ORNITOLOGIA NEOTROPICAL 16: 231–251, 2005© The Neotropical Ornithological Society

BIRD COMMUNITIES IN ANDEAN PREMONTANE FORESTS OF NORTHWESTERN ARGENTINA

Lucio R. Malizia1,2, Pedro G. Blendinger2,3, M. Eva Alvarez2, Luis O. Rivera4, Natalia Politi4, & Guillermo Nicolossi5

1International Center for Tropical Ecology, Department of Biology, University of Missouri-St. Louis, 8001 Natural Bridge Road, St. Louis, Missouri 63121, USA.

2Laboratorio de Investigaciones Ecológicas de las Yungas, Universidad Nacional de Tucumán, C.C. 34, (4107) Yerba Buena, Tucumán, Argentina.

3E-mail: blendinger@birdecology.com.ar 4Centro de Zoología Aplicada, Universidad Nacional de Córdoba, C.C. 122, (5000) Córdoba,

Argentina.5Parque Nacional Calilegua, Administración de Parques Nacionales, San Lorenzo s/n, (4514)

Calilegua, Jujuy, Argentina.

Resumen. – Comunidades de aves en selvas pedemontanas andinas del noroeste de Argentina. – Losestudios de aves andinas a la mesoescala (1000’s km2) son escasos, pero podrían ser importantes para plani-ficar estrategias de uso y conservación. Aquí describimos la riqueza, composición y abundancia de comuni-dades de aves de selvas pedemontanas del noroeste de Argentina (c. 2000 km2). Analizamos cambiosestacionales entre el período reproductivo y no-reproductivo, poniendo énfasis en migrantes y gremiostróficos. Además, evaluamos diferencias en la avifauna de áreas planas y de ladera. Muestreamos 12 comu-nidades de aves mediante conteos por puntos, transectas y redes de niebla. Las comunidades de aves cam-biaron espacialmente sin relación con la topografía (i.e., áreas planas o de ladera), y no registramos ningúngrupo de aves que caracterice estos sitios. La composición de especies varió marcadamente entre estacio-nes, con más especies durante la estación reproductiva, aunque la abundancia de aves fue similar entre esta-ciones. Los ensambles de la época reproductiva estuvieron caracterizados por migrantes latitudinales, y losde la época no-reproductiva por migrantes altitudinales. Los gremios de aves insectívoras y de aves graní-voras que forrajean en o encima del dosel caracterizaron la estación reproductiva, mientras que los gremiosfrugívoro-insectívoro e insectívoro que forrajean en todos los estratos del bosque caracterizaron la esta-ción no-reproductiva. De un total de 112 especies registradas, al menos 10 son de interés para su conserva-ción en Argentina, 14 son poco comunes en el Neotrópico y 31 son migrantes, señalando la importanciade la selva pedemontana para la conservación de aves. Gran parte (c. 75%) de la selva pedemontana hadesaparecido o está seriamente degradada. Esta situación debería llevarnos a unir fuerzas entre gobierno,propietarios, ONGs e instituciones académicas para cambiar las políticas de uso y conservación de losrecursos naturales de la región.

Abstract. – Studies of Andean avifaunas at the mesoscale (1000’s km2) are scarce, but they may prove crit-ical for understanding community structure and for conservation planning. We studied species diversity,composition and abundance of bird communities in premontane forests of northwestern Argentina (c.2000 km2). We measured changes between breeding and non-breeding seasons, focusing on migrants andtrophic guilds, and we compared flat and foothill areas. Twelve bird communities were sampled usingpoint counts, mist nets, and line transects. More species were recorded during the breeding than the non-breeding season, in part due to the arrival of migrants, but the abundance of birds did not differ between

231

MALIZIA ET AL.

seasons. We registered more individuals and species of migratory birds during the breeding season. Latitu-dinal and elevational migrants characterized breeding and non-breeding assemblages, respectively. Wefound no significant differences in bird species richness, composition or abundance in relation to foresttype (i.e., flat vs foothill), in part due to the patchy distributions of birds across the entire study area. Dif-ferent feeding guilds characterized each season: insectivorous and granivorous species that foraged in orabove the canopy were common during the breeding period, while frugivores-insectivores and insectivoresthat feed at all forest strata characterized the non-breeding period. Out of 112 species registered, at least 10are of conservation concern for Argentina, 14 are uncommon in the Neotropics, and 31 are migrants,highlighting the role of premontane forests for regional conservation. Most (c. 75%) premontane forestshave already disappeared or are largely degraded. This situation calls for a rapid shift in regional policies,only possible if government, landowners, NGO’s and academic institutions can share the common goal ofhabitat protection. Accepted 1 April 2005.

Key words: Andes, Argentina, feeding guild, montane forest, seasonal variation, species composition,species turnover.

INTRODUCTION nally occupied one third (c. 1 million ha) of

Tropical Andean forests contain high biodi-versity (Churchill et al. 1995, Brown & Kap-pelle 2001), particularly of birds (Haffer 1987,Fjeldså 1995, Stotz et al. 1996). Bird commu-nity structure (e.g., species diversity, composi-tion and abundance) and spatial distributionhas rarely been studied in Andean forests atthe mesoscale (1000’s km2). High bird diver-sity of Andean forests, difficult access tomany well-preserved forested sites within aregion, and limited economic and humanresources have largely prevented such studies.However, this scale of analysis is relevantbecause it is compatible with current conser-vation strategies (e.g., reserve size), and it mayhelp understand how bird communities areinfluenced by variations in spatial and envi-ronmental conditions of the landscape. In thisstudy, we describe the avifauna of 2000 km2

of premontane Andean forests in Argentina.Premontane forests, located at the base of

forested mountain ranges along the Andes,represent the most endangered forest type ofthe Andean region (Stotz et al. 1996, Brown etal. 2001). This is due to a massive replacementof forested areas by agriculture, coupled witha narrow original distribution of this foresttype. In Argentina, premontane forests origi-

Andean forests, but now only 20–30%remains forested, and what is left is being rap-idly deforested (Grau & Brown 2000). Thedisappearance of premontane forests is ofgreat concern because complete elevationalforest gradients may be needed for maintain-ing species diversity over the long term (Kap-pelle & Brown 2001). This is particularlycritical for bird species which have beenshown to perform seasonal movementsacross many Neotropical montane-forest gra-dients (Terborgh 1974; Loiselle & Blake 1991,1992; Winker et al. 1997). If part of the avi-fauna is temporally present at premontaneforests and seasonally uses other elevationallevels (i.e., elevational migrants), then the dis-appearance of the lower end of the gradientmay also cause disruptions at montane areas(Stiles 1988, Blake & Loiselle 2000). Otherspecies potentially affected by the disappear-ance of premontane forests are latitudinal andshort-distance migrants, species withrestricted distribution, species with lowregional abundance, and species of specialconservation concern such as game birds orpets. For example, the Military Macaw (Aramilitaris), a species of global conservation con-cern (CITES, Appendix 1), disappeared fromAndean premontane forests of Argentina,

232

BIRDS OF ANDEAN PREMONTANE FORESTS

and several others, such as the Toco Toucan(Ramphastos toco) and the Giant Antshrike (Ba-tara cinerea), already reduced their latitudinal

range by ≥ 150 km (pers. observ.). Thus, manyspecies may become regionally extinct as pre-montane forests continue to vanish.

