BIOLOGICAL CONSERVATION

Biological Conservation 123 (2005) 507519 www.elsevier.com/locate/biocon

Eects of structural and functional connectivity and patch size on the abundance of seven Atlantic Forest bird species
Alexandre Uezu a, Jean Paul Metzger
a

a,*

, Jacques M.E. Vielliard

ncias, Universidade de Sa Departamento de Ecologia, Instituto de Biocie o Paulo, Rua do Mata o, 321, travessa 14, 05508-900, Sa o Paulo, SP, Brazil b Departamento de Zoologia, Instituto de Biologia, Universidade Estadual de Campinas, C.P. 6109, 13084-971, Campinas, SP, Brazil Received 10 July 2004

Abstract We studied the importance of fragment size and structural and functional connectivity on the occurrence and abundance of seven Atlantic Forest bird species in 13 patches (13275 ha) and three sites within a continuous forest (10,000 ha). We sampled birds with point counts and evaluated structural connectivity considering the presence of corridors and the degree of isolation. We dened functional connectivity by analyzing species movements using playbacks in forest corridors between fragments and in the surrounding matrix. Species diered in their responses to fragmentation. For the frugivorous species, Trogon surrucura, Carpornis cucullatus and Triclaria malachitacea, patch size was the main factor determining abundance. Two understory insectivorous species, Basileuterus leucoblepharus and Pyriglena leucoptera, were more aected by the degree of patch connectivity, the former by the presence of corridors and the latter by the distance between patches. The capacity of P. leucoptera to use corridors and open areas (i.e. functional connectivity) shaped its abundance pattern. Fragmentation had no eect on the abundance of Chiroxiphia caudata and had a positive eect on Batara cinerea. This study emphasizes the importance of considering species perceptions of landscape, especially functional connectivity, in understanding the eects of habitat fragmentation. 2005 Elsevier Ltd. All rights reserved.
Keywords: Habitat fragmentation; Birds; Corridors; Patch size; Functional connectivity

1. Introduction Connectivity and patch size are important parameters for the persistence of species in fragmented landscapes (Karr, 1982; Blake and Karr, 1987; Bierregaard and Stouer, 1997; Stratford and Stouer, 1999; Crooks et al., 2000). While connectivity is associated with migration rates, and thus with the (re)colonization probability and the rescue eect, patch size is mainly related to the probability of local extinction (Levins, 1970; Hanski and Gilpin, 1997).
* Corresponding author. Tel.: +55 11 30917564; fax: +55 11 38134151. E-mail addresses: aleuezu@usp.br (A. Uezu), jpm@ib.usp.br (J.P. Metzger), jacques@unicamp.br (J.M.E. Vielliard).

Connectivity can be dened as the capacity of the landscape to facilitate biological uxes (Taylor et al., 1993; Tischendorf and Fahrig, 2000). In structural terms, it can be evaluated by measuring landscape patterns, such as density and complexity of corridors (Beier and Noss, 1998), distance between patches, and inter camps, habitat matrix permeability (Metzger and De 1997; Gascon et al., 1999; Antongiovanni and Metzger, 2005). Functional connectivity is more complex. It depends not only on the landscape pattern, but also on the interactions between this pattern and the biological characteristics of the target species, such as their ability to move in areas of non-habitat (Greenberg, 1989; Sieving et al., 1996). For example, understory insectivorous birds, which are able to use more intensively deforested countryside, are less aected by fragmentation and have

0006-3207/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2005.01.001

508

A. Uezu et al. / Biological Conservation 123 (2005) 507519

lower extinction risk (Sekercioglu et al., 2002). Interest in how species use dierent landscape elements to disperse is increasing (Tischendorf and Fahrig, 2000; Gra lisle and Desrochers, 2002; Adriaensen ham, 2001; Be et al., 2003) as it is linked with the probability of an individual nding patches of habitat that are spread across the landscape and, therefore, is related with species capacity to persist in such environments. In the present study, we evaluated the abundance of seven Atlantic rainforest bird species in relation to patch size and connectivity, considering not only structural characteristics of the landscape but also movement behavior of species. The Brazilian Atlantic rainforest has one of the highest levels of biodiversity and rates of endemism in the world, and is among the worlds top ve threatened hotspots (Myers et al., 2000; Mittermeier et al., 1999). Despite strict environmental legislation, today less than 8% of the forest remains and deforestation persists (Dean, ntica/INPE, 2002). 1996; Fundac a o SOS Mata Atla

Due to habitat loss and the consequent fragmentation, populations are becoming isolated in small forest patches and many species, especially endemics (Ribon et al., 2003), are locally disappearing. For example, the Atlantic Forest has 188 endemic bird species (Pacheco and Bauer, 2000) and among them 102 are considered threatened (Pacheco and Bauer, 2000; IUCN, 2003). There is a growing literature concerning the eects of deforestation and fragmentation on bird communities in this biome (Willis, 1979; Aleixo and Vielliard, 1995; Christiensen and Pitter, 1997; Anjos and Boc on, 1999; Brooks et al., 1999a; Maldonado-Coelho and Marini, 2000; Ribon et al., 2003), but few authors have considered aspects of connectivity. Since the response to fragmentation depends on the interaction between spatial characteristics of the landscape and the species behavior (Tischendorf and Fahrig, 2000), we analyzed connectivity in structural and functional terms. Structurally, we considered the presence or absence of connections (corridors) between small and larger patches in the land-

Fig. 1. Location of the study region in Sa o Paulo State, Brazil. The more detailed map shows the continuous forest (Morro Grande Reserve) and the adjacent fragmented landscape, with the 16 study sites (patch codes and spatial characteristics in Table 1).

A. Uezu et al. / Biological Conservation 123 (2005) 507519

509

scape as well as the distance between patches. In functional terms, we measured the species capacity to use corridors and to cross open areas of the matrix. We considered only strict forest species because, by denition, they are more sensitive to forest fragmentation. An understanding of the mechanisms that maintain these species in fragmented landscapes will likely have applications to other less demanding species (Lambeck, 1997; Simberlo, 1998).

