Revista Brasileira de Ornitologia, 22(4), 363-373 article December 2014 rvs g s bs b sps h amz

ax M. Fs1,5, M chH2, tms Hbk3,4 iz Ps Fs3

1 Universidade Federal Rural de Pernambuco, UFRPE, Unidade Acadêmica de Serra alhada, Serra alhada, PE, Brazil. 2 Instituto Nacional de Pesquisas da Amazonia, INPA, Manaus, AM, Brazil. 3 Universidade Federal do Amazonas, UFAM, Manaus, AM, Brazil. 4 University o Puerto Rico, Rio Piedras, UPR-RP, Puerto Rico, USA. 5 Corresponding author: [email protected]

Received on 19 November 2014. Accepted on 10 December 2014.

aBStract: Morphological, vocal and genetic studies have shown that the Madeira River and its right bank tributaries delimit populations o primates and . We sequenced the cytochrome b gene (approx. 950 bp) or individuals o three suboscine species, Glyphorynchus spirurus (Furnariidae), Willisornis poecilinotus (Tamnophilidae) and Schiornis turdina (ityridae), on opposite banks o the Madeira River and two o its right-bank tributaries, the Aripuanã and Jiparaná rivers. Phylogenetic hypotheses (parsimony, maximum likelihood and Bayesian analysis) revealed clades that have over 3.1% genetic dierentiation on opposite banks o the Madeira River or G. spirurus, W. poecilinotus and S. turdina, suggesting that this river restricts gene ow among populations o these three species. Te Jiparaná and Aripuanã rivers apparently separate distinct populations o G. spirurus, the smallest species we examined, but not those o the other two heavier bodied species, W. poecilinotus and S. turdina. In G. spirurus our clades with high levels o genetic dierentiation (3.2–5.5%) were ound to be delimited by the three rivers evaluated, whereas in W. poecilinotus and S. turdina no genetic structure across the Jiparaná and Aripuanã rivers was detected. In general, birds that are known to show population structure across the Madeira tributaries (Glyphorynchus spirurus, Hemitriccus minor, Hypocnemis rondoni, Herpsilochmus stotzi, and naevius) have body masses smaller than those o both Willisornis poecilinotus and Schiornis turdina, but some exceptions are discussed. Future studies controlling or several variables are necessary to determine the extent to which body mass is a useul predictor o genetic population structure in understory suboscine .

KeY-WordS: Areas o endemism, body mass, comparative phylogeography, conservation, dispersal rate, suboscine birds.

introduction Recent studies o primates and birds in the Madeira-apajós interuvium (M-), also known as Avian distribution patterns are reasonably well known the Rondônia area o endemism (Cracrat 1985), have and inuential in studies o evolutionary processes. suggested that smaller rivers also limit the distributions Jürgen Haer was one o the frst authors to compile o some taxa, thus orming smaller areas o endemism in bird distribution data to describe biogeographic patterns what was reerred to as “mini-interuvia” (Cohn-Hat et in South America (Haer 1974). His work made a al. 2007). Willis (1969), in a study o birds o the genus great contribution to evolutionary studies and to the Rhegmatorhina, was one o the frst to document complex ormulation o speciation hypotheses in Amazonia. patterns o bird distributions in this area. He discussed Te “centres o species endemism” he described remain the parapatric occurrence o Rhegmatorhina berlepschi largely unchanged in analyses o distribution patterns or and R. homannsi within the M- and suggested that the many avian groups. Tey have been generally accepted in Madeira and apajós rivers have occasionally changed subsequent works, and in the Amazon basin these regions their courses, resulting in the separation o populations are oten delimited by large rivers (Haer 1974; Cracrat and subsequent speciation. Van Roosmalen and 1985; da Silva & Oren 1996). Bird species are usually collaborators (1998) described geographic substitutions separated by the Amazon River and its major tributaries o species in primates o the genera Callithrix and such as the Negro, Madeira, apajós and ocantins rivers Callicebus on opposite banks o small rivers within this (Cohn-Hat 2000; Ribas et al. 2012; D’Horta et al. 2013; interuvium and described a new species o marmoset, Fernandes et al. 2012, 2013, 2014). Similar patterns are Callithrix humilis, that occurs only on the west bank also ound in other Amazonian vertebrate taxa, including o the Aripuanã River. Subsequently, several other bird primates and butteries (Wallace 1852; van Roosmalen et species in this region have been ound to contain vocally, al. 1998; Hall & Harvey 2002), suggesting that rivers are morphologically or genetically distinct populations, important barriers to dispersal. with restricted distributions and geographic substitution 364 Rivers acting as barriers or bird dispersal in the Amazon Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

on opposite banks o Madeira tributaries, such as the rivers along their lower courses are roughly 3.0, 0.8, 0.4 Aripuanã and Jiparaná (or Machado) rivers (Cohn-Hat and 0.3 km, respectively. Each collection point had a et al. 2007; Isler et al. 2007; obias et al. 2008; Fernandes corresponding point located on the opposite bank and et al. 2012, 2013, 2014; Whitney et al. 2013a, b, c, d, thereore in a dierent interuvium. For purposes o e). Similar geographic patterns, with dierent races being sampling and analyses, we suggest the existence o three separated by small Amazonian rivers, have also been mini-interuvia within the M-: Madeira-Jiparaná (MJ), ound or butteries (Hall & Harvey 2002). Aripuanã-Jiparaná (AJ) and Aripuanã-apajós (A) Sardelli (2005) ound genetic dierentiation (Fig. 1). A maximum o 5 individuals per species were (cytochrome b, 500 bp) among morphologically collected at each sampling point, with the total sample indistinguishable populations o the Snethlage’s ody- as ollows: Willisornis poecilinotus (n = 45), Schiornis yrant (Hemitriccus minor) apparently bounded by turdina (n = 23), and Glyphorynchus spirurus (n = 25) the Jiparaná and Aripuanã rivers. Tis study raised the (see Appendix). Specimens were deposited in the bird possibility o the existence o cryptic endemism in mini- collection o the National Institute or Amazonian interuvia, which was subsequently investigated or Research (INPA), Manaus, Brazil, where tissue samples the other three species o passerine birds (Myrmeciza (muscle, heart and liver) were stored in liquid nitrogen hemimelaena, Glyphorynchus spirurus, Hylophylax naevius or molecular analyses. (Fernandes et al. 2012, 2013, 2014). Fernandes (2013) cited in a review a number o publications corroborating the importance o the mini-interuvia, highlighting that these diversity patterns are a key (and possibly unique) Amazonian eature and that despite the act that this fne- scale endemism is well known and recognized among systematists working in the Amazon, it is not taken into account in conservation plans. Fernandes (2013) pointed out that many taxa in this region, including those yet to be given ormal scientifc names, may now be endangered or even extinct. Tus it is o utmost importance to consider species that present this kind o fne scale dierentiation in uture conservation proposals. Te objective o our study was threeold: 1) describe phylogeographic patterns or three species o suboscine passerines across the Rondônia area o endemism; 2) compare these patterns to those o other species know to have populations delimited by the Madeira, Aripuanã and Jiparaná rivers; and 3) investigate the relationship between degree o phylogeographic structure and ecological attributes in the light o the FiGure 1. Collection points and the interuvia sampled. Madeira- riverine barrier hypothesis. Jiparaná (MJ), Aripuanã-Jiparaná (AJ), Aripuanã-apajós (A) and Let bank o the Madeira River (LM).