FIG. 1. Map of the Upper Bermejo River Basin (Salta and Jujuy Provinces, northwestern Argentina),showing study sites and environmental units. Study sites: (1) Finca Abra Grande (flat, 430 m, 23º06’S,64º25’W), (2) Finca El Oculto (foothill, 600 m, 23º05’S, 64º32’W), (3) Finca El Tabacal (foothill, 550 m,23º13’S, 64º28’W), (4) Colonia Santa Rosa (flat, 500 m, 23º23’S, 64º32’W), (5) Finca Urundel (flat, 390,23º29’S, 64º25’W), (6) Finca Urundel (foothill, 500 m, 23º28’S, 64º29’W), (7) Finca Urundel (foothill, 660m, 23º26’S, 64º33’W), (8) Finca Yuchán (foothill, 730 m, 23º52’S, 64º55’W), (9) Colonia Guayacán (foot-hill, 630 m, 24º00’S, 64º53’W). Map modified from Grau & Brown (2000).

233

MALIZIA ET AL.

The specific objectives of this study werefourfold. First, we described bird communi-ties in terms of seasonal changes in speciescomposition and abundance. We expectedbird communities composed by large propor-tion of migrants, considering the marked sea-sonal changes in climatic conditions andforest phenology that characterize premon-tane forests (Malizia 2001), and the highmobility of birds shown for other mountainareas (Loiselle & Blake 1991, Winker et al.1997). Second, we analyzed changes in speciescomposition and abundance in relation totopography, elevation, and distance amongsites. As a first approximation, premontaneforests in Argentina can be divided into flatand foothill forests based on topography andelevation. This division may have biologicalrelevance because most avian communitieschange along elevational gradients (Terborgh1977, Young et al. 1998, Blake & Loiselle2000, Kessler et al. 2001), even for short ele-vational ranges (c. 300 m) (Poulin et al. 1993,Patterson et al. 1998). We expected high ratesof species turnover between flat and foothillforests and low rates among sites within thesame forest type. Third, we examined birdcommunity patterns considering feedingguilds in order to highlight functional aspectsand to facilitate comparisons with studiesconducted in other premontane areas, wherespecies composition may differ. Finally, wedetermined the importance of premontaneforests to regionally threatened species,uncommon species, and migrant birds.

MATERIALS AND METHODS

Study area. This study was conducted in pre-montane forests of the Upper Bermejo RiverBasin, provinces of Jujuy and Salta, north-western Argentina (23°–24° S, Fig. 1). Pre-montane forest is the lowest elevation (400–700 m) forest type of Andean montane for-ests (400–2500 m) and is known as Selva

Tucumano-Boliviana (Cabrera 1976). Thisforest type ranges from Santa Cruz depart-ment in Bolivia (19° S) south to Tucumánprovince in Argentina (28° S). From an orni-thological perspective, the study area lies inthe Santa Cruz-Tucumán subregion, of theCentral Andes region (Stotz et al. 1996). Adetailed description of premontane forestscan be found in Prado (1995) and Brown et al.(2001). As previously mentioned, premontaneforests can be subdivided into “flat forests”,located at the base of the Andes over rela-tively flat areas at c. 400–500 m, and “foothillforests”, located along the eastern slopes ofthe mountains, between c. 500–700 m.

Climate in the study area is highly sea-sonal. Dry and cool weather (winter) occursfrom May to September, while wet and hotconditions (summer) last from November toMarch. Total annual rainfall averages c. 1000mm (Bianchi & Yañez 1992), and is concen-trated (75–80%) during the summer (Hunz-inger 1995). Mean annual temperatureaverages 21.5°C (Marmol 1995).

Fieldwork was conducted between April2000 and October 2001. April to June sam-ples correspond to the non-breeding season,whereas October to January samples corre-spond to the breeding season. We samplednine premontane-forest sites (Fig. 1); six siteswere sampled once (sites 3–4 and 6–9) andthree sites were sampled twice (sites 1, 2 and5), totaling 12 samples, six during each season(sites 1–6 during the non-breeding season,and sites 1, 2, 5, 7–9 during the breeding sea-son). In this study we treat the 12 samples asindependent samples, although if strictly wevisited nine sites corresponding to 12 sam-ples, because the three sites visited twice weresampled only once during the breeding sea-son and once during the non-breeding sea-son. Out of the nine sites, five were located infoothills and four in flat areas (Fig. 1).

Bird sampling. We used three sampling meth-

234

BIRDS OF ANDEAN PREMONTANE FORESTS

ods to survey bird communities to overcomelimitations and biases of individual techniques(Bibby et al. 2000). We used point counts dur-ing both the breeding and non-breeding sea-sons, line transects only during the breedingseason, and mist nets only during the non-breeding season. At each site, we sampled theavifauna of a relatively small area (≤ 50 ha).Although differences in sampling methodsused between seasons call for a careful inter-pretation of results, we consider that our sea-sonal comparisons provide useful informationabout bird community structure. We focusedour study on all diurnal species that composelocal communities at premontane forests. Weestablished our study sites in forest-interiorand relatively well-preserved areas. We fol-lowed Remsen et al. (2004) for systematic andtaxonomic arrangements and for Englishcommon names (Appendix 1).

Point counts. We established 12 point-count sta-tions at each site visited during the breedingseason and 10 point-count stations at eachsite visited during the non-breeding season.We used only the first 10 point counts per sitein analyses that required equal sample sizes(i.e., seasonal comparisons). All point countswere established at > 150-m intervals in for-est-interior transects. All birds seen or heardin a 50-m radius were counted for 10 min atevery point (72 points during the breedingseason and 60 points during the non-breedingseason). Each point count was visited onlyonce. All counts were conducted for 2.5 hstarting at sunrise

Line transects. During the breeding season, ateach site, we surveyed 10 line transects dis-tributed as uniformly as possible over a 50-haforest-interior plot located at least 200 mfrom forest edges and roads, and from othertransects. At each transect we recorded allspecies seen or heard without restriction ofdistance from the observer. Each transect was

walked at a slow and steady pace during a 30-min period (totaling 5 h per site and 30 h forthe breeding season), without restriction oftransect length. Since our objective was todetect as many bird species as possible, linetransects were surveyed at different times dur-ing the day.

Mist nets. We used ground-level mist nets (12m x 2.8 m, 36-mm mesh) during the non-breeding season, when some species are lesslikely to be singing and therefore to bedetected by point-counts or transect methods.At each site, 15 mist nets were placed at > 40-m intervals along forest-interior transects.Mist nets were operated during three consecu-tive days at each site; the first two days forabout 10 h/day and the third day for about 5h, always starting right after dawn. All birdswere banded with numbered aluminum legbands and released close (~ 150 m) to mist-net locations to prevent disturbance at thenets. Feeding guilds. We assigned bird species to feed-ing guilds (Appendix 1) based on a combina-tion of three attributes: 1) main food type inthe diet (carnivore, frugivore-insectivore,granivore, granivore-insectivore, insectivore,nectarivore, and omnivore), 2) forest stratawhere birds forage (ground, understory, can-opy, all forest strata, above the canopy, andforest edge), and 3) primary foraging substratefrom which food is obtained (air, air-foliage,bark, foliage, and ground) (Blake & Rougès1997, Giannini 1999, Rougès & Blake 2001, L.R. Malizia & P. G. Blendinger pers. observ.). Data analyses. We used Bray-Curtis ordinationanalyses to graphically compare similarity inspecies composition among communities(PC-ORD version 4.01, McCune & Mefford1999). We used the variance-regressionapproach for endpoint selection and theSørensen-distance measure for all analyses.

235

MALIZIA ET AL.

Bray-Curtis ordinations were based on pres-ence-absence data of species recorded with allsampling methods. Ordination results basedon species abundance are not presentedbecause they were very similar to thoseobtained with presence-absence data. Weused number of species per guild to assesssimilarity in guild composition among com-munities; the use of presence-absence data ofguilds per site was precluded because mostguilds were present at all sites.

We used number of species recorded ateach site (combining all sampling methods) toanalyze differences in species richnessbetween seasons and forest types. Addition-ally, we estimated total species richness persite based on abundance data from pointcounts using EstimateS MMMean methodwith 1000 randomizations (Colwell 2000).These estimations tended to underestimatespecies richness per site when compared withtotal number of species recorded consideringall methods (point counts, transects and mistnets), but provided similar results in compari-sons between breeding and non-breedingperiods, and between flat and foothill forests.Therefore, we do not include here resultsbased on estimations of species richness.

