439

The Wilson Journal of Ornithology 119(3):439–449, 2007

AVIAN BIOGEOGRAPHY OF AMAZONIAN FLOODED FORESTS INTHE RIO BRANCO BASIN, BRAZIL

LUCIANO NICOLAS NAKA,1,4,5 MARIO COHN-HAFT,1 ANDREW WHITTAKER,2

JUAN MAZAR BARNETT,3 AND MARCELA DE FATIMA TORRES1

ABSTRACT.—Flooded forests represent an important part of Amazonian diversity, yet the distribution, ecol-ogy, and evolutionary history of the avifauna of these forests have received little attention. We conductedornithological surveys in the Rio Branco Basin, which is entirely in the Brazilian State of Roraima. In this paper,we discuss the presence of 20 bird species recorded along the lower Rio Branco, 16 of which represent newrecords for the State of Roraima and the entire Rio Branco Basin. Among our most interesting records are fourspecies of white-water river specialists (Synallaxis propinqua, Stigmatura napensis, Serpophaga hypoleuca, andConirostrum bicolor) that have populations on the lower Rio Branco, isolated from other Amazonian white-water river systems by the black waters of the Rio Negro where they do not occur. We also discovered newlocalities for the endemic and endangered Rio Branco Antbird (Cercomacra carbonaria), doubling the size ofits known range. We discuss the implications of these records in a biogeographic perspective for better under-standing the distributional patterns of the flooded-forest avifauna in Amazonia. Received 11 May 2006. Accepted12 December 2006.

The Amazon Basin contains the most spec-tacular river system on earth. Not only is theAmazon River the longest and largest river inthe world, but some of its tributaries also rateamong the 10 largest (Goulding et al. 2003).These rivers create the Amazonian floodplains, a mosaic of habitats which remainflooded for as much as 8 months every year(Junk 1989). These seasonally flooded habi-tats, known in Brazil as varzeas and igapos,include floating meadows, sandbar scrub, riv-er-edge forest, and tall flooded forest eachhosting a diverse and highly specialized avi-fauna that accounts for at least 15% of Ama-zonia’s overall non-aquatic bird diversity(Remsen and Parker 1983).

Despite the continental extent of this eco-system, which ranges thousands of kilometersacross northern South America, from the baseof the Andes to the Atlantic Ocean, the dis-tribution, ecology, and evolutionary history ofits avifauna have been poorly studied. Until

1 Instituto Nacional de Pesquisas da Amazonia -INPA, Colecoes Zoologicas e Coordenacao de Pes-quisas em Ecologia, C.P. 478, Manaus, Amazonas69011-970, Brazil.

2 Rua Samaumas 214, Aleixo, Conjunto Acariquara,Manaus, Amazonas 69085-410, Brazil.

3 Av. Forest 1531 1o B, (1430), Buenos Aires, Ar-gentina.

4 Current address: Museum of Natural Science, Lou-isiana State University, 119 Foster Hall, Baton Rouge,LA 70803, USA.

5 Corresponding author; e-mail: lnaka1@lsu.edu

recently, flooded forests have not captured theattention of avian biogeographers (Cohn-Haftet al. in press), who have focused mostly ondescribing bird distribution patterns and gen-erating evolutionary hypotheses for Amazo-nian terra firme forests (e.g., Sick 1967, Haf-fer 1969, Nores 1999). We attribute this lackof interest in the biogeography of flooded for-est birds, in part, to the general assumptionthat these birds, by living in dynamic and rel-atively ephemeral habitats, must have strongdispersal abilities. Strong dispersal abilities incontinuous, linear environments should not fa-vor genetic isolation and differentiation ofpopulations, leading to the expectation ofwidespread species with little or no geograph-ic variation (Remsen and Parker 1983). Thelack of precise distributional data for varzeabirds has also prevented ornithologists frommaking historical interpretations that go be-yond a handful of species (e.g., Haffer andFitzpatrick 1985, Haffer 1997, Isler et al.1999, Aleixo 2006).

The Rio Branco is unique in Amazonia, be-cause it is a nutrient-rich, semi-muddy river,surrounded by black-water rivers (poor in sed-iments and nutrients, but rich in organic mat-ter). The ‘color’ of the water has importantecological implications, because muddy andsemi-muddy rivers (hereafter ‘white-water’rivers) differ from black-water rivers not onlyin the amount of sediments they carry, butalso in the structure and composition of the

440 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 119, No. 3, September 2007

vegetation along their margins (Klinge andFurch 1991), affecting the species richnessand composition of its associated avifauna(Rosenberg 1990, Petermann 1997, Borgesand Carvalhaes 2000). Many widespreadflooded-forest birds occur in riverine habitatscreated by both types of rivers, but some arerestricted to either black-water or white-waterrivers. White-water river specialists that arepresent on the Rio Branco are necessarily iso-lated from other white-water drainages ofAmazonia, potentially allowing genetic andpopulation differentiation.

The Rio Branco Basin lies in a region ofhigh habitat heterogeneity and of great eco-logical and biogeographical interest, importantfor our understanding of distributional pat-terns of Amazonian birds (Naka et al. 2006).The only major ornithological explorations onthe Rio Branco until the mid 20th centurywere those of Natterer in 1831 and 1832 (vonPelzeln 1868–1871). More recently, Stotz(1997) and Zimmer et al. (1997) surveyed theriver, but like Natterer only explored the upperRio Branco, where river-created habitats con-sist of narrow stretches of gallery forest. Theavifaunas of gallery and flooded forests aregenerally distinct; the former being moreclosely related to dry forests in the savannasthan to Amazonian flooded forests. The firstcomments on the birds of the lower Rio Bran-co were published by Pacheco (1995) after ashort visit to the mouth of the river and bySantos (2004), who worked on several riverislands near Caracaraı.