Material and MetHodS We chose to study these three target species primarily because they are common, easy to collect, and widely Sps s smpg sg distributed across the entire Amazon basin. Although all three are suboscine passerines, they represent three distinct We studied three species o passerine birds belonging amilies and, as such, the results obtained in this study to three dierent amilies: Glyphorynchus spirurus can be assumed to be instances o independent evolution, (Furnariidae), Willisornis poecilinotus (Tamnophilidae), and, thus, support the generality o our conclusions. All and Schiornis turdina (ityridae). We sampled birds three species can be ound in the same habitat (terra rme at 12 sites between the apajós and Madeira rivers and orest) but they dier in a variety o ecological attributes: fve sites on the let bank o the Madeira River (LM; Glyphorynchus spirurus – Tis is a polytypic species Figure 1), with the fnal number o localities sampled per widely distributed in Neotropical lowland orests, species diering among the three species (see Results). occurring in Amazonia, Central America and along the Individuals were collected along the Madeira, Aripuanã, Atlantic coast o Brazil (Ridgely & udor 1994). Marantz Jiparaná and Roosevelt (the latter representing the largest et al. (2003) recognized thirteen subspecies, six o which tributary o the Aripuanã) rivers; the widths o these occur in the Brazilian Amazon. Tree o these occur

Revista Brasileira de Ornitologia, 22(4), 2014 Rivers acting as barriers or bird dispersal in the Amazon 365 Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

within or adjacent to the Madeira basin: G. s. castelnaudii volumes using a Termo Hybaid PCR Express thermal (west o the Madeira River to the Andes), G. s. albigularis cycler under the ollowing conditions: (1) an initial (south-eastern Bolivia and Peru), and G. s. inornatus, denaturing step at 94°C or 5 min; (2) 35 cycles o the which occurs throughout the Brazilian portion o the ollowing: 1 min at 92°C, 1 min at 48°C, and 1 min at Madeira-apajós interuvium (Peters 1951; Marantz et 72°C; (3) a 10-min extension step at 72°C. Following al. 2003). With a body mass ranging rom 10.5 to 21g PCR, correct ragment size and the presence o a single (typically 12.6-14.8 g in central Amazonia; Bierregaard, amplifcation product was confrmed via electrophoresis 1988), this is the smallest woodcreeper (Marantz et al. on 1% agarose gel. Ater amplifcation, the PCR 2003), and it is the smallest o the three species we studied. products were purifed using a salt protocol (Sambrook It occurs in both terra rme and seasonally ooded et al. 1989). Sequencing was perormed by the chain orests (várzea and igapó) (Marantz et al. 2003) and it termination method (Sanger et al. 1977), using a Big is moderately sensitive to environmental perturbation Dye ermination Kit (Applied Biosystems) ollowing (Ferraz et al. 2007). Recently, Fernandes et al. (2013) the manuacturer’s specifcations. Te products o the ound that populations o G. s. inornatus are delimited by sequencing reaction were precipitated with ris-HCl and the Aripuanã and Jiparaná rivers. alcohol, and resuspended in ormamide and resolved by Willisornis poecilinotus – A species endemic to the capillary electrophoresis in an ABI 3130xl automatic Amazon basin, with seven subspecies recognized (Peters sequencer (Applied Biosystems). All sequences have been 1951; Zimmer & Isler 2003). Only one subspecies (W. deposited in GenBank (accession numbers: HM164938 p. griseiventris) is recognized rom the middle and upper – HM165034). Madeira River basin; and there is no evidence o vocal or morphological dierentiation across the Madeira, agm Aripuanã and Jiparaná rivers (Isler & Whitney 2011), although Bates (2000) ound genetic dierentiation Sequences o DNA were visualized and edited using (based on analyses o isozymes) across the Madeira River. the Bioedit program (Hall, 1999). Alignments were Occurs in the understory o terra rme orest, where it is perormed in Clustal X within Bioedit (Hall, 1999). We a regular ollower o army ant swarms (Zimmer & Isler used recommended precautions and are confdent that all 2003). It is larger on average than Glyphorynchus, with sequences represent mitochondrial DNA or the ollowing a body mass ranging rom 15 to 19 g (Zimmer & Isler reasons: (1) DNA was extracted only rom tissue samples, 2003). which have high ratios o mitochondria to nuclei relative Schiornis turdina – Nyári (2007), with no samples to blood or skin samples; (2) no stop codons occurred rom the middle or lower Madeira River basin described within the cytochrome b o any o the sequences; (3) close geographic proximity in the upper Madeira o two sequences contain no insertions or deletions relative genetically distinct orms, showing no obvious vocal or to one another or to other known avian cytochrome b plumage dierences; the author proposed recognizing sequences; (4) sequences in both DNA ragments rom them as distinct species (S. amazona and S. turdina), each individual were identical and unambiguous in their as adopted by the Brazilian Ornithological Records region o overlap; (5) in phylogenetic analyses, no samples Committee (2014). Schiornis turdina (in the polytypic appeared in unexpectedly basal portions o the tree or sense used here) occurs in the understory o terra rme had exceptionally short or long branch lengths, both o and sandy-belt campinarana orests. Body mass averages which, i present, would indicate a ast evolving gene or 31 g (Snow 2004). Tis species is sensitive to orest an early diverged gene (a pseudogene, or example). ragmentation, disappearing rom small orest ragments (Ferraz et al. 2007). Phyg yss

