Molecular Phylogenetics and Evolution 59 (2011) 675–684

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier .com/ locate /ympev

Molecular systematics of two enigmatic genera Psittacella and Pezoporusilluminate the ecological radiation of Australo-Papuan parrots (Aves: Psittaciformes)

Leo Joseph a,⇑, Alicia Toon a,b, Erin E. Schirtzinger c, Timothy F. Wright c

a Australian National Wildlife Collection, CSIRO Ecosystem Sciences, GPO Box 284, Canberra, ACT 2601, Australiab Australian Rivers Institute, Griffith University, Nathan, Qld 4111, Australiac Department of Biology, MSC 3AF, New Mexico State University, Las Cruces, NM 88003-0001, USA

a r t i c l e i n f o

Article history:Received 1 October 2010Revised 11 March 2011Accepted 15 March 2011Available online 29 March 2011

Keywords:AustraliaAustralo-PapuaNew GuineaParrotsPezoporusPlatycercinesPsittacellaSystematics

1055-7903/$ - see front matter Crown Copyright � 2doi:10.1016/j.ympev.2011.03.017

⇑ Corresponding author. Fax: +61 2 6242 1689.E-mail address: Leo.joseph@csiro.au (L. Joseph).

a b s t r a c t

The platycercine parrots of Australia, usually recognized as the Platycercinae or Platycercini, are thebroad-tailed parrots and their allies typified by the rosellas Platycercus spp. Debate concerning their cir-cumscription has most recently centerd on the position of four genera, Neophema, Neopsephotus, Pezop-orus and Psittacella, the last two having never been adequately included in sequence-based analyses.We use broad taxon sampling, mitochondrial and nuclear DNA sequence data from seven independentloci (two linked mitochondrial loci and six nuclear loci), and both gene tree and species tree approachesto reconstruct phylogenies and so determine the systematic placement all four genera. Analyses of twodata sets, one of 48 taxa and five loci and one of 27 taxa and the same five plus three additional lociproduced broadly congruent and consistently well-resolved phylogenies. We reject placement of any ofthese four genera within core platycercines. Pezoporus is closely allied to Neophema and Neopsephotus.These three genera are the likely sister group to core platycercines and we advocate their recognition asa subfamily. Psittacella is the sole extant representative of a lineage that branched very early in the his-tory of Australo-Papuan parrot fauna and is not closely related to any of the mostly south-east Asianand Indonesian psittaculine taxa with which it is more often linked. We present a revised view ofthe extraordinary phylogenetic, phenotypic and ecological diversity that is the adaptive radiation ofAustralo-Papuan parrots. Finally, our analyses highlight the likely paraphyly of Mayr’s (2008)Loricoloriinae.

Crown Copyright � 2011 Published by Elsevier Inc. All rights reserved.

1. Introduction

The parrots (Aves: Psittaciformes) are some of the world’s mostfamiliar and recognizable birds due to their popularity in captivity,their extraordinarily conserved morphology, and, increasingly,their widespread endangerment. Anatomical characters in livingand fossil parrots (osteology, myology, arterial structure) have longbeen used to address higher-order relationships (reviewed inBoles, 2002; Wright et al., 2008; Forshaw, 1989; Mayr, 2008,2010). More recently, analyses of allozymes and DNA sequenceshave been used to clarify these relationships (see Christidis et al.,1991; de Kloet and de Kloet, 2005; Miyaki et al., 1998; Schweizeret al., 2010; Tavares et al., 2004, 2006; Wright et al., 2008). Majordivisions among parrots, especially Neotropical species, are nowreasonably well understood (Eberhard and Bermingham, 2005;Ribas et al., 2005; Schweizer et al., 2010; Wright et al., 2008). Herewe focus on an Old World group with problematic limits, the platy-cercine or broad-tailed parrots of Australasia, familiarly known as

011 Published by Elsevier Inc. All r

the rosellas and their relatives. This group is usually recognizedas the Platycercinae within Psittacidae (e.g., Schodde, 1997) orPlatycercini within the Psittacinae (Forshaw, 2002, 2006).

There is general agreement that platycercines include theAustralian endemic genera found almost exclusively in temperateforests and semi-arid woodlands. We term these genera the coreplatycercines: Platycercus, Barnardius, Psephotus, Northiella,Purpureicephalus, and Lathamus as well as the three genera Cyan-oramphus, Eunymphicus and Prosopeia found in New Zealand andthe Pacific island regions (Boon et al., 2008; Christidis et al., 1991;Condon, 1941; Homberger, 1980; Joseph et al., 2008; Mayr, 2008;Ovenden et al., 1987; Schodde, 1997; Smith, 1975). Although someearly treatments of this group included the monotypic genus Melo-psittacus of the arid zone (the familiar budgerigar Melopsittacusundulatus), recent analyses of both DNA sequences and morphologyclearly show that it is more closely related to mostly mesic zonelorikeets and rainforest-inhabiting fig-parrots (de Kloet and deKloet, 2005; Mayr, 2008; Schweizer et al., 2010; Wright et al.,2008). There is lingering debate over the inclusion of four othergenera: Neophema (six species) Neopsephotus (one), Pezoporus(three; see Murphy et al., 2011) and Psittacella (four). Of these four

ights reserved.

676 L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684

genera, Neophema, Neopsephotus, and Pezoporus are found in tem-perate and arid Australia while Psittacella inhabits the tropicalrain-forests of New Guinea. Wright et al. (2008), based on se-quences from two mitochondrial loci and three nuclear introns,found robust support for Neophema and Neopsephotus being mostclosely related to Agapornis, Loriculus and Bolbopsittacus, whichrange from Africa to east of Wallace’s Line in New Guinea but notAustralia (see Wright et al., 2008). Conversely, Schweizer et al.(2010) analysed sequences from three nuclear exons and foundNeophema and Neopsephotus to be sister to all other platycercinesbut with relatively low support. From osteological evidence,Mayr (2010) also concluded that Neophema and Neopsephotus arenot closely related to core platycercine genera. Relationships ofPsittacella and Pezoporus, which have not been included in any pre-vious DNA sequence study, remain especially contentious and, withNeophema and Neopsephotus, are the primary foci of the presentstudy.

Resolution of the systematics of Pezoporus and Psittacella is ofparticular biogeographic interest. Pezoporus comprises three spe-cies with fragmented ranges across inland and subcoastal southernAustralia (Forshaw, 2002; McDougall et al., 2009; Murphy et al.,2011). Clarification of relationships within Pezoporus and identifi-cation of their closest relatives would inform how they have speci-ated on the Australian landscape and contribute to the broaderquestions of relationships among mesic and xeric biota in Austra-lia. Their long-presumed association with Melopsittacus and theplatycercines has been based on similarities in plumage patternand nestlings’ food-begging calls, traits that have not been assessedin using independently derived phylogenies (Christidis et al., 1991;Courtney, 1997). Leeton et al. (1994) aligned Pezoporus, Melopsitta-cus and Neophema with Platycercus based on partial mitochondrialDNA (mtDNA) sequences but that study’s sequences were laterwithdrawn from GenBank.

