowed very clear patterns of spatial distribu-tion and influence of environmental hetero-geneity in different species of owls (Barros &Cintra 2009). As the species included in thatstudy have different body sizes, probably dif-fer in the size of areas used, and occur atdifferent densities, it seemed interesting toevaluate whether at smaller spatial scales thesepatterns are similar to the spatial patternfound by Barros & Cintra (2009). Hence theaim of the current study was to describe thespatial distribution of six owl species (Glauci-dium hardyi, Lophostrix cristata, Megascops wat-sonii, Pulsatrix perspicillata, Strix virgata, Strixhuhula) in the RFAD at a smaller spatial scale(25 km2) than applied by Barros & Cintra(2009) while using the same sample designand the same 1-km spacing between plots, butwith twice as many samples in plots and usingadditional detection techniques.

Specifically, this study examined howseven forest-structure components influ-enced habitat use (occurrence and abun-

437


dance) by owls in a central Amazonian uplandforest. The forest-structure componentswere: 1) leaf-litter depth, 2) density of livingtrees, 3) standing dead-trunk density, 4) fallendead-trunk density, 5) canopy opening, 6) ter-rain elevation, and 7) distance to neareststream. The presence and density of the spe-cies was estimated using playback, a methodbarely used for owls in Neotropical forestenvironments.

Our hypothesis is that the spatial variationin the components of forest structure influ-ences the spatial distribution and use ofmicrohabitat by owls (verified on a species-by-species basis). We predict that the abun-dance and frequency of species occurrencewill increase with the presence of flatter areasin the forest and with increases in vegetationdensity, leaf-litter depth, fallen-log abundance,the extent of canopy opening, and proximityto more humid areas, such as water bodies.

Considering some possibly importantaspects for our focal species, e.g., known feed-ing habits and local preference for shelter andnest building, our hypothesis predicts: 1)increase in the density of the AmazonianPygmy Owl (G. hardyi) with increase in can-opy opening; 2) increase in the density of theNorthern Tawny-bellied Screech Owl (M. wat-sonii) with the decrease in the canopy openingand increase in live-tree density, standingdead-trunk density, fallen dead-trunk density,and leaf-litter depth; 3) increase in the densityof the Mottled Owl (S. virgata) with decreasein the canopy opening and increase in live-tree density, standing dead-trunk density,fallen dead-trunk density, and leaf-litterdepth; and 4) increase in the density of theBlack-banded Owl (S. huhula) with decrease inthe canopy opening, together with an increasein standing dead-trunk density and fallendead-trunk density, respectively. For the twolargest species, our hypothesis predicts for theCrested Owl (L. cristata) an increase in densitywith an increase in canopy opening, live-tree

density, standing dead-trunk density, andfallen dead-trunk density, and for the Specta-cled Owl (P. perspicillata) a decrease in densitywith an increase in distance to nearest streamand in canopy opening, and an increase indensity with the increase in standing/fallendead-trunk density (Gwynne et al. 2010, Kö-nig & Weick 2008).

METHODS

Study area. The study was conducted in theReserva Florestal Adolfo Ducke (RFAD),located near Manaus, Amazonas State, Brazil(02°55’–03°01’S, 59°53’–59°59’W). Adminis-tered by the National Institute of AmazonianResearch (INPA), the RFAD covers some10,000 ha of primary terra firme forest, and isone of the best-studied areas of the BrazilianAmazon (Ribeiro et al. 1999, Oliveira et al.2008).

Within the RFAD, average annual temper-ature is 26ºC, and annual rainfall ranges from1750 to 2500 mm, with a rainy season(November to May) and a dry season (June toOctober) (Oliveira et al. 2008). The dominantvegetation is mature evergreen forest with acanopy between 30–40 m, with emergenttrees reaching up to 55 m (Ribeiro et al. 1999).The local topography consists of undulatingplateaus, with predominantly a closed forestcanopy and poorly-lit understory (Oliveira etal. 2008), incised by stream-bearing valley bot-toms. In the central sector, an elevated plateauarea divides the local stream system into twodistinct basins. The clear-water streams of theeastern region flow into the Amazon River,while the black-water creeks on the westernside flow into tributaries of the Rio Negro(Fig. 1).

Sample design. This study involved six commonspecies of Strigiformes in the study area:Amazonian Pygmy Owl, Northern Tawny-bellied Screech Owl, Mottled Owl, Black-

438

ESCLARSKI & CINTRA

banded Owl, Crested Owl, and SpectacledOwl. Sampling occurred on clear nights withlittle wind, between 18:00 and 23:00 h. Dur-ing the breeding season, the owls are moresedentary, which guarded against possibledouble-counting of the same individual inmore than one point. Additionally, observers’movements between sampling points weremade as quickly as possible to minimizeeffects of individual bird movements betweensampling locations.

The PPBio grid used had 30 plots with250 m in length and each is separated fromthe next by 1000 m, and to minimize within-plot variation, each follows the local topo-graphic curve. Each plot in this study wasconsidered as a sampling unit. Playback(broadcasting recordings of the spontaneousvocalizations of each species, so that animalsrespond to these calls: Motta-Junior et al.2004, Braga & Motta-Junior 2009) was usedto test species’ presence. The center of eachplot was used as the listening point, and

meant the center of the radius of detection(125 m around the observer). The radius ofdetection of the observer was previouslydetermined by experimental simulations eval-uating the detection capability of the observerin the field.

Initially, each sampling sequence startedwith a listening session of five minutes, whichwas directed to identify and estimate whetherthe target owl species was vocalizing sponta-neously. Following an initial 5-minutes inter-val, the listener began the sequence ofplaybacks for the five remaining species, eachvocalization being played for three minuteswith an interval of another three minutes untilthe playback changed to the next species, fol-lowing from the smallest-sized (G. hardyi) tothe largest owl (P. perspicillata), in order tominimize the potential effects of dominancerank (Motta-Junior et al. 2004). A full sam-pling sequence lasted 35 minutes at each lis-tening point, including the initial playback-free minutes and the playback times. Some 25

FIG. 1. Reserva Ducke (RFAD), Central Amazon, Brazil – trail grid and plots used during the study(source: http://ppbio.inpa. gov.br/repositorio/imagens).

439


minutes were spent commuting between lis-tening points, so that 3–4 plots were sampledper night. Each plot was visited twice.

The methods were combined in order tominimize the effects of false absences. Thefollowing information was recorded for eachindividual point counts: date and time of reg-istration, species and number of individualsvocalizing. To estimate the number of owls,we considered simultaneous vocalizations,assuming vocalizations in different directionsnot to be simultaneous, as a result of displace-ment of same individual. We also notedwhether there were spontaneous vocalizationsor response to playback, the length ofresponse times, and responses of species vsplayback. A digital recorder (Panasonic RR-XS4410) with microphone (Yoga EM-9600)and a portable speaker (MaxPrint 2W R.M.S.)were used to broadcast owl calls, alwaysplayed using the same setting for volume (20)in the four directions (north, south, east, west)from the center point. In case of difficulties inspecies’ identification, vocalizations wererecorded for later laboratory analysis.

