Molecular Phylogenetics and Evolution: Sciverse Sciencedirect
Molecular Phylogenetics and Evolution: Sciverse Sciencedirect
a r t i c l e i n f o a b s t r a c t
Article history: The mockingbirds, thrashers and allied birds in the family Mimidae are broadly distributed across the
Received 1 January 2011 Americas. Many aspects of their phylogenetic history are well established, but there has been no previous
Revised 3 June 2011 phylogenetic study that included all species in this radiation. Our reconstructions based on mitochondrial
Accepted 11 July 2011
and nuclear DNA sequence markers show that an early bifurcation separated the Mimidae into two clades,
Available online 17 August 2011
the first of which includes North and Middle American taxa (Melanotis, Melanoptila, Dumetella) plus a small
radiation that likely occurred largely within the West Indies (Ramphocinclus, Allenia, Margarops, Cinclocer-
Keywords:
thia). The second and larger radiation includes the Toxostoma thrasher clade, along with the monotypic
Mimidae
Mockingbird
Sage Thrasher (Oreoscoptes) and the phenotypically diverse and broadly distributed Mimus mockingbirds.
Thrasher This mockingbird group is biogeographically notable for including several lineages that colonized and
Galapagos diverged on isolated islands, including the Socorro Mockingbird (Mimus graysoni, formerly Mimodes)
Phylogeny and the diverse and historically important Galapagos mockingbirds (formerly Nesomimus). Our recon-
Classification structions support a sister relationship between the Galapagos mockingbird lineage and the Bahama
Mockingbird (M. gundlachi) of the West Indies, rather than the Long-tailed Mockingbird (M. longicaudatus)
or other species presently found on the South American mainland. Relationships within the genus Toxos-
toma conflict with traditional arrangements but support a tree based on a preivous mtDNA study. For
instance, the southern Mexican endemic Ocellated Thrasher (T. ocellatum) is not an isolated sister species
of the Curve-billed thrasher (T. curvirostre).
Ó 2011 Published by Elsevier Inc.
1. Introduction et al., 1997; Derrickson, 1988; Botero et al., 2009a); the group as a
whole shows extensive variation in the extent and use of acoustic
The mockingbirds, thrashers and allied birds that comprise the signals, and this behavioral diversity has informed comparative
avian family Mimidae are well-known models in evolutionary biol- studies of how complex song traits evolve (Botero et al., 2009b).
ogy and behavioral ecology. The Galapagos mockingbirds inspired The group is likewise notable for its high interspecific variation in
Charles Darwin when he noted the morphological differences social and mating systems (Curry, 1989; Curry and Grant, 1990),
among forms from several islands and their differentiation from re- as even within the mockingbirds breeding groups vary across a
lated species that occur on the South American mainland (Darwin, sociality spectrum from socially monogamous pairs, to parents
1845, 1859). Other species of mockingbirds and thrashers have with helpers who are previous offspring, to plural breeding groups
some of the most extensive song repertoires known in all birds that contain multiple reproducing pairs (Curry, 1988).
(e.g., Kroodsma and Parker, 1977; Derrickson, 1987; Read and No previous phylogenetic survey has included a full sample of
Weary, 1992), and many functional and behavioral aspects of their mimid species, and this lack of a completely inclusive phylogenetic
vocal and learning abilities have been characterized (e.g., Kroodsma hypothesis has hindered some comparative analyses of trait varia-
tion in this group. Nonetheless, many aspects of the phylogenetic
⇑ Corresponding author. Fax: +1 607 254 2486. history of the Mimidae radiation are well established. The mimid
E-mail address: [email protected] (I.J. Lovette). clade is sister to the more diverse Old World starlings (Sturnidae;
Beecher, 1953; Stallcup, 1961; Sibley and Ahlquist, 1984, 1990; shown that the historical population structure within the Nesomi-
Malcarney et al., 1994; Voelker and Spellman, 2004; Ericson and mus group does not fully track the traditional species boundaries
Johansson, 2003; Cibois and Cracraft, 2004; Barker et al., 2004; that were defined based on phenotypic traits, suggesting instead
Zuccon et al., 2006; Lovette and Rubenstein, 2007). The Mimi- that this radiation has a complex within-archipelago history of dis-
dae + Sturnidae are in turn sister to the lineage comprising the persal, gene flow, and phenotypic differentiation (Arbogast et al.,
two species of Oxpeckers (Buphagidae), and fall within a broader 2006; Hoeck et al., 2010). Several studies have found the Nesomi-
songbird radiation that includes the thrushes and dippers (Cibois mus group to be embedded within the broader Mimus clade
and Cracraft, 2004; Zuccon et al., 2006; Lovette and Rubenstein, (Arbogast et al., 2006; Lovette and Rubenstein, 2007) and taxo-
2007). Several phenotypically odd New World taxa (e.g., Donacobi- nomically this group is now merged into Mimus (Remsen et al.,
us, Rhodinocichla) have at times been suggested to be allied to the 2010), but precisely where the Nesomimus lineage falls in relation
Mimidae, but phylogenetic evidence has since shown that they fall to these other mockingbirds remains uncertain.
elsewhere within the Passeriformes (Seutin and Bermingham, A somewhat analogous situation involves the Socorro Mocking-
1997; Alström et al., 2006; Johansson et al., 2008; Gelang et al., bird (Mimus graysoni), a morphologically distinct endangered spe-
2009). In reconstructions based on molecular data, a long and cies endemic to its remote namesake island. This species was
highly supported branch separates the Sturnidae from the earliest separated into the monotypic genus Mimodes until analyses of
node within the Mimidae (Zuccon et al., 2006; Lovette and mitochondrial DNA characters revealed that it falls well within
Rubenstein, 2007), thereby strongly supporting the monophyly of the Mimus mockingbird group (Barber et al., 2004).
the Mimidae as it is currently classified (e.g., A.O.U., 1998). Relationships within the Tropical (M. gilvus) and Northern (M.