We analyzed bird-species abundance con-sidereding the number of individualsrecorded with point counts (i.e., not consider-ing mist nets or transects) because they wereused at all sites. We assigned birds to catego-ries of relative abundance for the mesoscale(premontane forest) and for the Neotropicsbased on our data and on Parker et al. (1996),respectively (Appendix 1). We used one-wayanalysis of variance (ANOVA) to test for dif-ferences in mean bird abundance per site (allspecies considered together) within breedingand non-breeding communities. Differencesin mean bird abundance and species richnessbetween seasons and forest types were ana-lyzed with two-sample t-tests and Chi-squaretests. We analyzed the relationship between

species abundance and frequency (i.e., num-ber of sites occupied), and resident andmigrant bird abundance with correlations.Prior to the analyses we tested the assump-tions of normality (Kolmogorov-Smirnovtest) and homogeneity of variances (Levenetest) and used non-parametric tests wheneverthe assumptions were not met.

Species turnover was estimated using theSørensen coefficient of dissimilarity (Vellend2001), and was calculated as SD = 1 – (c/α),where c is the number of species shared bytwo communities and α is average number ofspecies in each community. Values of SDrange from 0, indicating that every speciesfound in one community is also found in theother one, to 1, indicating no species in com-mon. We analyzed the relationship betweenspecies turnover and geographic distance withMantel’s tests using 9999 permutations toestimate r and P values, because regular r2 andP values are not recommended due to thenon-independence of the data (Legendre &Legendre 1998). Mantel’s tests were per-formed using Passage version 1.1 (Rosenberg2003).

RESULTS

Species richness. We recorded 112 species, rang-ing from 37 to 57 species per site (Appendix1). Study sites were forest-interior and rela-tively well-preserved areas; therefore, wemainly recorded forest-interior birds. How-ever, we also recorded several non-forest spe-cies [e.g., Striped Cuckoo (Tapera naevia),Great Kiskadee (Pitangus sulphuratus), Rufous-collared Sparrow (Zonotrichia capensis), Red-crested Finch (Coryphospingus cucullatus), andShiny Cowbird (Molothrus bonariensis)], andseveral species that forage above the forestcanopy [i.e., White-collared Swift (Streptoprocnezonaris), Sick’s Swift (Chaetura meridionalis),Swallow-tailed Kite (Elanoides forficatus),Plumbeous Kite (Ictinia plumbea), Black Vul-

236

BIRDS OF ANDEAN PREMONTANE FORESTS

ture (Coragyps atratus), and Turkey Vulture(Cathartes aura)]. Mean number of species persite differed between seasons (two-sample t-test for unequal variances = 2.82, df = 6.8, P= 0.03), with more species during the breed-ing (mean = 49.8, range: 37–57) than the non-breeding period (41.0, 37–47). Accordingly,considering all sites combined per season, wefound more species in premontane forestsduring the breeding season (93 vs 66 species;χ2 = 4.25, df = 1, P = 0.03).

We found no differences in mean speciesrichness per site between flat (43.8, 37–51)and foothill (46.6, 37–57) forests (two-samplet-test for equal variances = –0.67, df = 10, P= 0.52). Considering all sites combined perforest type, we found similar numbers of birdspecies in flat (N = 89) and foothill forests (N= 96).

Seasonal changes in species composition. As previ-ously mentioned, premontane-forest bird

communities showed important seasonal dif-ferences in species composition. Althoughmany species (47) were present both duringthe breeding and non-breeding seasons, ahigh proportion of species (58%) were exclu-sive to one season, with 46 species exclusiveto the breeding season and 19 to the non-breeding season. Migrant species represented27% (31 species) of the premontane forestavifauna, with 19 latitudinal and 13 elevationalmigrant species (Appendix 1). From this, SlatyThrush (Turdus nigriceps) performs latitudinaland elevational movements and therefore wasincluded in both migrant categories. Twentymigrant species were registered during thebreeding season (18 latitudinal and 3 eleva-tional migrants) and 18 species during thenon-breeding season (6 latitudinal and 13 ele-vational migrants).

The first two axes of a Bray-Curtis ordina-tion largely separated sites in relation to sam-pling season (Fig. 2). The ordination

FIG. 2. Bray-Curtis ordination of bird communities based on presence/absence of species. Species moststrongly correlated with the first two axes are indicated. *Elevational migrants, **Latitudinal migrants.

237

MALIZIA ET AL.

accounted for 63.2% of the variation in spe-cies composition across sites (axis I: 39.7%,axis II: 23.5%). Study sites were mainlygrouped on the basis of species exclusive toeach season, which included several migrantspecies (Fig. 2). Elevational migrants charac-terized the non-breeding season, while latitu-dinal migrants characterized the breedingseason (Fig. 2).

Spatial distribution of species. Only 10 species outof 112 were present at all sites: two non-pas-seriforms [Tataupa Tinamou (Crypturellustataupa) and Green-cheeked Parakeet (Pyrrhuramolinae)] and eight passeriforms [VariableAntshrike (Thamnophilus caerulescens), Ochre-cheeked Spinetail (Synallaxis scutata), Oliva-ceous Woodcreeper (Sittasomus griseicapillus),Plush-crested Jay (Cyanocorax chrysops),Rufous-bellied Thrush (Turdus rufiventris), Saf-fron-billed Sparrow (Arremon flavirostris),Tropical Parula (Parula pitiayumi), and Two-banded Warbler (Basileuterus bivittatus)].Twenty-two bird species were recorded atonly one site, including one species of ibis,vulture, guan and hummingbird, two speciesof cuckoos, three of diurnal raptors, and 13passeriform species (Appendix 1).

Most species (72 out of 112) were presentin both flat and foothill areas. Seventeen wereregistered only in flat sites, and 23 were exclu-sive to foothill sites. However, more than halfof these species (22 out of 40) were rare (i.e.,

present at only one site), and therefore wereinadequate to characterize flat or foothill for-ests.

Species turnover. Mean dissimilarity in speciescomposition across all site pairs was 0.41,indicating intermediate levels of species turn-over among bird communities. As expected,mean dissimilarity between sites sampled dur-ing the same season was lower (0.31) than forsites sampled during different seasons (0.50)(two-sample t-test = 11.9, df, corrected toaccount for non-independence of pairs, = 12,P < 0.001), indicating the temporal turnoverin species composition across seasons. Com-paring seasons, mean dissimilarity was higherduring the breeding (0.36) than the non-breeding (0.26) period (two-sample t- test forequal variances = 4.56, df = 28, P < 0.001),suggesting more homogeneous spatial distri-butions of birds during the non-breedingperiod.

Turnover in species composition over theentire study area was not correlated with geo-graphic distance between sites (Mantel’s test, r= 0.08, P = 0.29). Considering each seasonseparately, species turnover and distance wereagain unrelated (Mantel’s tests: breeding sea-son, r = 0.02, P = 0.43; non-breeding season,r = 0.23, P = 0.22). Mean dissimilarity amongflat-forest sites (0.34) was slightly higher thanamong foothill sites (0.29). Mean dissimilar-ity between flat and foothill sites (0.31) was

TABLE 1. Analyses of variance of bird abundance per point count in breeding and non-breeding assem-blages. Sites with distinct superscript letters showed significant differences (Tukey test of multiple compar-isons, P < 0.05). See Figure 1 for site references.

