zoogeographical origins of eastern brazilian birds

Volume 3 1992 No.1

ORNITOWGIA NEaTROPICAL 3: 1-15, 1992@ The Neotropical Ornithological Society

ZOOGEOGRAPHICAL ORIGINS OF EASTERN BRAZILIAN BIRDS

Edwin O. Willis

Departamento de Zoologia, Universidade Estadual PaulistaCaixa Posta1199, 13500 -Rio Claro, S8o Paulo, Brazil.

Abstract. Atlantic forests of eastern Brazil are separated from Amawnian and Andean ones by a semiopen region(the caatinga-cerrado-chaco wne). Fifty-seven water birds, 166 semiopen-wne birds, and 647 other nonmarineeastern breeding species are analyzed for wogeographical patterns. Water birds rarely speciate or subspeciate, per-haps because few large areas of nesting habitat occur in eastern Brazil and because these birds wander widely. Se-miopen-wne species center in the caatinga/cerrado wne and show links to Andean-Argentine species, to birds ofdepauperate small open patches in Amawnia, or to birds in the llanos of Colombia:Venezuela northward.Speciation in the open wne is often between caatinga and cerrado or this and chaco, subspeciation additionallyacross upland forests of western Goias or across dry forests in northwestern Bahia. Forest-zone eastern birds someti-mes have crossed from Amawnia via the north coast, more often .centrally. via the dty-forested Bahian wne oreven further south, at times with double crossings forming a .ring.. A .Madeiran. relationship, the eastern birdrelated to upper-Amawnian rather than lower-Amawnian forms, may result from such eastward movements wellsouth of the Amawn. Absence of certain eastern birds north of Bahia may also result from south-central crossings.Subspeciation is frequent across dry zones in northeastern Bahia but also in southern Bahia. Various other easternbirds are related to Andean or Argentine groups, probably crossing through northern Argentina and central Para-guay rather than along forested edges of Bolivian-Brazilian cerrado plateaus. Andean birds tend to occur northalong uplands iri central Brazil, even to the central plateau or northeast Brazil. A few other forms seem closerto Guianan than to Andean ones. H not the result of central Andean rapid evolution (.leapfrog speciation.), cros-sing via former lower Amawnian dry forests is indicated. Amawnian and Andean forms may invade the .islands.of Central America and eastern Brazil, as if central species-rich wnes .export. species in .taxon cycles.. Invasionof the Andes from southeast Brazil may have occurred occasionally, just as northward spread of central Brazilianopen-wne species has been proposed. Accepted 1~ November 199L

Key words: Birds, Brazi~ Atlantic forests, cerrado, caatinga, zoogeography, taxon cycles, leapfrog speciation, speciation,ring species, Andes, Amazonia.

ones. Here, I analyze in preliminary fashion thezoogeographic relationships of eastern Brazilianforms, especially the forest species.

INTRODUCTION

Atlantic forests of Brazil are separated from Ama-zonian and Andean ones by a zone of semiopensavannas ("cerrado") and dry woodlands, the"chaco" southwestward and "caatinga" to thenortheast. The avifauna of eastern Brazil is rich,while many genera, species and subspecies areendemic (Haffer 1974), as is the case for plants(Smith 1962), reptiles and amphibians (Müller1968) and other fauna (Müller 1973). Birds ofeastern Brazil are often related to Amazonianor (Miranda-Ribeiro 1906, Sick 1985a) Andean

METHODS

Pinto (1944, 1978), Peters et al. (1934-1986), Bel-ton (1984, 1985), Sick (1985b), Teixeira et al.(1987) and Silva (1989) were the main sources.Yoshika Oniki and I have concentrated our stu-dies on sao Paulo birds since 1975, but have donesome work to Rio Grande do Sul, Paraiba, andMato Grosso after earlier field studies in Maran-

hao, Ceará, Amazonia and the Andes. Here I ex-clude birds found only Ln the Uruguaiana regionof Rio Grande do Sul, marine birds, and non-breeding species. Water and cerrado-caatinga(open-zone) birds are considered separately.

RESULTS.

Water birdsWidespread herons, ducks, and other birds ofla~e marshes and rivers are in Appendix 1. Alt-hough several show Andean and southern rela-tionships ("S"), others cross central Brazil. Re-markably little speciation or subspeciation (ex-cept northern versus southern Chloroceryle ame-ricana) is involved in eastern Brazil, althoughAndean speciation has occurred (Fjeldsa 1985).Mergus octosetaceus might be included here, but Iprefer to discuss it later as a "forest" bird.

Open-zone birdsBirds of the semiopen formations of centralBrazil, caatinga to the northeast and cerrado tothe southwest, meeting on the west the swampyBananal and Pantanal and grading into the Para-guayan and Argentine chaco, have been consid-ered by several authors (Cracraft 1985, Haffer1967, 1974, Short 1975). Included in the central-Brazilian part of this region. are gallery wood-lands, which become rather extensive woodlandsof Atlantic type at high elevation in south-western Goias (the "Mato Grosso de Goiás"),plus dry woodlands along the northern edgeof the Brasília plateau, to Maranhao and Bahia("arboreal caatinga" of the "Mangabeiras bridge",noted below).

I have included 166 species (Appendix 2) astypical of the semiopen zones of central Brazilrather than of eastern Brazil, although many ad-vance well into eastern Brazil in cerrado patches,in dry zones near Salvador, and wherever humanshave cut forests. Since these or related birds oftenextend southwest into the open pampas of Ar-gentina, thence northward along the Andes oralong eastern Brazilian mountains at high eleva-tions or in rain.shadow dry valleys, Andean/Argentinian relationships are common. In dryclimatic cycles (Snethlage 1928, Haffer 1967,1974), the open or semiopen formations con-nected, via Marajó Island and Santarém, to thesavannas of Amapá, Surinam, and Venezuela-Co-lombia. Haffer and others have considered the

birds of northern savannas as recent derivativesof the Brazilian cerrado birds, with littleendemism and several missing species (Cracraft1985 does not list a center of endemism in the

llanos.).Some speciation has occurred within the Bra-

zilian part of the semiopen zone, principally bet-ween deciduous caatinga and evergreen if patchycerrado: Bucco maculatuslchacuru, Penelope jacu-

caca/ochrogaster, Anodorhynchus leari/hyacinthi-nus, Coryphospingus pileatuslcucullatus (these oc-cur northward, too ), Phaethomis gounelli/pretrei,Furnarius figulus/leucopus, Phacellodomus rufi-

frons/ruber, He11Jsilochmus pectoralis/longirostris,Euscarthmus meloryphus/rufomarginatus, Casior-nis fusca/rufa, Empidonomus aurantioatrocrista-tus/varius. Certain of the preceding cases andPhaethomis subochraceus/nattereri, Poospiza me-lanoleuca/cinerea, Melanopareia maximiliani/tor-quata, and Suiriri suiriri/affinis are possible cha-co/cerrado examples. Emberizoides herbicola ofcerrado is related to E. ypiranganus in montanemarshes southeastward, Polystictus pectoralis to Rsuperciliaris in high open zones, Thamnophilustorquatus to ruficapilla, Knipolegus lophotes to ni-gerrimus, Neopelma pallescens of woodlands toaurifrons of upland woods, Pseudoleistes guira-huro to virescens (both marshy grassland), Cyano-corax cyanopogon to chrysops (woodlands), Helio-master furcifer to squamosus, Crax fasciolata toblumenbachii (woodlands), Schistochlamys mela-nopis to ruficapilla (scrub), plus a swarm of per-haps polymorphic seedeaters on the marshy tran-sition to pampas: Sporophila bouvreuil to S. hypo-xantha, S. ruficollis, S. palustris, S. melanogaster,S. cinnamomea (and several Argentine or outsideforms, S. zelichi, S. hypochroma, S. nigrorufa, S.minuta, S. castaneiventris). Anodorhynchus lea-ri/glaucus are a similar example.

