Journal of Biogeography (J. Biogeogr.) (2015)

ORIGINAL Land connectivity changes and global ARTICLE cooling shaped the colonization history and diversification of (Aves: : Odontophoridae) Peter A. Hosner1*, Edward L. Braun1,2,3 and Rebecca T. Kimball1,2,3

1Department of Biology, University of Florida, ABSTRACT Gainesville, FL, USA, 2Genetics Institute, Aim Range disjunctions are frequent in , but the relative roles of vicari- University of Florida, Gainesville, FL, USA, 3 ance and long-distance dispersal in producing them are debated. Odontophorid Florida Museum of Natural History, University of Florida, Gainesville, FL, USA quail are widespread in tropical and temperate habitats in the Americas, yet recent phylogenetic studies support the view that they are sister to sub-Saharan African Ptilopachus rather than the widespread as formerly believed. To understand how this 10,000 km range disjunction arose in rela- tively non-vagile birds, we reconstructed colonization history and diversifica- tion of odontophorids with respect to hypothesized dry-land connections between continents (North Atlantic, Beringian, Panamanian) that would have facilitated faunal exchange. Location Africa, Nearctic and Neotropics. Methods We inferred a fossil-calibrated odontophorid phylogeny from DNA sequences (three mitochondrial genes and eight nuclear introns) and modelled ancestral ranges with six probabilistic biogeographical models. We used the Akaike information criterion (AIC) to select the best-fit biogeographical model.

Results Ptilopachus and New World quail shared an Old World ancestor c. 32 Ma. During this period, Beringia connected the Nearctic and Palaearctic, and global temperatures were high, such that presence of temperate organisms at high latitudes and direct dispersal across land connections were feasible. The extant New World quail began diversifying in Central America c. 18 Ma; tim- ing estimates and ancestral range reconstructions support the hypothesis that New World quail colonized and diversified in South America following closure of the Isthmus of Panama.

Main conclusions The Africa/New World range disjunction between New World quail and Ptilopachus is the result of changes in Earth and climate his- tory, combined with range expansion and diversification in the New World, and range contraction in the Old World. We find no evidence for overwater

*Correspondence: Peter A. Hosner, dispersal in New World quail. Department of Biology, PO Box 118525, Keywords University of Florida, Gainesville, FL 32611, USA. Beringia, climate change, dispersal, faunal interchange, Isthmus of Panama, E-mail: hosner@ufl.edu phylogeny, Ptilopachus, range disjunction, vicariance.

(Wiley, 1988; de Queiroz, 2005); that is, does long-distance INTRODUCTION dispersal or ancient vicariance produce these distributions? Organisms with limited dispersal capacity often exhibit Beyond understanding historical biogeography, understand- widely disjunct distributions. Ideas about the evolutionary ing the relative roles of dispersal and vicariance in creating processes that produce range disjunctions date to the foun- disjunct distributions is critical to understanding macroevo- dations of biogeography (Nelson, 1978; Sanmartın et al., lutionary patterns, such as geographical disparity in diversifi- 2001) and are central to the dispersal–vicariance debate cation rates (Rabosky, 2009). If driven by dispersal, range

ª 2015 John Wiley & Sons Ltd http://wileyonlinelibrary.com/journal/jbi 1 doi:10.1111/jbi.12555 P. A. Hosner et al. disjunctions are indicative of range expansion and increased connected Asia to North America (Simpson, 1947; Hopkins, potential for diversification. Alternatively, if driven by 1967; Marincovich & Gladenkov, 1999; Brikiatis, 2014). The ancient vicariance, disjunct distributions are relictual, and history and timing of North Atlantic land bridges is complex, indicative of range contraction and extinction. confusing and not wholly understood (e.g. Denk et al., 2011; In birds, continental range disjunctions are frequent in Brikiatis, 2014). In general, geological evidence supports the higher taxa. Common patterns include groups that are view that eastern North America and Europe were largely diverse in Afrotropical, Indomalayan and Neotropical regions connected until at least 55 Ma via the De Geer Route, and (e.g. barbets, Capitonidae; and trogons, Trogoniformes; later the Thulean Route (Brikiatis, 2014). Subaerial connec- Moyle, 2004; Hosner et al., 2010). In other groups, range tions established across Beringia around 100 Ma facilitated disjunctions are coupled with disparity in diversity, with faunal turnover until the late Miocene (Hopkins, 1967; Mar- species-rich clades inhabiting one continent and ‘relict’ incovich & Gladenkov, 1999; Brikiatis, 2014) when continu- lineages in another (Moyle et al., 2006, 2012; Reddy & ous land connections were interrupted by elevated sea levels Cracraft, 2007). There are similarly notable continental range (3–7.5 Ma, Marincovich & Gladenkov, 1999; 10–12 Ma, disjunctions documented in the fossil record (Mayr, 2004; Hopkins, 1967). From the late Pliocene until the present, Ksepka & Clarke, 2010; Nesbitt et al., 2011). land connections were periodically restored during glacial Birds are justly viewed as vagile organisms, and long-dis- maxima, the most recent of which ended c. 10 ka (Marinco- tance dispersal is a cause of range disjunction in some vich & Gladenkov, 1999; Miller, 2005). groups and species (Clegg et al., 2002; Billerman et al., Distributions of tropical/temperate organisms occurring 2011). However, despite having powered flight, birds fre- across high latitudes in Europe, Asia and North America in quently evolve limited ability to disperse to and colonize the warm early Tertiary would later be restricted to lower novel environs. In continental systems, tropical forest under- latitudes owing to massive global cooling events (first in the storey birds are often poor dispersers adapted to continuous early/middle Oligocene, and later in the late Miocene). Thus, habitats (Moore et al., 2008; Burney & Brumfield, 2009). climate-driven vicariance could widely explain avian distribu- Reduction of dispersal is considered advantageous for land tional disjunctions between the New World and the Old birds in island systems where dispersal ability may have World. Coincident with global cooling in the late Miocene, extreme fitness costs (Cody & Overton, 1996; Moyle et al., land connections initiated between Africa and the Mideast 2009). An extreme example of reduced dispersal is complete allowed dispersal and faunal interchange between Eurasia loss of flight, which has evolved multiple times in avian and Africa (Bibi, 2011). Similar faunal exchange occurred groups as diverse as rails (Rallidae), ducks (Anatidae) and between North and South America across the Isthmus of grebes (Podicipedidae) (Livezey, 1989; Fulton et al., 2012; Panama, in two pulses. An initial, minor pulse was the result Kirchman, 2012). of a brief or near (as a peninsula or archipelago) formation Avian biogeographers have historically favoured vicariant of an isthmus in the late Miocene (8–9 Ma, Coates et al., explanations for range disjunctions, particularly in pan-tropi- 2004; Montes et al., 2015). The second, major pulse was the cal groups. This view was fuelled by putative Gondwanan result of complete and permanent isthmus closure 3–4Ma distributions of poorly dispersing groups like the flightless (Webb, 1976; Weir et al., 2009; Leigh et al., 2013; Barker ratites (Cracraft, 2001). Molecular phylogenies now reject a et al., 2015). Gondwanaland origin for ratites, and instead support the Galliformes (landfowl) are almost worldwide in distribu- hypothesis that flighted ratite ancestors colonized continents tion, yet are among the least vagile orders of flying birds. across marine barriers, followed by convergent losses of flight There are two exceptions to this generalization. Megapodes (Smith et al., 2013; Mitchell et al., 2014). However, more (which are sister to all other Galliformes) are extremely vag- recent vicariance-driven disjunction remains a plausible ile and multiple lineages have colonized islands across the explanation for avian groups distributed across Africa, Asia Indo-Pacific region (Harris et al., 2014). Also, small-bodied and the New World. Facilitated by changing land connectiv- species in or related to the genus Coturnix (Old World quail) ity and climates, now-disjunct representatives of tropical/ are frequent on oceanic islands and have crossed Wallace’s temperate groups of Laurasian origin could conceivably and Lydekker’s lines. They are often nomadic or capable of occupy all continents except Australia and Antarctica without trans-continental seasonal migration (Johnsgard, 1988). Out- ever having crossed a marine barrier. side of megapodes and Old World quail there are no clear Land connections existed between the Palaearctic and examples of Galliformes occurring naturally on oceanic Nearctic through much of the Tertiary (reviewed in Brikiatis, islands (those without recent mainland connections during 2014). This extended connectivity, in combination with mild low sea-level stands, or connected by arctic sea ice in the global temperatures (Zachos, 2001), permitted broad distri- unique case of ptarmigans, Lagopus), suggesting that most butions of tropical and temperate organisms across Europe, Galliformes are incapable of crossing marine barriers. Based Asia and North America in the Palaeocene and . Two on a time-calibrated phylogeny, Pereira & Baker (2006) sug- routes provided opportunity to disperse between the Palae- gested that major galliform lineages were old enough to fit arctic and Nearctic: North Atlantic land bridges connected the Gondwanan vicariance model, but Harris et al. (2014) Europe to North America, and the Beringia land bridge suggested that their diversification was too recent.