2. Methods 2.1. Study region na, The study region is situated in the Plateau of Ibiu a Pre-Cambrian formation situated 40 km from the city of Sa o Paulo (2335 0 S2350 0 S; 4645 0 W4715 0 W; Fig. 1). The elevation ranges from 870 to 1030 m, and the topography is dominated by small round hills (Ponc ano et al., 1981). The weather is warm and humid, with mean monthly temperature varying between 11 and 27 C. The annual precipitation is about 13001400 mm with seasonal variation, April to August being cooler and

drier. The vegetation is classied as Lower Montane Rainforest (Oliveira-Filho and Fontes, 2000). Floristic surveys in the region showed a high tree richness (362 species in the region with diameter at breast height >5 cm), with a dominance of Myrtaceae (79 species), Lauraceae (38) and Fabaceae (31) (Bernacci et al., 2004). The region has well-preserved forest within the Morro Grande Reserve (10,000 ha), with an adjacent rural landscape where the forest occurs in patches and corridors (Fig. 1). Both the Morro Grande Forest Reserve and the rural landscape have similar geologic, geomorphologic, climatic and vegetational conditions, which make them suitable for comparison. In the fragmented area, the matrix is composed mainly of open areas (such as horticultural crops, fallow elds, pastures and pioneer vegetation; 60% of the matrix), rural installations and houses (17%) and Eucalyptus and Pinus plantations (11%). Widespread deforestation began in the 19th century, and was particularly intense during World War II (Seabra, 1971). The entire study area, including the Morro Grande Reserve, is composed of intermediate to old second-growth forest reestablished after logging and burning that occurred 5080 years ago (Seabra,

na Plateau, Southeastern Brazil). Points were Fig. 2. Location of playback points used to test the functional connectivity of small fragments (Ibiu chosen in corridors linking small connected patches (c) to large patches (L) to test the use of those corridors. Points were also placed around small isolated patches (i) to test possible functional connection with adjacent large patches. Light gray corresponds to young secondary forests (56 m canopy height) and dark gray to late and intermediate secondary forests (1520 m canopy height). Dashed lines represent the most likely routes for bird movement between small isolated and large patches.

510

A. Uezu et al. / Biological Conservation 123 (2005) 507519

1971). At present, approximately 31% of the fragmented landscape is composed of intermediate/old second growth forest and 6% of young secondary forest (ca. 10 years-old forest, 56 m high) (Fig. 2). Most patches are relatively small (212 ha) because of the small average lot size (<30 ha) in the area (Metzger et al., 2002). 2.2. Study sites We sampled 16 study sites, 13 forest patches and 3 sites in the reserve. All sample sites were composed of intermediate second growth forest (ca. 5060 years of regeneration) and patches were primarily surrounded by open areas. To ensure a large range of patch size, we sampled the ve largest patches in the fragmented landscape (50275 ha) and eight smaller ones (13 48 ha; Table 1). Among the smaller patches, forest corridors connected four of them to the largest patches sampled while four were structurally isolated. Corridors were <100 m wide and mainly composed of intermediate successional forest stages, but in one case the corridor was partially composed of young forest (connection between L3 and S6; Fig. 2). We established a paired design, where isolated and connected patches did not dier signicantly in size (t = 0.99; p = 0.36) although, on average, connected patches were larger (24.8 ha) than isolated ones (16.8 ha). The distance between small and large patches also did not dier signicantly (t = 1.83; p = 0.12), however, on average, connected patches were closer to large patches (225.5 m) than isolated ones (1185.0 m; Fig. 2). The large patches were not connected with each other. In the Reserve, we randomly selected three areas at least 2.2 km apart, resulting in an average distance of 3.1 km, equivalent to the distance between fragments (3.6 km; t test, F = 0.34, p = 0.65; Fig. 1).

2.3. Selected species We selected seven species based on their sensitivity to habitat loss and fragmentation and our capacity to detect them throughout the year (Willis, 1979; Aleixo and Vielliard, 1995; Magalha es, 1999). These species were: Carpornis cucullatus, Triclaria malachitacea, Trogon surrucura, Batara cinerea, Basileuterus leucoblepharus, Pyriglena leucoptera and Chiroxiphia caudata (Table 2). Except for T. malachitacea, all are inner forest species, territorial, have low dispersion capacity, and vocalize during the entire year (Vielliard personal experience; Willis, 1979; Silva, 1991; Aleixo and Vielliard, 1995). Triclaria malachitacea was chosen because it is a rare species and it is threatened by habitat fragmentation (IUCN, 2003). These species dier in feeding behavior, strata occupied, sensitivity to habitat alteration and degree of endemism (Ridgely and Tudor, 1989; Stotz et al., 1996; Sick, 1997; Pacheco and Bauer, 2000; Table 2). 2.4. Bird surveys 2.4.1. Point counts To estimate species abundance, we used unlimitedradius point counts (Blondel et al., 1970; Vielliard and Silva, 1990). Surveys were undertaken by AU every other month over 14 months, from September 2000 to September 2001, in a total of seven visits to each study site. The points were regularly distributed along a grid of 200 400 m (8 ha). Points were at least 200 m apart from each other and 50 m away from the forest edges. All sites had six points except for the two smallest patches, where it was possible to survey only four points due to the proximity to forest edges. In each visit, ve points were randomly selected and surveyed. For the

Table 1 na Plateau, Southeastern Brazil) Size and structural connectivity parameters for the studied patches of Atlantic Forest (Ibiu Patch code S1 S2 S3 S4 S5 S6 S7 S8 L1 L2 L3 L4 L5 Size classa Small Small Small Small Small Small Small Small Large Large Large Large Large Structural connectivity classb Isolated Isolated Isolated Isolated Connected Connected Connected Connected AREA (ha) 14.00 14.08 19.58 28.88 12.92 18.33 31.22 47.87 52.17 53.07 99.39 175.10 274.33 MNN (m) 89.59 32.04 71.67 59.29 31.36 38.10 61.50 48.38 81.28 44.80 48.27 47.24 53.54

AREA: patch size; MNN: mean nearest neighbor. a Large: >50 ha; small: 1050 ha. b Connected: small fragment structurally connected to a large fragment through forest corridors; isolated: small fragment without any structural connection to large fragments.

