Species characteristics associated with ... - Wiley Online Library

Animal Conservation. Print ISSN 1367-9430

Species characteristics associated with extinction vulnerability and nestedness rankings of birds in tropical forest fragments K. J. Feeley1, T. W. Gillespie2, D. J. Lebbin3 & H. S. Walter4 1 Center for Tropical Forest Science, Harvard University Arnold Arboretum, Cambridge, MA, USA 2 Department of Geography, University of California, Los Angeles, CA, USA 3 Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, NY, USA 4 Department of Geography, University of California, Los Angeles, CA, USA

Keywords extinction vulnerability; forest fragmentation; faunal relaxation; Lago Guri; natural history characteristics; nestedness; Venezuela. Correspondence Kenneth J. Feeley, Center for Tropical Forest Science, Harvard University Arnold Arboretum, 22 Divinity Avenue, Cambridge MA 02138 USA. Email: [email protected] Received 22 March 2007; accepted 13 August 2007 doi:10.1111/j.1469-1795.2007.00140.x

Abstract Following habitat fragmentation, species are predicted to go locally extinct from remnant patches in a predictable order due to differential extinction vulnerabilities. This selective species loss will result in nested distributions of species such that species found in depauperate patches will also tend to be found in larger, more speciose patches. Therefore, it should be possible to determine the relationship between species-specific characteristics and extinction vulnerability by comparing the order in which species are nested [i.e. nestedness ranking (NR)] with various natural history characteristics available from the literature and/or collected in the field. In this study, we investigate the relationship between the NRs of 41 resident forest-interior bird species inhabiting recently isolated landbridge islands in Lago Guri, Venezuela, with a large number of natural history characteristics collected from the literature (regional abundance, body length, habitat specificity, trophic guild, sensitivity to disturbance, range size) and from the field (local population density). In a comparison of the best regression models generated using just variables available through the literature (i.e. no local population density) with the best model generated using all possible variables, we found that the inclusion of field-based data significantly improved the amount of variation explained. The best overall model (r2 = 0.40, Po0.001) included body size, habitat specificity, zoogeographic distribution (a measure of range size) and local population density as predictors of NR. Understanding the factors that influence extinction vulnerability has important implications for conservation and could be used to help direct management efforts.

Introduction Despite increasing conservation efforts, the loss of tropical forests continues at an alarming rate (Fearnside, 2005; Jha & Bawa, 2006). This deforestation inevitably results in the fragmentation of primary habitat into isolated patches (Skole & Tucker, 1993; Wade et al., 2003). Numerous studies have shown that following fragmentation, the resident faunal community will undergo a period of species loss (sometimes referred to as ‘relaxation’ or ‘downsizing’; Walter, 2004) before a new equilibrium community is achieved (Diamond, 1972; Terborgh, 1974; Crooks et al., 2001; Ferraz et al., 2003; Sodhi et al., 2006; Stouffer et al., 2006). During relaxation, fragments have been observed to lose species in a predictable order on the basis of differential extinction vulnerabilities. This selective species loss will generally result in a nested structure among the communities such that the species found on depauperate islands also tend

to be found on larger, more speciose islands (Patterson, 1987; Blake, 1991; Bolger, Alberts & Soule`, 1991; Patterson & Atmar, 2000). As such, it should be possible to determine the relationship between species-specific attributes and extinction vulnerability by comparing the order in which species are nested with various natural history characteristics (Bolger et al., 1991; Martinez-Morales, 2005). In a study of the avian communities inhabiting recently isolated landbridge islands in Lago Guri, Venezuela, Feeley (2003) found that the resident forest-interior bird communities displayed a significantly nested distributional pattern that was hypothesized to be the result of species’ differential extinction rates (Feeley, 2003). Here, we compare the order in which the species of Lago Guri are nested [nestedness ranking (NR); species found on only the more speciose islands, and hence believed to have a high extinction vulnerability, are assigned a low rank] with a large number of natural history characteristics that have been hypothesized

c 2007 The Authors. Journal compilation c 2007 The Zoological Society of London Animal Conservation 10 (2007) 493–501