ex, mp sqg dna Phylogenetic analysis o DNA sequence data was perormed using maximum parsimony (MP) and DNA was extracted rom breast muscle (approximately maximum likelihood (ML) via PAUP* 4.0b10 (Swoord 0.2 g) using a standard phenol chloroorm protocol 2002) and Bayesian inerence (BI) implemented in (Sambrook et al. 1989). he mitochondrial MRBAYES 3.0b4 (Hulsenbeck & Ronquist 2001). cytochrome b was amplifed via the polymerase chain Maximum parsimony analysis was perormed using a reaction (PCR) using the primers: orward H16064 heuristic search with the ollowing options: BR branch- 5-ACCARCCGAGAAAYYGG-3, reverse swapping with 10 trees held at each step. Support or L14993 5-AAGGGAAGCCAGCGG-3, nodes was assessed using 1000 bootstrap replicates. both o which were designed exclusively or this project. Maximum likelihood was perormed using the model All amplifcation reactions were perormed in 25 μl parameters determined in the program Modeltest

Revista Brasileira de Ornitologia, 22(4), 2014 366 Rivers acting as barriers or bird dispersal in the Amazon Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

(Posada & Crandall 1998). Te support or nodes in reSultS the likelihood tree was assessed using 500 bootstrap iterations. For BI analyses, two independent runs o We ound signifcant phylogeographic structure among 8,000,000 generations each were perormed; or each run populations within all three study species. Te Madeira our Markov chains were simulated. rees were sampled River clearly separates genetically distinct populations in every 500 generations and the frst 4,000 samples were all o them. Within the Madeira-apajós interuvium, discarded as burn-in. the degree o structure varied among species (see below). Because there is strong evidence that geographic In each species, tree topologies were identical or all our distributions o Amazonian birds are bounded by tree-building algorithms, thus we only show the tree large rivers that orm areas o Neotropical endemism, resulting rom the Bayesian inerence analyses. Te pair- we used individuals rom populations rom other wise genetic p-distance between individuals rom opposite interuvia as outgroups or the three species studied. banks o the three rivers ranged rom 3.1 to 5.5% but the For the analysis o Glyphorynchus spirurus we used two variation within interuvia was low (0.0–0.09%). Results individuals collected in the headwaters o the Negro or each species were as ollows: River (Appendix) as the outgroup. For Willisornis poecilinotus, we used one individual collected in the Glyphorynchus spirurus Solimões-Negro River interuvium and or Schiornis turdina, we used as outgroups one individual collected We sequenced a total o 946 bp or 27 individuals o G. north o Manaus and also one sequence o Schiornis spirurus. Parsimony, maximum likelihood, and Bayesian virescens rom GenBank (accession number AF453816; inerence analyses suggested a genetic structure in the Appendix). orm o monophyletic groups on opposite banks o the Madeira, Aripuanã and Jiparaná rivers, each supported by Phyg vg yss high bootstrap values (MP = 100, ML = 100, BI = 1.00). No barrier eect was ound on opposite banks o the Phylogenetic divergence was estimated in the program Roosevelt River. Parsimony analysis yielded two equally BEAS v1.6.1 (Drummond & Rambaut 2007) using parsimonious trees (length = 170, CI = 0.8235, RI = the coalescent constant population size tree prior 0.9504). From 130 variable sites, 110 were parsimony (Drummond et al. 2002), the uncorrelated lognormal inormative. Maximum likelihood (-ln L = 2012.00097) relaxed molecular clock model (Drummond et al. and Bayesian inerence resulted in a topology very similar 2006), and the HKY (Hasegawa et al. 1985) models to that o the parsimony analysis. Levels o genetic o molecular evolution, including gamma-distributed divergence (uncorrected p-distance) between individuals rate heterogeneity among sites and invariant sites. o dierent clades ranged rom 3.2% (populations o Ater preliminary runs, we adjusted priors and MCMC AJ versus MJ) to 5.5% (populations o LM versus A) operators to assure optimum perormance. o assess the and levels o divergence between individuals within the robustness o estimates and investigate the inuence same interuvium ranged rom 0.0–0.03% (Figure 2). o the tree prior, we also perormed analyses under the Coalescent analyses in the program BEAS indicate a exponential (Drummond et al. 2002) and the Bayesian 6.5 mya (1.9 – 32.7, 90% HDP) divergence between skyline (Drummond et al. 2005) tree priors. o convert populations on the let and right banks o the Madeira divergence time estimates into units o millions o River. years, we used the mean substitution rate o 0.01105 substitutions/site/lineage/million years as proposed by Willisornis poecilinotus Weir & Schluter (2008). For each set o priors, two independent MCMC We sequenced a total o 956 bp or 46 Willisornis analyses were run or 100 million generations, sub- poecilinotus individuals. Te results o parsimony, sampling every 100 thousand generations. Ater a 10% maximum likelihood, and Bayesian inerence analyses burn-in, convergence o parameter estimates was assessed were concordant, thus indicating a strong phylogenetic using the Gelman-Rubin statistic implemented in the signal supported by high bootstrap values (MP = 100, module coda in the statistical package R (R Development ML = 100, BI = 1.00) (Figure 3). Parsimony analysis Core eam 2011). Independent chains were combined, yielded 100 equally parsimonious trees (length = 94, and marginal posterior parameter means and their CI = 0.8723, RI = 0.9634). From 71 variable sites, 41 associated 90% highest probability density intervals (90% were parsimony inormative. Maximum likelihood (-ln HPD) together with eective sample size (ESS) or each L = 1757.25307) and Bayesian inerence resulted in a divergence time estimate were calculated in the statistical topology very similar to that o the parsimony analysis. package R (R Development Core eam 2011). Te level o genetic divergence (uncorrected p-distance)

Revista Brasileira de Ornitologia, 22(4), 2014 Rivers acting as barriers or bird dispersal in the Amazon 367 Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

between individuals o the two clades separated by the interuvium ranged rom 0.0–0.09%. Coalescent Madeira River, RM (right bank o Madeira River) versus analyses in the program BEAS indicate a 2.6 mya (0.8 – LM (let bank o Madeira River), was 3.4% (Figure 3). 13.8, 90% HDP) divergence between populations on the Levels o divergence between individuals in the same let and right banks o the Madeira River.

