- Email: [email protected]
Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies
Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies q Gary Voelker a,⇑, Jerry W. Huntley a, Joshua V. Peñalba b, Rauri C.K. Bowie b a b
Department of Wildlife and Fisheries Sciences and Texas Cooperative Wildlife Collections, Texas A&M University, College Station, TX 77843, USA Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Science Building, University of California, Berkeley, CA 94720, USA
a r t i c l e
i n f o
Article history: Received 9 June 2015 Revised 18 September 2015 Accepted 29 September 2015 Available online xxxx Keywords: Africa Eurasia Historical biogeography Muscicapa Muscicapidae Systematics
a b s t r a c t Muscicapa flycatchers and their allies (Bradornis, Dioptornis, Empidornis, Fraseria, Myioparus, Namibornis, and Sigelus) are widely distributed in Africa, Europe and Asia. This broad distribution and the wide variety of habitats occupied by the group, ranging from arid to tropical forests, presents an interesting opportunity to explore the evolution of biogeographic patterns and habitat associations. Sequence data (up to 3310 base pairs from two mitochondrial and two nuclear genes) were generated for 36 of 42 species which comprise the assemblage. Complementary data from an additional species was retrieved from GenBank, as was an additional gene which was available for 21 of our included taxa. Using modelbased phylogenetic methods and molecular clock dating, we constructed a time-calibrated molecular phylogenetic hypothesis for the lineage. Ancestral area reconstructions were performed on the phylogeny using LaGrange and BioGeoBEARS. Our results indicate that Bradornis, Fraseria, and Muscicapa are each non-monophyletic, with the latter being shown to comprise five separate clades each more closely related to other genera. Two new genera (Chapinia and Ripleyia) are erected to account for these results. Muscicapa and allies originated c. 7.4 Ma, most likely in Africa given that their sister lineage is almost entirely from there, and rapidly achieved a Eurasian distribution by c. 7.1 Ma. A second divergence at c. 6.1 Ma resulted in two clades. The first is a largely Eurasian clade that subsequently recolonized Africa, perhaps as the result of the loss of migration. The second is an African clade, and ancestral reconstructions suggest a Congolian (e.g. tropical forest) origin for this clade, with several subsequent diversifications into more arid habitats. This is a unique result, as most tropical forest lineages are confined to that habitat. As with other studies of African bird lineages, Afrotropical forest dynamics appear to have played a significant role in driving diversification in Muscicapa and allies, and our results include just the second recorded case of southern to northern African colonization patterns. Ó 2015 Elsevier Inc. All rights reserved.
1. Introduction Higher-level avian systematic relationships have undergone considerable reshuffling over the past few decades, as molecular phylogenetic analyses have made it comparatively easier to resolve homoplasy resulting from convergent evolution (Wake et al., 2011). The overarching result has tended to be more rigorous formulations of relationships, as indicated by consistency of inferred relationships across studies. This rigor is in marked contrast to older taxonomic work which, as pointed out by Sangster et al. (2010), tended to formulate suprageneric classifications based on often misleading assessments of phenotype and morphology, rather than quantitative assessments of relationships. q
This paper was edited by the Associate Editor Edward Louis Braun.
⇑ Corresponding author. Fax: +1 9798454096.
E-mail address: [email protected] (G. Voelker).
One example of higher level systematic and considerable taxonomic reshuffling centers on the Muscicapidae (Old World chats and flycatchers). This diverse and speciose assemblage has been the subject of a number of molecular phylogenetic studies that have focused primarily on assigning genera to higher level monophyletic groups. Results of these studies include, or confirm, the removal of chats from Turdidae to Muscicapidae, the separation of chats and flycatchers into tribes within Muscicapidae (Saxicolini and Muscicapini), and the erection of a subfamily (Muscicapinae) to house these tribes (e.g., Cibois and Cracraft, 2004; Voelker and Spellman, 2004; Sangster et al., 2010). Other studies have moved genera between these tribes, and to Saxicolini from Turdinae (e.g., Voelker and Spellman, 2004) and have erected new tribes to accommodate still other monophyletic phylogenetic groups (e.g., Sangster et al., 2010). Suprageneric relationships are not the only systematic and taxonomic issues in Muscicapidae. For example, in their extensive
http://dx.doi.org/10.1016/j.ympev.2015.09.026 1055-7903/Ó 2015 Elsevier Inc. All rights reserved.
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
2
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
molecular assessment of that family, Sangster et al. (2010) demonstrated that the genus Melaenornis (Muscicapini) is not monophyletic, and indeed taxonomic treatments have disagreed on the limits of ‘‘Melaenornis” by including some taxa from that genus in other genera (e.g., Mayr and Cottrell, 1986; Taylor, 2006; Dickinson and Christidis, 2014). Thus, while we have a better idea of suprageneric relationships in Muscicapidae, we are still working towards assigning species membership to genera with confidence, and the relationships of species within those genera. This has also impacted other types of analyses (e.g., historical biogeography studies) which rely on well-resolved phylogenies to better understand the evolution of the group. Indeed, the intercontinental distribution of Muscicapini, the diversity of habitats collectively occupied by members of the tribe and the presence of migratory behavior in a small subset of species combine to suggest that interesting biogeographic scenarios are likely. We focus here on the systematics and biogeography of the Muscicapini: Old World chats and flycatchers. This is a diverse assemblage of about 37 species, which collectively have widespread distributions centered in Africa and Eurasia (Taylor, 2006). Our goals are as follows: (1) to resolve the taxonomic issues related to genus-level relationships and genus membership by species, and (2) to assess biogeographic patterns in the group. Based on the distributional patterns of most Muscicapini (as well as the African ancestral area of its sister clade, the Copsychini; Voelker et al., 2014), we expect that: (1) Eurasian taxa will be derived from African ancestors; (2) colonization of Eurasia will be attributable to expansion through the Saudi Peninsula, and not overwater dispersals as we found in Copsychini (Voelker et al., 2014); and (3) gains or losses of migratory behavior by Eurasian taxa will be relevant to explaining intercontinental distribution patterns.
2. Material and methods 2.1. Taxon sampling and sequencing Primarily following Dickinson and Christidis (2014; exceptions noted below) our analyses included 37 Muscicapini species representing the following genera: Muscicapa (19 of 23 species), Melaenornis (3/4), Myioparus (2/2), Namibornis (1/1), Dioptrornis (3/3; all considered Melaenornis by Dickson and Christidis), Empidornis (1/1), Fraseria (2/2), Sigelus (1/1) and Bradornis (4/4) (Supplementary Table S1). We note here that the last five of these genera, until recently, have been considered Melaenornis (e.g., Mayr and Cottrell, 1986; Taylor, 2006), but that genus has been rendered non-monophyletic in a recent molecular systematic study (Sangster et al., 2010), which included a subset of the Muscicapini species (n = 21) included here as part of our broader sampling. Zuccon and Ericson (2010) similarly found Melaenornis to be non-monophyletic, but included just nine Muscicapini species. We did not include the monotypic Humblotia, which Dickinson and Christidis (2014) place as a Muscicapini species (but see Jønsson and Fjeldså, 2006), but we did include Namibornis, which is not a member of Copsychini (see Mayr and Cottrell, 1986, and results below) as previously suggested. As outgroup taxa, we included a suite of Erythropygia and Copsychus species which, as members of Copsychini, form the sister clade to Muscicapini (Sangster et al., 2010; Voelker et al., 2014). Whole genomic DNA was extracted from tissue or toepads using the DNeasy tissue extraction kit (Qiagen, Valencia, CA, USA). We used the polymerase chain reaction (PCR) to amplify the mitochondrial cytochrome-b (CYB) and NADH dehydrogenase subunit 2 (ND2) genes, as well as the autosomal nuclear Myoglobin (MB intron 2) and Beta-Fibrinogen (FGB intron 5) genes. Standard primers and reaction conditions were employed. PCR products
were cleaned with ExoSAP-IT (Affymetrix, Santa Clara, CA, USA). Sanger sequencing was performed at the Beckman Coulter Genomics facility (Danvers, MA, USA). The mtDNA sequences were aligned using Sequencher v. 4.9 (Gene Codes Corporation, Ann Arbor, MI, USA). Introns were aligned using MAFFT (Katoh, 2013). 2.2. Phylogenetic analyses and divergence dating To determine the best-fit model(s) for the mitochondrial (mtDNA) sequence data, we assessed three alternate partitioning schemes, each of which relied on appropriate partition models derived from MrModelTest (Nylander, 2004). In our first partitioning scheme (two partitions) the ND2 and CYB genes were unlinked. In the second (four partitions), first and second codon positions were linked for ND2, linked for CYB, and third codon positions for each gene were treated as independent partitions. In the third scheme (six partitions), each codon position was unlinked across both genes. For each mixed-model partition scheme, we used MRBAYES (Huelsenbeck and Ronquist, 2001) to initiate two runs of four Markov-chain Monte Carlo (MCMC) chains of 5,000,000 generations, each starting from a random tree and sampling every 100 generations. Each run resulted in 50,000 trees and converged on the same topology. The first 5000 trees from each analysis were removed as ‘‘burn-in”, and the remaining 90,000 trees were used to generate a majority rule consensus tree. Bayes factors were computed using the harmonic means of the likelihoods calculated from the sump command within MRBAYES. A difference of 2 ln Bayes factor >10 was used as the minimum value to discriminate between mixed-model partitioning schemes (Brandley et al., 2005; Brown and Lemmon, 2007), and the six partition scheme was identified as the best-fit to the mitochondrial data. We then used the sixpartition mtDNA scheme in combination with the nuclear gene data. Best-fit models for nuclear data were also derived from MrModelTest analyses. This combined analysis set each nuclear gene and each mtDNA codon position as unlinked in two runs of four MCMC chains in MRBAYES as described above. We also incorporated best fit models for each gene in BEAST 2.0 (Drummond et al., 2006, 2012) to reconstruct a tree using all loci, and simultaneously estimated divergence times across all species. We employed a lineage substitution rate of 0.014 per lineage/ million years for CYB using a relaxed, uncorrelated lognormal clock. This substitution rate translates to 2.8% per million years, and is generally applicable to the CYB gene in songbirds (Weir and Schluter, 2008; Lerner et al., 2011). We used a rate of 0.029 per lineage/million years for ND2 (Lerner et al., 2011). We applied a slightly broader prior (0.002) for FGB than the 0.0017 rate used by Lerner et al. (2011), to reflect the greater variation of rates among introns. We applied the same 0.002 rate to MB. Standard deviations for the mitochondrial genes (CYB: 0.001; ND2: 0.0025) followed Lerner et al. (2011), while we used a 0.002 standard deviation for the introns. A Yule process speciation prior was implemented in each analysis. Two separate MCMC analyses were run for 10,000,000 generations with parameters sampled every 1000 steps, with a conservative 20% burn-in. Independent runs were combined using LogCombiner v.1.6.1 (Drummond et al., 2012). Tracer v.1.5 (Rambault and Drummond, 2007) was used to measure the effective sample size of each (all >200) and calculate the mean and upper and lower bounds of the 95% highest posterior density interval (95% HPD) for divergence times. Tree topologies were assessed using TreeAnnotator v.1.7 (Rambault and Drummond, 2007; Drummond et al., 2012) and FigTree v.1.3.1 (Rambault, 2008). Finally, we also performed the above analyses on a five gene dataset. These analyses included the ODC1 gene from 21 ingroup taxa for which data was available on GenBank. Nexus trees of both datasets are available as Supplementary material in the online version.