This paper reports our most interesting ob-servations obtained during several avifaunalsurveys on the lower Rio Branco since 1990,including several significant range extensionsand 16 species new to the Rio Branco Basinand the State of Roraima. We discuss the bio-geographical implications of these records andcall attention to the presence of previouslyoverlooked and possibly isolated populationsof white-water river specialists on the RioBranco.

METHODS

Study Area.—The Rio Branco Basin iswithin the Brazilian State of Roraima (Fig. 1)and the river is the largest tributary of the RioNegro, itself the second-largest tributary ofthe Amazon and the largest black-water river

in the world. The basins of the Rio Negro andRio Branco jointly contribute nearly 15% ofthe total annual discharge of the Amazon Riv-er and occupy an area of roughly 700,000km2, more than 10% of the Amazon Basin(Goulding et al. 2003). The Rio Branco is notas turbid as the Amazon or the Madeira rivers,but its waters contrast strongly with the darkwaters of the Negro. On satellite imagery, this‘‘meeting of the waters’’ can be seen for morethan 50 km downstream from their conflu-ence.

We divide the Rio Branco in two main geo-graphical and ecological regions: the upperand lower Rio Branco with the city of Cara-caraı (01� 50� N) as the dividing point. About70 km north of Caracaraı, coinciding with theconfluence of the Rio Branco and the Rio Mu-cajaı, rainfall decreases sharply (Barbosa1997) and soil conditions change drastically(Brown and Prance 1987). Savannas replacelowland Amazonian forests in this region and,along the rivers, gallery forests replace flood-ed forests. We have not sampled the Rio Bran-co between Caracaraı and Mucajaı, and do notknow how gradually these changes occur.However, apparently many flooded-forest birdspecies readily found around Caracaraı are notpresent in the gallery forests of the Rio Mu-cajaı or Rio Branco from Mucajaı north.

The terms varzea and igapo are local termsused to describe flooded forests and are oftenused interchangeably by local inhabitants.Here, we follow Prance (1979) using theseterms to distinguish forests based on the kindof water that floods them: varzea for forestsflooded by white (muddy) waters and igapofor forests flooded by black (translucent, acid-ic) waters. The banks of the lower Rio Brancoare covered by extensive varzeas, whereasmany of its tributaries (i.e., Agua Boa do Uni-vini, Irua, and Ajarani) are black-water rivers,whose margins are covered by igapo.

The Rio Branco has an average river-levelfluctuation of 5.7 m, which is enough to floodlarge areas along the river, and to ‘dam’ small-er tributaries (Goulding et al. 2003). Vegeta-tion structure and floristic composition dependon the duration and level of the flooding; areasflooded for longer periods tend to supportfewer tree species (Junk 1989). The varzeasof the Rio Branco have a well-developed can-opy reaching 20–25 m in height, and the un-

441Naka et al. • BIRDS OF THE RIO BRANCO FLOODED FORESTS

FIG. 1. The Rio Branco Basin and study localities, enclosed by the Brazilian State of Roraima, and includedwithin the Rio Negro Basin. White-water rivers are white; black-water rivers are black. Stippled areas representreserves. Numbers represent the following rivers or localities: (1) Rio Mucajaı, (2) Rio Ajaranı, (3) NiquiaEcological Station, (4) Virua National Park, (5) Rio Irua, (6) Rio Anaua, (7) Rio Catrimani, (8) Rio Agua Boado Univini, (9) village of Samauma on the Rio Jauaperı, (10) Parana da Floresta, (11) Marari Island, and (12)Inajatuba island. Digital data obtained from Ministerio do Meio Ambiente (2001).

derstory is generally open and poorly devel-oped. Tree species diversity and complexity ofvertical structure on river islands increase asan island ages (Robinson and Terborgh 1997)creating a series of successional stages that

range from sandbars to river-edge forest(dominated by Cecropia) to mature varzeaforest, which has the highest number of treespecies (Remsen and Parker 1983). Igapos areusually associated with sandy soils and have

442 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 119, No. 3, September 2007

fewer tree species than varzea forests, narrow-er tree diameters, and a lower canopy height(15–20 m). The physiognomy and floristiccomposition of the igapo resemble that ofwhite-sand forests (Anderson 1981).

Field Work.—AW visited the lower RioBranco near the mouth of the river on 10–11June and 22–24 December 1990, 6–7 April1991, 19–20 August 1993, 25–26 January and25–26 July 1994, 24–25 March and 15–16July 1995, and 9–10 June 1996. MCH madea short visit to the mouth of the Rio Brancoon 6 July 1999. In 2001 and 2006 we joinedthree expeditions sponsored by the InstitutoBrasileiro do Meio Ambiente e dos RecursosNaturais Renovaveis (IBAMA) to survey theavifauna of the lower Rio Branco and someof its tributaries. Between 6 and 11 July 2001,LNN and MFT covered �100 km of the RioBranco (from Santa Maria do Boiacu to themouth of the river), a short stretch of the RioNegro influenced by the Rio Branco (Paranada Floresta), and the west bank of the RioJauaperı (from its mouth to the village of Sa-mauma). Between 5 and 18 October 2001,LNN and JMB covered �170 km of the RioBranco from Caracaraı south to the mouth ofthe Rio Catrimani and some Rio Branco trib-utaries (Agua Boa do Univini, Irua, Anaua,and Ajarani rivers). Between 1 and 15 October2006, LNN and MCH, accompanied by Mar-celo dos Santos Jr. and Catherine L. Bechtoldt,studied �80 km of the Rio Branco from Car-acaraı to the confluence with the Rio Anaua,and another 35 km up the white-waters of theRio Anaua and a similar distance along theblack-waters of the Rio Irua.