2 Journal of Biogeography ª 2015 John Wiley & Sons Ltd Phylogeny of New World quail

The New World quail comprise 32 temperate and tropical World quail diversity in South America the result of rapid species that inhabit open and forested landscapes from diversification following colonization of a new continent? extreme southern Canada to southern Brazil and north- eastern Argentina (Gill & Donsker, 2015). The serrated MATERIALS AND METHODS ‘toothed’ edge of the mandible is a unique character that has long defined the clade (Johnsgard, 1988). The diversity of Data collection New World quail reaches its peak in Mexico and northern Central America, where 17 species occur and up to eight We extracted genomic DNA from frozen or ethanol pre- species can be found in proximity. Fourteen species occur in served tissue from 23 of 34 currently recognized odontopho- South America, and six species are found in temperate North rid species (see Table S1 in Appendix S1 of the Supporting America. Species such as California Quail (Callipepla Information; Gill & Donsker, 2015) using the Gentra Pure- californica) and Northern Bobwhite (Colinus virginianus) are gene DNA Purification Kit (Qiagen Inc., Valencia CA, common, familiar and well-studied birds. USA). Sampling included all recognized genera, and at least Recent molecular studies generally support the hypothesis 50% of recognized species within genera (Gill & Donsker, that New World quail are sister to the two species of Ptilopa- 2015). We included five phasianid species as outgroups chus partridges distributed in sub-Saharan Africa (Crowe (including New and Old World representatives), and rooted et al., 2006; Cohen et al., 2012; Meiklejohn et al., 2014). to the guineafowl Numida meleagris. We selected three mito- Recent classifications have combined New World quail and chondrial genes [12S rRNA (12S), NADH dehydrogenase Ptilopachus into a single family, Odontophoridae (Bowie subunits 2 (ND2) and 5 (ND5)] and eight nuclear introns et al., 2013; Gill & Donsker, 2015). The position of New [beta-fibrinogen intron 5 (FGB-5), clathrin heavy chain World quail within the Galliformes is now established (Cox intron 7 (CLTC)], clathrin heavy chain-like 1 intron 7 et al., 2007; Wang et al., 2013; Kimball & Braun, 2014; but (CLTCL1), eukaryotic translation elongation factor 2 intron see Pereira & Baker, 2006), but molecular phylogenetic 5(EFF2), muscle skeletal receptor tyrosine kinase (MUSK), hypotheses within New World quail are poorly developed. rhodopsin intron 1 (RHO), ovalbumin intron 3 (SER- Previous studies have included a few individuals of select PINB14), transforming growth factor beta 2 intron 5 genera to resolve higher-level relationships with respect to (TGFb2-5)] for sequencing. Primers and PCR amplification other galliform clades, are based on small fragments of mito- protocols follow Cox et al. (2007) and Kimball et al. (2009). chondrial DNA (Williford, 2013), or focused solely on the Cycle sequencing of PCR products was completed using ABI tm genus Callipepla (Zink & Blackwell, 1998). Given the lack of BigDye Terminator v.3.1 sequencing kits, products were tm a robust molecular framework, hypotheses about the system- analysed on an ABI Prism 3100-Avant or a 3730xl DNA atics and biogeography of New World quail (Fig. 1) are lar- Analyzer (Applied Biosystems, Foster City, CA, USA). Result- gely based on similarity of skeletal characters, natural history ing chromatograms were quality-controlled and reconciled in and distributional patterns (Holman, 1961; Johnsgard, 1988). Geneious 6 (Kearse et al., 2012). Sequences generated for Jetz et al. (2012) included Odontophoridae in their syn- this study (GenBank KR732672–KR732896) were combined thetic, constrained supermatrix/supertree phylogeny of all with previously published data from two higher-level galli- birds (Fig. 1) – although only 10 of 34 odontophorid species form phylogenetic studies (Cox et al., 2007; Kimball & and five of 10 odontophorid genera had available genetic Braun, 2014). data. The phylogeny of Jetz et al. (2012) placed Philortyx and Dendrortyx sister to all other New World quail, even though Phylogenetic analysis both genera lacked DNA sequence data and the relationship was unsupported by previous morphological studies (Hol- We aligned sequences in mafft 7 (Katoh & Standley, 2013) man, 1961; Johnsgard, 1988). Otherwise, the results of Jetz invoked in Genious 6. To identify an optimal partitioning et al. (2012) appeared to be driven by a constraint following scheme and corresponding model of sequence evolution for the Odontophorus group of Holman (1961) and the mito- the dataset, we implemented PartitionFinder 1.1 (Lanfear chondrial study of Zink & Blackwell (1998). et al., 2012) using the Bayesian information criterion (BIC). Here, we present a multilocus molecular phylogenetic Data subsets examined for partitioning included 12S, each hypothesis of New World quail. With this phylogeny, we codon for position of each mitochondrial gene, and each address two key biogeographical questions raised by the real- nuclear intron. Model testing included JC, F81, HKY and ization that New World quail are sister to African Ptilopa- GTR; with and without invariant sites and gamma- chus. (1) Does isolation of Ptilopachus and New World quail distributed rate heterogeneity. fit an Oligocene/late Miocene cooling vicariant model, or We used Garli 2.0 (Zwickl, 2006) to infer maximum like- was long-distance colonization across marine barriers by vag- lihood (ML) trees and sets of 500 bootstrap replicates for: ile ancestors necessary? (2) If (some) New World quail or (1) the entire concatenated dataset, (2) concatenated their ancestors were more vagile, did colonization of South mtDNA, (3) concatenated nuclear DNA, and (4) each America pre-date closure of the Isthmus of Panama with a nuclear intron separately. Each ML search consisted of 10 slow build up of diversity in South America, or is New replicates from random seeds, using the optimal partitioning

Journal of Biogeography 3 ª 2015 John Wiley & Sons Ltd P. A. Hosner et al.