493

Attributes associated with extinction vulnerability

to be associated with local extinction of forest interior birds in the tropics. Natural history characteristics included regional abundance (Diamond, Bishop & Balen, 1987; Newmark, 1991; Warburton, 1997), body size (Willis, 1974; Leck, 1979; Terborgh & Winter, 1980; Karr, 1982; Gillespie, 2000), habitat specificity (Diamond et al., 1987; Newmark, 1991; Kattan, Lopez & Giraldo, 1994; Warburton, 1997; Robinson, 1999), trophic guild (Terborgh, 1974; Kattan et al., 1994; Christiansen & Pitter, 1997), sensitivity to disturbance (Canaday, 1997; Estrada, Coates-Estrada & Meritt, 1997; Thiollay, 1997; Thiollay, 1999; Lens et al., 2002), range size (Terborgh & Winter, 1980; Simberloff, 1994; Turner, 1996; Manne, Brooks & Pimm, 1999) and local population density (Terborgh & Winter, 1980; Diamond et al., 1987; Pimm, Jones & Diamond, 1988; Diamond & Pimm, 1993; Walter, 2004). Specifically, we test the hypotheses that species with high regional abundances, large range sizes and low habitat specificity would be relatively resistant to local extinction and would thus have high NR. We also hypothesized that certain foraging guilds should be relatively prone to extinction from the habitat fragments (Terborgh, 1974). Several studies have shown frugivores to be particularly vulnerable to the deleterious effects of fragmentation. Fruit resources are often locally rare and temporally patchy (Levey & Stiles, 1992). Small fragments often do not contain sufficient reliable supplies of fruit resources to sustain resident populations of frugivores (Terborgh & Winter, 1980; Faaborg, 1982; Howe, 1984; Kattan et al., 1994). Insectivores and omnivores, on the other hand, depend on resources that are generally more evenly distributed and temporally dependable and thus may be able to persist in smaller habitat patches. Also, small birds generally tend to be less extinction prone than large birds (Terborgh, 1974; Terborgh & Winter, 1980; Gaston & Blackburn, 1995). This may be due in part to the association between body size and other ecological factors such as population density and guild status. In addition, small-bodied species tend to have higher reproductive rates than large-bodied species and therefore may be more capable of recovering from population crashes that could otherwise lead to local extinction (Sæther et al., 2005). Finally, species with low local population densities often require large areas to meet their resource requirements and thus we hypothesized that these species would be relatively prone to extinction and would have correspondingly low ranks (Bolger et al., 1991). Low population densities may be due to specialization on spatially or temporally dispersed resources (Terborgh & Winter, 1980), in which case, many habitat patches may be too small to meet these species’ resource needs, leading to a high risk of local extinction. The potential value of any model for conservation application is dependent not only on the amount of variation it can explain but also on how much work is required to collect the necessary data. Therefore, we compared the best possible model of NRs generated using just the natural history information that is available from the published literature (and thus relatively easy to obtain and available for other regions) with the best model generated from the literature494

K. J. Feeley et al.

based data plus information on local population densities. Understanding the relationships between species attributes and extinction vulnerability will have important implications for conservation, improving predictions about which bird species are most threatened by future disturbances and helping direct management efforts.

Methods Study site Lago Guri is a large hydroelectric reservoir created in 1986 upon closing of the Raul Leoni Dam on the lower Rio Caroni in east–central Venezuela (Fig. 1). The inundation of over 4300 km2 of hilly terrain resulted in the fragmentation of once continuous forest into hundreds of land-bridge islands. The habitat of all the islands used in this study is semi-deciduous tropical dry forest (Huber, 1986; Terborgh et al., 1997b), with a single dry season generally lasting from October to May. The height of the forest ranges from 15 to 20 m, with occasional emergent trees reaching 425 m (Huber, 1986; Aymard, Norconk & Kinzey, 1997).

Field methods This study was conducted using data collected from spotmap censuses of the forest-interior bird communities residing on 26 islands (ranging in area from 0.2 to 190 ha; Table 1, Fig. 1; Feeley, Gillespie & Terborgh, 2005) at the beginning of the wet season (May through August) of 2000. Detailed descriptions of the census methods have been presented elsewhere (Terborgh, Lopez & Tello, 1997a; Feeley, 2003), and thus here we only provide a brief description. Small (o2.5 ha) and medium islands (o25 ha) were all censused in their entirety using the trails that either bisected the islands along the primary axes (small islands) or circumscribed the perimeter (medium islands). The single large island (#26= 190 ha; Fig. 1) was sub-sampled with a 26 ha plot gridded with parallel trails located c. 100 m apart. All islands were censused five to nine times during hours of peak vocal activity (between 5:30 and 8:00 h). The order in which islands were censused and the direction in which the trails were walked were alternated in order to minimize any potential bias. Censuses were not conducted during inclement weather. Following the protocol of Terborgh et al. (1997a), all vocalizations and sightings of birds were identified and mapped. We divided the species into five categories: (1) presumptive visitors or non-breeding individuals; (2) species that utilize large territories incorporating multiple islands (such as raptors, macaws, large pigeons and large woodpeckers); (3) edge species; (4) aquatic species; and (5) resident forest-dwelling and presumably breeding species. Only species belonging to category 5, forest-interior residents, were included in the analyses. To be counted as a resident, a pair/male had to be observed during at least 2/3 of the census visits. Birds observed in less than 2/3 of the census visits were classified as visitors and excluded from the analyses (Terborgh et al., 1997a). Lago Guri islands are