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
3
2.3. Historical biogeography
3.2. Divergence dating and biogeographic history
For biogeographic analysis of the entire ingroup we used likelihood analysis of geographic range evolution (dispersal-extinction cladogenesis, DEC) implemented in LaGrange v. 2.0.1 (Ree and Smith, 2008). The DEC model as implemented in LaGrange has two free parameters specifying the rate of ‘‘dispersal” (i.e., range expansion) and ‘‘extinction” (i.e., range contraction), but the cladogenesis model remains fixed. This has an important consequence because the geographical range of the ancestral lineage is inherited by the two daughter lineages through a variety of plausible cladogenetic scenarios (e.g., sympatric, parapatric, vicariance) that have equal probability. We also made use of a new method implemented in the BioGeoBEARS package in the R statistical program (Matzke, 2013a,b) that adds a third free parameter to the DEC framework: long-distance dispersal (parameter j DEC+j model). This additional parameter enables one daughter lineage to disperse to an area outside (i.e. not adjoining) the ancestral range, thereby effectively mimicking the process of founder-event speciation. Because the classic DEC model of LaGrange is nested within the DEC+j model implemented in BioGeoBEARS, we are able to perform model comparison (DEC vs. DEC+j) with the Likelihood Ratio Test and by using AIC weights. We followed the statistically defined African sub-regions of Linder et al. (2012) for birds, and using distribution maps in Taylor (2006) categorized species as being present or absent in the following five areas or area combinations: Southern African, Zambezian, Congolian, Ethiopian + Somalian, and Sudanian. A sixth area, Europe + Asia, was included to account for the distribution of non-African taxa.
Molecular clock calibration sets the origin of the Muscicapa and allies clade to c. 7.4 Ma (Fig. 2). Within the clade, Muscicapa ruficauda diverged from a clade representing all other taxa at c. 7.1 Ma (Fig. 2). Ancestral reconstructions using LaGrange ( ln value of 118.6) suggested a Muscicapa and allies ancestral area comprising all defined areas, whereas the BioGeoBEARS DEC model ( ln value of 119.7) suggested a Europe + Asian ancestral area (Fig. 3, and see Fig. S1 for an alternative depiction of BioGeoBEARS output). The DEC + j model in BioGeoBEARS was not significantly better ( ln 111.2; P = 0.079). The next lineage divergence at c. 6.1 Ma resulted in two main clades. The first of these is the Muscicapa striata clade which comprises 11 species, six of which have exclusively Europe + Asia distributions. Europe + Asia is reconstructed by both methods as the ancestral area for this clade. Within this clade, there are two subsequent colonizations of Africa resulting in five taxa, and a back colonization of Europe + Asia by one of those taxa (Muscicapa gambagae; Fig. 3). All but one of the divergences in this clade are dated in the Pliocene (Fig. 2). The second main clade comprises all other ingroup taxa. The Congolian region is reconstructed by both methods as the ancestral area for the clade (Fig. 3). Several subsequent, rapid divergences in the latest Miocene/earliest Pliocene are also biogeographically reconstructed as Congolian, or include that area in the most likely reconstruction (Figs. 2 and 3). All species in the Fraseria ocreata–Fraseria cinerascens clade occur in the Congolian region, with several having achieved slightly broader (Myioparus griseigularis) or very broad distributions (Myioparus plumbeus and Muscicapa caerulescens). Other main divergences coincide with colonizations of the Southern African, Zambezian and Ethiopian + Somalian regions by the Empidornis–Meleanornis edolioides clade, of those same regions plus the Sudanian region by Bradornis infuscatus + B. pallidus, and of the Southern African, Zambezian and Ethiopian + Somalian regions by the Bradornis mariquensis–Muscicapa comitata clade (Fig. 3). Most divergences in these four subclades occur in the Pliocene, with just three having occurred more recently in the Pleistocene (Fig. 2).
3. Results 3.1. Phylogenetic analyses We sequenced up to a total of 3310 bp from four gene regions, for each individual sampled. Complete data from each gene was generated from most taxa, or retrieved from GenBank (Appendix). In a few instances, the sequences retrieved for a given taxon were not from the same individual (Appendix). Across all ingroup taxa (Fig. 1) there were 1210 variable sites of which 808 were parsimony informative. Bayesian analysis of the combined four gene dataset resulted in a phylogeny with mixed nodal support, where 25 of 36 nodes (69%) received strong support of P0.95 posterior probability and 11 of 36 nodes (31%) received moderate to weak support (Fig. 1). The addition of the ODC gene increased support values at just two nodes (Fig. 1). Overall, most nodes receiving moderate to poor support were basal in the phylogeny. Our BEAST analysis of the four gene dataset resulted in a phylogeny with similarly mixed support, where just 21 nodes (58%) were strongly supported (Fig. 2). There are six nodes in the species tree that conflict with the concatenated Bayesian topology; three of these conflicting nodes are strongly supported in the Bayesian phylogeny (Fig. 1). However, there are also two strongly supported nodes in the tree that were less well supported in the Bayesian concatenated phylogeny: Dioptrornis chocolatinus is supported as sister to D. brunneus + D. fischeri, and Muscicapa comitata is supported as sister to the Bradornis pallidus clade (Fig. 2). Across both analysis methods then, 81% of nodes received posterior probability support of P0.95. It is clear from our results that the taxonomy of this group requires revision. Fraseria (two species) is not monophyletic, nor is Bradornis (five species in two clades). The most obvious case of non-monophyly is evident in Muscicapa, which is composed of five distinct lineages across the phylogeny (Fig. 1).
4. Discussion 4.1. Molecular phylogeny and taxonomic implications Using molecular data from 20 taxa, Sangster et al. (2010) defined Muscicapini as comprising the genera Muscicapa, Melaenornis and, Fraseria; they did not include Myioparus in their analysis but it is clearly also a member of Muscicapini (Fig. 1). Other genera ascribed to Muscicapini historically were excluded by their results, which also indicated issues with the monophyly of each of the remaining three genera (Muscicapa, Melaenornis, and Fraseria; Sangster et al., 2010). Our study represents the most extensively sampled molecular phylogenetic analysis of Muscicapini species to date. Overall, we included 37 of 42 Muscicapini species (see Section 2). Of the five taxa we are missing in our analyses, two are insular forms from Southeast Asia (Muscicapa randi – Philippines and Muscicapa segregata – Lesser Sundas), and both have been considered as conspecific with Muscicapa dauurica (Taylor, 2006). The remaining three species have patchy or restricted ranges in Afrotropical rainforests (Muscicapa lendu, Muscicapa epulata, Melaenornis annamarulae). Our attempts to extract DNA from museum specimens of Muscicapa lendu and Muscicapa epulata were unsuccessful. Although comprising just four species in recent taxonomy (e.g., Taylor, 2006), the genus Melaenornis has previously been considered to include species now ascribed to Empidornis, Sigelus,
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
4
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
*
0.59
Sigelus Namibornis
*
*
* *
0.74
Dioptrornis chocolatinus Dioptrornis brunneus Dioptrornis fischeri Melaenornis ardesiacus
*
0.91 (0.95)
Melaenornis pammelaina Melaenornis edolioides Fraseria ocreata
0.63
*
0.64
Myioparus griseigularis Myioparus plumbeus Muscicapa olivascens (Apatema)
* *
0.90
Empidornis
*
Muscicapa tessmani Muscicapa caerulescens
(Cichlomyia or Butalis)
Fraseria cinerascens (Chapinia)
*
0.87
Bradornis infuscatus Bradornis pallidus
Bradornis mariquensis
*
0.55
(Haganopsornis)
*
Bradornis pumilus Bradornis microshynchus Muscicapa comitata (Pedilorhynchus)
0.90
*
*
*
*
*
* * *
*
(Artomyias)
(Bradyornis)
Muscicapa boehmi
0.82
*
Muscicapa ussheri Muscicapa infuscata Muscicapa striata
Muscicapa gambagae
0.94 (0.97)
*
*
Muscicapa aquatica Muscicapa cassini Muscicapa adusta Muscicapa sethsmithi Muscicapa daaurica Muscicapa muttui Muscicapa sibirica Muscicapa ferruginea Muscicapa griseisticta Muscicapa ruficauda
(Ripleyia)
Fig. 1. Molecular phylogeny of Muscicapa and allies based on Bayesian analysis of concatenated data (two mitochondrial and two nuclear genes). Asterisks denote Bayesian posterior probabilities P0.95, with lower posterior probability values shown. Numbers in parentheses are posterior probabilities derived from a reduced taxa five gene analysis (see Section 2), which provide stronger support for several relationships than support values returned from the four gene analysis. Note that several genera are rendered non-monophyletic in these analyses, necessitating nomenclatural revisions at the genus level. We elaborate on these revisions elsewhere (see text), but parenthetically note suggested alternative generic designations on the phylogeny.
Dioptrornis, Bradornis, and Fraseria (Mayr and Cottrell, 1986). Our results indicate, (1) that a larger Melaenornis (to include the aforementioned four genera) would be non-monophyletic (see also Sangster et al., 2010), (2) that Melaenornis (edolioides is the type) could be restricted to as little as three and perhaps four species (depending on the eventual systematic placement of annamarulae), and (3) that Melaenornis could be expanded to include Dioptrornis, Empidornis, Sigelus and Namibornis (Figs. 1 and 2). Due to the morphological distinctiveness of the latter four genera (three of which are monotypic) relative to Melaenornis, which are all black or dark gray in color, we agree with the more strict usage of Melaenornis (e.g., Taylor, 2006).