A. Uezu et al. / Biological Conservation 123 (2005) 507519 Endemic from Atlantic Forest

511

Sensitivity to human disturbance

patches with four points, a period that would correspond to a fth point was randomly chosen. During this time we did not make any observations. The survey included 546 point counts considered as independent samples: 28 for the two smallest patches and 35 for each of the other fourteen sites. The sampling period began a few minutes before dawn and ended approximately 4 h later. Observations were undertaken for 20 min at each point. During each visit, the sequence of the survey was randomly chosen to avoid non-independence of time and sampling points. For each species, a punctual abundance index (PAI) was calculated for each site where the number of contacts, visual and/or auditory, was divided by the number of sampled point counts. 2.4.2. Playback technique We surveyed the areas surrounding the small patches to test the eectiveness of corridors and degree of patch isolation on each species studied (Fig. 2). Playback technique was used to increase individual detection by reproducing the pre-recorded species-specic vocalizations to induce a response (Vielliard, 1989; Parker, 1991). All species studied here are territorial, so they typically respond very well to playback. The eectiveness of the playback technique has previously been veried for the species considered here (Boscolo, 2002). This technique eciently detected the presence of a species when its abundance was not too low (PAI > 0.05), and after three visits we have a condence of 95% to conrm species presence. We played the bird songs in pre-determined points in the corridors that links the small connected patches to the larger forests and around the small isolated fragments as well (Fig. 2). For the latter, we assumed the most likely routes used by birds to reach nearby large patches to be the shortest distance through open areas (see routes in Fig. 2) and placed the playback points along these lines. The number of points depended on the length of the corridor and on the context (variety of routes) where isolated fragments were situated. The test was done during October 2002, from 6:00 to 8:00 and from 11:00 to 13:00. This was considered as the best period of the year and time of day to detect the selected species using this method (Boscolo, 2002). To reach 95% condence in our results, each point was surveyed at least three times when the species were not detected. If all species were detected in a point before three visits, then the survey was stopped. At each point, the vocalization of each species was played continuously for one minute and followed by a 30-s listening interval. This procedure was repeated three times for each species. The sequence of species at each point was determined randomly. We considered the presence of a species in a corridor to indicate that such a connection facilitates dispersal

No Yes No Yes 20 30 140 25 Insectivores Insectivores Insectivores Frugivores Emberizidae Formicariidae Formicariidae Pipridae Low Medium High Low

Weight (g)

Feeding behavior

Table 2 Relevant biological characteristics of the seven bird species studied

Trogonidae Cotingidae Psittacidae Source: Willis (1979), Stotz et al. (1996) and Sick (1997). Trogon surrucura Carpornis cucullatus Triclaria malachitacea Surucua Trogon Hooded Berry-eater Blue-bellied Parrot

Family

Basileuterus leucoblepharus Pyriglena leucoptera Batara cinerea Chiroxiphia caudata

Species

White-browed Warbler White-shouldered Fire-eye Giant Antshrike Blue Manakin

Frugivores Frugivores Frugivores

70 60 90

Understory Understory Midstory Midstory/ Undestory Canopy Canopy Canopy

Strata

Medium High Medium

Yes Yes Yes

512

A. Uezu et al. / Biological Conservation 123 (2005) 507519

and thus gene ow. This assumption can be partly wrong if the territory of an individual in the corridor and can obstruct the movement of conspecics (Soule Gilpin, 1991). However, even if there is no gene ow at a specic time, we assume that the existence of a connection between territories should facilitate, through successive generations, gene ow between patches, and for this reason the patches should be considered functionally connected. Similarly, we considered that the lack of response (after three playback sessions) at a chosen point on the line between an isolated and a large patch indicates that this particular species does not use this route to move between those patches. Playback stimulation around isolated patches was mainly focused in open matrix, which represents most of the landscape around study patches. We did not control for the dierences between open matrix types (e.g. agricultural or pasture areas), because we assumed that they have the same eect on bird movement, since they are predominantly composed of herbaceous species. Except for urban areas, the open matrix should be the most signicant impediment to forest bird movement. Movement through urban areas was not considered because those areas represent less than 7% of the matrix and rarely occur surrounding study patches. However, we considered bird movement through young secondary forests because these can potentially connect isolated patches due to a higher forest structural similarity with intermediate/late secondary forests. Although the non-detection of a species in the matrix does not mean that it does not move through it, this information gives a rough idea of the matrix permeability for these species. When an individual was observed to cross open areas from a structurally isolated patch to a large patch, the former was considered functionally connected for this species. We thus used playback technique data to reclassify the small patches. If no birds were observed using a corridor that connects a small patch to a large one, the former was considered functionally isolated for the respective species. Conversely, if individuals of a species were conrmed to leave a structurally isolated forest and reach a large one, the patch was considered functionally connected for this species. 2.5. Quantitative analysis of landscape We selected two independent measures for the quantitative landscape analysis: patch size (AREA), excluding areas with less than 100 m in length (corridors); and the mean nearest neighbor (MNN). The latter index is the mean distance of patches to their respective nearest neighbor. We considered all patches totally or partially inside a radius of 1 km from the center of the studied patch. MNN was measured based on the shortest Euclidean edge-to-edge distance between fragments.

Using a 1 km radius that includes the areas that are likely to strongly inuence the presence of the study species, since most have territories of about 510 ha (Develey, 1997) and low dispersal capacity (Results). Larger radii would lead to a large overlap (>50%) among the radii of dierent patches. The metrics were calculated in ArcView 3.2 using the Patch Analyst 2.0 extension (Elkie et al., 1999). 2.6. Data analysis Categorical (size class and connectivity) and continuous data (landscape indices) were used to relate bird abundance and landscape parameters. We used analysis of variance (ANOVA) and the t test (Zar, 1996) to test for dierences between abundances considering patch size classes and the presence of structural or functional connections. We undertook four tests for each species: Morro Grande (forest landscape, n = 3) vs. patches (fragmented landscape, n = 13); large patches (n = 5) vs. small patches (n = 8); structurally connected patches (n = 4) vs. structurally isolated patches (n = 4); functionally connected patches vs. functionally isolated patches. In instances where species abundance did not satisfy ANOVA or t test assumptions, we used the respective equivalents non-parametric analyses, KruskalWallis and MannWhitney. For signicant results (i.e. p < 0.05), we applied an a posteriori test to identify which variables accounted for signicant dierences. s test for parametric analyses We applied the Schee and the Dunns test (Zar, 1996) for non-parametric tests. For the continuous data, we used regressions with bird abundances and landscape indices. The AREA index was log (base 10) transformed to present normal distribution characteristics. We used Statistica 5.01 for all analyses.

3. Results 3.1. Landscape structure Mean nearest neighbor (MNN) values ranged from 30 to 90 m, with an average of 54 m (Table 1). Fragment size (log transformed) and isolation were uncorrelated (Pearson, r = 0.027, p = 0.930), which allowed the evaluation of the eect of these indices with bird abundances independently. 3.2. Sensitivity to landscape structure The species showed varying patterns of abundance, indicating dierent degrees of sensitivity to fragmentation (Table 3). The comparison between forest and fragmented landscapes revealed that T. malachitacea and C. cucullatus were the most sensitive to alterations

A. Uezu et al. / Biological Conservation 123 (2005) 507519

513

in forest cover, since they were present only in the forest landscape (Table 3). Even inside the Reserve, they were restricted to a limited number of sites. Trogon surrucura occurred in both landscapes, but was more abundant in the Reserve than in the forest patches (Table 4). Conversely, B. cinerea was more abundant in the fragmented landscape than in the Reserve (Table 4). Within forest patches, T. surrucura was particularly sensitive to habitat loss (Table 4). It showed a positive and signicant correlation to patch size (Table 5). Basileuterus leucoblepharus and P. leucoptera were both sensitive to structural parameters of landscape connectivity. However, the former was signicantly more abundant in small connected patches than in small isolated forests (Table 4), while P. leucoptera was less abundant as the mean distance between patches increased (Table 5). Chiroxiphia caudata seems to be unaected by any of the landscape parameters considered in this study (Table 4).