We usually conducted bird surveys duringearly morning (0500–1200 hrs), late afternoon(1500–1830 hrs), and sporadically at night us-ing binoculars and tape-recorders. We ex-plored river-created habitats (varzea and iga-po) using small motorboats, which allowedaccess to the forest interior. LNN and MCHcollected bird specimens, housed at the Insti-tuto Nacional de Pesquisas da Amazonia(INPA) Bird Collection in Manaus, Brazil.Tape-recordings by LNN and MCH are ar-chived at the INPA Bird Collection; those byAW are at the British Library National SoundArchive. Taxonomy and species nomenclaturefollow the Brazilian Ornithological Records

Committee (Comite Brasileiro de RegistrosOrnitologicos 2006).

RESULTS

We present details of 20 species recordedduring our fieldwork which represent eithernew records for the Rio Branco Basin and theState of Roraima (species marked with an as-terisk), or note-worthy range extensions.

Species Accounts

*Leucopternis schistaceus (Slate-coloredHawk).—AW found two single individuals onriver islands on the lower Rio Branco (15 and27 km north of the river’s mouth) on 10 June1990 and 7 April 1991, respectively. In Oc-tober 2001 LNN and JMB observed and tape-recorded two birds in tall varzea forest alongthe east bank of the lower Rio Branco (01�36� 20� N, 61� 13� 08� W) at Virua NationalPark. This species seems to be restricted totall flooded forests in Brazil (Bierregaard1994) and probably occurs in the State of Ro-raima only along the lower Rio Branco.

*Asio stygius (Stygian Owl).—AW ob-served one bird perched on a sand bar of ariver island 25 km north of the mouth of theRio Branco on the night of 22 December1990. In subsequent visits to that island thefollowing day and following year (April 1991)AW repeatedly found an individual there (pre-sumably the same bird). This species is widelydistributed from Mexico to Argentina, butseems to be rare and local (Marks et al. 1999).This owl generally inhabits savannas in Am-azonian Brazil, but in recent years AW hasobtained several records from areas associatedwith black-water rivers and river islands.

*Monasa nigrifrons (Black-fronted Nun-bird).—AW found this species along the lowerRio Branco on 23 December 1990 and in suc-cessive trips to the same general area. Sub-sequently, LNN, JMB, and MFT observed andtape-recorded several individuals in varzeaforest along both banks of the lower Rio Bran-co. Our records ranged from the mouth of theriver north to the island of Inajatuba (01� 17�22� N, 61� 18� 04� W), nearly 60 km south ofCaracaraı. This species is widespread in Ama-zonia, mostly restricted to flooded forests(Rasmussen and Collar 2002), but seems to berestricted in the State of Roraima to the var-zeas of the lower Rio Branco, with no known

443Naka et al. • BIRDS OF THE RIO BRANCO FLOODED FORESTS

records from gallery forests further north orfrom Venezuela (Hilty 2003).

*Myrmotherula klagesi (Klages’ Ant-wren).—AW heard and observed several pairsin mature varzea forest �20 km north of themouth of the Rio Branco on 23 December1990. Birds were subsequently tape-recordedby AW along both banks of the Rio Jauaperıin August 1993 and other sites along the lowerRio Branco in June 1993 and July 1995. MCHtape-recorded the species near the mouth ofthe Rio Branco in July 1999, and in July andOctober 2001, LNN and MFT, and LNN andJMB, respectively, documented its presencethroughout the lower Rio Branco, obtainingtape-recordings from the mouth of the river asfar north as the island of Inajatuba, 300 km tothe north. On 6 October 2006 MCH tape-re-corded and collected a male from a river is-land on the Rio Branco, opposite the mouthof the Rio Anaua, representing the first spec-imen for the State of Roraima. Previously, thisspecies was known to occur on the lower RioNegro and along the Amazon, including theRio Madeira and the mouth of the Rio Tapajos(Zimmer and Isler 2003). Our records extendthe known distribution of M. klagesi well intothe Rio Branco Basin and more than 300 kmnorthward from previous known sites on theRio Negro.

*Myrmotherula assimilis (Leaden Ant-wren).—AW observed this species in tall var-zea forest �20 km north of the mouth of theRio Branco on 10 June 1990 and several othertimes in the same general area. In July andOctober 2001, LNN and MFT, and LNN andJMB, respectively, tape-recorded this speciesthroughout the lower Rio Branco, as far northas the island of Inajatuba, and also along theAnaua and Jauaperı rivers. This species isknown to be widespread in flooded foreststhroughout much of Amazonia (Zimmer andIsler 2003), but was previously unrecordedfrom the Rio Branco Basin.

*Cercomacra nigrescens (Blackish Ant-bird).—MCH tape-recorded this species inRoraima at the mouth of the Rio Branco on 6July 1999. Subsequently, in July and October2001, LNN and MFT, and LNN and JMB, re-spectively, saw and tape-recorded several in-dividuals in mid-successional vegetation onriverine islands and forest edges along bothbanks of the lower Rio Branco from the

mouth of the river north to the island of Ina-jatuba where it was fairly common. Addition-ally, LNN and MFT tape-recorded several in-dividuals along the Parana da Floresta on theRio Negro. On 15 October 2006 LNN tape-recorded and collected a pair on the left bankof the Rio Branco (01� 31� 06� N, 61� 14� 25�W). Three other individuals had been collect-ed earlier by M. P. D. Santos from Virua Na-tional Park and Sao Joao da Baliza in August2002 and January 2003, respectively. Most ofthe known range of C. nigrescens is restrictedto areas south of the Amazon from easternColombia to the Brazilian State of Para (Zim-mer and Isler 2003). The only known popu-lations north of the Amazon are representedby the nominate race, which occurs in Suri-nam, French Guiana, and along the north bankof the Amazon in the Brazilian State of Para(Pinto 1978, Haverschmidt and Mees 1994).These are the first published records of C. ni-grescens from the Rio Branco and are sepa-rated by �360 km and more than 600 kmfrom nearest localities to the south and north,respectively, apparently representing a dis-junct population.