Our results also indicate that Bradornis is non-monophyletic, with species falling into two distinct clades (Figs. 1 and 2). The Bradornis type is mariquensis (Mayr and Cottrell, 1986), and thus that genus should be applied to mariquensis, pumilus and microrhyncus. There is an available synonym, Haganopsornis, which was applied to infuscatus (Roberts, 1922 fide Mayr and Cottrell, 1986) and we suggest resurrecting that genus to include infuscatus and pallidus (Figs. 1 and 2). The genus Fraseria is also nonmonophyletic, with ocreata being more closely related to the genus Myioparus; our results conflict as to the phylogenetic position of cinerascens (Figs. 1 and 2). Regardless, Fraseria would apply to ocreata (Mayr and Cottrell, 1986). We find no synonym to apply
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
5
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
0.39
*
*
*
*
*
0.85
*
0.93 0.38
*
0.44 0.82
*
Fraseria cinerascens 0.90
Muscicapa olivascens Muscicapa tessmani Muscicapa caerulescens Bradornis infuscatus Bradornis pallidus Bradornis mariquensis Bradornis pumilus Bradornis microshynchus Muscicapa comitata Muscicapa ussheri Muscicapa infuscata Muscicapa boehmi Muscicapa gambagae Muscicapa striata Muscicapa aquatica Muscicapa cassini Muscicapa adusta Muscicapa sethsmithi Muscicapa daaurica Muscicapa muttui Muscicapa sibirica Muscicapa ferruginea Muscicapa griseisticta Muscicapa ruficauda
* *
* 0.88
*
*
*
*
0.91
0.43
0.69
*
* *
*
0.43
*
*
*
0.78 0.42
Miocene 7.0
6.0
Pliocene 5.0
Empidornis Namibornis Sigelus Dioptrornis chocolatinus Dioptrornis brunneus Dioptrornis fischeri Melaenornis ardesiacus Melaenornis pammelaina Melaenornis edolioides Fraseria ocreata Myioparus griseigularis Myioparus plumbeus
4.0
3.0
Pleistocene 2.0
1.0
0.0
Time in Ma Fig. 2. BEAST and molecular clock analysis based on the four-gene dataset. Asterisks denote Bayesian posterior probabilities P0.95, with lower posterior probability values shown. Arrows at nodes indicate points of conflict with the concatenated Bayesian analysis (Fig. 1). Molecular clock rates were based on rates derived by Lerner et al. (2011; see Section 2). Bars at nodes indicate 95% highest posterior density intervals. Time is millions of years before present.
to cinerascens, which then requires the designation of a new genus for that species which we propose as: Chapinia, new genus Voelker & Bowie Type species. – Chapinia cinerascens. Diagnosis – A genus of muscicapid flycatcher differing from all other genera of the family Muscicapidae by the following combination of characters: large size, diagnostic white supra-loral spot, dark upperparts, and mottled gray underparts with dark but poorly demarcated crescents on the breast. Etymology – This name honors Dr. James P. Chapin, for his extensive documentation of, and research on, the birds of the Belgian Congo. Finally, the genus Muscicapa appears to have been a taxonomic dumping ground for any small to medium sized Muscicapini flycatcher, as our results show it to be comprised of five distinct lineages (Fig. 1). The type for the genus is striata, thus Muscicapa would apply to the large clade of 11 species (Fig. 1). There are several synonyms available for other clades. The genus Apatema could be applied to olivascens, and Cichlomyia or Butalis (it is unclear to us which has priority) could apply to the closely related caerulescens, and thus to tessmani as well (Fig. 1). For comitata, the genus Pedilorhynchus is available, and Artomyias is available for infuscata
and thus also for the closely related ussheri (Fig. 1). Although sister to infuscata + ussheri, boehmi is highly distinct from them morphologically (Sinclair and Ryan, 2010). We therefore suggest applying the name Bradyornis to boehmi, following the original description of this species (Reichenow, 1884, fide Mayr and Cottrell, 1986). We find no available synonym for ruficauda, which is the first species to diverge within Muscicapini (Fig. 1). This requires the designation of a new genus for that species which we propose as: Ripleyia, new genus Voelker & Bowie Type species. – Ripleyia ruficauda. Diagnosis – A genus of muscicapid flycatcher differing from all other genera of the family Muscicapidae by the following combination of characters: rufous uppertail-coverts and tail, faint supercilium, and entirely orange lower mandible. Etymology – This name honors Dr. S. Dillon Ripley, former Secretary of the Smithsonian Institution, for his extensive work on the birds of India and southern Asia. 4.2. Biogeographic history – Intercontinental movements Our results indicate three movements between Africa and Eurasia. The first of these assumes movement from Africa to
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
6
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
SA SA SA Ethiopian + Somalian
Sudanian
Congolian Zambezian
CO ZSACO CO CO CO
Southern African
Z Z Z
Z Z Z
CO CO
CO CO CO
Z Dioptrornis brunneus Dioptrornis fischeri Z Melaenornis ardesiacus
Z
ESOZSACO
ESOZSACO
Melaenornis pammelaina
ESOSU
Meleanornis edolioides CO Fraseria ocreata
ZCO Myioparus griseigularis CO ESOSUZSACO CO ESOSUZSACO Myioparus plumbeus
CO CO CO
CO CO
ESOZ
Empidornis semipartitus
SA Sigelus silens Namibornis herero ESO Dioptrornis chocolatinus
SA
SA ESOZSA Z Z Z Z
ESO
ESOSA
CO
CO CO CO
Muscicapa olivascens CO Muscicapa tessmani
ESOZSACO CO
Muscicapa caerulescens
Fraseria cinerascens
ESOZSACO
CO CO
ESOSUZCOSAEAs EAs
SA ESOSUZSACO ESOSUZCO
CO
CO CO
SA ESOSA ESO
CO CO CO ESOSACO
CO Z
Bradornis infuscatus Bradornis pallidus SA Bradornis mariquensis
ESO Bradornis pumilus Bradornis microrhynchus CO Muscicapa comitata CO Muscicapa ussheri
ESO
CO CO CO
ZCO
SA
ESOSUZCO
Z
ESOSACOEAs
ESO
CO Muscicapa infuscata Muscicapa boehmi EAs
EAs SUEAs EAs
SUEAs Muscicapa gambagae Muscicapa aquatica ZCO Muscicapa cassini ESOSUZSACO Muscicapa adusta
SU
EAs
EAs EAs
EAs EAs
EAs EAs EAs
SU
SU+EAs EAs+CO SUZCO SU CO SUCOEAs SU CO CO ESOSUZSACO EAs CO
EAs
EAs EAs EAs EAs EAs EAs
Muscicapa striata
CO EAs
Muscicapa sethsmithi
Muscicapa daaurica EAs Muscicapa muttui EAs
Muscicapa sibirica Muscicapa ferruginea EAs Muscicapa griseisticta EAs
EAs
Muscicapa ruficauda
Fig. 3. Ancestral area reconstructions based on the DEC model implemented in BioGeoBEARS (reconstructions to the right of nodes), and LaGrange (reconstructions to the left of nodes). LaGrange returns two reconstructions, thus the upper reconstruction is the most likely area reconstruction derived from all node reconstructions from lineages ‘‘above” the node, and the lower is the most likely reconstruction derived from all node reconstructions from lineages ‘‘below” the node. Areas are defined as follows: Southern African (SA), Zambezian (Z), Congolian (CO), Ethiopian + Somalian (ESO), Sudanian (SU) and Europe + Asia (EAs). The coded distribution for each species is indicated next to that species’ name.