The broad outline of phylogenetic relationships within the polyglottos) Mockingbird complex are complex and not fully re-
Mimidae is also well established (Hunt et al., 2001; Lovette and solved. These taxa hybridize in their zone of contact in central
Rubenstein, 2007). Species in the family fall into two clades, the Mexico (Wetmore, 1943), and they have sometimes been consid-
first of which comprises three species found in Mexico and north- ered conspecific. They are generally classified as separate species
ern Central America (Melanoptila and Melanotis), plus a group that (e.g., Davis and Miller, 1960; Mayr and Short, 1970; A.O.U., 1998)
has radiated mostly within the West Indian archipelago (Cinclocer- and mitochondrial DNA-based phylogenetic reconstructions sug-
thia, Margarops, Allenia, Ramphocinclus). The migratory and mono- gest that some populations of these taxa are well differentiated
typic Gray Catbird (Dumatella carolinensis) which breeds in North (Hunt et al., 2001). Morphological traits suggest this complex fur-
America is embedded within this otherwise West Indian clade ther includes the San Andrés Mockingbird, a distinctively large-
(Hunt et al., 2001; Lovette and Rubenstein, 2007). At the intraspe- billed population found only on its namesake islands in the wes-
cific level, the phylogeography of all endemic West Indian species tern Caribbean that has most often been considered a subspecies
found on more than one island (Cinclocerthia ruficada, C. gutturalis, of gilvus (e.g., Davis and Miller, 1960; A.O.U., 1998; Dickinson,
Margarops fuscatus, Allenia fusca) has been explored using mito- 2003), but which is sometimes treated as the full species M. mag-
chondrial markers (Hunt et al., 2001). nirostris (e.g., Bond, 1956).
The second and larger of the two subclades within the Mimidae Here we use mitochondrial and nuclear intron DNA sequence
comprises the continental thrashers (Toxostoma and Oreoscoptes) markers to reconstruct the phylogenetic relationships among all
and the mockingbirds (Mimus, plus the species formerly assigned Mimidae species, which we hope will aid in future comparative
to Nesomimus and Mimodes). Relationships among Toxostoma pop- analyses of morphological, ecological, and behavioral trait varia-
ulations and species have been reconstructed in an extensive series tion in this group.
of studies using allozyme, mitochondrial DNA, morphological, and
behavioral characters (Zink et al., 1997, 1999; Zink and Blackwell-
Rago, 2000; Zink et al., 2001; Rojas-Soto, 2003; Sgariglia and Burns, 2. Materials and methods
2003; Rojas-Soto et al., 2007). Collectively, these studies have sup-
ported the monophyly of the species assigned to Toxostoma, clari- 2.1. Taxonomic and molecular marker sampling
fied the relationships among many of their populations, and
identified some populations that might warrant future recognition We included in our analyses sequences from all 34 currently
as full species (e.g., Rojas-Soto et al., 2007). At a slightly deeper recognized species of Mimidae (A.O.U., 1998; Dickinson, 2003),
phylogenetic level, the Toxostoma clade is sister to the remaining including multiple representatives of most species (Appendix
taxa in this thrasher/mockingbird group. The monotypic Sage A). The majority of the samples obtained for this study allowed
Thrasher (Oreoscoptes montanus) is in turn sister to a group that in- the extraction of high-quality DNA, and from these we generally
cludes all of the Mimus mockingbirds (Lovette and Rubenstein, obtained sequences of five protein-coding mtDNA genes (ND2,
2007). CO1, CO2, ATPase8, and ATPase6) with a combined length of
The Mimus mockingbird radiation is relatively recent and pre- 4118 nucleotides, and sequences from three autosomal nuclear
sents a number of phylogenetic and systematic challenges. Mock- intron loci, beta-fibrinogen introns 5 (FGB-5; Kimball et al.,
ingbird species are distributed throughout much of North, 2009) and 7 (FGB-7; Prychitko and Moore, 1997), transforming
Central, and South America, and different lineages have colonized growth factor beta-2 intron 5 (TGFB2–5; Primmer et al., 2002),
both archipelagos (the West Indies and the Galapagos) and conti- and rhodopsin intron 1 (RHO-1; Primmer et al., 2002). The com-
nental satellite islands (Isla Socorro in the Pacific Ocean off of Mex- bined length of these nuclear loci is 2963 aligned nucleotides
ico; the Islas San Andrés in the Caribbean off of southern Central (not including regions of ambiguous alignment). Our laboratory
America). As is often the case for insular birds, many of the island methods for the generation of these sequences have been de-
mockingbird lineages have differentiated substantially in morpho- tailed elsewhere (Lovette and Rubenstein, 2007; Lovette et al.,
logical and behavioral traits. This differentiation is pronounced 2010). In some cases (see Appendix A) we did not obtain the full
within the Galapagos, where the four generally recognized species set of loci from a sample, as was also true in the few cases when
have evolved differences in bill shape and size, plumage pattern, the sequences were obtained from studies conducted in other
body size, song, and mating systems, among other traits (Gulledge, laboratories (Appendix A; Barber et al., 2004; Arbogast et al.,
1975; Abbott and Abbott, 1978; Curry, 1989). Phylogeographic sur- 2006).