Systematics and biogeography of New Guinean Psittacella areeven more poorly understood. Its placement in the Psittaculinior alternatively, treatment as platycercine, has been based onmorphological traits not analysed in a phylogenetic context(Christidis et al., 1991; Forshaw, 1989; Schodde, 2006; Smith,1975). Christidis et al.’s (1991) distance-based analyses of allo-zyme data could not resolve relationships of Psittacella. Basedon similarities in a number of morphological characters, they con-sidered it platycercine but of ‘‘disparate affinity’’ to that group.Resolution of the systematic position of Psittacella, and the ques-tion of whether it is platycercine, could inform a range of evolu-tionary and ecological topics. If platycercine, it would be the onlymember of that group in New Guinea and indeed in mountainrainforest habitats generally above 1100 m altitude. It could ad-dress whether the relatively slender-bodied, long-tailed morphol-ogy of platycercines is derived from stouter, short-tailedPsittacella-like ancestors and test the hypothesis that arid andsemi-arid zone platycercines, which mostly feed terrestrially inAustralia have been derived from montane arboreal rainforestancestors in New Guinea (see Schodde, 1982; Schodde and Calab-y, 1972).

The primary aim of the present study is to clarify the limits ofthe platycercine parrots by including Pezoporus and Psittacella forthe first time in a multilocus DNA sequence-based data set withcomprehensive sampling of other platycercine parrots, includingNeophema and Neopsephotus. A secondary aim is to clarify relation-ships within Pezoporus by testing (1) whether Murphy et al.’s(2011) recognition of Pezoporus flaviventris as a sister species toPezoporus wallicus based on mtDNA and phenotypic data is sup-ported by nuclear data and (2) whether Pezoporus occidentalis isin turn their sister. Lastly, we discuss implications of our findingsto understanding the ecological history of parrots in the Austra-lo-Papuan region.

2. Materials and methods

2.1. Taxon sampling

Source tissues were frozen samples from vouchered museumspecimens or blood from birds held in zoos for some non Austra-lo-Papuan species. We sampled single individuals of 48 parrot taxarepresenting 31 Australasian genera and 7 genera from other re-gions (Table 1). Psittacella was represented by Psittacella brehmiiand Psittacella picta, Pezoporus by Pe. occidentalis, Pe. wallicus andPe. flaviventris. Core platycercine genera were Platycercus, Barnardi-us, Purpureicephalus, Northiella, Psephotus, Cyanoramphus, Eunym-phicus, Lathamus, and Prosopeia. Neophema and Neopsephotus alsowere included to investigate their relationship to Pezoporus andthe core platycercines. Other genera included as potential relativesof Psittacella (Forshaw, 2006; Smith, 1975) were Tanygnathus, Lori-culus, Agapornis, Bolbopsittacus, Prioniturus, Psittacula, Micropsitta,Geoffroyus, Eclectus, Alisterus, Aprosmictus and Polytelis. New Zea-land taxa Strigops and Nestor were used as outgroups (de Kloetand de Kloet, 2005; Schweizer et al., 2010; Wright et al., 2008).Nomenclature follows Forshaw (2006) except for recent species-le-vel recognition of Pe. flaviventris based on mtDNA and morphology(Murphy et al., 2011).

2.2. Character sampling

We isolated genomic DNA using DNeasy extraction kits follow-ing manufacturer’s protocols (Qiagen, Valencia CA). We sampled alltaxa at two linked mitochondrial protein-coding genes [cyto-chrome oxidase I (COI) and NADH dehydrogenase 2 (ND2)] andthree non-coding nuclear introns [tropomyosin alpha-subunit in-tron 5 (TROP), and transforming growth factor b-2 intron 5(TGFB2), rhodopsin intron 1 (RDPSN)]. In addition, we sampled27 of these taxa at three single-copy nuclear exons: the proto-oncogene c-mos (c-mos), recombination activating protein (RAG-1) and the transcription factor ZENK, a homolog of the early growthresponse-1 gene (ZENK). For the mtDNA loci and nuclear intronswe used primers given in Wright et al. (2008); primers for the nu-clear exons were from Schweizer et al. (2010). Amplification for allloci followed protocols in Wright et al. (2008) excepting sequencesfrom Schweizer et al. (2010) downloaded from GenBank.

All amplicons were sequenced in both directions at the Univer-sity of Chicago Cancer Sequencing Facility using an ABI 3730 DNAAnalyzer and ABI Big Dye chemistry. We did not obtain RDPSN se-quences for Barnardius barnardi, Barnardius zonarius, Coracopsisvasa, Eunymphicus uvaeensis, Pe. occidentalis, Platycercus elegans,Polytelis anthopeplus, Psephotus dissimilis, Psephotus haematonotus,Psephotus varius, Psittacella picta, and Psittacus erithacus; c-moswas not obtained for M. undulatus. Sequences were combined intoa single consensus sequence using Sequencher 4.7 (Gene Codes,Ann Arbor, MI). See Table 1 for GenBank accession numbers.

2.3. Phylogenetic analysis

We aligned the sequences for each gene region with Clustal W(Chenna et al., 2003) using the default parameters for gap openingand extension penalties with the exception of RDPSN where we useda gap opening penalty of 5 and an extension penalty of 2.5 to improvealignment of the terminal ends of the sequences. We combined theresulting alignments in PAUP� version 4.0b10 (Swofford, 1999) tocreate two datasets. The ‘Primary dataset’ included 48 taxa and fiveloci (COI, ND2, TROP, TGFB2, and RDPSN); the ‘Secondary dataset’ in-cluded 27 taxa and three additional loci (c-mos, Rag-1 and ZENK).

Previous analyses of parrot relationships employing the samefive loci as in the Primary dataset detected little conflict in

Table 1Species and loci sampled for study of evolutionary relationships of the Australo-Papuan parrot genera Pezoporus and Psittacella.

Taxon Museum/collectiona Specimen numbera Sample typea GenBank accession numbersc