(a) (d)

(b)

(c)

Figure 1 Previous hypotheses of New World quail evolutionary relationships based on morphology, plumage, biogeography and molecular markers. (a) Johnsgard’s (1988) phyletic hypothesis of odontophorid genera, (b) Johnsgard’s (1988) Odontophorus plumage groups, (c) Zink & Blackwell’s (1998) mitochondrial study, and (d) Jetz et al.’s (2012) constrained supermatrix/supertree. Taxa represented in this study are identified by *, taxa represented by molecular data in Jetz et al. (2012) are identified by #. Red branches identify Holman’s (1961) Dendrortyx group; blue branches identify Holman’s (1961) Odontophorus group. strategy and models of sequence evolution identified by Par- implemented two replicate 10 million generation Markov titionFinder. Bootstrap replicate searches were conducted chain Monte Carlo (MCMC) runs from random seeds, sam- with a similar approach, except they employed a single pling every 10,000 generations. This produced 2000 trees search for each replicate. We did not collapse near zero- sampled from the posterior distribution; we discarded the length branches to polytomies, because fully bifurcating gene first 25% of samples of each run as burn-in. For concate- trees were needed for downstream analyses using multispe- nated datasets, we increased MCMC chain length to 100 mil- cies coalescent approaches. lion generations and sampled every 100,000 generations, Bayesian phylogenetic analyses of each dataset were producing 2000 trees sampled from the posterior distribu- inferred with beast 2.1 (Bouckaert et al., 2014). We parti- tion; we discarded the first 25% of samples of each run as tioned according to locus and codon position (mtDNA), burn-in. To assess convergence of independent MCMC runs selected sequence models using PartitionFinder, selected we visually examined MCMC traces and parameter estimates, an uncorrelated lognormal relaxed clock (the standard devia- and determined that effective samples sizes (ESS) for each tion of the lognormal distribution included zero in prelimin- parameter estimate were > 200 in Tracer 1.5 (Rambaut & ary runs, justifying use), and selected a birth–death process Drummond, 2007). We examined topological convergence tree prior for each analysis. For separate loci, we with the program awty (Nylander et al., 2008).

4 Journal of Biogeography ª 2015 John Wiley & Sons Ltd Phylogeny of New World quail

In addition to phylogenetic inference of concatenated sympatric speciation, only speciation within the defined matrices, which could produce positively misleading results area). In ‘vicariant’ events, the ancestor range is divided under some circumstances, we inferred odontophorid phy- between daughter species without overlap. DIVA-LIKE is logeny using gene tree reconciliation analyses consistent similar to DEC, but it does not allow ‘sympatric’ modes of under the multispecies coalescent. Concatenated mtDNA was inheritance. BAYAREA-LIKE is also similar to DEC, but it treated as a single non-recombining locus. We implemented does not allow ‘vicariant’ modes of inheritance. BioGeo- two multispecies coalescent methods: (1) NJst (Liu & Yu, BEARS allows each model to be modified by adding a foun- 2011) via the STRAW server (Shaw et al., 2013), and (2) der-event speciation parameter (+j, or ‘jump’), which models ASTRAL (Mirarab et al., 2014). We used two tree sets with one daughter species inheriting the entire range of the ances- each method: the ML bootstrap treeset (500 bootstraps tor, and the second daughter dispersing to a new geographi- inferred for each of nine loci), and the Bayesian posterior cal area (Matzke, 2014). Fitting all six models permits testing treeset (500 trees randomly sampled from the post-burn-in assumptions of different biogeographical models in a statisti- posterior trees for each of nine loci). cal framework using the Akaike information criterion (AIC). We defined the following geographical regions as areas: (A) Old World, (B) Nearctic (including Baja California and Biogeographical reconstructions the Sonoran/Chihuahuan deserts, (C) Central America south To explore timing of odontophorid biogeography and diver- to Central Panama, (D) South America west of the Andes sification in relation to Earth and climate history, we used (including the Darien region of Panama/Colombia and the fossils to calibrate trees inferred from the full-concatenated Choco region of Colombia/Ecuador), and (E) South America matrix. Fossil elements attributed to New World quail date east of the Andes. We limited the maximum number of areas from the Oligocene and Miocene; however, these older fossils inferred at a node to three (the maximum number inhabited are not suitable for calibration of molecular clocks because by extant odontophorid species). they are fragments of single bones and are not placed phylo- genetically within the crown Odontophoridae (Ksepka, RESULTS 2009). Two more recent fossils are ascribed to extant genera, making it possible to place them phylogenetically: Colinus Sequence characteristics hibbardi from the late Pliocene of Kansas is described from 32 elements, and Callipepla shotwelli from the middle Plio- The concatenated DNA sequence alignment included cene of Oregon is described from 4 elements (Holman, 7055 bp, 3388 variable sites and 1942 parsimony informative 1961). In concatenated beast runs, we used Callipepla shotw- characters. Three mitochondrial genes yielded 247–360 infor- elli, which supersedes Colinus hibbardi, to calibrate the mini- mative sites each, totalling 872 informative sites. Eight mum age of the Colinus/Callipepla ancestor: a uniform prior nuclear introns yielded 75–328 informative sites each, total- of 3.5–65 Ma. This calibration explicitly assumes that Callip- ling 1070 informative sites (see Table S2 in Appendix S1). epla shotwelli is a stem lineage of extant Callipepla. It differs PartitionFinder identified five partitions as the optimal in four characters shared by extant Callipepla, justifying this strategy: (1) 12S, ND2 1st codon positions and ND5 1st assumption. We also placed a secondary calibration (normal codon positions; (2) CLTC, CLTCL1, FBG-5, MUSK, SER- prior; mean 27 Ma, sigma 1) on the most recent common PINB14 and TGFb2-5; (3) EFF2 and RHO; (4) ND2 and ancestor (MRCA) of Meleagris and Coturnix to match diver- ND5 2nd codon positions; and (5) ND2 and ND5 3rd codon gence timing inferred by (Jarvis et al., 2014) based on 19 fos- positions (Table S2). PartitionFinder identified GTR+I+G, sil calibrations from throughout the avian tree of life. HKY+G, HKY+G, HKY+I+G and GTR+I+G as the optimal To reconstruct ancestral areas, we fitted six probabilistic models of DNA sequence evolution, respectively (Table S2). biogeographical models on the concatenated beast maxi- mum clade credibility tree using BioGeoBEARS (Matzke, Concatenated analyses and gene trees 2014). BioGeoBEARS implements likelihood versions of the biogeographical models DEC (dispersal–extinction–cladogen- Maximum likelihood and Bayesian analysis of the concate- esis; Clark et al., 2008; Ree & Sanmartın, 2009), DIVA (dis- nated dataset strongly supported odontophorid monophyly persal–vicariance analysis; Ronquist, 1997; Yu et al., 2010), (Fig. 2, Figs S1–2 in Appendix S2). Ptilopachus was sister to and BAYAREA (Landis et al., 2013). Each model allows geo- all New World quail, and Rhynchortyx was sister to all other graphical range to evolve through specific types of events New World quail with strong support. The remaining genera (Ronquist & Sanmartın, 2011; Matzke, 2014). DEC-LIKE formed two clades: one comprising Cyrtonyx, Dactylortyx and allows two anagenetic events: ‘dispersal’ as range expansion, Odontophorus, and one comprising Oreortyx, Dendrortyx, and ‘extinction’ as range contraction. DEC-LIKE also allows Colinus, Philortyx and Callipepla. Monophyly of all genera two types of ‘cladogenesis’ events, where each daughter spe- was strongly supported (Fig. 2), as were relationships within cies inherits the range of an ancestor. In ‘sympatric’ events genera with the exception of Odontophorus. each daughter species inherits the entire, or partial but over- Analyses of single loci (see Figs S3–S11 in Appendix S2) lapping distribution from its ancestor (this does not imply inferred moderately to strongly supported topologies similar