K. J. Feeley et al.


= 1000 m = North

Mainland

N W

E S

Figure 1 Map of Venezuela with the location of Lago Guri (first inset) and study islands (second inset; black = land, gray = water) indicated. Island numbers refer to the information in Tables 1 and 2.

surrounded by the snags of trees killed by inundation. The snags form a distinct matrix from the islands and are used by a number of bird species for nesting, foraging and roosting (e.g. some falcons, woodpeckers, flycatchers, martins, swallows). Such ‘edge’ species were excluded from the analyses. Additional details regarding census methodology and classifications are contained in Terborgh et al. (1997a). Previous analyses of Guri census data (Terborgh et al., 1997a) included large pigeons (Columba spp.) and excluded the house wren Trogolydes ae¨don from the category of forest interior residents (category 5). Here, we followed the protocol of Feeley (2003) and excluded Columba spp. (because they were frequently observed to be flying between islands) but included T. ae¨don after observations indicated that they forage primarily in the forest interior of Lago Guri islands (K. Feeley, D. Lebbins and J. Hardesty, unpubl. data). In addition, two species, Tolmomyias polyocephalus and Monasa atra, were erroneously included in the initial species lists for 2000 (Feeley, 2003) due to misidentifications and are excluded from the current analyses.

Natural history characteristics Data on relative regional abundances, habitat specificity, range size and body length were collected from published sources for all 41 bird species inhabiting the study islands. These data were then transformed into ordinal rankings according to the hypothesized relationships. Species were classified as either uncommon (1), fairly common (2) or common (3) following the relative regional abundance

classifications for the Neotropics in Stotz et al. (1996). Stotz et al. (1996) identified 41 habitat or vegetation types in the Neotropics and habitat specificity was based on the incidence or number of habitats used by a given species (Stotz et al., 1996). Trophic guilds followed Meyer de Schauensee & Phelps (1978) and Stiles & Skutch (1989) and were quantified as frugivores (1); nectivores (2); insectivores (3); omnivores (4). There were no strict carnivores included in this study. Sensitivity to disturbance was classified as low (1); medium (2); high (3) following Stotz et al. (1996). We used two indices to estimate the range size for each species: latitudinal extent and zoogeographic distribution. Latitudinal extent is the straight-line distance (in 1) between the northern and southern extremes of a species’ breeding range with migratory ranges excluded (Gaston, 1996). We calculated latitudinal extent using published range maps and breeding records. Zoogeographic distribution was based on species’ incidence in 22 zoogeographic regions within the Neotropics according to Stotz et al. (1996). Body length (cm) was taken from Meyer de Schauensee & Phelps (1978). In addition to the character states collected from the literature, we also calculated the population density (pairs ha1) for each species based on encounter rates during censuses of the largest island (#26; Fig. 1).

Data analysis Using the census data, the nestedness of the Lago Guri bird community was calculated (Feeley, 2003) using Atmar and


495


Table 1 Area (ha) and number of forest-interior bird species residing on the study islands at Lago Guri, Venezuela Island ID#

Area (ha)

Number of bird species

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

0.2 0.3 0.3 0.3 0.5 0.5 0.6 0.6 0.7 0.7 0.8 0.9 1.0 1.4 1.5 1.7 2.3 2.4 9.3 9.7 11.1 12.0 12.3 14.7 23.3 180.0

2 9 1 1 2 5 7 8 4 2 3 7 10 12 14 4 3 10 16 6 18 7 8 15 13 40

Island identification numbers match those in Table 2 and Fig. 1.

Patterson’s nestedness calculator (Atmar & Patterson, 1995). This program ranks species (and fragments) such that the system’s degree of nestedness is maximized (Atmar & Patterson, 1993; Sfenthourakis, Giokas & Tzanatos, 2004). We used this NR as a proxy for extinction vulnerability such that species found only on the larger islands have a low NR and were considered to be relatively more prone to local extinction. NRs cannot distinguish between species that have identical distributions (i.e. occur on all of the same islands) and thus these species were assigned the average rank for all ‘equivalent’ species. The nestedness calculator and its measure of communitywide nestedness (temperature) have recently come under scrutiny (Cook & Quinn, 1998; Fischer & Lindenmayer, 2002; Ulrich & Gotelli, 2007). Criticisms have focused primarily on the sensitivity of the temperature index to matrix size, fill and structure as well as its purported inability to distinguish between communities that are nested due to passive sampling versus selective extinction or other ecological processes. The sorting of species within communities on the basis of nestedness has received less attention, but in order to ensure that our results were not biased by the use of NR, we also calculated two alternative sets of species rankings based on (1) the number of islands occupied by each species and (2) the area of the smallest island occupied by each species. NR was very highly correlated with both alternative rankings (no. of islands: R= 0.99, Po0.0001; 496