veys of the Galapagos populations using mitochondrial (Arbogast We generated sequences from three species (Mimus graysoni,
et al., 2006) and microsatellite (Hoeck et al., 2010) markers have Toxostoma ocellatum, Toxostoma guttatum) and four Mexican Mimus
I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229 221
gilvus individuals (a species from which we also included to each of the three independent nuclear loci (FGB-5 and FGB-7
high-quality samples from other locations) from toe-pads taken are linked and were grouped into a single partition) and separate
from museum skin samples. Toe-pad samples were extracted and partitions for each codon position in the mtDNA sequences. Like-
amplified in a dedicated ancient-DNA laboratory facility using wise, the four mtDNA genes are a single linkage group and were
stringent protocols as described in Lovette and Rubenstein analyzed together. For each partition (mtDNA1stcodon,
(2007). From the M. graysoni and T. guttatum samples we se- mtDNA2ndcodon, mtDNA3rdcodon, FGB-5 + FGB-7, TGFB2–5, RHO-
quenced nearly the entire set of mtDNA coding genes; from the 1), we used the program MrModeltest v2 (Nylander, 2004) to iden-
T. occellatum and M. gilvus samples we sequenced only the mtDNA tify the best-fitting model under the Bayesian information criterion
ND2 gene. (BIC; Akaike, 1973; Kass and Raftery, 1995; Posada and Buckley,
Preliminary analyses showed that in all cases, sequences from 2004). The preferred model for all mtDNA partitions was
conspecific samples had very high similarity. Because we are pri- GTR + G + I and the preferred model for the Fib 5&7, TGFB5 and
marily concerned here with interspecific relationships, and to re- Rho1 partitions was HKY + G. These two models were applied
duce computation times, we selected a single representative of respectively to all analyses involving these data partitions.
each species to include in some phylogenetic analyses (Appendix
A), although we report additional analyses (see below) with all 2.4. Gene tree analyses
samples included.
Previous studies have provided robust evidence that the Mimi- To assess congruence among independent gene trees, we ran
dae species included here form a monophyletic group that is the separate Bayesian MCMC phylogenetic analyses on each of the
sister-group of the starlings (Sturnidae; e.g., Sibley and Ahlquist, four unlinked data partitions (mtDNA, FGB-5 + FGB-7, TGFB2–5,
1990; Barker et al., 2004; Cibois and Cracraft, 2004; Zuccon et al., RHO-1). To facilitate comparisons across datasets, these analyses
2006; Voelker and Spellman, 2004; Lovette and Rubenstein, included only single representatives of 30 Mimidae species from
2007). We included single representatives of five Sturnidae subc- which we sampled all loci (Mimidae species entirely missing from
lades (see Lovette and Rubenstein, 2007; Lamprotornis superbus, these nuclear gene trees are M. graysoni, T. guttatum, T. ocellatum,
Onychognathus morio, Aplonis panayensis, Sturnus vulgaris, Rhabdor- and T. curvirostre). Each gene-tree analysis was run in MrBayes for
nis inornatus) as outgroups, along with one representative (Bupha- 1 x 107 generations sampled every 1 x 104 generation, with five
gus africanus) of the sister group to the Mimidae + Sturnidae, and replicated runs of four chains with default heating. Plots of –ln
one more distant outgroup taxon (Catharus guttatus). likelihood scores, parameter values, and posterior probability
scores for individual nodes suggested that each analysis had con-
2.2. Bayesian phylogenetic methods verged before (usually well before) the 5 106 generation. The
topologies of all well-supported nodes of the two independent
We based all phylogenetic reconstructions on Markov chain runs also converged by the 5 106 generation, as did the poster-
Monte Carlo (MCMC) analyses conducted using MrBayes 3.1.2. ior probability scores of individual nodes as assessed graphically
(Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, using AWTY. We discarded the first 9 10 generations as ‘burn-
2003). The number of MCMC generations, number of concurrent in’ and combined the remaining 100 samples from each run to
chains, and number of separate runs varied among analyses, as de- calculate 90% majority-rule consensus topologies for each gene
scribed below. All MCMC runs were sampled every 1 104 gener- tree. The .con file created by MrBayes was visualized in FigTree
ations and were run using the parallel computing resources of the v.1.2.2. This consensus tree was confirmed in PAUP by combining
Computational Biology Service Unit at the Cornell University The- the last 90% of each .t file into one .t file. A majority-rule consen-
ory Center. sus tree was then visualized.
For each set of MCMC analyses, we evaluated the ‘burn-in’ per- We used a similar approach to analyze the mitochondrial se-
iod by plotting variation in log likelihood scores and individual quences alone.
parameter values across the duration of the run, and by evaluat-
ing variation in the posterior probability scores of the most vari-
2.5. Combined data analyses
able nodes using the web-based program AWTY (Wilgenbusch
et al., 2004). We then discarded the pre-stationarity output, and
We conducted further MCMC analyses that included all loci
to be conservative at least the initial 25% of the apparently sta-
sampled from all samples listed in Appendix A. These more com-
tionary samples, based on the slowest-converging metric. We
prehensive analyses were run with six partitions, three for the
evaluated convergence among runs by comparing the resulting
mitochondrial codon positions and one each for the three unlinked
consensus topologies, and by plotting the posterior probability
nuclear intron loci. Five independent runs, each with four concur-
scores of individual nodes using AWTY. The retained post-sta-
rent chains, were run for 1.3 107 generations. The first 1 106
tionarity samples from separate but convergent runs were then
generations were discarded as burn-in; the remaining samples
combined to calculate majority-rule consensus trees, posterior
from the independent runs were highly congruent, and the com-
probability values for each node, mean branch lengths, and
bined post-burn-in samples from all five runs were used to derive
parameter estimates.
