Abstract

INTRODUCTION

‘But as far as the problem of the relationship of the orders of birds is concerned, so many distinguished investigators have labored in this field in vain, that little hope is left for spectacular break-throughs.’ (Stresemann, 1959: 277)

‘It must be remembered that the basic avian structure was determined at an early stage in the evolutionary history of birds because of the rigorous limitations placed upon a flying vertebrate. Consequently, adaptations in the birds have been along lines that are not always indicated by the details of anatomy, a fact that makes these vertebrates highly interesting to the student of recent animals but difficult subjects for the palaeontologist.’ (Colbert, 1980: 187)

Current status of avian phylogenetics

‘… the currently accepted arrangement of birds in no way reflects the probable evolutionary history of the class…. The arrangement used here is predicated mainly on the assumptions that birds originated on land rather than in the water, and that highly specialized waterbirds are more derived than less specialized ones…. a consensus has emerged that birds originated, if not in trees, certainly on land. Therefore, we should look for the most primitive taxa among the land birds.’ (Olson, 1985: 83, 84)

‘If one had to summarize the current state of knowledge, the most pessimistic view would see the neoavian tree as a “comb,” with little or no resolution among most traditional families and orders.’ (Cracraft et al., 2004: 475)

‘Perhaps the greatest unsolved problem in avian systematics is the evolutionary relationships among modern higher-level taxa.’ (James, 2005: 1052)

Harrison et al. (2004: 974) concluded: ‘It is almost an offense against birds that the deep mammalian tree is virtually resolved … whilst there are still major uncertainties about many aspects of the avian evolutionary tree.’ In support of this sentiment, the authors cited fundamental discordance among phylogenetic inferences for birds based on mitochondrial and nuclear genomes, an assessment at odds with a contemporary review by García-Moreno et al. (2003). Discussion of morphological efforts by Harrison et al. (2004) was limited to the uncertainties raised by Cracraft (1981, 2001) but verified increasingly by analyses (Cracraft, 1982a, 1986, 1988; Cracraft & Mindell, 1989; Cracraft et al., 2004; Mayr, 2005a). Reconstructions of the higher-order relationships of birds based on morphological characters, in turn, have been depauperate in both characters and taxa and seldom genuinely cladistic (e.g. Cracraft, 1986, 1988, 2001; Cracraft & Clarke, 2001; Mayr & Clarke, 2003).

Regardless of the taxonomic group considered, however, the sobering truth is that the goal of phylogenetics is extremely ambitious and without easy or uniformly reliable means of accomplishment. It is beyond debate that the conceptual framework of morphological cladistics (Hennig, 1966) and ever-increasing computational power has led to significant progress. Nevertheless, it is also clear that many phylogenetic problems have proven resistant to all attempts at solution and seem destined to controversy. Also, phylogenetic endeavours are replete with disagreements in method (both for reconstruction and for evaluation of estimates) and types of evidence considered most reliable. Currently, the tendency is to consider molecular reconstructions as representing the future of avian phylogenetics, and that it is simply a matter of time, perhaps less than a decade, before a global consensus is achieved within the systematic community (Barrowclough, 1992; Livezey & Zusi, 2001; Stanley & Cracraft, 2002).

Deficiencies in taxa or characters typically render comparisons among investigations problematic (Bledsoe & Raikow, 1990), and attempts to reconcile the phylogenetic evidence for Aves substantiate this generality (Cracraft & Mindell, 1989; Mayr, Manegold & Johansson, 2003, 2004a; Dyke & Van Tuinen, 2004; Griffiths et al., 2004). Indicative of disappointing progress in mid- and lower-order avian phylogenetics is the conclusion that basal (higher-order) nodes may be irresolvable or accurate approximations of genuine, explosive radiations (Poe & Chubb, 2004). While demonstrably true of analyses confined to few characters or limited taxonomically (Kumazawa & Nishida, 1995), a single decade of uninspiring inference is insufficient to judge solution to be beyond hope.

The current status of molecular resolution of deepest neornithine nodes, however, serves to underline the likelihood that many genes provide inadequate phylogenetic signal for the problems at hand, a deficiency exacerbated by basal polarities necessarily based on closest extant relatives that are unfortunately comparatively distantly related, e.g. Crocodylia and Testudines (Larhammar & Milner, 1989; Iwabe et al., 2004). The fact that ‘nearest’ outgroup(s) for molecular analysis need to be be extant has had unfortunate implications for rooting, in that for Neornithes these outgroups are comparatively distantly related and may converge on ‘white noise’ as indicators of avian polarities, especially for rapidly evolving mitochondrial data.

Like morphological estimates, a number of potential pitfalls (rooting aside) afflict molecular reconstructions, e.g. serial homoplasy by misalignment, distortions related to silent substitutions, unrealistic treatment of ‘gaps’, and unequal evolutionary rates over extended intervals of geological time and among lineages. Furthermore, disagreement persists if not expands regarding methodological preferences – e.g. classes of data employed, protocols for alignment (i.e. diagnosis of serial homology), choice among reconstructive methods, and assessment of resolution and support (Felsenstein, 2004). Until substantial agreement concerning methods is attained and accurate synergism among molecular and morphological methods secured, the field will remain vulnerable to methodological bias and a tolerance for poorly supported hypotheses of phylogeny, in which even the best-supported works disagree significantly (see Figs 4–9).

Open in a separate window

Figure 4

Molecular phylogenetic trees proposed in previous studies (see Fig. 1 for details), IV. A, Sibley & Ahlquist (1990: figs 354–356), simplified to orders, wherein parenthetical ‘para’ indicates paraphyly of sampled members, and ‘aug’ indicates unconventional content; B, Mindell et al. (1997).

Open in a separate window

Figure 9

Molecular phylogenetic trees proposed in previous studies (see Fig. 1 for details), IX. A, Pereira & Baker (2006a); B, Slack et al. (2006a).

Open in a separate window

Figure 2

Morphological phylogenetic trees proposed in previous studies (see Fig. 1 for details), II. A, Mayr & Clarke (2003); B, Bourdon et al. (2005).

Open in a separate window

Figure 5

Molecular phylogenetic trees proposed in previous studies (see Fig. 1 for details), V. A, Espinosa de los Monteros (2000); B, Johansson et al. (2001).

Goals and objectives

The primary purpose of this paper is to present a morphologically based phylogenetic hypothesis of higher-order relationships of Neornithes. A compendium of characters is provided within the companion work (Livezey & Zusi, 2006), including a bibliographic synthesis, annotations of prior uses of synonymous and related characters, and a compact disc of the data matrix for refinement and augmentation. The secondary objective of this work is to provide a cladistic alternative to the molecular phenetics of Sibley & Ahlquist (1990), at least for non-passeriform families, and to serve as a framework for lower-level studies of included families. An earlier paper on philosophical and methodogical issues (Livezey & Zusi, 2001), despite an explicit disclaimer to the contrary, frequently has been cited as a phylogenetic hypothesis appropriate for comparison with works considered complete by their authors, even regarding positions of individual taxa (e.g. Cracraft et al., 2003). We began the present study with the opinion that the phylogenetic signal encoded within avian anatomy is, with adequate study of both definitive and ontogenetic variation of an adequate sample of modern lineages, more than sufficient for the reconstruction of the higher-order phylogeny of Neornithes. We remain at least as optimistic concerning this goal.

