MOLECULAR

PHYLOGENETICS
AND
EVOLUTION
Molecular Phylogenetics and Evolution 31 (2004) 139–152
www.elsevier.com/locate/ympev

Towards a phylogenetic framework for the evolution

of shakes, rattles, and rolls in Myiarchus tyrant-flycatchers
(Aves: Passeriformes: Tyrannidae)
Leo Joseph,a,* Thomas Wilke,a,1 Eldredge Bermingham,b
Deryn Alpers,a,2 and Robert Ricklefsc
a
Department of Ornithology, The Academy of Natural Sciences, 1900 Benjamin Franklin Parkway, Philadelphia, PA 19103-1195, USA
b
Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Panama
c
Department of Biology, University of Missouri – St. Louis, 8001 Natural Bridge Road, St. Louis, MO 63121-4499, USA

Received 1 May 2003; revised 25 June 2003

Abstract

A phylogeny of 19 of the 22 currently recognized species of Myiarchus tyrant-flycatchers is presented. It is based on 842 bp of
mitochondrial DNA (mtDNA) sequences from the ATPase subunit 8 and ATPase subunit 6 genes. Except for the morphologically
distinct M. semirufus, mtDNAs of the remaining 18 species fall into either of two clades. One comprises predominantly Caribbean
and Central and North American taxa (Clade I), and the other is of predominantly South American taxa (Clade II). The phylogeny
is only very broadly concordant with some vocal characters and also with the limited morphological diversity for which the group is
well known. Paraphyly in several species (M. swainsoni, M. tuberculifer, M. ferox, M. phaeocephalus, M. sagrae, M. stolidus) suggests
that morphological evolution, albeit resulting in limited morphological diversity, has been more rapid than that of mtDNA, or that
current taxonomy is faulty, or both. A South American origin for Myiarchus is likely. Dispersal and vicariance both appear to have
been involved in generating the present-day distribution of some species. Relatively recent dispersal events out of South America are
inferred to have brought species of Clades I and II into broad sympatry. Jamaica has been colonized independently at least twice by
members of Clades I and II. The phylogeny brings a historical perspective that in turn suggests that ecological study of closely
related species from within each major clade where they are sympatric will be especially rewarding.
Ó 2003 Elsevier Inc. All rights reserved.

Keywords: Myiarchus; Systematics; Phylogeny; Passerines; Suboscines

1. Introduction research has traditionally focussed on higher-level rela-

tionships within the assemblage and has been based on
Tyrant-flycatchers are a diverse assemblage of several internal and external anatomy (e.g., Lanyon, 1984, 1985,
hundred species of suboscine passerine birds that range 1986; Traylor, 1977) with contributions from DNA
across the Americas (Ridgely and Tudor, 1994). Their hybridization (McKitrick, 1985; Sibley and Ahlquist,
taxonomic, ecological, and behavioural diversities have 1985). More recently, DNA sequencing has made it
received much attention (e.g., Fitzpatrick, 1985; Lan- feasible to examine relationships among species in some
yon, 1984; Smith, 1971; Vuilleumier, 1971). Systematic of the larger, morphologically homogeneous genera
(Chesser, 2000; Roy et al., 1999). The systematics and
biogeography of one such genus, Myiarchus Cabanis,
*
Corresponding author. Fax: 1-215-299-1182. 1844, is the focus of the present paper.
E-mail address: joseph@acnatsci.org (L. Joseph). Twenty-two species are currently recognized in My-
1
Present address: Department of Microbiology and Tropical iarchus (AOU, 1998; Lanyon, 1967, 1978). Their centre
Medicine, George Washington University Medical Center, Ross Hall,
Room 731, 2300 Eye Street NW, Washington, DC 20037, USA.
of diversity is the humid Neotropics though some species
2
Present address: Department of Veterinary Biology, Murdoch range north into temperate parts of the United States
University, Perth, WA 6150, Australia. and south as far as Argentina and Uruguay. They are

1055-7903/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S1055-7903(03)00259-8
140 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152

found mainly in secondary growth, rainforest edges and (mtDNA) sequences and explore its implications in
open, wooded habitats. Myiarchus has long been re- testing LanyonÕs (1967, 1978, 1985) definitions of the
nowned for its extreme morphological homogeneity monophyly of Myiarchus, species limits and historical
(e.g., Coues, 1872). Most species have a grey back and biogeography. We also address the groupÕs ancestral
chest with yellow belly (white in two species), and vary- distribution at a broad geographical scale (South
ing degrees of rufous or brown in the wings and tail. America or North America or Central America or Ca-
M. semirufus of arid coastal Peru is the only exception ribbean) and review our data for instances of dispersal
being entirely rufous and black. Lanyon (1978) noted into and out of these broad regions. We plan separate,
that the ‘‘remarkable uniformity of plumage coloration more detailed discussions of intraspecific phylogeogra-
and pattern, a lack of appreciable sexual dimorphism, phy for later publications but have already examined the
interspecific overlap in most mensural characters, a phylogeography and evolution of migration in the M.
misunderstanding with respect to geographical variation, swainsoni complex (Joseph et al., 2003). Renewed con-
and inadequate or unreliable diagnostic characters cern has been expressed lately (e.g., Ballard et al., 2002;
have led many museum-based workers to despair that Weckstein et al., 2001) about making taxonomic chan-
a reasonably satisfactory classification of the group ges based solely on mtDNA and so we offer our findings
would ever be achieved.’’ Lanyon (1967, 1978) analysed with this in mind.
Caribbean and South American taxa for vocal characters
that he termed rattles, rolls, hiccups, and so on. His
analyses of vocal and plumage variations are the basis 2. Materials and methods
of current taxonomy of Myiarchus (AOU, 1998; Lanyon,
1967, 1978). Unless indicated otherwise, we follow 2.1. Specimens
LanyonÕs (1967, 1978) taxonomy. Development of a
phylogenetic framework for Myiarchus within which the Of the 123 sequences studied, 117 were from 19
history of variation in these characters can be examined nominal Myiarchus species (Appendix A). Mean sample
is therefore a challenge well-suited to molecular methods. size per species was six, with a range from 1 (for 10
Lanyon (1985) formally rediagnosed Myiarchus with species) to 49 (one species). Tissue samples of liver and
characters of the nest and structure of the nasal capsule heart were used for all DNA extractions except one, M.
and syrinx. He found that neither of three problematic venezuelensis 01.VE, which was only available as blood.
species, M. validus, M. magnirostris and M. semirufus, The remaining six sequences were outgroups chosen
each of which has at some time been assigned to a from within the Tyranninae (sensu Traylor, 1977) and
monotypic genus, had characters contradicting their on the basis of LanyonÕs (1982, 1985) assessments of the
placement in Myiarchus. M. validus is an unusually large nearest relatives of Myiarchus. These were Tyrannus
but typically plumaged Myiarchus from Jamaica. M. melancholicus (n ¼ 4), T. caudifasciatus (n ¼ 1), and
magnirostris is an unusually small but also unexception- Rhytipterna immunda (n ¼ 1).
ally plumaged Myiarchus endemic to the Galapagos Is-
lands. Relative to all other Myiarchus spp, M. semirufus 2.2. DNA isolation, PCR amplification, and sequencing
from arid coastal Peru and Ecuador is uniquely rufous-
and black-plumaged though it is a typical Myiarchus in DNA extraction protocols and PCR conditions var-
vocalizations, behaviour, nest and eggs (Lanyon, 1975). ied slightly in our two laboratories but were typically as
The circumscription of Myiarchus with respect to the in- follows (variant protocols are described in Joseph et al.,
clusion of these species needs to be verified by indepen- 2003). DNA was extracted by digesting 0.1–0.5 g of
dent sources of data, such as DNA sequences. Lanyon ground tissue (pectoral muscle or liver) or 20–40 ll
(1967) also offered specific hypotheses of the historical of blood in 500 ll of 2
CTAB buffer solution and 10 ll
biogeography of Caribbean taxa, especially those of Ja- of a 10 mg/ml Proteinase K solution at 54 °C for 6–14 h.
maica. He proposed that M. stolidus evolved from a first, A typical DNA extraction yielded a volume of 400 ll of
late Tertiary colonization of Jamaica, probably from which 1 ll was used to seed polymerase chain reaction
Central America, and that M. sagrae, M. antillarum, and (PCR) amplifications. Amplification was carried out in
M. oberi were then derived from M. stolidus. He next 50 ll reactions under the following conditions: 1 ll DNA
proposed that M. validus and then M. barbirostris resulted (15–20 ng), 1
PCR Buffer II (Perkin–Elmer, Forest
from two later invasions in the early and late Pleistocene, City, CA), 2.0 mM MgCl2 , 2 lM each of dATP, dCTP,
respectively, the latter derived from Central American M. dGTP, and dTTP, 2 pM of each primer, and 0.25 ll of
tuberculifer. Lastly, he proposed a fourth colonization of Amplitaq polymerase (Perkin–Elmer, Forest City, CA).
the Caribbean resulting in the evolution of M. nugator of Reactions were denatured for 3 min at 94 °C, followed
Grenada from M. tyrannulus in Venezuela. by 25 thermal cycles of 94 °C denaturing for 45 s, 54 °C
This paperÕs first main goal is to present a phylogeny annealing for 45 s and 72 °C extension for 1 min, and
of Myiarchus based on partial mitochondrial DNA terminated with a 5-min extension at 72 °C. We used the
 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152 141

