Content uploaded by Péricles Sena do Rêgo
Author content
All content in this area was uploaded by Péricles Sena do Rêgo on Mar 06, 2015
Content may be subject to copyright.
SYSTEMATIC REVISION OF THE SPOTTED ANTPITTA
(GRALLARIIDAE: HYLOPEZUS MACULARIUS), WITH DESCRIPTION OF
A CRYPTIC NEW SPECIES FROM BRAZILIAN AMAZONIA
LincoLn SiLva carneiro,
1
Luiz Pedreira GonzaGa,
2
PéricLeS S. rêGo,
3
iraciLda SamPaio,
3
Horacio ScHneider,
3
and aLexandre aLeixo
4,5
1
Curso de Pós-Graduação de Zoologia, Universidade Federal do Pará–Museu Paraense Emílio Goeldi, Belém, Pará, Brazil;
2
Departamento de Zoologia, Instituto de Biologia, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil;
3
Laboratório de Genética e Biologia Molecular, Campus Universitário de Bragança, Universidade Federal do Pará, Bragança-PA, 68600-000, Brazil;
and
4
Coordenação de Zoologia, Museu Paraense Emílio Goeldi, Caixa Postal 399, CEP 66040-170, Belém, Pará, Brazil
A.—We present a systematic revision of the polytypic Spotted Antpitta (Grallariidae: Hylopezus macularius) based on
morphometric, plumage, vocal, and molecular characters. Morphological and vocal analyses were based, respectively, on specimens
and recordings. Molecular phylogenies were inferred on the basis of , base pairs of the mitochondrial DNA genes S, ND, and
cytochrome b from specimens, including several outgroups. Our results revealed the existence of an undescribed taxon endemic to
the Madeira–Xingu interfluvium, similar in morphology to paraensis, but vocally and genetically readily distinguished from the latter
and any other taxon grouped under H. macularius. We also found that populations from the Negro River basin (currently treated in
paraensis) and those from northern Peru and southern Venezuela (placed in diversus) should be treated as a single taxon, for which the
name dilutus is available. Reconstructed phylogenies recovered, with overall strong support, the reciprocal monophyly among four
main lineages of the Spotted Antpitta, three corresponding to already named taxa (dilutus, macularius, and paraensis), and one to
the unnamed taxon, which we describe. We show that those four taxa are also mutually diagnosed by a combination of both vocal and
morphological features, and we recommend treating them as separate species under alternative species concepts. Received July ,
accepted February .
Key words: Hylopezus, molecular systematics, song evolution, species limits, Spotted Antpitta, taxonomy, vocal characters.
Revisão Sistemática do torom-carijó Hylopezus macularius (Grallariidae), com a descrição de uma nova espécie
críptica da Amazônia Brasileira
R.—Uma revisão sistemática da espécie politípica Hylopezus macularius (Grallariidae), baseada em caracteres morfométricos,
de plumagem, vocais e moleculares, é apresentada. As análises morfológicas e vocais foram baseadas, respectivamente, em espécimes
e gravações. As filogenias moleculares basearam-se em . pares de bases de DNA dos genes mitocondriais S, ND, e cyt b de
espécimes, incluindo diversos táxons como grupos externos. Nossos resultados revelaram a existência de um táxon não descrito, endêmico
do interflúvio Xingu - Madeira, similar morfologicamente a paraensis, mas distinguível vocal e geneticamente do último e de todos os
demais táxons agrupados sob H. macularius. Também verificamos que as populações da bacia do rio Negro (atualmente tratadas como
paraensis) e aquelas no norte do Peru e sul da Venezuela (alocadas em diversus) devem ser tratadas dentro de um único táxon, para o qual o
nome dilutus está disponível. As árvores moleculares obtidas recuperaram com um forte apoio o monofiletismo recíproco entre as quatro
linhagens principais de H. macularius, três das quais correspondem a táxons já nomeados (dilutus, macularius, e paraensis), e uma ao
táxon anônimo, que é descrito neste trabalho. Nós mostramos que estes quatro táxons são mutuamente diagnosticáveis através de uma
combinação de caracteres vocais e morfológicos, portanto recomendamos tratá-los como espécies separadas com base em conceitos de
espécie alternativos.
—
338
—
e Auk 129(2):338−351, 2012
e American Ornithologists’ Union, 2012.
Printed in USA.
e Auk, Vol.
, Number , pages −. ISSN -, electronic ISSN -. by e American Ornithologists’ Union. All rights reserved. Please direct all
requests for permission to photocopy or reproduce article content through the University of California Press’s Rights and Permissions website, http://www.ucpressjournals.
com/reprintInfo.asp. DOI: ./auk..
5
Address correspondence to this author. E-mail: aleixo@museu-goeldi.br
T H was described by Ridgway () and
currently includes eight species distributed throughout most of
the Neotropics (Honduras to northeastern Argentina; Krabbe
and Schulenberg , Remsen et al. ). e endemic Spotted
Antpitta (H. macularius) inhabits both upland and seasonally
flooded lowland humid forests of the Amazon Basin. Currently,
the Spotted Antpitta is treated as a polytypic species with three
recognized subspecies: H. m. macularius, H. m. paraensis, and
aPriL 2012 — SyStematic reviSion of Hylopezus macularius — 339
Amplification conditions for the segment of S rRNA gene
consisted of initial denaturation at °C for min, followed by
cycles of : s at °C, min at °C, and min at °C. Final ex-
tension was at °C for min. Amplification of the cyt-b and ND
segments had the following steps: cycles of s at °C, min at
°C, and min at °C, followed by a final step of min at °C.
e PCR products were purified with ExoSap-It (Amersham Biosci-
ences, Piscataway, New Jersey) and sequenced with Big Dye reagent
Kit (Applied Biosystems, Foster City, California) following the man-
ufacturer’s protocols. Reagents not incorporated during the cycle
sequencing reaction were eliminated by washing with isopropanol
and products were run on an ABI Prism sequencer.
e sequences of each gene region were checked by eye,
edited manually using the BIOEDIT software (Hall ), and
aligned using the ClustalW application (ompson et al. ).
e following measures were taken to ensure that the DNA frag-
ments sequenced were accurate and of mitochondrial origin
(not pseudogenes): () both DNA strands were sequenced; () se-
quences were inspected using BIOEDIT (Hall ) for insertions,
deletions, and stop codons that would result in a nonfunctional
protein; and () possible saturation among ingroup sequences was
examined by plotting the number of transition and transversion
substitutions against p-distances for each pairwise comparison
using the program DAMBE (Xia and Xie ).
Phylogenetic analysis.—Phylogenetic analyses were per-
formed on the concatenated S–cyt b–ND data set, after a par-
tition-homogeneity test implemented in PAUP*, version .b
(Swofford ), failed to detect significantly different phyloge-
netic signals among the three genes analyzed (P > .). Genetic
distances were also calculated in PAUP. We used PAUP to run a
maximum-parsimony (MP) analysis on the concatenated data set
using a heuristic search with tree bisection–reconnection (TBR)
branch swapping and random-addition replicates. Support
for each node was assessed by , bootstrap replicates using a
heuristic search, TBR branch swapping, and random additions
per replicate. Maximum-likelihood (ML) analyses were also car-
ried out with PAUP under a general time-reversible (GTR) model
of sequence evolution with invariable sites and gamma rate het-
erogeneity (GTR+Γ+I), which, according to JMODELTEST, ver-
sion .. (Posada ), was the best model of sequence evolution
according to Akaike’s information criterion (AIC). Nodal sup-
port in ML analyses was assessed by bootstrap replicates
with TBR branch swapping and one random addition per repli-
cate. For Bayesian analyses (BA), we used JMODELTEST (Posada
) and MRMODELTEST, version . (Nylander ), to find
the best model of evolution for each partitioning scheme and then
tested different partitioning schemes with Bayes factor analysis
(Kass and Raftery ) as follows: () all data combined, () three
partitions (S, cyt-b, and ND), and () nine partitions (first, sec-
ond, and third codon positions of the S, cyt-b, and ND genes).
e Bayes factor analysis selected nine partitions as the best par-
titioning scheme. Each partition was assigned its own likelihood
model on the basis of JMODELTEST and MRMODELTEST re-
sults. All parameters were unlinked between partitions (except for
the topology and branch-length parameters) and were estimated
as part of the analysis. Using MRBAYES, version .. (Ronquist
and Huelsenbeck ), two parallel runs were carried out, each
with four Markov chains and for million generations, sampling
H. m. diversus (Krabbe and Schulenberg ). is taxonomic
treatment began with Snethlage (a), who described par-
aensis as a subspecies of H. macularius and was later consoli-
dated by Zimmer (), who lumped macularius, paraensis,
and the new taxon he described, diversus, into a single biological
species because plumage differences separating them were very
subtle, suggesting only subspecific differentiation. In addition to
the subspecies above, a fourth taxon, H. m. dilutus, was described
by Hellmayr () but was later synonymized with paraensis by
Cory and Hellmayr ().
