Phylogeny of eagles, Old World vultures, and other Accipitridae based on nuclear and mitochondrial DNA

https://doi.org/10.1016/j.ympev.2005.04.010Get rights and content

Abstract

We assessed phylogenetic relationships for birds of prey in the family Accipitridae using molecular sequence from two mitochondrial genes (1047 bases ND2 and 1041 bases cyt-b) and one nuclear intron (1074 bases β-fibrinogen intron 7). We sampled representatives of all 14 Accipitridae subfamilies, focusing on four subfamilies of eagles (booted eagles, sea eagles, harpy eagles, and snake eagles) and two subfamilies of Old World vultures (Gypaetinae and Aegypiinae) with nearly all known species represented. Multiple well-supported relationships among accipitrids identified with DNA differ from those traditionally recognized based on morphology or life history traits. Monophyly of sea eagles (Haliaeetinae) and booted eagles (Aquilinae) was supported; however, harpy eagles (Harpiinae), snake eagles (Circaetinae), and Old World vultures were found to be non-monophyletic. The Gymnogene (Polyboroides typus) and the Crane Hawk (Geranospiza caerulescens) were not found to be close relatives, presenting an example of convergent evolution for specialized limb morphology enabling predation on cavity nesting species. Investigation of named subspecies within Hieraaetus fasciatus and H. morphnoides revealed significant genetic differentiation or non-monophyly supporting recognition of H. spilogaster and H. weiskei as distinctive species.

Introduction

Accipitridae is a diverse avian family, comprising up to 14 subfamilies, 65 genera, and 231 species (see Table 1, Dickinson, 2003, Stresemann and Amadon, 1979). Of the Accipitridae species, some of the largest and most threatened by anthropogenic factors belong to four eagle subfamilies (Circaetinae, Haliaeetinae, Aquilinae, and Harpiinae) and two Old World vulture subfamilies (Gypaetinae and Aegypiinae). All Accipitridae species are protected under the Convention on International Trade in Endangered Species (CITES) and four eagles are listed as top priority species (CITES I, CITES-Secretariat, 2003). As ecologically sensitive predators, birds of prey are valuable indicators of habitat quality. The Accipitridae are found in a variety of habitats from primary rainforest to arctic tundra throughout the world. Some taxa are restricted in distribution such as the snake eagles (Circaetinae) which are found only in the Old World, while others, such as the sea eagles (Haliaeetinae), are global in distribution. Thorough phylogenetic analyses are needed to delineate the genetic and overall biological diversity of this family, and to inform conservation programs which aim to preserve genetic diversity of distinguishable taxonomic units.

Phylogeny for Accipitridae based on morphological traits has been difficult to resolve (e.g., Brown and Amadon, 1968, Jollie, 1976, Jollie, 1977a, Jollie, 1977b). The few published molecular studies have been limited in sampling and have proposed some previously unrecognized relationships (see below). The goal of the present study is to identify phylogenetic relationships within and among the six subfamilies of eagles and Old World vultures in the context of the other primary accipitrid groups using molecular data.

The booted eagles (Aquilinae) are one of the largest accipitrid groups containing 35–36 species in 8–9 genera and are distributed worldwide. The majority of the species fall into three genera, Aquila, Hieraaetus, and Spizaetus, while the remaining five genera are all monotypic. All species have “boots,” or feathered tarsi, a trait that separates this group from most other accipitrid taxa. The booted eagles have been considered to be monophyletic (Jollie, 1977b) or polyphyletic (Holdaway, 1994) with morphological data, and only a few species in one genus have been studied phylogenetically with molecular data (cyt-b, Seibold et al., 1996; control region, Vali, 2002). Monophyly of the three Aquilinae genera is not well supported with morphological characters, such that the Hieraaetus species and some Spizaetus species have been placed in the genus Aquila by various authors (described by Brown and Amadon, 1968; and Thiollay, 1994). The two molecular studies included about half of the species in the genus Aquila, and both found that A. chrysaetos was genetically distant from four other Aquila species. Sister relationships for A. clanga and A. pomarina, A. nipalensis and A. heliaca or A. heliaca and A. adalberti were also proposed.

The sea eagles (Haliaeetinae) are a much smaller and more easily defined group of large eagles found in coastal and riverine areas worldwide except South America and Antarctica. The two sea eagle genera, Haliaeetus and Ichthyophaga, share some morphological traits with two genera of kites (Milvus and Haliastur), suggesting a close relationship (Holdaway, 1994, Jollie, 1977b, Thiollay, 1994). The sea eagles also share some traits with the palmnut vulture (Gypohierax angolensis), suggesting a relationship between them and Old World vultures (Brown and Amadon, 1968). Using cyt-b sequence data, Seibold and Helbig (1996) studied eight of the nine species of sea eagles in the genus Haliaeetus. They supported a clear split between species with temperate versus tropical distributions, and a close relationship between the sea eagles and two Milvus kites. The relationship between the two genera of sea eagles has not been investigated with molecular sequence data and the possibility of paraphyly of the genera remains unresolved.