FiGure 2. Species–area relationships (a) and Bayesian inerence phylogeny (b) estimated or Glyphorynchus spirurus. Numbers at the tips o branches reer to localities where individuals were sampled. Bayesian inerence (BI) posterior probabilities and genetic distance (uncorrected p-distance) values are indicated in the branches. Note grouping o a sample rom the Aripuanã-apajós interuvium (location 15) with those o the Madeira-Jiparaná interuvium (see Discussion).

FiGure 3. Species–area relationships (a) and Bayesian inerence phylogeny (b) estimated or Willisornis poecilinotus. Numbers at the tips o branches reer to localities where individuals were sampled. Bayesian inerence (BI) posterior probabilities and genetic distance (uncorrected p-distance) values are indicated in the branches.

Revista Brasileira de Ornitologia, 22(4), 2014 368 Rivers acting as barriers or bird dispersal in the Amazon Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

Schifornis turdina likelihood (-ln L = 2043.26577) and Bayesian inerence resulted in a topology very similar to that o the parsimony We sequenced a total o 968 bp or 25 Schiornis turdina analysis consensus topology. Te maximum divergence individuals. Parsimony, maximum likelihood and (uncorrected p-distance) between individuals o the RM Bayesian inerence analyses suggested genetic structure on and LM clades was 3.1% (Figure 4). Levels o divergence opposite banks o the Madeira River supported by high among individuals o the same interuvium ranged rom bootstrap values (MP = 100, ML = 98, BI = 1.00) (Figure 0.0–0.3%. Coalescent analyses in the program BEAS 4). Parsimony analysis yielded 48 equally parsimonious indicate a 3.1 mya (1.0 – 15.2, 90% HDP) divergence trees (length = 175, CI = 0.9371, RI = 0.9214). From 159 between populations on the let and right banks o the variable sites, 67 were parsimony inormative. Maximum Madeira River.

FiGure 4. Species–area relationships (a) and Bayesian inerence phylogeny (b) estimated or Schiornis turdina. Numbers at the tips o branches reer to localities where individuals were sampled. Bayesian inerence (BI) posterior probabilities and genetic distance (uncorrected p-distance) values are indicated in the branches.

diScuSSion taxa separated by the same geographic barrier (Ribas et al. 2012). Based on coalescent analyses (see Methods), Strong genetic dierentiation in the ace o highly we estimated mean divergences o 6.5 mya, 2.6 mya and conserved phenotype is at the heart o numerous 3.1 mya between populations on let and right banks o descriptions o “cryptic species” in recent years the Madeira River or G. spirurus, W. poecilinotus and S. (Whitney et al. 2013a, b, c, d, e) and appears to be a turdina, respectively. Te separation o the lineages in all requent phenomenon in the Amazon. In S. turdina, three species o passerines are clearly ancient, all lineages dierentiation on opposite banks o the middle and are diagnosable by multiple molecular synapomorphies, lower reaches o the Madeira River is consistent with and all lineages are parapatrically distributed and likely that detected earlier in the upper Madeira (Nyári represent phylogenetic species. However, it is also clear 2007) and associated with species level taxa. In all that a more detailed analysis evaluating species status and three studied species, the observed molecular groups establishing species boundaries is necessary. are monophyletic and parapatrically distributed, their Irrespective o , the pattern o geographic geographic distributions are delimited by rivers, and variation delimited by rivers is clear or all three taxa the observed phylogenetic divergence between clades studied. Our results indicate genetically distinct on opposite banks o the Madeira River (3.1–5.5%) is populations on opposite banks o the Madeira River. For consistent with interspecifc divergences in other avian all three species analyzed in this study we ound sister

Revista Brasileira de Ornitologia, 22(4), 2014 Rivers acting as barriers or bird dispersal in the Amazon 369 Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

clades on opposite banks o the Madeira River, and or separated by the Madeira River not to be reciprocally G. spirurus, as documented previously (Fernandes et al. monophyletic (Aleixo 2004, Patané et al. 2009, Sousa- 2013), the data urther indicated sister clades on opposite Neves et al. 2013), as recovered herein or G. spirurus, banks o the smaller Aripuanã and Jiparaná rivers. Our W. poecilinotus, and S. turdina, hence supporting a more data thereore reinorce the importance o rivers as complex scenario o dierentiation and a broad range o geographic barriers, and suggest a hierarchical eect in phylogeographic patterns or the same region. which larger rivers divide older clades whereas smaller Despite the importance o rivers or avian rivers are associated with more recent divergences. For dierentiation, even the largest Amazonian rivers are G. spirurus, an individual collected on the right bank o not barriers or all species and smaller rivers are less the Aripuanã River that grouped in the clade Madeira/ likely to be barriers than larger rivers. Tere are several Jiparaná (MJ) provides evidence o upstream gene ow potential explanations or this phenomenon. Molecular across both o the same rivers (Jiparaná and Aripuanã) studies suggest that populations o canopy species are less that delimit dierentiated populations in their lower structured than those o understory birds (Capparella reaches. Since rivers naturally tend to be narrower in 1988; Burney & Brumfeld 2009). Te latter authors the upper reaches, this result suggests that river width is showed that genetic divergence is signifcantly smaller important in determining a river’s likelihood o delimiting across the Andes and two Amazonian rivers (Amazon and distributions and urther strengthens the hypothesis Madeira rivers) in canopy birds than in understory species. o a hierarchical eect o river width in structuring Burney & Brumfeld (2009) urther suggested that there populations (Haer 1974, 1997). is a negative relationship between dispersal propensity Assuming roughly equal rates o substitution, and genetic structure. Species that occupy the understory then Willisornis poecilinotus and Schiornis turdina are supposed to be less eective dispersers, which may populations may have dierentiated across the Madeira be one reason why there are more species o understory River at about the same time; however, Glyphorynchus birds, and that they are more locally distributed. spirurus would appear to have dierentiated much However, we ound dierences in genetic structure earlier. Tis implies that not all sympatric bird taxa among understory species, suggesting that other actors necessarily share the same evolutionary scenario. may also inuence the diversifcation o birds. One Although the Madeira River currently delimits the might also expect the degree o sensitivity to disturbance distributions o the let- and right-bank clades o all or habitat specialization on primary terra rme orest to three species, the Madeira River might not necessarily predict the importance o rivers in driving or maintaining be the primary agent that has driven the observed allopatric dierentiation. Ferraz et al. (2007) analyzed divergence. It may simply represent current limits o thirteen years o capture/recapture data or birds in the distribution or clades that have diverged due to other reserves managed by the Biological Dynamics o Forest abiotic or biotic orces, independent o the ormation Fragmentation Project (BDFFP), located in the Brazilian o the Madeira River itsel. Another non-exclusive state o Amazonas north o Manaus. Tese authors derived possibility is that rates o molecular substitutions are measures o the vulnerability o a species to isolation and 2–3 times aster in G. spirurus than in W. poecilinotus sensitivity due to ragment size. Tese two measures and S. turdina; however, such an elevated substitution reect sensitivity to environmental change. Among the rate appears to be a rare phenomenon in passerine birds, 54 species examined by Ferraz et al. (2007), G. spirurus and has been suggested only or one case o an Old was the least sensitive to the size o the ragment and one World species (Nectarinia humbloti; Warren et al. 2003). o the ten species least vulnerable to isolation. By contrast, Finally, a third explanation suggested previously (Willis S. turdina was among the most vulnerable and most 1969, Fernandes et al. 2012, 2014) is that changes in the sensitive species. Willisornis poecilinotus was not included courses o rivers might conuse the phylogenetic pattern. in the analysis. One would thereore expect G. spirurus, Tere is evidence that the course o rivers in the Madeira the species least aected by isolation and ragmentation, basin changed throughout history, but remained stable to have lower genetic divergence across the rivers than or long periods o time (Latrubesse 2002). Te period the other two species; however, our results contradict o stability could be enough to cause dierentiation the expected pattern. Glyphorynchus spirurus, although until their course was modifed again and became stable occurring in dierent types o orests and not being or another long period o time thus causing both spatial especially sensitive to disturbance, has populations that and temporal incongruences among phylogenies o co- are much more strongly structured than are those o the distributed species (Fernandes 2013). A comparative other two species. In this case, sensitivity to disturbance analysis including additional species and sampling and degree o specialization on primary terra rme orest nuclear markers is likely to shed more light on this issue, were not good predictors o the degree o population but at least three other studies have ound populations genetic structure.