Within each sampling plot from whichplayback occurred, the following forest-struc-ture variables were measured: 1) leaf-litterdepth, 2) fallen dead-log density, 3) standingdead-log density, 4) live-tree density (DBH >10 cm), 5) canopy opening, 6) terrain eleva-tion, and 7) distance to nearest stream. Allthese variables were shown by a previousstudy (Barros & Cintra 2009) in the same areato impact the owls’ micro-habitat use.

Leaf-litter depth, fallen dead-log, standingdead-trunk densities, and canopy openingwere measured during the same period offield work to avoid any seasonal variation(Luizão & Schubart 1987, Luizão 1989, Rodri-gues et al. 2000, Vital et al. 2004, Nascimento etal. 2006). Five records of leaf-litter depth weretaken every 5 m within each plot, using a rulergraduated in millimeters. The resulting 51 per-plot records were then averaged. Fallen logs

and standing dead trunks were counted bydirect observation for an area within 20 mfrom both sides of the plot’s center line, andalong the entire plot length. Canopy coverwas measured with a densiospherometer, viafour records (north, south, east, and west)every 10 m along the plot’s central line, and atotal of 26 measures averaged for each plot.For the other variables, we used the existinginformation available in the LTER database(http//peld.inpa.gov.br).

The correlation between the independentvariables (forest-structure components) wastested with a Pearson correlation matrix.

To test whether forest-structure compo-nents influence habitat use by owl species, weperformed an analysis using a multiple linearmodel for the density of each species, and amultiple logistic model analysis for occur-rence (using as dependent variables the cate-gorical variables presence = 1, absence = 0)with Systat 13.0. Multiple linear models wereassessed with a quantitative response variableand continuous Y for explanatory variables(forest-structure components). The generallogistic equation is p = 1/ 1+e-z.

A qualitative array was generated withpresence/absence, and analyses conductedseparately for each owl species.

Raw densities were converted to a densityindex by multiplying the maximum number ofindividuals recorded per species by the timespent on each plot (35 minutes/plot), so thatspecies’ density ranged from 0–0.11. To esti-mate the total density per species for theentire study area, we used the density given bythe number of plots where the speciesoccurred divided by the total number of plotsstudied, multiplied by the value obtained forthe density index per sample unit.

RESULTS

Across all plots, forest-structure variablesshowed the following patterns of variation:

440

ESCLARSKI & CINTRA

1) Fallen dead-trunk number ranged from9–40, and standing dead-trunk number from0–16; 2) leaf-litter depth varied from2.94–5.84 cm; 3) distance to nearest streamranged from 20–493.33 m; 4) canopy openingvaried from 1.29–3.94%; 5) density of treeswith DBH > 10 cm ranged from 293–393individuals; and 6) overall terrain elevationranged from 46–105 m a.s.l..

Of the six owl species that responded toplayback, Amazonian Pygmy Owl andCrested Owl were recorded in 28 of the 30plots (93.4 % of the sampled area), and werethe most abundant species. The NorthernTawny-bellied Screech Owl was recorded in18 plots (60 %) and the Spectacled Owl in13 plots (43.4 %). The Mottled Owl occurredin 7 plots (23.4 %) and the Black-banded Owlin only 3 plots (10 %). The latter two specieswere excluded from the statistical analysisbecause of low numbers of records. Threespecies were not considered by a previousstudy (Barros & Cintra 2009) for a similarreason (Strix virgata, S. huhula, Pulsatrix perspi-cillata). However, the playback method usedhere allowed the inclusion of the SpectacledOwl, which Barros & Cintra (2009) excludedbecause of low detectability.

All owl species were detected in at leastthree sampling plots (10%) in each field visit.In seven plots of the PPBio grid, the MottledOwl, a species not included in the study, res-ponded to playbacks of the other owl species.

Densities of the owl species in the PPBiogrid ranged from 0.002 to 0.102 (Table 1).However, the species distribution mapswithin RFAD were made for the six species(Fig. 2).

From a correlation test (Table 2), we sepa-rate the environmental variables in two mod-els, keeping the variables most strongly andsignificantly correlated separated in differentmodels. Thus, for each species, the modelswere constructed as following, Model 1: deadlogs on the ground, standing dead trunks, dis-tance from stream, and canopy opening, andModel 2: litter depth, abundance of trees withDBH greater than 10 cm, and terrain eleva-tion (Tables 3–6).

The results of multiple linear modelsshow that the 1) density of G. hardyi increaseswith increasing distance from stream andincrease in terrain elevation (Fig. 3); 2) densityof M. watsonii increases with decreasing depthof leaf litter (Fig. 4); 3) density of L. cristataincreases with the increase in the abundanceof standing dead trunks (Fig. 5); and 4) den-sity and occurrence of P. perspicillata increasewith the increase in the abundance of fallendead trunks (Figs 6, 7).

DISCUSSION

This is probably the first study that used twosimultaneous methods minimizing falseabsences (direct observation by countingpoints, and playback) to demonstrate how thespatial variation in the structural componentsof a terra firme upland forest influences theuse of micro-habitats by nocturnal predatorybirds. The key interest from the ecologicalperspective is that they are of various sizesand in the same family, thus representing apotentially competing array (Marshal 1939).The results of this study support existing evi-dence (e.g., Terborgh 1985, Barros & Cintra2009, Cintra & Naka 2012) that, by variationin forest-structure components, the hetero-

TABLE 1. Densities of owl species in the 25 km2

PPBio grid at Reserva Ducke, Central Amazon.

Species Density per plot

Density by grid

Glaucidium hardyiMegascops watsoniiStrix huhulaStrix virgataLophostrix cristataPulsatrix perspicillata

0–0,110–0,080–0,02 0–0,020–0,110–0,05

0,1020,0480,0020,0040,1020,021

441


geneity of tropical forests can influence thespatial distribution of bird species, indicatingthat such effects have broad impacts through-out bird communities, irrespective of theirdiet and time of activity.

Owls, like other predatory birds, select ter-ritories according to their potential for breed-ing and nesting areas (Motta-Junior et al.2004). Cavities for nesting and shelter arerarely constructed by owls, with existing cavi-

FIG. 2. Species distributions in the PPBio plots at Reserva Ducke, Central Amazon: a) Glaucidium hardyi,b) Megascops watsonii, c) Strix huhula, d) Strix virgata, e) Lophostrix cristata, f) Pulsatrix perspicillata.

442

ESCLARSKI & CINTRA

ties (natural rot holes; holes made by wood-peckers, armadillos, and others) exploitedopportunistically, sometimes being slightlyenlarged or otherwise modified (Glinski &Ohmart 1983).