Journal of Biogeography 5 ª 2015 John Wiley & Sons Ltd P. A. Hosner et al.

Figure 2 Concatenated maximum likelihood (ML) phylogeny of New World quail, inferred from 11 gene alignments. Support for each node is the ML bootstrap percentage (Garli) and the Bayesian posterior probability (beast). to the concatenated tree, with the exception of nodes per- Oligocene), and the crown of New World quail at 15–21 Ma taining to Callipepla/Colinus species, and relationships within (early Oligocene to early Miocene). New World quail genera Odontophorus. Single loci often lacked power to resolve rela- diverged largely in the Miocene, with most divergences tionships within these groups, and in a few cases topology between recognized species occurring in the Pliocene and differed from the concatenated tree with moderate or strong Pleistocene (Fig. 4). Given the calibrations, the substitution support (e.g. placement of O. guttatus in the mtDNA tree; rate of mitochondrial ND2 was estimated at 1.15% (cor- Fig. S3). rected) per lineage per million years, with a 95% highest posterior density of 0.88–1.41%. The posterior distribution of the Colinus/Callipepla ancestor was not bound by the Multispecies coalescent analyses upper limit on the uniform prior (65 Ma), justifying its use. Overall, topology and statistical support inferred with multi- Biogeographical model selection (Table 1, Figs S1–6in species coalescent methods (NJst and ASTRAL) were similar Appendix S2) indicated that DEC-LIKE+j was the best-fit to one another (Fig. 3b), and neither analysis method nor model, with 89% of the relative weight. This model sup- tree source consistently scored higher or lower support ported an Old World odontophorid ancestor and a Central across inferred trees. Additionally, results with coalescent American ancestor of New World quail (Fig. 4, Fig. S7 in approaches were similar to those inferred from the concate- Appendix S3). Temperate North America and South America nated matrix (Fig. 3a), with the exception of relationships were each colonized separately by three genera (Oreortyx, within Odontophorus – multispecies coalescent methods Callipepla, Colinus; Rhynchortyx, Odontophorus, Colinus, reconstructed O. capueira sister to O. stellatus, and O. ballivi- respectively). The DEC-LIKE+j model supports a Central ani sister to O. leucolaemus + O. speciosus, albeit with moder- American Odontophorus ancestor, with at least three indepen- ate support (Fig. 3). dent colonization events into South America. Speciation events within exclusively South American Odontophorus were estimated at c. 3.6 and c. 3.0 Ma. Other biogeographical Timing and geography of New World quail models receiving substantial relative weight were the DIVA- diversification LIKE+j model (8%) and the BAYAREA+j model (3%); over- Our estimates date the divergence of Ptilopachus and New all results of these models (Figs S1–S12 in Appendix S3) were World quail between 27 and 38 Ma (mid-Eocene to mid- similar to the DEC-LIKE+j model.

6 Journal of Biogeography ª 2015 John Wiley & Sons Ltd Phylogeny of New World quail

(a) (b)

Figure 3 Phylogenetic reconstructions of New World quail comparing (a) concatenated and (b) coalescent approaches. Concatenated analyses included maximum likelihood (ML) bootstrapping (above branches) and Bayesian posterior probabilities (below branches). Coalescent approaches included NJst and ASTRAL. Values above branches are inferred from ML bootstrapping; values below branches are inferred from Bayesian posterior tree sets. Taxa for which concatenated and coalescent topologies differ are highlighted in grey.

because of its small body size and heavy bill; our phyloge- DISCUSSION netic results suggest that this unique morphology is explained in part by unique ancestry. Phylogeny and gene tree discordance Within each clade, topology differed considerably from Our sequence data allowed us to infer a robust phylogenetic Johnsgard’s (1988) phyletic hypothesis among genera hypothesis for New World quail that differs considerably (Fig. 1). Within the Dendrortyx group, we found Philortyx from previous assessments of evolutionary relationships to be sister to Callipepla and Colinus. For overlapping taxa, (Holman, 1961; Johnsgard, 1988; Jetz et al., 2012). We phylogenetic results within the Dendrortyx clade were com- identified two major clades of New World quail: a parable to those of Zink & Blackwell (1998). We found ‘Dendrortyx clade’ (Dendrortyx/Oreortyx/Colinus/Philortyx/ Callipepla gambelii and C. californica to be sister taxa, and Callipepla) and an ‘Odontophorus clade’ (Dactylortyx/Cyrton- our concatenated analyses support that C. squamata and C. yx/Odontophorus). These clades corresponded to Holman’s douglasii are sister taxa. However, coalescent analyses of (1961) morphological groups with one exception – we found the ML bootstrap treeset received reduced support for C. Rhynchortyx to be sister to all other New World quail rather squamata + C. douglasii, and coalescent analyses of the than within the Odontophorus clade. Holman (1961) did Bayesian treeset recovered an alternative topology, with C. note that Rhynchortyx possessed unique morphology when douglasii sister to all other Callipepla, albeit without strong compared with other members of his Odontophorus group support.

Journal of Biogeography 7 ª 2015 John Wiley & Sons Ltd P. A. Hosner et al.