primers COIIGQL (50 -GGACAATGCTCAGAAATC throughout the dataset (see Fig. 1), we did not attempt
TGCGG-30 ) and COIIIHMH (50 -CATGGGCTGGGG to remove problematic lineages or to otherwise correct
TCRACTATGTG-30 ) developed by Seutin and Ber- for rate heterogeneity.
mingham (details at http://nmg.si.edu/bermlab.htm to Given the potential limitations of Bayesian inference
amplify the overlapping genes for subunits 8 and 6 of (see above), we also performed Maximum Parsimony
ATPase (hereafter ATPase 8 and 6). (MP), Maximum Likelihood (ML), and Neighbor-join-
ing (NJ) analyses in PAUP* v. 4.0b10 (Swofford, 2002)
2.3. Phylogenetic analyses either on the whole dataset (NJ, Tamura and Nei (1993)
distances) or on a reduced version in order to test
Prior to the phylogenetic analysis, Modeltest 3.06 whether different optimality criteria and different models
(Posada and Crandall, 1998) was used in order to find of sequence evolution affect our phylogenetic hypothe-
the optimal model of DNA substitution. The model sis.TBR (tree bisection-reconstruction) branch swapping
selected was F81 + C (Felsenstein, 1981) with base fre- was used in MP and ML analyses. The reduced dataset
quencies of A ¼ 0:4043, C ¼ 0:1639, G ¼ 0:16541, T ¼ comprised the minimum number of haplotypes to rep-
0:26639; and a gamma distribution shape parameter of resent all clades found in all previous analyses (n ¼ 31).
a ¼ 1:6296. For phylogenetic reconstruction, we used It was analyzed with MP and ML under the best-fit
MrBayes 3.0b4 (Huelsenbeck and Ronquist, 2001). The model selected by Modeltest (HKY + C model with
program is based on Bayesian inference, an approach gamma shape parameter of 0.2442 and transi-
that performs exceptionally well in supporting correct tion:transversion ratio of 10.7065). Individuals used in
groupings compared to traditional non-parametric the reduced datasets are indicated in Appendix A. MP
bootstrap methods (Alfaro et al., 2003). It should be and ML analyses used 100 bootstrap pseudoreplicates.
pointed out that Bayesian inference appears to be less
conservative than bootstrap approaches and seems to be 2.4. Other analyses
sensitive to small model mis-specifications (see Douady
et al., 2003). The ancestral distribution of Myiarchus was recon-
First, we compared several independent runs using structed using parsimony in MacClade 4.0 (Maddison
the default random tree option and the starting pa- and Maddison, 2002). The distribution of the taxon to
rameters suggested by Modeltest to monitor the con- which each sample belonged was first coded as South
vergence of the ln likelihoods of the trees. The ln America, Central America, North America, Caribbean,
likelihoods started at around )16,100 and converged or combinations of these as appropriate using AOU
upon a stable value of about )5900 after 75,000 gener- (1998) and Lanyon (1967, 1978) (Appendix A). Distri-
ations. It can be assumed that once this stable value is butions were then reconstructed on various MP (un-
reached, MrBayes is sampling trees according to their weighted and 1:5 transition:transversion) and ML
posterior probabilities. Based on these preliminary topologies. These analyses were repeated with Mexico
analyses, we did a final run using the Markov chain coded as within North America, within Central Amer-
Monte Carlo approach without Metropolis-coupling ica, or as a separate character state. Use of dispersal-
(single chain) with 1,000,000 sampled generations and vicariance analysis using DIVA 1.1 (Ronquist, 1996)
with the current tree saved at intervals of 10 generations. was inhibited by that programÕs need for strictly bifur-
Note that there is no good criterion that could be used cating trees and the need to use all equally parsimonious
to determine the optimal number of generations neces- trees where a consensus includes polytomies (Ronquist,
sary for obtaining a good consensus tree. However, gi- 1996, 1997). This was impractical in the present case,
ven that the ln likelihood values reached a plateau after e.g., 523 equally parsimonious trees from the n ¼ 31
75,000 generations, the number of 1,000,000 chosen here dataset. Accordingly, DIVA was used only to address
appears to be sufficient. Finally, a 50% majority rule tree the ancestral area of Myiarchus overall and not the
was constructed from all sampled trees with the first number of dispersal events in the history of individ-
10,000 trees ( ¼ 100,000 generations) ignored as burn-in. ual species. This was feasible because all phyloge-
In order to test whether our dataset is suitable for netic analyses produced bifurcating trees at the base of
assessing the relative timing of cladogenic events, we Myiarchus.
performed a log-likelihood ratio test (Huelsenbeck and
Rannala, 1997) with our full dataset based on 1,000,000
sampled Bayes generations with and without the 3. Results
constraint of the molecular clock. The test clearly re-
jected the molecular clock (log L0 ¼ 5944:35, log L1 ¼ 3.1. Sequence characteristics
5862:64, 2 log K ¼ 163:43, df ¼ 75:2, p ¼ 0:00),
indicating that the rates among lineages are not Eight hundred and forty two base pairs (bp) were
equal. Moreover, as highly unequal rates were obvious sequenced from the overlapping ATPase 8 (165 bp plus
142 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152

Fig. 1. Bayes phylogram for Myiarchus and outgroups based on 842 bp of mitochondrial sequences from the ATPase subunits 8 and 6 genes showing
the 50% majority-rule consensuses of topologies sampled during the Bayesian search. The scale bars indicate the expected number of substitutions
according to the model of sequence evolution applied. Posterior probabilities are provided for each clade. The phylogram shows the separation of M.
semirufus from all other Myiarchus species and the division of the remaining species into two clades (numbered with Roman numerals). *For
subspecies assignment see Joseph et al. (2003). Specimen codes are as in Appendix A.

TAA stop codon) and ATPase 6 genes (681 bp plus ATPase sequences in GenBank (results not shown).
TAA). All sequences showed the expected 10 bp overlap Relative to the homologous amino acid sequence in the
in different reading frames (ATGAACYTAA) between chicken Gallus gallus, Myiarchus and outgroups had a
the ATG start codon of ATPase 6 and the TAA stop single amino acid insertion in ATPase 8 between
codon of ATPase 8. Base frequencies were 0.29 (A), 0.34 positions 44 and 45, as in other passerines (e.g., H€
aarlid
(C), 0.08 (G), 0.28 (T), and showed the deficiency of and Arnason, 1999; Hunt et al., 2001; Lovette and
guanine typical of avian mtDNA. All sequences were Bermingham, 1999). This site was polymorphic for
translated into amino acids and aligned with other avian threonine (most Myiarchus, R. immunda), alanine (one
 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152 143