Revista Brasileira de Ornitologia, 22(4), 2014 370 Rivers acting as barriers or bird dispersal in the Amazon Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

Te Jiparaná and Aripuanã rivers separate populations reFerenceS o G. spirurus, the smallest species we examined (average body mass 13.7 g), but not the populations o two other ax, a. 2004. Historical diversifcation o a terra-frme orest species, W. poecilinotus (15–19 g) and S. turdina (30–35.5 bird superspecies: a phylogeographic perspective on the role o dierent hypotheses o Amazonian diversifcation. Evolution, 38: g). Te Jiparaná and Aripuanã rivers also appear to limit 1303-1317. the distributions o populations in other very small birds, Bs, J. M. 2000. Allozymic genetic structure and natural habitat including Hemitriccus minor (Sardelli 2005), Hypocnemis ragmentation: data or fve species o Amazonian orest birds. rondoni (Isler et al. 2007; obias et al. 2008; Whitney et Condor, 102: 770-783. Bg, r. o. 1988. al Herpsilochmus stotzi et al Morphological data rom understory . 2013a), (Whitney . 2013b), birds in terra frme orest in the central Amazonian basin. Revista Hylophylax naevius (Fernandes et al. 2014) and Picumnus Brasileira de Biologia, 48: 169-178. aurirons (Cohn-Hat et al. 2007), all o which weigh on By, c. W. & Bm, r. t. 2009. Ecology predicts levels o average less than 13 g each. In other parts o the world, genetic dierentiation in neotropical birds. American Naturalist, tiny birds make spectacular long-distance migrations 174: 358-68. cpp, a. P. 1988. Genetic Variation in Neotropical Birds: and even in Amazonia, where most species tend to be Implications or the Speciation Process. sedentary (Stotz et al. 1996), certain species, such as chH, M. 2000. A case study in Amazonian biogeography: vocal those adapted to river islands (Remsen & Parker, 1983; and DNA-sequence variation in Hemitriccus fycatchers. Ph.D. Rosenberg, 1990), are likely to be excellent dispersers, Tesis, Louisiana State University, Baton Rouge, Louisiana. chH, M.; Ph, a. M. F.; Bh, c. l.; ts, M. F. independent o size. However, there is evidence that at n. M.; Fs, a. M.; S, c. H. & Mê, i.t. 2007. least two other heavier bodied species have structured Inventário ornitológico, Biodiversidade do médio Madeira: Bases populations delimited by the Aripuanã and Jiparaná cientícas para propostas de conservação (ed. by L.R. Py-Daniel, rivers as well: Tamnophilus aethiops (23-30g; Tom C.P. Deus, A.L. Henriques, D.M. Pimpão, M.O. Ribeiro), pp. & Aleixo 2015) and Malacoptila rua (36-44g; Ferreira 145-178. Inpa, Manaus. c, J. 1985. Historical biogeography and patterns o 2013). As discussed by Smith et al. (2014) dierences in dierentiation within South American birds: areas o endemism. lie history attributes, eective population sizes, lineage Ornithological Monographs, 36: 49-84. ages, and dispersal rates can together account or highly d’H, F. M.; cv, a. M.; rbs, c. c.; Bm, r. t.; & disparate responses o avian lineages across important Myk, c. Y. 2013. Phylogeny and comparative phylogeography o Sclerurus (Aves: Furnariidae) reveal constant and cryptic physical barriers in the Neotropics such as the Andes diversifcation in an old radiation o rain orest understorey and some large Amazonian rivers, including the Madeira specialists. Journal o Biogeography, 40: 37-49. River. Further tests, controlling or phylogeny, habitat, d Sv, J. M. c. & o, d. c. 1996. Application o parsimony wing shape and loading, and behavioral responses to open analysis o endemicity in Amazonian biogeography: an example spaces, will be necessary to determine the extent to which with primates. Biological Journal o Linnean Society, 59: 427-437. dmm, a. J. & rmb, a. 2007. BEAS: Bayesian body mass is a useul predictor o genetic population evolutionary analysis by sampling trees. BMC Evolutionary structure in suboscine passerines. Biology, 7: 214. dmm, a. J.; H, S. Y. W.; Phps, M.J. & rmb, a. 2006. Relaxed phylogenetics and dating with confdence. PLoS Biology 4: 88. acKnoWledGMentS dmm, a.J.; rmb, a.; Shp, B. & Pybs, o. G. 2005. Bayesian coalescent inerence o past population dynamics rom AMF was supported during his masters studies by a molecular sequences. Molecular Biology and Evolution, 22: 1185- ellowship rom Conselho Nacional de Desenvolvimento 1192. dmm, a. J.; nhs, G. K.; rg, a. G. & Sm, W. Cientifco e ecnologico (CNPq). Financial support 2002. Estimating mutation parameters, population history and or both laboratory and feld work was granted by IEB genealogy simultaneously rom temporally spaced sequence data. (Instituto de Educação do Brasil, programa BECA). Genetics, 161: 1307-1320. Te research was conducted in the Laboratório de Fs, a. M.; Wk, M.; S, c. H. & ax, a. 2014. Evolução e Genética (LEGAL) with portion Multiple speciation across the Andes and throughout Amazonia: the case o the spot-backed species complex (Hylophylax o the analyses being done in the High Perormance naevius/Hylophylax naevioides). Journal o Biogeography, DOI: Computational Facility o the University o Puerto Rico. 10.1111/jbi.12277 We want to thank F. A. Carvalho or essential help and Fs, a. M. 2013. Fine-scale endemism o Amazonian birds in aiding in preparing the fgures, L. N. Naka, . C. H. a threatened landscape. Biodiversity and Conservation, 22, 2683- 2694. Cole, C. A. Marantz and an anonymous reviewer or Fs, a. M.; Gzs, J.; Wk, M. & ax, a. 2013. useul and inspiring comments on earlier versions o this Multilocus phylogeography o the Wedge-billed Woodcreeper manuscript. Collecting permits were provided to AMF Glyphorynchus spirurus (Aves, Furnariidae) in lowland Amazonia: and to MCH by IBAMA (Instituto Brasileiro do Meio Widespread cryptic diversity and paraphyly reveal a complex Ambiente). diversifcation pattern. Molecular Phylogenetics and Evolution, 66: 270-282.