Eurasia c. 7.4 Ma to explain the Eurasian distribution of Muscicapa ruficauda (Figs. 2 and 3). We assume this direction of movement given that the basal taxa in the sister clade (Erythropygia and allies) to Muscicapini are almost entirely of African origin (Voelker et al., 2014). A second Africa to Eurasian movement at c. 6.8 Ma is required to explain the basal Eurasian ancestral reconstructions for the Muscicapa striata clade (Figs. 2 and 3). Finally, movement from Eurasia to Africa at c. 4.7 Ma is required to explain the distributions of five African taxa imbedded in the otherwise Eurasian Muscicapa striata clade (Figs. 2 and 3). Given the distribution of Muscicapa ruficauda in Central Asia, and the broader Eurasian (i.e., both European and Asian) distributions of many of the taxa in the Muscicapa striata clade, we assume land-based colonization from Africa, rather than the repeated overwater dispersals which we invoked to explain African-East Asian connections in the sister group to Muscicapini (Erythropygia and allies; Voelker et al., 2014). Land-based movements are also a rea-
sonable conclusion for the recolonization of Africa by this clade, as several internal ancestral reconstructions include the Sudanian or Ethiopian + Somalian regions (Fig. 3) which are geographically adjacent to the Saudi Peninsula, and thus Eurasia generally. Further, the dates of these movements are during a period (10– 5.3 Ma) when an Afro-Arabian landbridge was in place. This connection allowed for extensive bi-directional dispersal of vertebrate lineages between the Horn of Africa and the Arabian Peninsula (Amer and Kumazawa, 2005; Pook et al., 2009; Wong et al., 2010; Metallinou et al., 2012; Portik and Papenfuss, 2012; Šmíd et al., 2013). Palearctic-derived lineages also began to colonize the Arabian Peninsula during this time (Portik and Papenfuss, 2015), suggesting large scale habitat changes across the region. These land based movements may also be related to evolution (gains or losses) of migration. The initial colonization of Eurasia from Africa could be explained via the gain of migratory behavior (the sister clade is non-migratory), by Muscicapa ruficauda (or more
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
likely an ancestor) which breeds in the Himalayas but winters in India (Taylor, 2006). The second colonization of Eurasia from Africa, and the recolonization of Africa from Eurasia could be related to the gain and subsequent loss of migration in the Muscicapa striata clade. All Eurasian taxa in this clade (Fig. 3) are migratory, and all but one winter in India or Southeast Asia; the exception is striata which winters in Africa (Taylor, 2006). This would suggest that the striata clade (and an ancestor of Muscicapa ruficauda) expanded to Eurasia to exploit suitable breeding habitat, and subsequently gained migratory behavior in response to winter climates (e.g., the southern home hypothesis of Cox, 1985). A subsequent loss of migratory behavior coincides with the recolonization of and subsequent diversification by African members of the striata clade (Fig. 3). The biogeographic connection that striata has with Africa (wintering range) and its close relationship to the African taxa implies a migratory drop-off pattern similar to what has been shown in other studies of avian genera with species distributed in both Africa and Eurasia (e.g., Voelker et al., 2009; Voelker and Light, 2011). 4.3. Biogeographic history – intra-African patterns After the initial recolonization of Africa at c. 4.7 Ma, the African members of the striata clade began to diversify at c. 3.6 Ma. Both of these dates coincide with a period from 5–3 Ma, when global climate changes caused significant expansions of the Afrotropical forests into Kenya, Tanzania, and Ethiopia (Hamilton and Taylor, 1991; Cane and Molnar, 2001; Feakins et al., 2005; Sepulchre et al., 2006), which, in turn, had a significant role in generating avian diversity, to include later diversifications during forest retractions (e.g., Outlaw et al., 2007; Fjeldså and Bowie, 2008; Voelker and Outlaw, 2008; Voelker et al., 2010, 2012; Fuchs et al., 2011). The Congolian region is reconstructed as the ancestral area for several basal nodes in the other major clade of Muscicapini (Fig. 3), which is comprised of three subclades (and Bradornis infuscatus + Bradornis pallidus). Two of these subclades, the Fraseria cinerascens–Fraseria ocreata clade (seven species) and the Muscicapa boehmi-Bradornis mariquensis (seven species) clade, are inferred to have a Congolian ancestral area (Fig. 3). The distribution of the Fraseria cinerascens–Fraseria ocreata clade can be explained via expansion from the Congolian region into adjacent areas primarily during the Pliocene (Figs. 2 and 3). In the Muscicapa boehmi–Bradornis mariquensis subclade, the divergence between the Muscicapa taxa (four species) and the Bradornis taxa (three species) is inferred to have occurred at c. 4.7 Ma, indicating another divergence that could be linked to Afrotropical forest dynamics (see above). The subsequent divergences of Bradornis lineages began at c. 2.9 Ma, as the tropical forest was retracting westward. This date, and the fact that the Congolian region and Ethiopian + Somalian regions are not adjacent (the Sudanian and Zambezian regions are between them) suggest that Bradornis pumilus and B. microrhynchus are the result of isolation due to forest retraction. Indeed, this divergence date matches well with other taxa inferred to have been the result of forest retraction and subsequent isolation in eastern Africa (Voelker et al., 2010). We discuss the anomalous distribution of Bradornis mariquensis in Southern Africa below. The ancestral area at the base of the remaining clade (Melearnornis edolioides–Empidornis; nine species) is reconstructed as South African + Zambezian (Fig. 3). This clade diverged from the Congolian Fraseria cinerascens–Fraseria ocreata clade at c. 5.2 Ma. The first divergence within this clade at c. 4.1 Ma reflects a biogeographic split between the adjacent South African and Zambezian regions (Fig. 3). Three subsequent movements to (primarily) the Ethiopian + Somalian region follow: one from Southern Africa at
7
c. 3 Ma, giving rise to Empidornis, and two from the Zambezian region, giving rise to Dioptrornis chocolatinus and Melaenornis pammelaina + Melaenornis edolioides (c. 1.4 and 0.8 Ma, respectively). The distributions of Dioptrornis chocolatinus, Melaenornis pammelaina, and Melaenornis edolioides can likely be explained by range expansion and diversification into adjacent biogeographic regions. The distribution of Empidornis (Ethiopian + Somalian) with respect to its closest relatives in Southern Africa is less easy to explain, because none have a distribution in the intervening Zambezian region (Fig. 3). The same pattern is also evident between Bradornis mariquensis (Southern Africa) and its closest relatives in the Ethiopian + Somalian regions (Fig. 3). It is however interesting to note that both of these divergences occur at c. 3 Ma (3.1 and 2.9 Ma, respectively). Here again, it might be possible to link these divergences to Afrotropical forest dynamics (see above). In these cases we would argue that the taxa in question, which occupy arid habitats (i.e., not tropical forest), diverged from one another as the result of tropical forest expansion to coastal Kenya and Tanzania. This would have isolated widely distributed taxa in arid regions to the north (Ethiopian + Somalian) and to the south (Southern African). And, the lack of a distributional presence in the Zambezian region may simply be due to the preference for more arid habitats such as Acacia savannahs, rather than the comparatively less arid habitats (e.g., Miombo woodlands) that occur throughout much of the Zambezian region. Overall, the pattern of diversification in Africa for Muscicapini is exceptional for two reasons. First, the Congolian region is reconstructed as ancestral for the more speciose clade, thus suggesting that the Afrotropical forests were the ancestral habitat for much of Muscicapini, with subsequent movements by a number of lineages into arid habitats; most African forest adapted lineages are restricted to that habitat. Second, several more recent diversifications indicate instances of movement from southern to northern regions, a pattern that appears to be rare. Indeed, we have previously documented this just once for African birds (Voelker et al., 2014), making the instances of south to north colonizations in Muscicapini just the second instance (collectively) that we are aware of. Acknowledgments We thank the collectors and curators that provided samples from their collections (see Appendix). For assistance with taxonomic issues, we thank K.W. Conway. Graphical Abstract pictures from wiki commons and attributed to: Kimmo Simomaa (Empidornis), Tom Tarrant (Muscicapa aquatica), Alan Manson (Melaenornis pammelaina), Lip Kee Yap (Sigelus silens), Ventus55 (Dioptrornis fischeri). This work was supported by NSF DEB-0613668, and by collaborative grants NSF DEB-1120356 and 1119931 (all to G.V. and R.C.K.B.). This is publication number 1510 of the Biodiversity Research and Teaching Collections at Texas A&M University. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ympev.2015.09. 026. References Amer, S.A.M., Kumazawa, Y., 2005. Mitochondrial DNA sequences of the AfroArabian spiny-tailed lizards (genus Uromastyx; family Agamidae): phylogenetic analyses and evolution of gene arrangements. Biol. J. Linn. Soc. 85, 247–260. Brandley, M.C., Schmitz, A., Reeder, T.W., 2005. Partitioned Bayesian analyses, partition choice, and the phylogenetic relationships of scincid lizards. Syst. Biol. 54, 373–390.
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
8
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
Brown, J.M., Lemmon, A.R., 2007. The importance of data partitioning and the utility of Bayes factors in Bayesian phylogenetics. Syst. Biol. 56, 643–655. Cane, M.A., Molnar, P., 2001. Closing of the Indonesian Seaway as a precursor to East African aridification around 3–4 million years ago. Nature 411, 157–162. Cibois, A., Cracraft, J., 2004. Assessing the passerine ‘‘Tapestry”: phylogenetic relationships of the Muscicapoidea inferred from nuclear DNA sequences. Mol. Phylogenet. Evol. 32, 264–273. Cox, G., 1985. The evolution of avian migration systems between temperate and tropical regions of the New World. Am. Nat. 126, 451–474. Dickinson, E.C., Christidis, L. (Eds.), 2014. The Howard & Moore Complete Checklist of the Birds of the World, fourth ed. Aves Press, Eastbourne, UK. Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics and dating with confidence. PLoS Biol. 4, 699–710. Drummond, A.J., Suchard, M.A., Xie, D., Rambaut, A., 2012. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Mol. Biol. Evol. 29, 1969–1973. Feakins, S.J., de Menocal, P.B., Eglinton, T.I., 2005. Biomarker records of late Neogene changes in northeast African vegetation. Geology 33, 977–980. Fjeldså, J., Bowie, R.C.K., 2008. New perspectives on the origin and diversification of Africa’s forest avifauna. Afr. J. Ecol. 46, 235–247. Fuchs, J., Fjeldså, J., Bowie, R.C.K., 2011. Diversification across an altitudinal gradient in the tiny greenbul (Phyllastrephus debilis) from the Eastern Arc Mountains of Africa. BMC Evol. Biol. 11, 117. Hamilton, A.C., Taylor, D., 1991. History of climate and forests in tropical Africa during the last 8 million years. Clim. Change 19, 65–78. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. Jønsson, K.A., Fjeldså, J., 2006. Determining biogeographical patterns of dispersal and diversification in oscine passerine birds in Australia, Southeast Asia and Africa. J. Biogeogr. 33, 1155–1165. Katoh, S., 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780. Lerner, H.R.L., Meyer, M., James, H.F., Hofreiter, M., Fleischer, R.C., 2011. Multilocus resolution of phylogeny and timescale in the extant adaptive radiation of Hawaiian honeycreepers. Curr. Biol. 21, 1838–1844. Linder, H.P., de Klerk, H.M., Born, J., Burgess, N.D., Fjeldså, J., Rahbek, C., 2012. The partitioning of Africa: statistically defined biogeographical regions in subSaharan Africa. J. Biogeogr. 39, 1189–1205. Matzke, N.J., 2013a. Probabilistic Historical Biogeography: New Models for Founder-Event Speciation, Imperfect Detection, and Fossils Allow Improved Accuracy and Model-Testing. Ph.D. Thesis, University of California, Berkeley. Matzke, N.J., 2013b. BioGeoBEARS: BioGeography with Bayesian (and Likelihood) Evolutionary Analysis in R Scripts. CRAN: The Comprehensive R Archive Network, Berkeley, California.. Mayr, E., Cottrell, G.W., 1986. Check-list of Birds of the World, vol. XI. Museum of Comparative Zoology, Cambridge, Massachusetts. Metallinou, M., Arnold, N.E., Crochet, P.A., Geniez, P., Brito, J.C., Lymberakis, P., El Din, S.B., Sindaco, R., Robinson, M., Carranza, S., 2012. Conquering the Sahara and Arabian deserts: systematics and biogeography of Stenodactylus geckos (Reptilia: Gekkonidae). BMC Evol. Biol. 12, 258. Nylander, J.A.A., 2004. MrModeltest, Version 2. Program Distributed by the Author. Evolutionary Biology Centre, Uppsala University, Uppsala. Outlaw, R.K., Voelker, G., Outlaw, D.C., 2007. Molecular systematics and historical biogeography of the Rock-Thrushes (Muscicapidae: Monticola). Auk 124, 561– 577. Pook, C.E., Joger, U., Stümpel, N., Wüster, W., 2009. When continents collide: phylogeny, historical biogeography and systematics of the medically important
viper genus Echis (Squamata: Serpentes: Viperidae). Mol. Phylogenet. Evol. 53, 792–807. Portik, D.M., Papenfuss, T.J., 2012. Monitors cross the Red Sea: the biogeographic history of Varanus yemenensis. Mol. Phylogenet. Evol. 62, 561–565. Portik, D.M., Papenfuss, T.J., 2015. Historical biogeography resolves the origins of endemic Arabian toad lineages (Anura: Bufonidae): evidence for ancient vicariance and dispersal events with the Horn of Africa and South Asia. BMC Evol. Biol. 15, 152. Rambault, A., 2008. FigTree. Distributed by the Author
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies q Gary Voelker a,⇑, Jerry W. Huntley a, Joshua V. Peñalba b, Rauri C.K. Bowie b a b
Department of Wildlife and Fisheries Sciences and Texas Cooperative Wildlife Collections, Texas A&M University, College Station, TX 77843, USA Museum of Vertebrate Zoology and Department of Integrative Biology, 3101 Valley Life Science Building, University of California, Berkeley, CA 94720, USA
a r t i c l e
i n f o
Article history: Received 9 June 2015 Revised 18 September 2015 Accepted 29 September 2015 Available online xxxx Keywords: Africa Eurasia Historical biogeography Muscicapa Muscicapidae Systematics
a b s t r a c t Muscicapa flycatchers and their allies (Bradornis, Dioptornis, Empidornis, Fraseria, Myioparus, Namibornis, and Sigelus) are widely distributed in Africa, Europe and Asia. This broad distribution and the wide variety of habitats occupied by the group, ranging from arid to tropical forests, presents an interesting opportunity to explore the evolution of biogeographic patterns and habitat associations. Sequence data (up to 3310 base pairs from two mitochondrial and two nuclear genes) were generated for 36 of 42 species which comprise the assemblage. Complementary data from an additional species was retrieved from GenBank, as was an additional gene which was available for 21 of our included taxa. Using modelbased phylogenetic methods and molecular clock dating, we constructed a time-calibrated molecular phylogenetic hypothesis for the lineage. Ancestral area reconstructions were performed on the phylogeny using LaGrange and BioGeoBEARS. Our results indicate that Bradornis, Fraseria, and Muscicapa are each non-monophyletic, with the latter being shown to comprise five separate clades each more closely related to other genera. Two new genera (Chapinia and Ripleyia) are erected to account for these results. Muscicapa and allies originated c. 7.4 Ma, most likely in Africa given that their sister lineage is almost entirely from there, and rapidly achieved a Eurasian distribution by c. 7.1 Ma. A second divergence at c. 6.1 Ma resulted in two clades. The first is a largely Eurasian clade that subsequently recolonized Africa, perhaps as the result of the loss of migration. The second is an African clade, and ancestral reconstructions suggest a Congolian (e.g. tropical forest) origin for this clade, with several subsequent diversifications into more arid habitats. This is a unique result, as most tropical forest lineages are confined to that habitat. As with other studies of African bird lineages, Afrotropical forest dynamics appear to have played a significant role in driving diversification in Muscicapa and allies, and our results include just the second recorded case of southern to northern African colonization patterns. Ó 2015 Elsevier Inc. All rights reserved.