M. tuberculifer, ANSP 10481), isoleucine and leucine M. semirufus was either the sister to all other My-
(Tyrannus spp). Among the 117 Myiarchus samples, 560 iarchus (Bayesian analysis), sister to the outgroup R.
characters were constant, 193 were parsimony infor- immunda (ML), or one lineage of a trichotomy with R.
mative, and 89 were variable and parsimony uninfor- immunda and all other Myiarchus (MP). Divergences
mative. Percentages of variable sites were 35.1% in between M. semirufus and the two Myiarchus clades
ATPase 8 and 33.1% in ATPase 6. These values are (14.00  0.37%, 13.99  0.72%) were the same as that
comparable with those reported by Hunt et al. (2001) for between it and R. immunda (13.90%). M. semirufus was
Caribbean mimids (33.9, 30.7%, respectively). Mean tree 18.47  1.27% divergent from the two Tyrannus species.
length of 10,000 random trees was 3164.22  95.88 and Within Clade I, the pairs M. oberi–M. antillarum and
the data contained significant phylogenetic signal M. stolidus–M. sagrae were strongly supported but M.
(g1 ¼ 0:32, p < 0:01; Hillis and Huelsenbeck, 1992). stolidus and M. sagrae were each paraphyletic with re-
GenBank accession numbers of the sequences used in spect to each other (Fig. 1). M. nugator nested within M.
this study are listed in Appendix A. tyrannulus. Sequence divergences among Clade I species
ranged from maxima between 4 and 9% (M. validus, M.
3.2. Phylogeny: overview oberi, andM. antillarum vs other Clade I species) to
lower values of ca. 2% (M. oberi vs M. antillarum) and
Results of NJ, MP, and ML analyses were consis- <1% with paraphyly (M. stolidus vs M. sagrae) or
tent with the Bayesian analysis of the entire dataset without it (M. tyrannulus). Divergences of 6% or greater
and usually with bootstrap values >80%. For brevity, in Clade I involved pairwise comparisons with M. yu-
we here focus on the results of Bayesian analysis and catanensis and M. validus.
cite MP, ML, and NJ analyses only where they pro- Within Clade II, M. ferox, M. panamensis, M. vene-
duced relationships that differed from the Bayesian zuelensis, and one M. phaeocephalus sample (ANSP
analysis with bootstrap support >70%. Elsewhere we 1687) were monophyletic (Fig. 1) thus rendering M.
will present more detailed discussions of our data in phaeocephalus paraphyletic. Next, M. swainsoni (except
terms of how they relate to the phylogeography and the nominate subspecies M. s. swainsoni) was sister to
population history of M. tyrannulus, M. tuberculifer, M. tuberculifer and M. barbirostris. With MP and ML,
and M. ferox. however, these same sequences formed a polytomy (72–
All Myiarchus taxa except M. semirufus fell in either 75% MP, 70% ML). Divergences among these taxa were
of two main clades. Clade I had 34 sequences from 2–2.75%. Closest relatives of M. s. swainsoni and M.
Caribbean, Central American and North American cephalotes within Clade II could not be discerned in any
species (M. validus, M. stolidus, M. sagrae, M. oberi, M. analysis.
antillarum, M. yucatanensis, M. cinerascens, and M.
nugator) and all sampled populations ofM. tyrannulus, 3.3. Phylogeny: individual species complexes
which included widely dispersed South American sam-
ples. Clade II had 82 sequences from South American M. s. swainsoni is not closely related to the rest of the
species (M. swainsoni, M. ferox, M. phaeocephalus, M. M. swainsoni complex. MtDNAs from remaining
cephalotes) and M. barbirostris of Jamaica and all se- members of the M. swainsoni complex fell into two
quences of Central and South American M. tuberculifer. clades, which were more closely related to M. tubercu-
Tamura–Nei divergences between these clades and out- lifer–M. barbirostris than to M. s. swainsoni. This com-
groups ranged from 13 to 20%. Estimates of average plex has been discussed more fully elsewhere (Joseph
pairwise sequence divergence within Clades I and II et al., 2003; see Section 4).
made from the n ¼ 31 dataset removed bias from mul- M. tuberculifer was paraphyletic with respect to M.
tiple intraspecific comparisons in species with larger barbirostris although the posterior probability of 0.77
sample sizes. These were similar at 5.04 and 4.42%, re- for uniting these two species was low (Fig. 1). Net di-
spectively (Table 1). Pairwise divergences within Clade I vergence between the two clades of M. tuberculifer was
were either less than 2% or greater than 4%. 2.04%. M. t. atriceps was paraphyletic. M. barbirostris,

Table 1
Sequence divergence comparisons in Clades I and II
Clade I Clade II
Pairwise Mean sequence Pairwise Mean sequence
comparisons divergence (%) comparisons divergence (%)
Full dataset, n ¼ 123 561 3.53  2.09 3240 2.56  1.84
Reduced dataset, n ¼ 31 21 5.04  1.24 66 4.42  1.15
Note that the n ¼ 31 dataset excludes 44 of 49 M. swainsoni individuals within which net diversity is zero.
144 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152

though itself monophyletic, was sister to a predomi- 3.4. Biogeography

nantly Central and North American clade of M. tuber-
culifer samples. Reconstruction of distributions showed: a South
The single M. venezuelensis sample nested within M. American origin for Myiarchus and Clade II (arrow 1,
ferox making the latter paraphyletic (Fig. 1). The 0.2% Fig. 2), and dispersal out of South America to the Ca-
divergence between M. venezuelensis and M. ferox was ribbean with the origin of Clade I (arrow 2, Fig. 2). How
the same order of magnitude as that among M. ferox Mexico was coded did not affect these results. Recon-
(0.5%). MP bootstrapping supported only a polytomy of struction on the weighted transversions MP bootstrap
three clades within M. ferox (62–81%) and posterior topology with the n ¼ 31 dataset yielded the same re-
probabilities for two of these clades were relatively low sults except when Mexico was coded as a separate
at 0.80 and 0.62, also giving weak support for structure character state. Under that coding scheme, the ancestral
within this species. One had three western South distribution of Myiarchus generally and Clade 1 specif-
American M. f. ferox and the single M. venezuelensis. ically were equivocal. Reconstructions on the un-
The second had Guyanan and north-east Brazilian weighted MP bootstrap and ML topologies with the
(Par
a) samples from the zone of intergradation between n ¼ 31 dataset were similarly inconsistent; when an area
M. f. ferox and M. f. brunnescens identified by Lanyon of origin for Myiarchus was unequivocally suggested it
(1978). The third had more Guyanan M. f. ferox–M. f. was always South America. DIVA also found a South
brunnescens intergrades as well as a south central Bra- American origin for Myiarchus (not shown). Dispersal
zilian (Rond^ onia) M. f. ferox. into and out of South America was evident in the history
One M. phaeocephalus sample, ANSP 1687, and M. of some species. M. tuberculifer and M. barbirostris
panamensis were basal to M. ferox–M. venezuelensis showed dispersal out of South America (arrows 3 and 4,
(Fig. 1) in all analyses, thus making M. phaeocephalus Fig. 2). Whether the evolution of the Caribbean popu-
paraphyletic. The associated voucher, ANSP 181729, is lations of M. tyrannulus involved dispersal into or out of
correctly identified as M. phaeocephalus interior (pers. South America is unclear (arrow 5, Fig. 2). A South
obs., L.J.). Its mtDNA sequence is translatable and we American area of origin for the entire M. tyrannulus
are satisfied that a valid, uncontaminated mtDNA se- complex was never suggested. A relatively complex hi-
quence has been obtained. Net divergence between story of movement among South, North, and Central
ANSP 1687 and M. ferox–M. venezuelensis was 3.2% America is suggested for M. tyrannulus complex re-
and that between ANSP 1687 and the two other M. gardless of how Mexico was coded.
phaeocephalus, with which it was paraphyletic, is 4.75%.
Divergence between M. ferox–M. venezuelensis and the
latter M. phaeocephalus is 4.26%. 4. Discussion
M. nugator from the Lesser Antilles was nested within
M. tyrannulus in a clade with samples of that species 4.1. Monophyly and composition of Myiarchus
from Trinidad and Guara
unos, Venezuela. This clade in
turn was sister to non-Venezuelan South American The uniquely rufous- and black-plumaged M. semiru-
samples. All remaining samples of M. tyrannulus, which fus from arid coastal Peru and Ecuador was consis-
were from North America and Central America (Hon- tently sister to all other Myiarchus, or allied with the
duras, Mexico, US) formed a further clade that was outgroup R. immunda making Myiarchus paraphyletic.
sister to the Trinidad and South American samples Further, M. semirufus is as divergent from other My-
(Fig. 1). Maximum pairwise sequence divergence within iarchus (14%) as they are from outgroups (13–20%). We
it was 1.69%. Net divergence ranged from 0.60% be- provisionally retain it in Myiarchus because our taxon
tween Caribbean–Venezuela and non-Venezuelan South sampling cannot determine whether it should be placed
America to 1.17% between North America–Central in monotypic Muscifur Bangs and Penard, 1921, or
America and Caribbean–Venezuela. Net divergence elsewhere. ZimmerÕs (1938) recommendation of treating
among non-Venezuelan South America (Guyana and Muscifur as a subgenus would be an appropriate interim
northern Brazil vs Paraguay, Bolivia) was zero and no taxonomic measure. The relationships of M. semirufus
phylogenetic structure was recovered among these need to be clarified with respect to taxa such as the
samples. similarly rufous-plumaged species of Rhytipterna, Attila,
Jamaican M. stolidus stolidus and Cuban M. sagrae and Casiornis and the Myiarchus-like Deltarhynchus
sagrae formed a clade that was sister to another with flammulatus (Lanyon, 1982), which we have been unable
Dominican Republic M. stolidus dominicensis and Bah- to examine. Our data affirm with strong support the
aman M. sagrae lucaysiensis (Fig. 1). Divergences monophyly of all other Myiarchus taxa examined and
among M. stolidus andM. sagrae were <1% but both furthermore the inclusion of M. validus in Myiarchus
were at least 4% divergent from all other members of from our data is unremarkable. Furthermore, it is part
Clade I. of Clade I and consistently the sister to all remaining
 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152 145