Cercomacra carbonaria (Rio Branco Ant-bird).—This taxon was considered to be strict-ly associated with gallery forests along the up-per Rio Branco and some of its tributaries(Stotz 1997, Zimmer et al. 1997, Santos 2003,Zimmer and Isler 2003). In July and October2001, LNN and MFT, and LNN and JMB, re-spectively, found and tape-recorded severalindividuals along the lower Rio Branco whereit was locally common in young riverine is-lands dominated by Cecropia spp. In October2006, LNN and MCH also found the speciesto be common on riverine islands from Car-acaraı south to the mouth of the Rio Anaua.Our records extend the known range of thisspecies south of Caracaraı (previously consid-ered its southernmost limit) to only 13 kmnorth from the mouth of the Rio Branco (01�16� 19� S, 61� 50� 21� W). We do not expect tofind this species anywhere outside the RioBranco Basin or in adjacent black-water riverswhere successional river islands do not occur.We failed to find C. carbonaria in tall varzeaforests where C. nigrescens was present sug-gesting that species interactions and habitatpreferences may limit its distribution withinthe Rio Branco Basin. Our records along the

444 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 119, No. 3, September 2007

lower Rio Branco suggest that C. carbonariahas greater ecological plasticity than previ-ously thought and represent a range extensionof more than 300 km southward, doubling thesize of its global distribution. At present, C.carbonaria is considered Vulnerable to ex-tinction by BirdLife International (2000), butwas recently excluded from the Brazilian listof threatened birds (Ministerio do Meio Am-biente 2003) based on lack of data. More re-cently, Vale et al. (in press) suggested down-listing the species to near-threatened, in partdue to the significant range extensions herepresented.

*Sclateria naevia (Silvered Antbird).—AWfound a vocalizing male 25 km north of themouth of the Rio Branco on 16 July 1995.Subsequently, LNN and MFT observed andtape-recorded a pair along a dry stream nearSao Joao da Baliza at the edge of the Wai-WaiIndian Reservation (01� 05� 21� N, 59� 56�37� W) in March 2001 and, in October 2006,LNN and MCH found several birds along theRio Anaua. These are the first records for theregion and fill what had appeared as a some-what inexplicable gap in its distribution (Zim-mer and Isler 2003).

Synallaxis propinqua (White-bellied Spine-tail).—AW found several individuals vocaliz-ing in early-successional vegetation on a riverisland 27 km north of the mouth of the RioBranco on 21 March 1993. One individualwas tape-recorded by Pacheco (1995) in Oc-tober 1993 near the mouth of the Rio Branco.In October 2006, LNN and MCH found sev-eral individuals on a young river island alongthe Rio Branco, opposite the mouth of the RioAnaua, and on two other river islands. On 14October 2006 MCH tape-recorded and col-lected one male on the island of Marari (01�28� 42� N, 61� 14� 58� W) where two pairswere building nests. Previous to these records,this spinetail was known to occur locally inearly-successional growth on river islandsalong the Amazon River and some of itswhite-water tributaries (Remsen 2003). Ourrecords not only document the presence of thisspecies in the Rio Branco Basin, but also sug-gest the presence of a breeding population iso-lated from the main populations distributedcontiguously along the Amazon River and itstributaries.

*Cranioleuca gutturata (Speckled Spine-

tail).—LNN and JMB found and tape-record-ed one individual in mature varzea forest atthe Niquia Ecological Station (00� 50� 08� N,61� 25� 45� W) on 6 October 2001. On 9 Oc-tober 2006 LNN tape-recorded and collecteda male in tall varzea forest along the Rio An-aua (00� 57� 20� N, 61� 09� 22� W). These arethe first records from the Rio Branco Basinalthough this species is relatively widespreadin Amazonian flooded forests and is presentalong the lower Rio Negro (Cintra et al.2007).

*Hemitriccus minor (Snethlage’s Tody-ty-rant).—AW saw and tape-recorded one indi-vidual 20 km north of the mouth of the RioBranco on 20 August 1993. Subsequently,MCH found this species in Roraima at themouth of the Rio Branco in July 1999, and inJuly and October 2001, LNN and MFT, andLNN and JMB, respectively, found it alongmost of the lower Rio Branco and other black-water tributaries, including the Agua Boa doUnivini and Ajaranı rivers (our northernmostlocality in Roraima). Our records seem to re-fer to the race pallens, which is a flooded-forest inhabitant common along the middleand lower Rio Negro.

Serpophaga hypoleuca (Riverside Tyran-nulet).—AW found a pair on a young riverisland 27 km north from the mouth of the RioBranco on 21 March 1993. This species wasalso recorded on the lower Rio Branco by Pa-checo (1995). Prior to these records, this ear-ly-successional vegetation inhabitant wasknown from several white-water rivers in theAmazon Basin, but not from the Rio Brancoor the Rio Negro basins (Fitzpatrick 2004).

*Stigmatura napensis (Lesser Wagtail-ty-rant).—AW found a pair in early-successionalvegetation on a river island 27 km upstreamfrom the mouth of the Rio Branco on 21March 1993. On 13 October 2006 MCH andLNN tape-recorded and collected a pair in ear-ly successional vegetation on the island ofMarari. Several other pairs were located onthe same island the following day. The femalecollected had unshelled eggs in the ovary,suggesting that individuals were reproducingon this island. Prior to our records, the Ama-zonian form (nominate race) of this specieswas known to occur exclusively along white-water Amazonian rivers south of the Amazon,especially on newly-formed sandbars with

445Naka et al. • BIRDS OF THE RIO BRANCO FLOODED FORESTS

bushes of the genus Tessaria (Fitzpatrick2004). We believe this species is absent fromthe Rio Negro and that birds recorded on theRio Branco represent an isolated population.Another isolated population was recentlyfound on the upper Orinoco River (Hilty2003).