1. Introduction Higher-level avian systematic relationships have undergone considerable reshuffling over the past few decades, as molecular phylogenetic analyses have made it comparatively easier to resolve homoplasy resulting from convergent evolution (Wake et al., 2011). The overarching result has tended to be more rigorous formulations of relationships, as indicated by consistency of inferred relationships across studies. This rigor is in marked contrast to older taxonomic work which, as pointed out by Sangster et al. (2010), tended to formulate suprageneric classifications based on often misleading assessments of phenotype and morphology, rather than quantitative assessments of relationships. q
This paper was edited by the Associate Editor Edward Louis Braun.
⇑ Corresponding author. Fax: +1 9798454096.
E-mail address: [email protected] (G. Voelker).
One example of higher level systematic and considerable taxonomic reshuffling centers on the Muscicapidae (Old World chats and flycatchers). This diverse and speciose assemblage has been the subject of a number of molecular phylogenetic studies that have focused primarily on assigning genera to higher level monophyletic groups. Results of these studies include, or confirm, the removal of chats from Turdidae to Muscicapidae, the separation of chats and flycatchers into tribes within Muscicapidae (Saxicolini and Muscicapini), and the erection of a subfamily (Muscicapinae) to house these tribes (e.g., Cibois and Cracraft, 2004; Voelker and Spellman, 2004; Sangster et al., 2010). Other studies have moved genera between these tribes, and to Saxicolini from Turdinae (e.g., Voelker and Spellman, 2004) and have erected new tribes to accommodate still other monophyletic phylogenetic groups (e.g., Sangster et al., 2010). Suprageneric relationships are not the only systematic and taxonomic issues in Muscicapidae. For example, in their extensive
http://dx.doi.org/10.1016/j.ympev.2015.09.026 1055-7903/Ó 2015 Elsevier Inc. All rights reserved.
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
2
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
molecular assessment of that family, Sangster et al. (2010) demonstrated that the genus Melaenornis (Muscicapini) is not monophyletic, and indeed taxonomic treatments have disagreed on the limits of ‘‘Melaenornis” by including some taxa from that genus in other genera (e.g., Mayr and Cottrell, 1986; Taylor, 2006; Dickinson and Christidis, 2014). Thus, while we have a better idea of suprageneric relationships in Muscicapidae, we are still working towards assigning species membership to genera with confidence, and the relationships of species within those genera. This has also impacted other types of analyses (e.g., historical biogeography studies) which rely on well-resolved phylogenies to better understand the evolution of the group. Indeed, the intercontinental distribution of Muscicapini, the diversity of habitats collectively occupied by members of the tribe and the presence of migratory behavior in a small subset of species combine to suggest that interesting biogeographic scenarios are likely. We focus here on the systematics and biogeography of the Muscicapini: Old World chats and flycatchers. This is a diverse assemblage of about 37 species, which collectively have widespread distributions centered in Africa and Eurasia (Taylor, 2006). Our goals are as follows: (1) to resolve the taxonomic issues related to genus-level relationships and genus membership by species, and (2) to assess biogeographic patterns in the group. Based on the distributional patterns of most Muscicapini (as well as the African ancestral area of its sister clade, the Copsychini; Voelker et al., 2014), we expect that: (1) Eurasian taxa will be derived from African ancestors; (2) colonization of Eurasia will be attributable to expansion through the Saudi Peninsula, and not overwater dispersals as we found in Copsychini (Voelker et al., 2014); and (3) gains or losses of migratory behavior by Eurasian taxa will be relevant to explaining intercontinental distribution patterns.
2. Material and methods 2.1. Taxon sampling and sequencing Primarily following Dickinson and Christidis (2014; exceptions noted below) our analyses included 37 Muscicapini species representing the following genera: Muscicapa (19 of 23 species), Melaenornis (3/4), Myioparus (2/2), Namibornis (1/1), Dioptrornis (3/3; all considered Melaenornis by Dickson and Christidis), Empidornis (1/1), Fraseria (2/2), Sigelus (1/1) and Bradornis (4/4) (Supplementary Table S1). We note here that the last five of these genera, until recently, have been considered Melaenornis (e.g., Mayr and Cottrell, 1986; Taylor, 2006), but that genus has been rendered non-monophyletic in a recent molecular systematic study (Sangster et al., 2010), which included a subset of the Muscicapini species (n = 21) included here as part of our broader sampling. Zuccon and Ericson (2010) similarly found Melaenornis to be non-monophyletic, but included just nine Muscicapini species. We did not include the monotypic Humblotia, which Dickinson and Christidis (2014) place as a Muscicapini species (but see Jønsson and Fjeldså, 2006), but we did include Namibornis, which is not a member of Copsychini (see Mayr and Cottrell, 1986, and results below) as previously suggested. As outgroup taxa, we included a suite of Erythropygia and Copsychus species which, as members of Copsychini, form the sister clade to Muscicapini (Sangster et al., 2010; Voelker et al., 2014). Whole genomic DNA was extracted from tissue or toepads using the DNeasy tissue extraction kit (Qiagen, Valencia, CA, USA). We used the polymerase chain reaction (PCR) to amplify the mitochondrial cytochrome-b (CYB) and NADH dehydrogenase subunit 2 (ND2) genes, as well as the autosomal nuclear Myoglobin (MB intron 2) and Beta-Fibrinogen (FGB intron 5) genes. Standard primers and reaction conditions were employed. PCR products
were cleaned with ExoSAP-IT (Affymetrix, Santa Clara, CA, USA). Sanger sequencing was performed at the Beckman Coulter Genomics facility (Danvers, MA, USA). The mtDNA sequences were aligned using Sequencher v. 4.9 (Gene Codes Corporation, Ann Arbor, MI, USA). Introns were aligned using MAFFT (Katoh, 2013). 2.2. Phylogenetic analyses and divergence dating To determine the best-fit model(s) for the mitochondrial (mtDNA) sequence data, we assessed three alternate partitioning schemes, each of which relied on appropriate partition models derived from MrModelTest (Nylander, 2004). In our first partitioning scheme (two partitions) the ND2 and CYB genes were unlinked. In the second (four partitions), first and second codon positions were linked for ND2, linked for CYB, and third codon positions for each gene were treated as independent partitions. In the third scheme (six partitions), each codon position was unlinked across both genes. For each mixed-model partition scheme, we used MRBAYES (Huelsenbeck and Ronquist, 2001) to initiate two runs of four Markov-chain Monte Carlo (MCMC) chains of 5,000,000 generations, each starting from a random tree and sampling every 100 generations. Each run resulted in 50,000 trees and converged on the same topology. The first 5000 trees from each analysis were removed as ‘‘burn-in”, and the remaining 90,000 trees were used to generate a majority rule consensus tree. Bayes factors were computed using the harmonic means of the likelihoods calculated from the sump command within MRBAYES. A difference of 2 ln Bayes factor >10 was used as the minimum value to discriminate between mixed-model partitioning schemes (Brandley et al., 2005; Brown and Lemmon, 2007), and the six partition scheme was identified as the best-fit to the mitochondrial data. We then used the sixpartition mtDNA scheme in combination with the nuclear gene data. Best-fit models for nuclear data were also derived from MrModelTest analyses. This combined analysis set each nuclear gene and each mtDNA codon position as unlinked in two runs of four MCMC chains in MRBAYES as described above. We also incorporated best fit models for each gene in BEAST 2.0 (Drummond et al., 2006, 2012) to reconstruct a tree using all loci, and simultaneously estimated divergence times across all species. We employed a lineage substitution rate of 0.014 per lineage/ million years for CYB using a relaxed, uncorrelated lognormal clock. This substitution rate translates to 2.8% per million years, and is generally applicable to the CYB gene in songbirds (Weir and Schluter, 2008; Lerner et al., 2011). We used a rate of 0.029 per lineage/million years for ND2 (Lerner et al., 2011). We applied a slightly broader prior (0.002) for FGB than the 0.0017 rate used by Lerner et al. (2011), to reflect the greater variation of rates among introns. We applied the same 0.002 rate to MB. Standard deviations for the mitochondrial genes (CYB: 0.001; ND2: 0.0025) followed Lerner et al. (2011), while we used a 0.002 standard deviation for the introns. A Yule process speciation prior was implemented in each analysis. Two separate MCMC analyses were run for 10,000,000 generations with parameters sampled every 1000 steps, with a conservative 20% burn-in. Independent runs were combined using LogCombiner v.1.6.1 (Drummond et al., 2012). Tracer v.1.5 (Rambault and Drummond, 2007) was used to measure the effective sample size of each (all >200) and calculate the mean and upper and lower bounds of the 95% highest posterior density interval (95% HPD) for divergence times. Tree topologies were assessed using TreeAnnotator v.1.7 (Rambault and Drummond, 2007; Drummond et al., 2012) and FigTree v.1.3.1 (Rambault, 2008). Finally, we also performed the above analyses on a five gene dataset. These analyses included the ODC1 gene from 21 ingroup taxa for which data was available on GenBank. Nexus trees of both datasets are available as Supplementary material in the online version.