Revista Brasileira de Ornitologia, 22(4), 2014 Rivers acting as barriers or bird dispersal in the Amazon 371 Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

Fs, a. M.; Wk, M. & ax, a. 2012. Phylogeography o Smbk, J.; Fsh, e. F. & Ms, t. 1989. Molecular the chestnut-tailed antbird (Myrmeciza hemimelaena) clarifes the cloning: a laboratory manual, 2nd ed, Spring. Cold Spring Harbor role o rivers in Amazonian biogeography. Journal o Biogeography, Laboratory Press, Cold Spring Harbor, NY. 39: 1524-1535. Sg, F.; nk, S. & cs, a. r. 1977. DNA sequencing Fz, G.; nhs, J. d.; Hs, J. e.; Sf, P. c.; Bg with chain-terminating inhibitors. Proceedings o the National J, r. o. & lvjy, t. e. 2007. A large-scale deorestation Academy o Sciences o the United States o America, 24: 5463-5467. experiment: Eects o patch area and isolation on Amazon Birds. S, c. H. 2005. Variação genética e geográca de Hemitriccus Science, 80: 238-241. minor (Aves-yrannidae) na Bacia do Madeira, AM/Brasil. Masters F, M. 2013. Sistemática molecular e logeograa do gênero Tesis, Instituto Nacional de Pesquisas da Amazônia, Manaus, Malacoptila (Aves: Bucconidae). Master Tesis, Universidade Brazil. Federal do Pará, Belém, Brazil Smh, B. t.; Mcmk, J. e.; cv, a. M.; Hks, M. J.; Hf, J. 1997. Alternative models o vertebrate speciation in ax, a.; c, c. d.; Pzem, J.; By, c. W.; X, Amazonia – an overview. Biodiversity and Conservation, 476: 451- X.; Hvy, M. G.; Fh, B. c.; G, t. c.; dyby, 476. e. P.; Pj, J.; Fs, S. & Bm, r. t. 2014. Te drivers Hf, J. 1974. Avian speciation in tropical South America. Nuttall o tropical speciation. Nature, 515: 406-409. Ornithological Club, Cambridge, Massachusetts. Ssnvs t; ax a. & Sq F. 2013. Cryptic patterns o H, J. P. W. & Hvy, d. J. 2002. Te phylogeography o Amazonia diversifcation o a widespread Amazonian Woodcreeper species revisited: new evidence rom riodinid butteries. Evolution, 56: complex (Aves: dendrocolaptidae) inerred rom multilocus 1489-1497. phylogenetic analysis: implications or historical biogeography H, t. a. 1999. Bioedit: A user-riendly biological sequence and taxonomy. Molecular Phylogenetics Evolution, 68: 410-424. alignment editor and analysis program or windows 95/98/N. Sz, d. F.; Fzpk, J. W.; Pk, t. a. & Mskvs, d. Nucleic Acids Symposium, 41: 95-98. K. 1996. Neotropical birds: ecology and conservation. Chicago Hsgw, M.; Ksh, H. & Y, t. a. 1985. Dating o the University Press. human-ape splitting by a molecular clock o mitochondrial DNA. Swf, d. l. 2002. PAUP*: phylogenetic analysis using parsimony Journal o Molecular Evolution 22: 160-174. (* and other methods). Version 4.0b10. Hsbk, J. P. & rqs, F. 2001. MrBayes: Bayesian inerence Tm, G. & ax, a. 2015. Cryptic speciation in the White- o phylogeny. Bioinormatics, 17: 754-755. shouldered Antshrike (Tamnophilus aethiops, Aves – is, M. l. & Why, B. M. 2011. Species limits in Tamnophilidae): the tale o a transcontinental radiation across (Tamnophilidae): the scale- backed antbird (Willisornis rivers in lowland Amazonia and the northeastern Atlantic Forest. poecilinotus) complex. Te Wilson Journal o Ornithology, 123: Molecular Phylogenetics and Evolution, 82: 95-110. 1-14. tbs, J. a.; Bs, J. M.; Hk, S. J.; S, n.; W, J. t. is, M. l.; is, P. r. & Why, B. M. 2007. Species limits in & Sh, d. 2008. Response to comment on “Te latitudinal antbirds (Tamnophilidae): the warbling antbird (Hypocnemis gradient in recent speciation and extinction rates o birds and cantator) complex. Auk, 124: 11-28. mammals”. Science, 319: 901. lbss, e. 2002. Evidence o quaternary paleohydrological V rsm, M. G. M.; v rsm, t.; Mm, r. a. & changes in middle Amazonia: Te Aripuanã-Roosevelt and Fs, G. a. 1998. A new and distinct species o Marmoset Jiparaná “ans”. Zeitschrit ür Geomorphologie, 129; 61-72. (Callitrichidae, Primates) rom the lower Rio Aripuanã, state o Mz, c. a.; ax, a.; Bv, l. r. & P, M. a. 2003. Amazonas, central Brazilian Amazonia. Goeldiana Zoologia, 22: Family Dendrocolaptidae (Woodcreepers), In: del Hoyo, J., 1-27. Elliott, A., Christie, D.A. (Eds.), Handbook o the Birds o the Sw, d.W. 2004. Family ityridae (Manakins), In: del Hoyo, J., World. Vol. 8. Broadbills to apaculos. Lynx Edicions, Barcelona, Elliott, A., Christie, D.A. (Eds.), Handbook o the birds o the pp. 358-447. world, Vol. 