K. J. Feeley et al.

area of smallest island: R =0.93, Po0.0001). Analyses based on these ranking schemes produced equitable results and are not discussed further here. We further tested the performance of the NR through comparisons with the rates of extinction (k) observed for species occurring on a subset of the study islands. Eleven of the 26 study islands (numbers 4, 8, 10, 12, 13, 14, 17, 21, 22, 23 and 26; Fig. 1) were censused in 1993, 2000, 2001, 2002 and 2003 using protocols identical to those described above (Feeley, 2005). Using these temporal data, we calculated k for each species based on the exponential decay function It ¼ I0 eðktÞ where I0 is the number of islands inhabited by a species before fragmentation (here assumed to be 11), It is the number of islands inhabited after t years and k is a decay constant describing the rate of extinction. NR scores were highly correlated with k (Spearman’s rank correlation, R = 0.70, Po0.001), thereby validating its use as a measure of relative extinction vulnerability. We were initially concerned that the NRs may be overly influenced by the inclusion of island # 26, which is markedly larger and more diverse than the other study islands. We therefore recalculated the NRs excluding island #26. Species’ NRs based on the full and reduced sets of study islands were very highly correlated (R= 0.94, Po0.001). Here, we only present results from the analyses using the rankings based on the full set of islands. To test the hypothesized relationships between the species-specific characteristics and extinction vulnerability, we conducted multiple regression analyses using NRs as the response variable and natural history characteristics as the main effects. The best overall models were identified by comparing all possible model subsets (comparisons based on AIC) using two different full models: (1) just character information obtained from the literature (i.e. no local population density) and (2) all variables including the local population density. We compared the fit of the two resultant ‘best’ models using a two-tailed F-test (Sokal & Rohlf, 1995).

Results There were a total of 41 forest-interior bird species recorded as residing on the 26 study island in Lago Guri, Venezuela. Ten of the 41 bird species were found to reside only on the largest study island and thus were all assigned the same NR (6.5). All other species had unique distributional patterns and inhabited between two and 22 of the islands (no single species inhabited all 26 islands; Table 2). The resident bird species exhibited a wide range of natural history characteristics. The body size ranged from 6.8 to 96 cm. Birds covered a wide range of latitudinal extents from 4 to 941, and the incidence in zoogeographic regions ranged from one to 20 regions, with a majority of the species occurring in one to six regions. Habitat specificity ranged from one to six habitat types, with a majority of the species occurring in four or less habitat types. There were four


K. J. Feeley et al.


Table 2 Presence–absence matrix for resident forest-interior birds recorded on study islands at Lago Guri, Venezuela, arranged to maximize nestedness according to Atmar & Patterson’s (1995) nestedness calculator (shaded= species occurrence)

SPECIES

ISLAND

NR 26 21 19 24 15 25 14 18 13

Melanerpes rubricapillus

41

Troglodytes aedon

40

Atalotriccus pilaris

39

Lepidocolaptes souleyetii

38

Myiarchus tyrannulus

37

Chlorostilbon mellisugus

36

Formicivora grisea

35

Myiarchus venezuelensis

34

Hylophilus flavipes

33

Tolmomyias flaviventris

32

Leptotila verreauxi

31

Vireo olivaceus

30

Thryothorus leucotis

29

Myiodynastes maculatus

28

Polioptila plumbea

27

Momotus momota

26

Penelope jacquacu

25

Myrmeciza longipes

24

Colibri delphinae

23

Hypnelus ruficollis

22

Thamnophilus punctatus

21

Pitangus sulphuratus

20

Basileuterus culicivorus

19

Chiroxiphia pareola

18

Myiopagis viridicata

17

Phaethornis superciliosus

16

Platyrinchus mystaceus

15

Ortalis motmot

14

Myiozetetes similis

13

Crypturellus erythropus

6.5

Cyclarhis gujanensis

6.5

Crax alector

6.5

Myiozetetes cayanensis

6.5

Megarynchus pitangua

6.5

Pteroglossus viridis

2 23 22

8

7 20 12

6 16 17 11

9 10

3

1

5

4

7

Piaya cayana

6.5

Camptostoma obsoletum

6.5

Tityra cayana

6.5

Geotrygon montana

6.5

Anthracothorax nigricollis

6.5

Xiphorynchus guttatus

6.5

. Species that are vulnerable to extinction are only found on the most speciose islands and have low nestedness rankings (NR). Island identification numbers refer to the information in Table 1 and Fig. 1.