COI ND2 TROP TGFB2 RDPSN C-MOS RAG-1 ZENK

Agapornis roseicollis NMNH B08798 T EU621593 EU327596 EU665562 EU660234 EU665501 GQ505086 GQ505194 GQ505141Alisterus amboinensis NMNH B06399 T EU621594 EU327597 EU665563 EU660235 EU665502 HQ316816 HQ316830 HQ316844Aprosmictus erythropterus NMNH B06424 T EU621596 EU327599 EU665565 EU660237 EU665504 HQ316817 HQ316831 HQ316845Barnardius barnardi ANSP 10702 T HQ316859 HQ316872 HQ316885 HQ316898Barnardius zonarius ANSP ANSP10669 T EU621599 EU327602 EU665568 EU660240Bolbopsittacus lunulatus NMNH B03677 T EU621600 EU327603 EU665569 EU660241 EU665507Calyptorhynchus banksii ANWC 50042 T HQ316860 HQ316873 HQ316886 HQ316899 HQ316809Calyptorhynchus funereus NMNH B06460 T EU621603 EU327606 EU665572 EU660244 EU665510 HQ316818 HQ316832 HQ316846Chalcopsitta duivenbodei NMNH B06396 T EU621604 EU327607 EU665573 EU660245 EU665511 HQ316819 HQ316833 HQ316847Coracopsis vasa LPF LPF07-29 (UV) B EU621608 EU327612 EU665578 EU660250 GQ505113 GQ505223 GQ505167Cyanoramphus auriceps VU FT3310 B EU621611 EU327615 EU665581 EU660252 EU665516 GQ505104 GQ505213 GQ505158Cyanoramphus novaezelandiae AMNH DOT 11060 T HQ316861 HQ316874 HQ316887 HQ316900 HQ316810Cyclopsitta diophthalma AMNH DOT 7799 T EU621612 EU327616 EU665582 EU660253 EU665517Eclectus roratus NMNH B06393 T EU621615 EU327619 EU665585 EU660256 EU665520 GQ505135 GQ505244 GQ505187Eolophus roseicapillus NMNH B06415 T EU621617 EU327621 EU665587 EU660258 EU665522Eunymphicus uvaeensis LPF LPF07-34 (UV) B EU621619 EU327623 EU665589 EU660260Geoffroyus heteroclitus AMNH DOT 6635 T EU621622 EU327626 EU665591 EU660263 EU665525 HQ316820 HQ316834 HQ316848Lathamus discolor ANWC 34174 T HQ316862 HQ316875 HQ316888 HQ316901 HQ316811 GQ505102 GQ505211 GQ505156Loriculus galgulus NMNH B06817 T EU621627 EU327631 EU665596 EU660267 EU665527 HQ316821 GQ505196 HQ316849Melopsittacus undulatus NMNH B06360 T EU621629 EU327633 EU665598 EU660269 EU665529 HQ316835 HQ316850Micropsitta finschii NMNH B04046 T EU621630 EU327634 EU665599 EU660270 EU737241 GQ505128 GQ505240 GQ505182Myiopsitta monachus NMNH B02706 T EU621631 EU327635 EU665600 EU660271 EU665531Neophema elegans NMNH B06444 T EU621634 EU327638 EU665603 EU660273 EU665533 HQ316822 HQ316836 HQ316851Neopsephotus bourkii ANSP ANSP11213 T EU621635 EU327639 EU665604 EU660274 EU665534 GQ505112 GQ505221 GQ505165Nestor notabilis NMNH B02885 T EU621637 EU327641 EU665606 EU660276 EU665536Northiella haematogaster NMNH B06432 T EU621638 EU327642 EU665607 EU660277 EU665537Pezoporus flaviventrisb ABBBS 230-06707 B, D HQ316863 HQ316876 HQ316889 HQ316902 HQ316812 HQ316823 HQ316837 HQ316852Pezoporus occidentalis QM QMO32613 D HQ316864 HQ316877 HQ316890 HQ316903 HQ316824 HQ316838 HQ316853Pezoporus wallicus ANWC 45982 T HQ316865 HQ316878 HQ316891 HQ316904 HQ316813 HQ316825 HQ316839 HQ316854Platycercus adscitus NMNH B06434 T EU621647 EU327651 EU665616 EU660285 EU665544 HQ316826 HQ316840 HQ316855Platycercus elegans NMNH B06370 T HQ316866 HQ316879 HQ316892 HQ316905Poicephalus robustus NMNH B06395 T EU621648 EU327652 EU665617 EU660286 HQ316814Polytelis alexandrae NMNH B02887 T EU621649 EU327653 EU665618 EU660287 EU665545 GQ505093 GQ505201 GQ505147Polytelis anthopeplus ANWC 50250 T HQ316867 HQ316880 HQ316893 HQ316906Prioniturus luconensis NMNH B03676 T EU621650 EU327654 EU665619 EU660288 EU665546 HQ316827 HQ316841 HQ316856Prosopeia tabuensis NMNH B02877 T EU621652 EU327656 EU665621 EU660290 EU665548Psephotus dissimilis ANWC 33421 T HQ316868 HQ316881 HQ316894 HQ316907 GQ505101 GQ505210 GQ505155Psephotus haematonotus ANWC 34210 T HQ316869 HQ316882 HQ316895 HQ316908Psephotus varius ANSP ANSP10641 T EU621653 EU327657 HQ378188 EU660291Psittacella brehmii KUMNH 4600 T HQ316870 HQ316883 HQ316896 HQ316909 HQ316815 HQ316828 HQ316842 HQ316857Psittacella picta AM AM0.59744 T HQ316871 HQ316884 HQ316897 HQ316910 HQ316829 HQ316843 HQ316858Psittacula columboides NMNH B06812 T EU621655 EU327659 EU665623 EU660293 EU665550Psittaculirostris edwardsii NMNH B06383 T EU621656 EU327660 EU665624 EU660294 EU665551 GQ505132 GQ505243 GQ505185Psittacus erithacus NMSU 987 T EU621657 EU327661 EU665625 EU660295 GQ505115 GQ505225 GQ505168Psittrichas fulgidus AMNH DOT 9597 T EU621658 EU327662 EU665626 EU660296 EU665552 GQ505127 GQ505239 GQ505181Purpureicephalus spurius ANSP ANSP11001 T EU621659 EU327663 EU665627 EU660297 EU665553Strigops habroptilus VU CD1139 B EU621663 EU327667 EU665631 EU660301 EU665557Tanygnathus lucionensis NMNH B06807 T EU621664 EU327668 EU665632 EU660302 EU665558

a Sample information refers only to sequences obtained by Wright et al. (2008) or in the present study. Sample information for sequences obtained by Schweizer et al. (2010) is available therein. Abbreviations: ABBBS, AustralianBird and Bat Banding Scheme, Canberra; AM, Australian Museum, Sydney; AMNH, American Museum of Natural History, New York; ANSP, Academy of Natural Sciences, Philadelphia; ANWC, Australian National Wildlife Collection,Canberra; LP, Loro Parque Fundación, Tenerife, Spain; MV, Museum Victoria, Melbourne; NMNH, Smithsonian National Museum of Natural History, Washington DC; NMSU, New Mexico State University Vertebrate Museum, LasCruces; QM, Queensland Museum, Brisbane; VU, Victoria University, Wellington. Abbreviations for sample type: T, tissue; B, blood, D, DNA extract provided by collection.

b Sample obtained from live bird that was banded and released.c Accession numbers starting with HQ identify sequences obtained in this study, those starting with EU were obtained by Wright et al. (2008); those starting with GQ were obtained by Schweizer et al. (2009).

L.Josephet

al./Molecular

Phylogeneticsand

Evolution59

(2011)675–

684677

678 L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684

phylogenetic signal between these loci (Wright et al., 2008).Accordingly, we used a combined phylogenetic analysis approachin which we concatenated or partitioned the five loci for searchesusing maximum likelihood (ML) and Bayesian (MB) criteria,respectively. For ML searches, we used the general time reversiblemodel plus invariant sites and a gamma distribution (GTR + I + G)model selected by the Akaike Information Criteria (AIC) in Model-Test 3.8 (Posada, 2006). The ML heuristic search was conducted inPAUP� using a random starting topology, TBR branch swapping,and proportion of invariant sites of 0.2944 and gamma shape dis-tribution of 0.51. Nodal support was assessed with 100 bootstrapreplicates in GARLI version 0.951 (Zwickl, 2006) using the sameML model, random starting trees and default parameters for thegenetic search algorithm.