The use of dead and/or broken trunks arecommon among owls, and several studieshave demonstrated their importance in habi-tat selection by owls (La Haye 1988, Hersheyet al. 1988, Carrey 1990, Thorstrom 2011,Barros & Cintra 2009, Slaght et al. 2013).However, the current study found no rela-tionship between the abundance of standingdead trunks and density of owls at the micro-habitat level, suggesting that owls may useother forest-structure components in thestudied habitat and probably depend on theavailability of food resources.

Food availability is one of the many fac-tors limiting habitat use. In owls, low foodavailability may interfere with reproduction,both extending inter-breeding intervals andreducing clutch size. Additionally, pair forma-tion may be tied to cycles of prey abundance(McInvaille & Keith 1974, Ellinson 1980,Smith et al. 1981, Village 1982, Rice et al. 1983,Smith & Gilbert 1984, Sparks et al. 1994,Motta-Junior et al. 2004). In the present study,we used two proxies for food availability -depth of litter and number of fallen deadtrunks - and found that these variables weresignificantly influencing the presence of three

owl species (Megascops watsonii, Lophostrix cris-tata, Pulsatrix perspicillata).

For the Crested Owl, the frequency ofmicro-habitat use varied between differentareas within the study site. Though studieswere conducted in the same location, thisresult differs from the study of Barros & Cin-tra (2009) conducted at a higher spatial scale(covering almost the entire area of RFAD). Inour study, the species were also widely distrib-uted in the grid and not influenced by thecentral plateau areas as in the aforementionedstudy.

It is commonly accepted that body size isdirectly related to the species’ home-rangesize (Schoener 1968, Holling 1992), which, inturn, reflects differences in how individualsfulfill their basic survival requirements andperceive differences in the distribution ofmicro-habitat patches (Ziv 2000, Haskell et al.2002). An earlier study in the Central Amazoncomparing occupancy in owls discriminateddifferent species assemblages in secondaryand primary terra firme and in seasonallyflooded forest, and proposed that habitat-structure differences might be influential(Borges et al. 2004). This suggests that theway in which the species involved perceivedifferences in habitat characteristics may notoccur at the level of simple spatial variation,but at a finer scale related to the structuralcomponents of forest (e.g., Wiens 1976,

TABLE 2. Pearson Correlation Matrix for the forest-structure components recorded in 30 PPBio plots atReserva Ducke, Central Amazon.

Logs Litter Snags Distance to stream

Canopyopening

Tree abundance

Terrain elevation

LogsLitterSnagsDistance to streamCanopy openingTree abundanceTerrain elevation

1,0000,1720,061-0,030-0,208-0,243-0,135

1,0000,029-0,2530,339-0,369-0,381

1,0000,0890,0920,1180,180

1,0000,0170,3590,757

1,0000,246-0,087

1,0000,410 1,000

443

FORE

ST-STRUCTU

RE E

FFECTS O

N O

WLS

Multiple linear models Multiple logistic models

Effect Coefficient Standarderror

Standardcoefficient

Tolerance t p-value Estimate Standarderror

Z p-value 95% Confidence Interval

Lower Upper

Model 1Constant LogsSnagsDistance to streamCanopy opening

Model 2Constant LitterTree abundanceTerrain elevation

2,254-0,032-0,0330,0030,049

0,854-0,3350,0030,018

1,0040,0230,0520,0010,258

2,5880,2380,0060,009

0,000-0,248-0,1100,4200,034

0,000-0,2520,1000,370

-0,9490,9770,9910,946

-0,8000,7790,771

2,2461,4020,6312,4240,191

0,3301,4040,5522,024

0,0340,1730,5340,0230,850

0,7440,1720,5860,053

-340,27918,217-27,408-7,672

112,312

395,751114,4621,566

-13,524

1,23E+0105,150E+0088,119E+0081,660E+0083,103E+009

9,124E+0113,159E+0102,384E+0091,160E+009

0,0000,0000,0000,0000,000

0,0000,0000,0000,000

1,0001,0001,0001,0001,000

1,0001,0001,0001,000

2,412E+0101,009E+0091,591E+0093,254E+0086,082E+009

1,788E+0126,192E+0104,673E+0092,274E+009

2,41E+0101,009E+0091,591E+0093,254E+0086,082E+009

1,788E+0126,192E+0104,673E+0092,274E+009

TABLE 3. Results of Multiple Linear Models and Multiple Logistic Models of variation of Glaucidium hardyi density in relation to forest-structure compo-nents (Model 1 and 2) recorded in 30 plots at Reserva Ducke, Central Amazon.

444

ESCLA

RSKI &

CINTRA



Standardcoefficient



Lower UpperModel 1Constant LogsSnagsDistance to streamCanopy opening


1,648-0,001-0,0540,001-0,235

5,328-0,482-0,0070,001

0,9540,0220,0490,0010,246

2,4880,2290,0060,008

0,000-0,012-0,2110,118-0,187

0,000-0,421-0,2540,029

-0,9490,9770,9910,946

-0,8000,7790,771

1,728-0,060-1,0970,618-0,956

2,141-2,103-1,2530,142

0,0960,9530,2830,5420,348

0,0420,0450,2210,888

-4,3080,0540,1320,0000,717

-12,6870,6670,029-0,010

2,5150,0560,1280,0030,611

7,3400,6110,0180,023

-1,7130,9761,032-0,1491,173

-1,7281,0921,622-0,430

0,0870,3290,3020,8820,241

0,0840,2750,1050,667

-9,237-0,055-0,119-0,005-0,481

-27,074-0,530-0,006-0,056

0,6200,1630,3830,0051,915

1,6991,8650,0640,036

TABLE 4. Results of Multiple Linear Models and Multiple Logistic Models of variation of Megascops watsonii density in relation to forest-structure compo-nents (Model 1 and 2) recorded in 30 plots at Reserva Ducke, Central Amazon.

445

FORE

ST-STRUCTU

RE E

FFECTS O

N O

WLS



Standardcoefficient





1,4370,054-0,0420,000-0,254

0,2760,2720,0000,007

1,1020,0250,0570,0010,284

3,3290,3070,0080,011

0,0000,389-0,1310,058-0,162

0,0000,191-0,0080,135

-0,9490,9770,9910,946

-0,8000,7790,771

1,3052,146-0,7310,327-0,894

0,0830,887-0,0350,617

0,2040,0420,4710,7460,380

0,9350,3830,9720,542

592,839-26,981-18,768-0,425-20,149

1.000,619-144,626

2,330-22,375

6,142E+0102,227E+0095,505E+00974,595,600

2,967E+010

8,193E+0101,157E+0101,495E+0081,091E+009

0,0000,0000,0000,0000,000

0,0000,0000,0000,000

1,0001,0001,0001,0001,000

1,0001,0001,0001,000

-1,204E+011-4,366E+009-1,079E+010-1,462E+008-5,816E+010

-1,606E+011-2,268E+010-2,930E+008-2,137E+009

1,204E+0114,366E+0091,079E+0101,462E+0085,816E+010

1,606E+0112,268E+0102,930E+0082,137E+009

TABLE 5. Results of Multiple Linear Models and Multiple Logistic Models of variation of Lophostrix cristata density in relation to forest-structure compo-nents (Model 1 and 2) recorded in 30 plots at Reserva Ducke, Central Amazon.