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
3
2.3. Historical biogeography
3.2. Divergence dating and biogeographic history
For biogeographic analysis of the entire ingroup we used likelihood analysis of geographic range evolution (dispersal-extinction cladogenesis, DEC) implemented in LaGrange v. 2.0.1 (Ree and Smith, 2008). The DEC model as implemented in LaGrange has two free parameters specifying the rate of ‘‘dispersal” (i.e., range expansion) and ‘‘extinction” (i.e., range contraction), but the cladogenesis model remains fixed. This has an important consequence because the geographical range of the ancestral lineage is inherited by the two daughter lineages through a variety of plausible cladogenetic scenarios (e.g., sympatric, parapatric, vicariance) that have equal probability. We also made use of a new method implemented in the BioGeoBEARS package in the R statistical program (Matzke, 2013a,b) that adds a third free parameter to the DEC framework: long-distance dispersal (parameter j DEC+j model). This additional parameter enables one daughter lineage to disperse to an area outside (i.e. not adjoining) the ancestral range, thereby effectively mimicking the process of founder-event speciation. Because the classic DEC model of LaGrange is nested within the DEC+j model implemented in BioGeoBEARS, we are able to perform model comparison (DEC vs. DEC+j) with the Likelihood Ratio Test and by using AIC weights. We followed the statistically defined African sub-regions of Linder et al. (2012) for birds, and using distribution maps in Taylor (2006) categorized species as being present or absent in the following five areas or area combinations: Southern African, Zambezian, Congolian, Ethiopian + Somalian, and Sudanian. A sixth area, Europe + Asia, was included to account for the distribution of non-African taxa.
Molecular clock calibration sets the origin of the Muscicapa and allies clade to c. 7.4 Ma (Fig. 2). Within the clade, Muscicapa ruficauda diverged from a clade representing all other taxa at c. 7.1 Ma (Fig. 2). Ancestral reconstructions using LaGrange ( ln value of 118.6) suggested a Muscicapa and allies ancestral area comprising all defined areas, whereas the BioGeoBEARS DEC model ( ln value of 119.7) suggested a Europe + Asian ancestral area (Fig. 3, and see Fig. S1 for an alternative depiction of BioGeoBEARS output). The DEC + j model in BioGeoBEARS was not significantly better ( ln 111.2; P = 0.079). The next lineage divergence at c. 6.1 Ma resulted in two main clades. The first of these is the Muscicapa striata clade which comprises 11 species, six of which have exclusively Europe + Asia distributions. Europe + Asia is reconstructed by both methods as the ancestral area for this clade. Within this clade, there are two subsequent colonizations of Africa resulting in five taxa, and a back colonization of Europe + Asia by one of those taxa (Muscicapa gambagae; Fig. 3). All but one of the divergences in this clade are dated in the Pliocene (Fig. 2). The second main clade comprises all other ingroup taxa. The Congolian region is reconstructed by both methods as the ancestral area for the clade (Fig. 3). Several subsequent, rapid divergences in the latest Miocene/earliest Pliocene are also biogeographically reconstructed as Congolian, or include that area in the most likely reconstruction (Figs. 2 and 3). All species in the Fraseria ocreata–Fraseria cinerascens clade occur in the Congolian region, with several having achieved slightly broader (Myioparus griseigularis) or very broad distributions (Myioparus plumbeus and Muscicapa caerulescens). Other main divergences coincide with colonizations of the Southern African, Zambezian and Ethiopian + Somalian regions by the Empidornis–Meleanornis edolioides clade, of those same regions plus the Sudanian region by Bradornis infuscatus + B. pallidus, and of the Southern African, Zambezian and Ethiopian + Somalian regions by the Bradornis mariquensis–Muscicapa comitata clade (Fig. 3). Most divergences in these four subclades occur in the Pliocene, with just three having occurred more recently in the Pleistocene (Fig. 2).
3. Results 3.1. Phylogenetic analyses We sequenced up to a total of 3310 bp from four gene regions, for each individual sampled. Complete data from each gene was generated from most taxa, or retrieved from GenBank (Appendix). In a few instances, the sequences retrieved for a given taxon were not from the same individual (Appendix). Across all ingroup taxa (Fig. 1) there were 1210 variable sites of which 808 were parsimony informative. Bayesian analysis of the combined four gene dataset resulted in a phylogeny with mixed nodal support, where 25 of 36 nodes (69%) received strong support of P0.95 posterior probability and 11 of 36 nodes (31%) received moderate to weak support (Fig. 1). The addition of the ODC gene increased support values at just two nodes (Fig. 1). Overall, most nodes receiving moderate to poor support were basal in the phylogeny. Our BEAST analysis of the four gene dataset resulted in a phylogeny with similarly mixed support, where just 21 nodes (58%) were strongly supported (Fig. 2). There are six nodes in the species tree that conflict with the concatenated Bayesian topology; three of these conflicting nodes are strongly supported in the Bayesian phylogeny (Fig. 1). However, there are also two strongly supported nodes in the tree that were less well supported in the Bayesian concatenated phylogeny: Dioptrornis chocolatinus is supported as sister to D. brunneus + D. fischeri, and Muscicapa comitata is supported as sister to the Bradornis pallidus clade (Fig. 2). Across both analysis methods then, 81% of nodes received posterior probability support of P0.95. It is clear from our results that the taxonomy of this group requires revision. Fraseria (two species) is not monophyletic, nor is Bradornis (five species in two clades). The most obvious case of non-monophyly is evident in Muscicapa, which is composed of five distinct lineages across the phylogeny (Fig. 1).
4. Discussion 4.1. Molecular phylogeny and taxonomic implications Using molecular data from 20 taxa, Sangster et al. (2010) defined Muscicapini as comprising the genera Muscicapa, Melaenornis and, Fraseria; they did not include Myioparus in their analysis but it is clearly also a member of Muscicapini (Fig. 1). Other genera ascribed to Muscicapini historically were excluded by their results, which also indicated issues with the monophyly of each of the remaining three genera (Muscicapa, Melaenornis, and Fraseria; Sangster et al., 2010). Our study represents the most extensively sampled molecular phylogenetic analysis of Muscicapini species to date. Overall, we included 37 of 42 Muscicapini species (see Section 2). Of the five taxa we are missing in our analyses, two are insular forms from Southeast Asia (Muscicapa randi – Philippines and Muscicapa segregata – Lesser Sundas), and both have been considered as conspecific with Muscicapa dauurica (Taylor, 2006). The remaining three species have patchy or restricted ranges in Afrotropical rainforests (Muscicapa lendu, Muscicapa epulata, Melaenornis annamarulae). Our attempts to extract DNA from museum specimens of Muscicapa lendu and Muscicapa epulata were unsuccessful. Although comprising just four species in recent taxonomy (e.g., Taylor, 2006), the genus Melaenornis has previously been considered to include species now ascribed to Empidornis, Sigelus,
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
4
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
*
0.59
Sigelus Namibornis
*
*
* *
0.74
Dioptrornis chocolatinus Dioptrornis brunneus Dioptrornis fischeri Melaenornis ardesiacus
*
0.91 (0.95)
Melaenornis pammelaina Melaenornis edolioides Fraseria ocreata
0.63
*
0.64
Myioparus griseigularis Myioparus plumbeus Muscicapa olivascens (Apatema)
* *
0.90
Empidornis
*
Muscicapa tessmani Muscicapa caerulescens
(Cichlomyia or Butalis)
Fraseria cinerascens (Chapinia)
*
0.87
Bradornis infuscatus Bradornis pallidus
Bradornis mariquensis
*
0.55
(Haganopsornis)
*
Bradornis pumilus Bradornis microshynchus Muscicapa comitata (Pedilorhynchus)
0.90
*
*
*
*
*
* * *
*
(Artomyias)
(Bradyornis)
Muscicapa boehmi
0.82
*
Muscicapa ussheri Muscicapa infuscata Muscicapa striata
Muscicapa gambagae
0.94 (0.97)
*
*
Muscicapa aquatica Muscicapa cassini Muscicapa adusta Muscicapa sethsmithi Muscicapa daaurica Muscicapa muttui Muscicapa sibirica Muscicapa ferruginea Muscicapa griseisticta Muscicapa ruficauda
(Ripleyia)
Fig. 1. Molecular phylogeny of Muscicapa and allies based on Bayesian analysis of concatenated data (two mitochondrial and two nuclear genes). Asterisks denote Bayesian posterior probabilities P0.95, with lower posterior probability values shown. Numbers in parentheses are posterior probabilities derived from a reduced taxa five gene analysis (see Section 2), which provide stronger support for several relationships than support values returned from the four gene analysis. Note that several genera are rendered non-monophyletic in these analyses, necessitating nomenclatural revisions at the genus level. We elaborate on these revisions elsewhere (see text), but parenthetically note suggested alternative generic designations on the phylogeny.