Speciation between caatinga/cerrado and Ar-gentina or the Andes occured in 21 cases: Nothu-ra minor/darwinii or maculatus/darwinii, Theri-sticus caudatus/melanopis, Buteo albicaudatus/poecilochrous, Har¡ryhaliaetus coronatuslsolitarius,Milvago chimachimalchimango, Vanellus chilensis/resplendens, Gallinago undulatalstricklandii, Co-lumba picazuro/maculosa, Colibri serrirostris/coruscans, Colaptes campestris/rupicola, Picoidesmixtus/lignarius, Geobates poecilopteralcunicula-ria, Pseudoseisura cristata/lophotes, Automolusrectirostris/erythrocephalus, Phyllomyias reiseri/

EAST BRAZILIAN BIRD ORIGINS

cases fit no pattern, one isolate in Rio de Janeiroand one in eastern sao Paulo as well as a main di-vision across the "Mato Grosso". Laterallus railsand Phacellodomus in marshes have also beenhard to place in preceding sections.

Strong subspecies occur in differing placesnorthward ( Crax fasciolata, Picumnus limae,Gnorimopsar chopi), southward ( Colaptes campe-stris), and near western sao Paulo (Picoides mix-tus, Philydor dimidiatus) as well as in S. bouvreu-il. Arremon jlavirostris, almost an open-wnebird, has strong subspecies near western saoPaulo, and Cyanocorax chrysops/cyanopogon meetthere.

Forest speciesNorthem-crossing widespread species. Of the 870species of east-Brazilian birds under considera-tion (647 when the 57 water and 166 open-wnespecies are excluded), 8 scrub, mangrove and coa-stal birds probably carne along the coast, viaexisting habitat: Nycticorax violaceus, Eudocimusruber, Buteogallus aequinoctialis, Rallus longi-rostris, Aramides mangle, Columbina passerina,

virescens, Phylloscartes roquettei/jlaviventris, Xol-mis cinerea/coronata, Knipolegus franciscanus/aterrimus, Alectrurus tricolor/risoria, Anthus nat-tereri/hellmayri, and Basileuterus leucophrys/leu-coblepharus. These examples suggest an earliercorridor of wet forest separating the open belt innorthern Árgentina or nearby (see below).

Directl'on of relationship is uncertain in Pene-lope (jacucaca and ochrogaster via Andean obscuraor Amazonian jacquacu?) and Xiphocolaptes (fal-cirostris via southern albicollis or central-nor-thern promeropirhynchus?).

Speciation between caatinga-cerrado and Ama-zonian semiopen zones, forest edges or the llanosoccurred in 20 cases: Cathartes burrovianus/me.lambrotus, Crax fasciolata/globulosa, Laterallusxenopterus/melanophaius, A ratinga cactorum/per-tinax, Caprimulgus candicans/cayennensis, Hy-dropsalis brasiliana/climacocerca, Polytmus guai-numbi/theresiae, Lophornis magnifica/stictolo-pha, Hylocharis chrysura/sapphirina, Picumnuspygmaeus/varzeae, Synallaxis scutata/kollari, For-micivora melanogaster/grisea, Sublegatus mode-

stus/obscurior, Stigmatura budytoides/napensis,Neopelma pallescens/sulphureiventer, Tyrannusalbogularis/melancholicus, Polioptila dumicola/plumbea, Paroaria dominicana/gularis, Saltatorcoerulescens/maximus, and Ammodramus hume-

ralis/aurifrons.In genera, about as many are Amazonian (54)

as Andean/Argentine (52) ("F" and "G" in Ap-pendix 2). Of species without obvious relatives,32 seem closest to Andean/Argentine and 24 toAmazonian forms ("S" and "N" in Appendix 2).

Subspeciation (asterisks in Appendix 2) is fre-quent in the chaco and Argentina (16 cases, plus3 just northward in Mato Grosso do Sul), theAndes or Peru (16 cases), and in the open zonesof Amazonia or the llanos (31 cases, plus one onthe Serra do Cachimbo). Internal subspeciation("I" in Appendix 2) occurs in 39 species, in fourof them twice and in Sporophila bouvreuil threetimes. The major patterns are of subspeciationacross the "Mato Grosso de Goiás", or region,some 19 cases, or across the dry wood1ands of theTocantins-Bahia region ("Mangabeiras bridge"),13 cases. These cases suggest the former existenceof major forest barriers in these regions (seebelow). Nine cases occur near western sao Paulo,where the Atlantic forest dips into Paraguay asfar as Asunci6n. The marshy-prairie S. bouvreuil

WILLIS

Mimus gilvus, and Conirostrum bicolor. Nonehave subspeciated. Noctilio fishing-bats have asimilar coastal distrib~tion (Emmons 1990), asdo marine birds not considered here.

Thirteen other widespread birds range ineastern Brazil only to the Pernambucan region,north of the sao Francisco River: Crypturellusstrigulosu.-, Mitu m. mitu, Glaucidium hardyi,Chaetura spinicauda, Pteroglossus inscriptus (Fig.1 ), Dendrocincla fuliginosa taunayi, Dendrocolap-tes certhia, 1hamnophilus aethiops, pyriglena leu-conota pernambucensis, Cercomacra laeta sabinoiby R. Blerregaard, pers. comm.), Zimmerius gra-

cilipes, Hemitriccus zosterops naumburgae, andProcnias averano. In the five indicated cases, localsubspecies are different from Amazonian ones.Unless these 13 came from the southwest, per-haps with later extinction in Bahia, a northerncrossing seems possible. H. Snethlage (1928: 539)suggested that earlier dry-forest links allowedR averano to cross to or from Venezuela. Somemammals (Emmons 1990) show a similar pat-tern: a pygmy anteater, a howler monkey, Ptero-notus bats.

Teixeira et al. ( 1987) show that Cissopis I. leve-riana of northern Amazonia occurs in Alagoas,

4


likely to have been a bridge for Amazonian birdsunless climates were much warmer, there wereon~e large forest areas in northern Minas Geraisas far south as Januária, forests which could haveformed a bridge (the "Januária bridge") from theParana basin to the headwaters of the rios Jequi-tinhonha and Doce. Before deforestation, thesewoodlands would have showed up on satellitephotos. We know this zone from 1977-78 (Wil-lis & Y. Oniki, in press) and found the habitateast of Brasília rather scruffy cerrado, the arbore-al caatinga near Januária rather patchy, and themountain ridge to the east (Serra do Cip6) ratheropen and full of southern birds nowadays. Ho-wever, with wetter forests in warm and humidepochs, Amazonian birds could have crossed theJanuária bridge directly to Espirito Santo andspread north to Bahia, rather than south fromBahia as suggested by the Mangabeiras hypothe-SlS.

not C. I. major of the south. Veniliomis affinisruficeps occurs from Amazonia to Alagoas, then'v: a. affinis south to Ri~ de Janeiro. Dysithamnusmentalis emiliae occurs from Pará to Alagoas, D.m. mentalis southward.