Seasons SitesBreeding F5, 66 = 10.95, P < 0.001 (12 counts/site)Non-breeding F5, 54 = 1.46, P = 0.22 (10 counts/site)

Site 1(flat)9.7ab

Site 1(flat)10

Site 5(flat)16.9bc

Site 4(flat)13

Site 2(foothill)

13.2ab

Site 5(flat)18.1

Site 7(foothill)

14.0b

Site 2(foothill)

16.1

Site 8(foothill)

6.7a

Site 3(foothill)

12.1

Site 9(foothill)

23.3c

Site 6(foothill)

12.3

238

BIRDS OF ANDEAN PREMONTANE FORESTS

intermediate to the previous values, instead ofhigher as we expected from comparisons ofdifferent forest types.

Species abundance. Bird abundance was esti-mated at 132 point counts in which we regis-tered 1819 individuals corresponding to 88species (Appendix 1). Total number of indi-viduals per species was correlated with thenumber of sites occupied, i.e., species thatwere more abundant tended to be present inmore communities (rS = 0.46, N = 88, P <

0.001).Mean bird abundance per point count did

not differ between seasons (two-sample t-test= 0.02, df = 10, P = 0.32) or between foresttypes (two-sample t-test = 0.50, df = 10, P =0.21). During the breeding season, mean birdabundance per point count differed amongsites (Table 1), due both to the presence ofmigrant and resident birds. Latitudinalmigrants represented in average 26.7% of thebirds recorded per site during the breedingseason, ranging among sites from an average

FIG. 3. Bray-Curtis ordinations of bird communities based on number of species per feeding guild. Circles’size is proportional to the number of species per feeding guild present at each community.

239

MALIZIA ET AL.

of 2.2 to 7.9 individuals per point count. Thenumber of resident birds also varied markedlyamong sites (F5, 66 = 10.27, P < 0.001), butmean abundance of resident and migrantsspecies were not correlated (r = 0.70, N = 6,P = 0.11). In contrast to the breeding season,there were no differences among sites inmean bird abundance per point count duringthe non-breeding season (Table 1). Eleva-tional migrants represented in average 14.9%of the birds recorded per site during the non-breeding season, ranging among sites from anaverage of 0.8 to 4.1 individuals per pointcount.

Feeding guilds. Combining data of primary for-aging substrata, foraging forest strata, anddiet, we classified the 112 bird speciesrecorded in premontane forests into 35 feed-ing guilds (Appendix 1). Due to our detailedguild classification, most guilds included oneor a few species, and only four guilds includedseven or more species: canopy insectivoresthat capture prey in the foliage (11), under-story insectivores that capture prey in the foli-age (9), canopy insectivores that capture preyboth in the foliage and the air (7), and forest-edge granivore-insectivores that forage in thefoliage (7).

The results of a Bray-Curtis ordinationbased on feeding guilds explained 52.2% ofthe variation and separated breeding fromnon-breeding communities (axis I: 36%, axisII: 16%). Six feeding guilds, which included34 species (13 migrants), explained most ofthe differences among seasons (χ2 = 25.4, df= 5, P < 0.001; Fig. 3). The breeding seasonwas characterized by bird species that foragein or above the canopy. They included i)medium and high forest-strata insectivores,such as Streaked Flycatcher (Myiodynastes macu-latus), Green-backed and White-wingedbecards (Pachyramphus viridis and P. poly-chopterus, respectively), Black-capped Antwren(Herpsilochmus atricapillus), Red-eye Vireo

(Vireo olivaceus), and Tropical Parula (Fig. 3a);ii) canopy granivores, such as Golden-collaredMacaw (Propyrrhura auricollis), White-eyed Par-akeet (Aratinga leucophthalma), Green-cheekedParakeet (Pyrrhura molinae), Scally-headed Par-rot (Pionus maximiliani), Blue-fronted Parrot(Amazona aestiva), and Pale-vented Pigeon(Columba cayennensis) (Fig. 3b); iii) insectivoresthat forage above the canopy, such as White-collared and Sick’s Swifts, Swallow-tailed andPlumbeous Kites, and Tropical Kingbird (Tyr-annus melancholicus) (Fig. 3c); and iv) canopyinsectivores that forage on tree bark, such asGolden-olive and Cream-backed woodpeck-ers (Piculus rubiginosus and Campephilus leuco-pogon, respectively), Stripe-crowned Spinetail(Cranioleuca pyrrhopia), and Streaked Xenops(Xenops rutilans) (Fig. 3d). In contrast, the non-breeding season was characterized by birdspecies that used all vegetation strata, fromunderstory to canopy. They included i) frugi-vore-insectivores, such as Dusky-legged Guan(Penelope obscura), Slaty, Rufous-bellied andCreamy-bellied thrushes (Turdus nigriceps, T.rufiventris and T. amaurochalinus, respectively),Common Bush-tanager (Chlorospingus ophthal-micus), and Sayaca Tanager (Thraupis sayaca)(Fig. 3e); and ii) insectivores, such as Mottled-cheeked Tyrannulet (Phylloscartes ventralis),Buff-browed Foliage-Gleaner (Syndactyla rufo-superciliata), Brown-capped Redstart (Myioborusbrunniceps), and Rusty Flowerpiercer (Diglossasittoides) (Fig. 3f).

Common vs uncommon species. Using the abun-dance assessments for Neotropical birds pub-lished by Parker et al. (1996), we recorded 50species as common, 48 as fairly common and14 as uncommon (Appendix 1). At themesoscale, considering all premontane forestssites together, we found 23 uncommon spe-cies (i.e., species present only at one site perseasons and with ≤ 2 individuals recorded atpoint counts), most (16) of them during thebreeding season (Appendix 1). The Hepatic

240

BIRDS OF ANDEAN PREMONTANE FORESTS

Tanager (Piranga flava) was the only uncom-mon species recorded during both seasons,although the number of uncommon speciespresent year round was probably underesti-mated, since uncommon species are by defini-tion hard to detect and could be lost duringshort-term sampling periods. Most (79%) lati-tudinal and elevational migrants were com-mon species, suggesting that premontaneforest harbored large populations of thesemigrants, and reinforcing the importance ofpremontane forests for their conservation.Uncommon migrants included two eleva-tional migrants specialized in nectar [PlanaltoHermit (Phaetornis pretrei) and Rusty Flower-piercer], a resource patchily distributed andscarce during the non-breeding season. Theother uncommon migrant species were insec-tivorous latitudinal migrants mainly relatedwith more open environments [Dark-billedCuckoo (Coccyzus melacoryphus), PlumbeousKite, White-winged Black-Tyrant (Knipolegusaterrimus), Tropical Kingbird, and VariegatedFlycatcher (Empidonomus varius)].

DISCUSSION

To our knowledge, this is the first study toreport abundance, composition and diversityof bird communities at premontane forests,which cover 2.5 million ha of the Santa Cruz-Tucumán subregion. Despite the limitationsof this study, that include modest sample sizeat each site, it may well represent the mostdetailed report currently available to guideregional conservation strategies.

This study showed important seasonalchanges in bird-species richness and composi-tion, in part due to the arrival of migrants.More speciose communities were recordedduring the breeding than the non-breedingseason, although total bird abundance did notvary between periods. However, we registeredmore individuals and species of migratorybirds during the breeding season. Breeding

assemblages were characterized by the pres-ence of latitudinal migrants, e.g., Sick’s Swift,Pale-vented Pigeon, Euler’s Flycatcher (La-throtriccus euleri), three Myiarchus flycatchers,Streaked Flycatcher, Red-eye Vireo, andSwainson’s Thrush (Catharus ustulatus), whilenon-breeding assemblages harbored eleva-tional migrants, e.g., Red-tailed Comet (Sapphosparganura), Buff-banded Tyrannulet (Meco-cerculus hellmayri), Andean Tyrant (Knipolegus sig-natus), Brown-capped Redstart, CommonBush-Tanager, and Black-backed Grosbeak(Pheucticus aureoventris).