For the MB searches, we used MrModelTest version 2.0(Nylander, 2004) to select the best fit model under the AIC for eachof the five gene regions (Table 2). We tested whether to partitionby codon within the mtDNA coding gene regions by partitioningeach region using first, second and third positions then comparingwith a non-partitioned gene dataset using Bayes factor analysis(Nylander, 2004). Analyses were run in MrBayes version 3.1.2(Ronquist and Huelsenbeck, 2003) and the harmonic mean of theestimated likelihood for nested analyses were compared. Partition-ing of the mtDNA gene regions (ND2, COI) by codon was stronglysupported by the Bayes factor and used in further analysis (Table 2).We also included indels as characters using the simple gap codingmethod following Simmons and Ochoterena (2000) and imple-mented in SeqState 1.4.1 (Müller, 2005). We then used a Bayesianframework with Markov chain Monte Carlo (MB) inference to esti-mate phylogenies using a dataset of five gene regions plus codedgaps. The MB analysis was run in MrBayes version 3.1.2 (Ronquistand Huelsenbeck, 2003) with 1 � 107 generations sampled every1000 generations. All analyses were run on the CIPRES 2.0 portal(Stamatakis et al., 2008) and the Computational Biology ServiceUnit from Cornell University. Each analysis was run three timesstarting from random trees. Convergence and mixing was assessedfor all the parameters in Tracer 1.4 (Drummond and Rambaut,2007). Each run was also evaluated for stationarity by comparinglog-likelihood values over time and 10% of generations (1 � 106)were discarded as burn-in. The posterior probabilities for individ-ual clades obtained from separate analyses were compared for con-gruence and then combined and summarized on a 50% majority-rule consensus tree (Huelsenbeck and Imennov, 2002; Huelsen-beck et al., 2002).

Based on results from these initial searches with the Primarydataset, we refined our analysis by reducing the number of taxato 27 and adding three additional loci in order to focus on cladesof particular interest and improve resolution at basal nodes ofthe trees (Table 1). We analyzed this Secondary dataset with MB

Table 2Model and partition data for each gene region included in the phylogenetic analyses.

Dataset Gene region Aligned

Primary (48 Taxa) COI 570ND2 1041TROP 548TGFB2 646RDPSN 817Concatenated 3622

Secondary (27 Taxa) COI 570ND2 1041c-mos 603RAG1 1458ZENK 1146TROP 536TGFB2 639RDPSN 810

searches using the eight gene regions plus coded gaps as separatepartitions. As with the Primary dataset, partitioning by codon wasstrongly supported for the mtDNA gene regions (ND2, COI) andwas used in further analyses; partitioning of the nDNA by codonwas not supported and was not employed in further analyses (Ta-ble 2). MB analyses were conducted on the entire concatenateddataset of eight gene regions and also on the mtDNA gene regionsand nuclear gene regions separately to assess conflict between thenuclear and mitochondrial loci.

We also implemented the species tree approach in BEST (Liu,2008) to incorporate gene tree heterogeneity (Edwards, 2008; Liuand Edwards, 2009). The Best analysis was run in mbbest (Liu,2008), in MrBayes (Ronquist and Huelsenbeck, 2003) at the Com-putational Biology Service Unit at Cornell University. Two analyseswere run, one with the combined mtDNA and nDNA, and one withnDNA gene regions only. Model parameters estimated from Model-Test were included as unlinked partitions for each gene region. Ini-tially we used priors suggested in the manual (inverse gammadistribution (3, 0.003) for theta and a uniform distribution (0.2,2) for gene mutation) but our searches were unable to reach con-vergence. We did reach convergence by selecting a broader inversegamma distribution prior (1, 1) (Wiens et al., 2010). We sampledevery 1000 generations for 2 � 107 generations. Convergence andmixing was assessed for all the parameters in Tracer v1.4 (Drum-mond and Rambaut, 2007). Species trees were constructed fromcombined runs.

3. Results

3.1. Sequence characteristics

Sequences obtained for the mtDNA genes were consistent inlength across all 48 taxa: 570 base pairs (bp) for COI and 1041 bpfor ND2. No stop codons or frameshift mutations were detected,suggesting that sequences were mitochondrial in origin and notnuclear pseudogenes. In contrast, sequence lengths were more var-iable in the three nuclear intron loci. Across the 48 taxa sampled,TROP sequences varied from 518 to 540 bp, TGFB2 sequences var-ied from 610 to 628 bp, and RDPSN sequences varied from 732 to795 bp in length. The three nuclear coding regions, which wereonly sampled for the 27 taxa in the Secondary dataset, showedintermediate levels of length variation, with C-MOS sequencesranging from 600 to 603 bp, RAG-1 sequences ranging from 1449to 1458 bp, and ZENK sequences ranging from 1137 to 1146 bpin length. No unexpected stop codons were detected and all indelswere of entire codons. After alignment and concatenation, the Pri-mary dataset was 3622 bp in length plus 100 coded gap charactersand the Secondary dataset was 6803 bp in length plus 79 codedgap characters (Table 2).

characters Model Partition

GTR + I + G CodonGTR + I + G CodonGTR + G Gene regionHKY + G Gene regionHKY + G Gene regionGTR + I + G Entire alignment

GTR + I + G CodonGTR + I + G CodonGTR + I + G Gene regionGTR + I + G Gene regionGTR + I + G Gene regionGTR + G Gene regionHKY + G Gene regionHKY + G Gene region

L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684 679

3.2. Phylogenetic analysis

For the Primary dataset, the ML analysis produced a single treewith a score of �32,914 that exhibited broad congruence with themajority consensus tree recovered using the MB analysis (Fig. 1).The ML and MB trees for the Primary dataset were consistent inshowing monophyly of Pezoporus, with strong support for a cladeof Pe. wallicus and Pe. flaviventris, which was in turn sister to Pe.occidentalis. Divergence between Pe. wallicus and Pe. flaviventriswas 3.0 % and that between this pair and Pe. occidentalis was

Fig. 1. Phylogenetic reconstruction of relationships of the Australo-Papuan parrots usinPosterior probabilities are indicated first followed by bootstrap support values from 100below 0.7 or 70% are not shown. Taxa names are colored by distribution: Australia in blueNew Zealand in black, Africa, Madagascar and south-east Asia west of Wallace’s Line in

6.1% (uncorrected p distances using mtDNA only). Also stronglysupported was a clade comprising Neophema–Neopsephotus–Pezoporus,which was in turn sister to a clade comprising Agapornis–Loriculus–Bolbopsittacus, albeit with slightly lower support values. We termthe clade uniting all of these taxa in analyses of the Primary dataset‘Clade 1’. The core platycercines formed ‘Clade 2’. Strongly sup-ported as sister to Clades 1 and 2 was ‘Clade 3’, which comprisedthe lorikeet-fig-parrot-budgerigar assemblage (Melopsittacus, Chal-copsitta, Cyclopsitta, Psittaculirostris). Finally, Psittacella comprised‘Clade 4’ and was sister to Clades 1, 2 and 3 (Fig. 1).

g the Bayesian criteria and the 48 taxa and five loci used in the Primary dataset.maximum likelihood runs. Values of 1.0 or 100% are indicated with asterisks; values, New Guinea, Indonesia east of Wallace’s Line and the south-west Pacific in purple,red, and South America in yellow.

680 L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684

Bayesian analyses of the Secondary dataset recovered somesimilar and some critically different relationships according towhich loci were used. Analyses of all eight loci and of the nuclearloci alone produced a well-resolved topology, and higher nodalposterior probabilities than the Primary dataset (Fig. 2). Analysisusing mtDNA recovered similar relationships among closely re-lated species but had poor resolution at deeper nodes. The key dif-ference in these results relative to the Primary dataset was thatAgapornis–Loriculus was sister to Clade 3 whereas Pezoporus, Neop-hema and Neopsephotus formed a well-supported clade (Fig. 2) thatwas sister to the core platycercines. In all of the Secondary data-set’s analyses, Psittacella tended to be sister to this entire assem-blage; the exception was in analysis of mtDNA plus nuclearcoding genes in which it was sister to what we term the expandedplatycercines: core platycercines + Pezoporus, Neophema andNeopsephotus. The mtDNA plus nuclear intron analysis agreed withthe Primary dataset in recovering Clade 1. Examination of individ-ual gene tree topologies in the Primary dataset suggested thatrecovery of Clade 1 was being driven primarily by the nuclear in-tron RDPSN (Supplementary material Figs. S1–S6).