*Inezia subflava (Amazonian Tyrannu-let).—AW heard several pairs in flooded veg-etation at a lake edge 20 km north of themouth of the Rio Branco on 23 December1990 and 23 March 1995. MCH, LNN, andMFT encountered and tape-recorded this spe-cies in Virua National Park in March 2001.Subsequently, in July and October 2001, LNNand MFT, and LNN and JMB, respectively,found I. subflava to be fairly common in var-zea and igapo edges along the lower RioBranco and some black-water tributaries suchas the Agua Boa do Univini, Irua, and Jauap-erı rivers. Previous specimens from the Stateof Roraima collected by Natterer (von Pelzeln1868–1871), Pinto (1966), Emılio Dente (heldat Museu Paraense Emılio Goeldi, MPEG),and D. F. Stotz (held at the Field Museum ofNatural History) were taken from the upperRio Branco (and other rivers within the sa-vannas domain) and apparently refer to I. cau-data, now considered a separate species (Zim-mer and Whittaker 2000). Our records repre-sent the first for I. subflava on the Rio Brancoand in the State of Roraima. Inezia subflava(a flooded-forest inhabitant) and I. caudata (agallery forest species) seem to replace eachother ecologically and geographically, withthe region of turnover somewhere near Cara-caraı.

*Conopias trivirgatus (Three-striped Fly-catcher).—AW saw a pair at the mouth of theRio Jauaperı on 25 July 1994. Subsequently,AW tape-recorded a few individuals in July1995 in transitional varzea forest �20 kmnorth of the mouth of the Rio Branco. Thedistribution of this species is poorly knownwith several disjunct populations throughouttropical South America (Fitzpatrick 2004).

*Schiffornis major (Varzea Schiffornis).—AW heard and observed several birds in var-zea forest near the mouth of the Rio Brancoon 7 April 1991. In October 2001, LNN andJMB saw and tape-recorded individuals in tallvarzea along the lower Rio Branco (00� 50�08� N, 61� 25� 45� W) and in igapo along the

Rio Agua Boa do Univini. One bird was col-lected in Virua National Park in August 2002by M. P. D. Santos (held at MPEG) repre-senting the first specimen for the State of Ro-raima. Several other birds were also heardalong the Rio Anaua by LNN and MCH inOctober 2006 indicating the presence of thisspecies was previously overlooked in the re-gion.

*Hylophilus semicinereus (Gray-chestedGreenlet).—AW heard one bird vocalizingnear the mouth of the Rio Branco on 21 De-cember 1990. Subsequently, in July 2001,LNN and MFT observed and tape-recordedseveral individuals in tall varzea and igapoforest along the lower Rio Branco and RioJauaperı and, in October 2001, LNN and JMBtape-recorded several individuals within theNiquia Ecological Station and Virua NationalPark. On 9 October 2006 LNN tape-recordedand collected a male from a campinarana-var-zea transition zone on the right bank of theRio Anaua (00� 57� 33� N, 61� 09� 30� W),finding several other individuals along thisriver and the Rio Irua. This species is rela-tively abundant along the Rio Negro and ourrecords fill a sampling gap in its distribution(Ridgely and Tudor 1989).

Conirostrum bicolor (Bicolor Conebill).—LNN and MFT found and tape-recorded twoindividuals on a river island (01� 16� S, 61�50� W) dominated by Cecropia. On 6 October2006 MCH tape-recorded and collected amale from a group of three birds on an earlysuccessional river island on the Rio Brancoopposite the mouth of the Rio Anaua. LNNand MCH found this species to be abundanton other river islands between Caracaraı andthe mouth of the Rio Anaua suggesting thatC. bicolor has permanent populations alongthis stretch of the river. This species has twodistinct populations in the Neotropics; one in-habiting mangroves in coastal areas in north-ern South America and the other inhabitingearly-successional vegetation in Amazonia,mainly along the Amazon and some tributar-ies (Ridgely and Tudor 1989). This specieshas been previously recorded along the lowerRio Branco by Pacheco (1995) and our rec-ords further extend its distribution 260 km tothe north, representing the northernmost lo-calities within the Amazon Basin.

*Ammodramus aurifrons (Yellow-browed

446 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 119, No. 3, September 2007

Sparrow).—AW saw two pairs along the edgeof a river island 27 km north from the mouthof the Rio Branco on 14 April 1991. Subse-quently, in July 2001, LNN and MFT saw andtape-recorded several individuals around thevillages of Floresta and Samauma. In October2001, LNN and JMB found this species to belocally common along sand-banks and sand-bar scrub on river islands along the lower RioBranco from the mouth of the river extendingto 50 km south of Caracaraı (01� 36� 19� N,61� 13� 27� W). On 13 October 2006, LNNtape-recorded and collected a male on a sandybeach on a river island on the Rio Branco(01� 14� 00� N, 61� 19� 01� W). These recordsindicate this species is widespread in the re-gion and was previously overlooked due to asampling gap.

*Sicalis columbiana (Orange-fronted Yel-low-finch).—AW found several individualsalong the edge of a river island on the lowerRio Branco on 21 December 1990. Subse-quently, in July 2001, LNN and MFT saw andtape-recorded more than 20 individuals at thevillage of Floresta on the Parana da Floresta.These seem to represent the first and only rec-ords of its occurrence within the Rio BrancoBasin and the State of Roraima.