We found no significant differences inbird species richness, composition or abun-dance in relation to elevation and topographiclocation (i.e., flat vs foothill) within premon-tane forest. Flat-forest sites were not charac-terized by any particular set of species, in partdue to the patchy distributions of birds acrossthe entire study area. If this pattern holdsafter detailed studies, both an encouragingand a warning message should be stated. Theencouraging message is that no bird speciesseems to depend exclusively on flat premon-tane forest, the most endangered habitat ofthis part of the Andean region. The warningmessage is that we know little about the inter-dependence of flat and foothill premontaneforests in terms of diversity and functionalaspects. If complete elevational gradients areneeded to maintain the diversity of Andeanforests (see below), then the rapid destructionof flat forests may hamper all efforts of habi-tat conservation in the region.

Species turnover between communitiesranged from intermediate (0.50, different sea-sons) to low values (0.26, within the non-breeding season) in absolute terms, but wasrelatively high considering that all sites wheresampled within limited latitudinal (100 km)and elevational ranges (390–730 m). Ourresults are similar to those of Poulsen &Krabbe (1998) for other Andean forests,where they interpreted low dissimilarity values

241

MALIZIA ET AL.

(mean = 0.33) as evidence of moderately highdifferences in bird species composition. Inour study, geographical distance did notexplain dissimilarity in species compositionbetween communities, as suggested by non-significant correlations between inter-site dis-tance and species turnover. Environmentalfactors, such as forest type, neither explainsuch dissimilarities. We do not think the dis-similarity levels found are a consequence ofour sampling methods. This idea is supportedby extra observations conducted during thebreeding season, when we made an extraeffort to register all birds present at each site,besides the systematic methods presentedhere. This initiative showed that our samplingmethods detected on average 88% of all birdspresented at each site. This result supportsthe notion that we sampled a high percentageof the species present at each site and that thedissimilarity levels found may be caused bypatchily-distributed species at the mesoscale,a pattern also reported for other forest birdcommunities (Gaston & Blackburn 2000).The causes promoting patchy distribution ofbirds are unclear; it has been proposed thatthey are largely species specific and that thefloristic structure and composition of thehabitat may play an important role (Bierre-gaard & Stouffer 1997, Renjifo 2001).Although we sampled relatively well-pre-served sites (no fragments were considered,see Fig. 1), the characteristics of the matrixaround each site varied largely due to forestdegradation and transformation. Perhapsenvironmental variables acting at small (e.g.,0.1 km2) and landscape scales (e.g., 100 km2)may have been responsible for the amount ofspecies turnover found.

The spatial distribution of feeding guildsdid not show any specific pattern, but thetemporal distribution did reveal interestingseasonal differences. The breeding period wascharacterized by insectivorous and granivo-rous species that primarily forage in or above

the canopy, while the non-breeding seasonwas characterized by frugivore-insectivoresand insectivores that move across all foreststrata to feed. Premontane forests are highlyseasonal in terms of leaf phenology, and mostcanopy trees lose their leaves during the dryand cold period. This may explain the virtualabsence of feeding guilds that primarily for-age in the canopy and the prevalence of guildsthat utilize all strata during the non-breedingperiod.

Regional distribution and abundance ofspecies can strongly influence the composi-tion and structure of local communities. Inpremontane forest, species abundance andfrequency (i.e., sites occupied) were corre-lated, indicating that regionally abundant spe-cies also tended to occupy many localcommunities. This seems to be a general rulefor temperate and subtropical areas (e.g.,Blackburn & Gaston 2001), but probably notfor tropical areas where many narrowly dis-tributed species may show high abundancewithin their small ranges. In our case, all butone (Ochre-cheeked Spinetail) common spe-cies (i.e., 36 species with > 0.5% of totalcounts and present at more than 1 site perseason) at premontane forests are also com-mon or fairly common in the Neotropics(sensu Stotz et al. 1996). Most (22/36) com-mon species at premontane forests also tendto occupy large areas in South America (> 1/4 of the continent), and only some (6/36)have relative small (< 1/8) continental distri-butions [Green-cheeked Parakeet, Ochre-faced Tody-Flycatcher (Poecilotriccus plum-beiceps), Azara’s Spinetail (Synallaxis azarae),Two-banded Warbler (Basileuterus bivittatus),Common Bush-Tanager, and Black-backedGrosbeak] (Ridgely & Tudor 1989, 1994; delHoyo et al. 1992, 1997, 1999, 2001).

In this study, we found that at least 10species of birds inhabiting premontane for-ests may be threatened in one or several waysfor Argentina. The Toco Toucan, Giant Ant-

242

BIRDS OF ANDEAN PREMONTANE FORESTS

shrike (P. Capllonch com. pers.) and Hook-billed Kite (Blendinger et al. 2004) contractedtheir distribution range during the lastdecades. The Dusky-legged Guan is heavilyhunted for food, while the Blue-fronted Par-rot and Black-backed Grosbeak are widelyused as pets in legal and illegal markets. Thepremontane forest may represent an impor-tant habitat for elevational migrant speciesrestricted to montane forest during the breed-ing season, such as the Andean Tyrant and theBuff-banded Tyrannulet. Several speciesrestricted in Argentina to the Santa Cruz-Tucumán subregion are residents of premon-tane forests, such as the Dot-fronted Wood-pecker (Veniliornis frontalis) and the White-throated Antpitta (Grallaria albigula). More-over, the White-throated Antpitta, subspeciescinereiventris, is endemic to the study region(Jujuy and Salta) in Argentina (Chebez 1999).This list highlights the need for premontaneforest to be included in a national conserva-tion strategy.

In this study we recorded 112 bird speciesat Andean premontane forests, representing c.11% of Argentina’s bird diversity. From this,10 species are of conservation concern for thecountry, 14 are uncommon for the Neotro-pics as a whole, and 31 are migrants, either lat-itudinal (19) or elevational (13). As previouslymentioned, conservation concerns haveincreased in the last decades with accumulatedevidence suggesting the importance of com-plete elevational forest gradients for maintain-ing bird species diversity at the long term inthe Neotropics (Terborgh 1974, Stiles 1988;Loiselle & Blake 1991, 1992; Winker et al.1997, Blake & Loiselle 2000). This informa-tion, coupled with premontane forest’s endan-gered status (Stotz et al. 1996, Brown et al.2001), call for the protection of this habitat.As a future challenge, we need to betterunderstand the role of premontane forest atthe regional scale as a key component foravian migrant movements, both to upper

montane forests and to lowland forests as theChaco thorny woodland.

ACKNOWLEDGMENTS

LRM is in debt with M. Echenique and C.Aguirre for fieldwork assistance, with J. G.Blake and A. D. Brown for general advice andLaboratorio de Investigaciones Ecológicas delas Yungas (LIEY) for logistical support. Wethank A. Caro for the design of Figure 1. Weacknowledge comments of J. Blake, S. K.Robinson, J. Perez, B. Loiselle, M. Rougès, J.Lill, I. Jiménez, L. Lohmann, D. Cadena, L.Arroyo, L. Griz, K. Holbrook, J. Uribe, G.Servat, and F. Wolff, which improved differ-ent versions of the manuscript. We acknowl-edge the authorities of Secretaría de MedioAmbiente of Provincia de Salta and Adminis-tración de Parques Nacionales for work per-mission, and Finca owners for allowing us towork in their properties. This study wasfunded by the William Belton Program of theAmerican Bird Conservancy and U.S. Fishand Wildlife Service, by Programa ProYungas(LIEY and Fundación Vida Silvestre Argen-tina), by the International Center for TropicalEcology (University of Missouri-St. Louis)through a John Denver Memorial Scholarship,and by Fundación Capacitar (Salta, Argen-tina).

REFERENCES

Bianchi, A. R., & C. Yañez. 1992. Las precipita-ciones en el noroeste argentino. InstitutoNacional de Tecnología Agropecuaria, Salta,Argentina.