3.3. Structural vs. functional connectivity Using the playback protocol we observed 157 individuals. Of these, 95 were using corridors (44 B. leucoblepharus, 22 C. caudata, 15 B. cinerea, 14 P. leucoptera). We also noted 24 individuals crossing open areas (eight P. leucoptera, six C. caudata, four B. leucoblepharus, four B. cinerea and one T. surrucura). These open areas varied from a small road (10 m length) to gaps of up to 130 m. We veried that individuals crossed through open areas towards other forest patches. If we played recorded songs in a direction where there was no forest, the individual could eventually travel some meters from the forest, but then went back to its original place. Observations in corridors and around small isolated patches yielded qualitative data about the use of these landscape elements, providing indicators of functional connectivity for each species (Table 6). With the exception of T. surrucura, all species observed in the fragmented

Table 3 na Plateau, Southeastern Brazil Punctual abundance index (PAI) based on point count of bird species in 16 sampled sites from Atlantic Forest, Ibiu Study sites S1 S2 S3 S4 S5 S6 S7 S8 L1 L2 L3 L4 L5 Ct1 Ct2 Ct3 Size S S S S S S S S L L L L L Ct Ct Ct Con. I I I I C C C C T. mala 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.03 0.17 C. cucu 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.63 T. surr. 0 0 0 0 0 0 0 0.03 0.11 0.11 0.23 0.34 0.23 0.23 0.51 0.17 B. leuc 0.25 0.96 0.89 0.34 1.00 1.43 1.69 1.40 0.71 1.43 0.89 1.69 0.31 2.09 1.34 1.37 P. leuc 0.04 0.32 0.20 0.17 0.43 0.29 0.54 0.23 0.14 0.34 0.26 0.37 0.26 0.34 0.31 0.26 C. caud 1.46 1.18 0.77 1.03 1.18 0.94 0.63 0.89 1.03 1.60 1.31 1.14 1.03 1.11 1.40 1.26 B. cine 0.38 0.50 0.63 0.13 0.50 0.37 0.43 0.47 0.73 0.70 0.53 0.30 0.30 0.23 0.27 0.23

T. mala, Triclaria malachitacea; C. cucu, Carpornis cucullatus; T. surr, Trogon surrucura; B. leuc, Basileuterus leucoblepharus; P. leuc, Pyriglena leucoptera; C. caud, Chiroxiphia caudata and B. cine, Batara cinerea. See species characteristics in Table 2. Sites are classied in size: S, small; L, large and Ct, control; and in connectivity (Con.): C, connected and I, isolated.

Table 4 Signicance values for analyses of variance when comparing the abundance obtained with point counts of ve species considering habitat loss and fragmentation (patches vs. reserve), fragment size (large vs. small), and structural and functional connectivity (isolated vs. connected) treatments na Plateau, Southeastern Brazil) (Ibiu Patches vs. reserve n df Trogon surrucuraa Basileuterus leucoblepharus Pyriglena leucoptera Chiroxiphia caudata Batara cinerea 16 1, 14 0.041 () 0.077 0.717 0.325 0.051 Large vs. small 13 1, 11 <0.001 (+) 0.971 0.975 0.172 0.390 Connected vs. isolated structural 8 1, 6 0.439 0.016 (+) 0.085 0.314 0.786 Connected vs. isolated functional 8 1, 6 0.021 (+) 0.016 (+) 0.792 0.528

Signs + and indicate that the rst treatment has a bigger or lower value of abundance, respectively. a Non-parametric tests.

514

A. Uezu et al. / Biological Conservation 123 (2005) 507519

Table 5 Simple linear regression (n = 13, df = 1, 11) of species abundance with patch size (AREA) and with isolation (MNN), considering only data from the na Plateau, Southeastern Brazil) fragmented landscape (Ibiu Log (AREA) R2 Trogon surrucura Basileuterus leucoblepharus Pyriglena leucoptera Chiroxiphia caudata Batara cinerea 0.817 0.003 0.005 0.003 0.020 p <0.001 0.859 0.816 0.869 0.642 MNN R2 0.008 0.218 0.397 0.007 0.010 p 0.763 0.107 0.021 0.791 0.741

Table 6 na Capacity to use corridors and young secondary forests, and to cross open areas of the ve species observed in the fragmented landscape (Ibiu Plateau, Southeastern Brazil) Use of corridors Trogon surrucura Basileuterus leucoblepharus Pyriglena leucoptera Chiroxiphia caudata Batara cinerea No Yes Yes Yes Yes Use of young secondary forests No Spontaneous Spontaneous Spontaneous With stimulation Capacity to cross open areas with playback stimulation (m) 10 100 60 130 60

area appear to use corridors as habitat (Table 6). This was particularly apparent in circumstances where the corridor was characterized by intermediate or old successional vegetation. These species were also able to use young secondary forest and cross open areas varying from 60 to 130 m (Table 6). Batara cinerea was normally associated with intermediate/old forest areas and was found in young secondary forest only when crossing from one patch to another as a result of the playback stimulation. In only one occasion was a T. surrucura individual observed to cross an open area, a small road approximately 10 m length. Omitting T. surrucura, the average maximum distance that individuals of all species traveled in open areas is greater than the mean distance between patches (54 m, Table 1). This suggests that most patches are eectively connected for these species. The surveys around isolated patches and in corridors suggest that structural connection is not always indicative of functional connectivity for some species (Table 7). Thus, although forests S1, S2, S3 and S4 are structurally isolated (Table 7), S2 (14.1 ha) is not functionally isolated for B. leucoblepharus, B. cinerea and C. caudata since we observed individuals of those species moving between this patch and other forest areas that were midway to patch L1 (52.1 ha; Fig. 2). We did not observe these species crossing from the other isolated patches (S1, S3 and S4) to large forest areas. Similarly, we did not observe P. leucoptera in the corridor between S8 and L3 (Table 7), although it was seen in corridors between other structurally connected areas (Table 7). The redenition of patch category concerning functional connectivity was especially important for P. leucoptera. The new t test, considering the S8 patch as isolated, resulted in a signicant dierence (t = 3.33;

p = 0.016; Table 4). Species abundance was higher in the connected patches, although this dierence was not detected when using the structural denition for connectivity (Table 4).