The present phylogenetic hypothesis is intended to serve both as a baseline estimate and ‘scaffold’ for finer-scale reconstructions of terminal clades (i.e. families), as attempts at broad reconstructions of the phylogeny of Neornithes to date have been limited, at the very least, in taxonomic representation (e.g. Slack et al., 2006b) or discredited methods of inference (Sibley & Ahlquist, 1990). We also sought to provide robust nodes supplemental to the few phylogenetic hypothesis currently employed for calibrations of age based on fossils (e.g. Dyke & Van Tuinen, 2004; Pereira & Baker, 2006a) or their surrogates (Van Tuinen, Stidham & Hadly, 2006). Integration of these data with a rich matrix of DNA-sequence data (Cracraft et al., 2004) is planned to explore the power of ‘total evidence’ to recover both higher-level and lower-level avian phylogeny. Perhaps most importantly for the facilitation of future analyses, be these morphological or molecular, is the identification of sister-groups (optimal outgroups) for purposes of rooting analyses of single orders or families. The comparatively sparse representation of taxa in the present analysis reflected logistical limits, but was considered adequate for achieving the stated objectives. Findings herein principally were compared with modern higher-order reconstructions (e.g. Mindell et al., 1997; Mayr & Clarke, 2003; Mayr et al., 2003), the most critical of which are summarized graphically here (Figs 1–9). Works of narrower scope are considered where issues of familial monophyly persist, with emphasis on truly phylogenetic works as opposed to eclectic or phenetic assessments (Raikow, 1985a).

Open in a separate window

Figure 1

Morphological phylogenetic trees proposed in previous studies, I. A, Cracraft (1988); B, Mayr et al. (2003). Some trees were subjected to topologically neutral modifications of taxa to facilitate comparisons (also Figs 2–9). See corresponding papers for analytical methods and topological statistics.

METHODS

Phylogenetic analysis

Critical concepts and terminology

‘It is possible[50% likelihood] that – 1) a distant relationship exists between Apteryx and a tinamou-galliform assemblage; … (5) the diurnal birds of prey may be allied to the owls through the Falconidae … It is improbable[formerly widely believed, since discredited] that – 1) a close relationship exists between Rhea and the tinamous; … (3) Pandion deserves familial status in the Falconiformes …’ (Sibley & Ahlquist, 1972: 241), emphasis added.

‘… the mousebirds, or colies, [i] have no close living relatives, … [ii] they are the only survivors of an ancient divergence … Their [iii] closest living relatives are probably …’ (Sibley & Ahlquist, 1990: 363)

Before considering specific findings in the present study, a clarification of critical terms is essential. The first of the foregoing quotes comprises four statements of perceived probability that either make no objective sense or are self-contradictory by conventional standards. Also, the second quote contains three conclusions (i–iii) for a single group based on a single data set that are: either mutually contradictory (i and iii), or of undetermined meaning (ii vs. either i or iii). In cladistic terms, ‘most closely’ implies ‘closely’ in that hierarchy defines relative relationships. Sister-groups are by definition the ‘most closely related’ of any taxa compared. For example, in cladistic terms, an assumption of monophyly of life on earth implies that every taxon has a close relative and/or closest relative, regardless of extinctions. In other words, degree of relatedness is relative: all lineages have a closest relative and therefore a close relative. Sister-groups need not meet some standard of similarity or absolute antiquity of divergence to qualify. However, under an expectation of at least a limited correlation between evolutionary change in morphology with time – neither ‘clock-like’ nor wildly heterogeneous and completely disassociated – sister-taxa can be expected to share degrees of similarity broadly related to time since divergence, such that recency of divergence between sister-taxa tends to be associated with similarity, and antiquity of such divergence to be associated with dissimilarity.

RESULTS

Minimal-length trees or MPTs

The search for MPTs recovered 97 trees of minimal length (19 553 steps) under standard ordering of multistate characters and rooting by outgroup taxa as given (see Methods). This solution set (2.04 × 10¹¹ rearrangements assessed) had the following summary statistics: CI = 0.2432; RC = 0.1664; RI = 0.6842; and skewness, g₁|10⁵ = −0.4258.

A strict consensus tree of the MPTs (Figs 10–18) was completely resolved for the Neornithes with the exception of six polytomies (mostly trichotomies, some nested, discussed below), uncertainties sufficiently limited so as to obviate a majority-rule consensus tree for the primary solutions set, or to delimit ambiguity where one or more ‘rogue taxa’ may be influential (Sumrall, Brochu & Merck, 2001). The strict consensus tree for the 97 MPTs shared the following summary statistics: (i) component information, 173; (ii) Nelson–Platnick term information, 4367; (iii) Nelson–Platnick total information, 4540; and (iv) Mickevich consensus information, 0.168.

Open in a separate window

Figure 10

Ordinal-level strict consensus tree for orders of Neornithes based on 2954 morphological characters, indicating delimitations of segments detailed in Figures 12–18.

Open in a separate window

Figure 18

Detailed segment of strict consensus tree of all MPTs recovered in present study. Part G. Neornithes: Piciformes, and Passeriformes. Nodes are labelled above by percentages of bootstrapped replicates in which node was retained (italics), and below by Bremer support indices (bold type).

Outgroup taxa: Mesozoic roots of Aves

Non-Neornithine Aves

In light of the growing consensus regarding fossil lineages of the Mesozoic and widely employed characters thereof, broad agreement between our findings and those of others treating pre-neornithine birds was not unexpected. Relationships among outgroup taxa in this analysis generally were consistent with recent analyses (Martin, 1983; Witmer, 1991; Holtz, 1998; Padian & Chiappe, 1998; Clarke & Chiappe, 2001; Chiappe, 2001, 2002; Clarke & Norell, 2002, 2004; Clark, Norell & Makovicky, 2002; Chiappe & Dyke, 2002; Maryanska, Osmólska & Wolsan, 2002; Pisani et al., 2002; Snively, Russell & Powell, 2004; Mayr, Pohl & Peters, 2005; Zhou & Zhang, 2005).

Critical for empirical rooting of ingroup taxa, as opposed to hypothetical ancestors or other synthetic means of proposing polarities, this congruence lends credence to assessments of polarities of characters at the most basal of neornithine nodes (e.g. the divergence of neognathous from palaeognathous taxa). Crocodylians fell as predicted among the basal Archosauria (Larhammar & Milner, 1989; Hedges, 1994). Principal exceptions from a growing consensus of palaeontologists were reversed positions or irresolution within two pairs (Fig. 12): (i) Troodontoidea (Troodon and Saurornithoides) and Dromaeosauroidea; and (ii) Rahonavis and Apsaravis, the latter couplet being equally parsimonious whether paraphyletic to other taxa or as sister taxa. Details of positions among outgroups are of secondary interest here, but it is noteworthy that the few instances of incongruence with other studies were associated with exceptionally poorly supported nodes or polytomies in the present work (Fig. 12). It is likely that the generally lower support indices among pre-Neornithes reflect missing key taxa and poor preservation of those coded.

Open in a separate window

Figure 12

Detailed segments of strict consensus tree of all MPTs recovered in present study. Part A. Outgroup (non-neornithine) taxa. Nodes are labelled by percentages of bootstrapped replicates in which node was retained (numerator), and below by Bremer support indices (denominator).

Neornithes

‘… it is probable that the majority of living genera [of birds] were in existence by the end of the Tertiary… . Most, perhaps all, of the [modern] orders of birds had become established by the end of the Eocene.’ (Brodkorb, 1971a: 42)

‘The phylogenetic position of Palaeogene birds thus indicates that diversification of the crown-groups of modern avian “families” did not take place before the Oligocene, irrespective of their relative position within Neornithes (crown-group birds).’Mayr (2005a: 515)

Strong support for monophyly of the Neornithes (Table 2; Figs 10, ​,11)11) was conferred. Notable, however, in the present reconstruction was its poor congruence with the ‘tapestry’ depicted by Sibley & Ahlquist (1990), in which only three higher-order taxa – their Ratitae, Galloanserae and Procellarioidea, and monophyly of one contentious order (Caprimulgiformes) – were in significant agreement with corresponding clades in the present analysis. Points of disagreement, however, were abundant and included much of the topological (diagrammatic) structure in the two works, and notably included the following groupings depicted by Sibley & Ahlquist (1990: fig. 4A): (i) monophyly of {Ratitae, Galloanserae}; (ii) provisional, exceptionally basal placement of Turnix; (iii) very basal positions and interposition of Piciformes, Coraciiformes, Coliiformes, Trogoniformes and Passeriformes; (iv) multiple discrepancies associated with hypotheses of polyphyly of Pelecaniformes and Ciconiiformes, and (v) inclusion of some Gaviiformes, Podicipediformes, Sphenisciformes and Falconiformes among these groups. Topological dichotomies that hierarchically group modern orders of Neornithes were sought (Fig. 10), and these formed the ordinal basis for a higher-order classification (Appendix 1).