Some species that now range coastally southof Alagoas, such as 'v: affinis, may aIso have en-tered fron1 the north. Several mammals, notablybats and two opossums, have this completelycoastal pattern (Emmons 1990). Pteroglossus ara-cari (Fig. 2) shows one such pattern, where P. ca-stanotis ~eems to block more southern entry.This "ring" pattern, discussed later, brings us tothe next category.

Central crossing. As Haffer (1974) suggested,most east-Brazilian species of wide distributionprobably crossed the central Brazilian semiopenzones. P. castanotis (Fig. 2), though a forest-edgebird, wanders in groups across wide clearings andopen cerrados between gallery woodlands, easilyreaching eastern Brazil today. Other speciesrarely leave forest, and must have needed forest tocross in wet epochs.

A composite satellite photograph of world ve-getation (Van Sant & Warren 1991) clearly showsa dark green dry-forest zone up the Grande andPreto rivers just south of the Chapada das Man-gabeiras in northwestern Bahia, aImost reachingAmazonian forests just northwest. On flights toand from the I Congress of Brazilian Ornitho-logy in Belém in July, 1991, via Brasília, I notedforests on north escarpments at the border ofGoiás and Tocantins states as a possible linkwestward (the "Parana depression", studied by Sil-va 1989 and authors cited therein; in 1989 I sawforests below the nearby Parque Nacional Cha-pada dos Veadeiros, on a visit courtesy of R. B.Cavalcanti), but then a very long stretch of drycerrados aIong the unforested Tocantins, and fi-nally gallery forests on side rivers south of Ara-guaiana. I propose that the link northwesternBahia/ Amazonia may have been forested in wetclimatic cycles, and that this formed a "Manga-beiras bridge" for forest bird movement from theAraguaiana and/or Parana regions.

The Mangabeiras bridge (Fig. 3) is not theonly possible crossing between east and west.The "Mato Grosso de Goiás", now deforested,does not show up on satellite photographs.While it now seems an area of highland birds, un-

Some 318 east-Brazi1ian birds fa11 in the "cen-tral" group: of these, 92 probab1y entered north-ward, on the Mangabeiras bridge (whether viathe southern escarpments or Araguaina region),21 southwestward 1ike P. castanotis, and the restcross wide1y or cannot yet be assigned. Mammalssuggesting the Mangabeiras pattern are a sloth,an agouti, a marmoset (?), and severa1 bats, whi1ethe southwestern "Mato Grosso de Goiás" or Ja-nuária pattern inc1udes another agouti, a how1ermonkey, and other bats (Emmons 1990). One 10-wer-Amazonian opossum ( Caluromys philander)today even traverses southward over the cerrado,east of the Araguaia, to Rio de Janeiro/Bahia.

In some bird cases, the pattern is of a "Madei-ran" type, first noted in Heliothryx aurita: thePurus-Araguaia H a. auriculata reappears inBahia to sao Pau10, whi1e another subspecies(H a. phainolaema) seems c10ser geographica11yon the 10wer Amazon. Hinke1mann (1989) re-gistered Phaethomis malaris (?) margarettae ofcoastal Brazi1 as more 1ike P. m. (?) bolivianus ofthe upper Madeira-Tapajoz than 1ike P. supercilio-sus muelleri of the 10wer Amazon. We shou1dwatch for other possib1e examp1es, which seemto indicate entry into eastern Brazi1 via theescarpments north of Brasí1ia.

I shou1d note that 1ittle is known aboutspecies or subspecies distributions in northernMato Grosso, and that certain Madeiran birds go

further east in the headwaters than they do alongthe Amazon. In Hylophylax poicilinota, the Ma-deiran subspecies swirfgs eastward south of Ama-zon subspecies. Haffer & Fitzpatrick (1985)show a tongue of higher rainfall from Peru acrossthe southern Amazonian headwaters, separatedfrom the Amazon-mouth Guianan tongue ofhigh rairtfall. However, the wide cerrado zone ofthe central Tocantins today suggests that connec-tion with the Mangabeiras bridge was south-y;rard, at the Brasília escarpments.

Several Amazonian or widespread species,found in Bahia or southward, do not occur in thePernambucan region. Mammals apparently pre-sent only north to Bahia include paca, Oxymycte-rus mice, tapir, jaguar, margay, puma, otter, coati,a bush dog, and a free-tailed bat (Emmons 1990).If not a result of undercollecting, recent humaninterference, or past extinctions northward, so-me birds may have been unable to spread northfrom B~hia or have not yet done so. It is not ne-cessary to presume that these birds ever occurredin the Pernambucan region, for they could havecrossed along the Mangabeiras bridge in north-western Bahia or southward.

Southern cro55ing. Miranda-Ribeiro (1906) andSick (1982a) emphasized that some easte¡;n birdsextend to the Andes or have close relatives there.The former, based on the connection, evenargued that Scytalopu5 speluncae must be thesame species as Andean S. magellanicu5. Evenexcluding this case as unlikely, I count 79 wide-spread species that extend from eastern Brazilinto Argentina or along the Andes (Appendix 3,including Bubulcu5 ibis as a species that carne thisway recently), 53 to or along the Andes. It is truethat 8 of the 53 (19 of 79) barely enter easternBrazil in the south (Rio Grande do Sul, "G").Among mammals (Emmons 1990), only oneagouti suggests this pattern or a very south-of-center one, but she probably excludes Andeanand southern forms.

Endemic speciesWith out5ide allopatric 5i5ter 5pecie5. In 49 cases,the outside sister species of an endemic ( east Bra-zilian region) bird could be identified. In fourcases, the crossing seems to have been northward,in twelve ¿ases southward, and in 33 cases cen-trally. Among the 33, 8 seemed results of a

crossing rather southwestward and 18 via theMangabeiras or Januária regions, northward.

The four possible northern cases are Tinamussolitarius, Automolus leucophthalmus, Iodopleurapipra (Pernambucan subspecies showing some in-termediate features to L tao, A. infuscatusand I.isabellae) and Coccyzus euleri (C. americanus ismigrant from North America).

These could have come by the Mangabeirasbridge rather than coastally, of course. SinceCeará montane birds are often southeasternrather than Amazonian, one can even doubt anycoastal movement of forest birds.