4. Discussion 4.1. Importance of area The species in this study showed dierent degrees of sensitivity to habitat area and connectivity, supporting previous ndings with birds and other taxonomic groups (Willis, 1979; Laurance, 1991; Newmark, 1991; n, 1994; Stouer and Bierregaard, Pearson, 1993; Andre 1995; Jokima ki and Huhta, 1996). Patch size was particularly important to large canopy frugivorous species. This is not surprising as they probably have a large area requirement due to high energetic demand and to the spatial and temporal variation of their food resources (Goerck, 1997). Although it is not a consensus among authors that higher body size increases vulnerability (Gaston and Blackburn, 1996; Henle et al., 2004), it seems that such a correlation does exist for frugivorous birds (Willis, 1979; Kattan, 1994; Christiensen and Pitter, 1997; Renjinfo, 1999; Castelletta et al., 2000). Thus the three most sensitive species in this study, T. malachitacea, C. cucullatus and T. surrucura, belong to this guild. Although we found no data in the literature concerning habitat requirements, these species may need large territories to secure sucient food during the entire year. For example, a study of other Trogon species in the Amazon Rainforest suggests they need territories, on average, of 8.5 ha (Terborgh et al., 1990).

A. Uezu et al. / Biological Conservation 123 (2005) 507519

515

Table 7 na Plateau, Classication of small patches according to structural and functional connectivity for the four species observed in these patches (Ibiu Southeastern Brazil) Study patches Structural connectivity Functional connectivity Basileuterus leucoblepharus S1 S2 S3 S4 S5 S6 S7 S8 I I I I C C C C I C I I I C C C Pyriglena leucoptera I I I I C C C I Chiroxiphia caudata I C I I C C C C Batara cinerea I C I I C C C C

C, connected to a large patch (>50 ha); I, isolated from large patches.

Rarity (low abundance within a habitat) also has a strong relationship with extinction susceptibility (Leck, 1979; Diamond et al., 1987; Laurance, 1990; Newmark, 1991; Goerck, 1997; Manne and Pimm, 2001). Triclaria malachitacea is naturally rare (Stotz et al., 1996) and is considered sensitive to even small habitat alterations (Magalha es, 1999). We found this species in only two areas inside the forest reserve, possibly due to the heterogeneity of forest structure. The degree of endemism, naturally low abundance and high sensitivity to habitat reduction are probably responsible for the status of this species as globally threatened (IUCN, 2003). The species may also suer from pet trade trac (as do other Psittacidae), particularly in the fragmented region where their habitat substantially interfaces with human activities. However, we do not have any evidence of this activity in the region. Alternatively, the low abundance of C. cucullatus might be related to the low forest quality in the region, which was substantially altered during the past 50 years. The species is common in other localities where habitat is well preserved (Ho ing and Lencioni, 1992). Trogon surrucura is also highly sensitive to habitat loss, although it is not as vulnerable as the previous two species. It was more often observed in larger patches, conrming results from previous studies (Willis, 1979; Christiensen and Pitter, 1997; Anjos and Boc on, 1999; Marini, 2001). Although, there is no consensus about the minimum patch size needed by the species, we obtained a similar result (12 ha) to that found by Anjos and Boc on (1999, 19.6 ha). Note however that some authors have found T. surrucura only in patches larger than 63 ha (Willis, 1979; Christiensen and Pitter, 1997; Marini, 2001). This may be due to diering conditions in the spatial structure of the landscapes. Thus the species may be able to survive in smaller fragments if they are part of a well-connected landscape. Our data also conrm previous observations that this species actively risks crossing open areas (Magalha es, 1999). For instance, we observed T. surrucura crossing open areas to feed in isolated fruit trees. Thus, although there is evi-

dence that large patches are essential for the maintenance of this species, our results suggest that smaller patches may also maintain this species. It seems to use several patches to nd resources, but visits larger patches more frequently, possibly because of the greater quantity of resources. Price et al. (1999) found a similar pattern for frugivorous birds in the tropical forests of Australia. Batara cinerea seems to benet from fragmentation at our study site since it was more abundant in patches than in the Reserve. These results run opposite to other observations that classify this species as sensitive to habitat alteration (Stotz et al., 1996; Aleixo and Vielliard, 1995; Anjos and Boc on, 1999). Although it is a relatively large and territorial species, qualities that theoretically should make it more demanding in habitat selection (Terborgh, 1974; Leck, 1979), fragmentation did not appear to negatively aect B. cinerea. Its capacity to cross open areas (Willis, 1979) may minimize the impact of forest reduction, allowing it to use resources from other nearby areas (Rolstad, 1991). Batara cinerea was observed crossing open areas of up to 60 m. This may allow individuals to explore many patches in a landscape such as Caucaia, where the mean distance between patches is small (54 m). Moreover, the species does not seem to be restricted by vegetation quality as it was frequently seen at forest edges. Perhaps the species suers negative eects from fragmentation in cases where the mean distance between patches is substantially larger than the one in Caucaia, as it was seen by other researchers (Aleixo and Vielliard, 1995; Anjos and Boc on, 1999). Thus T. surrucura and B. cinerea respond dierently to fragmentation, besides their large body size and thus high energy demand. While T. surrucura is concentrated mainly in large patches, B. cinerea had a reasonably even distribution among all patches. This dierence suggests that the dispersal capacity of T. surrucura is probably attenuated through non-forest areas. Therefore, it may be more dependent on local resources and more restricted to patches where it already occurs in comparison

516

A. Uezu et al. / Biological Conservation 123 (2005) 507519

to B. cinerea, which could search for resources in more than one fragment. 4.2. Structural and functional connectivity The species studied show distinct responses to structural connectivity. The distance between patches is an important parameter accounting for the variation in P. leucoptera abundance. However, since patches were relatively close to each other, the 60 m dispersion capacity demonstrated by this species is sufcient to the species to eectively disperse between most patches. From the perspective of P. leucoptera, most patches are probably linked. However, the results also suggest that this parameter varies among the habitat patches surrounding each studied area in a 1 km radius, and that such a factor is related with P. leucoptera abundance. Thus, besides the capacity of the species to cross open areas, this movement may be less frequent as the distance between patch increases. Similar results were observed with a toucan species (Ramphastos sulfuratus) in Southern Mexico, where individual movement between habitat patches was positively correlated with distance between them (Graham, 2001). Although for P. leucoptera functional connectivity is better dened when we consider its capacity to cross open areas, corridors may play a more important role for connecting local populations as the distances between patches increase. Structural connections seems to be relevant only for B. leucoblepharus, an understory insectivorous species that forages near the ground, prefers humid areas and avoids open areas (Silva, 1991). These peculiarities should make it more averse to cross open areas. Thus, we expected it to use corridors more frequently than the other species to move from one patch to another. However, the use of corridors seems not to be related to a low capacity to cross open areas or to use young secondary forests, since B. leucoblepharus showed similar characteristics as P. leucoptera and C. caudata. Moreover, the use of corridors is neither related to the capacity to use forest edge habitats (which characterize the corridors) since most of the contact with the species in point counts occurred in patch cores (i.e. >50 m from the forest edges). The option to cross longer distance using corridors or undertake shorter movements through the matrix is at least partly related to the energetic cost of moving and the risk of predation/mortality (Tischendorf and Fahrig, 2000). Studies have demonstrated that some bird species opt for the longer route (i.e. going around the forest) and only when the ratio between the distance through the forest and through the open area is high, they prefer lisle and Desrochers, 2002). The the shorter option (Be capacity to cross non-habitat areas is also intrinsically related to species characteristics such as: body size lar-