a 50% majority rule consensus tree and to calculate mean branch
In addition to the phylogenetic analyses covered in detail be-
lengths.
low, we also ran a variety of preliminary analyses using other
phylogenetic methods (maximum likelihood and maximum parsi-
mony), alternative data partitioning schemes, and subsets of the 3. Results
overall dataset. Those results were highly consistent with the re-
sults of the comprehensive Bayesian analyses reported here. All sequences new to this study have been deposited in Gen-
Bank; Appendix A gives accession numbers for both these and all
2.3. Data partitions and model selection previously reported sequences.
The Bayesian analyses of all concatenated loci from all samples
Our dataset comprised a heterogeneous mix of mtDNA protein- produced a topology (Fig. 1) in which nearly all nodes are highly
coding and nuclear non-coding loci, and hence we used six data supported, and which is largely consistent with the relationships
partitions in our likelihood analyses, assigning single partitions found in previous studies that included subsets of these taxa as
222 I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229
Fig. 1. Relationships among species of Mimidae estimated by Bayesian MCMC analyses of concatenated mitochondrial and nuclear intron sequences subdivided into six
partitions (three mitochondrial partitions by codon position; nuclear locus partitions FGB-5 + FGB-7, TGFB2-5, RHO-1). The topology shown here is a 50% majority rule
consensus of post-burn-in trees. Numbers adjacent to nodes indicate posterior probability values >75. Topology is rooted to outgroup taxa (not shown) in the sister family
Sturnidae as described in the text. Geographic descriptors at right indicate primary area of breeding distribution.
I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229 223
described above. Novel or potentially important aspects of this concatenated analysis is driven overwhelmingly by the signal in
tree include the placement of the Galapagos mockingbird lineage the mitochondrial data partition.
as sister to the Bahama mockingbird (Mimus gundlachi) with a
posterior probability score of 100; this relationship was found
4. Discussion
with lower support in previous studies (Arbogast et al., 2006).
As has been found previously, both the Galapagos Mockingbird
4.1. Phylogenetic relationships within the Mimidae
group (previously separated as Nesomimus) and the Socorro
Mockingbird (previously in Mimodes) are embedded within a
Fig. 1 presents our most comprehensive summary of the phylo-
shallow clade that also includes all Mimus mockingbird taxa. Taxa
genetic relationships within the Mimdae based on our analyses of
that have not been included in earlier molecular phylogenetic
mitochondrial and nuclear DNA sequence markers. Under this
reconstructions include the Brown-backed Mockingbird (Mimus
phylogenenetic hypothesis, the Mimidae can be subdivided by a
dorsalis) of the southwestern Andes which falls within a clade
basal split into two daughter clades, one of which comprises the
of other continental South American mockingbirds, and the San
‘blue’ mockingbirds, New World catbirds, and Antillean thrashers,
Andrés Mockingbird (M. [gilvus] magnirostris) which falls within
the other containing the Mimus mockingbirds and the continental
the Northern/Tropical (M. polyglottos/gilvus) mockingbird
thrashers.
complex.
As expected at this level of divergence, all phylogenetic analyses
based on single nuclear loci had low resolution (Fig. 2), although 4.1.1. Blue mockingbirds, new world catbirds, and antillean thrashers
each provided some support for relationships that receive higher The blue mockingbird/catbird/antillean thrasher clade contains
support in the in the concatenated analysis. In contrast, the analy- nine species that are currently assigned to seven genera (Melanotis,
ses based on mitochondrial sequences alone (Fig. 3) produced a Melanoptila, Dumetella, Ramphocinclus, Allenia, Margarops, Cinclo-
strongly supported topology that was identical at the species and certhia; A.O.U., 1998). Relationships within this group were ex-
higher level to that of the concatenated analysis (Fig. 1). Consid- plored by Hunt et al. (2001) using mitochondrial markers and by
ered together, these results indicate that the topology in the Lovette and Rubenstein (2007) using both mitochondrial and
Fig. 2. Three single-locus gene tree hypotheses based on nuclear intron sequences. Topologies represent 50% majority rule consensus of post-burn-in trees. Numbers adjacent
to nodes indicate posterior probability values >75%. All reconstructions were rooted using Sturnidae outgroup taxa (not shown).
224 I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229
Dumetella carolinensis
100 98
Ramphocinclus brachyurus
100
Melanoptila glabrirostris
Toxostoma lecontei
91
Toxostoma redivivum
100
Toxostoma rufum
100
97
Toxostoma longirostre
91
Toxostoma guttatum
99
Toxostoma ocellatum
Toxostoma curvirostre
Fig. 3. Bayesian likelihood reconstruction of relationships among Mimidae species based on mitochondrial sequences alone. Numbers adjacent to nodes indicate posterior
probability values >75%. All reconstructions were rooted using Sturnidae outgroup taxa (not shown).
I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229 225
nuclear DNA data; our topologies here are consistent with those 4.1.2. Toxostoma thrashers
earlier results and the historical interpretations based on them. Our reconstructions of species-level relationships of the Toxos-
An early split within this clade separates the blue mockingbird toma thrashers are highly consistent with previous phylogenetic
lineage from a clade containing the remaining catbirds and Antil- hypotheses for this group (Zink et al., 1999; Lovette and Ruben-
lean thrashers. The two Melanotis blue mockingbirds have allopat- stein, 2007). We found these birds to be subdivided into two clades
ric distributions in Mexico and northern Central America, and they (Fig. 1), each comprised of five currently recognized species
are clearly sister taxa although they are separated by substantial (A.O.U., 1998, and later supplements). Within the first of these
genetic divergence in comparison to many other sister species groups, the Gray (T. cinereum) and Bendire’s (T. bendirei) thrashers
pairs within the Mimidae (Fig. 3). are sister species separated by relatively low genetic divergence,
All but one of the remaining taxa in the catbird/antillean consistent with the expectation that these taxa are closely allied
thrasher group have distributions within the Caribbean basin. (e.g., Mayr and Short, 1970; A.O.U., 1998; Zink et al., 1999). This
The basal split within this clade separates the Black Catbird (Mela- group also contains the more divergent sister-species pair formed
noptila glabrirostris) of the Yucatan peninsula from the remaining by the Crissal (T. crissale) and Le Conte’s (T. lecontei) thrashers,
species. This general topology, with two successive early-splitting which together are in turn sister to the California Thrasher
lineages now found on the Central American mainland, may hint (T. redivivium).