North-central cases are (eastern bird first, anygeographically intermediate species being indi-

cated) Crypturellus noctivagus/erythropus, Leucop-ternis lacernulata/kuhli, Leucopternis polionota/albicollis, Odontophorus capueira/gujanensis, Craxblumenbachii/fasciolata/globulosus, pyrrhura leu-cotis/pfrimeri/picta (species according to Silva1989), Chlorostilbon aureoventris/mellisuga, Ma-lacoptila striata/rufa, Selenidera maculirostris/

gouldii, Colaptes melanochloros/punctigula, Lepi-docolaptes squamatus/albolineatus, Myrmotherulaurosticta/longipennis, Myrmotherula jluminen-sis/iheringi, Cotinga maculata/cotinga, Xipholenaatropurpurea/lamellipennis, Corythopis delalan-di/torquata, Mionectes rufiventris/macconnelli,and Pitylus fuliginosus/grossus.

Examples of a possible Argentine crossing areAmazona pretrei/tucumana, Pulsatrix koeniswal-diana/melanota, Otus atricapillus/hoyi, Cinclodespabsti/fusca, Chamaeza meruloides/turdina (Wil-lis, in press), Elaenia parvirostris/albiceps, Elaenia

mesoleuca/strepera, Cyanocorax caeruleus/cyano-melas, Haplospiza unicolor/rustica, Amaurospizamoesta/concolor, and Cacicus chrysopterus/leuco-ramphus. The new Asthenes luizae from MinasGerais has southern relatives (Vielliard 1990).

Central-southern sister species are Pipile jacu-tinga/pipile, Ara maracana/aurocapilla, pyrrhurafrontalis/devillei, pyrrhura cruentata/molinae, Scle-rurus scansor/albigularis, Myiornis auricularis/albiventris, Turdus amaurochalinus/ignobilis and

Sporophila falcirostris/schistacea.Difficult to assign, but central or central-

northern are Phaethornis pretrei/augusti, Celeus

jlavescens/lugubris, Dendrocolaptes platyrostris/pi-cumnus, Xiphocolaptes albicollis/falcirostris/pro-meropirhynchus, Cercomacra brasiliana/ferdinan-

WILLIS

Seven groups of Andean relationship are Lep-tasthenura setaria/striolata as forms of platensis,Synallaxis ruficapilla/infuscata as forms of super-ciliosa, Cranioleuca semicinerea/pallida/obsoletaending in pyrrhophia from Rio Grande do Sulwest, Scytalopus novacapitalis/speluncae to An-dean magellanicus, Phylloscartes ceciliae/sp. nov.(Willis & Oniki, MS) to Andean/southeasternP. ventralis, Hemitriccus mirandae/kaempferi/Andean cinnamomeipectus, and Hemitriccus diops/obsoletus to Andean jlammulatus.

In addition, 1halurania glaucopis/watertoniare related to furcata, which reaches the regioncentrally; Veniliomis spilogaster/maculifrons mayrepresent the widespread affinis, pyriglena leucop-tera/atra the Amawnian-Pernambucan leuco-nota, and 1angara preciosa/peruviana local formsof widespread cayana.

Thirteen other eastern groups are at themoment hard to relate to outside birds: Augastesscutatus/lumachellus may be related to AndeanSchistes, Phacellodomus erythrophthalmus/ferrugi-neigula to cerrado ruber; but Philydor novaesi/atricapillus, Mackenziaena severa/leachi, Terenura

sicki/maculata, Dryrhophila ferruginea/rubricol-lis, Myrmeciza ruficauda/loricata/squamosa, Scy-talopus psychopompus/indigoticus, Merulaxis stre-semanni/ater, Tijuca condita/atra, Ca1pomis me-

lanocephala/cucullata, Hylophilus amaurocepha-lus/poicilotis, and 1angara cyanoventris/desmare-sti are of uncertain relationship, mostly ratherAndean.

Endemic without obvious relatives

Eighty-eight of 229 east-Brazilian regional species

have no obvious relative. Some 24 represent en-

demic genera (plus Augastes, Mackenziaena, Me-

rulaxis, Tijuca, and Carpornis above): Triclaria,

Macropsalis, Eleothreptus, Ramphodon (Amazo-nian Glaucis ?), Melanotrochilus (Amazonian Flo-

risuga), Leucochloris (Andean Schistes ?), Stephano-

xis (Andean?), Aphantochroa (Andean Haplophae-

dia ?), Clytolaema (Andean Heliodoxa ?), Baillo-

nius (Andigena), Oreophylax (Andean Schizoeca),

Hypoedaleus, Biatas, Rhopornis (Amazonian Pyri-

glena ?), Psiloramphus, Calyptura, Ilicura, Muscipi-

pra, Satrapa, Stephanophorus (Andean Anisogna-thus?), Orchesticus, Orthogonys (Andean Chloro-

thraupis ?), pyrrhocoma (Andean Thlypopsis), and

Donacospiza (Andean Poospiza).

di, Sporophila collaris/americana, and Euphoniapectoralis/ ru fi ventris .

"Reinvasion" of siblingIn 31 pairs, an endemic species represents a wide-spread bird that reaches eastern Brazil. In manyof these pairs, the members are in contact oralmost meet. In one case, the outside bird hasbeen regist'ered as a northern entrant, in threecases as a southern entrant, and in 27 cases cen-trally. Among the 27, six seem north-central, onesouth-central.

Amazon-mouth Dendrocincla fuliginosa meetsor nearly meets its counterpart D. turdina acrossthe lower sao Francisco river. To the south Phily-dor lichtensteini meets P. rufus, Saltator maxillo-sus meets aurantiirostris and Embernagra longi-cauda meets platensis as "Andean" pairs.

North-centrally, Harpagus diodon/bidentatus,Glaucis dohrni/hirsuta, Lipaugus lanioides/vocife-

rans, Schiffornis virescens/turdinus, Chiroxiphiacaudata/pareola, and Polioptila lactea/plumbeaare pairs in which the second Amazonian bird re-aches east Brazil. In Brachygalba tridactyla/lugu-bris the crossing was probably south-central.Central but unspecified are Leptodon forbesi/cay-anensis, Aramides saracura/cajanea, Brotogeris ti-

rica/versicolurus, Streptoprocne biscutata/zonaris,Ramphastos dicolurus/vitellinus, Campylorham-

phus falcularius/trochilirostris, Synallaxis spixi/hypospodia, Dysithamnus stictothorax/mentalis,Formicivora serrana/grisea, 7Odirostrum polioce-

phalum/cinereum, Thryothorus longirostris/leuco-tis, Basileuterus hypoleucus/culicivorus, Hemith-raupis ruficapilla/guira, Dacnis nigripes/cayana,Nemosia rourei/pileata, Thraupis cyanoptera/saya-ca, Ramphocelus bresilius/carbo, Tachyphonus co-ronatus/rufus and Saltator similis/maximus.

Multiple endemic speciesIn some 29 cases, there are two or more endemicspecies in eastern Brasil, which are either sympa-tric or allopatric. Amazona brasiliensis/rhodoco-rytha and extralimital dufresniana are coastal-woodland birds, following the northern route.Tangara seledon/fastuosa and Amazonian chilen-sis suggest a north-central route, as do 7Ouit mela-nonota/surda to northern purpurascens, Aratingaauricapillus/jandaya to solstitialis and Hylopezusnattereri/ochroleucus (Vielliard & H. Camargo,MS) to berlepschi.