trophic guilds represented (frugivores, insectivores, nectivores and omnivores), with insectivores being the most common (18 spp.). Most species (29 spp.) were classified as

having a low sensitivity to disturbance. Local densities, as recorded from the largest island, ranged from approximately one pair per every 3 ha to less than one pair per


497


26 ha. The natural history characteristics of all species are listed in Supplementary Material Appendix S1. When only the published information on natural history characteristics was included as potential explanatory variables, the best model included body length, habitat specificity and zoogeographic distribution [NR = 20.33–0.19 (body length)+2.69 (habitat specificity) 0.87 (zoogeographic regions) (r2 = 0.20, Po0.05)]. When local population density was added to the list of potential variables, the best model included the same variables as the previous model, plus local population density [NR = 11.27–0.13 (body length)+2.81 (habitat specificity) 0.79 (zoogeographic regions), +53.48 (population density), (r2 = 0.40, Po0.001)]. The model containing the local population density did describe a greater proportion of the variance and based on an F-test was significantly better at describing the data even after accounting for the decreased degrees of freedom (F37,36 = 12.00, Po0.001). The full model including all possible variables (field and literature based) explained 41% of the variation in NRs (Po0.05).

Discussion The best model describing relative extinction vulnerability using just the literature-based data included body size, habitat specificity and the number of zoogeographic regions (a measure of range size). If we also included the field-based measure of local population density in the list of potential variables, the best model contained all of the same explanatory variables plus local population density from the largest island. As hypothesized, NR was negatively associated with body length (indicating that large-bodied bird species are relatively prone to local extinction) and positively associated with habitat specificity. The fact that habitat specialists tend to be lost from small fragments is possibly due to the fact that smaller fragments inherently contain fewer microhabitat types (Terborgh & Winter, 1980). However, contrary to predictions, NR was negatively associated with the number of zoogeographic zones inhabited. The reason why species with wider geographic ranges are more prone to local extinction from small Guri islands is unknown. Not surprisingly, we found that the variable most strongly associated with NR was local population density. Population density was the only system-specific variable included and has been identified previously as one of the most important factors in influencing the distribution of bird species in fragmented tropical forests (Diamond et al., 1987; Pimm et al., 1988; Newmark, 1991). Local density has also been found to be an important factor influencing the persistence of a variety of taxa in other ecosystems (e.g. Soule` Allison & Bolger, 1992; Foufopoulos & Ives, 1999). The fact that the single best model incorporated local population density, even after taking into account the increased complexity and decreased degrees of freedom, indicates that while published natural history data can be potentially used as a first-order approximation of the relative extinction vulnerability of bird species, especially 498

K. J. Feeley et al.

for regions where little other data exist (Lindell et al., 2006), the inclusion of field data from even a single large fragment does significantly improve the model. Unfortunately, the experience, time and resources needed to calculate the local population densities of tropical birds are not always available (Brown, 1995; Margules & Pressey, 2000; Purvis & Hector, 2000). For example, this research required over 60 person hours to collect population densities in the largest fragments and over 500 person hours for all study islands. The value of including field-based data will depend on the resources available, the specific questions being addressed and the degree of accuracy required. Even with a full model including all of the possible variables, we could not explain more than half of the total variation in NRs. The relatively weak association between NRs and natural history characteristics may indicate that there is a large stochastic component to the order in which species are lost from fragments. This stochasticity is reflected in the fact that the Lago Guri bird communities do not form perfectly nested subsets (Terborgh et al., 1997a; Feeley, 2003) and that there is some turnover in species composition from year to year (Feeley, 2005). In addition, some noise may be due to inaccurate characterizations of the island communities; it is possible that some species were overlooked on some islands or that censuses were biased for or against certain groups of species. Finally, we did not include information on all of the factors that are likely associated with extinction vulnerability. Other natural history characteristics such as habitat preference and dispersal ability have been identified as potential predictors of species distributions (Beier, Drielen & Kankam, 2001; Lens et al., 2002; Van Houtan et al., 2007); by collecting data on these and other variables, it may be possible to increase the predictive power of our models. For this study, we based our estimates of extinction vulnerability and NRs on the distribution of bird species 14 years post-isolation. By this time, a considerable number of species may have already been lost from even the largest of our study islands (Feeley, 2005). It is therefore possible that our results were biased by inadvertently excluding some of the most vulnerable species. This may explain why the majority of species included in the analyses were classified as having a low sensitivity to disturbance and no species were classified as regionally rare (Supplementary Material Appendix S1). This problem could potentially have been avoided by including avian species occurrence information from a comparable undisturbed mainland control site; unfortunately, no such site exists at Lago Guri. Over the past two decades, the surrounding mainland forests have been burned repeatedly due to the action of ranchers attempting to open new grazing land and illegal hunters trying to flush game. In contrast, the islands have rarely burned (possibly due to their isolation and protected status). As a consequence of these different fire regimes, the mainland forest now differs significantly in structure and composition from the island forests. The faunal and floral community of the largest island included in this study (#26; Fig. 1) is believed to reflect more closely the general