Dioptrornis, Bradornis, and Fraseria (Mayr and Cottrell, 1986). Our results indicate, (1) that a larger Melaenornis (to include the aforementioned four genera) would be non-monophyletic (see also Sangster et al., 2010), (2) that Melaenornis (edolioides is the type) could be restricted to as little as three and perhaps four species (depending on the eventual systematic placement of annamarulae), and (3) that Melaenornis could be expanded to include Dioptrornis, Empidornis, Sigelus and Namibornis (Figs. 1 and 2). Due to the morphological distinctiveness of the latter four genera (three of which are monotypic) relative to Melaenornis, which are all black or dark gray in color, we agree with the more strict usage of Melaenornis (e.g., Taylor, 2006).
Our results also indicate that Bradornis is non-monophyletic, with species falling into two distinct clades (Figs. 1 and 2). The Bradornis type is mariquensis (Mayr and Cottrell, 1986), and thus that genus should be applied to mariquensis, pumilus and microrhyncus. There is an available synonym, Haganopsornis, which was applied to infuscatus (Roberts, 1922 fide Mayr and Cottrell, 1986) and we suggest resurrecting that genus to include infuscatus and pallidus (Figs. 1 and 2). The genus Fraseria is also nonmonophyletic, with ocreata being more closely related to the genus Myioparus; our results conflict as to the phylogenetic position of cinerascens (Figs. 1 and 2). Regardless, Fraseria would apply to ocreata (Mayr and Cottrell, 1986). We find no synonym to apply
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
5
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
0.39
*
*
*
*
*
0.85
*
0.93 0.38
*
0.44 0.82
*
Fraseria cinerascens 0.90
Muscicapa olivascens Muscicapa tessmani Muscicapa caerulescens Bradornis infuscatus Bradornis pallidus Bradornis mariquensis Bradornis pumilus Bradornis microshynchus Muscicapa comitata Muscicapa ussheri Muscicapa infuscata Muscicapa boehmi Muscicapa gambagae Muscicapa striata Muscicapa aquatica Muscicapa cassini Muscicapa adusta Muscicapa sethsmithi Muscicapa daaurica Muscicapa muttui Muscicapa sibirica Muscicapa ferruginea Muscicapa griseisticta Muscicapa ruficauda
* *
* 0.88
*
*
*
*
0.91
0.43
0.69
*
* *
*
0.43
*
*
*
0.78 0.42
Miocene 7.0
6.0
Pliocene 5.0
Empidornis Namibornis Sigelus Dioptrornis chocolatinus Dioptrornis brunneus Dioptrornis fischeri Melaenornis ardesiacus Melaenornis pammelaina Melaenornis edolioides Fraseria ocreata Myioparus griseigularis Myioparus plumbeus
4.0
3.0
Pleistocene 2.0
1.0
0.0
Time in Ma Fig. 2. BEAST and molecular clock analysis based on the four-gene dataset. Asterisks denote Bayesian posterior probabilities P0.95, with lower posterior probability values shown. Arrows at nodes indicate points of conflict with the concatenated Bayesian analysis (Fig. 1). Molecular clock rates were based on rates derived by Lerner et al. (2011; see Section 2). Bars at nodes indicate 95% highest posterior density intervals. Time is millions of years before present.
to cinerascens, which then requires the designation of a new genus for that species which we propose as: Chapinia, new genus Voelker & Bowie Type species. – Chapinia cinerascens. Diagnosis – A genus of muscicapid flycatcher differing from all other genera of the family Muscicapidae by the following combination of characters: large size, diagnostic white supra-loral spot, dark upperparts, and mottled gray underparts with dark but poorly demarcated crescents on the breast. Etymology – This name honors Dr. James P. Chapin, for his extensive documentation of, and research on, the birds of the Belgian Congo. Finally, the genus Muscicapa appears to have been a taxonomic dumping ground for any small to medium sized Muscicapini flycatcher, as our results show it to be comprised of five distinct lineages (Fig. 1). The type for the genus is striata, thus Muscicapa would apply to the large clade of 11 species (Fig. 1). There are several synonyms available for other clades. The genus Apatema could be applied to olivascens, and Cichlomyia or Butalis (it is unclear to us which has priority) could apply to the closely related caerulescens, and thus to tessmani as well (Fig. 1). For comitata, the genus Pedilorhynchus is available, and Artomyias is available for infuscata
and thus also for the closely related ussheri (Fig. 1). Although sister to infuscata + ussheri, boehmi is highly distinct from them morphologically (Sinclair and Ryan, 2010). We therefore suggest applying the name Bradyornis to boehmi, following the original description of this species (Reichenow, 1884, fide Mayr and Cottrell, 1986). We find no available synonym for ruficauda, which is the first species to diverge within Muscicapini (Fig. 1). This requires the designation of a new genus for that species which we propose as: Ripleyia, new genus Voelker & Bowie Type species. – Ripleyia ruficauda. Diagnosis – A genus of muscicapid flycatcher differing from all other genera of the family Muscicapidae by the following combination of characters: rufous uppertail-coverts and tail, faint supercilium, and entirely orange lower mandible. Etymology – This name honors Dr. S. Dillon Ripley, former Secretary of the Smithsonian Institution, for his extensive work on the birds of India and southern Asia. 4.2. Biogeographic history – Intercontinental movements Our results indicate three movements between Africa and Eurasia. The first of these assumes movement from Africa to
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
6
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
SA SA SA Ethiopian + Somalian
Sudanian
Congolian Zambezian
CO ZSACO CO CO CO
Southern African
Z Z Z
Z Z Z
CO CO
CO CO CO
Z Dioptrornis brunneus Dioptrornis fischeri Z Melaenornis ardesiacus
Z
ESOZSACO
ESOZSACO
Melaenornis pammelaina
ESOSU
Meleanornis edolioides CO Fraseria ocreata
ZCO Myioparus griseigularis CO ESOSUZSACO CO ESOSUZSACO Myioparus plumbeus
CO CO CO
CO CO
ESOZ
Empidornis semipartitus
SA Sigelus silens Namibornis herero ESO Dioptrornis chocolatinus
SA
SA ESOZSA Z Z Z Z
ESO
ESOSA
CO
CO CO CO
Muscicapa olivascens CO Muscicapa tessmani
ESOZSACO CO
Muscicapa caerulescens
Fraseria cinerascens
ESOZSACO
CO CO
ESOSUZCOSAEAs EAs
SA ESOSUZSACO ESOSUZCO
CO
CO CO
SA ESOSA ESO
CO CO CO ESOSACO
CO Z
Bradornis infuscatus Bradornis pallidus SA Bradornis mariquensis
ESO Bradornis pumilus Bradornis microrhynchus CO Muscicapa comitata CO Muscicapa ussheri
ESO
CO CO CO
ZCO
SA
ESOSUZCO
Z
ESOSACOEAs
ESO
CO Muscicapa infuscata Muscicapa boehmi EAs
EAs SUEAs EAs
SUEAs Muscicapa gambagae Muscicapa aquatica ZCO Muscicapa cassini ESOSUZSACO Muscicapa adusta
SU
EAs
EAs EAs
EAs EAs
EAs EAs EAs
SU
SU+EAs EAs+CO SUZCO SU CO SUCOEAs SU CO CO ESOSUZSACO EAs CO
EAs
EAs EAs EAs EAs EAs EAs
Muscicapa striata
CO EAs
Muscicapa sethsmithi
Muscicapa daaurica EAs Muscicapa muttui EAs
Muscicapa sibirica Muscicapa ferruginea EAs Muscicapa griseisticta EAs
EAs
Muscicapa ruficauda
Fig. 3. Ancestral area reconstructions based on the DEC model implemented in BioGeoBEARS (reconstructions to the right of nodes), and LaGrange (reconstructions to the left of nodes). LaGrange returns two reconstructions, thus the upper reconstruction is the most likely area reconstruction derived from all node reconstructions from lineages ‘‘above” the node, and the lower is the most likely reconstruction derived from all node reconstructions from lineages ‘‘below” the node. Areas are defined as follows: Southern African (SA), Zambezian (Z), Congolian (CO), Ethiopian + Somalian (ESO), Sudanian (SU) and Europe + Asia (EAs). The coded distribution for each species is indicated next to that species’ name.