Fig. 2. Reconstruction of distributional changes in Myiarchus on the Bayesian tree derived from the full dataset. Five arrows point to distributional
shifts found in all or most analyses and discussed in Section 3. Specimen codes are as in Appendix A.

taxa in that clade. We have been unable to examine M. ified huit notes in its dawn song (Lanyon, 1967). Few
nuttingii (Central America) and M. apicalis (Colombia) other characters show any such degree of congruence.
but predict on the basis of morphology and distribution Notable among them is extensive cinnamon ( ¼ Antique
that they will be members of Clades I and II, respec- Brown of Lanyon, 1978) being present in the inner vanes
tively (AOU, 1998; Lanyon, 1978). Similarly, we were of the rectrices in M. semirufus and Clade I but absent in
unable to examine the unusually small but again Clade II. In contrast, simple, unmodulated whistles in
otherwise unexceptionally plumaged M. magnirostris the dawn song and the roll, an element of the vocal re-
endemic to the Galapagos Islands. We predict from sponse to intruding conspecifics, show no correlation
Lanyon (1978) that its closest relative will be within the with membership of Clade I or II (see Lanyon, 1978, pp.
M. tyrannulus complex (Clade I). 443–448) whereas ‘‘hiccup’’ notes are present in Clade II
Congruence among characters of mtDNA, voice, and and variable in Clade I (e.g., absent in M. tyrannulus,
plumage is limited. The phylogenetic structure in present in M. antillarum). All members of Clade II lack
mtDNA of Clades I and II being sisters and M. what Lanyon (1978) termed ‘‘rapid huit’’ notes, ‘‘whay-
semirufus being their sister is paralleled in other char- burg’’ notes and and a dawn song with huits but no
acters only in plumage. Thus M. semirufus is distinc- rasping or plaintive whistles though each is present in
tively rufous and black in plumage and all other species M. tyrannulus (Clade I). Despite LanyonÕs (1967, 1978)
are essentially similar to each other in being dorsally and herculean work in documenting vocal characters, we are
pectorally grey, and yellow or creamy bellied. Although reluctant to attempt further mapping of vocal characters
M. semirufus is the only strictly South American species on to the phylogeny. Homologies in many of the vocal
having a dawn song with what Lanyon (1967, 1978) elements involved are not readily identifiable without a
termed ‘‘huit’’ notes and ‘‘rasping whistles,’’ Jamaican broader phylogeny of tyrannids. The problem is further
M. stolidus from Clade I, for example, does have mod- compounded by a formal nomenclature of the different
146 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152

elements of the Myiarchus vocal repertoire being more 4.2.2. M. tuberculifer

readily available for South American species (Lanyon, M. tuberculifer comprised two clades in most analy-
1978) than for Caribbean ones (Lanyon, 1967) though ses. While not closest relatives, their relationships to
we have attempted to circumvent this in some of the each other and to M. swainsoni (except its nominate
observations reported above. Finally here, we note that subspecies M. s. swainsoni) are unresolved. Caribbean
the basic plumage pattern shown by most Myiarchus M. barbirostris was most closely related to the pre-
species except M. semirufus (i.e., dorsal and pectoral dominantly Central and North American clade of M.
grey, yellow belly) is probably ancestral. It occurs in tuberculifer.
R. immunda and T. melancholicus as well as in Del- We plan a more complete discussion of the M. tu-
tarhynchus flammulatus, a species we have not examined berculifer group elsewhere but the paraphyly of M. t.
but which Lanyon (1967, 1985) considered not closely atriceps warrants mention here. The taxonomic status
related to Myiarchus, and still other tyrannids. of M. t. atriceps is among the most challenging prob-
lems in South American Myiarchus (Lanyon, 1978) and
4.2. Species-level systematics our findings reinforce this. LanyonÕs (1978) arguments
for treating M. t. atriceps as a subspecies of M. tuber-
Despite limited morphological diversity among My- culifer were complex but carefully reasoned interpreta-
iarchus species, the mtDNA phylogeny is often not tions of playback experiments and altitudinal effects on
concordant with current taxonomy. Paraphyly is wide- morphological and vocal characters. The central, novel
spread and suggests in some taxa a decoupling of issue arising in our study is the paraphyly of mtDNA
mtDNA and morphological evolution. To illustrate from opposite ends of M. t. atricepsÕs range, in Ecuador
concordances and discordances between mtDNA and and Argentina. Obviously, more complete sampling
morphology, we review cases of paraphyly and their will be needed before this can be explained but some
systematic implications. Also, because renewed concern possibilities emerge. MtDNA of M. t. atriceps at the
has been expressed lately (e.g., Ballard et al., 2002; northern and southern ends of its range could be int-
Weckstein et al., 2001) about making taxonomic chan- rogressed from adjoining populations of other subspe-
ges based solely on mtDNA, we offer our findings as a cies. This would cause paraphyly in M. t. atriceps
focus for further questions that should be addressed because those adjoining populationsÕ mtDNAs are
with characters from nuclear DNA, morphology and themselves paraphyletic (Fig. 1). Alternative explana-
voice. tions include incomplete sorting of mtDNA, and inac-
curate taxonomy. We will examine these possibilities
4.2.1. M. swainsoni more fully elsewhere.
We have discussed this complex in detail elsewhere
(Joseph et al., 2003). Here it is worth noting LanyonÕs 4.2.3. M. ferox
(1978, pp. 472–474) finding that playback of M. s. Paraphyly of M. ferox with respect to M. venezuel-
ferocior elicited a behavioural response in Argentinean ensis and the divergence between them (0.24%), which
M. tuberculifer atriceps and that our mtDNA data is the same order of magnitude as among M. ferox
have identified a similar, unexpected relationship be- alone (0.55%), suggest either incomplete sorting or that
tween M. tuberculifer and all subspecies examined of they are conspecific or that the M. venezuelensis spec-
M. swainsoni except its nominate subspecies M. s. imen is descended from hybridization between female
swainsoni. Playback results may have indicated a re- M. ferox and male M. venezuelensis at an unknown
sponse to plesiomorphic vocal characters (Ryan and time in the past. The locality of the M. venezuelensis
Rand, 1995). Pending acquisition of nuclear DNA sample (Guara
unos, Sucre, Venezuela: 10°330 N,
0
(nDNA) data, we note that taxonomic changes will 63°07 W) is from within a zone of overlap between that
almost certainly be needed in the M. swainsoni com- species and M. ferox (Lanyon, 1978) so hybridization
plex. Even if M. s. swainsoni sensu Lanyon (1978) is is feasible. (The M. venezuelensis sample was available
separated as M. swainsoni Cabanis and Heine, 1859, only as a blood sample and so we have been unable to
however, as seems reasonable based on its mtDNA check its external phenotype.) A close relationship
and on morphological differences that presumably re- among these taxa, however, is expected (see Lanyon,
flect nDNA, taxonomic changes to the remaining 1978).
members of the group depend on: (1) examination of
the single M. swainsoni subspecies we were unable to 4.2.4. M. phaeocephalus
study, M. s. phaeonotus, (2) resolution of why two Hybridization between M. phaeocephalus and M. fe-
Guyanan individuals are so divergent for mtDNA, and rox might explain the paraphyly of M. phaeocephalus.
(3) determination of priority among the epithets fe- The sample causing it (ANSP 1687) is of the subspecies
rocior, pelzelni, and phaeonotus, all of which were de- M. p. interior, which is locally sympatric with M.
scribed in 1883. ferox (Lanyon, 1978, Figs. 1 and 3). We consider
 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152 147