8


groups is not significantly different, but the twodiffer from the Andean group (Chi2). Only 7 ofthe 23 other southern birds, extending only toArgentina, have subspeciatedj Serpophaga sub-cristata has a strong subspecies (S. s. munda)westward. Of birds that cross northward, onlyMitu mitu has strong subspecies; centrally, Mi-crastur ruficollis, Brotogeris versicolurus, Ciccaba

virgata, Nyctibius leucopterus, Lurocalis semitor-quata, Chlorestes notatus, Baryphthengus rufica-pillus, Bucco macrorhynchus, Colaptes melano-chloros, Formicarius colma, Myiodynastes macula-tus, Tityra cayana, Arremon taciturnus, Arremonjlavirostris, Habia rubica, and Icterus cayanensiscome to mind.

Internal subspeciation within eastern Brazilhas occurred in 70 of the 340 northerly and cen-trally crossing species, only 6 of 78 southerlycrossing birds, and in barely 37 of the 229 en-demic east-Brazilian species (113 cases in 647).

Subspeciation within eastern Brazil is com-mon (17 cases) across the dry region of the lowersao Francisco River and in southern Bahia (47cases). Strong subspecies northward are in Tina-mus, Iodopleura pipra, Turdus albicollis, Automo-lus leucophthalmus, and Cissopis; Sittasomus,Contopus and Crypturellus noctivagus are ratherdistinct in the south-Bahian break. A dry zonealmost reaches the coast at Salvador, along thelower Paragua~u River, and separates Pyriglenaatra from R leucoptera of the southeast (Willis &Oniki 1984). However, some subspecies seem tocross this gap and change in southern Bahia; de-tails need study.

In some cases, the wet Serra do Mar of south-eastern Brazillongitudinally separates coastal andinterior subspecies or species of less wet forests:Amazilia fimbriata nigricauda inland and A. ftephrocephala coastally in dune ("restinga") vege-tation, 7hryothorus leucotis inland and T longiro-stris coastally, Phylloscartes ventralis and R sp.nov. (Willis & Oniki, MS), etc. Formicivora ant-birds have speciated (E erythronota) and subspe-ciated (Gonzaga & Pacheco 1990) coastally indry patches, Reptile, frog, and insect papers sug-gest many local barriers, dry zones in dunes oracross to the coast for dry-woodland groups, wetzones coast:wise for humid-zone groups.

In birds, subspeciation parallel to the coast israther common, somewhat in the dry wood-lands of the interior versus the wetter coastal~

Since Augastes seems Andean, Tijuca seemsrelated to Amazonian Lipaugus and Caryornis toAndean Pipreola (Snt>w 1982), Eleothreptus issouthern, and the rhinocryptids (related to Psilo-ramphus and Merulaxis) generally Andean, moreendemic genera have Andean relatives than Ama-zonian. This may indicate an old Andean forest

connectiDn.Among 61 local endemic species, 30 with

possible Andean links (Mergus octosetaceus, Ama-zona vinacea, Cypseloides fumigatus, C. senex, Pi-cumnus nebulosus, 7hripophaga macroura, Phily-dor amaurotis, Xenops contaminatus, Drymophilasquamata, D. malura, D. ochropyga, D. genei, Myr-motherula unicolor, Phyllomyias fasciatus, R grise-ocapillus, Hemitriccus nidipendulus, H orbita-tus,H furcatus, Phylloscartes sylviolus, R difficilis,R eximius, R paulistus, R oustaleti, Turdus rufiven-

tris, Tangara cyanocephala, Euphonia chalybea,7hraupis ornata, Poospiza lateralis, R thoracica,Carduelis yarrelli) about equal 31 with seeminglyAmazonian links (Laterallus leucopyrrhus, Pio-nopsitta pileata, Chaetura andrei, Phaethornis eu-rynome, R idaliae, R squalidus, Heliomaster squa-mosus, Trogon surrucura, Piculus aurulentus, Me-

lanerpes jlavifrons, Dryocopus galeatus, Campephi-lus robustus, Lepidocolaptes fuscus, Synallaxis cine-rascens (via «Poecilurus» scutatus), Phacellodomus

dendrocolaptoides, Philydor leucophrys, R fuscus,

Myrmotherula gularis, Dysithamnus xanthopterus,Formicivora erythronota, E iheringi, Conopopha-ga melanops, C. lineata, Procnias nudicollis, Pipri-tes pileatus, Neopelma aurifrons, Platyrinchus leu-coryphus, Attila phoenicurus, A. rufus, Sporophilafrontalis, and Schistochlamys ruficapillus). SomeAndean birds may have crossed Amazonia viathe Guiana highlands, however (see below).

Endemic subspecies in eastern BrazilSubspeciation is frequent between the Andes andeastern Brazil -of 53 Andean/eastern species, allbut Bubulcus ibis, Accipiter striatus, Buteo leucor-rhous, Rallus nigricans, R. sanguinolentus, Asiojlammeus, Aegolius harristii, Notiochelidon cyano-leuca (subspeciation southward), Euphonia mu-sica, and Trichothraupis melanops have subspe-ciated, or 43 of 53. Strong subspecies occur inStrix hylophila, Turdus nigriceps, and Poospiza ni-grorufa. Subspeciation is less common for cen-trally crossing species, 126 of 318, and for nor-thern species, 5 of 22. Subspeciation in the two

9

WILLIS

number of possible patterns, especially as sugge-sted by plant and reptile or amphibian workers,is great.

Otherwise, open-zone species of birds showlinks with open areas in Argentina and theAndes, as well as to the Amazon open zones andllanos, more or less as proposed by variousauthors.

ones: 23 cases. An additiona1 9 cases separate sub-species of northwestern Bahia from coasta1 ones,but these might also be-grouped with birds twoparagraphs back. All resemble subspeciationacross the open center of Brazil, but are displacedeastward. Strong forms occur in Crypturellus noc.tivagus, Amazilia fimbriata and versicolor, Colap-tes melanothloros, Dysithamnus mentalis, 1Olmo-myias sulphurescens, Contopus cinereus, Cnemo-triccus fuscatus, Myiarchus swainsoni, Pirangaflava, and Cyclarhis gujanensis.

Minor subspeciation breaks occur in saoPaulo ( 14 cases, including rather different formsin Aramides cajanea [D. Stotz, pers. comm.],Otus atricapillus, Stephanoxis, Picumnus cirratus,Xenops contaminatus U.M.C. Silva, pers. comm.],and Lepidocolaptes squamatus) in Espirito Santo(8 cases) and Rio de Janeiro (11; jointly, 19; diffe-rent forms in Myiozetetes cayanensis but othersweak). Only 3 cases are from Paraná and 3 fromSanta Catarina.

Forest birds

Amazonian linksThe analysis of east-Brazilian forest birds suggeststhat a few species entered along the north coast,but that most Amazonian species crossed to orfrom the region off to the south, across what arecalled "central" links.