Open in a separate window

Figure 11

Simplified summary tree for uppermost, supraordinal ranks of avian classification. Dashed internodes correspond to marginally supported clades. For complete classification, see Appendix 1.

In the following, descriptions of findings, statistics of support, etc., were presented in figures, and reference to these was employed in place of repetition of metrics in the text. Consequently, readers are directed to the appropriate figures and tables where narratives refer to robustness, support and relative parsimony of alternative hypotheses.

Modern palaeognathous birds

This analysis revealed the relationships among the palaeognathous birds to be exceptionally resolved, well supported, virtually unambiguous, empirically rich, markedly traditional, and supported by an unprecedented sample of outgroups. The ratites or flightless modern palaeognathous birds have been the subject of more anatomical and molecular study than any other avian group, an important motivation for which concerned diagnoses of plesiomorphic and apomorphic morphological characters in a group widely recognized to represent an early branch among Neornithes but for which useful outgroups were lacking (Balouet, 1984; Zusi, 1993). Basal polarities of characters of plesiomorphic condition among modern and closely related fossil palaeognathous taxa (Houde & Olson, 1981; Houde, 1988; Leonard et al., 2005) awaited resolution by means of the most primitive Aves, many recovered only recently (Appendix 1).

Taxonomically orientated anatomical studies, emphasizing ratites or more inclusive in scope, ensued during the 19th and 20th centuries (Fürbringer, 1888; Feduccia, 1980; Houde & Haubold, 1987), and investigations of phylogenetic emphasis were among the earliest for Neornithes (Verheyen, 1960a; Sibley & Ahlquist, 1972; Cracraft, 1974a; Wattel, Stapel & de Jong, 1988). In some cases, inference of the primary grade of divergences of palaeognathous, galloanserine and other neognathous taxa aided in the recovery of historical patterns and broad outlines of phylogeny of palaeognathous taxa, patterns that were to prove beyond the limits of mtDNA for resolution (Härlid, Janke & Arnason, 1997, 1998).

Most prior studies regardless of method – notably excepting early works conceptually confined by the dated biogeographical paradigm of static continents (Briggs, 2003) or phenetic perspectives on affinities (McDowell, 1948; de Beer, 1956; Storer, 1960a, 1971a, b; Sibley & Frelin 1972) – have hypothesized that the palaeognathous birds are the sister-group of other Neornithes, the Tinamiformes are the sister-order of the ratites among palaeognathous taxa (Caspers, Wattel & de Jong, 1994; de Kloet & de Kloet, 2003), and accordingly the ratites are monophyletic (Bock, 1963b; Prager et al., 1976; Stapel et al., 1984; Bock & Bühler, 1988; Härlid et al., 1997; Lee et al., 1997; Van Tuinen, Sibley & Hedges, 1998; Dyke, 2001a; Dyke & Van Tuinen, 2004; Slack et al., 2006a, b). These findings counter early disputes based in part on biogeography, isolated interpretations of fossils (Houde & Olson, 1981), speculations regarding heterochrony (Feduccia, 1985) and (subsequently admitted) analytical anomalies (Härlid & Arnason, 1998). Notable in the last of the foregoing categories was the initial inference of a sister-relationship between a neognathous group comprising the Galliformes and Anseriformes and the palaeognathous birds by Sibley & Ahlquist (1990), a topology rendering at the outset the polyphyly of neognathous taxa; subsequently these authors depicted the neognathous birds as monophyletic.

Monophyly of the Tinamiformes was supported by the molecular analyses by Paton et al. (2002) and Harrison et al. (2004), but minimal taxonomic sampling diminished the generality of these inferences. Sister-group relationships of palaeognathous orders – Struthioniformes and Rheiformes, and Dromaiiformes and Casuariiformes – were supported strongly here (Fig. 13) and elsewhere (Lee, Feinstein & Cracraft, 1997; Leonard, Dyke & Van Tuinen, 2005). A minority of earlier findings (Figs 7A, ​,8B)8B) provided weak evidence of paraphyly of the Struthioniformes and Rheiformes with respect to a sister-grouping of Dromaiiformes and Casuariiformes and also provided weak support for the Apterygiformes as as sister-group to the latter (Van Tuinen, Sibley & Hedges, 2000; Cooper et al., 1992, 2001; Paton et al., 2002; Harrison et al., 2004). Despite support indices suggestive of robustness in several of the molecular works, questions regarding Bayesian bootstrap values (Simmons, Pickett & Miya, 2004) justify caution in such assessments.

Open in a separate window

Figure 7

Molecular phylogenetic trees proposed in previous studies (see Fig. 1 for details), VII. A, Paton et al. (2002); B, Sorenson et al. (2003).

Open in a separate window

Figure 8

Molecular phylogenetic trees proposed in previous studies (see Fig. 1 for details), VIII. A, Chubb (2004a); B, Harrison et al. (2004).

Open in a separate window

Figure 13

Detailed segment of strict consensus tree of all MPTs recovered in present study. Part B. Neornithes: Palaeognathae and Galloanserae. Nodes are labelled above by percentages of bootstrapped replicates in which node was retained (italics), and below by Bremer support indices (bold type).

The Apterygiformes, herein placed as sister-group to all other ratites (Fig. 13), have been inferred to occupy a marked diversity of positions in prior studies (Cracraft, 1974a, 2001; Lee et al., 1997; Cooper et al., 2001; Haddrath & Baker, 2001; Paton et al., 2002; Harrison et al., 2004). Also, the position of the Apterygiformes relative to the extinct Dinornithiformes varied (Vickers-Rich et al., 1995). The Apterygiformes are the most speciose and genetically subdivided of extant orders of ratites (Baker et al., 1995; Burbridge et al., 2003), but are significantly less diverse than the formerly sympatric Dinornithiformes.

The position of the Dinornithiformes also remains a point of controversy, in part because of missing data for this extinct, diverse group; monophyly and relationships among members have been confirmed (Baker et al., 2005). Cracraft (1974a, 2001) considered the Dinornithiformes to be the sister-group of the Apterygiformes, contrary to Cooper et al. (1992, 2001), Van Tuinen et al. (1998, 2000), Haddrath & Baker (2001) and the present provisional inferences. In most respects, the topologies for ratites inferred by Lee et al. (1997) and Dyke & Van Tuinen (2004: fig. 4) most closely approximated that inferred here (Fig. 13).

Missing data for two orders of ratites – Dinornithiformes and Aepyornithiformes – proved analytically problematic if included unconditionally with extant ratites. Unrestricted analysis of these extinct, moderately related, highly divergent, sparsely coded lineages resulted in a suspicious placement of these two orders as sister taxa. The large numbers of missing data in the two extinct lineages, many lacking in both taxa, prompted two alternative analyses to be performed. Global searches of Dinornithiformes (excluding the poorly known Aepyornithiformes) and placements within the MPT as backbone-constraint placed the moas to be the sister-group of other ratites exclusive of Apterygiformes (Fig. 13), contrary to a sister-relationship between these New Zealand endemics as advocated by Cracraft (1974a, 2001). By backbone-constraints or exclusion of the Dinornithiformes, the Aepyornithiformes were placed as the sister group of the clade comprising Struthionidae and Rheidae (Fig. 13).

Galliformes and Anseriformes: land and water fowl

Galliformes

The pioneering myological works by Hudson, Lanzillotti & Edwards (1959) and Hudson & Lanzillotti (1964) provided early hints concerning relationships of Galliformes, but unfortunately these surveys were not cladistic and followed Peters (1934) in considering unique Opisthocomus as an aberrant galliform. Studies of galliform fossils continue to be phenetic in approach (Mourer-Chauviré, 2000; Göhlich & Mourer-Chauviré, 2005). Fortunately, this pattern is likely to change with the increasingly common phylogenetic analyses of galliforms (Dyke, Gulas & Crowe, 2003) and an improved fossil record (Mayr & Weidig, 2004; Mayr, 2005a).