hybridization unlikely, however, because ANSP 1687 is the mtDNA data suggesting paraphyly due to hybrid-
consistently basal to M. ferox rather than nested within ization and/or incomplete sorting.
it as would be expected from hybridization cf M. vene-
zuelensis unless, perhaps, the hybridization event was an 4.2.5. M. tyrannulus
ancient one relative to the divergence of these species Separation of Central and North American mtDNAs
from their common ancestor. Furthermore, the associ- within the M. tyrannulus complex supports their recog-
ated voucher shows no phenotypic tendency towards M. nition either as M. Cooperi Baird, 1858 or two species
ferox (Fig. 3). Alternatively, mtDNAs of M. phaeo- M. magister Ridgway, 1884 and M. cooperi (see AOU,
cephalus and M. ferox may be incompletely sorted. The 1998; Cardiff and Dittmann, 2000). We cannot address
close relationship of M. panamensis to M. ferox and M. whether Pacific slope populations of Central America
venezuelensis was not unexpected (Lanyon, 1978). should be further separated as M. brachyurus Ridgway,
Arguably, the mtDNAs of M. phaeocephalus, M. pan- 1887 (e.g., AOU, 1998) because our Central American
amensis, M. ferox, and M. venezuelensis are all incom- samples are from the Atlantic coast. Our two samples of
pletely sorted. Field study of ecological interactions and M. t. insularum from the island of Utila, Honduras, were
niche separation among these four taxa where they are monophyletic. On the basis of voice playback experi-
parapatric and sympatric would be rewarding in light of ments and minor differences in colour of mouth-lining

Fig. 3. Map of Ecuador and adjacent parts of Peru showing distributions of M. ferox, M. phaeocephalus, and M. tuberculifer in that region. The
mtDNA of sample M. phaeocephalus interior 01.EC ( ¼ skin ANSP 181729 ¼ tissue ANSP 1687) is more closely related to that of M. ferox than to
other M. phaeocephalus (Fig. 1). Note, however, that the same bird is phenotypically typical of M. phaeocephalus, e.g., its bill is narrow as in M.
phaeocephalus, not broad as in M. ferox, despite its proximity to M. ferox localities. Localities are derived from Lanyon (1978) except where in-
dicated. Specimen codes are as in Appendix A.
148 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152

and wing feathers, Lanyon (1967) argued that M. nu- two major radiations within Myiarchus. One is pri-
gator of the southern Lesser Antilles is specifically dis- marily of West Indian and Central and North American
tinct from M. tyrannulus. Divergence in these characters species (Clade I), and the other is primarily of South
has likely been extremely rapid and recent given M. American species (Clade II). The former comprises all
nugatorÕs 0.03% divergence from its closest relatives in West Indian endemics as well as M. tyrannulus, which is
M. tyrannulus from nearby Trinidad and Venezuela. widely distributed across North, Central, and South
Relevant to the status of the Caribbean endemic M. America. Similarly, the South American radiation
nugator is that our Venezuelan sample of M. tyrannulus comprises all other species occurring there that we have
was unexpectedly more closely related to M. nugator examined except M. semirufus, which is not closely re-
and Trinidad M. tyrannulus than to M. tyrannulus lated to other Myiarchus. We do find support for
samples from Guyana and the rest of South America LanyonÕs suggestion of M. barbirostris having evolved
and that biogeographic reconstruction suggested this through relatively recent colonization from a M. tu-
clade had dispersed into northern South America. The berculifer-like ancestor. This represents dispersal of
possibility arises that M. nugator Riley, 1904 and Ven- Clade II into the Caribbean from South or Central
ezuelan and Trinidad populations should all be recog- America. However, in the history of Clade I and with
nized as M. erythrocercus Sclater and Salvin, 1868 (type respect to M. stolidus and M. validus, we find support
locality Caracas, Venezuela). M. tyrannulus (Muller, for a sequence of evolution that is the reverse of what
1776) (type locality Cayenne, French Guiana) would Lanyon (1967) proposed, i.e., that M. validus is the
then apply to remaining South American populations. older species. In the M. tyrannulus complex, a more
The respective ranges in northern South America of M. complex history is suggested. Colonization of the Ca-
erythrocercus and M. tyrannulus as so construed need to ribbean has clearly been involved in the evolution of M.
be clarified. The mtDNA data again imply that pheno- nugator. From our data, we cannot rule out the possi-
typic divergence in the M. tyrannulus complex has been bility that M. tyrannulus has not itself back colonized
rapid. far northern mainland South America from the Carib-
bean. It appears highly likely that the occurrence of M.
4.2.6. M. stolidus superspecies tyrannulus elsewhere in South America other than far
Lanyon (1967) recognized M. antillarum, M. oberi, northern Venezuela is the result of an earlier, indepen-
M. sagrae, and M. stolidus as the M. stolidus superspe- dent colonization. Thus Clade I has secondarily entered
cies to emphasize close inferred relationship. In the South America at least once.
present study, the four formed a single monophyletic Rejection of a molecular clock in our dataset pre-
clade only in the Bayesian and NJ analyses of all data. cludes most attempts to estimate the timing of clado-
Individually, M. sagrae and M. stolidus were not genic events in the history of Myiarchus. Accordingly,
monophyletic. Bahaman M. sagrae lucaysiensis aligns we only make the following statements. If non-control
with Dominican Republic M. stolidus dominicensis region mtDNA is evolving at roughly 2% per million
(Fig. 1) and in this mtDNA and morphology are con- years (e.g., Fleischer et al., 1998; Krajewski and King,
cordant to some extent. They share smaller size and 1996; see also Avise and Walker, 1998), then sequence
paler yellow bellies relative to M. stolidus stolidus divergences suggest that speciation within Clades I and
(Lanyon, 1967). Cuban M. sagrae sagrae aligns with II began before the Pleistocene (Table 1). M. validus is
Jamaican M. stolidus stolidus. In this alignment, likely the oldest member of Clade I, and M. s. swainsoni
mtDNA and morphology are discordant. M. stolidus and M. cephalotes likely are among the oldest of Clade
stolidus is larger and more brightly yellow-bellied than II. Maximum pairwise divergences in Clade I (6%–10%)
M. sagrae sagrae (Lanyon, 1967). Pallor of the belly suggest a longer history of diversification in that Clade
increases in the sequence M. stolidus dominicensis–M. (Table 1). Identification of factors causing speciation
sagrae lucaysiensis–M. antillarum, the last-named having within Clades I and II and the history of speciation in
an almost white belly. Rather than interpret the para- both clades are topics beyond the scope of the present
phyly of M. stolidus and M. sagrae sensu Lanyon (1967) paper.
as being solely due to incomplete sorting and obvious
decoupling of the evolution of mtDNA and morphol-
ogy, we suspect that the taxonomy requires re-exami- 5. Conclusions
nation by way of a systematic reassessment of the entire
M. stolidus ‘‘superspecies.’’ The monophyly of Myiarchus has been robustly
supported with the only uncertainty being the place-
4.3. Biogeography ment of the uniquely rufous and black M. semirufus
among the groupÕs close relatives. Evolution within a
A South American origin of Myiarchus is plausible primarily South American clade and a primarily Ca-
given the diversity of Tyrannidae there. Our data reveal ribbean/North American one has occurred over broadly
 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152 149