lisle ger species tend to have a lower predation risk (Be and Desrochers, 2002); degrees of demand for specic habitats generalists are less averse to open areas (Greenberg, 1989); and behavior - some species move faster through areas of non-habitat and thus reduce predation risk (Sieving et al., 1996). Nonetheless, our results suggest that it is dicult to dene which biological characteristics predispose a species to use corridors or to cross an open matrix. Thus, although B. leucoblepharus and P. leucoptera have distinct responses to landscape structure, they show many similar biological characteristics. Both are small understory insectivores, are able to cross open areas and can use young forests. We did not detect any characteristic that produced a different perception of landscape elements for these two species. Such perception might reect ancestral dierences and/or may be due to some behavioral characteristics. For example, P. leucoptera developed a specialization to forage following ants and thus abundance of this resource might reect in its capacity to survive in habitat patches or to use corridors. For B. cinerea and C. caudata, neither parameter of structural connectivity seems to explain variation in abundance. The distance between patches is probably insucient to impact abundance and does not correspond to a signicantly elevated cost of dispersion for these species. In this situation, all patches are functionally connected and species distribution and abundance are probably more associated with local factors, such as resource availability. 4.3. Conservation implications Although fragmented landscapes support a proportion of bird community, local extinctions occur when there is intense fragmentation (Ribon et al., 2003; Develey and Metzger, 2005). Triclaria malachitacea and C. cucullatus illustrate the importance of continuous areas, such as the Morro Grande Reserve, for the maintenance of a complete bird community. These two area demanding species might be considered umbrella-species potentially useful in dening appropriate Atlantic rainforest conservation areas for similarly sensitive birds. Although this biome has been substantially diminished, the time lag between forest loss and extinction (Brooks and Balmford, 1996; Brooks et al., 1999b) and the distribution patterns of the remnant fragments may explain the lack of signicant mass extinctions. The remaining forest areas tend to be comprised of large fragments (e.g. Morro Grande) surrounded by many smaller patches. Species loss could have been substantially higher if the fragmentation had been more homogeneous. The larger patches (>60 ha) were extremely important, guaranteeing the presence of demanding species, such as T. surrucura, in this fragmented landscape. Con-

A. Uezu et al. / Biological Conservation 123 (2005) 507519

517

sequently, their conservation should be a priority, particularly since many other species have a similar pattern. This results is also important in the larger Atlantic Rainforest context, because many remnants of this biome are smaller than 30 ha. Our results suggest that the distance between patches is also an important consideration since shorter interpatch distances facilitate biological uxes through the landscape and thus promote species maintenance. Thus, small patches may act as stepping-stones for dispersion between larger patches. For this reason, landscape management should also incorporate additional elements such as small patches, corridors and dierent types of matrices (e.g. young secondary forest). Our study suggests that the response to habitat alteration and fragmentation is species-specic. Some taxa were more aected by fragment size, while others by connectivity. The limited dispersal capacity of understory insectivorous birds, for example, may be a key factor determining their sensitivity to fragmentation (Sekercioglu et al., 2002). With respect to connectivity, some species may be more sensitive to inter-patch distance, whereas others are more sensitive to the presence of corridors. These dierent perceptions are particularly relevant since they can elucidate which landscape components play major roles for each species. Functional connectivity is thus a key property to dene functional species groups in the context of fragmentation, which allow us to make generalizations about its eects. Conservation may be more eective if we consider the perception and demands of these dierent functional groups.

Acknowledgments We thank Clinton Jenkins, Pedro Develey, Alexandre Martensen, and Pedro M. Pedro for revising a previous version of the manuscript. We are also grateful to Denis Saunders, Cagan Sekercioglu, and an anonymous reviewer for their detail analysis and eciency throughout the review process. We appreciate the cooperation during all the steps of this study of members of the project Biodiversity conservation in fragmented landscapes at the Atlantic Plateau of Sa o Paulo. We thank SABESP, which allowed us access to the Morro Grande Reserve and also all private owners of the properties containing patches. Without their agreement this study could not have been undertaken. Financial support was provided by FAPESP (99/05123-4; 00/01120-0).

References
Adriaensen, F., Chardon, J.P., De Blust, G., Swinnen, E., Villalba, S., Gulinck, H., Matthysen, E., 2003. The application of least-cost