446

ESCLA

RSKI &

CINTRA



Standardcoefficient





-0,8830,0400,0490,0000,099

-0,7540,2190,001-0,002

0,6160,0140,0320,0010,159

1,9520,1800,0050,007

0,0000,4790,260-0,1110,107

0,0000,2570,066-0,068

-0,9490,9770,9910,946

-0,8000,7790,771

-1,4332,8221,554-0,6690,627

-0,3861,2160,306-0,314

0,1640,0090,1330,5090,536

0,7020,2350,7620,756

6,656-0,151-0,3480,001-0,433

8,668-0,920-0,0140,006

3,3850,0770,2040,0030,640

6,7110,6250,0160,022

1,967-1,963-1,7060,345-0,677

1,292-1,472-0,9000,271

0,0490,0500,0880,7300,499

0,1960,1410,3680,787

0,023-0,302-0,747-0,005-1,687

-4,485-2,145-0,045-0,037

13,2900,00010,0520,0070,821

21,8200,3050,0170,048

TABLE 6. Results of Multiple Linear Models and Multiple Logistic Models of variation of Pulsatrix perspicillata density in relation to forest-structure compo-nents (Model 1 and 2) recorded in 30 plots at Reserva Ducke, Central Amazon.

447


-20 -10 0 10 20

Fallen dead-trunk abundance-2

-1

0

1

2

Gla

ucid

ium

har

dyi

-10 -5 0 5 10

Standing dead-trunk abundance-2

-1

0

1

2

-3

00-2

00-1

00 010

0200 30

0400

Distance to nearest stream (m)

-2

-1

0

1

2

Gla

ucid

ium

har

dyi

-2 -1 0 1 2Canopy opening (%)

-2

-1

0

1

2

-100 -50 0 50

Tree abundance (DBH > 10 cm)-2

-1

0

1

2

FIG. 3. Results of multiple linear models of Glaucidium hardyi density in relation to forest-structure compo-nents (Model 1 and 2). Some numbers of the axes are negative because the partial relationships deviationsof the expected results as all the other variables are maintained constant with their observed means.

Model 1

Model 2

-40

-30

-20

-10 0 10 20 30 40

Terrain elevation (m)

-2

-1

0

1

2

Gla

ucid

ium

har

dyi

-1 0 1 2

Leaf-litter depth (cm)-2

-1

0

1

2

Gla

ucid

ium

har

dyi

-100 -50 0 50

Tree abundance (DBH > 10 cm)

-2

-1

0

1

2

448

ESCLARSKI & CINTRA

-20 -10 0 10 20


-1

0

1

2

3

Meg

asco

ps w

atso

nii

-10 -5 0 5 10Standing dead-trunk abundance

-1

0

1

2

3

-3

00-2

00-1

00 010

0200 30

0400


-2

-1

0

1

2

3

Meg

asco

ps w

atso

nii

Model 2


-2

-1

0

1

2

3

-1 0 1 2Leaf-litter depth (cm)

-1

0

1

2

3

Meg

asco

ps w

atso

nii

-100 -50 0 50Tree abundance (DBH > 10 cm)

-2

-1

0

1

2

Model 1

-40

-30

-20

-10 0 10 20 30 40

Terrain elevation

-2

-1

0

1

2

Meg

asco

ps w

atso

nii

FIG. 4. Results of multiple linear models of Megasciops watsonii density in relation to forest-structure compo-nents (Model 1 and 2).

449


Model 1

-300-20

0-10

0 0100

20030

0400


-2

-1

0

1

2

Loph

ostri

x cr

ista

ta


-2

-1

0

1

2

Model 2


-2

-1

0

1

2

3

Loph

ostri

x cr

ista

ta

-100 -50 0 50Tree abundance (DBH > 10 cm)

-2

-1

0

1

2

3

-40 -30 -20 -10 0 10 20 30 40


-2

-1

0

1

2

3

Loph

ostri

x cr

ista

ta

FIG. 5. Results of multiple linear models of Lophostrix cristata density in relation to forest-structure compo-nents (Model 1 and 2).

-20 -10 0 10 20


-1

0

1

2

3

Loph

ostri

x cr

ista

ta

-10 -5 0 5 10

Standing dead-trunk abundance-2

-1

0

1

2

450

ESCLARSKI & CINTRA

-20 -10 0 10 20Fallen dead-trunk abundance

-1

0

1

2

Pul

satri

x pe

rspi

cilla

ta

-10 -5 0 5 10Standing dead-trunk abundance

-1

0

1

2

Model 1


-1

0

1

2

Pul

satri

x pe

rspi

cilla

ta

-40 -30 -20 -10 0 10 20 30 40


-1

0

1

2

Pul

satri

x pe

rspi

cilla

ta

-100 -50 0 50Tree abundance (DBH > 10cm)

-1

0

1

2

Model 2

FIG. 6. Results of multiple linear models of Pulsathrix perspicillata density in relation to forest-structurecomponents (Model 1 and 2).


-1

0

1

2

-300

-200

-100 0

100

200

300

400


-1

0

1

2

Pul

satri

x pe

rspi

cilla

ta

451


Kotliar & Wiens 1990, Tews et al. 2004). Forexample, Lophostrix and Pulsatrix are largerthan Glaucidium and Megascops, which implieshigher energy demands in the former onesand hence more extensive home ranges (Kelt& Vuren 1999, Brown 2007). These size dif-ferences may explain why Lophostrix and Pulsa-trix exploit larger areas, which likely include agreater variety of micro-habitats than thoseembraced by the smaller home ranges ofsmaller-sized owl species.

Unlike the study by Barros & Cintra(2009), where Crested Owl density was foundto be related to the abundance of standingdead trunks, the current study found habitatuse by Spectacled Owl and Crested Owl cor-related with an increase in fallen-log abun-dance. As forest-floor logs attract potentialowl-prey items, such as invertebrates andsmall mammals (Kiltie, 1981), a general posi-tive relationship between forest logs and owlabundance is indicated even when habitat,prey type, and owl size are very different (seeSmith & Gilbert 1984). This suggests that thestudied owl species preferably use particularareas within the available habitat. In addition,the different components of forest structure

had different effects on use frequency by suchsmaller species as Amazonian Pygmy Owl andNorthern Tawny-bellied Screech Owl.

Obviously, use frequency by Crested Owland Spectacled Owl is increased in areas witha greater abundance of fallen logs. Forest-floor areas with increased log abundance arewidely considered as places important to for-age and find potential prey, such as rodents aswell as invertebrates within decaying trunks(del Hoyo 1999). Moreover, the understoryenvironment near such trunks tends to bemore humid, and this may be important in thedry season, which at RFAD coincides with thenesting season of most species of ground-breeding birds (see appendix Cintra & Naka2012). Owls at RFAD are likely to find fooditems more promptly when visiting areas withhigher than those with lower abundance offallen logs. However, areas with these charac-teristics can increase the risk of exposure ofindividuals to potential predators, like snakesthat use such locations as refuges and forbreeding, and hunting wild cats (ocelots, jag-uars) that use them for concealment. There-fore, owls visiting such profitable foragingsites must balance visit length and frequency

FIG. 7. Results of multiple linear models of Pulsatrix perspicillata density in relation to abundance of fallenlogs.