Bibby, C. J., N. D. Burgess, D. A. Hill, & S. Mustoe.2000. Bird census techniques. Academic Press,London, UK.

Bierregaard, R. O., & P. C. Stouffer. 1997. Under-story birds and habitat mosaics in Amazonianrainforests. Pp. 138–155 in Laurance W. F., & R.O. Bierregaard (eds.). Tropical forest remnants:ecology, management, and conservation of

243

MALIZIA ET AL.

fragmented communities. Univ. of ChicagoPress, Chicago, Illinois.

Blackburn, T. M., & K. J. Gaston. 2001. Localavian assemblages as random draws fromregional pools. Ecography 24: 50–58.

Blake, J. G., & B. A. Loiselle. 2000. Diversity ofbirds along an elevational gradient in the Cor-dillera Central, Costa Rica. Auk 117: 663–686.

Blake, J. G., & M. Rougès. 1997. Variation in cap-ture rates of understory birds in El ReyNational Park, northwestern Argentina. Orni-tol. Neotrop. 8: 185–193.

Blendinger, P. G., P. Capllonch, & M. E. Alvarez.2004. Abundance and distribution of raptors inthe Sierra de San Javier Biological Park, north-western Argentina. Ornitol. Neotrop. 15: 501–512.

Brown, A. D., & M. Kappelle. 2001. Introducción alos bosques nublados del Neotrópico: una sín-tesis regional. Pp. 25–40 in Kappelle, M., & A.D. Brown (eds.). Bosques nublados delNeotrópico. Instituto Nacional de Biodiver-sidad, San José, Costa Rica.

Brown, A. D., H. R. Grau, L. R. Malizia, & A.Grau. 2001. Argentina. Pp. 623–659 in Kap-pelle, M., & A. D. Brown (eds.). Bosques nubla-dos del Neotrópico. Instituto Nacional deBiodiversidad, San José, Costa Rica.

Cabrera, A. L. 1976. Regiones fitogeográficasargentinas. Enciclopedia Argentina de Agricul-tura y Jardinería. 2da ed. Fascículo 1, EditorialACME, Buenos Aires, Argentina.

Chebez, J. C. 1999. Los que se van. Especies argen-tinas en peligro. Editorial Albatros, BuenosAires, Argentina.

Churchill, S. P., H. Balslev, E. Forero, & J. L.Luteyn. 1995. Biodiversity and conservation ofNeotropical montane forests. The New YorkBotanical Garden, New York, New York.

Colwell, R. K. 2000. EstimateS: statistical estima-tion of species richness and shared speciesfrom samples, Version 6 (http://vice-roy.eeb.unconn.edu/estimates).

del Hoyo, J., A. Elliott, & J. Sargatal. 1992. Hand-book of the birds of the world. Volume 1:Ostrich to ducks. Lynx Edicions, Barcelona,Spain.

del Hoyo, J., A. Elliott, & J. Sargatal. 1997. Hand-book of the birds of the world. Volume 4:

Sandgrouse to cuckoos. Lynx Edicions, Barce-lona, Spain.

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

del Hoyo, J., A. Elliott, & J. Sargatal. 2001. Hand-book of the birds of the world. Volume 6:Mousebirds to hornbills. Lynx Edicions, Barce-lona, Spain.

Fjeldså, J. 1995. Geographical patterns of neoen-demic and older relict species of Andean forestbirds: the significance of ecologically stableareas. Pp. 89–102 in Churchill, S. P., H. Balslev,E. Forero, & J. L. Luteyn (eds.). Biodiversityand conservation of Neotropical montane for-ests. The New York Botanical Garden, NewYork, New York.

Gaston, K. J., & T. M. Blackburn. 2000. Patternand process in macroecology. Blackwell Sci-ence, Oxford, UK.

Giannini, N. P. 1999. La interacción de aves-mur-ciélagos-plantas, en el sistema de frugívoros ydispersión de semillas en San Javier, Tucumán,Argentina. Tesis Doc., Univ. Nacional deTucumán, Tucumán, Argentina.

Grau, A., & A. D. Brown. 2000. Developmentthreats to biodiversity and opportunities forconservation in the mountain ranges of theUpper Bermejo River Basin, NW Argentinaand SW Bolivia. Ambio 29: 445–450.

Haffer, J. 1987. Biogeography of Neotropical birds.Pp. 105–150 in Whitmore, T. C. G., & T. Prance(eds.). Biogeography and Quaternary history intropical America. Clarenden Press, Oxford,UK.

Hunzinger, H. 1995. La precipitación horizontal:su importancia para el bosque y a nivel decuencas en la Sierra de San Javier, TucumánArgentina. Pp. 53–58 in Brown, A. D., & H. R.Grau (eds.). Investigación, conservación ydesarrollo en selvas subtropicales de montaña.Proyecto de Desarrollo Forestal, Laboratoriode Investigaciones Ecológicas de las Yungas,Tucumán, Argentina.

Kappelle, M., & A. D. Brown. 2001. Bosquesnublados del Neotrópico. Instituto Nacional deBiodiversidad, San José, Costa Rica.

Kessler, M., S. K. Herzog, J. Fjeldså, & K. Bach.

244

BIRDS OF ANDEAN PREMONTANE FORESTS

2001. Species richness and endemism of plantand bird communities along two gradients ofelevation, humidity and land use in the BolivianAndes. Divers. Distrib. 7: 61–77.

Legendre, P., & L. Legendre. 1998. Numerical ecol-ogy. 2nd ed. Elsevier Science, Amsterdam.

Loiselle, B. A., & J. G. Blake. 1991. Temporal varia-tion in birds and fruits along an elevational gra-dient in Costa Rica. Ecology 72: 180–193.

Loiselle, B. A., & J. G. Blake. 1992. Population vari-ation in a tropical bird community. BioScience42: 838–845.

Malizia, L. R. 2001. Seasonal fluctuations of birds,fruits, and flowers in a subtropical forest ofArgentina. Condor 103: 45–61.

Marmol, L. A. 1995. Enriquecimiento forestal deselva degradada en las Yungas de Yuto (Prov.de Jujuy). Pp. 85–92 in Brown, A. D., & H. R.Grau (eds.). Investigación, conservación ydesarrollo en selvas subtropicales de montaña.Proyecto de Desarrollo Forestal, Laboratoriode Investigaciones Ecológicas de las Yungas,Tucumán, Argentina.

McCune, B., & M. J. Mefford. 1999. PC-ORD:Multivariate analysis of ecological data. Version4.01. MjM Software Design, Gleneden Beach,Oregon.

Parker, T. A., III, D. F. Stotz, & J. W. Fitzpatrick.1996. Ecological and distributional databasesfor Neotropical birds. Pp. 118–478 in Stotz, D.F., J. W. Fitzpatrick, T. A. Parker, & D. K. Mosk-ovits ( eds.). Neotropical birds: ecology andconservation. Univ. of Chicago Press, Chicago,Illinois.

Patterson, B. D., D. F. Stotz, S. Solari, J. W. Fitz-patrick, & V. Pacheco. 1998. Contrasting pat-terns of elevational zonation for birds andmammals in the Andes of southeastern Peru. J.Biogeogr. 25: 593–607.

Poulin, B., G. Lefebvre, & R. McNeil. 1993. Varia-tions in bird abundance in tropical arid andsemi-arid habitats. Ibis 135: 432–441.

Poulsen, B. O., & N. Krabbe. 1998. Avifaunal diver-sity of five high-altitude cloud forests on theAndean western slope of Ecuador: testing arapid assessment method. J. Biogeogr. 25: 83–93.

Prado, D. E. 1995. Selva pedemontana: contextoregional y lista florística de un ecosistema en

peligro. Pp. 19–52 in Brown A. D., & H. R.Grau (eds.). Investigación, conservación ydesarrollo en selvas subtropicales de montaña.Proyecto de Desarrollo Forestal, Laboratoriode Investigaciones Ecológicas de las Yungas,Tucumán, Argentina.