K. J. Feeley et al.

conditions of the pre-inundation forest (J. Terborgh, pers. comm.). This assertion is supported by the fact that Terborgh et al. (1997a) found fewer bird species inhabiting the mainland than on the large island during their 1993 bird censuses. Tropical forests are becoming increasingly fragmented (Skole & Tucker, 1993; Wade et al., 2003), and there is a growing interest in predicting the effects of future disturbances, especially in biodiversity hotspots that have high levels of endemism and are highly endangered (Myers et al., 2000). Understanding the factors that influence extinction vulnerability has important conservation implications. By assessing the risk of local extinction for different target species, we can more successfully implement conservation projects designed at preventing species loss. Also, knowledge about the relative amount of time that various species will persist within a given set of fragments before going locally extinct could be used to determine the priority of restoration projects and direct conservation efforts.

Acknowledgements We would like to thank J. Terborgh, L. Lopez, D. Esclasans and L. Davenport for their assistance in the field. We also thank L. Balbas for logistical support. Funding for field work was provided by the American Museum of Natural History and the Georgia Ornithological Society.

References Atmar, W. & Patterson, B.D. (1993). The measure of order and disorder in the distribution of species in fragmented habitat. Oecologia 96, 373–382. Atmar, W. & Patterson, B.D. (1995). The nestedness temperature calculator. In A visual basic program, including 294 presence–absence matrices. University Park, NM: AICS Research Inc. Aymard, G., Norconk, M.A. & Kinzey, W.G. (1997). ´ flor ´ıstica de comunidades vegetales en islas Composicion en el embase de Guri, R ´ıo Caron ´ı, Estado Bolivar, Venezuela. BioLania 6, 195–233. Beier, P., Drielen, M.V. & Kankam, B.O. (2001). Avifaunal collapse in West African forest fragments. Conserv. Biol. 13, 1140–1150. Blake, J.G. (1991). Nested subsets and the distribution of birds on isolated woodlots. Conserv. Biol. 5, 58–66. Bolger, D.T., Alberts, A.C. & Soule`, M.S. (1991). Occurrence patterns of bird species in habitat fragments: sampling, extinction, and nested species subsets. Am. Naturalist 137, 155–166. Brown, J.H. (1995). Macroecology. Chicago: University of Chicago Press. Canaday, C. (1997). Loss of insectivorous birds along a gradient of human impact in Amazonia. Biol. Conserv. 77, 63–77.


Christiansen, M.B. & Pitter, E. (1997). Species loss in a forest bird community near Lagoa Santa in southeastern Brazil. Biol. Conserv. 80, 23–32. Cook, R.R. & Quinn, J.F. (1998). An evaluation of randomization models for nested species subsets analysis. Oecologia 113, 584–592. Crooks, K.R., Suarez, A.V., Bolger, T.G. & Soule`, M.S. (2001). Extinction and colonization of birds on habitat islands. Conserv. Biol. 15, 159–172. Diamond, J. & Pimm, S. (1993). Survival times of bird populations: a reply. Am. Naturalist 142, 1030–1035. Diamond, J.M. (1972). Biogeographic kinetics: estimation of relaxation times for avifaunas of Southwest Pacific islands. Proc. Natl. Acad. Sci. USA 69, 3199–3203. Diamond, J.M., Bishop, K.D. & Balen, S.V. (1987). Bird survival in an isolated Javan woodland: island or mirror? Conserv. Biol. 1, 132–142. Estrada, A., Coates-Estrada, R. & Meritt, D.A. Jr. (1997). Anthropogenic landscape changes and avian diversity at Lox Tuxtlas, Mexico. Biodivers. Conserv. 6, 19–43. Faaborg, J. (1982). Trophic and size structure of West Indian bird communities. Proc. Nat. Acad. Sci. USA 79, 1563–1567. Fearnside, P.M. (2005). Deforestation in Brazilian Amazonia: history, rates, and consequences. Conserv. Biol. 19, 680–688. Feeley, K. (2003). Analysis of avian communities in Lake Guri, Venezuela, using multiple assembly rule models. Oecologia 137, 104–113. Feeley, K.J. (2005). The effects of habitat fragmentation on tropical floral and faunal communities as mediated through trophic interactions. PhD dissertation, Duke University. Feeley, K.J., Gillespie, T.G. & Terborgh, J.W. (2005). The utility of spectral indices from Landsat ETM+for measuring the structure and composition of tropical dry forests. Biotropica 37, 508–519. Ferraz, G., Russell, G.J., Stouffer, P.C., Bierregaard, R.O., Pimm, S.L. & Lovejoy, T.E. (2003). Rates of species loss from Amazonian forest fragments. Proc. Nat. Acad. Sci. USA 100, 14069–14073. Fischer, J. & Lindenmayer, D.B. (2002). Treating the nestedness temperature calculator as a ‘‘black box’’ can lead to false conclusions. Oikos 99, 193–199. Foufopoulos, J. & Ives, A.R. (1999). Reptile extinctions on land-bridge islands: life history attributes and vulnerability to extinction. Am. Naturalist 15, 1–25. Gaston, K.J. (1996). Species-range-size distributions: patterns, mechanisms and implications. Trends Ecol. Evol. 11, 197–201. Gaston, K.J. & Blackburn, T.M. (1995). Birds, body size and the threat of extinction. Philos. Trans. Roy. Soc. Lond. B 347, 205–212. Gillespie, T.W. (2000). Rarity and conservation of forest birds in the tropical dry forest region of Central America. Biol. Conserv. 96, 161–168.