In each analysis there is a different relationship among Apros-mictus, Polytelis and Alisterus although none have high posteriorprobabilities and none impact our central focus on platycercines,Pezoporus and Psittacella. The species tree analysis using BESTachieved good mixing and convergence but, in contrast to theBayesian analysis, showed poor resolution for both the entire data-set and for nDNA only (Supplementary material, Fig. S7).

In summary, the trees obtained using the Primary and Second-ary datasets supported the monophyly of Pezoporus and the sisterrelationship between this genus and Neophema and Neopsephotus.They differed, however, in whether this clade was sister to theclade containing Agapornis and Loriculus. In the Primary dataset allthese taxa were combined in Clade 1 of Fig. 1. In the Secondarydataset, however, (1) Agapornis and Loriculus were sister to Clade3 and (2) the clade of Pezoporus, Neophema and Neopsephotuswas sister to core platycerines. None of the analyses on either data-set recovered a sister relationship of Pezoporus with Melopsittacus,or of Agapornis and Loriculus with Micropsitta. Sister taxon relation-ships of Psittacella could not be determined confidently in anyanalysis other than to infer it as sister either to the entire assem-blage represented by Clades 1, 2 and 3 or, in one analysis, as sisterto expanded platycercines. No analysis included either Pezoporusor Psittacella within the core platycercine clade nor were they everfound to be sister to any of the genera typically classified in thetribes Polytelini (Alisterus, Aprosmictus, Polytelis, the last namednot being monophyletic as Schweizer et al., 2010 also found) orPsittaculini (Tanygnathus, Eclectus, Geoffroyus, Psittacula, Prionitu-rus). The latter two clades were consistently found to be monophy-letic clades and sister to each other.

4. Discussion

We set out to clarify the limits of platycercine (broad-tailed)parrots by determining whether two genera, Pezoporus of Australiaand Psittacella of New Guinea, are closely related to core platycer-cine genera (Platycercus, Barnardius, Psephotus, Northiella, Purpurei-cephalus, Lathamus, Cyanoramphus, Eunymphicus and Prosopeia).Both Pezoporus and Psittacella have been associated with Australianplatycercine parrots in distance-based analyses (see Section 1), butto date no phylogenetic analysis has adequately tested eitherhypothesis. We also sought to clarify relationships within Pezopo-rus, an enigmatic and little known genus of Australian parrots.Our results show that Psittacella is not closely aligned to the coreplatycercines. The composition of Pezoporus is clarified and we af-firm its close relationship to, but not within, core platycercines.

Based on these results, we describe below a rigorous phylogeneticand biogeographic circumscription of platycercine parrots andconsider the evolutionary history of the Australo-Papuan parrotradiation.

4.1. Systematics of Psittacella

For Psittacella, we confidently reject that it is a core platycercineor, for that matter, that it is close to any of the other genera withwhich it has been associated in the Psittaculini (Forshaw, 2006;Smith, 1975). Almost all analyses placed it on a long branch arisingearly in the history of a broad, predominantly Australian assem-blage comprising core platycercines (Clade 2) and a diversity ofother groups including Clade 3 as well as Pezoporus, Neophemaand Neopsephotus. Only one analysis recovered it as sister to an ex-panded platycercine assemblage that included Pezoporus, Neophe-ma and Neopsephotus. Furthermore, we found no support foraligning it with psittaculine groups with which it has more oftenbeen placed. We acknowledge that we have only examined thetwo larger species of Psittacella, having been unable to locate anytissue samples of the two smaller species.

Determining the systematic position of Psittacella is fundamen-tal to understanding the biogeography and ecology of Australo-Papuan parrots. Christidis et al. (1991) confidently consideredPsittacella as platycercine but with ‘‘disparate affinity’’ to thegroup. To reconcile their view with ours, we more closely examinethe details of the earlier study’s findings. First, Christidis et al.(1991) found that a UPGMA phenogram and distance Wagner tree,both built on allozyme-based genetic distances, aligned Psittacellawith core platycercines (and in one case with Pezoporus). In con-trast, their consensus trees generated using parsimony criteriaand the allozyme data could not resolve its position. Second, theyenumerated in some detail morphological traits Psittacella shareswith some but not all platycercines (e.g., blue cheek patches of Pspicta and red undertail-coverts of all Psittacella species). Given thattheir Wagner tree and our results at least agree in placing Psittacel-la well outside, not within, core platycercines, we ascribe thesemorphological similarities to convergence or retention of ancestral,plesiomorphic traits. Further, their earlier distance analyses werelikely ill-suited to clarifying relationships at the appropriate phylo-genetic depths. In summary, no previous analysis has appreciatedPsittacella’s likely systematic position and biogeographical signifi-cance. Far from being a New Guinean representative of the primar-ily south-east Asian and Indonesian psittaculine parrots, it is anextant representative in New Guinea of a very early branch in amajor radiation of Australo-Papuan parrots including platycercinesand Clade 3, the lorikeet-fig-parrot-budgerigar assemblage.

4.2. Systematics of Pezoporus: species level corollaries

The extreme rarity of all populations of Pezoporus precludedsampling of multiple individuals of the three relevant taxa. Withthat caveat, we find strong support for a sister group relationshipbetween eastern and western populations of the ground parrot(Pe. wallicus sensu lato) and that Pe. occidentalis is their sister taxon.Our findings are consistent with Murphy et al.’s (2011) species-le-vel recognition based on mtDNA and phenotype of Pe. wallicus(eastern ground parrot) and Pe. flaviventris (western ground par-rot). Whereas adults are typically sedentary, data from juvenilesof eastern populations of Pe. wallicus indicates their capacity to dis-perse up to 80 km (reviews in Higgins, 1999; Forshaw, 2002). Thusdispersal may have been important in the group’s evolutionary his-tory. We suggest, however, that when viewed against the broadertopography and biogeography of the Australian continent, the evo-lution of Pezoporus is most simply explained by sequential vicari-ance, first between the arid and coastal zones leading to the

Fig. 2. Phylogenetic reconstruction of relationships of the Australo-Papuan parrots using the Bayesian criteria and the 27 taxa and eight loci used in the Secondary dataset.Posterior probabilities are indicated above branches; values of 1.0 or 100% are indicated with asterisks and values below 0.7 or 70% are not shown. Images of birds painted byFrank Knight reproduced with permission (see Acknowledgments). The species depicted, from top to bottom are shown approximately to scale, and are: Pezoporus flaviventris,Neophema elegans, Psephotus dissimilis, Melopsittacus undulatus, Loriculus galgulus, Psittacella brehmi, Eclectus roratus (male, left and female, right), Micropsitta finschii.