0 10 20 30 40 50

Fallen dead-trunk abundance0,0

0,2

0,4

0,6

0,8

1,0

1,2

Pul

satri

x pe

rspi

cilla

ta

452

ESCLARSKI & CINTRA

in order to reduce their vulnerability to preda-tors (e.g., Lima & Dill 1990, Masse et al.2013). We had expected that the Crested Owloccurrence would be positively related to ele-vation, since a previous study (Barros & Cin-tra 2009) had demonstrated a strong asso-ciation of this large owl with higher altitude,probably because in such areas the foresttends to be more open. However, the currentstudy found that the Crested Owl also useslow-lying and sloped areas, a novelty for thebiology of this species, previously believed touse only flat forested areas in the plateaus.

Arthropod abundance increases with litterdepth, which, in turn, influences the occur-rence of understory-living species of insectiv-orous birds (for the tropics: Pearson 1977,Pearson & Derr 1986; for RFAD: Cintra &Cancelli 2008, Cintra & Naka 2012). Con-firming results from a study conducted withinthe reserve but at a larger spatial scale (Barros& Cintra 2009), our results show that theNorthern Tawny-bellied Screech Owl prefersforest environments with shallow leaf-litterlayers. Though specialists exist, many owl spe-cies are opportunistic and have a generalistdiet (Schubart et al. 1965, Sick 1997). TheNorthern Tawny-bellied Screech Owl eatssmall rodents, birds, and invertebrates, so thefact that it most frequently uses areas withshallow leaf-litter may be due to the ease withwhich such prey is both recognized and cap-tured in these places. Such foraging may beenergetically effective since the use of siteswith less litter increases predation-successrates, hence economizing the predator’s timeand energy budgets (Amaral 2007).

In our study, the Amazonian Pygmy Owlhad a higher density in areas closer to creeks.This was also found by Barros & Cintra(2009), who believe this may be explained bythe more open canopy and understory closeto streams, which increase visibility as well asthe frequency of small animals (lizards, frogs,small mammals) in such areas. In a study of

Eastern Screech Owls (Megascops asio), Smith& Gilberd (1984) found that they used differ-ent habitats during the breeding and non-breeding season, preferring more open areas,which increased successful predation.

The Amazonian Pygmy Owl respondedpromptly and aggressively to playback vocal-izations, even to those of three-times largerspecies (i.e., Spectacled Owl). This suggeststhat G. hardyi, being the smallest representa-tive of owls at RFAD, is highly territorial andinvests more energy to defend its resourcesthan other (i.e., larger) species. Alternatively,this aggressive behavior could be interpretedas reaction towards the presence of similar-sized or larger, physically more dominantowls that may represent both competitorsand/or potential predators. Generally, in thecontext of signaling theory aggressive callingcan indicate towards an intruder defensecapability and readiness for the next level ofaggression, respectively, which could help toavoid further, e.g., more costly interactions(Georgiev et al. 2013; cf. Gill 2007), includingthe risk of being predated (Jakobsson et al.1995).

Crested Owl and Amazonian Pygmy Owlwere the commonest owl species at RFAD,being absent from only two sampled plots butabundant in the remaining 28 plots. Giventheir spatial overlaps, these species probablyare likely not competitors and thus can coex-ist in the same habitat. The current studyfound the Crested Owl to be uniformly dis-tributed within the study area. This contra-dicts the results of Barros & Cintra (2009),who indicated that its abundance is clumpedon the RFAD central plateau.

The current study found that the North-ern Tawny-bellied Screech Owl was abundantand widely distributed across the study grid.This is also in contrast to the findings of Bar-ros & Cintra (2009), who found it to berestricted to one area, Ipiranga, in the south-western part of the RFAD (not included in

453


this study). Thus, the currently observed pat-tern might be linked to methodological differ-ences or to a recent expansion and coloni-zation of new areas by this owl.

According to maps generated by the study,the Spectacled Owl occurs in areas close tostreams below 100 m a.s.l.. This may berelated to its diet that includes small mam-mals, birds, large spiders, insects, and freshwa-ter crabs (Sick 1997, Sigrist 2009). Similarresults have been recorded for other owl spe-cies; e.g., a study of the diet of the Ural Owl(Strix uralensis), Korpimaki & Sulkava (1987)found a positive correlation between the pro-portion of water bodies and an increased pre-dation on rodents, possibly because preyanimals are more vulnerable to the predationat such localities. The frequency, with whichthe Spectacled Owl was recorded in the sam-pled area, corroborates the view of Stotz et al.(1997) that this species is common in areas ofcentral Amazonian terra firme, a fact that isstill is no consensus among researchers. Ourresults also confirm hitherto untested occa-sional observations that the species prefersareas closer to water bodies (Sick 1997).

The Black-banded Owl was recorded inonly three plots, all in a single hydrologicalmicro-basin, while the Mottled Owl was pres-ent only in plots with minimal canopy open-ing.

Although the current study covered onlyone breeding season, our results demonstratethat six species of owls at RFAD, being dis-similar in body size as well as in their habitatrequirements, food, and reproductive behav-iors, use different micro-habitats. Accordingto the relationships between the componentsof forest structure and its importance to thespecies, the components related to the avail-ability of food resources were significantlyassociated with the occurrence and density ofspecies of owls. This was possibly founded inthe fact that they already had chicks in thenest, which also contributes to the increase in

the density of species in the sampling points,as well as parents seeking food for themselveshave to seek to feed their offspring. Compar-ing our results with those from Barros & Cin-tra (2009) it is indicated that the differences instudy results mentioned before might be dueto the use of different sampling methods, butthere may be also substantial inter-annual vari-ation in habitat preferences. Consequently, wewould recommend a multi-seasonal monitor-ing program that includes habitat use andmovement of individuals in order to deter-mine whether the presence of individuals ofvarious species in a given area is related to theinfluence of vegetation structure or might bea result of population dispersal events, withexpansion of the species and occupation ofterritories across consecutive breeding sea-sons (Ritchison et al. 1988).

ACKNOWLEDGMENTS

PE thanks the following: Instituto Nacionalde Pesquisas da Amazônia (INPA) for logistichelp during this study; Programa de Pesquisaem Biodiversidade, Programa de PesquisasEcologicas de Longa Duração, and the Con-selho Nacional de Desenvolvimento Científi-co e Tecnológico (PPBio/PELD/CNPq) forthe study grant and the project financing. Fur-ther thanks go to Rafael S. Guerta, for help atall stages of the research; to my friendsGlauco Kohler, Matheus Montefusco, andRaiclicia Morais for their company and help inthe field, and especially the field guides Sr.Ayres and Srta. Ivanery, without whom itwould have been impossible to collect thedata required for this study. Adrian Barnetthelped with the English.