similar periods of time, as judged by similar levels of Natural History, Lawrence (A. Peterson, M. Robbins,
divergence within the two clades. Some species of the N. Rice); Laboratory of Molecular Systematics and
primarily South American clade have dispersed out of Unites States National Museum, Smithsonian Institu-
that continent whereas M. tyrannulus of a primarily tion, Washington, DC (M. Braun, M. Schmidt, C.
Caribbean and North American clade has secondarily Milensky, J. Dean); Louisiana State University, Baton
entered it. Closer resolution of relationships and rates of Rouge (J. Babin, S. Cardiff, D. Dittmann, F. Sheldon,
evolution within both clades of Myiarchus will require J.V. Remsen); Marjorie Barrick Museum of Natural
study of nuclear DNA and loci with a range of evolu- History, Las Vegas (G. Voelker); Jose Maria Cardoso
tionary rates. da Silva and his students answered queries concerning
some specimens. G. Voelker (MBM, Las Vegas) assisted
with DIVA. Additional support and helpful discussions
Acknowledgments and comments came from L. Belasco, D. Cohen, J.
Huelsenbeck, T.D. Pedersen, J. Reich, N. Rice, T.
The pioneering studies of Wesley Lanyon paved the Schulenberg, C. Spolsky, T. Uzzell, and two anonymous
way for this work and we record our debt to his dedi- reviewers.
cation and scholarship. Work at ANSP for this project
was funded by the Rodolphe Meyer de Schauensee Appendix A
Fund (ANSP), National Science Foundation Grant No.
9871363 to the Molecular Systematics and Evolution Details of specimens examined with latitudes and
group (ANSP), and the generous support of D. Cohen. longitudes where available from primary specimen data.
L.J. is indebted to the late Shane Parker for accidentally An asterisk (*) indicates individuals used in the n ¼ 31
kindling his interest in Myiarchus. L.J.Õs contribution to dataset. Full details of the 49 specimens examined from
the work began in Uruguay with the help, encourage- the M. swainsoni complex are in Joseph et al., 2003); for
ment and field company of Dr. Enrique Lessa and brevity, they are here accompanied only with taxon
Santiago Claramunt, Maria Noel Cortinas, Guillermo codes corresponding to those used in Figs. 1 and 2 of
DÕElia, Federico Hoffman, Tito Oliveira, Ana Luz this paper, their GenBank accession codes, and their
Porzecanski, and Ivanna Tomasco, all of the Universi- museum accession numbers. Coding used for biogeo-
dad de Republica, Montevideo. We are grateful to all graphic analyses shown in Fig. 2 are: SA, South
permit-granting authorities and all collectors involved America; NA, North America; CA, Central America;
for their work. We are grateful to the following insti- Car, Caribbean; MX, Mexico. See Section 2 for details
tutions and their curatorial and collection management of how Mexico was incorporated into various coding
staff for making samples readily available to us and, in schemes. Museum and collection acronyms: AMNH,
some cases, for collecting samples specifically at our American Museum of Natural History; ANSP, Acad-
request: Academy of Natural Sciences, Philadelphia (D. emy of Natural Sciences; FMNH, Field Museum of
Agro, R. Ridgely, M. Robbins; F. Sornoza); American Natural History; KU, University of Kansas; LSUMZ,
Museum of Natural History, New York (C. Blake, A. Louisiana State University Museum of Zoology; MBM,
L. Porzecanski, J. Weicker, P. Sweet, G. Barrowc- Marjorie Barrick Museum of Natural History; STRI,
lough); Field Museum of Natural History, Chicago (J. Smithsonian Tropical Research Institute; USNM, Uni-
Bates, D. Willard); University of Kansas Museum of ted States National Museum.

Appendix A

Taxon and code Locality Distribution GenBank Museum code

in Fig. 1 code
M. antillarum 01.PR* Puerto Rico: Guanica Car AY115182 STRI PR MAN 2
M. antillarum 02.PR Puerto Rico: Guanica Car AY115181 STRI PR MAN 1
M. barbirostris 01.JM Jamaica: Luana Point Car AF497962 STRI JA-MBA 1
M. barbirostris 02.JM* Jamaica: Paradise Park Car AY266207 STRI JA-MBA 5
M. cephalotes 01.EC* Ecuador: Depto Napo, 12 km NNE El SA AF497966 ANSP 4807
Chaco, Mirador
M. cephalotes 02.EC Ecuador: Napo, 3 km N Cosanga SA AY266183 ANSP 5124
M. cinerascens 01.MX* Mexico: Queretaro, Pena Bernal Mex AY266222 FMNH 394371
M. crinitus 01.US* Honduras: Cochino Pequeno CA AY266197 STRI HA MCR HA168
M. ferox 01.GY* Guyana: Surama, 04°100 N, 59°050 W SA AY266184 ANSP 8369
M. ferox 02.EC Ecuador: Sucumbios, SA AY266185 ANSP 5498
ca 14 km N Tigre Playa, 00°200 N, 76°400 W
150 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152

Appendix A (continued)
Taxon and code Locality Distribution GenBank Museum code
in Fig. 1 code
M. ferox 03.GY Guyana: Iwokrama, SA AY266186 KU B 1233
ca. 41 road km SW Kurupukari,
04°200 N, 58°510 W
M. ferox 04.GY Guyana: GunnÕs Landing, 01°390 N, 58°570 W SA AY266187 USNM B 10968
M. ferox 05.GY Guyana: Berbice River, SA AY266788 USNM B 4377
Dubulay Ranch, 05°400 N, 57°530 W
M. ferox 06.PE Peru: Dpto Ucayali, SA AY266189 LSUMZ B 10730
65 km ENE Pucallpa
M. ferox 07.BR Brazil: Rondonia, Cachoeira Nazare, SA AY266191 FMNH 389972
W bank Rio Jiparan
a,
M. ferox 08.BR Brazil: Para, Monte Alegre, Colonia do Erere, SA AY266192 FMNH 392547
M. ferox 09.BR Brazil: Para, Monte Alegre, Colonia do Erere, SA AF497965 FMNH 392548
M. ferox 10.BO Bolivia: Laguna Suarez, 14°480 S, 64°460 W SA AY266193 FMNH 394469
M. ferox 11.GY Guyana: 01°390 N, 58°370 W SA AY266194 USNM B 10972
M. ferox 12.GY Guyana: near Iwokrama SA AY266195 ANSP 8162
Reserve, 04°100 N, 59°050 W
M. ferox 13.GY Guyana: 01°390 N, 58°370 W SA AY266196 USNM B 10971
M. ferox 14.GY Guyana: GunnÕs Landing: 01°390 N, 58°570 W SA AY266190 USNM B 11551
M. nugator 01.VC* St. Vincent: Cumberland Valley Car AY266198 STRI SV MOB 1
M. oberi 01.AG* Antigua and Barbuda: 5 km E Cadrington Car AY115173 STRI BU MOB 1
M. panamensis 01.CR* Costa Rica: Puntarenas, 1 km NW Tarcoles CA AF497959 LSUMZ B9935
M. phaeocephalus 01.EC* Ecuador: Zaruma, 04°530 S, 79°080 W SA AF497963 ANSP 1687
M. phaeocephalus 02.EC* Ecuador: 3 km N Zapotillo, SA AF497964 ANSP 4147
ca. 04°020 S, 79°010 W
M. phaeocephalus 03.EC Ecuador: Loja, 16 road km SW Sabanilla CA AY266171 ANSP 1745
M. sagrae 01.BS* Bahamas: Airport Car AY266203 STRI BH MAN 1
M. sagrae 02.CU* Cuba: Pinar del Rio Car AY266204 STRI CU MAN 5546
M. sagrae 03.CU Cuba: Pinar del Rio Car AY266205 STRI CU MAN 5549
M. semirufus 01.PE* Peru: Lambayeque: Las Pampas, km 885 Pan SA AY266170 LSUMZ B 5203
American Hwy, 11 road km from Olmos
M. stolidus 01.JM* Jamaica: Luana Point Car AY266199 STRI JAMYS 1
M. stolidus 02.JM Jamaica: Portland Ridge Car AY266200 STRI JAMYS 2
M. stolidus 03.DO* Dominican Republic: Club Dominicus Car AY266201 STRI RDMSYS 1
M. stolidus 04.DO Dominican Republic: Car AY266202 STRI RDMSYS 7
8 December 1995
M. tuberculifer Mexico: Campeche, 24 km S Silvituc, MX AY266172 KU B 1932
platyrhynchus 01.MX 18°140 N, 90°120 W
M. tuberculifer Guyana: Sipu River, 01°250 N, 58°570 W SA AY266173 USNM B 10481
tuberculifer 01.GY
M. tuberculifer Argentina: Tucum
an SA AF497961 USNM B 5785
atriceps 01.AR*
M. tuberculifer Panama: Bocas del Toro CA AY266174 USNM B 458
nigricapillus 01.PA*
M. tuberculifer Panama: Bocas del Toro, Cayo Agua, CA AY266175 USNM B 1157
nigricapillus 02.PA near Punta Limon
M. tuberculifer Panama: Bocas del Toro, Rio Changuinola CA AY266176 USNM B 1917
nigricapillus 03.PA
M. tuberculifer Ecuador: 6 km NW San Andres, E slope Cord. SA AF497960 ANSP 5043
cf. atriceps (see text) 01.EC Lagunillas
M. tuberculifer Panama: Chiriqui, Fortuna Dam, CA AY266177 USNM B 5343
nigricapillus 04.PA 08°450 N, 82°150 W
M. tuberculifer Mexico: Yucatan, MX AY266178 USNM B 1860
platyrhynchus 02.MX 18 km E Dzilam de Bravo, 21°280 N, 88°340 W
M. tuberculifer Ecuador: Imuya Cocha, SA AY266179 ANSP 3241
tuberculifer 02.EC Sucumbios, 00°340 S, 75°170 W
M. tyrannulus tyrannulus 01.TT* Trindidad and Tobago: Car AY115170 STRI CC-MTY1
Chacachacare Island
M. tyrannulus tyrannulus 02.TT Trindidad and Tobago: Car AY115169 STRI CC-MTY2
Chacachacare Island
M. tyrannulus insularum 01.HN* Honduras: Utila, Bay Islands CA AY266217 STRI HA-MTY-HA214
M. tyrannulus insularum 02.HN Honduras: Utila, Bay Islands CA AY266218 STRI HA-MTY-HA215
M. tyrannulus cooperi 01.US USA: Louisiana: Cameron Parish NA AY266219 STRI US-MTY4069
M. tyrannulus cooperi 02.US USA: Louisiana: Cameron Parish NA AY266220 STRI US-MTY5678
 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152 151