Recent field work that we and others have conducted indi-
cates that pronounced vocal variation exists among subspecies
of H. macularius and that major vocal patterns conflict strongly
with all proposed taxonomic treatments based on plumage char-
acters (Snethlage , Cory and Hellmayr , Zimmer ,
Krabbe and Schulenberg ). For instance, Maijer ()
showed that H. auricularis (until then regarded as a subspecies
of H. macularius) was vocally very distinct from any other taxa
grouped under the Spotted Antpitta and, thus, deserved full spe-
cies status, a recommendation that has been followed ever since
(Krabbe and Schulenberg , Remsen et al. ). Other simi-
lar examples of cryptic undescribed variation still persist in the
polytypic H. macularius complex, suggesting that further splits
are probably warranted (Krabbe and Schulenberg , Remsen
et al. ). us, a major multicharacter taxonomic revision is
long due for the Spotted Antpitta complex to objectively base
taxonomic decisions concerning the ranking of its taxa. Here,
we use a combination of morphological, vocal, and molecular
characters to review the taxonomy and interspecific limits in the
Spotted Antpitta.
Methods
Genetic sampling.—Molecular analyses were based on muscle
tissue samples belonging to all currently recognized H. macu-
larius subspecies (Krabbe and Schulenberg ), except diver-
sus, and outgroups that included representatives of four other
Hylopezus species (H. auricularis, H. berlepschi, H. nattereri, and
H. ochroleucus) and another genus in Grallariidae (rush-like
Antpitta [Myrmothera campanisona]). We also obtained samples
from two study skins housed at the Museu Paraense Emílio Goeldi
(MPEG; specimen numbers and ) through the re-
moval of digital and metatarsal pads (Fig. ; Appendix S in online
supplementary materials [see Acknowledgments]).
Genetic samples obtained from skins were extracted from
digital and metatarsal pads using the Qiagen (Valencia, Califor-
nia) DNeasy kit, following the manufacturer’s protocol, whereas
those obtained from muscle tissue were extracted following a phe-
nol-chlorophorm method as described in Sambrook et al. ().
Polymerase chain reaction (PCR) was used to amplify a segment
of ~ base pairs (bp) of the rRNA S mitochondrial gene (s
rRNA) with the following primers as described by Palumbi et al.
(): (L-) and (H-). For the segment of ~ bp of the
cytochrome-b (cyt-b) gene, the following primers developed by So-
renson et al. () were used: (L-) and (H-), whereas
for the segment of ~ bp of the NADH dehydrogenase subunit
(ND), the following primers developed by Hackett () were
used: (L-) and (H-).
340 — carneiro et aL. — auk, voL. 129
the Markov chains every , generations. We used the resulting
, parameter point-estimates minus the burn-in generations
(,) to create a % majority-rule consensus tree and to calcu-
late Bayesian posterior probabilities (PP) to assess nodal support.
Using TRACER, version . (Drummond and Rambaut ), we
determined that the chosen burn-in setting (%) was sufficient
for the log likelihood values of parallel runs to reach stationar-
ity, with all parameters meeting benchmark effective sample-size
values (>).
Morphological analyses.—We examined study skins,
belonging to all taxa described so far and grouped under the
polytypic H. macularius (dilutus, diversus, macularius, and
paraensis), housed at various institutions (Appendix S in on-
line supplementary materials; see Acknowledgments). In ad-
dition to those specimens, high-resolution digital pictures of
H. m. macularius and H. m. dilutus type specimens were exam-
ined (see Fig. and Appendix S). To avoid discrepancies between
researcher measurements, only data on the specimens mea-
sured by L.S.C. were included in the statistical analyses. For the
remaining specimens examined only by A.A. at the Ameri-
can Museum of Natural History (AMNH), Carnegie Museum
(CM), and Naturhistorisches Forschunginstitut, Museum für
Naturkunde (ZMB), only information on plumage and color were
incorporated into the study (Appendix S).
Measurements of the following characters were taken to the
nearest . mm with an electronic caliper: wing length, tail length,
tarsus length, bill length from the distal points of the nostrils to the
tip of the bill, bill depth and width at the distal point of the nostrils,
average length of the whitish portion of the lower pectoral feathers,
average width of the whitish portion of the lower pectoral feathers,
fiG. 1. Geographic distribution of specimens, vocalizations, and tissues of Hylopezus macularius taxa analyzed in the present study. We followed
Krabbe and Schulenberg’s (2003) subspecies definitions. Squares = H. m. paraensis, circles = H. m. diversus, and triangles = H. m. macularius. Type
localities for each individual taxon are shown as white symbols. The lone white square with a small dot in the center indicates the type locality of
dilutus, synonymized previously into paraensis, and thus not considered by Krabbe and Schulenberg (2003). Letters next to a symbol represent ma-
terials available for that given locality: P = photographs only; S = skins only; V = tape-recordings only; T = tissues only; S,V = skins and vocalizations
only; S,T = skins and tissues only; and C = tape-recordings, skins, and tissues. Dashed lines delimit main lineages recovered by a molecular phylogeny
and interpreted as natural populations as follows: A = paraensis, B = whittakeri, C = dilutus, and D = macularius. The question mark denotes an area
between the Branco and Negro rivers where the presence and taxonomic identity of any H. macularius population are unknown. The conspicuous
lack of any historical and modern records of H. macularius in southwestern Amazonia (Inambari area of endemism, sensu Da Silva et al. 2005) appear
to reflect a true absence from this area (Krabbe and Schulenberg 2003, Whittaker et al. 2008).
aPriL 2012 — SyStematic reviSion of Hylopezus macularius — 341
and average width of the blackish terminal area of the lower pecto-
ral feathers. Averages for the three last characters were calculated on
the basis of measurements of five individual feathers each. All mor-
phological nomenclature follows Proctor and Lynch (). We used
Smithe (, ) as a standard color reference when describing
plumage tones. Groupings for statistical analysis of morphological
data were based on the molecular phylogeny obtained (see below),
which recognized four main lineages in H. macularius (Fig. ). We as-
sessed normality of morphometric data with Kolmogorov-Smirnov
tests and used discriminant function analyses (DFA) to test for dif-
ferences in the morphometric space among lineages. We combined
both sexes in the analyses because there was no evidence of sexual di-
morphism for any character. Missing morphometric values for some
specimens were estimated using a missing value analysis, based on
the linear regression of the observed variables. Missing values never
represented >% of the total measurements obtained for each char-
acter. All statistical analyses were performed with SYSTAT, version
, for Windows (Systat Software, San Jose, California). In all tests,
statistical significance was accepted at P ≤ ..
Vocal analyses.—We analyzed different recordings, in-
cluding > distinct songs, from localities throughout the
Amazon, belonging to all subspecies of H. macularius (dilutus, di-
versus, macularius, and paraensis); only one song or call per in-
dividual was used in the analyses (Fig. ; Appendix S in online
supplementary materials [see Acknowledgments]).
e vocalizations were categorized as “loudsongs” and “calls”
(sensu Willis ) through auditory and visual comparisons of
spectrograms. e H. macularius loudsong usually consists of six
clear whistled notes. We measured the duration of each of the six
notes, the duration of each of the five intervals between notes, and
the interval between loudsongs, thus yielding different time-re-
lated characters for each individual song. However, because some
individuals, especially those found between the Madeira and Xingu
rivers ascribed by current taxonomy to paraensis, sometimes omit-
ted one or two final notes from their loudsongs, only time-related
vocal characters present in the majority of vocalizations (%) were
included in statistical analyses. ese measurements were made
in the waveform; when necessary, background noise was removed
through lowpass and highpass filtering.
Following the method proposed by Isler et al. (), we cal-
culated the mean value and standard deviations for all quantitative
vocal variables measured in three bouts of loudsong. Only one song
or call recording per individual was used in the analyses. To evaluate
the difference between any two populations in any of the quanti-
tative variables, we used the following criteria. First, we considered
two continuous and normally distributed variables diagnostic only if
their ranges did not overlap and if the means (X) and standard devia-
tions (SD) of the population with the smaller set of measurements (a)
and the population with the larger set of measurements (b) met the
requirement: X
a
+ t
a
SD
a
≤ X
b
– t
b
SD
b
, where t
i
is the t score at the .
percentile of the t distribution for n – df (Isler et al. ). In addi-
tion, we ran a DFA in SYSTAT to determine whether the taxa under
study could be diagnosed in the multivariate song space (based on
variables; see Table ). Again, to avoid pseudoreplication, only one
song or call recording per individual was used in the analyses. The
sample size of loudsongs obtained for H. m. diversus and H. m. dilutus
was not sufficient for statistical analyses (Fig. and Appendix S), so
they were analyzed only in a qualitative fashion.
Loudsong and call frequency measurements were made using
audiospectrograms and refer to the fundamental harmonic, which
in all vocalizations analyzed was also the dominant one. e “max
frequency” measurement corresponded to the frequency at which
the maximum power occurs within a given time interval (Charif
et al. ). us, the values given are the dominant frequency
measured while considering the entire duration of each note. e
syntax and note structure (sensu Isler et al. ) were analyzed
qualitatively through a blind inspection and grouping of printed
sonograms, followed by an assessment of whether the groupings
matched the populations under study. Audiospectrograms and all
song and call measurements were carried out using RAVEN PRO,
version . (Cornell Laboratory of Ornithology, Ithaca, New York),
with all vocalizations digitized at a sample rate of . kHz and
bits in the mono pattern.