In the present work, relationships of two families within the Galliformes – Megapodiidae (Birks & Edwards, 2002) and Cracidae (Pereira & Baker, 2004; Grau et al., 2005) as mutually monophyletic, sequential sister-groups to all remaining galliforms – agree with placements by other investigators (Prager & Wilson, 1976; Cracraft et al., 2004). Some workers (Hudson et al., 1966), however, suggested a sister-group relationship between the two families (superfamily Cracoidea), as opposed to placement as successive sister-groups (paraphyletic) to other galliforms (Fig. 13).

The robust placement of Meleagrididae as sister-group to the Phasianidae sensu lato in the present work (Fig. 13) opposes inclusion of the family among the enormous complement of other galliforms (reviewed by Sibley & Ahlquist, 1990). The present finding also differs with the indeterminate placement of this distinctive group from most galliforms by Dyke et al. (2003). Dyke et al. (2003: fig. 3) depicted the Megapodiidae and Cracidae as basal, successive sister-groups to the diverse and speciose ‘Phasianoidea’; the latter group included Numida and Acryllium (Numidinae) as members of a polytomous assemblage immediately basal to Meleagris, Agriocharus, Tetraonidae, and a clade comprising 39 taxa of other galliforms inviting taxonomic subdivision. Most of the large-bodied genera of phasianoids (e.g. Gallus, Phasianus) and the ‘Old World quail and partridges’ were among a large, basal polytomy of the ‘phasianoids’ exclusive of the guineafowl (Numidinae). Some of the nodes within this large group, including those resolving Meleagridae and Tetraonidae relative to megapodiid and cracid galliforms, were not sustained by Dyke et al. (2003: fig. 3) in a strict consensus of 1700 MPTs based on 102 characters. Also, the tree inferred here (Fig. 13) departed from those recovered using molecular data (Dimcheff, 2002; Dimcheff, Drovetski & Mindell, 2002).

Open in a separate window

Figure 3

Morphological phylogenetic trees proposed in previous studies (see Fig. 1 for details), III. A, Mayr (2005b); B, Mayr (2005f: fig. 9), excluding fossils Prefica and Paraprefica.

The vast majority of galliform taxa are members of a morphologically conservative group (Holman, 1961), many formerly included among the Perdicidae or Odontophoridae (Sibley & Ahlquist, 1990). These taxa also posed problems of resolution in the present work (Fig. 13), and nodes among these taxa were sufficiently weak as to permit alternative local topologies (i.e. a terminal polytomy). Armstrong, Braun & Kimball (2001) found that mitochondrial and nuclear DNA similarly resolved groupings within a sparse but broad sample of Galliformes. Basal nodes of the latter taxa are broadly consistent with some higher-order topologies (Prager & Wilson, 1976; Helm-Bychowski & Wilson, 1986; Crowe et al., 1992; Kimball et al., 1999; Gutiérrez, Barrowclough & Groth, 2000; Lucchini et al., 2001; Dimcheff et al., 2002; Pereira, Baker & Wajntal, 2002). The single exception among this group (based on included genera) is the strongly supported sister-group relationship between Gallus (Phasianidae) and Numida (Numidinae). The Numidinae were inferred to be the sister-group of the Phasianidae by Kimball et al. (1999) and Pereira & Baker (2006a).

Marine assemblage

A diversity of mutually distinctive groups of aquatic birds have been the focus of much early speculation regarding the potentially misleading effects of similarities of locomotion leading to morphological convergence. Most evocative of these speculations concerned the Gaviiformes and Podicipediformes (e.g. Stolpe, 1935; Storer, 1956, 1960b), foot-propelled diving specialists that prompted arguments based on phenetics, assumptions of ancestral status for fossils, simplistic proposals of evolutionary trends and (most fundamentally) a failure to meet conventional standards of phylogenetic inference. These shortcomings notwithstanding, such proposals from this era gave rise to a general and uncritical acceptance of rampant convergence uniquely afflicting morphological characters, claims that persist to the present day.

Various alliances among the Gaviiformes, Podicipediformes and Procellariiformes were suggested by Mayr & Amadon (1951), and proved consistent with myological data analysed by McKitrick (1991a, b) and molecular patterns recovered by Watanabe et al. (2006). A relationship between the Gaviidae and Charadriiformes was considered plausible by Storer (1956). Without explanation, however, Storer (1971b) listed the loons and grebes together immediately following the Charadriiformes, in apparent contradiction to his previous opinion. Foreshadowing a natural radiation of marine birds, Ho et al. (1976) inferred a comparatively close relationship of the Sphenisciformes with other primarily marine orders, and fossil evidence for loons – of only marginal quality, optimistic appraisals by Olson (1992a) and Mayr (2004a) notwithstanding – suggests an early origin at least for the Gaviiformes. A phylogenetic alliance among the Sphenisciformes, Procellariiformes, Gaviiformes and Podicipediformes was substantiated as well by Cracraft (1982a), and this was indicated by Nunn & Stanley (1998) and Slack et al. (2006a) on molecular grounds.

The comparatively robust skeletal elements of penguins predispose them to fossil preservation, and recently recovered remains hold promise for stratigraphic chronology (Slack et al., 2006b). The clade of basal marine taxa inferred herein evolved myriad modes of foraging (Storer, 1971a): (i) Gaviiformes and Podicipediformes being extremely specialized foot-propelled diving birds; (ii) Sphenisciformes and Pelecanoididae (Procellariiformes) being wing-propelled diving birds, submarine ‘flight’ of the former rendering members aerially flightless (Livezey, 1989a); and (iii) Procellariiformes, comprising hover-foraging Oceanitidae and other families combining wind-powered gliding and plunge-diving (Del Hoyo, Elliott & Sardgatal, 1992). Some fossil groups remain of uncertain ordinal affinity – e.g. the wing-propelled Plotopteridae (Olson & Hasegawa, 1979, 1996; Olson, 1980; Goedert, 1988; Goedert & Cornish, 2002; Mayr, 2004b) – and did not merit analysis herein, where states for cranial characters are critical but specimens are woefully incomplete. Early descriptions suggested the inclusion of the Plotopteridae among Pelecaniformes is competitive with an alternative relationship to Sphenisciformes for which pectoral similarities were emphasized (Mayr, 2004b). Dissent regarding the ordinal relationships of the Plotopteridae is consistent, to a point, with the interordinal relationships of the Pelecaniformes and Sphenisciformes inferred herein (Fig. 14).

Open in a separate window

Figure 14

Detailed segment of strict consensus tree of all MPTs recovered in present study. Part C. Neornithes: nodes are labelled above by percentages of bootstrapped replicates in which node was retained (italics), and below by Bremer support indices (bold type).

Monophyly of the Sphenisciformes seldom has been doubted, and resolution of relationships among modern and fossil species was achieved (Ksepka, Bertelli & Giannini, 2006), but the position of this distinctive marine group remains a long-standing controversy. This duality of distinct synapomorphy and symplesiomorphy underlies a number of classificatory problems of Aves, in which marked distinction of groups tends to confound comparisons with other groups. Of the alternatives proposed, an affinity with the Procellariiformes has received broadest support, both in the present analysis (Fig. 14) and elsewhere (Cracraft, 1981, 1986, 1988).