Appendix A (continued)
Taxon and code Locality Distribution GenBank Museum code
in Fig. 1 code
M. tyrannulus magister 03.US* USA: Arizona: Cochise County, NA AY266221 LSUMZ B 4075
Silver Creek, near Paradise
M. tyrannulus tyrannulus 03.VE Venezuela: Guarapo SA AY115171 STRI VE MTY 1
M. tyrannulus tyrannulus 04.GY* Guyana: Mahaica-Berbice, Coast, near SA AF497958 ANSP 8399
Onverwagt, 06°250 N, 57°370 W
M. tyrannulus tyrannulus 05.PY Paraguay: San Luis N.P., 22°480 S, 57°210 W SA AY266214 KU B 176
M. tyrannulus cooperi 04.MX Mexico: Campeche, 24 km S Silvituc, MX AY266215 KU B 2094
18°140 N, 90°120 W
M. tyrannulus cooperi 05.MX Mexico: Campeche, 24 km S Silvituc, MX AY266216 KU B 2112
18°140 N, 90°120 W
M. tyrannulus bahiae 01.BR Brazil: Amap
a, Fazenda Casemiro SA AY266208 FMNH 391520
M. tyrannulus tyrannulus 06.BO Bolivia: Beni, 5 km SW Laguna Su
arez, SA AY266209 FMNH 396030
04°500 S, 64°510 W
M. tyrannulus tyrannulus 07.BO Bolivia: Chiquitos, 69 Km. N of San Jos
e de SA AY266210 FMNH 396038
Chiquitos, 17°400 S, 59°420 W
M. tyrannulus tyrannulus 08.BO Bolivia: Beni, 5 km SW Laguna Su
arez, SA AY266211 FMNH 396037
04°500 S, 64°510 W
M. tyrannulus tyrannulus 09.BO Bolivia: Beni, 5 km SW Laguna Su
arez, SA AY266212 FMNH 396051
04°500 S, 64°510 W
M. tyrannulus tyrannulus 10.BO Bolivia: near El Tun
a, 16°330 S, 59°390 W SA AY266213 AMNH 25665
M. validus 01.JM* Jamaica: Portland Ridge Car AY266206 STRI JA MVL 1
M. venezuelensis 01.VE* Venezuela: Guaraunos, 110°330 N, 63°070 W SA AY266182 STRI VE MVE 1
M. yucatanensis 01.MX* Mexico: Campeche, 24 km S Silvituc, MX AY266180 KU B 1950
18°140 N, 90°120 W
M. yucatanensis 02.MX Mexico: Yucatan, 18 km E Dzilam de Bravo, MX AY266181 KU B 1880
21°280 N, 88°340 W
Rhytipterna immunda 01.GY* Guyana: Rupununi, 02°120 N, 59°220 W SA AF497967 USNM B 12786
Tyrannus caudifasciatus 01.BS* Bahamas: Abaco, 19 October 1993 Car AF497968 STRI AB-TCF 1
Tyrannus melancholicus 01.GY Guyana: Iwokrama, 04°450 N, 59°010 W SA AY266223 ANSP 8341
Tyrannus melancholicus 02.UY Uruguay: 15 km N Mercedes, 33°070 S, 58°010 W SA AY266224 ANSP 10316
Tyrannus melancholicus 03.UY Uruguay: 15 km N Mercedes, 33°070 S, 58°010 W SA AY266225 ANSP 10317
Tyrannus melancholicus 04.TT* Trinidad and Tobago: Caroni Swamp Car AY266226 STRI TRTML 4
M. swainsoni 01 AR*, AF497948, MBM 5343; 02 AR, AF497949, MBM 5500; 03 AR, AF497950, MBM 5501; 04 PE,
(includes M. s. ferocior, AF497927, LSUMZ B 10644; 05 BO, AF497928, LSUMZ B 9563; 06 BO, AF497930, AMNH 833439; 07
M. s. pelzelni and BO, AF497931, AMNH 833291; 08 BO, AF497937, AMNH 833292; 09 BO, AF497932, AMNH 25670; 10
M. s. swainsoni
M. s. BO, AF497939, AMNH 833440; 11 BO, AF497933, AMNH 25671; 12 BO, AF497938, AMNH 833437; 13
pelzelni intergrades, BO, AF497934, AMNH 833438; 14 BO, AF497935, AMNH 25668; 15 BO, AF497936, AMNH 25669; 16
M. s. ferocior
M. s. PY, AF497922, KU B 4397; 17 PY, AF497946, KU B 137; 18 PY, AF497947, KU B 102; 19 UY, AF497951,
pelzelni intergrades, ANSP 10921; 20 UY, AF497952, ANSP 10331; 21 GY, AF497921, KU B 4336; 22 GY, AF497909, USNM B
and all but two M. s. 10975; 23 GY, AF497910, USNM B 10986; 24 GY, AF497911, USNM B 10993; 25 GY, AF497912, USNM
phaeonotus
M. s. B 10980; 26 GY, AF497913, USNM B 11108; 27 GY, AF497914, USNM B 10963; 28 GY, AF497915,
pelzelni intergrades). USNM B 10312; 29 BO, AF497940, FMNH 334508; 30 BO, AF497941, FMNH 394470; 31 BO, AF497929,
See Joseph et al. (2003) LSU B 9458; 32 PY, AF497942, KU B 3415; 33 GY, AF497916, USNM B 11387; 34 GY, AF497917, USNM
for further details. B 11554; 35 GY, AF497918, USNM B 11908; 36 GY, AF497919, USNM B 11964; 37 GY, AF497920,
USNM B 12220; 38 PY, AF497945, KU B 2984; 39 PY, AF497944, KU B 2986; 40 PY, AF497943, KU B
3101; 41 BR, AF497926, FMNH 391521 ¼ MPEG 53569; 42 BR, AF497923, FMNH 391516; 43 BR,
AF497924, FMNH 391517; 44 BR, AF497925, FMNH 391518
M. s. phaeonotus
M. s. pelzelni 01.GY, AF497954, USNM B 11581; 02.GY*, AF497953, USNM B 10990
intergrades
M. swainsoni swainsoni 01.UY*, AF497955, ANSP 10250; 02.UY, AF497956, ANSP 10269; 03.AR, AF497597, USNM B 5967