at the route of colonization into the West Indies taken by the The second Toxostoma clade includes the Curve-billed Thrasher
ancestor of the radiation now found in that island system. The sis- (T. curvirostre) as sister to the remaining four species; this topology
ter-clade to Melanotis comprises five species of thrashers resident is not consistent with the suggestion (Sibley and Monroe, 1990) that
within the West Indies, plus the Gray Catbird (Dumetella carolinen- T. curvirostre is most closely allied to the Ocellated Thrasher (T. ocell-
sis), a migratory species with a broad breeding range that spans atum). However, the clade formed by the Brown (T. rufum), Long-
much of the United States and southern Canada, and which winters billed (T. longirostre), and Cozumel (T. guttatum) thrashers matches
in the continental regions surrounding the Gulf of Mexico and on previous predictions about their close affinities (Mayr and Short,
the western islands in the Greater Antilles. As found previously 1970; A.O.U., 1998; Zink et al., 1999; Sibley and Monroe, 1990).
(Hunt et al., 2001; Lovette and Rubenstein, 2007), Dumetella is
embedded within the Antillean thrasher radiation as the sister tax- 4.1.3. Mimus mockingbirds
on of Ramphocinclus brachyurus, the White-breasted Thrasher, a The Mimus mockingbirds form a shallow clade of 14 or 15 spe-
highly endangered species found only on the Lesser Antillean is- cies that collectively are sister to the Sage Thrasher (Oreoscoptes
lands of Martinique and St. Lucia. Based on the phylogenetic evi- montanus) lineage (Fig. 1). Subgroups within this relatively recent
dence, it seems most likely that the ancestral Dumetella lineage radiation include a clade of five species endemic to South America.
recolonized North America from the West Indies, adding to the Within this South American group, the two southernmost species
growing list of avian examples of such ‘reverse colonization’ of (the Chilean Mockingbird M. thenca and Patagonian Mockingbird
continental areas from island source populations that have been M. patagonicus) are allopatric sister taxa separated by the heights
identified from phylogenetic evidence (Bellemain and Ricklefs, of the southern Andes. The range-restricted Brown-backed Mock-
2008; Sheldon et al., 2009; Sturge et al., 2009). The Dumetella line- ingbird (M. dorsalis), which has not been included in previous phy-
age likewise probably transitioned to long-distance migration from logenetic surveys of this genus, is sister to the White-banded
a sedentary ancestor, a change consistent with predictions that Mockingbird (M. triurus), and these two species are in turn sister
avian migratory behaviors have high evolutionary lability despite to the broadly distributed Chalk-browed Mockingbird (M.
the complex suite of traits necessary to sustain successful migra- saturninus).
tion (Berthold, 1999; Zink, 2002). A second group within Mimus comprises the Northern (M.
The five species of resident Antillean thrashers are all found in polyglottos) and Tropical (M. gilvus) Mockingbirds along with
the Lesser Antilles, with one (Margarops fuscatus) extending its two island forms, the Socorro Mockingbird (Mimus [Mimodes]
range to Puerto Rico and the Bahamian archipelago. These birds graysoni) and the San Andrés Mockingbird (M. [gilvus] magniros-
have a complex pattern of island occupancy, with some Lesser tris). As found previously (Barber et al., 2004), M. graysoni is
Antillean islands supporting all five species. The morphological embedded within the larger Mimus clade, where it sister to the
and behavioral variation within this shallow clade of thrashers is remaining members of the M. polyglottos/gilvus group. Relation-
correspondingly high, and this group may represent the sole avian ships among the polyglottos/gilvus group are complicated, as both
example of a Lesser Antillean adaptive radiation (Ricklefs and Ber- M. magnirostris and M. polyglottos fall within the M. gilvus cluster,
mingham, 2008). thereby rendering M. gilvus paraphyletic. At this level of diver-
The high morphological distinctiveness of the Antillean gence, nearly all of the phylogenetic signal in our analyses derives
thrasher species has resulted in all but Cinclocerthia being currently from the mitochondrial loci (see Fig. 2), so this paraphyly reflects
classified into monotypic genera (A.O.U., 1998; Banks et al., 2002). the history of the associated mitochondrial gene tree. In the ab-
Although many of these taxa are morphologically divergent from sence of new behavioral and genetic information from the M.
the other species in this clade, it could be argued that this group polyglottos/gilvus hybrid zone in Central Mexico, we recommend
is generically over-split, as this Antillean thrasher clade is likely retaining the current species-level taxonomy for these groups,
no older than the Mimus or Toxostoma clades which similarly con- recognizing both M. gilvus and M. polyglottos but considering M.
tain a range of species with high phenotypic variation, including magnirostris a subspecies of M. gilvus.
some island forms that have differentiated notably from their con- A final group within Mimus comprises the Galapagos mocking-
tinental allies. There is no intermediate generic reassignment that bird clade, and their apparent sister taxon, the Bahama Mocking-
recognizes monophyletic subgroups within this clade while also bird (M. gundlachii).