499


Howe, R.W. (1984). Local dynamics of bird assemblages in small forest habitat islands in Australia and North America. Ecology 65, 1585–1601. Huber, O. (1986). La vegetacion de la cuenca del Rio Caroni. Interciencia 11, 301–310. Jha, S. & Bawa, K.S. (2006). Population growth, human development, and deforestation in biodiversity hotspots. Conserv. Biol. 20, 906–912. Karr, J.R. (1982). Population variability and extinction in the avifauna of a tropical land bridge island. Ecology 63, 1975–1978. Kattan, G.H., Lopez, H.A. & Giraldo, M. (1994). Forest fragmentation and bird extinctions: San Antonio eighty years later. Conserv. Biol. 8, 138–146. Leck, C.F. (1979). Avian extinction in an isolated tropical wet-forest preserve, Ecuador. Auk 96, 343–352. Lens, L., Dongan, S.V., Norris, K., Githiru, M. & Matthysen, E. (2002). Avian persistence in fragmented rainforest. Science 298, 1236–1237. Levey, D.J. & Stiles, F.G. (1992). Evolutionary precursors of long-distance migration: resource availability and the movement patterns in Neotropical landbirds. Am. Naturalist 140, 447–476. Lindell, C.A., Chomentowski, W.H., Zook, J.R. & Kaiser, S.A. (2006). Generalizability of neotropical bird abundance and richness models. Anim. Conserv. 9, 445–455. Manne, L.L., Brooks, T.M. & Pimm, S.T. (1999). Relative risk of extinction of passerine birds on continents and islands. Nature 399, 258–261. Margules, C.R. & Pressey, R.L. (2000). Systematic conservation planning. Nature 405, 853–858. Martinez-Morales, M.A. (2005). Nested species assemblages as a tool to detect sensitivity to forest fragmentation: the case of cloud forest birds. Oikos 108, 634–642. Meyer de Schauensee, R.M. & Phelps, W.H. (1978). A guide to birds of Venezuela. Princeton, NJ: Princeton University Press. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B & Kent, J. (2000). Biodiversity hotspots for conservation priorities. Nature 403, 853–858. Newmark, W.D. (1991). Tropical forest fragmentation and the local extinction of understory birds in the eastern Usambara Mountains, Tanzania. Conserv. Biol. 5, 67–78. Patterson, B.D. (1987). The principle of nested subsets and its implications for biological conservation. Conserv. Biol. 1, 323–334. Patterson, B.D. & Atmar, W. (2000). Analyzing species composition in fragments. In Isolated vertebrate communities in the tropics: Proc. 4th Int. Symp: 1–16. Rheinwald, G. (Ed.). Bonn Zool. Monogr. 46, Bonn. Pimm, S.L., Jones, H.L. & Diamond, J. (1988). On the risk of extinction. Am. Naturalist 132, 757–785. Purvis, A. & Hector, A. (2000). Getting the measure of biodiversity. Nature 405, 212–219.