Results
Molecular phylogenetics.—We obtained , bp from segments
of the S rRNA ( bp), ND ( bp), and cyt-b ( bp) mi-
tochondrial genes for specimens of H. macularius and
outgroup taxa (Appendix S). Of these, bp (~.%) were phy-
logenetically informative. No stop codons were observed. Transi-
tion versus transversion plots do not indicate saturation among
ingroup taxa. All phylogenetic trees obtained by MP, ML, and BA
grouped all H. macularius samples in a strongly supported clade,
but the node connecting it to its sister group (H. ochroleucus) was
highly supported only in BA (Fig. ). Nevertheless, MP, ML, and
BA trees ruled out the possibility that H. auricularis (once re-
garded as a subspecies of H. macularius) is nested in the Spotted
Antpitta clade (Fig. ). Within H. macularius, four major recipro-
cally monophyletic clades were recovered with overall high statis-
tical support (Figs. and ), indicating that two of the subspecies
of H. macularius recognized by Krabbe and Schulenberg ()
and sampled genetically in this study are paraphyletic with strong
statistical support, because they include the non-sister clades A+D
(macularius) and B+C (paraensis; Fig. ). e phylogenetic posi-
tion of the third subspecies (diversus) was not assessed because
of the lack of tissues from nearby its type locality, but both mor-
phological and vocal evidence suggests that this taxon is included
in clade C (see below). Because the molecular phylogeny obtained
contrasted strongly with currently recognized subspecies limits
in H. macularius (Krabbe and Schulenberg ), clades or popu-
lations A–D (Figs. and ) will be referred to as natural evolutio-
nary units for a taxonomic reassessment of the entire group based
on morphological and bioacoustical characters.
Morphological characters.—A DFA analysis found significant
differences among some populations of H. macularius in the mor-
phometric space, with the first two canonical discriminant vari-
ables accounting for .% of the total variation based mainly on
the contributions of tarsus length, bill depth, bill width, and the
extent of white on the pectoral spots (Wilks’s lambda = .,
P= .; Fig. and Table ). Population C is distinguished from
the others by significantly shorter tarsi and greater bill depth and
width values, whereas population B differs from populations A,
C, and D by the highest bill-width values in the entire sampling
(Table ). Finally, populations A and D overlap broadly in the mor-
phometric space (Fig. and Tables and ).
342 — carneiro et aL. — auk, voL. 129
taxa are visually apparent on sonograms of representative loud-
songs (Fig. ).
A DFA based on the loudsong characters measured also
showed a clear separation between population B and populations
A and D, with the first two canonical discriminant variables ac-
counting for .% of the total variation (Wilks’s lambda = .,
P = .; Fig. ); the obtained DFA classified correctly most
loudsong samples, including all those of population B (Table ).
Because of a lack of sufficient vocal samples, population C was not
included in the DFA analysis, but loudsong structure in this pop-
ulation resembled closely those of population D (Tables and
and Fig. ). e duration of the second note and the duration of
the second interval (between the second and third notes) were the
most useful in discriminating populations A–D (Table ). Further-
more, the number of loudsong notes was an important character
setting apart birds of population B (in which >.% of the record-
ings sampled contained songs with five or four notes) from those
e four natural populations of H. macularius are also very
similar in plumage, although population C is distinguished from
the others by the following characteristics: () less olivaceous
and more brownish (color no. ) upperparts; () shaft-streaks
on the mantle absent or obsolete; () deeper ochraceous subter-
minal band of pectoral feathers; () darker black terminal band
of pectoral spots; and () wingbars extended only to the middle
rather than the tips of the wing (Fig. ). Population D, on average,
has deeper buff flanks (color no. B) than all remaining popu-
lations and less conspicuous shaft-streaks on the mantle than
populations A and B (Fig. ). Populations A and B could not be
mutually differentiated by any plumage character (Fig. ).
Vocal characters.—e pairwise diagnosability tests showed
diagnostic features in all (two) qualitative (syntax and notes
structure) and of the quantitative loudsong traits analyzed
(number of notes, pace, duration of second, third and fourth in-
tervals; Tables and ). Moreover, vocal differences among the
tabLe 1. Mean (± SD) values of 15 continuous and 3 discrete loudsong and call characters evaluated in the present study (left) and results of the pair-
wise test of diagnosability in loudsong characters among four taxa of the Hylopezus macularius complex (right). The taxonomy follows that proposed
for the H. macularius complex in the text.
Taxon Pairwise diagnosability test
a
Vocal variable
paraensis
(n = 27)
whittakeri
(n = 28)
dilutus
(n = 4)
macularius
(n = 45)
par-
aensis vs.
whittakeri
par-
aensis vs.
dilutus
paraensis
vs. mac-
ularius
whittak-
eri vs.
dilutus
whittaker
vs. macu-
larius
dilutus
vs. mac-
ularius
Number of notes 5.96 ± 0.19 4.93 ± 0.19 6.00 ± 0.00 5.97 ± 0.26 * N.D. 1
st
N.D. 1
st
* * N.D. 1
st
Duration of the song
b
(s) 2.34 ± 0.13 2.66 ± 0.36 2.08 ± 0.15 2.36 ± 0.15 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 2
nd
N.D. 1
st
N.D. 1
st
Pace (number of notes/
duration)
b
2.55 ± 0.13 1.85 ± 0.15 2.88 ± 0.20 2.53 ± 0.17 * N.D. 1
st
N.D. 1
st
* * N.D. 1
st
Duration of first note(s) 0.24 ± 0.03 0.24 ± 0.03 0.25 ± 0.06 0.24 ± 0.03 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
Duration of second note(s) 0.22 ± 0.02 0.25 ± 0.03 0.23 ± 0.01 0.21 ± 0.03 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
Duration of third note(s) 0.23 ± 0.03 0.26 ± 0.03 0.21 ± 0.03 0.21 ± 0.04 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
Duration of fourth note(s) 0.24 ± 0.03 0.26 ± 0.03 0.24 ± 0.02 0.21 ± 0.03 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
Duration of fifth note(s) 0.22 ± 0.03 0.24 ± 0.07 0.21 ± 0.02 0.21 ± 0.03 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
Duration of sixth note(s)
b
0.22 ± 0.04 0.22 ± 0.06 0.21 ± 0.05 0.20 ± 0.03 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
Duration of first interval(s) 0.16 ± 0.03 0.27 ± 0.05 0.13 ± 0.02 0.19 ± 0.04 N.D. 2
nd
N.D. 1
st
N.D. 1
st
N.D. 2
nd
N.D. 1
st
N.D. 1
st
Duration of second
interval(s)
0.19 ± 0.04 0.41 ± 0.09 0.14 ± 0.03 0.21 ± 0.04 N.D. 2
nd
N.D. 1
st
N.D. 1
st
* N.D. 2
nd
N.D. 1
st
Duration of third interval(s) 0.20 ± 0.04 0.39 ± 0.06 0.13 ± 0.02 0.24 ± 0.04 N.D. 2
nd
N.D. 2
nd
N.D. 1
st
* N.D. 2
nd
N.D. 2
nd
Duration of fourth interval(s) 0.21 ± 0.04 0.40 ± 0.08 0.14 ± 0.02 0.21 ± 0.05 N.D. 2
nd
N.D. 2
nd
N.D. 1
st
* N.D. 2
nd
N.D. 1
st
Duration of fifth interval(s)
b
0.22 ± 0.05 0.29 ± 0.13 0.16 ± 0.02 0.24 ± 0.05 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 2
nd
Interval between loudsongs
b
22.14 ± 11.93 12.43 ± 3.18 10.32 ± 8.16 17.64 ± 6.13 N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
N.D. 1
st
Notes structure
c
— — — — Yes Yes Yes Yes Yes Yes
Syntax
c
— — — — Yes Yes Yes Yes Yes No
Total differences in loud-
song characters
— — — — 4 2 2 7 4 1
Calls
b,d
n = 1
(7 calls)
n = 1
(5 calls)
e
n = 1
(3 calls)
— — — — — —
Number of notes 8 5–6 — 10–12 — — — — — —
Duration of the call (s) 0.97 ± 0.03 0.81 ± 0.06 — 1.01 ± 0.06 — — — — — —
Pace (number of notes/
duration)
8.27 ± 0.27 7.11 ± 0.10 — 10.87 ± 0.39 — — — — — —
a
N.D. 1
st
= not diagnosable according to the first criterion (i.e., ranges do not overlap); N.D. 2
nd
= diagnosable according to the first criterion but not diagnosable ac-
cording to the second criterion (i.e., means and standard deviations of the population with the smaller set of measurements [a] and the population with the larger set of
measurements [b] meet the requirement: X
a
+ t
a
SD
a
≤ X
b
– t
b
SD
b
where t
i
= the t-score at the 97.5 percentile of the t-distribution for n – 1 degrees of freedom; see text
for details). Asterisk indicates normally distributed vocal character in which populations differ diagnostically according to both first and second criteria of diagnosability
(see text for details).
b
Not included in multivariate analysis.
c
Pairwise diagnosability assessed qualitatively for these qualitative loudsong characters.
d
Not submitted to the pairwise diagnosability test.
e
No call sample available for this taxon.
aPriL 2012 — SyStematic reviSion of Hylopezus macularius — 343
of populations A, C, and D (in which birds often uttered songs
with six notes; Tables and ).