Figure 4 Estimates of divergence dates in Odontophoridae, with simplified timing estimates of Earth history events that favour odontophorid range expansion: dry land connections across Beringia, the North Atlantic and the Isthmus of Panama; and globally warm temperatures that would facilitate occurrence across high latitudes (highlighted in medium grey). Maximum likelihood ancestral range reconstructions under DEC+j are indicated at each node. Each biogeographical area defined is indicated on the map, along with present-day odontophorid distribution (inset).

In general, relationships within Odontophorus were poorly little support for Johnsgard’s (1988) species groups based on resolved, and placements of key taxa differed between con- plumage characters (Fig. 1). Concatenated results support catenated and coalescent approaches. Regardless, we found sister relationships between widely disjunct taxa: (1) between

8 Journal of Biogeography ª 2015 John Wiley & Sons Ltd Phylogeny of New World quail

Table 1 Biogeographical model selection in BioGeoBears for Odontophoridae. Models included dispersal–extinction–cladogenesis-like (DEC-LIKE), dispersal–vicariance analysis-like (DIVA-LIKE) and BAYAREA-LIKE, each with and without founder-event speciation (+j).

Biogeographical model Ln Parameters dej AIC Delta AIC Weight Relative weight (%)

DEC+j À58.06 3 0.0042 0.0000 0.0487 122.1 0.00 1.00E+00 88.5 DIVA-LIKE+j À60.37 3 0.0047 0.0000 0.0473 126.8 4.70 9.54E-02 8.4 BAYAREA-LIKE+j À61.31 3 0.0039 0.0000 0.0531 128.6 6.50 3.88E-02 3.4 DEC À68.74 2 0.0081 0.0057 – 140.9 18.80 8.27E-05 0.01 DIVA-LIKE À69.21 2 0.0093 0.0000 – 142.4 20.30 3.91E-05 0.00 BAYAREA-LIKE À76.53 2 0.0081 0.0526 – 157.1 35.00 2.51E-08 0.00

Ln, log-likelihood; parameters, number of parameters in each model; d, rate of dispersal; e, rate of extinction; j, rate of founder event speciation at cladogenesis; AIC, Akaike’s information criterion; weight, model weight calculated from AIC; relative weight, model relative weight calculated from AIC. northern Central American O. guttatus and southern Andean bal cooling in the Oligocene, which may have served to frag- O. balliviani, and (2) between Darıen/Choco O. erythrops ment the distribution of the Ptilopachus/New World quail and Atlantic Forest O. capueira (Figs 2 & 3). Coalescent ancestor following New World colonization. Otherwise, glo- approaches supported grouping without wide range disjunc- bal temperatures during much of the late Eocene and early tions: (1) O. guttatus sister to all other Odontophorus, (2) O. to mid-Miocene were much warmer than in the present day balliviani sister to the montane species pair of O. speciosus/ (Zachos, 2001), such that environmental conditions across O. leucolaemus, and (3) O. capueira sister to Amazonian O. Beringia were suitable. stellatus (Fig. 3). Failure to unambiguously resolve relation- Under the Laurasian colonization/global cooling vicariance ships in Odontophorus was possibly because of gene tree dis- model, odontophorids would have once been widespread in cordance, or simply because of the challenge of phylogenetic the Old World, yet present-day Old World odontophorids inference in rapidly diversifying lineages. Genomic sampling (Ptilopachus) are limited to Africa. Biogeographical recon- of hundreds or even thousands of loci, using coalescent structions support the idea that the two species of Ptilopa- methods that can utilize phylogenomic (Mirarab et al., 2014) chus are surviving members of a previously widespread or modified concatenation methods (Sun et al., 2014), may ancestral lineage, and their current distribution is the result be needed to resolve such problems with confidence. of range contraction. Unlike the New World, there are many Our multilocus phylogenetic hypothesis (Figs 2–4) con- lineages of small phasianids in the Old World tropics (e.g. tradicted all intergeneric relationships reported by Jetz et al. Arborophila, Coturnix, Perdicula, Pternistis, Rollulus), which (2012), with the exception of Callipepla + Colinus (Zink & may have limited odontophorid distribution and diversifica- Blackwell, 1998). This disparity in results is a reminder that tion there (Rabosky, 2009). Indeed, poorly known Ptilopa- the lack of primary sequence data is still a major limitation chus nahani is so superficially similar to various Pernistis in reconstructing the avian tree of life, and that genetic sam- species that it was misplaced in Phasianidae until recently pling is still sparse in many groups, both in breadth (taxon (Cohen et al., 2012). sampling) and depth (marker sampling). We encourage cau- Some palaeontologists have considered Palaeortyx (Oligo- tion when interpreting large-scale supermatricies, supertrees cene of Europe) to be an odontophorid based on skeletal and large-scale macroevolutionary studies based on them, characters, but this taxon lacks the distinctive bill structure when underlying sampling is insufficient. of extant New World quail (Mayr et al., 2006). Given the realization that Ptilopachus is an odontophorid, but also lacks the bill structure of New World quail, a robust morphologi- Biogeography and timing of New World quail cal phylogenetic analysis including Palaeortyx and numerous diversification extant galliforms is needed. Results from this analysis could Divergence dates and biogeographical reconstructions of clarify the timing and geography of odontophorid diversifica- New World quail (Fig. 4) supported an odontophorid colo- tion in the New World, and shed light on the notion of nization scenario free of long-distance dispersal across mar- odontophorid range contraction in the Old World. ine barriers. The best-fit model, DEC-LIKE+j, reconstructed At more recent time-scales, biogeographical reconstruc- an Old World odontophorid ancestor. Ptilopachus diverged tions (DEC+j) inferred a Central American Odontophorus from New World quail in the late Eocene (c. 36 Ma), well ancestor at 5.8 Ma, supporting the hypothesis that after the interruption of the Thulean Route (c. 55 Ma; Bri- Odontophorus colonized South America multiple times kiatis, 2014), rejecting the hypothesis that odontophorids following closure of the Isthmus of Panama at 3–4 Ma. New colonized the New World via the North Atlantic. The timing World quail diversity in South America is the direct result of of Ptilopachus/New World quail divergence is coincident with colonization from Central America, or recent in situ diversi- Beringia uniting Asia and North America (Hopkins, 1967; fication following isthmus closure. However, because of low Marincovich & Gladenkov, 1999), and with a period of glo- statistical support (both in topology and ancestral range

Journal of Biogeography 9 ª 2015 John Wiley & Sons Ltd P. A. Hosner et al.