reflecting the major phenotypic differences among these taxa. If
the genera we currently recognize as Melanoptila, Dumetella, Ram- 4.2. Historical biogeography of oceanic island mockingbirds
phocinclus, Allenia, Margarops, and Cinclocerthia were grouped into
a single genus, Dumetella Wood, 1837 would have nomenclatural The current distribution of Mimus mockingbirds suggests that
priority. these birds (or their recent ancestors) are accomplished
226 I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229
EF484115
EF484120
pendently colonized one or more oceanic islands. In addition to
TGFB2-4
Pending
Pending
Pending
EF484125
EF484119
being found across much of North America, the Northern Mock-
Pending
Pending
ingbird (M. polyglottos) is resident through the Bahamas, the large
—
—
islands of the Greater Antilles except for Jamaica, and the Cayman
Islands; phylogeographic studies have shown that these West In-
EF472856
EF472861
Pending
Pending
Pending
EF472866
EF472860
dian populations are closely allied to the more northerly conti-
Pending
Pending
Pending
Rho-1
nental populations, implying a recent set of island colonizations
(Hunt et al., 2001). The Tropical Mockingbird (M. gilvus) has sim-
—
ilarly colonized the most of the Lesser Antillean island chain from
EF471844
EF471849
South America (Hunt et al., 2001). In the western Caribbean, a
Pending
Pending
Pending
EF471854
EF471848
Nuclear intron loci
Pending
Pending
Pending
separate colonization by a lineage in the M. polyglottos/gilvus
Fib-7
group gave rise to the unusually large-billed population (spe-
—
cies/subspecies M. magnirostris) on San Andrés Island; the mito-
EF468323
EF468328
chondrial evidence suggests that the San Andrés population is
Pending
Pending
Pending
EF468333
EF468327
allied to M. gilvus lineages from southern Mexico (Fig. 1), thereby
Pending
Pending
Fib-5
suggesting its likely colonization source. The more divergent
—
—
mockingbird population (M. graysonii) endemic to Isla Socorro–a
small, arid island in the Pacific Ocean located about 600 km west
EF486776
EF486781
ATPasese
Pending
Pending
Pending
of the central Mexican coast – is apparently derived from the
EF486786
EF486780
Pending
Pending
Pending
lineage that is also ancestral to the entire M. polyglottos/gilvus
group, implying a substantially earlier colonization of that remote
—
island along with a substantial period of isolation of M. graysonii.
Taxa included in this study, tissue types, collecting localities, institutional sources, and GenBank accession numbers.
EF484317
EF484322
(Fig. 1).
Pending
Pending
Pending
EF484327
EF484321
Pending
Pending
Pending
The mockingbird dispersal event with the most notable evolu-
CO2
Mitochondrial coding genes
tionary consequences was the colonization of the Galapagos by
—
the ancestor of the lineages that then radiated in that archipelago.
Darwin noted the affinities of the Galapagos mockingbirds to spe-
EF484216
EF484221
Pending
Pending
Pending
EF484226
EF484220
cies of Mimus from the South American continent (Darwin, 1859),
Pending
Pending
Pending
and it was long assumed that the Galapagos lineage was most
closely related to species found on that relatively nearby source. CO1
—
Mitochondrial sequence-based analyses by Arbogast et al.
EF468192
EF468197
(2006) instead identified the Bahama Mockingbird (M. gundlachii)
Pending
Pending
Pending
EF468196
EF468202
Pending
Pending
Pending
Pending
as the sister lineage to the Galapagos clade, a finding mirrored in
ND2d
being recovered in our phylogenetic analyses of mitochondrial se- USA, New York, Tompkins
Trinidad, Chacachacare I.
recovered in the independent topologies from two of our three
Colombia
Colombia
Mexico
Co.
Acknowledgments
Typeb
B
B
T
T
T
LSUMNS-B21369
CUMV-50343
CUMV-50532
CUMV-50356
CUMV-33468
CUMV-50469
STRI-CCMGI1
glabrirostris
carolinensis
carolinensis
carolinensis
Melanoptila
gilvus
gilvus
gilvus
gilvus
Dumetella
Dumetella
Dumetella
Mimus
Mimus
Mimus
Mimus
Mimus
Mimus
227
228 I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229
Institutional sources of samples: ANSP: Academy of Natural Sciences, Philadelphia, Pennsylvania, USA; BMNH: Bell Museum of Natural History, University of Minnesota, Minnneapolis, Minnesota, USA; CUMV: Cornell
University Museum of Vertebrates, Ithaca, New York, USA; LSUMNS: Louisiana State University Museum of Natural Science, Baton Rouge, Louisiana, USA; MNHN: Museo Nacional de Historia Natural, Santiago, Chile; SDSU: San
NSF-DEB-0515981.
TGFB2-4
EF484118
EF484117
Pending
Pending
Pending
— References
EF472858
Akaike, H., 1973. Information theory and an extension of the maximum likelihood
Pending
Pending
Pending
Rho-1
Alström, P., Ericson, P.G.P., Olsson, U., Sundberg, P., 2006. Phylogeny and
classification of the avian superfamily Sylvioidea. Mol. Phylogenet. Evol. 38,
381–397.
EF471847
EF471846
Nuclear intron loci
Pending
Pending
Pending
American Ornithologists’ Union, 1998. Checklist of North American Birds, 7th ed.
Fib-7
EF468325
Banks, R.C., Cicero, C., Dunn, J.L., Kratter, A.W., Rasmussen, P.C., Remsen Jr., J.V.,
Pending
Pending
Pending
Barber, B.R., Martinez-Gomez, J.E., Peterson, A.T., 2004. Systematic position of the
Socorro mockingbird Mimus graysoni. J. Avian Biol. 35, 195–198.