modelling as a functional landscape model. Landscape and Urban Planning 64, 233247. mica da Aleixo, A., Vielliard, J.M.E., 1995. Composic a o e dina avifauna da mata de Santa Genebra, Campinas, Sa o Paulo, Brazil. Revista brasileira de Zoologia 12, 493511. n, H., 1994. Eects of habitat fragmentation on birds and Andre mammals in landscapes with dierent proportions of suitable habitat: a review. Oikos 71, 355366. Anjos, L., Boc on, R., 1999. Bird communities in natural forest patches in Southern Brazil. Wilson Bulletin 111, 397414. Antongiovanni, M., Metzger, J.P., 2005. Inuence of the matrix habitats on the occurrence of insectivorous bird species in Amazonian forest fragments. Biological Conservation 122, 441 451. Beier, P., Noss, R.F., 1998. Do corridors provide connectivity?. Conservation Biology 12, 12411252. lisle, M., Desrochers, A., 2002. Gap-crossing decisions by forest Be birds: an empirical basis for parameterizing spatially-explicit, individual-based models. Landscape Ecology 17, 219231. Bernacci, L.C., Durigan, G., Correia, G., Arbocz, G., Catharino, E., stica e estrutura da Metzger, J.P.M., 2004. Composic a o or na. Anexo vegetac a o em fragmentos orestais do Planalto de Ibiu 7. In: Metzger, J.P. (Ed.), Conservac a o da Biodiversidade em ntico de Sa Paisagens Fragmentadas no Planalto Atla o Paulo. FAPESP 99/05123-4 technical report n. 4, Sa o Paulo, Brazil. Bierregaard Jr, R.O., Stouer, P.C., 1997. Understory birds and dynamic habitat mosaics in Amazonian Rainforests. In: Laurance, W.F., Bierregaard, R.O. Jr. (Eds.), Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities. University of Chicago Press, Chicago, pp. 138155. Blake, J.G., Karr, J.R., 1987. Breeding birds of isolated woodlots: area and habitat relationships. Ecology 68, 17241734. thode des indices Blondel, J., Ferry, C., Frochot, B., 1970. La me s davifaune par ponctuels dabondance (I.P.A.) ou des releve coute. Alauda 38, 5571. stations de cnicas de playback no desenvolvimento de Boscolo, D., 2002. Uso de te todo capaz de atestar a presenc ncia de aves no um me a ou ause interior de fragmentos orestais. MS thesis, University of Sa o Paulo, Sa o Paulo, Brazil. Brooks, T., Balmford, A., 1996. Atlantic forest extinctions. Nature 380, 115. Brooks, T., Tobias, J., Balmford, A., 1999a. Deforestation and bird extinctions in the Atlantic Forest. Animal Conservation 2, 211 222. Brooks, T., Pimm, S.L., Oyugi, J.O., 1999b. Time lag between deforestation and bird extinction in tropical forest fragments. Conservation Biology 13, 11401150. Castelletta, M., Sodhi, N.S., Subaraj, R., 2000. Heavy extinctions of forest avifauna in Singapore: lessons for conservation biodiversity in Southeast Asia. Conservation Biology 14, 18701880. Christiensen, M.M., Pitter, E., 1997. Species loss in a forest bird community near Lagoa Santa in Southeastern Brazil. Biological Conservation 80, 2332. , M.E., 2000. Crooks, K.R., Suarez, A.V., Bolger, D.T., Soule Extinction and colonization of birds on habitat islands. Conservation Biology 15, 159172. ria E a Devastac Dean, W., 1996. A Ferro E Fogo A Histo a o Da ntica Brasileira. Companhia das Letras, Sa Mata Atla o Paulo, Brazil. Develey, P.F., 1997. Ecologia de bandos mistos de aves de Mata ntica na Estac gica Jure ia-Itatins, Sa Atla a o Ecolo o Paulo, Brazil. MS Thesis, University of Sa o Paulo, Sa o Paulo, Brazil. Develey, P.F., Metzger, J.P., 2005. Birds in Atlantic forest landscapes: eects of forest cover and conguration. In: Laurance, W., Peres, C.A. (Eds.), Emerging Threats to Tropical Forests. University of Chicago Press, Chicago (in press).

518

A. Uezu et al. / Biological Conservation 123 (2005) 507519 Metzger, J.P., Simonetti, C., Uezu, A., Diederichsen, A., Goulart, W., 2002. Estrutura da paisagem em Caucaia do Alto (Planalto de na, SP). Anexo 1. In: Metzger, J.P. (Ed.), Conservac Ibiu a o da ntico Biodiversidade em Paisagens Fragmentadas no Planalto Atla de Sa o Paulo. FAPESP 99/05123-4 technical report n. 2 Sa o Paulo, Brazil. Mittermeier, R.A., Myers, N., Gil, P.R., Mittermeier, C.G., 1999. Hotspots: Earths Biologically Richest and Most Endangered Terrestrial Ecoregions. CEMEX, Mexico City, Mexico. Myers, N., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853858. Newmark, W.D., 1991. Tropical Forest fragmentation and the local extinction of understory birds in the Eastern Usambara mountains, Tanzania. Conservation Biology 5, 6778. Oliveira-Filho, A.T., Fontes, M.A.L., 2000. Patterns of oristic dierentiation among Atlantic Forests in southeastern Brazil and inuence of climate. Biotropica 32, 793810. Pacheco, J.F., Bauer, E.C., 2000. Biogeograa e conservac a o da ntica e campos sulinos construc vel Avifauna da Mata Atla a o e n rias atual de conhecimento. Workshop Avaliac a o e Ac o es Priorita ntica e Campos para Conservac a o dos Biomas Floresta Atla ntese dos resultados e Relato rio Final do Grupo Sulinos. S tico Aves. Available from: <http://www.conservation.org.br/ Tema ma/rnais/rt_aves.html>. Parker, T.A., 1991. On the use of tape recorders in avifaunal surveys. Auk 108, 443444. Pearson, S.M., 1993. The spatial extent and relative inuence of landscape-level factors on wintering bird populations. Landscape Ecology 8, 318. Ponc ano, W.L., Carneiro, C.D.R., Bistrichi, C.A., Almeida, F.F.M. gico Do Estado De Sa de, Prandini, F.L., 1981. Mapa Geomorfolo o gicas, Sa Paulo, 1. IPT Instituto de Pesquisas Tecnolo o Paulo. Price, O.F., Woinarski, J.C.Z., Robinson, D., 1999. Very large area requirements for frugivorous birds in monsoon rain forests of the Northern Territory, Australia. Biological Conservation 91, 169 180. Renjinfo, L.M., 1999. Composition changes in a subandean avifauna after long-term forest fragmentation. Conservation Biology 13, 11241139. Ribon, R., Simon, J.E., De Mattos, G.T., 2003. Bird extinctions in Atlantic Forest Fragments of Vic osa Region, Southeastern Brazil. Conservation Biology 17, 18271839. Ridgely, R.S., Tudor, G., 1989. The birds of South AmericaThe Oscines Passerines, 1. Univ. Texas Press, Austin. Rolstad, J., 1991. Consequences of forest fragmentation for the dynamics of birds populations: conceptual issues and the evidence. Biological Journal of the Linnean Society 42, 149163. Seabra, M., 1971. Vargem Grande: Organizac a o E Transformac a o De Um Setor Do Cintura o Paulistano. USP, Instituto de Geograa, Sa o Paulo, Brazil. Sekercioglu, C.H., Ehrlich, P.R., Daily, G.C., Aygen, D., Goehring, , R.F., 2002. Disappearance of insectivorous birds from D., Sand tropical forest fragments. Proceedings of the National Academy of Sciences 99, 263267. Sick, H., 1997. Ornitologia Brasileira. Editora Nova Fronteira S.A, Rio de Janeiro. Sieving, K.E., Wilson, M.F., De Santo, T.L., 1996. Habitat barriers to movement of understory birds in fragmented South-Temperate Rainforest. The Auk 113, 944949. gicos, bioacu cos e sticos, biogeogra Silva, W.R., 1991. Padro es ecolo ticos do complexo Basileuterus culicivorus (aves, Parulidae) logene cies brasileiras do ge nero. Ph.D. Thesis, State e demais espe University of Campinas, Campinas, Brazil. Simberlo, D., 1998. Flagships, umbrellas, and keystones: is single in the landscape era?. Conservation species management passe Biology 83, 247257.