DISCUSSION

Our studies along the lower Rio Branco re-vealed that south of Caracaraı, the avifaunaassociated with the river floodplains is repre-sented by typical flooded-forest species, sim-ilar to those in the Anavilhanas Archipelagoand the lower Rio Negro (Cintra et al. 2007).Environmental variables such as soil, floodingregimes, and rainfall differ between the lowerRio Branco and the upper portions of the riverwhere tall varzea forests are replaced by gal-lery forests. Bird species composition on thelower Rio Branco is generally similar to thatof other Amazonian rivers, such as the Ama-zon, Madeira, and the lower Rio Negro. Theavifauna of the upper Rio Branco, however,seems to be more closely related to galleryand dry semi-deciduous forests typical of thesavannas of northern Roraima, Guyana, andVenezuela. The transition zone between thetwo habitats seems to be quite abrupt, occur-ring within the 70-km stretch between Cara-caraı and the mouth of the Rio Mucajaı, co-inciding with the region where savannas re-

place humid forests less than 100 km south ofBoa Vista.

Most typical varzea bird species such asLeucopternis schistaceus, Monasa nigrifrons,Xiphorhynchus kienerii, Thamnophilus nigro-cinereus, Myrmotherula klagesi, M. assimilis,Myrmoborus lugubris, and Hemitriccus minorseem to reach their northernmost limit in thisarea. The shift in habitat seems also to be re-sponsible for ecological substitutions betweenclosely related taxa from varzea and galleryforests (i.e., Inezia subflava and I. caudata,and Ammodramus aurifrons and A. humeral-is).

The Rio Branco is unique in representing awhite-water river surrounded by black-waterdrainages including the Negro, Jauaperı, AguaBoa do Univini, Ajarani, and Irua rivers (Fig.1). Despite the high similarity between aviancommunities of the varzeas of the Rio Brancoand the tall igapo of the lower Rio Negro,several elements typical of Amazonian white-water rivers are present on the Rio Branco,but seem to be absent from the entire Rio Ne-gro and its black-water tributaries. We discov-ered populations of Cercomacra nigrescensand Stigmatura napensis along the lower RioBranco, and also confirmed the presence ofspecies previously reported from the mouth ofthe Rio Branco by Pacheco (1995), such asSynallaxis propinqua, Serpophaga hypoleuca,and Conirostrum bicolor. These species, withthe exception of Cercomacra nigrescens,which also occurs in terra firme forest in partof its range, are typical white-water river-is-land specialists (Ridgely and Tudor 1989,1994; Rosenberg 1990; Robinson and Ter-borgh 1997; Remsen 2003). Repeated recordsfrom the area since 1993 suggest these speciesmaintain stable populations locally, ratherthan representing vagrants or dispersing indi-viduals. Supporting this hypothesis are datathat we obtained of unequivocal breeding ev-idence for Stigmatura napensis and Synallaxispropinqua.

Most varzea birds are believed to be excel-lent dispersers (Remsen and Parker 1983) andthese abilities should be enhanced by havinglinear and uninterrupted habitats. The exis-tence of isolated populations of white-waterspecialists along the lower Rio Branco couldeither be the result of remarkably long-dis-tance dispersal or represent relict populations

447Naka et al. • BIRDS OF THE RIO BRANCO FLOODED FORESTS

from once-continuous habitats. Whether thesepopulations are old or relatively recent, theirisolation provides an opportunity for geneticdifferentiation and endemism in varzea forest.

Geomorphological data are still equivocalon past connections of Amazonian rivers, butfish communities of the Rio Branco seem tobe more closely related to those of the Ama-zon River than to those of the Rio Negro (Jan-zen Zuanon, pers. comm.). Recent ParsimonyAnalysis of Endemicity (PAE) for characiformfishes produced an area of purported ende-mism including the Rio Branco, the lower RioNegro, and the lower Amazon excluding theupper Rio Negro and other portions of theAmazon River (Hubert and Renno 2006). Thisimplies the lower Rio Negro may not be asstrong a barrier for these fishes as we believeit is for white-water specialist birds. Rossettiet al. (2005) suggested the existence of a bodyof water connecting the upper Amazon Riverand the Rio Branco during the late Pleisto-cene, a scenario that could explain the exis-tence of recently isolated populations ofwhite-water specialists on the Rio Branco.The extent, age, size, and genetic distinctive-ness of these avian populations have not beenstudied, but represent important data for ouroverall understanding of the evolutionary his-tory of Amazonian rivers and related floodedforests.

ACKNOWLEDGMENTS

We are particularly grateful to the Instituto Brasi-leiro do Meio Ambiente e dos Recursos Naturais Re-novaveis (IBAMA) in Roraima for allowing collectingand access to parks and other conservation units underits jurisdiction. We extend our gratitude to AntonioGaldino and Antonio and Beatriz Lisboa for their sup-port in Caracaraı, and for inviting us to participate inthree expeditions sponsored by the Institute. We alsoextend our gratitude to A. F. Coutinho Mello, Otaciliode Souza Neves Jr., Inara Rocha Santos, Giovanna Pal-lazi, and Marina Kluppel for permitting and encour-aging our research activities. We are indebted to Rein-aldo Barbosa, Vincenzo Lauriola, and Sebastiao doNascimento from INPA in Roraima for providing lo-gistical support in Boa Vista. We thank Luiz Mestre,Catherine Bechtoldt, and Marcelo dos Santos Jr. fortheir company during some of the field trips. AWthanks Mark Baker of Eco Tours for allowing him theopportunity to travel up the Rio Branco and to conductfield work, and MCH is grateful to Scott Mori andCarol Gracie of the New York Botanical Garden fortaking him to the mouth of the Rio Branco for the firsttime. Sergio Borges, C. E. Braun, Santiago Claramunt,

Jurgen Haffer, Daniel Lane, J. V. Remsen Jr., and ananonymous reviewer made helpful comments on pre-vious versions of the manuscript that greatly improvedthis paper. LNN had a Conselho Brasileiro de Desen-volvimento Cientıfico e Tecnologico (CNPq)/INPA(DTI) fellowship during the period he developed fieldwork in Roraima in 2003, and a CNPq Overseas Doc-toral Fellowship in 2006 (grant #202131/2006-8).Birders Exchange kindly donated optical and recordingequipment to the INPA ornithological collection. Thisarticle represents Publication 4 in the Amazonian Or-nithology Technical Series of the INPA Scientific Col-lections Program.