From a quali-quantitative perspective, the loudsong of
population A (Fig. A) was characterized by notes of nearly
identical shape and dominant frequency (. Hz), although
occasionally some notes (especially the fourth) exhibited a
slightly lower frequency around . Hz. Unlike in the re-
maining populations, the third, fifth, and sixth notes showed a
richer harmonic structure, with subharmonics present, which
fiG. 2. Bayesian molecular phylogeny obtained for different taxa grouped under Hylopezus macularius (following Krabbe and Schulenberg 2003) and
outgroups based on 1,352 bp of the 16S rRNA, ND2, and cytochrome-b mtDNA genes. Numbers above branches denote indices of Bayesian poste-
rior probabilities (top), maximum parsimony (MP; before slash), and maximum likelihood (ML; after slash) nodal support. Values for poorly supported
nodes, as indicated by bootstrap values <75% for (MP and ML) and <95% (Bayesian), are not shown. A pound sign (#) denotes nodes with bootstrap
values ≥75% only for MP. Names in parentheses at the tips of branches of H. macularius taxa denote Amazonian areas of endemism as recognized by
Da Silva et al. (2005). All four main lineages of the H. macularius complex are labeled A–D next to the taxon name with priority applicable to each
clade, which we interpret as species-level taxa (see text for details).
fiG. 3. Graphic representation of scores of the first two axes of a dis-
criminant function analysis separating specimens belonging to natural
populations A–D of Hylopezus macularius based on measurements of 9
morphometric characters (see Figs. 1 and 2 and text for details).
tabLe 2. Summary of classification accuracy of specimens of Hylop-
ezus macularius belonging to natural populations A–D (see Figs. 1 and
2) obtained through a discriminant function analysis based on mea-
surements of 9 morphometric characters (see text for details). Numbers
before and after slashes represent, respectively, values obtained with-
out and with jackknife procedures. The number of specimens of each
population included in the analysis is shown in parentheses.
Populations A B C D Correctness (%)
A (n = 16)
11 / 8 2 / 5 0 / 0 3 / 3 68.7 / 50
B (n = 23)
8 / 3 21 / 18 0 / 0 1 / 2 91.3 / 78
C (n = 11)
0 / 1 0 / 6 11 / 4 0 / 0 100 / 36
D (n = 14)
0 / 2 1 / 3 0 / 0 13 / 9 92.8 / 64
Total (n = 64)
12 / 14 24 / 32 11 / 4 17 / 14 86 / 61
344 — carneiro et aL. — auk, voL. 129
gave them a distinct raspy quality (Fig. A). In population A,
multisyllabic calls (Table and Fig. A) differed from those of
populations B and D by the number of notes (); its duration is
slightly longer than in population B and nearly the same as in
population D.
In population B, two characters readily distinguished its
loudsong from those of the remaining populations: () number
of notes, which in .% of the samples consisted of only four or
(more often) five, and less often (.% of the samples) six notes,
as in the remaining populations; and () pace, which was shorter
than in any other population (mean ± SD = . ± . s; Table
and Fig. B). e interval between the second and third notes
was longer than in any other population and longer than those be-
tween the other notes of the song (. ± . s; Table and Fig.
B). Calls differed most conspicuously from those of populations
A and D by the smaller number of notes (five or six) and shorter
duration (Table and Fig. B).
In population C, loudsongs were readily distinguished from
those of the other populations by the shorter duration of all note
intervals (ranging, on average, from . to . s), thus also yield-
ing an overall shorter loudsong (. ± . s; Table and Fig. C)
than in any other population. Furthermore, the syntax of the loud-
song in this population differed markedly from that of populations
A and B by having two different types of notes, following the pat-
tern A A B A B B, which was found also in population D; the notes
of type A were flatter in frequency, whereas notes of type B showed
a strongly ascending–descending pattern of frequency modula-
tion (Fig. C, D). Unfortunately, only loudsongs belonging to
recordings obtained at different localities were analyzed, and no
calls for population C were available.
Finally, population D’s loudsong followed the same pattern
observed in population C, in which the first, second, and fourth
notes and the third, fifth, and sixth notes formed two groups of
notes very similar in shape (Fig. D). Calls (Fig. C and Table ) dif-
fered from those in populations A and B by the larger number of
notes (–), which were longer in duration than those of popula-
tion B but nearly equal in duration to those of population A.
Taxonomy.—Our results indicated that each natural popu-
lation of H. macularius recognized in the present study (A–D)
can be mutually diagnosed through a combination of vocal, mo-
lecular, and morphological features and, thus, can be interpreted
as basal taxa deserving formal taxonomic recognition (Figs. , ,
and ; Table ). As discussed in detail in Appendix S (in online
supplementary materials; see Acknowledgments), already exist-
ing names are applicable to populations A (paraensis), C (dilutus),
and D (macularius), but no available name exists for population B,
which is formally described below. We also discuss below the evi-
dence supporting the treatment of all four natural populations of
H. macularius as species-level taxa.
discussion
Species limits and loudsong evolution in the H. macularius com-
plex.—Despite the sampling limitations of our study, we have
demonstrated that the current taxonomy of the Spotted Antpitta
complex contrasts strongly with its evolutionary history (Figs.
and ). When interpreted together, these findings allow a redefi-
nition of species limits in the H. macularius complex, whose taxa
have been historically treated as subspecies (Cory and Hellmayr
, Peters , Krabbe and Schulenberg ).
e statistically well-supported reciprocal monophyly re-
covered for the four lineages of the Spotted Antpitta identified
here (Fig. ), added to their vocal diagnoses (Fig. ), which remain
constant within each clade, are indicative of species-level status
under the phylogenetic species concept (PSC) and the biological
species concept (BSC). Under the PSC, their reciprocal combined
tabLe 3. Standard measurements of selected morphometric characters and weight data of natural populations A–D of the Hylopezus macularius com-
plex (Figs. 1 and 2). Values are in millimeters and represent means, with ranges in parentheses where appropriate.
Population Sex
Sample
size Wing Tail Tarsus Bill length Bill width Bill depth Body mass (g)
a
Male 4 83.6 (81.1–86.8) 38.5 (34.1–39.2) 35.4 (34.3–37.2) 13.6 (12.2–14.0) 5.5 (5.0–6.0) 5.3 (5.1–5.8) 43 (42–44)
A Female 3 81.4 (80.8–87.0) 38.8 (34.9–40.8) 35.2 (33.8–38.5) 12.3 (12.3–14.2) 5.2 (5.2–6.4) 5.6 (5.1–5.9) 45.20 (42–48.4)
Unknown 9 83.9 (81.1–87,0) 36.3(34.1–38.6 35.6 (36.2–38.5) 15.4 (12.6–14.2) 5.6 (5.4–5.9) 5.8 (5.7–6.0) —
Total 16 83.4 37.4 35.5 14.6 5.5 5.7 44.1
Male 10 86.1 (82.5–90.5) 40.1 (37.4–42.7) 35.0 (32.4–36.4) 13.2 (12.5–13.9) 5.8 (5.3–6.3) 5.8 (5.5–5.9) 44.1 (40–47)
B Female 3 85.2 (83.9–86.9) 40.8 (36.6–44.6) 34.4 (33.8–34.7) 13.0 (12.1–14.0) 5.9 (5.7–6.1) 5.8 (5.7–5.9) 42.4
Unknown 10 86.1 (83.1–90.1) 37.0 (33.9–39.6) 35. 2 ( 33 . 7–37. 6 ) 13.4 (12.7–14.0) 5.9 (5.1–7.7) 6.2(5.6–6.7) —
Total 23 86.0 38.9 35.0 13.3 5.9 6.0 43.2
Male 3 84.2 (83.8–84.8) 36.8 (36.4–37.0) 32.8 (30.9–34.3) 12.7 (12.2–13.3) 6.2 (6.1–6.4) 5.8 (5.8–5.9) 42.8 (39–45.5)
C Female 4 82.8 (79.9–85.4) 37.0 (35.0–39.2) 33.6 (31.8–35.6) 12.9 (12.2–12.9) 6.0 (5.8–6.1) 6.0 (5.9–6.2) —
Unknown 4 82.6 (79.0–84.3) 34.0 (31.0–35.6) 34.5 (34.0–35.2) 13.3 (13.0 –13.8) 6.3 (5.7–6.9) 5.9 (5.7–6.1) —
Total 11 83.1 35.8 33.7 13.0 6.2 5.9 42.8
Male 5 84.6 (82.9–88.9) 39.7 (38.2–41.4) 36.2 (34.8–38.6) 12.7 (11.5–13.6) 5.7 (5.2–6.1) 5.8 (5.5–6.3) 42 (41–43)
D Female 2 80.6 (78.2–83.0) 38.0 (35.0–41.1) 33.9 (32.2–35.6) 12.6 (11.7–13.6) 5.8 (5.7–5.9) 6.0 (5.8–6,2) 39.5
Unknown 7 82.8 (78.8–87.9) 35.4 (33.9–36.6) 35.6 (33.7–36.9) 13.0 (11.9 –13.9) 5.9 (5.6–6.8) 5.4 (5.1–5.6) 39
Total 14 83.1 37. 3 35.5 12.8 5.8 5.6 40.1
a
Body mass information not available for some specimens.