REFERENCES

Allen, A. P., & J. F. Gillooly. 2006. Assessing latitu-dinal gradients in speciation rates and biodiver-sity at the global scale. Ecol. Lett. 9: 947–954.

454

ESCLARSKI & CINTRA

Amaral, K. F. 2007. Composição e abundânciade corujas em Floresta Atlântica e sua relaçãocom variáveis de habitat. M.Sc. thesis, Univ.Federale do Rio Grande do Sul, Porto Alegre,Brazil.

Barros, O. G., & R. Cintra. 2009. The effects offorest structure on occurrence and abundanceof three owl species (Aves: Strigidae) in thecentral Amazon forest. Zoology 26: 85–96.

Begon, M., J. L. Harper, & C. R. Townsend. 1986.Ecology: individuals, populations and commu-nities. Blackwell Scientific Publications, Ox-ford, UK.

Braga, A. C. R., & J. C. Motta-Junior. 2009.Weather conditions and moon phase influenceon Tropical Screech Owl and Burrowing Owldetection by playback. Ardea 97: 395–401.

Brown, W. P. 2007. Body mass, habitat generality,and avian community composition in forestremnants. J. Biogeogr. 34: 2168–2181.

Bull, E. L., A. L. Wright, & M. G. Henjum. 1989.Nesting and diet of Long-eared Owls in coniferforest, Oregon. Condor 91: 908–912.

Belthoff, J. R., & G. Ritchinson. 1990. Nest-siteselection by Eastern Screech-Owls in centralKentucky. Condor 92: 982–990.

Best, A. S., K. Johst, T. Münkemüller, & J. M. J.Travis. 2007. Which species will successfullytrack climate change? The influence ofintraspecific competition and density depen-dent dispersal on range shifting dynamics.Oikos 116: 1531–1539.

Borges, S. H., L. M. Henriques, & A. Carvalhaes.2004. Density and habitat use by owls in twoAmazonian forest types. J. Field Ornithol. 75:176–182.

Cintra, R., A. E. Maruoka, & L. N. Naka. 2006.Abundance of two Dendrocincla woodcreepers(Aves: Dendrocolaptidae) in relation to foreststructure in Central Amazonia. Acta Amazo-nica 36: 209–220.

Cintra, R., & L. N. Naka. 2012. Spatial variation inbird community composition in relation totopographic gradient and forest heterogeneityin a Central Amazonian rainforest. Int. J. Ecol.1–25.

Chase, J. M., & M. A. Leibold. 2003. Ecologicalniches. Univ. of Chicago Press, Chicago, Illi-nois, USA.

Day, T. 2000. Competition and the effect of spatialresource heterogeneity on evolutionary diver-sification. Am. Nat. 155: 790–803.

del Hoyo J., A. Elliott, & J. Sargatal. 1999. Hand-book of the birds of the world. Volume 5:Barn-owls to hummingbirds. Lynx Edicions,Barcelona, Spain.

Ellinson, P. T. 1980. Habitat use by residentScreech-Owls (Otus asio). M. Sc. thesis, Univ. ofMassachusetts, Amherst, Massachusetts, USA.

Enfjäll, K., & O. Leimar. 2009. The evolution ofdispersal – the importance of informationabout population density and habitat character-istics. Oikos 118: 291–299.

Enriquéz-Rocha, P. L., & J. L. Rangel-Salazar 2001.Owl occurrence and calling behavior in a tropi-cal rain forest. J. Raptor Res. 35: 107–114.

Filloy, J., & M. I. Bellocq. 2013. Spatial variation inbird species abundances: Environmental con-straints across southern Neotropical regions.Basic App. Ecol. 14: 263–270.

Fletcher, R. J. 2006. Emergent properties of con-specific attraction in fragmented landscapes.Am. Nat. 168: 207–219.

Fletcher, R. J. 2007. Species interactions and popu-lation density mediate the use of social cues forhabitat selection. J. Anim. Ecol. 76: 598–606.

Gaston, K. J. 2000. Global patterns in biodiversity.Nature 405: 220–227.

Gayne, J. L., & R. P. Balda. 1994. Habitat selectionby Mexican Spotted Owls in northern Arizona.Auk 111: 162–169.

Georgiev, A. V., A. C. Klimczuk, D. M. Traficonte,& D. Maestripieri. 2013. When violence pays: acost-benefit analysis of aggressive behavior inanimals and humans. Evol. Psychol. 11:678–699.

Gill, F. B. 2007. Ornithology, 3rd ed. W. H. Free-man and Company, New York, New York,USA.

Glinski, R. L., & R. D. Ohmart. 1983. Breedingecology of the Mississippi Kite. Condor 85:200–207.

Graham, C. H., C. Moritz, & S. E. Williams. 2006.Habitat history improves prediction of biodi-versity in a rainforest fauna. Proc. Natl. Acad.Sci. 103: 632–636.

Granzinolli, M. A., & J. C. Motta-Junior. 2010.Ciência aplicada, técnicas de pesquisa e levanta-

455


mento: Aves de rapina – levantamento, seleçãode habitat e dieta. Pp. 169–187 in Straube, F. C.,V. Q. Piacentini, I. A. Accordi, & J. F. Cândido-Junior (eds). Ornitologia e conservação - cien-cia aplicada, tecnicas de pesquisa e levanta-mento. Technical Books, Rio de Janeiro,Brazil.

Gwynne, J. A., R. S. Ridgely, G. Tudor & M. Argel,2010. Aves do Brasil, Volume 1: Pantanal &Cerrado. Ed. Horizonte, São Paulo, SP, Brasil.

Haskell, J. P., M. E. Ritchie, & H. Olff. 2002. Frac-tal geometry predicts varying body size scalingrelationships for mammals and bird homeranges. Nature 418: 527–530.

Hershey, K. T., E. C. Meslow, & F. L. Ramsey.1998. Characteristics of forest at Spotted Owlnest sites in the Pacific Northwest. J. Wildl.Manag. 62: 1398–1410.

Hildén, O. 1965. Habitat selection in birds. ActaZool. Fenn. 2: 53–75.

Holling, C. S. 1992. Cross scale morphology,geometry and dynamics of ecosystems. Ecol.Monogr. 62: 447–502.

Hughes, C. L., C. Dytham, & J. K. Hill. 2007. Mod-eling and analysing evolution of dispersal inpopulations at expanding boundaries. Ecol.Entomol. 32: 437–445.

Hunter, J. E., R. J. Guttiérrez, & A. B. Franklin.1995. Habitat configuration around SpottedOwl sites in northwestern California. Condor97: 684–693.

Jakobsson, S., O. Brick, & C. Kullberg. 1995. Esca-lated fighting behaviour incurs increased preda-tion risk. Anim. Behav. 49: 235–239.