Remsen, J. V., Jr., A. Jaramillo, M. A. Nores, M. B.Robbins, T. S. Schulenberg, F. G. Stiles, J. M. C.da Silva, D. F. Stotz, & K. J. Zimmer. 2004. Aclassification of the bird species of SouthAmerica. American Ornithologists’ Union. Ver-sion: June 2004 (http://www.museum.lsu.edu/~Remsen/SACBaseline.html).

Renjifo, L. M. 2001. Effect of natural and anthro-pogenic landscape matrices on the abundanceof subandean bird species. Ecol. Appl. 11: 14–31.

Ridgely, R., & G. Tudor. 1989. The birds of SouthAmerica. Volume 1: The oscine passeriformes.Univ. of Texas Press, Austin, Texas.

Ridgely, R., & G. Tudor. 1994. The birds of SouthAmerica. Volume 2: The suboscine passeri-formes. Univ. of Texas Press, Austin, Texas.

Rosenberg, M. 2003. Passage: Pattern analysis, spa-tial statistics, and geographic exegesis. Version1.1 (http://lsweb.la.asu.edu/rosenberg/Pas-sage).

Rougès, M., & J. G. Blake. 2001. Tasas de captura ydietas de aves del sotobosque en el ParqueBiológico Sierra de San Javier, Tucumán. Hor-nero 16: 7–15.

Stiles, F. G. 1988. Altitudinal movements of birdson the Caribbean slope of Costa Rica: implica-tions for conservation. Pp. 243–258 in Almeda,F., & C. M. Pringle ( eds.). Tropical rainforest:diversity and conservation. California Academyof Sciences, San Francisco, California.

Stotz, D. F., J. W. Fitzpatrick, T. A. Parker, & D. K.Moskovits. 1996. Neotropical birds: ecologyand conservation. Univ. of Chicago Press, Chi-cago, Illinois.

Terborgh, J. 1974. Faunal equilibria and the designof wildlife preserves. Pp. 340–348 in Colley, F.B., & E. Medina (eds.). Tropical ecological sys-tems: trends in terrestrial and aquatic research.Springer-Verlag, New York, New York.

Terborgh, J. 1977. Bird species diversity on anAndean elevational gradient. Ecology 58:1007–1019.

245

MALIZIA ET AL.

Vellend, M. 2001. Do commonly used indices ofβ-diversity measure species turnover? J. Veg.Sci. 12: 545–552.

Winker, K., P. Escalante, J. H. Rappole, M. A.Ramos, R. J. Oehlenschlager, & D. H. Warner.1997. Periodic migration and lowland forestrefugia in a “sedentary” Neotropical bird, Wet-

more’s Bush-Tanager. Conserv. Biol. 11: 692–697.

Young, B. E., D. DeRosier, & G. V. N. Powell.1998. Diversity and conservation of understorybirds in the Tilarán Mountains, Costa Rica. Auk115: 998–1016.

246

247

BIRDS O

F AN

DE

AN

PREM

ON

TAN

E FO

RESTS

APPENDIX 1. List of species detected in Andean premontane forests of northwestern Argentina, including bird attributes (migrant status, feeding guild,Neotropical abundance, and premontane-forest abundance), number of individuals counted (considering 10 point counts per site; not mist nets or transects)at fla g and non-breeding seasons. Numberof sit tes.

Spec Non-breeding

o. sites Ind. flat Ind. hill No. sitesTataDusBuffTurkBlacHooSwaPumGraRoaBarrPaleWhiLargGolWhiGreScalBlueDarSquiStripWhiSickPlanGlit

61121121143654466441212402

100000000000000061210000006

500000000000000044080000007

60000000030004006260000015

t (ind. flat) and foothill (ind. hill) forests, and number of sites occupied (considering all methods) during breedines sampled: breeding season, two flat and four foothill sites; non-breeding season, three flat and three foothill si

ies Migr. Feeding Neotrop. Premon. Breeding

status1 guild2 abund3 abund.4 Ind. flat Ind. hill Nupa Tinamou (Crypturellus tataupa)ky-legged Guan (Penelope obscura)-necked Ibis (Theristicus caudatus)ey Vulture (Cathartes aura)k Vulture (Coragyps atratus)k-billed Kite (Chondrohierax uncinatus)llow-tailed Kite (Elanoides forficatus)beous Kite (Ictinia plumbea)

y Hawk (Asturina nitida)dside Hawk (Buteo magnirostris)ed Forest-Falcon (Micrastur ruficollis)-vented Pigeon (Patagioenas cayennensis)te-tipped Dove (Leptotila verreauxi)e-tailed Dove (Leptotila megalura)

den-collared Macaw (Propyrrhura auricollis)te-eyed Parakeet (Aratinga leucophthalma)en-cheeked Parakeet (Pyrrhura molinae)ly-headed Parrot (Pionus maximiliani)-fronted Parrot (Amazona aestiva)k-billed Cuckoo (Coccyzus melacoryphus)rrel Cuckoo (Piaya cayana)ed Cuckoo (Tapera naevia)te-collared Swift (Streptoprocne zonaris)'s Swift (Chaetura meridionalis)alto Hermit (Phaethornis pretrei)tering-bellied Emerald (Chlorostilbon aureoventris)

RRRRRRLL?RRLRRRR?RRRLRRRLAR

GI.G.GFI.X.FI.G.GC.A.GC.A.GC.C.FI.A.AI.A.AC.C.FC.C.FC.X.AG.C.FG.G.GG.G.GG.C.FG.C.FG.C.FG.C.FG.C.FI.U.FO.X.FI.U.FI.A.AI.A.AN.U.FN.X.F

FFFCCUUCFCFCCUFCCFFFCCFCFC

CUUCUUCUUCCCCCCCCCCUCUCCUC

800000000003000284161000700

140000020010120867933260100103

248

MA

LIZIA

ET A

L.APPENDIX 1. Continued.

Spec Non-breeding

o. sites Ind. flat Ind. hill No. sitesWhiRedBlueBlueBlueTocoWhiDotGolCreaOlivGreBlacNarSooAzaOchStripBuffStreGianGreVariStripBlacWhiGreSmaHigh

30254243436046226253416252310

21041001007100061802001020101101

10040003031101103170410006000001

52064005066362366050306020324

ies Migr. Feeding Neotrop. Premon. Breeding

status1 guild2 abund3 abund.4 Ind. flat Ind. hill Nte-bellied Hummingbird (Amazilia chionogaster)-tailed Comet (Sappho sparganura)-tufted Starthroat (Heliomaster furcifer)-crowned Trogon (Trogon curucui)-crowned Motmot (Momotus momota) Toucan (Ramphastos toco)

te-barred Piculet (Picumnus cirratus)-fronted Woodpecker (Veniliornis frontalis)den-olive Woodpecker (Piculus rubiginosus)m-backed Woodpecker (Campephilus leucopogon)aceous Woodcreeper (Sittasomus griseicapillus)at Rufous Woodcreeper (Xiphocolaptes major)k-banded Woodcreeper (Dendrocolaptes picumnus)row-billed Woodcreeper (Lepidocolaptes angustirostris)ty-fronted Spinetail (Synallaxis frontalis)ra's Spinetail (Synallaxis azarae)re-cheeked Spinetail (Synallaxis scutata)e-crowned Spinetail (Cranioleuca pyrrhophia)-browed Foliage-gleaner (Syndactyla rufosuperciliata)aked Xenops (Xenops rutilans)t Antshrike (Batara cinerea)

at Antshrike (Taraba major)able Antshrike (Thamnophilus caerulescens)e-backed Antbird (Myrmorchilus strigilatus)k-capped Antwren (Herpsilochmus atricapillus)te-throated Antpitta (Grallaria albigula)enish Elaenia (Myiopagis viridicata)ll-billed Elaenia? (Elaenia sp. [parvirostris?])land Elaenia (Elaenia obscura)

RARRRRRRRRRRRRRRRRRRRRRRL?RRLA

N.C.FN.X.FN.E.FFI.C.FO.U.FO.C.FI.U.BI.U.BI.C.BI.C.BI.X.BI.X.BI.X.BI.X.BI.U.FI.U.FI.U.FI.C.BI.X.FI.C.BO.U.FO.U.FI.U.FI.U.FI.C.FI.G.GI.U.AF

FI.X.AFFI.U.AF

CFUFCFCCF

U?CUUFFFUFCFUCCFCUFCU

CCCCCCCCCCCCCCCCCCCCCUCCCCCCC

0000102000190221012010501333180200

30021010332802702174260504302500120

249

BIRDS O

F AN

DE

AN

PREM

ON

TAN

E FO

RESTS

APPENDIX 1. Continued.