The four species and genera of harpy eagles (Harpiinae) are some of the largest raptors and are found in tropical rain forests in the Americas, the Philippines and New Guinea. This group is generally considered monophyletic due to their large size, lack of feathers on the tarsi, and similarities in behavior (Brown and Amadon, 1968, Thiollay, 1994); however, some have suggested that the Old World species are not sister to the New World species (e.g., Jollie, 1977b). Holdaway (1994) removed one Old World (Pithecophaga) and one New World (Morphnus) species from the Harpiinae. A close relationship between the booted eagles and the harpy eagles has been proposed but not tested with molecular data.

The 14 species of snake eagles (Circaetinae) in five genera are found only in the Old World. Although usually considered monophyletic (Brown and Amadon, 1968, Friedmann, 1950), the possibility of polyphyletic origins for snake eagles has been raised (Jollie, 1977b, could not identify sister relationships for Eutriorchis and Dryotriorchis; Thiollay, 1994).

The final group we focused on is the Old World vultures, a diverse mix of scavengers including at least one species that uses tools (Egyptian vulture), and potentially including a frugivorous raptor (palmnut vulture). One or two subfamilies have been proposed for the Old World vultures. Three species are highly divergent from the remaining 11 and have been placed by some in a separate family called Gypaetinae (Mundy et al., 1992). The core 11 species are called the Aegypiinae. Seibold and Helbig (1995) used cyt-b sequence from 11 Old World vulture species and found evidence of polyphyly for the Old World vultures.

There are no previously published molecular studies that include representatives of all of the Accipitridae subfamilies; however, several molecular studies have used mitochondrial DNA to examine particular Accipitridae subgroups and have found evidence for polyphyly of some traditionally recognized taxa (e.g., polyphyly of the Perninae kites, Riesing et al., 2003; and the genus Buteo, Gamauf and Haring, 2004). Relationships among a small set of accipitrids based on mtDNA indicated a closer relationship between a representative sea eagle and kite in the genus Milvus, than between the sea eagle and a snake eagle in the genus Circaetus. A representative Old World vulture was more closely related to the snake eagle than other accipitrid taxa in the study, including species of Buteo, Haliaeetus, Milvus, Circus, Accipiter, and Pernis (Mindell et al., 1997). Increased sampling of species and molecular characters are needed to improve our understanding of phylogenetic relationships among the Accipitridae.

In this study, we focus on full or nearly complete taxonomic representation of five accipitrid subgroups (sea and fish eagles, harpy eagles, booted eagles, snake eagles, and Old World vultures), corresponding to six potential subfamilies. We use both mitochondrial and nuclear sequences for representatives of 51 out of 65 genera (78%) and just under half of the known Accipitridae species (n = 111). At least one representative of each previously proposed subgroup/subfamily within the Accipitridae have been included to help in phylogenetic placement of the focal taxa.

Section snippets

Taxon sampling

We include at least one representative from all genera and the majority of species of sea and fish eagles (2 genera, 10 species), snake eagles (4 genera, 12 species), harpy eagles (4 genera, 4 species), booted eagles (8 genera, 29 species), and Old World vultures (9 genera, 13 species), based on the taxonomy in Dickinson (2003). In two cases where significant morphological differences among geographical populations have been documented, multiple samples representing different subspecies were

Gene properties: sequence composition and divergence

We sequenced 1047 bases of ND2 and 1041 bases of cyt-b for 110 individuals representing 106–108 species and 1074 bases of BF-I7 for 68 of the same species. ND2 contained the most variable sites, the most parsimony informative sites, the highest transition–transversion ratio and had a higher maximum divergence among species as compared to BF-I7 and cyt-b (Table 4). BF-I7 had the lowest percent divergence among taxa and the lowest transition–transversion ratio of the three sequences. Cyt-b had

Discussion

We have presented data from both mitochondrial and nuclear sequences for approximately 50% of the recognized species in the Accipitridae, focusing on groups commonly known as eagles and Old World vultures with nearly complete species representation. This is the most complete systematic treatment of the Accipitridae family to date based on molecular data. We found strong evidence for non-monophyly of some existing genera and subfamilies. Although Accipitridae subfamilies are infrequently used in

Conclusions

This study takes a large step toward resolving the uncertain relationships among birds in the Accipitridae. Our analyses include over 3000 bases of nuclear and mitochondrial DNA and a sampling of almost half of the known Accipitridae species, with nearly complete species sampling for eagles and Old World vultures. We find support for a set of phylogenetic relationships among Accipitridae taxa that differ from previous hypotheses based on morphological data. Fourteen subfamilies, of which two