Despite agreement with the inferences by Cracraft (1982a), it is predictable that strong confirmation of a sister-group relationship between the Gaviiformes and Podicipediformes (Fig. 14) herein will engender concerns of artefactual pairing by convergence (Storer, 1956, 1971a, 2000, 2002). Storer (2002: 16) felt that the non-phylogenetic work by Stolpe (1935) ‘… demonstrated that the similarities among the loons, grebes, … resulted from convergent evolution …’ The inclusion of the Mesozoic Hesperornithiformes with modern Gaviiformes and Podicipediformes by Cracraft (1982a), a finding not supported here (Figs 10, ​,14),14), was the inference subjected to greatest criticism. Obvious similarities of form and life history have prompted exceptional attention to differences between the two orders (e.g. Sibley & Ahlquist, 1972: table 1), tallies without benefit of polarities or phylogenetics. In many cases, these rationalizations are undermined with respect to functional comparisons, e.g. the Gaviidae employ feet for primary propulsion but also use their wings (Olson, 1985), and members of the two orders also differ in the movements typical of the pelvic limb (Storer, 1956). Pairing of the Gaviiformes with the Podicipediformes as sister-groups has been championed by Cracraft (1982a, 1988), a proposal not without opposition (e.g. Storer, 1956, 1960b, 1971a; Sibley & Ahlquist, 1972, 1990). Additional support for this ordinal pairing has been reported (Cracraft & Mindell, 1989; Bourdon, Boya & Iarochène, 2005), but most other analyses excluded one or both of these key orders, rendering comparisons among such works regarding these orders impossible.

Without a consensus regarding a relationship between the Podicipedidae and Gaviidae, the former have been the subject of several extraordinary proposals, based on relatively weak evidence or mere speculation. Olson (1985: 168), under the subheading ‘Family Incertae Sedis Podicipedidae’, stated: ‘In looking beyond their obvious specializations for diving, I cannot see that the grebes (Podicipedidae) would be out of place in the Gruiformes.’ A more precise proposal for the latter is a possible affinity on myological grounds with the gruiforms Rhynochetos and Eurypyga (Zusi & Storer, 1969). An apparent variant of this speculation was a possible relationship with the Heliornithidae and the closely related Rallidae (Beddard, 1893; Olson, 1985; Houde, 1994). Also, a tenuous alliance between the Podicipedidae and Cuculidae was depicted by Van Tuinen et al. (2000), but subsequent works have failed to support this grouping. Another position recently inferred for the Podicipediformes relates to the Phoenicopteridae (Van Tuinen et al., 2001; Mayr, 2004c), a proposal considered further below.

In most respects, inferences herein regarding the Procellariiformes were among the least contentious for the marine assemblage, whether in comparison with traditional (Kuroda, 1954) or modern reconstructions (Nunn & Stanley, 1998; Kennedy & Page, 2002; Watanabe et al., 2006). A moderate departure from traditional arrangements is the finding herein of the Diomedeidae (albatrosses) as comparatively derived, with other Procellariiformes paraphyletic to the typical Procellariidae (Austin, 1996; Gómez-Díaz et al., 2006) and Diomedeidae (Nunn et al., 1996).

Pelecans and allies: totipalmate birds

The totipalmate or pelecaniform birds, as traditionally defined, remain a higher-order group of extraordinary controversy, but in reality the suite of unifying characters, stressed by Beddard (1898), has been expanded for decades beyond the totipalmy cited as sole uniting anatomical character for the order by Sibley & Ahlquist (1972). Polyphyly of the order was inferred subsequently by Sibley & Ahlquist (1990) and Hedges & Sibley (1994). The status of the Pelecaniformes has been debated since the core assemblage was included in widely recognized classifications (Mayr & Amadon, 1951; Wetmore, 1930, 1960), and points of controversy include those of monophyly, content and interordinal position, as empirically derived from metric (Verheyen, 1960b), neontological (Cracraft, 1985), palaeontological (Bourdon, 2005; Bourdon et al., 2005) and molecular perspectives (Siegel-Causey, 1997; Farris et al., 1999).

The exceptional heterogeneity of traditionally included families – e.g. frigatebirds, gannets and pelicans – render questions of membership especially problematic. Perhaps most intriguing of the debated memberships is that of the shoebill or Balaeniceps (Reinhardt, 1860, 1862; Cottam, 1957; Feduccia, 1977a; Mayr, 2003a). Purportedly intermediate features of ‘stork-like’ and ‘pelican-like’ forms (Van Tuinen et al., 2001; Bourdon et al., 2005) have extended to proposals of pelecaniform affinity of the hammerkop (Scopidae). In agreement with the present analysis, the consensus of available phylogenetic works places the distinct Phaethontidae as sister-group to other pelecaniforms exclusive of Balaeniceps (Mayr & Clarke, 2003), with an alternative position hypothesized for the Phaethontidae as an exceptional plesiomorph allied to some pelecaniforms and the Procellariiformes (Bourdon et al., 2005). The present study also resolved Balaeniceps as sister-group to the clade comprising Phaethontidae and other (traditional) Pelecaniformes. Scopus was not inferred here to be closely related to the Pelecaniformes (Fig. 14), contraMayr (2003a).

Relationships among traditional Pelecaniformes (excluding Balaeniceps), inferred cladistically by Cracraft (1985: figs 6, ​,7),7), agreed with the inferences presented herein (Figs 10, ​,14),14), whereas comparisons between the studies with respect to the orders Sphenisciformes, Gaviiformes, Podicipediformes and Procellariiformes were not possible. Sibley & Ahlquist (1990) proposed a ‘four-fold’ polyphyly of Pelecaniformes among the most notable departures of their analysis from contemporary arrangements, whereas several other traditional elements were conserved in their scheme. Hedges & Sibley (1994), based on an analysis impoverished in both data and taxa, also suggested polyphyly of taxa traditionally considered pelecaniform in a work remonstrated by Farris et al. (1999). Syntheses by Van Tets (1965) and Siegel-Causey (1997: fig. 6.3) reaffirmed ordinal monophyly (exclusive of Phaethontidae) using morpho-ethological data, whereas molecular reconstructions violated ordinal monophyly by topologically variable inclusions of the Diomedeidae, Procellariidae and Cathartidae (Siegel-Causey, 1997: fig. 6.2). One minor departure from tradition by Sibley & Ahlquist (1990) was a terminal triad in which the Phalacrocoracidae were placed as sister-group to the Anhingidae and Sulidae.

Open in a separate window

Figure 6

Molecular phylogenetic trees proposed in previous studies (see Fig. 1 for details), VI. A, Van Tuinen et al. (2000); B, Van Tuinen et al. (2001).

Kennedy & Spencer (2004) weakly confirmed monophyly of the traditionally constituted order, in part by use of appropriate outgroups but despite heterogeneous taxonomic sampling of ingroup families. Three weakly resolved departures by Kennedy & Spencer (2004) from the hypothesis inferred herein (Fig. 14) were: (i) reversal of the positions of the Phaethontidae relative to the Fregatidae + Pelecanidae; (ii) a sister-relationship between the Pelecanidae and Phaethontidae; and (iii) paraphyly of Phalacrocoracidae and Anhingidae to the Sulidae.

The Phalacrocoracidae and Anhingidae – families long considered closely related and strikingly similar in external and skeletal aspects (Siegel-Causey, 1988) – have been subjected to unexpected hypotheses of relationship. A series of related papers (Kennedy, Spencer & Gray, 1996; Kennedy, Gray & Spencer, 2000; Kennedy & Spencer, 2000, 2004; Kennedy et al., 2005), based on limited taxonomic representation of pelecaniform families and unconventional analytical methods, mustered mtDNA sequences and behavioural data that favoured paraphyly of these two families to the Sulidae, also inferred phenetically by Sibley & Ahlquist (1990). Based on the present analysis (Table 3), however, a sister-group relationship between Phalacrocoracidae and Anhingidae is strongly favoured.