aPriL 2012 — SyStematic reviSion of Hylopezus macularius — 345
sampling. Even though only a new study, with much denser sam-
pling in terms of molecular markers and individuals throughout
Amazonia, could document the presence and contribution of in-
trogressed individuals to the genetic make-up of the four main
mitochondrial lineages of the Spotted Antpitta identified in our
study, the following two lines of evidence suggest that they never-
theless represent biological species. First, our genetic sampling of
individuals came from localities (all three Paranaíta locali-
ties—see Appendix S—were lumped together in Fig. and count
as a single locality) scattered across all areas of endemism inhab-
ited by the Spotted Antpitta (Krabbe and Schulenberg ; Fig.
) and, thus, can be regarded as representative at least over a broad
geographic scale. e strong statistical support obtained for the
mitochondrial DNA (mtDNA) reciprocal monophyly of all four
main lineages of the Spotted Antpitta (Fig. ) offer unequivocal ev-
idence that, even if present, introgression between lineages is not
molecular, vocal, and morphological diagnoses provide the ba-
sis for considering populations A–D of the Spotted Antpitta as
separate species, whereas under the BSC the congruent evolution
between genetic and vocal characters, added to the absence of ge-
netic paraphyly and vocally intermediate individuals in our sam-
pling, are also indicative of species-level status for these lineages,
even though such congruence does not guarantee reproductive
isolation. For instance, it is possible that our relatively sparse ge-
netic sampling could simply have missed introgressed individu-
als, particularly in areas where the ranges of parapatric taxa abut,
such as the Negro–Branco interfluve (where the presence and
identity of the local Spotted Antpitta taxon remain unknown; Fig.
) and headwaters of the Xingu River.
Furthermore, the absence of nuclear markers in our mo-
lecular phylogenetics analysis represents another source of dif-
ficulty in properly detecting introgressed individuals in our
fiG. 4. Dorsal (top) and ventral (bottom) views of some of the few plumage characters that distinguish natural populations A (MPEG 55691), B (MPEG
56066), C (MPEG 42750), and D (MPEG 66340) of Hylopezus macularius. Scale bar = 30 mm. See text for details.
346 — carneiro et aL. — auk, voL. 129
widespread enough to promote extensive admixture and merg-
ing of their mtDNA haplotypes; more importantly, the acquisition
and maintenance of reciprocal monophyly between parapatric sis-
ter lineages typically indicate the final stages of the speciation pro-
cess and that reproductive isolation has been reached (de Queiroz
, Patten ). Second, mtDNA reciprocal monophyly among
lineages is mirrored by significant differences in vocal characters,
for which our sampling is more extensive ( localities; Fig. ). e
importance of vocal characters as a premating isolating mecha-
nism in Grallariidae was recently underscored by a study show-
ing that loudsong variation appears to have a strong genetic basis
in this family (Cadena et al. ), which agrees with the pattern
documented here for the Spotted Antpitta. Furthermore, pairwise
tests of diagnosability contrasting the loudsongs of the unnamed
population B birds (see description below) with those of the other
Spotted Antpitta populations always found more than three char-
acters distinguishing them (Table ), which is the minimum num-
ber found to distinguish syntopic and allopatric pairs of sister
biological species of antbirds (amnophilidae), a family closely
related to the Grallariidae (Isler et al. , , ; Braun et
al. ; Isler and Whitney ). On the other hand, diagnos-
ability tests contrasting the loudsongs of dilutus, macularius, and
paraensis involved a smaller number of diagnostic characters: two
between macularius and paraensis and between dilutus and par-
aensis, and only one between dilutus and macularius (Table ).
Interestingly, using the same diagnosability tests employed in
our study, Chaves et al. () showed that only two diagnosable
vocal characters can distinguish three reciprocally monophyletic
populations of the Dull-mantled Antbird (Myrmeciza laemosticta)
complex (amnophilidae) separated by large genetic distances,
prompting those authors to suggest that only two diagnostic char-
acters may be necessary for two populations to be considered dis-
tinct species in this complex. In agreement with their study, our
data further indicate that, in fact, only one loudsong character
passing the diagnosability tests of Isler et al. () can distinguish
fiG. 5. Representative loudsong audiospectrograms of populations A–D of Hylopezus macularius (window type Hamming, window size 1,300 sam-
ples, time grid 90% overlap, and DFT size 16,384 samples). (A) Population A (H. paraensis), Caxiuanã, Pará, Brazil (C. A. Marantz, MLS 127444). (B)
Population B (H. whittakeri), Alta Floresta, Mato Grosso, Brazil (P. R. Isler, MLS 48068). (C) Population C (H. dilutus), Maraã, Lago Cumapi, Amazo-
nas, Brazil (A.A., 1 PAC). (D) Population D (H. macularius), Rupununi, Guyana (T. A. Parker III, MLS 73054).
tabLe 4. Summary of loudsong characters that distinguish pairs of popu-
lations of the Hylopezus macularius complex treated as species in the
present study. The total number of vocal variables distinguishing each
pair is shown in parentheses.
H. paraensis H. whittakeri H. dilutus
H. whittakeri
(4)
Number of
notes
Song pace
Notes structure
Syntax
— —
H. dilutus
(2)
Notes structure
Syntax
(7)
Number of notes
Song pace
Interval lengths
(2
nd
, 3
rd
, 4
th
)
Notes structure
Syntax
—
H. macularius
(2)
Notes structure
Syntax
(4)
Number of notes
Song pace
Notes structure
Syntax
(1)
Notes
structure
aPriL 2012 — SyStematic reviSion of Hylopezus macularius — 347
reciprocally monophyletic mtDNA lineages separated by com-
paratively high (i.e., .%) uncorrected genetic p-distances, which
are also known to differ in morphometrics and, to some extent, in
plumage as well (i.e., dilutus and macularius; Figs. and ). How-
ever, given our poor sampling of dilutus, it seems premature to infer
reproductive isolation between dilutus and macularius, given the
results of the loudsong diagnosability tests (Table ) and the un-
certain phylogenetic affinities of the former, whose recovered sis-
ter relationship to paraensis and the unnamed population B lacked
significant statistical support, thus not allowing the rejection of a
putative monophyly involving macularius and dilutus (Fig. ).
erefore, we interpret our combined results as evidence sup-
porting the recognition of four phylogenetic (populations A–D)
and three biological (populations A, B, and C+D) species in the
Spotted Antpitta complex. It should be mentioned, though, that
the biological species recognized here would likely not pass the
phenotypic diagnosability test described by Tobias et al. () for
assigning species rank based on the BSC because of their high de-
gree of plumage and morphometric conservatism (Fig. and Table
). Tobias et al. () advocated a test whereby it is impossible
for any pair of taxa to reach a minimum fixed threshold (defined
as seven character points) to be classified as separate biological
species, if they have experienced heterogeneous rates of evolution
in phenotypic traits, such as that observed between morphologi-
cal and vocal characters in the Spotted Antpitta complex. Hence,
this test does not seem appropriate to define interspecific limits in
the Spotted Antpitta complex because it devalues the contribu-
tion of bioacoustical characters by comparatively overestimating
the degree of morphological differentiation necessary for repro-
ductive isolation in a dense-forest-interior lineage known to use
vocal characters as a premating isolating mechanism (Cadena
tabLe 5. Summary of classification accuracy of loudsong types of Hylop-
ezus macularius belonging to natural populations A, B, and D (see Figs. 1
and 2) obtained through a discriminant function analysis based on mea-
surements of 10 characters related to the duration of individual notes and
intervals between notes. Numbers before and after slashes represent re-
spectively values obtained without and with jackknife procedures. The
number of tape-recordings included in the analysis is shown in parenthe-
ses. Population C was not included in the analysis because of a very small
sample size. See text for details.