Table 2 Likelihood of ancestral ranges at select nodes employing the best-fit biogeographical model (DEC+j). Ancestral ranges are (A) Old World, (B) Nearctic, (C) Central America, (D) trans-Andes, and (E) cis-Andes. node A B C D E AC BC CD CE CDE

Odontophorid MRCA 67% – 6% 1% – 15% –––– New World quail MRCA ––64% 3% –– 4% 26% –– Odontophorus group MRCA ––97% –––– 1% 2% – Dendrortyx group MRCA – 7% 75% –– – 17% ––– Colinus, Callipepla, Philortyx MRCA – 1% 80% –– – 15% 1% 1% – Colinus MRCA – 2% 48% –– – 24% 10% 10% – Odontophorus MRCA ––39% – 10% –– 5% 32% 12% O. gujanensis clade ––7% 1% 19% –– 3% 38% 28% O. guttatus + O. balliviani ––62% – 34% ––– 5% – O. speciosus + O. leucolaemus ––60% – 33% ––– 8% –

MRCA, most recent common ancestor. estimation) and missing taxa, some caution is needed before ACKNOWLEDGEMENTS accepting these seemingly straightforward and plausible con- clusions. According to biogeographical reconstructions, a We thank collectors and museums that loaned tissues used Central American Odontophorus ancestor was most probable in this study: Florida Museum of Natural History (FLMNH), (39%), but a Central American + cis-Andes ancestor (32%), Field Museum of Natural History (FMNH), Louisiana State a Central American + trans-Andes + cis-Andes ancestor Museum of Natural Science (LSUMZ), University of Kansas (12%), or a solely South American Odontophorus ancestor Biodiversity Institute (KU), Universidad Nacional Autonoma (15%) were also possible (Table 2). In addition, six of seven de Mexico (UNAM) and University of Washington Burke Odontophorus species lacking in our analysis are South Museum (UWBM). Sarah Hyde and Michael Kusniak American, and their inclusion may alter ancestral area recon- assisted in sequence data collection. Funding was provided structions. Regardless of the specifics of Odontophorus colo- by the National Science Foundation (DEB-1118823) to nization, phylogeny broadly supports the hypothesis that R.T.K. and E.L.B. New World quail are another example of birds of Laurasian origin that diversified in Central America before colonizing REFERENCES South America as part of the Great American Biotic Inter- change (Dacosta & Klicka, 2008; Weir et al., 2009; Smith & Barker, F.K., Burns, K.J., Klicka, J., Lanyon, S.M. & Lovette, Klicka, 2010; Barker et al., 2015). I.J. (2015) New insights into New World biogeography: an Our results reinforce the idea that odontophorids lack the integrated view from the phylogeny of blackbirds, cardi- capacity to cross marine barriers. Then, how did endemic nals, sparrows, tanagers, warblers, and allies. The Auk, 132, subspecies of Colinus virginianus (Cuba) and Callipepla cali- 333–348. fornica (Santa Catalina, off California) arrive on oceanic Bibi, F. (2011) Mio-pliocene faunal exchanges and African islands? Authors have questioned the natural occurrence of biogeography: the record of fossil bovids. PLoS ONE, 6, these taxa on islands, yet they remain valid taxa in current e16688. taxonomic treatments (e.g. Gill & Donsker, 2015). Confusion Billerman, S.M., Huber, G.H., Winkler, D.W., Safran, R.J. & surrounds the validity of Colinus v. cubanensis because Coli- Lovette, I.J. (2011) Population genetics of a recent trans- nus v. floridanus was widely introduced for hunting through- continental colonization of South America by breeding out the Greater Antilles, including Cuba, as early as the 18th Barn Swallows (Hirundo rustica). The Auk, 128, 506–513. century (Gundlach, 1876). Mitochondrial DNA fragments Bouckaert, R., Heled, J., Kuhnert,€ D., Vaughan, T., Wu, amplified from Cuban Colinus specimens are not genetically C.-H., Xie, D., Suchard, M.A., Rambaut, A. & Drummond, distinct, as expected of a natural-occurring island population A.J. (2014) BEAST 2: a software platform for Bayesian (Williford, 2013). Furthermore, well-studied late Pleistocene evolutionary analysis. PLoS Computational Biology, 10, sites in Cuba lack Colinus (Olson, 1978), whereas Colinus e1003537. remains are common at sites in peninsular Florida (Holman, Bowie, R.C.K., Cohen, C. & Crowe, T.M. (2013) Ptilopachi- 1961). Similarly, studies of Callipepla c. catalinensis failed to nae: a new subfamily of the Odontophoridae (Aves: Galli- find evidence of genetic differentiation (Zink et al., 1987; formes). Zootaxa, 3670, 097–098. Williford, 2013), and thorough archaeological studies of Brikiatis, L. (2014) The De Geer, Thulean and Beringia middens on Santa Catalina failed to identify Callipepla routes: key concepts for understanding early Cenozoic bio- amongst thousands of subfossil elements (Porcasi, 1999). In geography. Journal of Biogeography, 41, 1036–1054. the light of genetic and fossil/subfossil evidence, we conclude Burney, C.W. & Brumfield, R.T. (2009) Ecology predicts lev- that natural occurrence of odontophorids is limited to conti- els of genetic differentiation in Neotropical birds. The nents and continental islands. American Naturalist, 174, 358–368.