Diamond, J.M., Bishop, K.D., Balen, S.V., 1987. Bird survival in an isolated Javan Woodland: island or mirror?. Conservation Biology 1, 132142. Elkie, P., Rempel, R., Carr. A., 1999. Patch Analyst Users Manual. Ontario Ministry of Natural Resources Northwest Science and Technology, Thunder Bay, Ont., Canada. ntica/INPE, 2002. Atlas da Evoluc Fundac a o SOS Mata Atla a o dos nio Remanescentes Florestais e Ecossistemas Associados no Dom ntica no Per odo de 19952000, Sa da Mata Atla o Paulo, SP. Gascon, C., Lovejoy, T.E., Bierregaard, R.O., Malcolm, J.R., Stouer, P.C., Vasconcelos, H., Laurance, W.F., Zimmerman, B., Tocher, M., Borges, S., 1999. Matrix habitat and species persistence in tropical forest remnants. Biological Conservation 91, 223229. Gaston, K.J., Blackburn, T.M., 1996. Rarity and body size: some cautionary remarks. Conservation Biology 10, 12951298. Goerck, J.M., 1997. Patterns of rarity in the birds of the Atlantic Forest of Brazil. Conservation Biology 11, 112118. Graham, C.H., 2001. Factors inuencing movement patterns of KeelBilled Toucans in a fragmented tropical landscape in Southern Mexico. Conservation Biology 15, 17891798. Greenberg, R., 1989. Neophobia, aversion to open space, and ecological plasticity in song and swamp sparrows. Canadian Journal of Zoology 67, 11941199. Hanski, I., Gilpin, M.E., 1997. Metapopulation Biology: Ecology, Genetics, and Evolution. Academic Press, San Diego, CA. Henle, K., Davies, K.F., Kleyer, M., Margules, C., Settele, J., 2004. Predictors of species sensitivities to fragmentation. Biodiversity and Conservation 13, 207251. ntica, regia Ho o ing, E., Lencioni, F., 1992. Avifauna da Floresta Atla polis, Estado de Sa de Saleso o Paulo. Revista Brasileira de Biologia 45, 361378. IUCN, 2003. 2003 IUCN Red List of Threatened Species. Available from: <www.redlist.org>. Jokima ki, J., Huhta, E., 1996. The eects of landscape matrix and habitat structure on a bird community in northern Finland: a multi-scale approach. Ornis Fennica 73, 97113. Kattan, G.H., 1994. Forest fragmentation and bird extinctions: San Antonio eighty years later. Conservation Biology 8, 138146. Karr, J.R., 1982. Avian extinction on Barro Colorado Island, Panama: a reassessment. The American Naturalist 119, 220239. Lambeck, R.J., 1997. Focal species: a multi-species umbrella for nature conservation. Conservation Biology 11, 849856. Laurance, W.F., 1990. Comparative responses of ve arboreal marsupials to tropical forest fragmentation. Journal of Mammalogy 71, 641653. Laurance, W.F., 1991. Ecological correlates of extinction proneness in Australian Tropical Rain Forest mammals. Conservation Biology 5, 7989. Leck, C.L., 1979. Avian extinctions in an isolated Tropical Wet-Forest preserve, Ecuador. The Auk 96, 343352. Levins, R., 1970. Extinction. In: Gerstenhaber, M. (Ed.), Some Mathematical Problems in Biology, Lectures on Mathematical in Life Sciences, 2. American Mathematical Society Providence, Rhode Island, pp. 77107. iade, Magalha es, J.C.R., 1999. As Aves na Fazenda Barreiro Rico, Ple Sa o Paulo. Maldonado-Coelho, M., Marini, M., 2000. Eects of Forest fragment size and successional stage on mixed-species bird ocks in southeastern Brazil. The Condor 102, 585594. Manne, L.L., Pimm, S.L., 2001. Beyond eight forms of rarity: which species are threatened and which will be next?. Animal Conservation 4, 221229. Marini, M.A., 2001. Eects of forest fragmentation on birds of the cerrado region, Brazil. Bird Conservation International 11, 1325. camps, H., 1997. The structural connectivity threshMetzger, J.P., De old: an hypothesis in conservation biology at the landscape scale. Acta cologica 18, 112.

A. Uezu et al. / Biological Conservation 123 (2005) 507519 , M.E., Gilpin, M.E., 1991. The theory of wildlife corridor Soule capability. In: Saunders, D.A., Hobbs, R.J. (Eds.), Nature Conservation 2: the Role of Corridors. Surrey Beatty and Sons Pty Limited, Chipping Norton, NSW, Australia, pp. 38. Stotz, D.F., Fitzpatrick, J.W., Parker III, T.A., Moskovits, D.K., 1996. Neotropical Birds. Ecology and Conservation. The University of Chicago Press, Chicago. Stouer, P.C., Bierregaard, R.O., 1995. Use of Amazonian forest fragments by understory insectivorous birds. Ecology 76, 2429 2445. Stratford, J.A., Stouer, P.C., 1999. Local extinctions of terrestrial insectivorous birds in a fragmented landscape near Manaus, Brazil. Conservation Biology 13, 14161423. Taylor, P.D., Fahrig, L., Henein, K., Merriam, G., 1993. Connectivity is a vital element of landscape structure. Oikos 69, 571 573. Terborgh, J., 1974. Preservation of natural diversity: the problem of extinction prone species. Bioscience 24, 715722.

519

Terborgh, J., Robinson, S.K., Parker III, T.A., Munn, C.A., Pierpont, N., 1990. Structure and organization of an Amazonian Forest bird community. Ecological Monographs 60, 213238. Tischendorf, L., Fahrig, L., 2000. On the usage and measurement of landscape connectivity. Oikos 90, 719. Vielliard, J., 1989. O registro dos sinais de comunicac a o sonora em metros etolo gicos e testes de playback. Anais de Etologia aves: para 7, 134149. Vielliard, J., Silva, W.R., 1990. Nova metodologia de levantamento quantitativo da avifauna e primeiros resultados do interior do Estado de Sa o Paulo, Brazil. Anais do IV Encontro Nacional dos Anilhadores de Aves. Universidade Federal Rural de Pernambuco, Recife, pp. 117151. Willis, E.O., 1979. The composition of avian communities in three is Avulsos de remanescent woodlots in southern Brazil. Pape Zoologia 33, 125. Zar, J.H., 1996. Biostatistical Analysis. Prentice-Hall, Upper Saddle River, NJ.