Eurasia c. 7.4 Ma to explain the Eurasian distribution of Muscicapa ruficauda (Figs. 2 and 3). We assume this direction of movement given that the basal taxa in the sister clade (Erythropygia and allies) to Muscicapini are almost entirely of African origin (Voelker et al., 2014). A second Africa to Eurasian movement at c. 6.8 Ma is required to explain the basal Eurasian ancestral reconstructions for the Muscicapa striata clade (Figs. 2 and 3). Finally, movement from Eurasia to Africa at c. 4.7 Ma is required to explain the distributions of five African taxa imbedded in the otherwise Eurasian Muscicapa striata clade (Figs. 2 and 3). Given the distribution of Muscicapa ruficauda in Central Asia, and the broader Eurasian (i.e., both European and Asian) distributions of many of the taxa in the Muscicapa striata clade, we assume land-based colonization from Africa, rather than the repeated overwater dispersals which we invoked to explain African-East Asian connections in the sister group to Muscicapini (Erythropygia and allies; Voelker et al., 2014). Land-based movements are also a rea-
sonable conclusion for the recolonization of Africa by this clade, as several internal ancestral reconstructions include the Sudanian or Ethiopian + Somalian regions (Fig. 3) which are geographically adjacent to the Saudi Peninsula, and thus Eurasia generally. Further, the dates of these movements are during a period (10– 5.3 Ma) when an Afro-Arabian landbridge was in place. This connection allowed for extensive bi-directional dispersal of vertebrate lineages between the Horn of Africa and the Arabian Peninsula (Amer and Kumazawa, 2005; Pook et al., 2009; Wong et al., 2010; Metallinou et al., 2012; Portik and Papenfuss, 2012; Šmíd et al., 2013). Palearctic-derived lineages also began to colonize the Arabian Peninsula during this time (Portik and Papenfuss, 2015), suggesting large scale habitat changes across the region. These land based movements may also be related to evolution (gains or losses) of migration. The initial colonization of Eurasia from Africa could be explained via the gain of migratory behavior (the sister clade is non-migratory), by Muscicapa ruficauda (or more
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
likely an ancestor) which breeds in the Himalayas but winters in India (Taylor, 2006). The second colonization of Eurasia from Africa, and the recolonization of Africa from Eurasia could be related to the gain and subsequent loss of migration in the Muscicapa striata clade. All Eurasian taxa in this clade (Fig. 3) are migratory, and all but one winter in India or Southeast Asia; the exception is striata which winters in Africa (Taylor, 2006). This would suggest that the striata clade (and an ancestor of Muscicapa ruficauda) expanded to Eurasia to exploit suitable breeding habitat, and subsequently gained migratory behavior in response to winter climates (e.g., the southern home hypothesis of Cox, 1985). A subsequent loss of migratory behavior coincides with the recolonization of and subsequent diversification by African members of the striata clade (Fig. 3). The biogeographic connection that striata has with Africa (wintering range) and its close relationship to the African taxa implies a migratory drop-off pattern similar to what has been shown in other studies of avian genera with species distributed in both Africa and Eurasia (e.g., Voelker et al., 2009; Voelker and Light, 2011). 4.3. Biogeographic history – intra-African patterns After the initial recolonization of Africa at c. 4.7 Ma, the African members of the striata clade began to diversify at c. 3.6 Ma. Both of these dates coincide with a period from 5–3 Ma, when global climate changes caused significant expansions of the Afrotropical forests into Kenya, Tanzania, and Ethiopia (Hamilton and Taylor, 1991; Cane and Molnar, 2001; Feakins et al., 2005; Sepulchre et al., 2006), which, in turn, had a significant role in generating avian diversity, to include later diversifications during forest retractions (e.g., Outlaw et al., 2007; Fjeldså and Bowie, 2008; Voelker and Outlaw, 2008; Voelker et al., 2010, 2012; Fuchs et al., 2011). The Congolian region is reconstructed as the ancestral area for several basal nodes in the other major clade of Muscicapini (Fig. 3), which is comprised of three subclades (and Bradornis infuscatus + Bradornis pallidus). Two of these subclades, the Fraseria cinerascens–Fraseria ocreata clade (seven species) and the Muscicapa boehmi-Bradornis mariquensis (seven species) clade, are inferred to have a Congolian ancestral area (Fig. 3). The distribution of the Fraseria cinerascens–Fraseria ocreata clade can be explained via expansion from the Congolian region into adjacent areas primarily during the Pliocene (Figs. 2 and 3). In the Muscicapa boehmi–Bradornis mariquensis subclade, the divergence between the Muscicapa taxa (four species) and the Bradornis taxa (three species) is inferred to have occurred at c. 4.7 Ma, indicating another divergence that could be linked to Afrotropical forest dynamics (see above). The subsequent divergences of Bradornis lineages began at c. 2.9 Ma, as the tropical forest was retracting westward. This date, and the fact that the Congolian region and Ethiopian + Somalian regions are not adjacent (the Sudanian and Zambezian regions are between them) suggest that Bradornis pumilus and B. microrhynchus are the result of isolation due to forest retraction. Indeed, this divergence date matches well with other taxa inferred to have been the result of forest retraction and subsequent isolation in eastern Africa (Voelker et al., 2010). We discuss the anomalous distribution of Bradornis mariquensis in Southern Africa below. The ancestral area at the base of the remaining clade (Melearnornis edolioides–Empidornis; nine species) is reconstructed as South African + Zambezian (Fig. 3). This clade diverged from the Congolian Fraseria cinerascens–Fraseria ocreata clade at c. 5.2 Ma. The first divergence within this clade at c. 4.1 Ma reflects a biogeographic split between the adjacent South African and Zambezian regions (Fig. 3). Three subsequent movements to (primarily) the Ethiopian + Somalian region follow: one from Southern Africa at
7
c. 3 Ma, giving rise to Empidornis, and two from the Zambezian region, giving rise to Dioptrornis chocolatinus and Melaenornis pammelaina + Melaenornis edolioides (c. 1.4 and 0.8 Ma, respectively). The distributions of Dioptrornis chocolatinus, Melaenornis pammelaina, and Melaenornis edolioides can likely be explained by range expansion and diversification into adjacent biogeographic regions. The distribution of Empidornis (Ethiopian + Somalian) with respect to its closest relatives in Southern Africa is less easy to explain, because none have a distribution in the intervening Zambezian region (Fig. 3). The same pattern is also evident between Bradornis mariquensis (Southern Africa) and its closest relatives in the Ethiopian + Somalian regions (Fig. 3). It is however interesting to note that both of these divergences occur at c. 3 Ma (3.1 and 2.9 Ma, respectively). Here again, it might be possible to link these divergences to Afrotropical forest dynamics (see above). In these cases we would argue that the taxa in question, which occupy arid habitats (i.e., not tropical forest), diverged from one another as the result of tropical forest expansion to coastal Kenya and Tanzania. This would have isolated widely distributed taxa in arid regions to the north (Ethiopian + Somalian) and to the south (Southern African). And, the lack of a distributional presence in the Zambezian region may simply be due to the preference for more arid habitats such as Acacia savannahs, rather than the comparatively less arid habitats (e.g., Miombo woodlands) that occur throughout much of the Zambezian region. Overall, the pattern of diversification in Africa for Muscicapini is exceptional for two reasons. First, the Congolian region is reconstructed as ancestral for the more speciose clade, thus suggesting that the Afrotropical forests were the ancestral habitat for much of Muscicapini, with subsequent movements by a number of lineages into arid habitats; most African forest adapted lineages are restricted to that habitat. Second, several more recent diversifications indicate instances of movement from southern to northern regions, a pattern that appears to be rare. Indeed, we have previously documented this just once for African birds (Voelker et al., 2014), making the instances of south to north colonizations in Muscicapini just the second instance (collectively) that we are aware of. Acknowledgments We thank the collectors and curators that provided samples from their collections (see Appendix). For assistance with taxonomic issues, we thank K.W. Conway. Graphical Abstract pictures from wiki commons and attributed to: Kimmo Simomaa (Empidornis), Tom Tarrant (Muscicapa aquatica), Alan Manson (Melaenornis pammelaina), Lip Kee Yap (Sigelus silens), Ventus55 (Dioptrornis fischeri). This work was supported by NSF DEB-0613668, and by collaborative grants NSF DEB-1120356 and 1119931 (all to G.V. and R.C.K.B.). This is publication number 1510 of the Biodiversity Research and Teaching Collections at Texas A&M University. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ympev.2015.09. 026. References Amer, S.A.M., Kumazawa, Y., 2005. Mitochondrial DNA sequences of the AfroArabian spiny-tailed lizards (genus Uromastyx; family Agamidae): phylogenetic analyses and evolution of gene arrangements. Biol. J. Linn. Soc. 85, 247–260. Brandley, M.C., Schmitz, A., Reeder, T.W., 2005. Partitioned Bayesian analyses, partition choice, and the phylogenetic relationships of scincid lizards. Syst. Biol. 54, 373–390.
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026
8
G. Voelker et al. / Molecular Phylogenetics and Evolution xxx (2015) xxx–xxx
Brown, J.M., Lemmon, A.R., 2007. The importance of data partitioning and the utility of Bayes factors in Bayesian phylogenetics. Syst. Biol. 56, 643–655. Cane, M.A., Molnar, P., 2001. Closing of the Indonesian Seaway as a precursor to East African aridification around 3–4 million years ago. Nature 411, 157–162. Cibois, A., Cracraft, J., 2004. Assessing the passerine ‘‘Tapestry”: phylogenetic relationships of the Muscicapoidea inferred from nuclear DNA sequences. Mol. Phylogenet. Evol. 32, 264–273. Cox, G., 1985. The evolution of avian migration systems between temperate and tropical regions of the New World. Am. Nat. 126, 451–474. Dickinson, E.C., Christidis, L. (Eds.), 2014. The Howard & Moore Complete Checklist of the Birds of the World, fourth ed. Aves Press, Eastbourne, UK. Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics and dating with confidence. PLoS Biol. 4, 699–710. Drummond, A.J., Suchard, M.A., Xie, D., Rambaut, A., 2012. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Mol. Biol. Evol. 29, 1969–1973. Feakins, S.J., de Menocal, P.B., Eglinton, T.I., 2005. Biomarker records of late Neogene changes in northeast African vegetation. Geology 33, 977–980. Fjeldså, J., Bowie, R.C.K., 2008. New perspectives on the origin and diversification of Africa’s forest avifauna. Afr. J. Ecol. 46, 235–247. Fuchs, J., Fjeldså, J., Bowie, R.C.K., 2011. Diversification across an altitudinal gradient in the tiny greenbul (Phyllastrephus debilis) from the Eastern Arc Mountains of Africa. BMC Evol. Biol. 11, 117. Hamilton, A.C., Taylor, D., 1991. History of climate and forests in tropical Africa during the last 8 million years. Clim. Change 19, 65–78. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. Jønsson, K.A., Fjeldså, J., 2006. Determining biogeographical patterns of dispersal and diversification in oscine passerine birds in Australia, Southeast Asia and Africa. J. Biogeogr. 33, 1155–1165. Katoh, S., 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780. Lerner, H.R.L., Meyer, M., James, H.F., Hofreiter, M., Fleischer, R.C., 2011. Multilocus resolution of phylogeny and timescale in the extant adaptive radiation of Hawaiian honeycreepers. Curr. Biol. 21, 1838–1844. Linder, H.P., de Klerk, H.M., Born, J., Burgess, N.D., Fjeldså, J., Rahbek, C., 2012. The partitioning of Africa: statistically defined biogeographical regions in subSaharan Africa. J. Biogeogr. 39, 1189–1205. Matzke, N.J., 2013a. Probabilistic Historical Biogeography: New Models for Founder-Event Speciation, Imperfect Detection, and Fossils Allow Improved Accuracy and Model-Testing. Ph.D. Thesis, University of California, Berkeley. Matzke, N.J., 2013b. BioGeoBEARS: BioGeography with Bayesian (and Likelihood) Evolutionary Analysis in R Scripts. CRAN: The Comprehensive R Archive Network, Berkeley, California.
viper genus Echis (Squamata: Serpentes: Viperidae). Mol. Phylogenet. Evol. 53, 792–807. Portik, D.M., Papenfuss, T.J., 2012. Monitors cross the Red Sea: the biogeographic history of Varanus yemenensis. Mol. Phylogenet. Evol. 62, 561–565. Portik, D.M., Papenfuss, T.J., 2015. Historical biogeography resolves the origins of endemic Arabian toad lineages (Anura: Bufonidae): evidence for ancient vicariance and dispersal events with the Horn of Africa and South Asia. BMC Evol. Biol. 15, 152. Rambault, A., 2008. FigTree. Distributed by the Author
Please cite this article in press as: Voelker, G., et al. Resolving taxonomic uncertainty and historical biogeographic patterns in Muscicapa flycatchers and their allies. Mol. Phylogenet. Evol. (2015), http://dx.doi.org/10.1016/j.ympev.2015.09.026