Table 3

Alternative topological inferences and minimal differences in tree length (additional steps) relative to placements in MPTs (Figs 11–17), conditional on other topological alterations being prohibited (optimizations of characters thereon permitted). Higher-order taxa correspond to classification proposed in Appendix 1

Taxon	Alternative hypothesis*	Δ length	References
Palaeognathae	Galloanseromorphae	54	Sibley & Ahlquist (1990)
Ratitae (global)†	Δ topology	31	Cracraft (1974a)
Ratitae (local)†	Δ topology	63	Cracraft (1974a)
	Δ topology‡	17	Cooper et al. (2001)
	Δ topology‡	13	Haddrath & Baker (2001)
Galloanserimorphae	Polyphyly§	[19]	Bourdon et al. (2005)
Galliformes	Polyphyly	90	Dyke et al. (2003)
Megapodiidae	Cracidae	10	Dyke et al. (2003)
Meleagrididae	Phasianidae	20	Dyke et al. (2003)
Anseriformes	Δ familial topology	54	Olson & Feduccia (1980a); Livezey (1997a)
Anhimae	Galliformes	41	Olson & Feduccia (1980a); Livezey (1997a)
Gaviomorphae	Charadriomorphae	72	Storer (1956); Olson (1985)
Podicipediformes	Phoenicopteridae	146	Mayr & Clarke (2003); Mayr (2004a)
	Charadriomorphae	54	Storer (1956)
	Eurypygidae	182	Zusi & Storer (1969)
	Ralliformes	159	Olson (1985); Houde (1994)
Pelecaniformes	Δ topology	344	Kennedy & Spencer (2004: fig. 1B)
Sulae	Δ topology	125	Kennedy et al. (2005: fig. 8)
Balaenicepitidae	¬ Pelecaniformes	30	Cracraft (1985); Mayr (2003a)
Scopidae	Pelecaniformes	23	Mayr (2003a)
Threskiornithidae	Charadriiformes	174	Olson (1978)
Ardeidae	(Turnices Eurypygae)	75	Olson (1978)
Phoenicopteridae	Anseriformes	107	Feduccia (1976, 1977b); Hagey et al. (1990)
	Cladorhynchini	154	Olson & Feduccia (1980b)
Gruiformes (traditional)	Monophyly¶	11	Livezey (1998b)
Charadriiformes	Δ topology	60	Strauch (1978) fideChu (1995: fig. 1)
	Δ topology	106	Sibley & Ahlquist (1990) fidePaton et al. (2003)
	Δ topology	82	Chu (1995: fig. 8), excluding Ibidorhyncha
Mesitornithidae	Cuculiformes	107	Mayr & Ericson (2004)
Strigiformes	Caprimulgiformes	43	Hoff (1966)
Cathartidae	Ciconiiformes	112	Ligon (1967); Rea (1983); Avise et al. (1994a)
Opisthocomidae	Galliformes	120	Hudson et al. (1959); Hudson & Lanzillotti (1964)
	Cuculiformes**	22	Avise et al. (1994b); Hughes & Baker (1999)
Caprimulgiformes	Polyphyly	31	Mayr (2002a, b)
	Cypselomorphae	42	Mayr (2002a, 2003c, 2004d, 2005f, g)
Aegothelidae	Apodiformes	31	Mayr (2002a, 2003c, 2004d, 2005f, g)
Steatornithidae	Trogoniformes	102	Mayr (2003b)
Hemiprocnidae	Apodidae, monophyly	5	Sibley & Ahlquist (1990: fig. 361)
Apodidae	Passeri (Hirundinidae)	193	Shufeldt (1889b); Van Tuinen (2002)
Galbulae	Coraciiformes	20	Olson (1983a)
Coraciiformes	Δ topology, Trogoniformes	199	Lowe (1946); Maurer & Raikow (1981)
Coracii	Δ topology	64	Cracraft (1971b)
Menura	Passeri	11	Irestedt et al. (2001); Barker et al. (2002)

Open in a separate window

^*Set-symbolism coopted for concise statement of phylogenetic hypotheses, as follows: , sister-group (disjoint) union; , included as subclade; , included as a member taxon; , or; Δ, change in; ¬, not (negation of predicate argument).

^†Local optima for Aepyornithiformes and Dinornithiformes (as bi-ordinal sister-group to ratites exclusive of Apterygiformes) and global optima (former as sister-group to Struthionidae and Rheidae, latter as sister-group to ratites exclusive of Apterygiformes).

^‡Comparisons excluded effects due to differences in outgroup taxa, as well as tentatively placed Aepyornithiformes.

^§Doubtful comparability given differences in taxonomic samples between studies.

^¶Corresponds to that proposed by Livezey (1998b), exclusive of Pedionomidae and fossil gruiforms (Cracraft 1969, 1971a, 1973a).

^**Alternative hypothesis compared sister-grouping with Cuculiformes exclusive of Musophagidae.

Storks, herons and allies

‘Wading birds’, as delimited here, comprise the typically long-legged, long-necked storks and herons, and exclude the morphologically reminiscent cranes and allies (Gruiformes) and the potentially allied shorebirds (Charadriiformes). Highest-order nodes resolved in the present study defined a primary division of (i) ‘herons’ from (ii) ‘storks’ and allies as sister-groups (Fig. 14). Among the ‘storks’, Scopus is the sister-group to other members, the latter comprising clades partitioning the (i) ibises and spoonbills, and (ii) flamingos and typical storks. Within the ‘herons’, the only notable finding is the placement of Cochlearius as sister-group to other herons (Fig. 14), an inference consistent with traditional classifications (e.g. Wetmore, 1960) and earlier findings (Cracraft, 1967a; Sheldon, Jones & McCracken, 2000).

Shufeldt (1901b) suggested affinities between the Phoenicopteridae (flamingos) and both the Anseriformes (waterfowl) and the Ciconiiformes (storks and traditional allies). Olson (1978) questioned the monophyly of the traditional Ciconiiformes on phenetic grounds, suggested charadriiform affinities of Phoenicopteridae and Threskiornithidae, and expressed uncertainty regarding the ordinal placement of the herons (Ardeidae). Van Tuinen et al. (2001), based on conventional molecular estimates and the phenetics of DNA–DNA hybridization, found no support for monophyly of the Ciconiiformes in an analysis including representatives from several other traditional groups. Molecular reconstructions by Slikas (1997), however, confirmed monophyly of the morphologically diverse, ‘true’ storks (Scopus and Balaeniceps not sampled), groupings that also were afforded significant ethological support (Slikas, 1998).

It has been hypothesized in recent years that the Phoenicopteridae may be the sister-group of the grebes (Podicipediformes), a proposal supported by tenuous molecular (Van Tuinen et al., 2001) and morphological evidence (Mayr & Clarke, 2003; Mayr, 2004c; but see Storer, 2006). Given the variable viewpoints expressed regarding the Phoenicopteridae as well (Gadow, 1877; Shufeldt, 1889a; Feduccia, 1976, 1977a), this couplet offered the hope of dispensing with two challenging taxonomic placements by means of a single union, a circumstance not uncommonly an artefact of long-branch attraction (Philippe et al., 2005). Both of these autapomorphic taxa have been subjected to classificatory confusion for more than a century (e.g. Weldon, 1883; Shufeldt, 1901b; Jenkin, 1957), with affinities of the flamingos considered plausible between either the Ciconiiformes or the Anseriformes. Despite robust support for the more traditional position in the present analysis (Tables 2, ​,3;3; Figs 10, ​,14)14) and the minimal evidence presented by others for the proposal of the Podicipediformes, the latter hypothesis merits examination on the grounds of its superficial implausibility and the marked rearrangements of higher-order avian relationships it would imply. Supplementary morphological support for a sister-group relationship between grebes and flamingos marshalled by Mayr & Clarke (2003), however, required the exclusion (in a second analysis) of the loons – heretofore the global sister-group of the grebes – to sustain the grouping in question. Both exclusion of the Gaviiformes and narrow sampling of characters and taxa with which the Phoenicopteriformes were evaluated by Mayr (2004c) weakened the resultant inferences regarding the relationships of flamingos.