One likely place for the north-central linkwith Amazonia is via the northern and easternedges of the central Brazilian uplands, notablyvia the dry-forested upper Tocantins and itsbranch, the Parana River of eastern Goiás (Silva1989). These dry forests almost meet north-western Bahian ones (the "Mangabeiras bridge")across the north-south Serra Geral between Bahiaand Tocantins. Connection across the north endof the Serra Geral, from the Chapada das Manga-beiras northwest in dry forests of southern Ma-ranhao and northern Tocantins, is also possible.Contacts in slightly wetter climatic cycles, use ofevergreen "grottoes" like ones we have seen nearBrejo de Amparo ("Brejo Januária") in northernMinas Gerais, or even entry in the green summerfollowed by desperate wandering in the dry win-ter, could provide a link for such species as Neo-morphus geo.lfroyi and others to reach Bahia butnot Pernambuco. Birds could also have used theJanuária bridge, across central Minas Gerais,from the Parana River.

Southwestward, the former semievergreenforest on the uplands of western Goiás ("MatoGrosso de Goiás") could be another link. How-ever, birds and plants of this region are often At-lantic forms, more of an extension of Andeanforms northward than a pathway for Amawnianbirds (Silva 1989). Still, some Amazonian birdslike Saltator maximus do reach Brasília, alongwith Andean birds like Euphonia musica (J. C.Motta Jr., pers. comm.). Several Amawnianmammals seem to occur across the Mato Grosso,but no Amawnian bird species seem to have

DlSCUSSION

Water and open-wne birdsWater birds are remarkably nonspeciose insoutheastern Brnzil, either because they flyeverywhere or never developed large local pOpU-lations due to relative lack of major freshwaterhabitats. Marsh birds have speciated somewhatmore noticeably, at least some rails and icterines.

Open-area birds have speciated rather stronglyin central to eastern Brnzil. Although their wo-geography is somewhat outside the range of thisreport, some cases of speciation and subspe-ciation could have been caused by former forestlinks which are here suggested to transverselyconnect Amazonia or the Andes with easternBrnzil.

The north-central "Mangabeiras" link pro-posed here would have separated caatinga fromcerrado birds, for instance. The possible "MatoGrosso de Goiás" link, Januária to south-centralAndean link, and the northern Argentine linkcould have separated others. A rather similar si-tuation is suggested for subspecies and species ofcoastal birds, where dry links perpendicular tothe coast would separate wet-forest birds at thetime of dry climatic cycles, while longitudinalmontane wet forests would cut coastal dry-wood-land birds off at the time of wet epochs. The

10


Double linksIn somecases, there may have been "ring" forma-tion or double entry into (or exit from) easternBrazil. The Pteroglossus of Fig. 2 are one likelycase of central + north crossings, as are the twosubspecies of Cissopis leveriana and of Rampha-stos vitellinus. Glaucidium hardyi seems Amazo-nian, G. minutissimum Andean, Dendrocinclaturdina calls more like Madeiran D. fuliginosaatrirostris than like other forms of D. fuliginosa,including taunayi of Alagoas. pyriglena leucop-tera females look rather like those of Madeiran P.leuconota maura, while P. leuconota pernambu-cencis is definitely northern. On the other hand,

Sporophila nigricollis/caerulescens show the un-certainties of this type of analysis, for the latteris more likely an invader of Bolivia than thereverse; the center of the ring seems in easternBrazil, with hybrid " s. ardeola" as the link. Ring

forms can be species generators, since any breakin the ring center causes separation of popula-tions.

Guianan linksIn a few cases, eastern Brazil and the northernAndes have similar birds; a more different birdmay live between, on the central Andes: Platy-cichla jlavipes (leucops centrally), Cichlopsis leuco-

genys (Andean subspecies different), Basileuterusjlaveolus (signatus in the Andes). rhese may becases of "leapfrog speciation", but could be caseswhere the species crossed directly from theGuianas to southeastern Brazil, perhaps viamountains like the serra dos Carajás in southernPará or via dry forests at the Amazon mouth(Snethlage 1928). The latter author suggestedcrossing for Procnias averano and Basileuterusculicivorus, this last being another form withvery different subspecies along the Andes.

Taxon cycles and Andean-Amazonian dominanceIt seems possible that eastern Brazilian forests arelike an island, being repeatedly invaded by speciesthat enter from Amazonia or from the Andes inwhat could be considered "taxon cycles". Haffer(1967a,b) has suggested that Amazonian formsare constantly moving into northwestern Co-lombia and Central America, taking the place ofbirds a1ready present, in more or less the sort oftaxon cycle proposed here. If so, the Amazon andAndes could be constantly "exporting species",

crossed. The mammal distribution may be alongthe Januária/Parana connection, of course.

The entry of A.mazonian birds north-cen-traliy, rather than on the coast, is further sug-gested by a few confirmed "Madeiran" sub-species. The possibility that eastern birds aremore closely related to Madeiran or upper Ama-zonian. subspecies or species than to birds fromthe lower Amazon should be kept in mind infuture analyses.

Andean/ Argentine linksThere are many species and genera related toAndean forms, an idea emphasized by Miranda-Ribeiro (1906) and briefly reviewed by Sick(1985a). In most cases these birds occupy mon-tane forests and limited alpine grasslands ofeastern Brazil, even to the uplands of 1000-1600meters in the Brasília region and others east ofthe sao Francisco River in Minas Gerais andBahia. A few reach the interior uplands of Ala-goas and Pernambuco, where Teixeira et al. (1987and elsewhere) have detected species of southernBrazilian birds only a few hundred meters aboveAmazonian lowland species.

Some Andean birds may have moved along asouth-central route, along the evergreen forests ofbases of the central Brazilian "chapadas" or table-lands. Migratory and nocturnal species couldpass along this route even today, as there are largechapadas even west of the Guaporé River, on theBrazilian-Bolivian border, close to the serra dosParecís and other western fingers of the Braziliantablelands. However, there is little evidence ofpassage. A more likely route is via gallery forestsof Paraguay and northean Argentina, direct fromeastern forests near Asunci6n, in wetter climaticperiods than today. M. Nores (1991 Ornitholo-gical Congress in Cidade del Este, Paraguay)noted that gallery forests west along the Pilco-mayo and Bermejo rivers, to within 400 km ofAndean woodlands, have forest birds even today.Other species may have crossed further south,through central Argentina.

Failure to realize the importance of Andeanlinks southeastward may have led to some incor-rect taxonomic decisions, as pointed out by Sick(1985a). For instance, southeastern pyrrhocomaruficeps is probably just another peri-Andean7hlypopsis, while Ca1pornis sp. of the southeastseem like Andean Pipreola.

11

WILLIS

which would either have some competitive ad-vantage over peripheral species or would takeover at climatically unfatorable periods.

It may be that Haffer and I have overlookedpossible reverse casf§., due to our "extra-Amazo-nian" points of view. One could argue that Ptero-glQSSUS aracari (Fig. 2) started in the east and inva-ded the Am'azon, for instance. I found few otherlikely cases northward, but there seem severalpossible cases southward: SpQrophila caerulescens,TrichQthraupis melanQPs, Phibalura jlavirostris,and Batara cinerea may have invaded the Andesrather than the reverse.