Barker, F.K., Cibois, A., Schikler, P., Cracraft, J., 2004. Phylogeny and diversification of
e
ATPases
EF486779
EF486778
the largest avian radiation. Proc. Nat. Acad. Sci. USA 101, 11040–11045.
Pending
Pending
Pending
Bond, J., 1956. Check-List of Birds of the West Indies. Academy of Natural Sciences of
Pending
Pending
Pending
Philadelphia.
CO2
Mitochondrial coding genes
Botero, C.A., Rossman, R.J., Caro, L.M., Stenzler, L.M., Lovette, I.J., de Kort, S.R.,
—
Vehrencamp, S.L., 2009a. Syllable type consistency is related to age, social status
and reproductive success in the tropical mockingbird. Anim. Behav. 77, 701–
706.
EF484219
EF484218
Pending
Pending
Pending
Botero, C.A., Boogert, N.J., Vehrencamp, S.L., Lovette, I.J., 2009b. Climatic patterns
CO1
Diego State University Museum of Biodiversity, San Diego, CA, USA; WFU: Wake Forest University, Winston-Salem, NC, USA.
predict the elaboration of song displays in mockingbirds. Curr. Biol. 19, 1151–
1155.
—
Cibois, A., Cracraft, J., 2004. Assessing the passerine ‘‘Tapestry’’: phylogenetic
relationships of the Muscicapoidea inferred from nuclear DNA sequences. Mol.
EF468195
EF468194
EF468191
Pending
Pending
Curry, R.L., 1988. Group structure, within-group conflict and reproductive tactics in
d
ND2
Countries Visited during the Voyage of H.M.S. Beagle round the world, second
Guadaloupe
Darwin, C., 1859. On the Origin of Species by Means of Natural Selection, or the
St. Lucia
St. Lucia
Preservation of Favoured Races in the Struggle for Life. John Murray, London.
Davis, J., Miller, A.H., 1960. Family mimidae. In: Mayr, E., Greenway, J.C., Jr. (Eds.),
Check-List of Birds of the World, vol. IX. Museum of Comparative Zoology,
Cambridge, Massachusetts, pp. 440–458.
Typeb
Derrickson, K.C., 1987. Yearly and situational changes in the estimate of repertoire
size in Northern Mockingbirds (Mimus polyglottos). Auk 104, 198–207.
T
T
T
T
based on nuclear and mitochondrial sequence data. Mol. Phylogenet. Evol. 29,
126–138.
STRI-DOMFU3
sample No.
STRI-GUCRU1
STRI-BUMFT1
STRI-SLMFU4
STRI-SLCRU2
Gelang, M., Cibois, A., Pasquet, E., Olsson, U., Alström, P., Ericson, P.G.P., 2009.
ANSP-B2134
Cinclocerthia
Cinclocerthia
Allenia fusca
Allenia fusca
Hunt, J.S., Bermingham, E., Ricklefs, R.E., 2001. Molecular systematics and
a (continued)
gutturalis
ruficauda
ruficauda
d
c
based on three nuclear intron markers. Mol. Phylogenet. Evol. 48, 858–876.
I.J. Lovette et al. / Molecular Phylogenetics and Evolution 63 (2012) 219–229 229
Kass, R.E., Raftery, A.E., 1995. Bayes factors. J. Am. Stat. Assoc. 90, 773–795. Seutin, G., Bermingham, E., 1997. Rhodinocichla rosea is an emberizid (Aves;
Kimball, R.T., Braun, E.L., Barker, F.K., Bowie, R.C.K., Braun, M.J., Chojnowski, J.L., Passeriformes) based on mitochondrial DNA analyses. Mol. Phylogenet. Evol. 8,
Hackett, S.J., Han, K.-L., Harshman, J., Heimer-Torres, V., Holznagel, W., 260–274.
Huddleston, C.J., Marks, B.D., Miglia, K.J., Moore, W.S., Reddy, S., Sheldon, F.H., Sgariglia, E.A., Burns, K.J., 2003. Phylogeography of the California Thrasher
Smith, J.V., Witt, C.C., Yuri, T., 2009. A well-tested set of primers to amplify (Toxostoma redivivum) based on nested-clade analysis of mitochondrial-DNA
regions spread across the avian genome. Mol. Phylogenet. Evol. 50, 654–660. variation. Auk 120, 346–361.
Kroodsma, D.E., Houlihan, P.W., Fallon, P.A., Wells, J.A., 1997. Song development by Sheldon, F.H., Lohman, D.J., Lim, H.C., Zou, F., Goodman, S.M., Prawiradilaga, D.M.,
grey catbirds. Anim. Behav. 54, 457–464. Winker, K., Braile, T.M., Moyle, R.G., 2009. Phylogeography of the magpie-robin
Kroodsma, D.E., Parker, L.D., 1977. Vocal virtuosity in brown thrasher. Auk 94, 783– species complex (Aves: Turdidae: Copsychus) reveals a Philippine species, an
785. interesting isolating barrier and unusual dispersal patterns in the Indian Ocean
Lovette, I.J., Rubenstein, D.R., 2007. A comprehensive molecular phylogeny of the and Southeast Asia. J. Biogeogr. 36, 1365–2699.
starlings (Aves: Sturnidae) and mockingbirds (Aves: Mimidae): congruent Sibley, C.G., Ahlquist, J.E., 1984. The relationships of the starlings (Sturnidae:
mtDNA and nuclear trees for a cosmopolitan avian radiation. Mol. Phylogenet. Sturnini) and the mockingbirds (Sturnidae: Mimini). Auk 101, 230–243.