10 Journal of Biogeography ª 2015 John Wiley & Sons Ltd Phylogeny of New World quail

Clark, J., Ree, R., Alfaro, M., King, M., Wagner, W. & Roal- Holman, J.A. (1961) Osteology of living and fossil New son, E. (2008) A comparative study in ancestral range World quails (Aves: Galliformes). Bulletin of the Florida reconstruction methods: retracing the uncertain histories State Museum, 6, 131–233. of insular lineages. Systematic Biology, 57, 693–707. Hopkins, D.M. (1967) The Bering Land Bridge. Stanford Clegg, S.M., Degnan, S.M., Moritz, C., Estoup, A., Kikkawa, University Press, Palo Alto, CA. J. & Owens, I. (2002) Microevolution in island forms: the Hosner, P.A., Sheldon, F.H., Lim, H.C. & Moyle, R.G. roles of drift and directional selection in morphological (2010) Phylogeny and biogeography of the Asian trogons divergence of a passerine . Evolution, 56, 2090–2099. (Aves: Trogoniformes) inferred from nuclear and mito- Coates, A.G., Collins, L.S., Aubry, M.-P. & Berggren, W.A. chondrial DNA sequences. Molecular Phylogenetics and (2004) The geology of the Darien, Panama, and the late Evolution, 57, 1219–1225. Miocene–Pliocene collision of the Panama arc with north- Jarvis, E.D., Mirarab, S., Aberer, A.J., Li, B., Houde, P. & Li, western South America. Geological Society of America Bul- C. (2014) Whole-genome analyses resolve early branches letin, 116, 1327–1344. in the tree of life of modern birds. Science, 346, 1320– Cody, M.L. & Overton, J.M. (1996) Short-term evolution of 1331. reduced dispersal in island plant populations. Journal of Jetz, W., Thomas, G.H., Joy, J.B., Hartmann, K. & Mooers, Ecology, 84,53–61. A.O. (2012) The global diversity of birds in space and Cohen, C., Wakeling, J.L., Mandiwana Neudani, T.G., time. Nature, 491, 444–448. Sande, E., Dranzoa, C., Crowe, T.M. & Bowie, R.C. Johnsgard, P.A. (1988) The quails, partridges, and francolins (2012) Phylogenetic affinities of evolutionarily enigmatic of the world. Oxford University Press, New York. African galliforms: the Stone Partridge Ptilopachus petro- Katoh, K. & Standley, D.M. (2013) MAFFT multiple sus and Nahan’s Francolin Francolinus nahani, and sup- sequence alignment software version 7: improvements in port for their sister relationship with New World quails. performance and usability. Molecular Biology and Evolu- Ibis, 154, 768–780. tion, 30, 772–780. Cox, W.A., Kimball, R.T. & Braun, E.L. (2007) Phylogenetic Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, position of the New World quail (Odontophoridae): eight M., Sturrock, S., Buxton, S., Cooper, A., Markowitz, S., nuclear loci and three mitochondrial regions contradict Duran, C., Thierer, T., Ashton, B., Meintjes, P. & Drum- morphology and the Sibley-Ahlquist tapestry. The Auk, mond, A. (2012) Geneious Basic: an integrated and extend- 124,71–84. able desktop software platform for the organization and Cracraft, J. (2001) Avian evolution, Gondwana biogeography analysis of sequence data. Bioinformatics, 28, 1647–1649. and the Cretaceous–Tertiary mass extinction event. Pro- Kimball, R.T. & Braun, E.L. (2014) Does more sequence data ceedings of the Royal Society B: Biological Sciences, 268, improve estimates of galliform phylogeny? Analyses of a 459–469. rapid radiation using a complete data matrix. PeerJ, 2, Crowe, T., Bowie, R., Bloomer, P., Mandiwana, T., Hedder- e361. son, T., Randi, E., Pereira, S. & Wakeling, J. (2006) Phy- Kimball, R.T., Braun, E.L., Barker, F.K., Bowie, R.C.K., logenetics, biogeography and classification of, and Braun, M.J., Chojnowski, J.L., Hackett, S.J., Han, K.-L., character evolution in, gamebirds (Aves: Galliformes): Harshman, J., Heimer-Torres, V., Holznagel, W., Huddle- effects of character exclusion, data partitioning and miss- ston, C.J., Marks, B.D., Miglia, K.J., Moore, W.S., Reddy, ing data. Cladistics, 22, 495–532. S., Sheldon, F.H., Smith, J.V., Witt, C.C. & Yuri, T. (2009) Dacosta, J.M. & Klicka, J. (2008) The Great American Inter- A well-tested set of primers to amplify regions spread change in birds: a phylogenetic perspective with the genus across the avian genome. Molecular Phylogenetics and Evo- Trogon. Molecular Ecology, 17, 1328–1343. lution, 50, 654–660. Denk, T., Grımsson, F., Zetter, R. & Sımonarson, L.A. (2011) Kirchman, J.J. (2012) Speciation of flightless rails on islands: The biogeographic history of Iceland: the North Atlantic a DNA-based phylogeny of the typical rails of the Pacific. land bridge revisited. Topics in Geobiology, 35, 647–668. The Auk, 129,56–69. Springer Netherlands, Dordrecht. Ksepka, D.T. (2009) Broken gears in the avian molecular Fulton, T.L., Letts, B. & Shapiro, B. (2012) Multiple losses of clock: new phylogenetic analyses support stem galliform flight and recent speciation in steamer ducks. Proceedings status for Gallinuloides wyomingensis and rallid affinities of the Royal Society B: Biological Sciences, 279, 2339–2346. for Amitabha urbsinterdictensis. Cladistics, 25, 173–197. Gill, F. & Donsker, D. (2015) IOC world bird names (version Ksepka, D.T. & Clarke, J.A. (2010) New fossil mousebird 4.0). Available at: http://www.worldbirdnames.org. (Aves: Coliiformes) with feather preservation provides Gundlach, J.C. (1876) Contribucion a la ornitologia cubana. insight into the ecological diversity of an Eocene North Imp. ‘La Antilla’, Habana, Cuba. American avifauna. Zoological Journal of the Linnean Soci- Harris, R.B., Birks, S.M. & Leache, A.D. (2014) Incubator birds: ety, 160, 685–706. biogeographical origins and evolution of underground Landis, M.J., Matzke, N.J., Moore, B.R. & Huelsenbeck, J.P. nesting in megapodes (Galliformes: Megapodiidae). Journal (2013) Bayesian analysis of biogeography when the num- of Biogeography, 41, 2045–2056. ber of areas is large. Systematic Biology, 62, 789–804.

Journal of Biogeography 11 ª 2015 John Wiley & Sons Ltd P. A. Hosner et al.