Kavanagh, R. P., S. Debus, T. Tweedie, & R. Web-ster. 1995. Distribution of nocturnal forestbirds and mammals in north-eastern NewSouth Wales: relationships with environmentalvariables and management history. Wildl. Res.22: 359–377.

Kelt, D. A., & D. V. Vuren. 1999. Energetic con-straints and the relationship between body sizeand home range area in mammals. Ecology 80:337–340.

Kiltie, R. 1981. Distribution of palm fruit on a rainforest floor: why White-lipped Peccaries foragenear objects. Biotropica 13: 141–145.

König, C., & F. Weick, 2008. Owls of the world.2nd ed. Christopher Helm, London, UK.

Korpimaki, E., & S. Sulkava. 1987. Diet and breed-ing performance of Ural Owls (Strix uralensis)under fluctuating food conditions. Ornis Fenn.64: 57–66.

Kotliar, N. B., & J. A. Wiens. 1990. Multiple scalesof patchiness and patch structure: a hierarchicalframework for the study of heterogeneity.Oikos 59: 253–260.

Kuch, J., & S. Idelberger. 2005. Spatial and tempo-ral variability of bat foraging in western Euro-pean low mountain range forest. Mammalia 69:21–33.

La Haye, W. S., & R. J. Gutiérrez. 1999. Nest sitesand nestling habitat of the Northern SpottedOwl in the northwestern California. Condor101: 324–330.

Le Galliard, J. F., R. Ferrière, & U. Dieckmann.2005. Adaptive evolution of social traits: origin,trajectories, and correlations of altruism andmobility. Am. Nat. 165: 206–224.

Levin, L. A. 1984. Life-history and dispersal pat-terns in a dense infaunal polychaete assem-blage: community structure and response todisturbance. Ecology 65: 1185–1200.

Lima, S. L., & L. M. Dill. 1990. Behavioral deci-sions made under the risk of predation: areview and prospectus. Can. J. Zool. 68:619–640.

Luizão, F. J., & H. O. R. Schubart. 1987. Litter pro-duction and decomposition in a terra-firme for-est of central Amazonia. Experientia 43:254–264.

Luizão, F. J. 1989. Litter production and mineralelement input to the forest floor in a CentralAmazonian forest. GeoJournal 19: 407–417.

MacKenzie, D. Y., J. D. Nichols, G. B. Lachman,S.Droege, J. A. Royle, & C. A. Lahgtimm. 2002.Estimating site occupancy rates: when detec-tion probabilities are less than one. Ecology 83:2248–2255.

Manzur, K. M., S. D. Frith, & P. C. James. 1998.Barred Owl home-range and habitat selectionin the boreal of central Saskatchewan. Auk 115:746–754.

Marshal, J. T. Jr. 1939. Territorial behaviour of theFlammulated Screech Owl. Condor 41: 71–78.

Martin, T. E. 1998. Are microhabitat preferences ofcoexisting species under selection and adaptive?Ecology 70: 656–670.

456

ESCLARSKI & CINTRA

Masse, R. J., B. C. Tefft, J. A. Amador, & S. R.McWilliams. 2013. Why woodcocks commute:testing the foraging-benefit and predation-riskhypotheses. Behav. Ecol. 24: 1348–1355.

McCallum, D. A., & F. R. Gehlbach. 1988. Nest-site preferences of Flammulated Owls in west-ern New Mexico. Condor 90: 653–661.

McInvaille, W. B., & L. B. Keith. 1972. Predator-prey relations and breeding biology of the greathorned owl and the red-tailed hawk in centralAlberta. Can. Field Nat. 88: 1–20.

Motta-Junior, J. C., A. A. Bueno, & A. C. R. Braga.2004. Corujas brasileiras. Departamento deEcologia, Instituto de Biociências da Universi-dade de São Paulo, São Paulo, Brazil.

Motta-Junior, J. C., & A. C. R. Braga, 2012. Estadodel conocimento sobre la ecologia y biologia debúhos em Brasil. Ornitol Neotrop. 23: 233.

Nascimento, R. E. S. A., U. Mehlig, M. M. O.Abreu, & M. P. M. Menezes. 2006. Litterfallyield in a tract of terra firme forest adjacent toa mangrove stand on the Ajuruteua Peninsula,Bragança, Pará. Bol. Mus. Para. Emílio GoeldiCienc. Nat. 1: 71–76.

Newton, I., M. Marquiss, D. N. Weir, & D. Moss.1977. Spacing of Sparrow-hawk nesting territo-ries. J. Anim. Ecol. 46: 425–441.

Oliveira, A. N. de, I. L. do Amaral, M. B. P. Ramos,S. D. Nobre, L. B. Couto, & R. M. Sahdo. 2008.Composição e diversidade florístico-estruturalde um hectare de floresta densa de terra firmena Amazônia Central, Amazonas, Brasil. ActaAmazonica 38: 627–642.

Orians, G. H. 1969. The number of bird species insome tropical forests. Ecology 50: 783–801.

Pearson, D. L. 1977. Ecological relationship ofsmall antbirds in Amazonian bird communities.Auk 94: 283–292.

Pearson, D. L., & J. A. Derr. 1986. Seasonal pat-terns of lowland forest floor arthropod abun-dance in southern Peru. Biotropica 18:244–246.

Peery, M. Z., R. J. Gutiérrez, & M. E. Seamans.1999. Habitat composition and configurationaround Mexican Spotted Owl nest site androost sites in the Talurosa Mountains, NewMexico. J. Wildl. Manag. 63: 36–43.

Pukinski, Y. B. 1973. On the ecology of the EagleOwl (Ketupa blakistoni doerriesi) in the basin of

the river Bikin. Bull. Mosc. Soc. Nat. Biol. Ser.78: 40–47.

Ribeiro, J. E. L. S., M. J. G. Hopkins, A. Vicentini,C. A. Sothers, M. A. Costa, J. M. de Brito, M. A.D. Souza, L. H. P. Martins, L. G. Lohmann, P.A. C. L. Assunção, E. C. Pereira, C. F. da Silva,M. R. Mesquita, & L. C. Procópio. 1999. Florada Reserva Ducke - guia de identificação dasplantas vasculares de uma floresta de terra-firme na Amazônia central. INPA-DFID,Manaus, Brazil.

Rice, J. 1983. Habitat selection attributes of anavian community: a discriminant analysis inves-tigation. Ecol. Monogr. 53: 263-290.

Ricklefs, R. E. 2004. A comprehensive frameworkfor global patterns in biodiversity. Ecol. Lett. 7:1–15.

Ritchison, G., P. M. Cavanagh, J. R. Belthoff, & E.J. Sparks 1988. The singing behavior of EasternScreech-Owls: seasonal timing and response toplayback of conspecific song. Condor 90:648–652.

Robinson, S. K. 1994. Habitat selection and forag-ing ecology of raptors in Amazonian Peru.Biotropica 26: 443–458.