500

K. J. Feeley et al.

Robinson, W.D. (1999). Long-term changes in avifauna of Barro Colorado Island, Panama, a tropical forest isolate. Conserv. Biol. 13, 85–97. Sæther, B.-E., Engen, S., Møller, A.P., Visser, M.E., Matthysen, E., Fiedler, W., Lambrechts, M.M., Becker, P.H., Brommer, J.E., Dickinson, J., Feu, C.D., Gehlbach, F.R., Merila, J., Rendell, W., Robertson, R.J., Thomson, D. & Torok, J. (2005). Time to extinction of bird populations. Ecology 86, 693–700. Sfenthourakis, S., Giokas, S. & Tzanatos, E. (2004). From sampling stations to archipelagos: investigating aspects of the assemblage of insular biota. Global Ecol. Biogeogr. 13, 23–35. Simberloff, D. (1994). Habitat fragmentation and population extinction of birds. Ibis 137, 105–111. Skole, D. & Tucker, C. (1993). Tropical deforestation and habitat fragmentation in the Amazon: satellite data from 1978 to 1988. Science 260, 1905–1910. Sodhi, N.S., Lee, T.M., Koh, L.P. & Prawiradilaga, D.M. (2006). Long-term avifaunal impoverishment in an isolated tropical woodlot. Conserv. Biol. 20, 772–779. Sokal, R.R. & Rohlf, F.J. (1995). Biometry. 3rd edn. New York: W.H. Freeman and Company. Soule`, M.S., Allison, A.C. & Bolger, D.T. (1992). The effects of habitat fragmentation on chaparral plants and vertebrates. Oikos 63, 39–47. Stiles, F.G. & Skutch, A. (1989). A guide to the birds of Costa Rica. New York: Cornell University Press. Stotz, D.F., Fitzpatrick, J.W., Iii, T.A.P. & Moskovits, D.K. (1996). Neotropical birds: ecology and conservation. Chicago: The University of Chicago Press. Stouffer, P.C., Bierregaard, R.O., Strong, C. & Lovejoy, T.E. (2006). Long-term landscape change and bird abundance in Amazonian rainforest fragments. Conserv. Biol. 20, 1212–1223. Terborgh, J. (1974). Preservation of natural diversity: the problem of extinction prone species. Bioscience 24, 715–722. Terborgh, J., Lopez, L. & Tello, J. (1997a). Bird communities in transition: the Lago Guri islands. Ecology 78, 1494–1501. Terborgh, J., Lopez, L., Tello, J., Yu, D. & Bruni, A.R. (1997b). Transitory states in relaxing ecosystems of land bridge islands. In Tropical forest remnants: 256–274. Laurence, W.F. & Bierregaard, R.O. (Eds). Chicago: University of Chicago Press. Terborgh, J. & Winter, B. (1980). Some causes of extinction. In Conservation Biology: an evolutionary– ecological perspective: 199–133. Soule`, M.S. & Wilcox, B.A. (Eds). Sunderland, MA: Siquer Associates Inc. Thiollay, J.M. (1997). Disturbance, selective logging and bird diversity: a Neotropical forest study. Biodivers. Conserv. 6, 1155–1173. Thiollay, J.M. (1999). Response of avian community to rainforest degradation. Biodivers. Conserv. 8, 513–534.


K. J. Feeley et al.

Turner, I.M. (1996). Species loss in fragments of tropical rain forest: a review of the evidence. J. Appl. Ecol. 33, 200–209. Ulrich, W. & Gotelli, N.J. (2007). Null model analysis of species nestedness patterns. Ecology 88, 1824–1831. Van Houtan, K.S., Pimm, S.L., Halley, J.M., Bierregaard, R.O. & Lovejoy, T.E. (2007). Dispersal of Amazonian birds in continuous and fragmented forest. Ecol. Lett. 10, 219–229. Wade, T.G., Ritters, K.H., Wickham, J.D. & Jones, K.B. (2003). Distribution and causes of global forest fragmentation. Conserv. Ecol. 7, 7. Walter, H.S. (2004). The mismeasure of islands: implications for biogeographic theory and the conservation of nature. J. Biogeogr. 31, 1–21. Warburton, N.H. (1997). Structure and conservation of forest avifauna in isolated rainforest remnants in tropical Australia. In Tropical forest remnants: 190–208. Laurence, W.F. & Bierregaard, R.O. (Eds). Chicago: University of Chicago Press.


Willis, E.O. (1974). Population and local extinction of birds on Barro Colorado Island, Panama. Ecological Monogr. 44, 153–169.

Supplementary Material The following supplementary material is available for this article online: Appendix S1 Characteristics of the 41 forest interior bird species inhabiting study islands in Lago Guri, Venezuela. This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/ j.1469-1795.2007.00140.x (This link will take you to the article abstract). Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.


501