L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684 681

evolution of Pe. occidentalis and the ancestor of Pe. wallicus + flavi-ventris, respectively, and later within coastal zones of southernAustralia, leading to Pe. wallicus and Pe. flaviventris. While recog-nizing current debate over the calibration of rates of molecularevolution (e.g., Lovette, 2004; Ho, 2007), we note that the conven-tional calibration for mtDNA evolution of 2% per million years (Tarrand Fleischer, 1993) would suggest that the split between Pe. occi-dentalis and Pe. wallicus/Pe. flaviventris likely occurred about3.3 mya, certainly before the Pleistocene. Assignment of Pe. occi-dentalis to monotypic Geopsittacus Gould, 1861(Brereton, 1963;

Courtney, 1997; Forshaw, 1969) though not a recent practice (For-shaw, 2002, 2006; Schodde, 1997) is a valid taxonomic option con-sistent with our data. We see no strong imperative or justificationfor the use of Geopsittacus, however: it would still be sister to Pe.wallicus + Pe. flaviventris and it is not especially divergent fromthem relative to many other congeneric sister taxa in birds (e.g., re-views in Avise and Walker, 1998; Johnson and Cicero, 2004; Josephand Omland, 2009; Klicka and Zink, 1997; Zink and Barrowclough,2008). We advocate retention of Pezoporus as the simplest meansof expressing the relationship of these three species to each other

682 L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684

and the likely history of simple vicariance events in their group’sevolution.

4.3. Systematics of Pezoporus and limits of the platycercine parrots

At the generic level, we reject a close relationship between Pezop-orus and Melopsittacus. Pezoporus is, however, firmly allied as sisterto Neophema and Neopsephotus. The main uncertainty in our findingsconcerns how closely related Pezoporus–Neophema–Neopsephotusare to the clade containing Agapornis, Loriculus and Bolbopsittacus.The linkage of Pezoporus–Neophema–Neopsephotus to non-Austra-lian but stouter bodied Agapornis, Loriculus and Bolbopsittacus camefrom analyses of our Primary dataset and was apparently drivenmostly by a single locus, the nuclear intron RDPSN. This result ac-cords with Wright et al. (2008), who found similar results usingthe same five loci as our Primary dataset but without samplingPezoporus. In contrast, Schweizer et al. (2010), also without samplingPezoporus, found weak support for Neophema and Neopsephotus assister to the core platycercine group using the three nuclear codinggenes we added into our Secondary dataset. Results from ourSecondary dataset were consistent with those of Schweizer et al.(2010), with Neophema–Neopsephotus as sister to the core platycer-cines and Agapornis–Loriculus as sister to the robustly supportedClade 3 (Melopsittacus–Chalcopsitta–Psittaculirostris) group. Accord-ingly, we consider the implications to the circumscription andevolution of platycercines of these two alternative placements ofAgapornis–Loriculus–Bolbopsittacus.

If affirmed, the radiation identified as Clade 1 from analyses ofour Primary dataset (Pezoporus–Neophema–Neopsephotus–Agapor-nis–Loriculus–Bolbopsittacus) would entail extraordinary ecological,biogeographical and phenotypic diversity. Its geographic rangewould span either side of Wallace’s Line. Its habitats would includerainforest, rocky coastlines, and extremely arid and climaticallyunpredictable Australian deserts and African savannas. Lastly, thephenotypic diversity in plumage patterning and morphology isextreme.

In contrast, if Pezoporus, Neophema and Neopsephotus (itself agroup with great phenotypic diversity; Fig. 2) are the sister groupto the core platycercines as in analyses of our Secondary dataset,then the platycercine radiation would be biogeographically cohe-sive within Australia and the Pacific, and it would have two broadlydivergent subclades (core platycercines and Pezoporus–Neophema–Neopsephotus).

4.4. Evolution of Australo-Papuan parrots: closing remarks

We conclude that Pezoporus–Neophema–Neopsephotus is mostlikely the sister group to the core platycercines. An alternativeplacement was supported largely by a single nuclear intron andpresents a biogeographic anomaly. We advocate continued restric-tion of the term platycercine and its formal nomenclatural corol-laries (Platycercinae, Platycercini) to the genera in Clade 2. Onesubclade of these core platycercines (Platycercus, Barnardius,Purpureicephalus, Psephotus, Northiella) is exclusively Australian.Their ecological radiation spans rainforest, wet temperate forests,temperate, semi-arid and monsoonal woodlands and even arid,desert shrublands. The other subclade includes three genera (Cyan-oramphus, Eunymphicus, Prosopeia) that have colonized deep intothe temperate and tropical south-west Pacific Ocean. Notably, theirsister, monotypic Lathamus of temperate south-eastern Australianwoodlands (present study; Schweizer et al., 2010), is migratory.If its migration evolved as early in the subclade’s history as theLathamus lineage itself, then this subclade may have been predis-posed to the dispersal and colonization that has clearly occurredin its history. It would represent an additional trans-oceanic dis-persal event in parrots (Schweizer et al., 2010). For the robustly

supported Pezoporus–Neophema–Neopsephotus clade advocate rec-ognition as a subfamily or tribe the nomenclatural details of whichwe will discuss elsewhere. Parenthetically, we note that crepuscu-lar or nocturnal activity characterizes a number of species in thisclade (Higgins, 1999; Forshaw, 2002) and suggest that this trait ar-ose early in the clade’s history.

Our results underscore and refine knowledge of the phyloge-netic and ecological diversity of parrots in the Australo-Papuan re-gion. Five clades or elements within them (cockatoos, polytelineparrots, core platycercines, Melopsittacus, and Neophema–Neopse-photus–Pezoporus) have diversified in Australia alone. Evolutionof one platycercine species, monotypic Lathamus, has involved aloss of seed-eating and concomitant origin of nectarivory and ana-tomical adaptations convergent on the nectarivorous lorikeets(Gartrell et al., 2000; Gartrell and Jones, 2001). Migration has aris-en at least twice (Lathamus and Neophema). There have been atleast three dispersals out of Australo-Papua itself: at least one bythe lories and lorikeets to the south-west Pacific islands, one bythe Agapornis–Bolbopsittacus–Loriculus lineage (Schweizer et al.,2010), and the other by core platycercines (Cyanoramphus–Eunym-phicus–Prosopeia; discussed above). The position of Psittacella assister to taxa in Clades 1 + 2 + 3, regardless of some uncertaintyabout relationships of non-Australian genera, is at least consistentwith Schodde’s ((2006); references therein) thesis of montane NewGuinean being a refugium for early lineages that may have beenancestral to more xeric Australian biota (see Jønsson et al., 2011).Our phylogenetic results, however, indicate more dynamism inthe evolution of the present-day ecology and biogeography ratherthan a simple ancestor–descendant pattern between New Guineaand Australia, respectively. Gardner et al. (2010) recently describedthe history of another major radiation of Australian birds, that ofthe passerine superfamily Meliphagoidea. They alluded to the like-lihood of extinction and complex patterns of relationships betweenAustralia and New Guinea as having been important in the evolu-tion of the present Meliphagoidea. Our findings indicate similarlydynamic histories for parrots in the region.