Chubb (2004a: 148) recovered 50% and 78% bootstrap support for this taxonomic couplet in analyses of different partitions of the ZENK gene, and joined Van Tuinen et al. (2001) in the speculation that: ‘… because both grebes and flamingos are highly derived morphologically and adapted to unique aquatic niches, their potential evolutionary alliance has previously gone unnoticed.’ Unfortunately, this rationalization is vulnerable to criticism because: (i) modifications for foot-propelled diving of grebes are comparable with those of several other groups of Neornithes – e.g. some Anatidae (Oxyurini, Mergini), Gaviidae, Phalacrocoracidae and Anhingidae; and (ii) the ‘unnoticed alliance’ between grebes and flamingos recognized by Chubb (2004a) instead was countered by a number of apomorphies in each genus that are shared with other taxa – e.g. Podiceps with Gavia, Phoenicopterus with (other) Ciconiiformes. The present data set (Livezey & Zusi, 2006) supports the rejection of this novel proposal involving the grebes and flamingos (Table 3; Fig. 14), and suggests that the taxonomic proposal for the couplet by Sangster (2005) is premature.

Cranes, rails, shorebirds and allies

The remaining long-legged, statuesque denizens of early successional, often wet habitats, together with the true shorebirds, compose the sister-group of remaining neornithine taxa (Fig. 15). These families, typically included within the traditional Charadriiformes and Gruiformes, have a long, perhaps unequalled history of debate in the ornithological literature (reviewed by Sibley & Ahlquist, 1972, 1990; Livezey, 1998b). Primary points of controversy concern the monophyly of the Gruiformes, and relationships between the taxa traditionally referred to the Gruiformes and the Charadriiformes; the latter order is known for especially great diversity in structure of the skull (Kozlova, 1961).

Open in a separate window

Figure 15

Detailed segment of strict consensus tree of all MPTs recovered in present study. Part D. Neornithes: Gruiformes and Charadriiformes. Nodes are labelled above by percentages of bootstrapped replicates in which node was retained (italics), and below by Bremer support indices (bold type).

Birds of prey–diurnal and nocturnal

Raptors or birds of prey – comprising the diurnal Falconiformes and (principally) nocturnal Strigiformes – share a primary reliance on carnivory, by scavenging or capture of prey and associated functional commonalities. The sister-relationship of these raptorial orders inferred herein (Figs 10, ​,16)16) and by Mayr et al. (2003) has been the subject of suspicion based on phenetic tallies of differences (Gadow, 1893; Beddard, 1898) and speculations concerning convergences and raptorial specializations (Sibley & Ahlquist, 1972; Cracraft, 1981). However, these orders differ in many respects and manifest substantial diversity within orders, conditions as suggestive of comparatively ancient divergence of sister-groups sharing general raptorial lifestyles and independent (order- and family-specific) morphological refinements. This clade is first in a sequence of four – the birds of prey, Opisthocomus, Cuculiformes, Psittaciformes and Columbiformes – that are sequential sister-groups of remaining Neornithes. Although all of these orders were robust with respect to individual monophyly, the four highest-order branches supporting these orders were not (Figs 10, ​,1515–17), rendering the branching sequence provisional.

Open in a separate window

Figure 16

Detailed segment of strict consensus tree of all MPTs recovered in present study. Part E. Neornithes: Falconiformes, Strigiformes, Cuculiformes and Psittaciformes. Nodes are labelled above by percentages of bootstrapped replicates in which node was retained (italics), and below by Bremer support indices (bold type).

Open in a separate window

Figure 17

Detailed segment of strict consensus tree of all MPTs recovered in present study. Part F. Neornithes: Columbiformes, Caprimulgiformes, Apodiformes, Coliiformes, Trogoniformes and Coraciiformes. Nodes are labelled above by percentages of bootstrapped replicates in which node was retained (italics), and below by Bremer support indices (bold type).

In addition to suspicions of convergence, several concerns may be seen as opposing the phylogeny inferred herein: (i) an alternative interordinal hypothesis that presumes the Strigiformes to be most closely related to the non-raptorial but similarly nocturnal Caprimulgiformes; (ii) an hypothesis that holds the New World vultures (Cathartidae) to be more closely related to the Ciconiidae than to typical birds of prey; and (iii) several counterproposals concerning certain families and genera of Falconiformes, notably positions of the terrestrially specialized secretary-bird (Sagittarius serpentarius), the piscivorous ospreys (Pandion haliaetus), and the distinctive Falconidae relative to other diurnal raptors.

Hoatzin, cuckoos, pigeons, parrots and allies

This group of medium-sized landbirds approximates part of the ‘Anomalogonatae’ of Garrod (1874) and Beddard (1898), largely synonymous with the earlier branches within the ‘higher landbird assemblage’ of Olson (1985). Clades informally included in this grade of modern landbirds are characterized by mutually exclusive apomorphies rendering many of the groups among the most readily recognized of birds. Members of this subterminal grade of avian orders have been the subject of numerous studies, but nonetheless a clear consensus regarding their interordinal affinities has failed to emerge (Bleiweiss, Kirsch & Lapointe, 1994; Bleiweiss, Kirsch & Shafi, 1995; Johansson et al., 2001). As noted previously, these taxa –Opisthocomus, Cuculiformes, Psittaciformes and Columbiformes – traditionally were accorded ordinal rank, and were resolved here as a grade in which defining nodes achieved only marginal support. Accordingly, this series of clades conservatively can be considered to compose a tri-ordinal grade or corresponding polytomy that bridges the Cuculiformes with the Caprimulgiformes and Apodiformes. The latter ambiguity primarily relates to the failure to resolve the order of branching of the Psittaciformes relative to the Columbiformes (Fig. 17). Nevertheless, the orders branching from this grade were each strongly supported.

The Opisthocomidae – solely comprising the unusual hoatzin (Opisthocomus hoazin) – has been allied with Tinamidae, Galliformes, Cuculiformes, Columbidae, Pteroclidae, Rallidae, Otididae and Coliidae, among other higher-order groups (Table 1). Recent attempts to resolve the uncertainty of position of this monotypic lineage by molecular means have proven largely unsuccessful, principally by mutual contradiction or ambiquity of findings (Avise, Nelson & Sibley, 1994b; Hedges et al., 1995; Marceliano, 1996; Sorenson et al., 2003), and also because of contaminated sequence data (Avise & Nelson, 1995; Hackett et al., 1995). A growing number of works are at least consistent with an affinity between Opisthocomus and the Cuculidae (Sibley & Ahlquist, 1972, 1990; Hughes & Baker, 1999), despite disputes regarding method and differences in taxonomic sampling. In the present analysis, Opisthocomus was placed as the sister-group of the Cuculiformes, the latter weakly including the Musophagidae (Veron, 1999) as sister-group to the Cuculidae (Table 2; Fig. 16).

Uncertainties of phylogenetic position and superficial plesiomorphy of Opisthocomus led some (e.g. Feduccia, 1980, 1996; Olson, 1985) to suggest that the taxon derives from the ‘roots’ of Neornithes. This proposal is consistent with a perception that the species descended from uniquely primitive ancestry, a view exemplified by its description as a ‘reptilian’ bird by Parker (1891), its use as the only neornithine explicitly figured with Archaeopteryx or non-avian Theropoda (Brodkorb, 1971a; Feduccia, Lingham-Soliar & Hinchliffe, 2005: fig. 26), and the much-publicised retention and use of weakly functional ungues alulares in the genus prior to fledging (Shufeldt, 1918). In actuality, such ‘wing claws’ are retained by members of many modern avian orders in variably vestigial states (Livezey & Zusi, 2006). Accordingly, morphological and molecular evidence for the purported plesiomorphy of Opisthocomus is ambiguous at best: most studies place the genus as closely related to the Cuculiformes (Hughes & Baker, 1999; present study), whereas a few analyses suggest a more distant relationship (Mayr et al., 2003; Mayr, 2005b).

Various other studies, most with only marginal taxonomic sampling, have inferred a sister-group relationship between Opisthocomus and the Cariamidae (Mayr & Clarke, 2003; Mayr, 2005c) or inclusion within an eclectic assemblage defying plausible explanation in light of other findings (Fain & Houde, 2004). The unique alimentary features of Opisthocomus, notably refinements for herbivorous or ruminant digestion (Dominguez-Bello, Ruiz & Michelangeli, 1993; Kornegay, Schilling & Wilson, 1994), are of little phylogenetic significance as they are autapomorphic among Neornithes. However, the lysozymes associated with fermentation by Kornegay et al. (1994, 2003) suggest Opisthocomus to be more similar to Columba than Gallus.