Evol. 44, 1031–1056. Sibley, C.G., Ahlquist, J.E., 1990. Phylogeny and Classification of Birds. Yale
Lovette, I.J., Pérez-Emán, J.L., Sullivan, J., Banks, R.C., Fiorentino, I., Córdoba-Córdoba, University Press, New Haven.
S., Echeverry-Galvis, M., Barker, F.K., Burns, K., Klicka, J., Lanyon, S.M., Sibley, C.G., Monroe, B.L., 1990. Distribution and Taxonomy of Birds of the World.
Bermingham, E., 2010. A comprehensive multilocus phylogeny for the Wood- Yale University Press, New Haven.
Warblers and a revised classification of the Parulidae (Aves). Mol. Phylogenet. Stallcup, W.B., 1961. Relationships of some families of the suborder Passeres
Evol. 57, 753–770. (songbirds) as indicated by comparisons of tissue proteins. J. Graduate Res.
Malcarney, H.L., Martinez, D.R.C., Apanius, V., 1994. Sucrose intolerance in birds: Center South. Method. Univ. 29, 43–65.
simple nonlethal diagnostic methods and consequences for assimilation of Sturge, R.J., Jacobsen, F., Rosensteel, B.B., Neale, R.J., Omland, K.E., 2009. Colonization
complex carbohydrates. Auk 111, 170–177. of South America from Caribbean islands confirmed by molecular phylogeny
Mayr, E., Short, L.L., 1970. Species Taxa of North American Birds. Publications of the with increased taxon sampling. Condor 111, 575–579.
Nuttall Ornithological Club No. 9. Voelker, G., Spellman, G.M., 2004. Nuclear and mitochondrial DNA evidence of
Nylander, J.A.A., 2004. MrModeltest v2. Program distributed by the author. polyphyly in the avian superfamily Muscicapoidea. Mol. Phylogenet. Evol. 30,
Evolutionary Biology Centre, Uppsala University. 386–394.
Parent, C.E., Caccone, A., Petren, K., 2008. Colonization and diversification of Wetmore, A., 1943. The birds of southern Veracruz, Mexico. Proc. US Natl. Mus. 93,
Galápagos terrestrial fauna: a phylogenetic and biogeographical synthesis. Proc. 215–340.
Roy. Soc. Lond. B 363, 3347–3361. Wilgenbusch, J.C., Warren, D.L., Swofford, D.L., 2004. AWTY: A System for Graphical
Posada, D., Buckley, T.R., 2004. Model selection and model averaging in Exploration of MCMC Convergence in Bayesian Phylogenetic Inference. <http://
phylogenetics: advantages of Akaike information criterion and bayesian ceb.csit.fsu.edu/awty>.
approaches over likelihood ratio tests. Syst. Biol. 53, 793–808. Zink, R., 2002. Towards a framework for understanding the evolution of avian
Primmer, C.R., Borge, T., Lindell, J., Sætre, G.-P., 2002. Single-nucleotide migration. J. Avian Biol. 33, 433–436.
polymorphism characterization in species with limited available sequence Zink, R.M., Blackwell-Rago, R.C., 2000. Species limits and recent population history
information: high nucleotide diversity revealed in the avian genome. Mol. Ecol. in the curve-billed thrasher. Condor 102, 881–886.
11, 603–612. Zink, R.M., Blackwell, R.C., Rojas-Soto, O., 1997. Species limits in the Le Conte’s
Prychitko, T.M., Moore, W.S., 1997. The utility of DNA sequences of an intron from Thrasher. Condor 99, 132–138.
the beta-fibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Zink, R.M., Dittmann, D.L., Klicka, J., Blackwell-Rago, R.C., 1999. Evolutionary
Picidae). Mol. Phylogenet. Evol. 8, 193–204. patterns of morphometrics, allozymes, and mitochondrial DNA in thrashers
Read, A.F., Weary, D.M., 1992. The evolution of bird song: comparative analyses. (genus Toxostoma). Auk 116, 1021–1038.
Philos. Trans. Roy. Soc. Lond. B 338, 165–187. Zink, R.M., Kessen, A.E., Line, T.V., Blackwell-Rago, R.C., 2001. Comparative
Remsen Jr., J.V., Cadena, C.D., Jaramillo, A., Nores, M., Pacheco, J.F., Robbins, M.B., phylogeography of some aridland bird species. Condor 103, 1–10.
Schulenberg, T.S., Stiles, F.G., Stotz, D.F., Zimmer, K.J., Version 19 August 2010. A Zuccon, D., Cibois, A., Pasquet, E., Ericson, P.G.P., 2006. Nuclear and mitochondrial
Classification of the Bird Species of South America. American Ornithologists’ sequence data reveal the major lineages of starlings, mynas and related taxa.
Union. <http://www.museum.lsu.edu/~Remsen/SACCBaseline.html>. Mol. Phylogenet. Evol. 41, 333–344.
Ricklefs, R.E., Bermingham, E., 2008. The West Indies as a laboratory of
biogeography and evolution. Philos. Trans. Roy. Soc. Lond. B 363, 2393–2413.
Rojas-Soto, O.R., 2003. Geographic variation of the curve-billed thrasher
Further reading
(Toxostoma curvirostre) complex. Auk 120, 311–322.
Ronquist, F., Huelsenbeck, J.P., 2003. MRBAYES 3: bayesian phylogenetic inference Posada, D., Crandal, K.A., 1998. ModelTest: testing the model of DNA substitution.
under mixed models. Bioinformatics 19, 1572–1574. Bioinformatics 14, 817–818.