Populations A B D Correctness (%)
A (n = 28)
26 / 15 0 / 1 2 / 12 92.8 / 54
B (n = 27)
0 / 0 27 / 27 0 / 0 100 / 100
D (n = 45)
4 / 6 1 / 3 40 / 36 88.8 / 80
Total (n = 100)
30 / 21 28 / 31 42 / 48 93 / 78
fiG. 6. Graphic representation of scores of the first two axes of a discrimi-
nant function analysis based on measurements of 10 characters related
to the duration of individual notes and intervals between notes distin-
guishing loudsongs of Hylopezus macularius natural populations A (H.
paraensis), B (H. whittakeri), and D (H. macularius; see Figs. 1 and 2 and
text for details). Population C (H. dilutus) was not included in the analysis
because of a very small sample size.
fiG. 7. Representative call audiospectrograms of populations A, B, and
D of Hylopezus macularius (window type Hamming, window size 800
samples, time grid 90% overlap, and DFT size 16,384 samples). (A) Pop-
ulation A, Caxiuanã, Pará, Brazil (A. Whittaker, 1 PAC). (B) Population
B, Parque Nacional da Amazônia, Pará, Brazil (A. Whittaker, 1 PAC). (C)
Population D, Bolívar, Venezuela (T. A. Parker III, MLS 34472).
348 — carneiro et aL. — auk, voL. 129
et al. ). Furthermore, species-rank diagnosability tests such
as those put forward by Tobias et al. () and Patten and Unitt
() do not incorporate genetic data in their calculations and,
thus, are not entirely suited to fully integrated multicharacter data
sets such as the one discussed here. On the other hand, popula-
tions A–D of the Spotted Antpitta all meet the requirements for
biological species ranking in the context of parapatry according
to the British Ornithologists’ Union (Helbig et al. ), given
that they () are diagnosable and () do not appear to hybridize.
However, as explained above, our poor sampling of population C
makes it still premature to infer that it is reproductively isolated
from the parapatric population D.
Taxonomic recommendations.—On the basis of the com-
bined character analyses presented and discussed above, we rec-
ommend the splitting of the H. macularius complex into four
phylogenetic (populations A–D) and three biological species (pop-
ulations A, B, and C+D), one of which (population B) has yet to be
named and is here described as
Hylopezus whittakeri, sp. nov.
Alta Floresta Antpitta
Torom-de-alta floresta (Portuguese)
Grallaria macularia (Temminck, ): Snethlage b:
(part: specimen from Cussary at ZMB examined).
Grallaria maculata diluta Hellmayr, : (part: speci-
men from Calama at AMNH examined).
Grallaria macularia paraensis Snethlage, : Snethlage
: (part: specimen from Cachoeira do Cahy at ZMB exam-
ined); Zimmer : (specimen from Limoal at AMNH exam-
ined); Cory and Hellmayr : (part: specimens from Calama
and Cachoeira do Cahy); Gyldenstolpe : (specimens from
Aveiro, Caxiricatuba, Marai, and Patinga at the Naturhistoriska
Riksmuseet, Stockholm, Sweden, not examined); Peters :
(part: specimens from the Madeira, Tapajoz, and the Jamauchim).
Hylopezus macularius paraensis (Snethlage, ): Krabbe
and Schulenberg : (part: specimens from south of the
Amazon, between the Madeira and Xingu rivers).
Hylopezus sp. nov.: Whittaker : .
Holotype.—MPEG . Skin, adult male, skull % ossi-
fied, testes × mm, collected in the understory of upland (terra
firme) forest on July by D. Davison, W. Figueiredo, and L. W.
Figueiredo in Belterra, Floresta Nacional do Tapajós, Sucupira base,
Km of the BR- highway, state of Pará, Brazil (°′′′S,
°′′′W). Tissue samples deposited at the Laboratório de Gené-
tica e Biologia Molecular, Campus Universitário de Bragança, Uni-
versidade Federal do Pará (LGBM) under accession number WN
. Mitochondrial S rRNA, ND, and cyt-b gene sequences de-
posited in GenBank (JQ, JQ, JQ).
Paratypes.—MN : skin, collected at Jamary, state of
Rondônia, Brazil. MPEG : skin, adult female, skull % ossi-
fied, collected in the understory of terra firme forest by G. P. Silva on
September at Rio Aripuanã, Humboldt, Cachoeira Darda-
nelos, state of Mato Grosso, Brazil (°′S, °′W). MPEG :
skin, adult male, skull % ossified, testes × mm, collected in
the understory of terra firme forest by M. S. Brígida on November
at Alvorada d’Oeste, Linha , Br Km , state of Rondônia,
Brazil (°′S, °′W). MPEG : skin, adult male, skull %
ossified, testes × mm, collected in the understory of terra firme
forest by T. Schulenberg on October at Cachoeira Nazaré,
west bank Rio Ji-paraná, state of Rondônia, Brazil (°′S, °′W).
MPEG : skin, adult male, skull % ossified, testes × mm,
collected in the understory of terra firme forest by A. T. Peterson on
October at Cachoeira Nazaré. MPEG : skin, female,
skull % ossified, ovarium × mm, collected in the understory of
terra firme forest by A. T. Peterson on October at Cachoeira
Nazaré. Tissue samples deposited at the Field Museum of Natural
History (FMNH) under accession number . Mitochondrial
S rRNA, ND, and cyt-b gene sequences deposited in GenBank
(JQ, JQ, JQ). MPEG : skin, adult male,
skull % ossified, testes × mm, collected in the understory of
terra firme forest by M. P. D. Santos and G. C. Silva on April
at Humaitá, Terra Indígena Parintintin, Aldeia Traíra-Chororó,
state of Amazonas, Brazil (°′S, °′W); prepared by M. Santa-
Brígida under field number MPDS . Tissue samples deposited at
MPEG. Mitochondrial S rRNA, ND, and cyt-b gene sequences
deposited in GenBank (JQ, JQ, JQ). MZUSP
: skin, male, collected at Fordlândia, state of Pará, Brazil.
Diagnosis.—Phenotypically, the new species can be unambigu-
ously assigned to the genus Hylopezus (Grallariidae) on the basis of its
relatively small size, tarsi without scutellations and inner edge con-
volutions, absence of rictal bristles, basal portion of primaries buffy
forming a contrasting wing bar, underparts predominantly white
with the chest more or less streaked with black, and tail less than half
as long as the wing (Lowery and O´Neill ). Within Hylopezus,
measurements and plumage coloration similar to those of other taxa
in the H. macularius complex, being closest to population A (to which
the name paraensis applies; Appendix S), from which it is separable
only through vocal and genetic characters (see below). Distinguish-
able from population C (to which the name dilutus applies; Appen-
dix S) by more conspicuous mantle shaft-streaks, more olivaceous
upperparts, much paler ochraceous subterminal bands of pectoral
spots, and wingbars occupying the entire length of the wing (Fig. ).
From population D (to which the name macularius is applicable; Ap-
pendix S), the new taxon is distinguished by much paler flanks and
more conspicuous mantle shaft-streaks (Fig. ).
Vocally, the new taxon is uniquely distinguished from all those
in the H. macularius complex by a loudsong normally composed of
five (rarely four or six) notes with identical shapes, in which the sec-
ond and third notes are separated by an unusually longer time inter-
val (Figs. and and Table ). Even though no tape-recordings that
could be unambiguously assigned to the holotype were obtained,
the characteristic loudsong of an individual of this new taxon was
tape-recorded at the exact same locality where the holotype was
collected on September by C. A. Marantz (MLS ;
Appendix S). Samples of loudsongs of the new taxon can be found
in Isler and Whitney () and Marantz and Zimmer ().
Description of holotype.—Crown and nape blackish neutral
gray (); lores buff–yellow (); broad eye ring yellow-ocher (c)
bordered by a continuous narrow blackish line; auricular region
with distinct black and buff streaks; malar region crossed by a con-
spicuous black streak, contrasting with the whitish throat and cen-
ter of chin. Upperparts olive () with conspicuous pale shaft-lines
on the central portions of mantle feathers. Breast strongly marked
with mixed black and buff-yellow () spots, with the black usu-
ally restricted to the v-shaped tips feathers. Flanks conspicuously
buff (); belly white; wing-coverts tipped yellow-ocher (c), faint
aPriL 2012 — SyStematic reviSion of Hylopezus macularius — 349
cinnamon (a) wing bars occupying the entire length of wing;
primary coverts blackish, contrasting with a well-defined tawny
() patch at the base of primaries; remainder of the primaries olive-
brown ().
Measurements of holotype.—Bill width at anterior end of
nostrils . mm; bill depth at anterior end of nostrils . mm; bill
length from anterior end of nostril to tip . mm; wing length .
mm; tail length . mm; tarsus length . mm; body mass . g.
Variation in the type series.—e type series includes eight spec-
imens: five males, two females, and one of unknown sex. No sexual
dimorphism was detected in any of the measurements taken (Table )
or characters studied. Two specimens (MPEG and ) pres-
ent more visible yellowish mantle shaft-streaks on the back than the
holotype. Other specimens (MZUSP and MPEG ) lack
the discrete black line that subdivides the wing bar as present in the
holotype and other specimens in the type series. Back color varies
in several specimens, going from black olive () (MZUSP ) to
olive-brown () (MN ); similarly, underwing coverts vary from
clay color (b) (MN ) to cinnamon () (MZUSP ).