Lanfear, R., Calcott, B., Ho, S.Y.W. & Guindon, S. (2012) Par- expansion of a “great speciator”. Proceedings of the titionFinder: combined selection of partitioning schemes National Academy of Sciences USA, 106, 1863–1868. and substitution models for phylogenetic analyses. Molecu- Moyle, R.G., Andersen, M.J., Oliveros, C.H., Steinheimer, lar Biology and Evolution, 29, 1695–1701. F.D. & Reddy, S. (2012) Phylogeny and biogeography of Leigh, E.G., O’Dea, A. & Vermeij, G.J. (2013) Historical bio- the core babblers (Aves: Timaliidae). Systematic Biology, geography of the Isthmus of Panama. Biological Reviews, 61, 631–651. 89, 148–172. Nelson, G. (1978) From Candolle to Croizat: comments on Liu, L. & Yu, L. (2011) Estimating species trees from unroot- the history of biogeography. Journal of the History of Biol- ed gene trees. Systematic Biology, 60, 661–667. ogy, 11, 269–305. Livezey, B.C. (1989) Flightlessness in grebes (Aves, Podici- Nesbitt, S.J., Ksepka, D.T. & Clarke, J.A. (2011) Podargiform pedidae): its independent evolution in three genera. Evolu- affinities of the enigmatic Fluvioviridavis platyrhamphus tion, 43,29–54. and the early diversification of Strisores (“Caprimulgifor- Marincovich, L. & Gladenkov, A.Y. (1999) Evidence for mes” + Apodiformes). PLoS ONE, 6, e26350. an early opening of the Bering Strait. Nature, 397, 149– Nylander, J.A.A., Wilgenbusch, J.C., Warren, D.L. & Swof- 151. ford, D.L. (2008) AWTY (are we there yet?): a system for Matzke, N.J. (2014) Model selection in historical biogeogra- graphical exploration of MCMC convergence in Bayesian phy reveals that founder-event speciation is a crucial pro- phylogenetics. Bioinformatics, 24, 581–583. cess in island clades. Systematic Biology, 61, 951–970. Pereira, S.L. & Baker, A.J. (2006) A molecular timescale for Mayr, G. (2004) Old world fossil record of modern-type galliform birds accounting for uncertainty in time esti- hummingbirds. Science, 304, 861–864. mates and heterogeneity of rates of DNA substitutions Mayr, G., Poschmann, M. & Wuttke, M. (2006) A nearly across lineages and sites. Molecular Phylogenetics and Evo- complete skeleton of the fossil galliform bird Palaeortyx lution, 38, 499–509. from the late Oligocene of Germany. Acta Ornithologica, Porcasi, J.F. (1999) Prehistoric bird remains from the south- 41, 129–135. ern Channel Islands. Pacific Coast Archaeological Society Meiklejohn, K.A., Danielson, M.J., Faircloth, B.C., Glenn, Quarterly, 35,38–59. T.C., Braun, E.L. & Kimball, R.T. (2014) Incongruence de Queiroz, A. (2005) The resurrection of oceanic dispersal among different mitochondrial regions: a case study using in historical biogeography. Trends in Ecology and Evolu- complete mitogenomes. Molecular Phylogenetics and Evolu- tion, 20,68–73. tion, 78, 314–323. Rabosky, D.L. (2009) Ecological limits and diversification Miller, K.G. (2005) The Phanerozoic record of global sea- rate: alternative paradigms to explain the variation in spe- level change. Science, 310, 1293–1298. cies richness among clades and regions. Ecology Letters, 12, Mirarab, S., Reaz, R., Bayzid, M.S., Zimmermann, T., Swen- 735–743. son, M.S. & Warnow, T. (2014) ASTRAL: genome-scale Rambaut, A. & Drummond, A.J. (2007) Tracer version 1.5. coalescent-based species tree estimation. Bioinformatics, 30, Available at: http://beast.bio.ed.ac.uk. i541–i548. Reddy, S. & Cracraft, J. (2007) Old World shrike-babblers Mitchell, K.J., Llamas, B., Soubrier, J., Rawlence, N.J., Wor- (Pteruthius) belong with New World vireos (Vireonidae). thy, T.H., Wood, J., Lee, M.S.Y. & Cooper, A. (2014) Molecular Phylogenetics and Evolution, 44, 1352–1357. Ancient DNA reveals elephant birds and kiwi are sister Ree, R.H. & Sanmartın, I. (2009) Prospects and challenges taxa and clarifies ratite bird evolution. Science, 344, for parametric models in historical biogeographical infer- 898–900. ence. Journal of Biogeography, 36, 1211–1220. Montes, C., Cardona, A., Jaramillo, C., Pardo, A., Silva, J.C., Ronquist, F. (1997) Dispersal–vicariance analysis: a new Valencia, V., Ayala, C., Perez-Angel, L.C., Rodrıguez-Parra, approach to the quantification of historical biogeography. L.A., Ramirez, V. & Nino,~ H. (2015) Middle Miocene clo- Systematic Biology, 46, 195–203. sure of the Central American Seaway. Science, 22,6–228. Ronquist, F. & Sanmartın, I. (2011) Phylogenetic methods in Moore, R.P., Robinson, W.D., Lovette, I.J. & Robinson, T.R. biogeography. Annual Review of Ecology and Systematics, (2008) Experimental evidence for extreme dispersal limita- 42, 441–464. tion in tropical forest birds. Ecology Letters, 11, 960–968. Sanmartın, I., Enghoff, H. & Ronquist, F. (2001) Patterns of Moyle, R.G. (2004) Phylogenetics of barbets (Aves: Picifor- dispersal, vicariance and diversification in the mes) based on nuclear and mitochondrial DNA sequence Holarctic. Biological Journal of the Linnean Society, 73, data. Molecular Phylogenetics and Evolution, 30, 187–200. 345–390. Moyle, R.G., Chesser, R.T., Prum, R.O., Schikler, P. & Cra- Shaw, T.I., Ruan, Z., Glenn, T.C. & Liu, L. (2013) STRAW: craft, J. (2006) Phylogeny and evolutionary history of Old Species TRee Analysis Web server. Nucleic Acids Research, World suboscine birds (Aves: Eurylaimides). American 41, W238–W241. Museum Novitates, 3544,1–22. Simpson, G.G. (1947) Holarctic mammalian faunas and Moyle, R.G., Filardi, C.E., Smith, C.E. & Diamond, J. (2009) continental relationships during the Cenozoic. Geological Explosive Pleistocene diversification and hemispheric Society of America Bulletin, 58, 613–687.

12 Journal of Biogeography ª 2015 John Wiley & Sons Ltd Phylogeny of New World quail

Smith, B.T. & Klicka, J. (2010) The profound influence of Zink, R.M., Lott, D.F. & Anderson, D.W. (1987) Genetic var- the Late Pliocene Panamanian uplift on the exchange, iation, population structure, and evolution of California diversification, and distribution of New World birds. Ecog- Quail. The Condor, 89, 395–405. raphy, 33, 333–342. Zwickl, D.J. (2006) GARLI: genetic algorithm for rapid likeli- Smith, J.V., Braun, E.L. & Kimball, R.T. (2013) Ratite non- hood inference. Available at: http://www. bio. utexas. edu/ monophyly: independent evidence from 40 novel loci. Sys- faculty/antisense/garli/Garli.html. tematic Biology, 62,35–49. Sun, K., Meiklejohn, K.A., Faircloth, B.C., Glenn, T.C., SUPPORTING INFORMATION Braun, E.L. & Kimball, R.T. (2014) The evolution of pea- fowl and other taxa with ocelli (eyespots): a phylogenomic Additional Supporting Information may be found in the approach. Proceedings of the Royal Society B: Biological Sci- online version of this article: ences, 281, 20140823–20140823. Appendix S1 Supplementary tables. Wang, N., Kimball, R.T., Braun, E.L., Liang, B. & Zhang, Z. Appendix S2 Supplementary tree figures. (2013) Assessing phylogenetic relationships among Galli- Appendix S3 Raw PDF outputs from BioGeoBEARS. formes: a multigene phylogeny with expanded taxon sam- pling in Phasianidae. PLoS ONE, 8, e64312. BIOSKETCHES Webb, D. (1976) Mammalian faunal dynamics of the great 2 – American interchange. Paleobiology, , 220 234. Peter Hosner is a postdoctoral associate at the University Weir, J.T., Bermingham, E. & Schluter, D. (2009) The Great of Florida. His research interests include biogeography, pop- American Biotic Interchange in birds. Proceedings of the ulation history, diversification, molecular systematics and 106 – National Academy of Sciences USA, , 21737 21742. conservation of tropical birds. Wiley, E.O. (1988) Vicariance biogeography. Annual Review of Ecology and Systematics, 19, 513–542. Edward Braun is an associate professor at the University of Williford, D.L. (2013) Molecular genetics of the Northern Bob- Florida. He is interested in computational biology, evolution- white, Scaled Quail, and Gambel’s Quail. Texas A&M ary genomics, and the theory and practice of phylogenetic Kingsville, TX. systematics, particularly as they relate to birds and reptiles. Yu, Y., Harris, A.J. & He, X. (2010) Molecular phylogenetics Rebecca Kimball is a professor at the University of Florida. 56 and evolution. Molecular Phylogenetics and Evolution, , She is interested in many aspects of avian evolution, includ- – 848 850. ing molecular phylogenetics, diversification, trait evolution, Zachos, J. (2001) Trends, rhythms, and aberrations in global population genetics and behaviour. climate 65 Ma to present. Science, 292, 686–693. Zink, R.M. & Blackwell, R.C. (1998) Molecular systematics of the scaled quail complex (genus Callipepla). The Auk, Editor: Michael Patten 115, 394–403.

Journal of Biogeography 13 ª 2015 John Wiley & Sons Ltd