References Avise, J.C., Walker, D., 1998. Pleistocene phylogeographic effects on

avian populations and the speciation process. Proc. R. Soc. Lond.
B 265, 457–463.
Alfaro, M.E., Zoller, S., Lutzoni, F., 2003. Bayes or bootstrap? A
Ballard, J.W.O., Chernoff, B., James, A.C., 2002. Divergence of
simulation study comparing the performance of Bayesian Markov
mitochondrial DNA is not corroborated by nuclear DNA,
chain Monte Carlo sampling and bootstrapping in assessing
morphology, or behavior in Drosophila simulans. Evolution 56,
phylogenetic confidence. Mol. Biol. Evol. 20, 255–266.
527–545.
AOU [American Ornithologists Union], 1998. Check-list of North
Cardiff, S.W., Dittmann, D.L., 2000. Brown-crested Flycatcher (My-
American Birds, seventh ed. American Ornithologists Union,
iarchus tyrannulus). In: Poole, A., Gill, F. (Eds.), The Birds of
Washington, DC.
152 L. Joseph et al. / Molecular Phylogenetics and Evolution 31 (2004) 139–152

North America, No. 496. The Birds of North America, Inc, Lanyon, W.E., 1985. A phylogeny of the myiarchine flycatchers. Orn.
Philadelphia, PA. Monogr. 36, 360–380.
Chesser, R.T., 2000. Evolution in the high Andes: the phylogenetics of Lanyon, W.E., 1986. A phylogeny of the thirty-three genera in the
Muscisaxicola ground-tyrants. Mol. Phylogenet. Evol. 15, 369–380. Empidonax assemblage of tyrant flycatchers. Am. Mus. Novitat.
Coues, E., 1872. Studies of the Tyrannidae—Part I. Revision of the 2846, 1–29.
species of Myiarchus. Proc. Acad. Nat. Sci. Phil. 24, 56–81. Lovette, I.J., Bermingham, E., 1999. Explosive speciation in the
Douady, C.J., Delsuc, F., Boucher, Y., Doolittle, W.F., Douzery, New World Dendroica warblers. Proc. R. Soc. Lond. B 266, 1629–
E.J.P., 2003. Comparison of Bayesian and maximum likelihood 1636.
bootstrap measures of phylogenetic reliability. Mol. Biol. Evol. 20, Maddison, D.R., Maddison, W.P., 2002. MacClade 4.0. Sinauer,
248–254. Sunderland.
Felsenstein, J., 1981. Evolutionary trees from DNA sequences: a McKitrick, M.C., 1985. Monophyly of the Tyrannidae (Aves:
maximum likelihood approach. J. Mol. Evol. 17, 368–376. comparison of morphology and DNA). Syst. Zool. 34, 35–45.
Fitzpatrick, J.W., 1985. Form, foraging behavior, and adaptive Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model
radiation in the Tyrannidae. Orn. Monogr. 36, 447–470. of DNA substitution. Bioinformatics 14, 817–818.
Fleischer, R., McIntosh, C., Tarr, C., 1998. Evolution on a volcanic Ridgely, R.S., Tudor, G., 1994. The Birds of South America: Oscines.
conveyor belt, using phylogeographic reconstructions and K–Ar Oxford University Press, Oxford.
based ages of the Hawaiian Islands to estimate molecular evolu- Ronquist, F., 1996. DIVA, ver. 1.1. Computer program available
tionary rates. Mol. Ecol. 7, 533–545. by anonymous ftp via <http://www.ebc.uu.se/systzoo/staff/ronquist.
H€aarlid, A., Arnason, U., 1999. Analyses of mitochondrial DNA nest html>. Uppsala University, Uppsala.
ratite birds within the Neognathae: supporting a neotenous origin of Ronquist, F., 1997. Dispersal-vicariance analysis: a new approach to the
ratite morphological characters. Proc. R. Soc. Lond. B 266, 305–309. quantification of historical biogeography. Syst. Biol. 46, 195–203.
Hillis, D.M., Huelsenbeck, J.P., 1992. Signal, noise, and reliability in Roy, M.S., Torres-Mura, J.C., Hertel, F., 1999. Molecular phylogeny
molecular phylogenetic analyses. J. Hered. 83, 189–195. and evolutionary history of the tit-tyrants (Aves: Tyrannidae).
Huelsenbeck, J.P., Rannala, B., 1997. Phylogenetic methods come of Mol. Phylogenet. Evol. 11, 67–76.
age: testing hypotheses in an evolutionary context. Science 276, Ryan, M.J., Rand, A.S., 1995. Female responses to ancestral adver-
227–232. tisement calls in T

ungara frogs. Science 269, 390–392.
Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference Sibley, C.G., Ahlquist, J.E., 1985. Phylogeny and classification of New
of phylogeny. Bioinformatics 17, 754–755. World suboscine passerine birds (Passeriformes: Oligomyodi:
Hunt, J.S, Bermingham, E., Ricklefs, R., 2001. Molecular systematics Tyrannides). Orn. Monogr. 36, 396–428.
and biogeography of Antillean thrashes, termblers, and mocking- Smith, J.W., 1971. Behavior of Muscisaxicola and related genera. Bull.
birds (Aves: Mimidae). Auk 118, 35–55. Mus. Comp. Zool. 141, 23–268.
Joseph, L., Wilke, T., Alpers, D., 2003. Independent evolution of Swofford, D., 2002. PAUP 4.0b10. Phylogenetic Analysis Using
migration on the South American landscape in a long-distance Parsimony. Sinauer, Sunderland.
temperate-tropical migratory bird, SwainsonÕs Flycatcher Myiar- Tamura, K., Nei, M., 1993. Estimation of the number of nucleotide
chus swainsoni. J. Biogeogr. 30, 925–937. substitutions in the control region of mitochondrial DNA in
Krajewski, C., King, D.G., 1996. Molecular divergence and phylog- humans and chimpanzees. Mol. Biol. Evol. 10, 512–526.
eny: rates and patterns of cytochrome b evolution in cranes. Mol. Traylor, M.E., 1977. A classification of the tyrant flycatchers (Tyran-
Biol. Evol. 13, 21–30. nidae). Bull. Mus. Comp. Zool. 148, 129–184.
Lanyon, W.E., 1967. Revision and probable evolution of the Myiar- Weckstein, J.D., Zink, R.M., Blackwell-Rago, R.C., Nelson, D.A.,
chus flycatchers of the West Indies. Bull. Am. Mus. Nat. Hist. 136, 2001. Anomalous variation in mitochondrial genomes of white-
329–370. crowned (Zonotrichia leucophrys) and golden-crowned (Z. atrica-
Lanyon, W.E., 1975. Behavior and generic status of the Rufous pilla) sparrows: pseudogenes, hybridization, or incomplete lineage
Flycatcher of Peru. Wilson Bull. 87, 441–608. sorting? Auk 118, 231–236.
Lanyon, W.E., 1978. Revision of the Myiarchus flycatchers of South Vuilleumier, F., 1971. Generic relationships and speciation patterns in
America. Bull. Am. Mus. Nat. Hist. 161, 429–627. Ochthoeca, Myiotheretes, Xolmis, Neoxolmis, Agriornis, and Mu-
Lanyon, W.E., 1982. Behavior, morphology, and systematics of the scisaxicola. Bull. Mus. Comp. Zool. 141, 181–232.
flammulated flycatcher of Mexico. Auk 99, 414–423. Zimmer, J.T., 1938. Studies of Peruvian birds. No 29. The genera
Lanyon, W.E., 1984. A phylogeny of the kingbirds and their allies. Myiarchus, Mitrephanes, and Cnemotriccus. Am. Mus. Novitat.
Am. Mus. Novitat. 2797, 1–28. 994.