9: Cotingas to Pipits and Wagtails. Lynx Edicions, nyá, a. 2007. Phylogeographic pattern, molecular and vocal Barcelona, pp. 110-169. dierentiation, and species limits in Schiornis turdina (Aves). W, a. 1852. On the monkeys o the Amazon. Proceedings o the Molecular Phylogenetics and Evolution, 44: 154-164. Zoological Society o London, 20: 107-110. Pé, J. S. l.; Wks, J. d.; ax, a. & Bs, J. M. W, B. H.; Bmghm, e.; Bw, r. c. K.; PysJs, r. P. 2009. Evolutionary history o Ramphastos toucans: molecular & Téb, c. 2003. Molecular phylogeography reveals island phylogenetics, temporal diversifcation, and biogeography. colonization history and diversifcation o western Indian Ocean Molecular Phylogenetics and Evolution, 53: 923-934. sunbirds (Nectarinia: Nectariniidae). Molecular Phylogenetics and Ps, J. l. 1951. Checklist o the Birds o the World. Vol. VIII. Museum Evolution, 29: 76-85. o Comparative Zoology, Cambridge, Massachusetts. W, J. t. & Sh, d. 2008. Calibrating the avian molecular Ps, d. & c, K. a. 1998. MODELES: testing the clock. Molecular Ecology, 17: 2321-2328. model o DNA substitution. Bioinormatics, 14: 817-818. Why, B. M.; is, M. l.; Bv, G. a.; aszáb, n.; Shk, rms, J. V. & Pk, t. a. 1983. Contribution o river-created F.; Sv, l. F.; P, V. Q.; chH, M. & rêg, M. habitats to bird species richness in Amazonia. Biotropica, 15, a. 2013. A new species o antbird in the Hypocnemis cantator 223-231. complex rom the Aripuanã-Machado interuvium in central rbs, c. c.; ax, a.; ng, a. c. r.; Myk, c. Y. & Amazonian Brazil. In: del Hoyo J, Elliott A, Sargatal J, Christie c, J. 2012. A palaeobiogeographic model or biotic D (Eds.), Handbook o the Birds o the World. Special volume: new diversifcation within Amazonia over the past three million years. species and global index. Lynx Edicions, Barcelona, pp. 282-285. Proceedings o the Royal Society B, 1729: 681-689. Why, B. M.; chH, M.; Bv, G. a.; & Sv, l. rgy, r. S. & t, G. 1994. Te birds o South America. Vol. II. F. 2013b. A new species o Herpsilochmus antwren rom the Te suboscine passerines. University o exas Press, Austin, exas. Aripuanã-Machado interuvium in central Amazonian Brazil. In: rsbg, G. H. 1990. Habitat specialization and oraging behavior del Hoyo J, Elliott A, Sargatal J, Christie D (Eds.), Handbook o by birds o Amazonian river islands in northeastern Peru. Condor, the Birds o the World. Special volume: new species and global index. 92: 427-443. Lynx Edicions, Barcelona, pp 277-281.

Revista Brasileira de Ornitologia, 22(4), 2014 372 Rivers acting as barriers or bird dispersal in the Amazon Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

Why, B. M.; is, M. l.; Bv, G. a.; aszáb, n.; Shk, the lower Sucunduri-apajós interuvium in central Amazonian F.; Sv, l. F. & P, V. Q. 2013. A new species o Brazil heralds a new chapter in Amazonian biogeography. In: del Epinecrophylla antwren rom the Aripuanã-Machado interuvium Hoyo J, Elliott A, Sargatal J, Christie D (Eds.), Handbook o the in central Amazonian Brazil with revision o the ‘‘stipple-throated Birds o the World. Special volume: new species and global index. antwren’’ complex. In: del Hoyo J, Elliott A, Sargatal J, Christie Lynx Edicions, Barcelona, pp 297-300. D (Eds.), Handbook o the Birds o the World. Special volume: new Ws, e. o. 1969. On the behavior o fve species o Rhegmatorhina, species and global index. Lynx Edicions, Barcelona, pp 263-267. ant-ollowing antbirds o the Amazon basin. Wilson Bulletin 81: Why, B. M.; Shk, F.; rêg, M.a. & Sv, l. F. 2013. 362-394. A new species o Zimmerius tyrannulet rom the upper Madeira- Zmm, K. J. & is, M. l. 2003. Family Tamnophilidae (ypical apajós interuvium in central Amazonian Brazil: Birds don’t antbirds), In: del Hoyo, J., Elliott, A., Christie, D.A. (Eds.), always occur where they ‘‘should’’. In: del Hoyo J, Elliott A, Handbook o the Birds o the World. Vol. 8. Broadbills to apaculos. Sargatal J, Christie D (Eds.), Handbook o the Birds o the World. Lynx Edicions, Barcelona, pp. 448-681. Special volume: new species and global index. Lynx Edicions, Barcelona, pp 286-291. Why, B. M.; Shk, F.; rêg, M. a. & Sv, l. F. 2013. A new species o ycatcher in the olmomyias assimilis radiation rom Associate Editor: Gustavo Sebastián Cabanne

aPPendiX

Inormation on the specimens analyzed.