4.5. Final remarks

Some final systematic implications of our study concern Mayr’s(2010) tentative conclusion that the hypotarsal morphology was oflimited utility in parrot systematics. He stressed that resolution ofits significance depended on the position of Neophema, Neopsepho-tus and Micropsitta. Our study, though designed to assess the posi-tion of Psittacella and Pezoporus, shows that Micropsitta is notclosely aligned with the other taxa that Mayr (2008) earliergrouped in the Loricoloriinae (Agapornis, Loriculus, lorikeets andfig-parrots and Melopsittacus). We conclude that Loricoloriinae asMayr (2008) construed it is paraphyletic. Exclusion of Micropsitta,however, which does not appear at all close to the other relevanttaxa in our analyses, still would not render Loricoloriinae mono-phyletic because of lingering uncertainty surrounding positionsof Agapornis and Loriculus. de Kloet and de Kloet (2005) groupedAgapornis and Loriculus as sister taxa to what we have termedClade 3 but this had relatively weak support and was based onone gene. What does seem abundantly clear from the present studyand all recent phylogenetic analyses of parrots whether on ana-tomical or molecular data (e.g., Wright et al., 2008; Schweizeret al., 2010; Mayr, 2008, 2010) is that nomenclatural recognitionis warranted for Clade 3, the lorikeet, fig-parrot-budgerigar assem-blage. We will discuss details of this elsewhere.

Acknowledgments

AT was supported by the Australian National Wildlife CollectionFoundation. Funding was provided by National Institutes of Health

L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684 683

grant S06 GM008136 to TFW. The following people kindly suppliedcritical tissue samples or assisted in data collection: A. Nyari, A.T.Peterson, M. Robbins (Kansas University Natural History Museum,Lawrence), J. Sumner and V. Thompson (Museum Victoria,Melbourne), S, van Dyck and H. Janetzki (Queensland Museum,Brisbane), S. Murphy (Australian Wildlife Conservancy) and A.Burbidge (Western Australian Department of Environment andConservation), G. Graves (Smithsonian National Museum ofNatural History, Washington DC), P. Houde (New Mexico StateUniversity’s Vertebrate Museum, Las Cruces), D. Waugh, S. Capelli,H. Müller, J. Scharpegge (Loro Parque Funcación, Gran Canaria),T. Matsumoto, C. Miyaki (Universidad de São Paulo, São Paulo), J.Sanchez, A. Hernandez (National Institute of Toxicology and Foren-sic Science, Spain), and G. Chambers (Victoria University, Welling-ton). A. Burbidge, J. Forshaw, B. Halliday, and S. Murphy helpedwith discussions and laboratory aspects. S. Debus helped by stress-ing John Courtney’s discussions of the affinities of Pe. occidentalis.We thank Angela Frost for her expert help in preparing the figures.K. Aplin, J. Oakeshott and J. Peters commented very helpfully ondrafts. Part of this work was carried out by using the resources ofthe Computational Biology Service Unit from Cornell Universitywhich is partially funded by Microsoft Corporation. Images of par-rots from Forshaw (2006) were painted by Frank Knight and wethank author, illustrator and publisher (R. Kirk, S. Wolf, PrincetonUniversity Press) for permission to use them.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ympev.2011.03.017.

References

Avise, J.C., Walker, D., 1998. Pleistocene phylogeographic effects on avianpopulations and the speciation process. Proc. Roy. Soc. Lond. B 265, 457–463.doi:10.1098/rspb.1998.0317.

Boles, W.E., 2002. The fossil history of parrots. In: Forshaw, J. (Ed.), AustralianParrots, Robina Press, Queensland, pp. 36-40.

Boon, W., Robinet, O., Rawlence, N., Bretagnolle, V., Norman, J.A., Christidis, L.,Chambers, G., 2008. Morphological, behavioural and genetic differentiationwithin the horned parakeet (Eunymphicus cornutus) and its affinities toCyanoramphus and Prosopeia. Emu 108, 251–260.

Brereton, J.L., 1963. Evolution within the Psittaciformes. Proc. Int. Ornithol. Congr.13, 499–517.

Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T.J., Higgins, D.G., Thompson,J.D., 2003. Multiple sequence alignment with the Clustal series of programs.Nucleic Acids Res. 31, 3497–3500.

Christidis, L., Schodde, R., Shaw, D., Maynes, S., 1991. Relationships among theAustralo-Papuan parrots, lorikeets and cockatoos (Aves: Psittaciformes):protein evidence. Condor 93, 302–317.

Condon, H.T., 1941. The Australian broad-tailed parrots. Rec. South Aust. Mus. 7,117–144.

Courtney, J., 1997. The juvenile food-begging calls and associated aspects in the‘‘broad-tailed’’ (platycercine) parrots. Aust. Bird Watcher 17, 169–184.

de Kloet, R., de Kloet, S., 2005. The evolution of the spindlin gene in birds: sequenceanalysis of an intron of the spindlin W and Z gene reveals four major divisions ofthe Psittaciformes. Mol. Phylogenet. Evol. 36, 706–721.

Drummond, A., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis bysampling trees. BMC Evol. Biol. 7, 214. doi:10.1186/1471-2148-7-214.

Eberhard, J.R., Bermingham, E., 2005. Phylogeny and comparative biogeography ofPionopsitta parrots and Pteroglossus toucans. Mol. Phylogenet. Evol. 36, 288–304.

Edwards, S.V., 2008. Is a new and general theory of molecular systematicsemerging? Evolution 63, 1–19.

Forshaw, J., 1969. Australian Parrots. Landsdowne Press, Melbourne.Forshaw, J., 1989. Parrots of The World. third ed. Illustrated by W.T. Cooper,

Landsdowne Press, Melbourne.Forshaw, J., 2002. Australian Parrots. third ed. Illustrated by W.T. Cooper, Robina

Press, Queensland.Forshaw, J., 2006. Parrots of the World. An Identification Guide. Illustrated by F.

Knight. Princeton University Press, Princeton.Gardner, J., Trueman, J., Ebert, D., Joseph, L., Magrath, R.D., 2010. Phylogeny and

evolution of the Meliphagoidea, the largest radiation of Australasian songbirds.Mol. Phylogenet. Evol. 55, 1087–1102. doi:10.1016/j.ympev.2010.02.00.

Gartrell, B.D., Jones, S.M., 2001. Eucalyptus pollen grain emptying by two Australiannectarivorous psittacines. J. Avian Biol. 32, 224–230.

Gartrell, B.D., Jones, S.M., Brereton, R.N., Astheimer, L.B., 2000. Morphologicaladaptations to nectarivory of the alimentary tract of the swift parrot Lathamusdiscolor. Emu 100, 274–279.

Higgins, P.J. (Ed.), 1999. Handbook of Australian, New Zealand and Antarctic Birds.Vol. 4. Parrots to Dollarbird. Oxford University Press, Melbourne.

Ho, S.Y.W., 2007. Calibrating molecular estimates of substitution rates anddivergence times in birds. J. Avian Biol. 38, 409–414. doi:10.1111/j.2007.0908-8857.04168.x.

Homberger, D.G., 1980. Funktionell-morphologische Untersuchungen zur Radiationder Ernährungs – und Trinkmethoden der Papageien (Psittaci). Bonner Zool.Monogr. 13, 1–192.

Huelsenbeck, J.P., Imennov, N.S., 2002. Geographic origin of human mitochondrialDNA: accommodating phylogenetic uncertainty and model comparison. Syst.Biol. 51, 155–165.

Huelsenbeck, J.P., Larget, B., Miller, R.E., Ronquist, F., 2002. Applications and pitfallsof Bayesian inference of phylogeny. Syst. Biol. 51, 673–688.

Johnson, N., Cicero, C., 2004. New mitochondrial DNA data affirm the importance ofPleistocene speciation in North American birds. Evolution 58, 1122–1130.

Jønsson, K.A., Fabre, P.-H., Ricklefs, R.E., Fjeldså, J., 2011. Major global radiation ofcorvoid birds originated in the proto-Papuan archipelago. Proc. Natl. Acad. Sci.USA 108, 2328–2333.