Spec Non-breeding

No. sites Ind. flat Ind. hill No. sitesSouBuffWhiWhiMotSepiPearOchYelloEuleFuscAndWhiGreStreVariTropRufoDusSwaBrowGreWhiRufoRedPlusHouSpoSwa

20026226440000511043412466104

026006022000001000100100108800

020010305600500000000000006800

04205505411512000400100206510

ies Migr. Feeding Neotrop. Premon. Breeding

status1 guild2 abund3 abund.4 Ind. flat Ind. hillthern Beardless-Tyrannulet (Camptostoma obsoletum)-banded Tyrannulet (Mecocerculus hellmayri)te-throated Tyrannulet (Mecocerculus leucophrys)te-crested Tyrannulet (Serpophaga subcristata)tled-cheeked Tyrannulet (Phylloscartes ventralis)a-capped Flycatcher (Leptopogon amaurocephalus)ly-vented Tody-Tyrant (Hemitriccus margaritaceiventer)re-faced Tody-Flycatcher (Poecilotriccus plumbeiceps)w-olive Flycatcher (Tolmomyias sulphurescens)r's Flycatcher (Lathrotriccus euleri)ous Flycatcher (Cnemotriccus fuscatus)ean Tyrant (Knipolegus signatus)te-winged Black-Tyrant (Knipolegus aterrimus)at Kiskadee (Pitangus sulphuratus)aked Flycatcher (Myiodynastes maculatus)egated Flycatcher (Empidonomus varius)ical Kingbird (Tyrannus melancholicus)us Casiornis (Casiornis rufus)

ky-capped Flycatcher (Myiarchus tuberculifer)inson's Flycatcher (Myiarchus swainsoni)

n-crested Flycatcher (Myiarchus tyrannulus)en-backed Becard (Pachyramphus viridis)te-winged Becard (Pachyramphus polychopterus)us-browed Peppershrike (Cyclarhis gujanensis)

-eye Vireo (Vireo olivaceus)h-crested Jay (Cyanocorax chrysops)se Wren (Troglodytes aedon)

tted Nightingale-Thrush (Catharus dryas)inson's Thrush (Catharus ustulatus)

RAARRARRRLRALRLLLRL?LLRLRLRRAL

I.C.AFI.C.AFI.C.AFI.E.AFI.X.F

I.U.AFI.U.FI.U.F

I.C.AFI.U.AFI.U.AFI.X.AFI.E.AFI.E.FI.C.FI.C.AI.A.AI.C.AFI.X.AFI.C.AFI.C.AFI.C.FI.C.FI.C.FI.C.FO.X.FI.U.F

FI.U.FFI.U.F

FFCFFFCFFFFUFCCFCFCF

F/CU/P

CCCFCF-

CCCCCCCCCCUCUCCUUCCCCUCCCCCUC

00004032551300008100011201205301

100341819830000320205330292616004

250

MA

LIZIA

ET A

L.APPENDIX 1. Continued.

Spec Non-breeding

No. sites Ind. flat Ind. hill No. sites

SlatyRufoCreaHooOraSayaGuiCheRustComHepRufoBlacDulSaffStripRedBlacGraUltrTropBrowTwoCresEpaShinPurpGol

2621145600102063013160620151

2183009020171000280013004923212040

1181005400046010024000005965000020

5630350116110261230066622060

ies Migr. Feeding Neotrop. Premon. Breeding

status1 guild2 abund3 abund.4 Ind. flat Ind. hill

Thrush (Turdus nigriceps)us-bellied Thrush (Turdus rufiventris)my-bellied Thrush (Turdus amaurochalinus)ded Tanager (Nemosia pileata)

nge-headed Tanager (Thlypopsis sordida)ca Tanager (Thraupis sayaca)ra Tanager (Hemithraupis guira)stnut-vented Conebill (Conirostrum speciosum)y Flowerpiercer (Diglossa sittoides)mon Bush-Tanager (Chlorospingus ophthalmicus)atic Tanager (Piranga flava)us-collared Sparrow (Zonotrichia capensis)

k-capped Warbling-Finch (Poospiza melanoleuca)l-colored Grassquit (Tiaris obscurus)ron-billed Sparrow (Arremon flavirostris)e-headed Brush-Finch (Buarremon torquatus)

-crested Finch (Coryphospingus cucullatus)k-backed Grosbeak (Pheucticus aureoventris)yish Saltator (Saltator coerulescens)amarine Grosbeak (Cyanocompsa brissonii)ical Parula (Parula pitiayumi)n-capped Redstart (Myioborus brunniceps)

-banded Warbler (Basileuterus bivittatus)ted Oropendola (Psarocolius decumanus)ulet Oriole (Icterus cayanensis)y Cowbird (Molothrus bonariensis)le-throated Euphonia (Euphonia chlorotica)

den-rumped Euphonia (Euphonia cyanocephala)

A,LRRRRRR?RAARRRR?RRRARRRARRRRRR

FI.X.FFI.X.FFI.X.FFI.C.FFI.U.FFI.X.FFI.C.FI.C.FI.X.F

FI.X.FFI.C.FGI.E.FGI.E.FGI.E.FGI.U.FGI.U.FGI.E.FGI.C.FGI.E.FGI.E.FI.C.FI.X.FI.U.FO.C.FI.C.F

GI.E.FFI.C.FFI.C.F

FCCFFCCCFCFCC

U/PFCCCCFCFF

C/PFCC

F/P

CCCUCCCCUCUUCCCCCCCUCCCCCUCU

011020841000000700010110900030

226102162200000019502314204950070

251

BIRDS O

F AN

DE

AN

PREM

ON

TAN

E FO

RESTS

APPENDIX 1. Continued.

Spec Non-breeding

No. sites Ind. flat Ind. hill No. sites

TotaTota

––

35145

46335

––

1Mig2Feed diet: C = carnivore, FI = frugivore-inse nd set of letters refers to forest stratawhe strata. iii) The third set of letters refersto fo

3Neo common.4Prem

ies Migr. Feeding Neotrop. Premon. Breeding

status1 guild2 abund3 abund.4 Ind. flat Ind. hill

l individualsl species

31947

68664

rant status: A = elevational migrant, L = latitudinal migrant, R = resident.ing guilds: composed as a combination of three categories separated by dots. i) The first set of letters refers to

ctivore, G = granivore, GI = granivore-insectivore, I = insectivore, N = nectarivore, O = omnivore. ii) The secore birds forage: A = above the canopy, C = canopy, E = forest edge, G = ground, U = understory, X = all forest raging substrate from where food is obtained: A = air, AF = air-foliage, B = bark, F = foliage, G = ground.tropical abundance (sensu Parker et al. 1996): C = common, F = fairly common, P = patchily distributed, U = unontane-forest abundance (at the mesoscale): C = common, U = uncommon.