In any such hypothesis, of course, high-altitu-de and open-country forms would invade fromtheir own centers of distribution. Fjeldsa (1985)suggests that Patagonian forms invaded the An-des during cool periods after warm epochs elimi-nated local forms. Haffer (1974) and others havesuggested that central Brazil furnishes species forthe llanosnorthward, where areas of habitat are

smaller and extinctions more likely. Old-Worldfamilies invading South America when the Pana-ma isthmus arose in the late Pliocene were mo-ving from their own rich faunas, which collecti-vely overshadow the Amazonian and Andeancenters.

An alternative would be that central regionsreceive a few successful new forms from each ofseveral side regions (such as eastern Brazil), pluslocally evolved forms, and that each side regionthen receives more than it gives.

ACKNOWLEDGEMENTS

I appreciate a fellowship from the Funda~aode Amparo a Pesquisa do Estado de sao Paulo(FAPESP) and Deutsche Akademische Aus-tauschdienst (DAAD) for preparation of the ma-nuscript at the Museum Alexander Koenig atBonn, Germany in 1990. I especially thank Karl-L. Schuchmann for help at the Museum, J. Haf-fer for suggestions on the manuscript and lettingme use his range maps, and J. M. C. Silva for un-published material based on his studies in centralBrazil. Aerial study of the Brasília-Araguaianaaxis was incidentally provided by the Museu Pa-raense "Emílio Goeldi" to attend the I Congres-so Brasileiro de Ornitologia in Belém.

REFERENCES

Belton, W 1984. Birds of Rio Grande do Sul, Brazil.Part 1. Rheidae through Furnariidae. Bulletin Ame-rican Museum of Natural History 178: 369-636.

Belton, W. 1985. Birds of Rio Grande do Sul, Brazil.Part 2. Formicariidae through Corvidae. BulletinAmerican Museum of Natural History 180: 1-241.

Cracraft, J. 1985. Historical biogeography and patternsof differentiation within the South Americanavifauna: areas of endemism. Orn. Monogr. 36:49-84.

Emmons, L. 1990. Neotropical rainforest mammals: afield guide. Chicago.

Fjeldsa, J. 1985. Origin, evolution, and status of theavifauna in Andean wetlands. Orn. Monogr. 36:85-112.

Gonzaga, L.P., & F.F. Pacheco. 1990. Two new sub-species of Formicivora serrana (Hellmayr) fromsoutheastern Brazil, and notes on the type localityof Formicivora deluzae Menetries. Bull. B.O. C.110: 187-193.

Haffer, J. 1967a. Zoogeographical notes on the .non-forest. lowland bird faunas of northwestern SouthAmerica. Hornero 10: 315-333.

Haffer, J. 1967b. Speciation in Colombian forest birdswest of the Andes. Amer. Mus. Novitates 2294:1-57.

Haffer, J. 1974. Avian speciation in tropical SouthAmerica. Publications Nuttall OrnithologicalClub, Cambridge, Massachusetts, No.14, 390 pp.

Haffer, J., & J. W. Fitzpatrick. 1985. Geographicvariation in some Amazonian forest birds. Orn.Monogr. 36: 147-168.

Hinkelmann, C. 1989. Anmerkungen zu .Phaethornismargarettae. und anderen nicht validen neu be-schriebenen Kolibriarten. Trochilus 10: 64-70.

Miranda-Ribeiro, A. 1906. Vertebrados do Itatiaia(peixes, serpentes, sáurios, aves e mam¡feros). Arq.Mus. Nac., Rio de Janeiro 13: 165-190.

Müller, P. 1968. Die Herpetofauna der Insel von saoSebastiao {Brasilien). Saarbrücker Zeitung, Saar-brücken.

Müller, P. 1973. The dispersal centers of terrestrial ver-tebrates in the neotropical realm. BiogeographicaNo.2, 244 pp.

Peters, J. L. et al. 1934-1986. Birds of the world. Vols.2-15, Vol. 1 (2nd edition). Museum of Compa-rative Zoology, Cambridge, Mass.

Pinto, 0. 1944. Catálogo das Aves do Brasil. Segundaparte. Departamento de Zoologia, Secretaria deAgricultura, sao Paulo.

Pinto,O. 1978. Novo Catálogo das Aves do Brasil. Pri-meira parte. sao Paulo, Empresa Gráfica da Revistados Tribunais.

12


Snethlage, H. 1928. Meine Reise durch Nordbrasi-lien. II. Biologische Beobachtungen. J. Orn. 76:503-581.

Snow, D. W 1982. The Cotingas. Oxford Univ. Press,Oxford.

Teixeira, D.M., Nacinovic, J.B., & F.B. Pontual. 1987.Notes on some birds of northeastern Brazil (2).Bull. B.O. C. 107: 151-157.

Van Sant, T., & V. Warren. 1991. Our world, our envi-ronment. Dialogue 93: 16-18.

Vielliard, J. 1990. Urna nova espécie de Asthenes daserra do Cipó, Minas Gerais, Brazil. Ararajuba 1:121-122.

Short, L.L. Jr. 1975. A zoogeographic analysis of theSouth American Chaco avifauna. Bull. Amer. Mus.Nat. Hist. 154: 16~-352.

Sick, H. 1985a. Observations on the Andean-Patago-nian component of southeastern Brazil's avifauna.Orn. Monogr. 36: 233-237.

Sick, H. 1985b. Ornitologia Brasileira, urna introdu~ao.Vols. 1-2. Brasília, Universidade de Brasília.

Silva, J. M. C. 1989. Análise Biogeográfica da Avifaunade Florestas do Interf1úvio Araguaia-Sao Francisco.M. S. Thesis, Brasília, 110 pp.

Smith, L. 1962. Origins of the flora of southern Brazil.Bull. U.S. Nat. 35: 215-249.

APPENDIX 1. Widespread v

Tachybaptus dominicus

Podilymbus podicepsPodiceps rolland S

Phalacrocorax brasilianus

Anhinga anhingaPilherodius pileatus

Ardea cocoi

Florida caerulea

Egretta alba

Egretta thula

Butorides striatus

Nycticorax nycticoraxCochlearius cochlearius

Tigrisoma lineatum

Ixobrychus involucris

Ixobrychus exilis

Botaurus pinnatus

Mycteria americana

Ciconia maguari

APPENDIX 2. Open-wne birds of eastern Brazil. Explanation of letters and asterisk see text.

Rhea americana *GIS Caracara plancus GS Columbina minuta NNothura boraquira GS Milvago chimachima G Uropelia campestrisNothura minor Falco sparverius GIS Anodorhynchus hyacinthinus FNothura maculosa *1 Falco femoralis S Anodorhynchus leari1aoniscus nanus GS Penelope ochrogaster F Anodorhynchus glaucusRhynchotus rufescens *GIS Penelope jacucaca Cyanopsitta spixi FSyrigma sibilatrix GS Crax fasciolata FI Ara nobilis FINPhimosus infuscatus *GS Aramides ypecaha FN Phacellodomus ruber FTheristicus caudatus G Laterallus xenopterus FN Phacellodomus rufi/rons *FCathartes burrovianus F Micropygia schomburgkii F Aratinga cactorum FGampsonyx swainsoni G Cariama cristata GS Aratinga aurea NElanus caeruleus Vanellus chilensis F Aratinga acuticaudata INCircus buffoni S Gallinago undulata pyrrhura pfrimeri FNButeogallus meridionalis FN Columba picazuro FI Amazona xanthops FNParabuteo unicinctus GS 7L'naida auriculata I Guira guira GSButeo albicaudatus F Columbina cyanopis Bubo virginianus GISHarpyhaliaetus coronatus G Columbina picui IS Athene cunicularia GS

13

r and marsh birds in E Brazil. Expl~

Jabiru mycteriaPlegadis chihi S

Platalea ajaja

Rostrhamus sociabilis

Buteogallus urubitingaBusarellus nigricollis

Dendrocygna bicolor S

Dendrocygna viduata

Dendrocygna autumnalis

Coscoroba coscoroba S

Cygnus melancoryphus S

Cairina moschata

Sarkidiornis melanotos

Amazonetta brasiliensis

Anas bahamensis

Anas versicolor S

Anas georgica S

Netta erythrophthalma

Netta peposaca S

Ion of letters and asterisk see text.