Rodewald, A. D., & R. H. Yahner. 2000. Bird com-munities associated with harvested hardwoodstands containing residual trees. J. Wildl.Manag. 64: 924–932

Rodrigues, W.A., H. Klinge, & E. J. Fittkae. 2000.Estrutura e funcionamento de um ecossistemaflorestal Amazônico de terra firme junto aReserva Florestal Walter Egles, municipio deRio Preto da Eva, Amazonas, Brasil. Acta Biol.Par. 29: 219–243.

Sberze, M., M. Cohn-Haft, & G. Ferraz. 2010. Oldgrowth and secondary forest site occupancy bynocturnal birds in a Neotropical landscape.Anim. Conserv. 13: 3–11.

Schilling, J. W., K. M. Dugger, & R. G. Anthony.2013. Survival and home-range size of North-ern Spotted Owls in southwestern Oregon. J.Raptor Res. 47: 1–14.

Schoener, T. W. 1968. Sizes of feeding territoriesamong birds. Ecology 49 : 123–141.

Schubart, O., A. C. Aguirre, & H. Sick. 1965. Con-tribuição para o conhecimento da alimentaçãodas aves brasileiras. Arq. Zool. São Paulo 12:95–249.

457


Sick, H. 1997. Ornitologia brasileira: uma intro-dução. Editora Nova Fronteira, Rio de Janeiro,Brazil.

Sigrist, T., 2009. Avifauna brasileira. 1st ed. AvisBrasilis, São Paulo, Brazil.

Slaght, J. C. 2011. Management and conservationimplications of Blakiston’s Fish Owl (Bubo bla-kistoni) resource selection in Primorye, Russia.Ph.D. diss., Univ. of Minnesota, St. Paul, Min-nesota, USA.

Slagth, J. C., S. G. Surmach, & R. J. Gutiérrez. 2013.Riparian old-growth forests provide criticalnesting and foraging habitat for Blakiston’s FishOwl Bubo blakistoni in Russia. Oryx 47: 1–8.

Smith, D. G., & R. Gilbert. 1984. Eastern Screech-Owl home range and use of suburban habitatsin southern Connecticut. J. Field Ornithol. 55:322–329.

Smith, D. G. 1987. Owl census techniques. Pp.304–307 in Nero, R. N., Clark, R. J., Knapton,R. J. & Hamre, R. H. (eds). Biology and conser-vation of northern forest owls. Forest ServiceGeneral Technical Report RM-142. USDA,Forest Service, North Central Forest Experi-ment Station, Winnipeg, Manitoba, Canada.

Spangenberg, Y. P. 1965. Birds of the Iman Riverbasin. Sbornik Trudov Zool. Mus. MGU 9:98–202.

Sparks, E. J., J. R. Belthoff, & G. Ritchison. 1994.Habitat use by Eastern Screech Owl in centralKentucky. J. Field Ornithol. 65: 83–95.

Srugley, R. B. & P. Chai. 1990. Flight morphologyof Neotropical butterflies: palatability and dis-tribution of mass to the thorax and abdomen.Oecologia 84: 491–499.

Sumarch, S. G. 1998. Present status of Blakiston’sFish Owl (Ketupa blakistoni Seebohm) in Ussuri-land and some recommendations for protec-tion of the species. Rep. Pro Natura Found. 7:109–123.

Terborgh, J. W., J. W. Fitzpatrick, & L. Emmons.1984. Annotated checklist of birds and mam-mal species of Cocha Cashu Biological Station,Manu National Park, Peru. Fieldiana Zool. 21:28.

Terborgh, J. 1985. Habitat selection in Amazonianbirds. Pp. 311–338 in Cody, M. L. (ed.). Habitatselection in birds. Academic Press Inc., Orlan-do, Florida, USA.

Tews, J., U. Brose, V. Grimm, K. Tielborger, M. C.Wichmann, M. Schwager, & F. Jeltsch. 2004.Animal species diversity driven by habitatheterogeneity/diversity: the importance of key-stone structures. J. Biogeogr. 31: 79–92.

Thiollay, J. M. 1996. Distributional patterns of rap-tors along altitudinal gradients in the northernAndes and effects of forest fragmentation. J.Trop. Ecol. 13: 535–560.

Thiollay, J. M. 2002. Avian diversity and distribu-tion in French Guiana: patterns across a largeforest landscape. J. Trop. Ecol. 18: 471–498.

Thiollay, J. M., & Z. Rahman. 2002. The raptorcommunity of Central Sulawesi: habitat selec-tion and conservation status. Biol. Conserv.107: 111–12.

Throstrom, R., J. D. Ramos, C. M. Morales. 2000.Breeding biology of Barred Forest Falcons(Micrastur rufficollis) in northeastern Guatemala.Auk 117: 781–786.

Valentin, J. L. 2012. Ecologia numérica - umaintrodução a analise multivariada de dadosecológicos. Editora Interciência, Rio de Janeiro,Brazil.

Van Daele, J., & H. A. Van Daele. 1982. Factorsaffecting the productivity of Ospreys nesting inwest-central Idaho. Condor 84: 292–299.

Village, A. 1982. The home range and density ofkestrels in relation to vole abundance. J. Anim.Ecol. 51: 413–428.

Vital, A. R. T., I. A. Guerrini, W. K. Franken, & R.C. B. Fonseca. 2004. Produção de serrapilheirae ciclagem de nutrients de uma floresta estacio-nal semidecidual em zona ripária. Rev. Árvore28: 793–800.

Whitacre, D. F., C. W. Turley, & E. C. Cleaveland.1990. Correlation of diurnal abundance withhabitat features in Tikal National Park. Prog-ress report III: Maya Project. The PeregrineFund, Boise, Idaho, USA.

Wiens, J. A. 1969. An approach to the study of eco-logical relationships among grassland birds.Ornithol. Monogr. 8: 1–93. American Orni-thologists’ Union, Washington, D.C., USA.

Wilkinson, L. 2010. Systat: the system for statistics.Systat. Inc., Evanston, Illinois, USA.

Willis, E. O. 1977. Lista preliminar das aves da par-te noroeste e áreas vizinhas da Reserva Ducke,Amazonas, Brasil. Rev. Bras. Biol. 37: 585–601.

458

ESCLARSKI & CINTRA

Yaber, M. C., & K. N. Rabenold. 2002. Effects ofsociality on short-distance, female-biased dis-persal in tropical wrens. J. Anim. Ecol. 71:1042–1055.

Ziv, Y. 2000. On the scaling of habitat specificitywith body size. Ecology 81: 2932–2938.

Zwank, P. J., K. W. Kroel, D. M. Levin, G. M.Southerland, & R. C. Romé. 1994. Habitatcharacteristics of Mexican Spotted Owls insouthern New Mexico. J. Field Ornithol. 65:324–334.

Documents

EFFECTS OF TERRA FIRME-FOREST STRUCTURE … 25 433-458.pdfcia da disponibilidade de recurso alimentar na percepção de habitat pelas corujas. Abstract. – Owls are a poorly-studied