Joseph, L., Omland, K., 2009. Phylogeography: its development and impact inAustralo-Papuan ornithology with special reference to paraphyly in Australianbirds. Emu 109, 1–23.

Joseph, L., Dolman, G., Saint, K., Donnellan, S., Berg, M., Bennett, A., 2008. Where andwhen does a ring start and end? Testing the ring species hypothesis in a speciescomplex of Australian parrots. Proc. Roy. Soc. Lond. B 275, 2431–2440.doi:10.1098/rspb.2008.0765.

Klicka, J., Zink, R.M., 1997. The importance of recent ice ages in speciation: a failedparadigm. Science 277, 1666–1669.

Leeton, P., Christidis, L., Westerman, M., Boles, W., 1994. Molecular phylogeneticaffinities of the night parrot (Geopsittacus occidentalis) and the ground parrot(Pezoporus wallicus). Auk 111, 833–843.

Liu, L., 2008. BEST: Bayesian estimation of species trees under coalescent model.Bioinformatics 24, 2542–2543.

Liu, L., Edwards, S., 2009. Phylogenetic analysis in the anomaly zone. Syst. Biol. 58,452–460.

Lovette, I., 2004. Mitochondrial dating and mixed support for the ‘2%’ rule in birds.Auk 121, 1–6.

Mayr, G., 2008. The phylogenetic affinities of the parrot taxa Agapornis, Loriculus andMelopsittacus (Aves: Psittaciformes): hypotarsal morphology supports theresults of molecular analyses. Emu 108, 23–27.

Mayr, G., 2010. Parrot interrelationships – morphology and the new molecularphylogenies. Emu 110, 348–357.

McDougall, A., Porter, G., Mostert, M., Cupitt, R., Cupitt, S., Joseph, L., Murphy, S.,Janetzki, H., Gallagher, A., Burbidge, A., 2009. Another piece in an Australianornithological puzzle – a second night parrot is found dead in Queensland. Emu109, 198–203.

Miyaki, C., Martioli, S., Burke, T., Wajntal, A., 1998. Parrot evolution andpaleogeographical events: mitochondrial DNA evidence. Mol. Biol. Evol. 15,544–551.

Müller, K., 2005. SeqState: primer design and sequence statistics for phylogeneticDNA datasets. Appl. Bioinf. 4, 65–69.

Murphy, S., Joseph, L., Burbidge, A., Austin, J., 2011. A cryptic and criticallyendangered species revealed by mitochondrial DNA analyses – the westernground parrot. Conserv. Genet. 12, 595–600. doi:10.1007/s10592-010-0161-1.

Nylander, J.A.A., 2004. MrModeltest v2. Program Distributed by the Author.Evolutionary Biology Centre, Uppsala University. <http://www.abc.se/’nylander/>.

Ovenden, J., Mackinlay, A.G., Crozier, R.H., 1987. Systematics and mitochondrialgenome evolution of Australian rosellas (Aves: Platycercidae). Mol. Biol. Evol. 4,526–543.

Posada, D., 2006. ModelTest Server: a web-based tool for the statistical selection ofmodels of nucleotide substitution online. Nucleic Acids Res. 34, W700–W703.

Ribas, C.C., Gaban-Lima, R., Miyaki, C.Y., Cracraft, J., 2005. Historical biogeographyand diversification within the Neotropical parrot genus Pionopsitta (Aves:Psittacidae). J. Biogeogr. 32, 1409–1427.

Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inferenceunder mixed models. Bioinformatics 19, 1572–1574.

Schodde, R., 1982. Origin, adaptation and evolution of birds in arid Australia. In:Barker, W.R., Greenslade, P.J.M. (Eds.), Evolution of the Flora and Fauna of AridAustralia. Peacock Publications, Adelaide, pp. 191–224.

Schodde, R., 1997. Aves (Columbidae to Coraciidae). In: Houston, W.M.K., Wells, A.(Eds.), Zoological Catalogue of Australia, vol. 37.2. CSIRO Publishing, Melbourne,pp. 1–440.

Schodde, R., 2006. Australasia’s bird fauna today – origins and evolutionarydevelopment. In: Merrick, J.R., Archer, M., Hickey, G., Lee, M. (Eds.), Evolutionand Biogeography of Australasian Vertebrates. AusciPub, Sydney,pp. 413–458.

Schodde, R., Calaby, J.H., 1972. The biogeography of the Australo-Papuan bird andmammal faunas in relation to Torres Strait. In: Walker, D. (Ed.), Bridge andBarrier: The Natural and Cultural History of Torres Strait. Department ofGeomorphology and Australian National University Press, Canberra, pp. 255–300.

Schweizer, M., Güntert, M., Seehausen, O., Hertwig, S.T., 2010. The evolutionarydiversification of parrots supports a taxon pulse model with multiple trans-

684 L. Joseph et al. / Molecular Phylogenetics and Evolution 59 (2011) 675–684

oceanic dispersal events and local radiations. Mol. Phylogenet. Evol. 54, 984–994. doi:10.1016/j.ympev.2009.08.021 3.

Simmons, M.P., Ochoterena, H., 2000. Gaps as characters in sequence-basedphylogenetic analyses. Syst. Biol. 49, 369–381.

Smith, G.A., 1975. Systematics of parrots. Ibis 117, 18–68.Stamatakis, A., Hoover, P., Rougemont, J., 2008. A fast bootstrapping algorithm for

the RAxML web-servers. Syst. Biol. 57, 758–771.Swofford, D.L., 1999. PAUP�. Phylogenetic Analysis Using Parsimony (�and Other

Methods). Sinauer Associates, Sunderland, MA.Tarr, C.L., Fleischer, R.C., 1993. Mitochondrial DNA variation and evolutionary

relationships in the Amakihi complex. Auk 110, 825–831.Tavares, E., Yamashita, C., Miyaki, C., 2004. Phylogenetic relationships among some

Neotropical parrot genera (Psittacidae) based on mitochondrial sequences. Auk121, 230–242.

Tavares, E., Baker, A., Pereira, S., Miyaki, C., 2006. Phylogenetic relationships andhistorical biogeography of Neotropical parrots (Psittaciformes: Psittacidae:

Arini) inferred from mitochondrial and nuclear DNA sequences. Syst. Biol. 55,454–470.

Wiens, J.J., Kuczynski, C., Arif, S., Reeder, T.W., 2010. Phylogenetic relationshipsof phrynosomatid lizards based on nuclear and mitochondrial data,and a revised phylogeny for Sceloporus. Mol. Phylogenet. Evol. 54,150–161.

Wright, T., Schirtzinger, E., Eberhard, J., Graves, G., Sanchez, J., Capelli, S., Muller, H.,Scharpegge, J., Chambers, G., Fleischer, R., 2008. A multilocus molecularphylogeny of the parrots (Psittaciformes): support for a Gondwanan originduring the Cretaceous. Mol. Biol. Evol. 25, 2141–2156.

Zink, R., Barrowclough, G., 2008. Mitochondrial DNA under siege in avianphylogeography. Mol. Ecol. 17, 2107–2121. doi:10.1111/j.1365-294X.2008.03737.x.

Zwickl, D.J., 2006. Genetic Algorithm Approaches for the Phylogenetic Analysis ofLarge Biological Sequence Datasets Under the Maximum Likelihood Criterion.Ph.D. Dissertation. The University of Texas at Austin, Austin.