Oxyura dominica

Neochen jubata

Chauna torquata S

Gallinula chloropus S

Fulica rufifrons s

Fulica armillata S

Fulica leucoptera S

Vanellus cayanus

Charadrius collarisHimantopus himantopus *

Sterna simplex

Sterna superciliarisRynchops niger *

Ceryle torquataChloroceryle amazona

Chloroceryle americana I

Chloroceryle inda

Chloroceryle aenea

Atticora melanoleuca

Chordeiles pusillus FIN

Caprimulgus candicans F

Caprimulgus maculicaudus N

Caprimulgus hirundinaceus N

Podager nacunda GS

Hydropsalis brasiliana FI

1achyornis squamata FN

Phaethornis gounellei F

Polytmus gua"inumbi F

Colibri serrirostris G

Eupetomena macroura FIN

Lophornis magnijlca F

Heliactin cornuta F

Heliomaster furcifer F

Hylocharis chrysura F

Bucco maculatus *FI

Bucco chacuru

Ramphastos toco FN

Picumnus limae FIN

Picumnus pygmaeus

Colaptes campestris *GI

Picoides mixtus *GI

Melanerpes candidus G

Xiphocolaptes falcirostris F

Lepidocolaptes angustirostris *FIN

Geositta poeciloptera G

Furnarius leucopus G

Furnarius figulus

Synallaxis phryganophila GISSynallaxis scutata *

Synallaxis hellmayriSynallaxis frontalis *

Xenopsaris albinucha F

Phaeoprogne tapera F

Alopochelidon fucata FN

Pseudoseisura cristata GI

Philydor dimidiatus FIN

Automolus rectirostris G

Megaxenops parnaguae

1hamnophilus torquatus GSakesphorus cristatus FNFormicivora rufa FIFormicivora melanogaster I

Herpsilochmus pileatusHerpsilochmus longirostrisHerpsilochmus pectoralisMyrmorchilus strigilatus FINMelanopareia torquata GIAntilophia galeata FNPolystictus pectoralis GCulicivora caudacuta GSPhyllomyias reiseri FPhylloscartes roquettei GElaenia cristata FNElaenia chiriquensis NSublegatus modestus *FStigmatura budytoides *F

Stigmatura napensisSuiriri affinis GIFluvicola nengeta FEuscarthmus rufomarginatusEuscarthmus meloryphus GXolmis velata GSXolmis cinerea IXolmis irupero ISKnipolegus lophotes GSKnipolegus franciscanus SAlectrurus tricolor GGubernetes yetapa GSCasiornis fusca FCasiornis rufaMachetornis rixosus GNeopelma pallescens FEmpidonomus varius FIE. aurantioatrocristatus ITyrannus albogularis FTyrannus savana N

Cyanocorax cristatellus

Cyanocorax cyanopogonMimus saturninus *GISAnthus nattereri GPolioptila dumicola GBasileuterus leucophrys G

Neothraupis fasciataCypsnagra hirundinacea ICompsothraupis loricata GSSchistochlamys melanopis *

Paroaria dominicana FCharitospiza eucosmaCoryphospingus cucullatus *

Coryphospingus pileatusSaltator atricollisSaltator caerulescens FISicalis citrina GSSicalis columbiana NSicalis luteola SSporophila bouvreuil FI

Sporophila hypoxanthaSporophila ruficollisSporophila palustrisSporophila melanogasterSporophila cinnamomea

Sporophila plumbeaSporophila albogularis SPophyrospiza caerulescens SPoospiza cinerea GCoryphaspiza melanotisAmmodramus humeralis GIEmberizoides herbicola GCuraeus forbesiGnorimopsar chopi IMolothrus badius GISMolothrus rufoaxillaris SPseudoleistes guirahuro GPseudoleistes virescens

APPENDIX 3. Andean-Argentine birds in east Brazil. Explanation of letters and asterisk see text.

Bubulcus ibis A Aegolius harristii A Lochmias nematura *ATigrisoma fasciatum *A Glaucidium minutissimum *A Philydor rufosuperciliatus *AlAccipiter striatus A Nyctibius aethereus * A Dysithamnus plumbeus *AGeranoaetus melanoleucus * Caprimulgus longirostris *A Thamnophilus ruficapillus *A

Buteo leucorrhous A Amazilia lactea *A Batara cinerea *AMilvago chimango G Picumnus cirratus *Al Chamaeza campanisona *APenelope obscura *Al Geositta cunicularia *AG Chamaeza sp. *ARallus nigricans A Phleocryptes melanops *A 1.aniisoma elegans *ARallus sanguinolentus A Phacellodomus striaticollis *AG Phibalura flavirostris *APo11'hyriops melanops *A Furnarius rufus *1 Pyroderus scutatus *ANycticryphes semicollaris Anumbius annumbi Xolmis dominicana *Myiopsitta monachus *G Certhiaxis sulphurifera G Knipolegus cyanirostris *Pionus maximiliani * Spartanoica maluroides G Serpophaga subcristata *

Strix hylophila *A Limnornis rectirostris G Serpophaga nigricansAsio flammeus AG Limnornis curvirostris G Phvlloscartes ventralis *


Turdus nigriceps *AAnthus hellmayri *AAnthus furcatus *AGAnthus correndera *GBasileuterus leucoblepharusEuphonia musica AChlorophonia cyanea *APipraeidea melanonota *A1hraupis bonariensis *ATrichothraupis melanops ASaltator aurantiirostris *AG

Phyllomyias burmeisteri *A

Elaenia obscura *ATachuris rubrogastra *A .

Pseudocolopteryx sclateri

1bdirostrum plumbeiceps *Al

Pyrocephalus rubinus *AG

Ramphotrigon megacephala *A

Tachycineta leucorrhoa

Notiocheliáon cyanoleuca A

Cistothorus platensis *A

Cichlopsis leucogenys *A

Platycichla flavipes *A

Paroaria coronata GPasserina glaucocerulea

Embernagra platensisPoospiza nigrorufa *AGPoospiza melanoleuca GSporophila caerulescens *AZonotrichia capensis *AlXanthopsar flavus GAgelaius thilius *AGAmblycercus holosericeus GCarduelis magellanica *A

15

zoogeographical origins of eastern brazilian birds

Documents