Author’s note

After this article was accepted for publication a treatment of a subset of booted eagles (Aquilinae) using cyt-b and additional nuclear sequences was published by Helbig et al. (2005). The findings of Helbig et al. are concordant with our study, as are the cyt-b sequences with the notable exception of Aquila pomarina. While both analyses place A. pomarina as sister to A. clanga with high support, the cyt-b sequences are relatively dissimilar (87.4% similarity index). Aquila pomarina has a

Acknowledgments

We are very grateful to the many institutions and individuals that contributed samples for this study (see Table 1) as well as those who helped in aspects of data collection: Mike Bunce and others at the ABC, Joshua Rest, Wendy Grus, Jeff Johnson, Jared Rubenstein, Matthew C. Klaver, Jill Ceitlin, Christie Chapin, Soochin Cho, Michael Lerner, Walter Boles, David Allen, Alistair McInnes, Ben Hoffman, Sarah Chabangu, Andrew Jenkins, Gerwin van Deventer, Kerri Wolters, Jeanne-Marie Pittman,

References (48)

  • CITES-Secretariat 2003. Update to Appendices I, II, III. Convention on International Trade in Endangered Species of...
  • N.J. Collar et al.

    Threatened birds of Asia: The birdlife international red data book. BirdLife international

    (2001)
  • R.A. Cooper

    DNA from museum specimens

  • R.A. Cooper et al.

    Ancient DNA: do it right or not at all

    Science

    (2000)
  • E.C. Dickinson

    The howard and moore complete checklist of the birds of the world

    (2003)
  • J. Ferguson-Lees et al.

    Raptors of the world

    (2001)
  • H. Friedmann

    The birds of north and middle america

    (1950)
  • Gadow, H., 1893. Vogel. In: Bronns, H.G. (Ed.), Klassen und Ordnungen des Thier-Reichs, wissenschaftlich dargestellt in...
  • A. Gamauf et al.

    Molecular phylogeny and biogeography of honey-buzzards (genera Pernis and Henicopernis)

    J. Zool. Syst. Evol. Res.

    (2004)
  • TA. Hall

    BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT

    Nucleic Acids Symp. Ser.

    (1999)
  • Holdaway, R.N., 1994. An exploratory phylogenetic analysis of the genera of the Accipitridae, with notes on the...
  • J.P. Huelsenbeck et al.

    MR BAYES: Bayesian inference of phylogenetic trees

    Bioinformatics

    (2001)
  • P. Jepson et al.

    The end for Indonesia’s lowland forests?

    Science

    (2001)
  • Johnson, J.A., Watson, R.T., Mindell, D.P., 2005. Prioritizing species conservation: does the Cape Verde kite exist?...
  • Cited by (189)

    • Scientific data laundering: Chimeric mitogenomes of a sparrowhawk and a nightjar covered-up by forged phylogenies

      2021, Biochemical Systematics and Ecology
      Citation Excerpt :

      This study found multiple instances where the relationships depicted in Liu et al. (2017) differed strongly from those inferred from re-analysis of the same (mitogenomic) or likely the same (12S/16S rRNA) sequences. These include: (i) the strongly-supported position of Accipiter gularis among Accipiter in their 12S/16S rRNA phylogeny, while the 12S/16S rRNA part of the relevant sequence corresponds largely to that of Streptopelia orientalis, (ii) the strongly-supported position of Caprimulgus indicus near other nightjars, while this sequence is actually a chimera of two owls and ended up among Otus owls in all relevant phylogenies inferred in the present study, (iii) the inclusion of Nisaetus bartelsi in the mitogenome tree, whereas no mitogenome sequence of this species exists, (iv) the inclusion of two N. alboniger in the 12S/16S rRNA tree, whereas only one such sequence exists, (v) the reciprocal monophyly of Buteo, Leucopternis and Buteogallus, while our re-analysis and three previous studies (Amaral et al. 2006, 2009; Lerner et al., 2008) showed that these genera are not monophyletic, and (vi) the puzzling, but strongly-supported, sister relationship of Geranoaetus melanoleucus and ‘Geranoaetus’ (=Geranospiza) caerulescens, contradicting the results of at least six other molecular phylogenetic studies (Lerner and Mindell 2005; Griffiths et al., 2007; Lerner et al., 2008; Amaral et al., 2009; Barrowclough et al., 2014; Mindell et al., 2018). Other suspicious aspects of the phylogenies in Liu et al. (2017) are (vii) the consistently high bootstrap values in the 12S/16S rRNA tree, which is surprising considering that the tree spanned multiple orders of birds; indeed, our re-analysis resulted in significantly lower bootstrap support values, (viii) the lack of conflicts between the 12S/16S rRNA trees obtained by ML and BI, whereas in our re-analysis many relationships in the ML and Bayesian trees were incongruent; and (viii) the much lower log likelihood values of the topology of the trees in Liu et al. (2017) than those obtained during re-analysis.

    • Macroevolutionary drivers of morphological disparity in the avian quadrate

      2024, Proceedings of the Royal Society B: Biological Sciences
    View all citing articles on Scopus
    View full text