LITERATURE CITED

ALEIXO, A. 2006. Historical diversification of flood-plain forest specialist species in the Amazon: acase study with two species of the avian genusXiphorhynchus (Aves: Dendrocolaptidae). Biolog-ical Journal of the Linnean Society 89:383–395.

ANDERSON, A. 1981. White-sand vegetation of Brazil-ian Amazonia. Biotropica 13:199–210.

BARBOSA, R. I. 1997. Distribuicao das Chuvas em Ro-raima. Pages 325–335 in Homem, ambiente eecologia no estado de Roraima (R. I. Barbosa, E.J. G. Ferreira, and E. G. Castellon, Editors). In-stituto Nacional de Pesquisas da Amazonia, Ma-naus, Brazil.

BIERREGAARD, R. O. 1994. Family Accipitridae (hawksand eagles); Slate-colored Hawk species account.Page 168 in Handbook of the birds of the world.Volume 2. New world vultures to guineafowl (J.del Hoyo, A. Elliot, and J. Sargatal, Editors). LynxEdicions, Barcelona, Spain.

BIRDLIFE INTERNATIONAL. 2000. Threatened birds ofthe world. Lynx Editions and BirdLife Interna-tional, Barcelona, Spain.

BORGES, S. H. AND A. CARVALHAES. 2000. Bird speciesof black water inundated forests in the Jau Na-tional Park (Amazonas State, Brazil): their contri-bution to regional species richness. Biodiversityand Conservation 9:201–214.

BROWN, K. S. AND G. T. PRANCE. 1987. Soils and veg-etation. Pages 19–45 in Biogeography and Qua-ternary history in tropical America (T. C. Whit-more and G. T. Prance, Editors). Clarendon Press,Oxford, United Kingdom.

CINTRA, R., T. M. SANAIOTTI, AND M. COHN-HAFT.2007. Spatial distribution and habitat of the Ana-vilhanas Archipelago bird community in the Bra-zilian Amazon. Biodiversity and Conservation 16:313–336.

COHN-HAFT, M., L. N. NAKA, AND A. M. FERNANDES.In press. Padroes de distribuicao da avifauna davarzea do rio Solimoes-Amazonas in Bases cien-tificas para a conservacao da varzea: identificacaoe caracterizacao de regioes biogeograficas (A. Al-bernaz, Editor). Ibama, Brazilia, Brazil.

COMITE BRASILEIRO DE REGISTROS ORNITOLOGICOS.2006. Lista das aves do Brasil. Online at http://www.cbro.org.br. (accessed 1 August 2006).

448 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 119, No. 3, September 2007

FITZPATRICK, J. W. 2004. Family Tyrannidae (tyrant-flycatchers). Pages 170–462 in Handbook of thebirds of the world. Volume 9. Cotingas to pipitsand wagtails (J. del Hoyo, A. Elliot, and D. A.Christie, Editors). Lynx Edicions, Barcelona,Spain.

GOULDING, M., R. BARTHEM, AND E. J. G. FERREIRA.2003. The Smithsonian atlas of the Amazon.Smithsonian Institution Press, Washington, D.C.,USA.

HAFFER, J. 1969. Speciation in Amazonian forest birds.Science 165:131–137.

HAFFER, J. 1997. Contact zones between birds ofsouthern Amazonia. Ornithological Monographs48:281–306.

HAFFER, J. AND J. W. FITZPATRICK. 1985. Geographicvariation in some Amazonian forest birds. Orni-thological Monographs 36:147–168.

HAVERSCHMIDT, F. AND G. F. MEES. 1994. Birds of Su-rinam. VACO, Paramaribo, Surinam.

HILTY, S. 2003. Birds of Venezuela. Second Edition.Princeton University Press, Princeton, New Jer-sey, USA.

HUBERT, N. AND J. F. RENNO. 2006. Historical bioge-ography of South American freshwater fishes.Journal of Biogeography 33:1414–1436.

ISLER, M. L., P. R. ISLER, AND B. M. WHITNEY. 1999.Species limits in antbirds (Passeriformes: Tham-nophilidae): the Myrmotherula surinamensis com-plex. Auk 116:83–96.

JUNK, W. R. 1989. Flood tolerance and tree distributionin central Amazon floodplain. Pages 23–46 inTropical forests: dynamics, speciation and diver-sity (L. B. Holm-Nielsen, I. C. Nielsen, and H.Balsev, Editors). Academic Press, London, UnitedKingdom.

KLINGE, H. AND K. FURCH. 1991. Towards the classi-fication of Amazonian floodplains and their for-ests by means of biogeochemical criteria of riverwater and forest biomass. Interciencia 16:196–201.

MARKS, J. S., R. J. CANNINGS, AND H. MIKKOLA. 1999.Family Strigidae (typical owls). Pages 76–242 inHandbook of the birds of the world. Volume 5.Barn Owls to hummingbirds (J. del Hoyo, A. El-liot, and J. Sargatal, Editors). Lynx Edicions, Bar-celona, Spain.