Distribution.—e new taxon corresponds to natural popula-
tion B of the Hylopezus macularius complex, whose distribution
is restricted to the Madeira–Xingu interfluve in Brazilian south-
central Amazonia (Fig. ). So far, no Bolivian records belonging to
any taxon of the H. macularius are known (Remsen et al. ),
even though a paratype of the new taxon from Alvorada d’Oeste in
Rondônia (MPEG ) represents the southernmost and closest
record to the Bolivian border available to date (~ km; Fig. ).
Etymology.—We are pleased to name this new taxon af-
ter our colleague Andrew (“Andy”) Whittaker, whose contribu-
tions to Amazonian ornithology over the past years resulted
in the description and rediscovery of several species, new coun-
try records, and many noteworthy range extensions. e common
names Alta Floresta Antpitta (English) and torom-de-alta floresta
(Portuguese) refer to a popular birding destination in Brazil where
the new species is regularly found.
Habitat.—e new taxon is found on or very close to the ground
in dense undergrowth of humid lowland forest (sea level to ~ m),
with an apparent preference for wet or flooded areas in upland terra
firme forest, but also in drier transitional forest on the southern limit
of its range in northern Mato Grosso (Lees et al. ). It also seems
to be more commonly found around treefall gaps and streams, but
rarely in more open and disturbed areas. In common with most Gral-
lariidae, the new species seems very sensitive to the effects of habitat
loss, fragmentation, and perturbation, given that it was found in only
% of a sample of variably sized (.–, ha) forest patches in
the Alta Floresta region, northern Mato Grosso, where the smallest
occupied patch was ha (Lees and Peres , A. Lees pers. comm.).
We also recommend the treatment of the other natural popu-
lations of the Spotted Antpitta recognized in the present study as
species-level taxa according to either the BSC or the PSC, as follows.
Spotted Antpitta (Hylopezus macularius, Temminck []).—
Distributed on the Guianan shield from the eastern bank of the Ne-
gro and Branco rivers in Brazil eastward through eastern Venezuela
(Bolívar), Guyana, French Guiana, and the state of Amapá in Brazil
(Fig. ; see also Hilty , Restall et al. , Naka ). Statistical
support for the reciprocal monophyly of macularius with respect to
all other Spotted Antpitta lineages is very high according to all phy-
logenetic criteria employed (Fig. ). Average pairwise uncorrected
p-distances between macularius and other lineages are as follows:
.% (dilutus), .% (whittakeri), and .% (paraensis). On the other
hand, pairwise p-distances within macularius range from to
.% (average .%). As shown earlier, there are two and four diag-
nostic loudsong features distinguishing macularius from paraensis
and whittakeri, respectively, but only one separating it from dilutus,
the reason that both should still be treated as subspecies under the
BSC but as separate species under the PSC (Figs. and ; Tables
and ; and see above). Morphologically, only minor differences in
flank color and the intensity of pale shaft streaking on the back dis-
tinguish H. macularius from the other taxa.
Zimmer´s Antpitta (Hylopezus dilutus, Hellmayr, []).—
Distributed north of the Amazon from the western banks of
the Negro and Branco rivers in the Brazilian state of Amazo-
nas through southern Venezuela (Amazonas), southern Colom-
bia (Amazonas), eastern Ecuador, and northern Peru west of the
Ucayali River (Fig. ; Hilty and Brown , Ridgely and Green-
field , Schulenberg et al. ). Statistical support for the
reciprocal monophyly of dilutus with respect to other Spotted
Antpitta lineages is very high according to all phylogenetic cri-
teria employed, but consistently poor when considering its sister
relationship with the paraensis–whittakeri clade (Fig. ). ere-
fore, the possibility that dilutus is actually sister to macularius
cannot be entirely ruled out, even though this relationship was
never recovered by any of the phylogeny estimates obtained in our
study; furthermore, the highest pairwise genetic distances in the
Spotted Antpitta complex involve dilutus and macularius, indi-
cating an advanced degree of evolutionary independence even if
they were sister taxa. Average pairwise uncorrected p-distances
between dilutus and other lineages were as follows: .% (macu-
larius), .% (whittakeri), and .% (paraensis), whereas no genetic
divergence was detected between the only two individuals of H.
dilutus sequenced. Vocally, dilutus is distinguished from all other
lineages by a significantly shorter loudsong lasting ~ s (in other
lineages, loudsong usually lasts .–. s; Tables and and Fig.
), whereas morphologically birds from this group can be consid-
ered the most distinct among all Spotted Antpitta lineages, given
their significantly shorter tarsi, greater bill depth and width val-
ues, and a brownish rather than greenish-olivaceous back with lit-
tle or no pale shaft streaking (Table and Figs. and ). All things
considered, we recommend recognizing dilutus as a phylogenetic
species or as a subspecies of macularius until new data provide a
more accurate picture of its phylogenetic position and reciprocal
vocal diagnosis within the Spotted Antpitta complex. e com-
mon name proposed here honors J. T. Zimmer’s early contribution
to the taxonomy of the H. macularius complex (Zimmer ).
Snethlage´s Antpitta (Hylopezus paraensis, Snethlage []).—
Distributed south of the Amazon in Brazil from the Xingu River east-
ward in the state of Pará to the western part of the state of Maranhão,
and southward to southern Pará (Snethlage , Pacheco et al. ).
Statistical support for the reciprocal monophyly of paraensis with re-
spect to all other Spotted Antpittas is high according to all phyloge-
netic criteria employed (Fig. ). Average pairwise uncorrected p genetic
distances between paraensis and other lineages were as follows: .%
(macularius), .% (dilutus), and .% (whittakeri), whereas within
paraensis it reached only .% (average .%). ere are two diagnos-
tic loudsong features distinguishing paraensis from each dilutus and
macularius and four distinguishing it from whittakeri (Fig. and
350 — carneiro et aL. — auk, voL. 129
Table), the reason that it should be treated as a separate species un-
der both the BSC and PSC. Morphologically, paraensis is hardly distin-
guished from whittakeri and macularius by slight differences in flank
color and bill measurements, respectively. e common name pro-
posed here honors E. Snethlage´s early contributions to the taxonomy
of the H. macularius complex (Snethlage a, ).
AcknowledgMents
Online supplementary materials are available at dx.doi.org/./
auk... is work was made possible only by the tenacious
effort of many specimen collectors and tape-recordists, some of
them anonymous, who by years of intense field work in the Amazon
amassed the raw data necessary for our analyses. We thank the cura-
tors and curatorial assistants of the following collections for allow-
ing us to study and sequence skins, tissues, and vocalizations under
their care (for acronyms not defined here, see text): AMNH, Acad-
emy of Natural Sciences (ANSP), British Library Sound Archive
(BLS), CM, FMNH, LGBM, Laboratório de Genética e Evolução Mo-
lecular, Universidade de São Paulo (LGEMA), Macaulay Library of
Natural Sounds (MLS), MN, MZUSP, and ZMB. M. S. Hoogmoed
and M. A. Raposo (through CNPq grant no. /-) kindly
sent us high-resolution digital pictures of type specimens deposited,
respectively, at the Nationaal Natuurhistorisch Museum (RMNH)
and Naturhistorisches Museum (NHMV). S. H. Borges, S. Dantas,
B. M. Whitney, and A. Whittaker generously allowed us to analyze
recordings from their personal archives. Field and laboratory work
related to this study was funded through the following agencies and
institutions: Conservation International (CI)-Brazil, World Wildlife
Fund (WWF)-Brazil, Ministério do Meio Ambiente (MMA), Con-
selho Nacional de Desenvolvimento Científico e Tecnológico (CNPq;
grants no. /-, /-, and “INCT em Biodivers-
idade e Uso da Terra da Amazônia” no. /-), FMNH Mar-
shall Funds, and the National Science Foundation (DEB- and
DEB-). We thank A. Whittaker, P. Peloso, M. Sturaro, A.
Maciel, A. Lees, L. Joseph, and five anonymous reviewers for contri-
butions to earlier versions of the manuscript. L.S.C. received a gradu-
ate fellowship from Coordenação de Aperfeiçoamento de Pessoal de
Nível Superior (CAPES) during the study. A.A., I.S., and H.S. are sup-
ported by productivity fellowships from CNPq.
liteRAtuRe cited
B, M. J., M. L. I, P. R. I, J. M. B, M. B. R-
.
. Avian speciation in the Pantepui: e case of the
Roraiman Antbird (Percnostola [schistocichla] “leucostigma” sat-
urata). Condor :–.
C, C. D., B. L-L, J. M. B, N. K, N. H.
R, F. G. S, J. D. P, P. S.
. A
rare case of interspecific hybridization in the tracheophone sub-
oscines: Chestnut-naped Antpitta Grallaria nuchalis × Chest-
nut-crowned Antpitta G. ruficapilla in a fragmented Andean
landscape. Ibis :–.
C, R. A., C. W. C, K. M. F.
. Raven .
User’s Manual. Cornell Laboratory of Ornithology, Ithaca, New York.
C, J. C., A. M. C, M. J. M, C. D. C.
. Revising species limits in a group of Myrmeciza antbirds
reveals a cryptic species within M. laemosticta (amnophili-
dae). Condor :–.
C, C. B., C. E. H.