Pp/ly tx: Vh/Gbk ss mbs

0/ AM: right bank o lower Juruá, RESEX Baixo Juruá, G. spirurus: INPA A 808/HM164938 comunidade Socó. 3°36’S; 66°4’W 1/ AM:”Campo do Lago Preto”, let bank o Madeira river, 39 W. poecilinotus: INPA A 420/HM164985,421/HM164980, km W Novo Aripuanã. 5°09’S; 60°44’W 422/HM164994, 424/HM164975. S. turdina: INPA A 395/ HM165029, 413/HM165028 2/ AM: right bank Purús river, Ussuã stream, tributary o W. poecilinotus: INPA A 101/HM164978 Mucuim river (right bank). 7°13’S; 64°10’W 3/ RO: let bank o the Madeira river, near Jacy Paraná, ca. 45 G. spirurus: INPA A 349/HM164942, 359/HM164941. W. km southwest Porto Velho. 9°10’S; 64°23’W poecilinotus: INPA A 345/HM164983, 347/HM164999, S. turdina INPA A 348/HM165032 4/ RO: let bank o the Madeira river, ca. 20 km N Abunã. G. spirurus: INPA A 173/HM164943, 191/HM164960. W. 9°31’S; 65°21’W poecilinotus: 182/HM164984 5/ AM: let bank o Aripuanã river, Arauazinho stream, 130 km G. spirurus: INPA A 461/HM164951, 466/HM164950, S Novo Aripuanã. 6°18’S; 60°24’W 510/HM164952, 553/HM164963, 562/HM164962. W. poecilinotus: INPA A 465/HM164974, 472/HM164967, 475/HM165009, 504/HM164971. S. turdina: INPA A 533/ HM165014 6/ AM: let bank o lower Roosevelt river, conuence with G. spirurus: INPA A 906/HM164945. W. poecilinotus: INPA A Aripuanã river. 7°35’S; 60°43’W 902/HM164976, 904/HM164992. S. turdina: INPA A 903/ HM165017.

Revista Brasileira de Ornitologia, 22(4), 2014 Rivers acting as barriers or bird dispersal in the Amazon 373 Alexandre M. Fernandes, Mario Cohn-Hat, omas Hrbek and Izeni Pires Farias

7/ AM: right bank o lower Roosevelt river, conuence with G. spirurus: INPA A 895/HM164948. W. poecilinotus: INPA Aripuanã river. 7°38’S; 60°40’W A 894/HM164997, 896/HM165007, 900/HM165003, 898/ HM165001. S. turdina: INPA A 893/HM165015 8/ RO: right bank lower Jiparaná river, comunidade G. spirurus: INPA A 875/HM164946, 877/HM164958. W. Demarcação, ca. 20 km southeast Calama. 8°09’S; 62°47’W poecilinotus: INPA A 876/HM164989, 884/HM164990, 907/ HM165008. S. turdina: INPA A 885/HM165016 9/ RO: let bank o lower Jiparaná river, opp. Comunidade G. spirurus: INPA A 870/HM164964, 886/HM164940. W. Demarcação, ca. 20 km southeast Calama. 8°14’S; 62°46’W poecilinotus: INPA A 871/HM164982, 872/HM164996, 873/HM165000, 878/HM164988, 881/HM165010, 882/ HM164977. S. turdina: INPA A 874/HM165034, 880/ HM165021, 883/HM165018 10/ RO: right bank o Madeira river, 9.5 km southeast Porto G. spirurus: INPA A 329/HM164959. W. poecilinotus: INPA Velho. 8°52’S; 64°0’W A 307/HM164966, 308/HM164970, 326/HM164968, 327/ HM164973, 334/HM164972 11/ RO: right bank o Madeira river, near Jacy Paraná, ca. 45 W. poecilinotus: INPA A 367/HM165002, 368/HM164969. km southwest Porto Velho S. turdina: INPA A 371/HM165030, 372/HM165025, 374/ HM165026 12/ RO: right bank o Madeira river, ca. 20 km N Abunã. G. spirurus: INPA A 208/HM164949. W. poecilinotus: INPA 9°35’S; 65°21’W A 248/HM165005, 264/HM165006, 265/HM164993. S. turdina: INPA A 249/HM165031, 266/HM165012 13/ AM: right bank o Aripuanã, Extremo stream, 135 km S G. spirurus: INPA A 536/HM164944, 559/HM164947, 561/ Novo Aripuanã. 6°18’S; 60°20’W HM164955. W. poecilinotus: INPA A 478/HM164987, 479/ HM164979. S. turdina: INPA A 525/HM165027, 527/ HM165023, 538/HM165024 14/ AM: right bank o middle Aripuanã, conuence with G. spirurus: INPA A 890/HM164954. W. poecilinotus: INPA Roosevelt river. 7°37’S; 60°40’W A 887/HM164995, 888/HM164981, 891/HM165004. S. turdina: INPA A 892/HM165033, 889/HM165019 15/ AM: Floresta Estadual do Sucunduri, right bank o upper G. spirurus: INPA A 845/HM164957. W. poecilinotus: INPA A Sucunduri river. 8°34.5’S; 59°08.5’W 849/HM164998. S. turdina: INPA A 846/HM165013, 848/ HM165022, 850/HM165020 16/ AM: Parque Estadual do Sucunduri; right bank o Bararati G. spirurus: INPA A 852/HM164953, 855/HM164956. W. river. 8°21’S; 58°37’W poecilinotus: INPA A 856/HM164986, 857/HM164991 Outgroup/ AM: right bank o upper Negro river, 3 km SW São G. spirurus: INPA A 1153/HM164939 Gabriel da Cachoeira. 0°8’S; 67°5’W Outgroup/ AM: let bank upper Negro river, 10 km east São G. spirurus: INPA A 1118/HM164961 Gabriel da Cachoeira. 0°10’S; 66°59’W Outgroup/ AM: let bank middle Solimões river; RDS Amanã, W. poecilinotus: INPA A 398/HM164965 Comunidade Nova Canaã, Centro Grande stream. 2°36’S; 64°52’W Outgroup/ AM: ca. 60 km N Manaus; highway BR-174, km S. turdina: INPA A 777/HM165011 43; Campina reserve/INPA

Revista Brasileira de Ornitologia, 22(4), 2014