MINISTERIO DO MEIO AMBIENTE. 2001. Avaliacao eacoes prioritarias para a conservacao, uso susten-tavel e reparticao de benefıcios da biodiversidadena Amazonia Brasileira. (CD-Rom). Instituto So-cioambiental (ISA), Brasilia, Brazil.

MINISTERIO DO MEIO AMBIENTE. 2003. Lista nacionaldas especies da fauna brasileira ameacadas de ex-tincao. Online at http://www.mma.gov.br/port/sbf/fauna/grupos3.html (accessed 15 April 2005).

NAKA, L. N., M. COHN-HAFT, F. MALLET-RODRIGUES, M.P. D. SANTOS, AND M. F. TORRES. 2006. The avifaunaof the Brazilian State of Roraima: bird distributionand biogeography in the Rio Branco Basin. RevistaBrasileira de Ornitologia 14:197–238.

NORES, M. 1999. An alternative hypothesis of the or-igin of Amazonian bird diversity. Journal of Bio-geography 26:275–285.

PACHECO, J. F. 1995. New distributional records forsome birds from varzea forest at Mamiraua Re-serve, western Brazilian Amazonia. Ararajuba 3:83–87.

PETERMANN, P. 1997. The Birds. Pages 299–314 in Thecentral Amazon floodplain, ecology of a pulsingsystem (W. J. Junk, Editor). Springer-Verlag, Ber-lin/Heidelberg, Germany.

PINTO, O. M. O. 1966. Estudo crıtico e catalogo re-missivo das aves do Territorio Federal de Rorai-ma. Instituto Nacional de Pesquisas da Amazonia,Manaus, Brazil.

PINTO, O. M. O. 1978. Novo catalogo das aves doBrasil e lista dos exemplares que as representamno Museu Paulista. Revista Brasileira do MuseuPaulista 22.

PRANCE, G. T. 1979. Notes on the vegetation of Ama-zonia. III. The terminology of forest types subjectto inundation. Brittonia 31:26–38.

RASMUSSEN, P. C. AND N. J. COLLAR. 2002. FamilyBucconidae (puffbirds). Pages 102–138 in Hand-book of the birds of the world. Volume 7. Jaca-mars to woodpeckers (J. del Hoyo, A. Elliot, andJ. Sargatal, Editors). Lynx Edicions, Barcelona,Spain.

REMSEN JR., J. V. 2003. Family Furnariidae (oven-birds). Pages 162–357 in Handbook of the birdsof the world. Volume 8. Broadbills to tapaculos(J. del Hoyo, A. Elliot, and D. A. Christie, Edi-tors). Lynx Edicions, Barcelona, Spain.

REMSEN JR., J. V. AND T. A. PARKER III. 1983. Contri-bution of river-created habitats to bird speciesrichness in Amazonia. Biotropica 15:221–231.

RIDGELY, R. S. AND G. TUDOR. 1989. The birds ofSouth America. Volume 1. The oscine passerines.University of Texas Press, Austin, USA.

RIDGELY, R. S. AND G. TUDOR. 1994. The birds ofSouth America. Volume 2. The suboscine passer-ines. University of Texas Press, Austin, USA.

ROBINSON, S. K. AND J. TERBORGH. 1997. Bird com-munity dynamics along primary successional gra-dients of an Amazonian whitewater river. Ornitho-logical Monographs 48:641–672.

ROSENBERG, G. H. 1990. Habitat specialization and for-aging behavior by birds of Amazonian river is-lands in northeastern Peru. Condor 92:427–443.

ROSSETTI, D. F., P. MANN DE TOLEDO, AND A. M. GOES.2005. New geological framework for western Ama-zonia (Brazil) and implications for biogeography andevolution. Quaternary Research 63:78–89.

SANTOS, M. P. D. 2003. Novos registros do chororo-do-Rio-Branco (Cercomacra carbonaria) no es-tado de Roraima, Brasil. Atualidades Ornitologi-cas 114:3.

SANTOS, M. P. D. 2004. New records of birds from theBrazilian State of Roraima. Bulletin of the BritishOrnithologists’ Club 124:223–226.

SICK, H. 1967. Rios e enchentes na Amazonia como

449Naka et al. • BIRDS OF THE RIO BRANCO FLOODED FORESTS

obstaculo para a avifauna. Atas Simposio BiotaAmazonia Zoolologica 5:495–520.

STOTZ, D. F. 1997. Levantamento preliminar da avifau-na em Roraima. Pages 581–608 in Homem, Am-biente e Ecologia no Estado de Roraima (R. I.Barbosa, E. J. G. Ferreira, and E. G. Castellon,Editors). Instituto Nacional de Pesquisas da Ama-zonia, Manaus, Brazil.

VALE, M. M., J. B. BELL, M. A. S. ALVES, AND S. L.PIMM. In press. Abundance, distribution, and con-servation of Cercomacra carbonaria and Synal-laxis kollari. Bird Conservation International 17:In press.

VON PELZELN, A. 1868–1871. Zur Ornithologie Brasi-

liens. Resultate von Johann Natterers Reisen inden Jahren 1817 bis 1835. A. Pichler’s Witwe undSohn, Wien, Austria.

ZIMMER, K. J. AND M. L. ISLER. 2003. Family Tham-nophilidae (typical antbirds). Pages 448–681 inHandbook of the birds of the world. Volume 8.Broadbills to tapaculos (J. del Hoyo, A. Elliott,and J. Sargatal, Editors). Lynx Editions, Barce-lona, Spain.

ZIMMER, K. J. AND A. WHITTAKER. 2000. Species limitsin Pale-tipped Tyrannulets (Inezia: Tyrannidae).Wilson Bulletin 112:51–66.

ZIMMER, K. J., A. WHITTAKER, AND D. F. STOTZ. 1997.Vocalizations, behavior and distribution of the RioBranco Antbird. Wilson Bulletin 109:663–678.