. Catalogue of birds of the
Americas. Field Museum of Natural History Zoological Series
vol. , part .
D S, J. M. C., A. B. R, G. A. B. D F.
.
e Fate of the Amazonian areas of endemism. Conservation
Biology :–.
Q, K.
. Ernst Mayr and the modern concept of spe-
cies. Proceedings of the National Academy of Sciences USA
:–.
D, R. W. R. J.
. Type specimens of birds in the National
Museum of Natural History, Leiden. Part . Passerines: Eury-
laimidae–Eopsaltriidae (Peters’s sequence). NNM Technical Bul-
letin :–.
D, A. J., A. R.
. BEAST: Bayesian evolu-
tionary analysis by sampling trees. BMC Evolutionary Biology :.
G, N.
. Preliminary diagnoses of some new birds
from Bolivia. Arkiv for Zoologi B, No. .
H, S. J.
.
Molecular phylogenetics and biogeography of
tanagers in the genus
Ramphocelus
(Aves).
Molecular Phyloge-
netics and Evolution :
–
.
H, T. A.
. BIOEDIT: A user-friendly biological sequence
alignment editor and analysis program for Windows //NT.
Nucleic Acids Symposium Series :–.
H, A. J., A. G. K, D. T. P, G. S, M.
C.
. Guidelines for assigning species rank. Ibis
:–.
H, C. E.
. e birds of the Rio Madeira. Novitates
Zoologicae :–.
H, S. L.
. Birds of Venezuela, nd ed. Princeton University
Press, Princeton, New Jersey.
H, S. L., W. L. B.
. A Guide to the Birds of
Colombia. Princeton University, Princeton, New Jersey.
I, M. L., P. R. I, R. T. B.
. Clinal varia-
tion in vocalizations of an antbird (amnophilidae) and impli-
cations for defining species limits. Auk :–.
I, M. L., P. R. I, B. M. W.
. Use of vocaliza-
tions to establish species limits in antbirds (Passeriformes: am-
nophilidae). Auk :–.
I, M. L., P. R. I, B. M. W.
. Species limits
in antbirds (amnophilidae): e Warbling Antbird (Hypocne-
mis cantator) complex. Auk :–.
I, M. L., B. M. W.
. Species limits in antbirds
(amnophilidae): e Scale-backed Antbird (Willisornis poecili-
notus) complex. Wilson Journal of Ornithology :–.
I, P. R., B. M. W.
. Songs of the Antbirds. [ree
CD set.] Cornell Laboratory of Ornithology, Ithaca, New York.
K, R. E., A. E. R.
. Bayes factor. Journal of the
American Statistical Association :–.
K, N., T. S. S.
. Family Formicar-
iidae (ground-antbirds). Pages – in Handbook of the
Birds of the World, vol. : Broadbills to Tapaculos (J. del Hoyo,
A. Elliott, and D. A. Christie, Eds.). Lynx Edicions, Barcelona,
Spain.
L, A. C., B. J. W. D, A. V. G. O, C. A. P.
. Avifauna of a structurally heterogeneous forest landscape
aPriL 2012 — SyStematic reviSion of Hylopezus macularius — 351
in the Serra dos Caiabis, Mato Grosso, Brazil: A preliminary
assessment. Cotinga :–.
L, A. C., C. A. P.
. Habitat and life history determi-
nants of antbird occurrence in variable-sized Amazonian forest
fragments. Biotropica :–.
L, G. H. J., J. P. ON.
. A new species of Grallaria
from Peru, and a revision of the subfamily Grallarinae. Auk :–.
M, S.
. Rediscovery of Hylopezus (macularius) auricularis:
Distinctive song and habitat indicate species rank. Auk :–
.
M, C. M., K. J. Z.
. Bird Voices of Alta Flo-
resta and Southeastern Amazonian Brazil. Macaulay Library,
Cornell Laboratory of Ornithology, Ithaca, New York.
N, L. N.
. Avian distribution patterns in the Guiana Shield:
Implications for the delimitation of Amazonian areas of ende-
mism. Journal of Biogeography :–.
N, J. A. A.
.
MRMODELTEST, version . Program
distributed by the author.
Evolutionary Biology Centre, Uppsala
University, Uppsala, Sweden.
P, J. F., G. M. K, A. A, B. M. W, J.
M, K. J. Z, A. W, P. S. M. F, M. F. C.
L, D. C. O.
. An avifaunal inventory of the
CVRD Serra dos Carajás project, Pará, Brazil. Cotinga :–.
P, S., A. M, S. R, W. O. MM, L. S,
G. G.
. e Simple Fool’s Guide to PCR. Uni-
versity of Hawaii Press, Honolulu.
P, M. A.
. Null expectations in subspecies diagnosis.
Pages – in Avian Subspecies (K. Winker and S. M. Haig,
Eds.). Ornithological Monographs, no. .
P, M. A., P. U.
. Diagnosability versus mean
differences of Sage Sparrow subspecies. Auk :–.
P, J. L.
. Check-list of the Birds of the World, vol. .
Museum of Comparative Zoology, Cambridge, Massachusetts.
P, D.
. MODELTEST: Phylogenetic model averaging.
Molecular Biology and Evolution :–.
P, N. S., P. J. L.
. Manual of Ornithology:
Avian Structure and Function.
Yale University Press, New
Haven, Connecticut.
R, J. V., J., C. D. C, A. J, M. N, J. F.
P, M. B. R, T. S. S, F. G. S, D. F.
S, K. J. Z.
. [Version: January .] A
classification of the bird species of South America. American
Ornithologists’ Union. [Online.] Available at www.museum.lsu.
edu/~Remsen/SACCBaseline.html.
R, R., C. R, M. L.
. Birds of North-
ern South America: An Identification Guide, vols. and . Yale
University Press, New Haven, Connecticut.
R, R. S., P. J. G.
. e Birds of Ecuador,
vol. . Cornell University Press, Ithaca, New York.
R, R.
. New genera, species and subspecies of Formicari-
idae, Furnariidae, and Dendrocolaptidae. Proceedings of the Bio-
logical Society of Washington :–.
R, F., J. P. H.
. MRBAYES : Bayes-
ian phylogenetic inference under mixed models. Bioinformatics
:–.
S, J., E. F. F, T. M.
. Molecular
Cloning: A Laboratory Manual, nd ed. Cold Spring Harbor Lab-
oratory Press, Cold Spring Harbor, New York.
S, T. S., D. F. S, D. F. L, J. P. O’N,
T. A. P. . Birds of Peru. Princeton University Press,
Princeton, New Jersey.
S, F. B. . Naturalist’s Color Guide. American Museum of
Natural History, New York.
S, F. B. . Naturalist’s Color Guide. Part III. American
Museum of Natural History, New York.
S, E. a. Neue Vogelarten aus Südamerika. Ornithol-
ogische Monatsberichte :–.
S, E. b. Über unteramazonische Vögel. Journal of
Ornithology :–.
S, E. . Neue Vogelarten aus
Amazonien. Ornitholo-
gische Monatsberichte :–.
S, E.
. Catálogo das aves amazônicas. Boletim do
Museu Paraense Emílio Goeldi :–.
S, M. D., J. C. A, D. E. D, T. Y, D. P.
M.
. Primers for a PCR-based approach to mitochon-
drial genome sequencing in birds and other vertebrates. Molecu-
lar Phylogenetics and Evolution :–.
S, D. L.
. PAUP*: Phylogenetic
Analysis Using Par-
simony (*and Other Methods). Beta Version v.b. Sinauer
Associates, Sunderland, Massachusetts.
T, J. D., D. G. H, T. J. G. .
CLUSTAL W: Improving the sensitivity of progressive multi-
ple sequence alignment through sequence weighting, position-
specific gap penalties and weight matrix choice. Nucleic Acids
Research :–.
T, J. A., N. S, C. N. S, J. D. P,
L. D. C. F, N. J. C. . Quantitative crite-
ria for species delimitation. Ibis :–.
W, B. M., J. F. P, P. R. I, M. L. I. .
Hylopezus nattereri (Pinto, ) is a valid species (Passeri-
formes: Formicariidae). Ararajuba :–.
W, A. . Pousada Rio Roosevelt: A provisional avi-
faunal inventory in south-western Amazonian Brazil, with
information on life history, new distributional data and com-
ments on taxonomy. Cotinga :–.
W, A., A. A, F. P. . Corrections
and additions to an annotated checklist of birds of the upper
rio Urucu, Amazonas, Brazil. Bulletin of the British Ornitholo-
gists’ Club :–.
W, M. P . . Beiträge zur Naturgeschichte von Brasil-
ien, Vögel . Landes-Industrie-Comptoirs, Weimar, Germany.
W, E. O. . e Behavior of Bicolored Antbirds. University
of California Publications in Zoology No. .
X, X., Z. X. . DAMBE: Software package for data
analysis in molecular biology and evolution. Journal of Hered-
ity :–.
Z, J. T. . Studies of Peruvian birds. XII. Notes on
Hylo-
phylax, Myrmothera
, and
Grallaria
. American Museum Novi-
tates :–.
Associate Editor: L. Joseph