Phylogenetic studies of the Cuculidae per se are surprisingly few, but include taxonomically inclusive attempts at morphological and ethological insights (Seibel, 1988; Hughes, 1996, 2000; Posso & Donatelli, 2001) as well as a molecular exploration (Sorenson & Payne, 2003). Berger (1960) compiled characters distinguishing the Cuclidae from the Musophagidae, many of which show homoplasy at wider scales of comparison. The molecular study by Johnson et al. (2000), the primary focus of which were the Malagasy couas, resulted in a topology within the family broadly similar to that inferred herein, differences in sampling notwithstanding (Fig. 16).

Goatsuckers, swifts and hummingbirds, mousebirds and trogons

Coraciiform, piciform, and passeriform birds

Branch lengths and evolutionary change

Morphological characters employed in cladistic analyses tend to be held to unrealistic standards, and to serve as sources of insights (not expected of molecular characters) beyond mere inference of phylogenetic relationships. For example, in some circles there is an expectation that, in addition to resolving phylogenetic relationships of multiple taxa, apomorphies supportive of nodes should make obvious functional sense (e.g. debates regarding aquatic lineages and possible convergences) and permit interpretation resembling lists of (semi)diagnostic characters for nested series of taxa. In some cases, particularly where taxonomic scale is low and a functional focus pertains, such patterns and trends are discernible. However, with increasing taxonomic scale, these are in the minority, and like DNA sequence data, such diagnostic transparency and functional interpretation is seldom attainable. Many subtle features possessed of phylogenetic signal may be structural artefacts of functionally neutral details of anatomy, historical accidents that prove variably reliable through the process of evolutionary modification with descent.

Nevertheless, quantification of evolutionary change is critical to estimates of rates and correlation of change among characters and related evolutionary topics (e.g. Omland, 1997a, b; Nunn & Stanley, 1998), and exploration of this aspect of reconstructions is intended to pre-empt misplaced expectations or distorted perspectives. The tempo and mode of morphological evolution and cladogenesis have held the interests of systematists for decades (Simpson, 1944; Cracraft, 1984), pre-dating the advent of molecular methods or assumptions made for them (e.g. uniform or ‘clock-like’ evolutionary rates). Antiquity of lineages provides opportunity, other parameters being equal, for increased expectation of evolutionary changes (probabilistic, not deterministic expectation), and where such lineages comprise only modest numbers of members – i.e. limited evolutionary opportunities for departures from uniformity or reversals within a given lineage – such change also tends to lead to comparatively direct diagnosticity of terminal lineages. Intuitive relevance of origins, ages, longevities of lineages and expectations of evolutionary divergence notwithstanding, these topics have been underserved by newly acquired empirical evidence. Van Tuinen et al. (2006: table 2) listed 14 avian families construed to show molecular variation significantly lower than that expected on the basis of current taxonomic status. Given present findings, however, this issue appears illusory, with virtually all taxa in question having early origins, including the Anhimidae (Anseriformes), Podicipedidae and Spheniscidae, three being members of the Pelecaniformes, and the remaining examples members of either the Ciconiiformes, the Gruiformes or the Charadriiformes.

Unlike molecular evolution, no strict assumptions or dependence on constant or uniform rates of change have been made for morphological characters. In the present analysis, branch lengths varied substantially depending on specific optimizations, and therefore comparisons of lengths, like the internodes in trees, were not restricted to unambiguous changes. Instead, central tendencies of branch lengths of MPTs were quantified by median lengths, and variation among optimizations by standard deviations and ranges of lengths recovered. For Neornithes, numbers of changes optimized as autapomorphies averaged 41 (SD = 33, range 4–186). By contrast, for single lineages, maximal lengths of terminal branches were: 186 for Spheniscus, 132 for Phoenicopterus and 131 for Mesitornis. Minimal lengths of terminal branches were: 4 for Francolinus and Alectoris. At higher phylogenetic scales (interordinal and interfamimlial), branch lengths had the following summary statistics: mean = 36, SD = 33, range 5–133. In general, then, terminal branch lengths were approximately 10% greater than those of the branches subtending them (i.e. deeper internodes). A pattern of short internodes has been inferred previously (Cracraft et al., 2004), but the attribution of cause to realities of evolutionary intervals vs. diminished power of resolution remains contentious.

A survey of the minimal branch lengths included in the MPT revealed that branches among higher-order nodes were extraordinarily similar to associated terminal branches (latter being those subtending individual taxa) in means and variances of branch lengths. However, comparative numbers of the more critical diagnostic and supportive characters within the Neornithes revealed that character-based definitions of highest-order clades (corresponding to the most ancient of synapomorphies) were disappointingly low, whereas those for superorders and orders (Appendix 2) were comparatively robust and included suites of diagnostic character-states (Table 2). However, the correspondence among ‘raw’ branch lengths, statistics of nodal support and numbers of ‘diagnostic’ apomorphies generally was poor (Table 2), in agreement with the findings of Farris et al. (2001) and Wilkinson (2003).

DISCUSSION

Acknowledgments

APPENDIX 1

The following proposal for a higher-order classification of Class Aves is intended to encode natural groups as recovered in the foregoing phylogenetic analysis of the companion morphological data (Livezey & Zusi, 2006). Procedures of phylogenetic classification followed Wiley (1981) and Cracraft (1974b, 1978). We avoided the current divergence between rank-free Phylocode and traditional Linnean formats, as well as the palaeontological penchant for ‘stem’ and ‘crown’ groups. The four principles considered here were: (i) hierarchical grouping by phylogenetic relationship; (ii) preference to familiar, available taxa; (iii) preference given to names based on included type genera, where all other considerations are equal; and (iv) coordination of taxonomic ranks by similar emendation of names. Higher-order group names were chosen to conform most closely with others published comparatively recently and conformal with several conventions: (i) incertae sedis (indicative of unconfirmed monophyly and/or content); and (ii) sedis mutabilis, where a taxon comprises three or more members of equal rank (i.e. lineages in polytomy). Among Neornithes, the sequencing convention (Wiley, 1981) was used only for ordinal ranks for some taxa traditionally considered to be Gruiformes.

The comparatively simplified phylogeny upon which this classification is based is depicted in Figure 11. Exemplary taxa (often nominate genera) actually coded and analysed are shown explicitly in trees (Figs 13–18), and the higher-order taxa (mostly families) that correspond to the exemplars appear in the following classification. Families included in higher taxa are limited largely to those represented by exemplars analysed, e.g. two subfamilies of Anatidae as opposed to all recognized by Livezey (1997b). This convention is most notable with respect to the exceptionally diverse and minimally represented Passeriformes and embracing superorder. However, inclusion of comparatively recently recognized family group names within two orders of Superorder Psittacimorphae (Psittaciformes and Columbiformes), not represented among exemplars, was intended to counter under-representation of non-passeriform clades as well as to accommodate uncertainty of phylogenetic placement of exemplary genera with respect to recognized (sub)families.

No protocol for derivation of taxa of higher rank has been codified; a recent attempt was that by Sibley et al. (1988, 1990). The proposal made here is but one of many alternatives, including 150 years of provisional classifications. Although the ‘sequence convention’ might be applied to the very highest taxonomic ranks, we elected to retain distinct, dichotomous taxa to draw attention to these highest-order ranks within Neornithes; this permits hierarchical clarity, but it also results in some redundancy of higher-order names (Table 2). However, the convention was applied to names for some ranks listed prior to the Neornithes – parvclasses, sections, etc. (cf. Ratitae). We chose to use historical names over proliferation of new (semi)synonyms to preserve taxonomic history and despite the fact that this perpetuated some names of variably different content and inappropriate etymology and diagnosis (as understood by the original authors of these taxa). In some cases, acceptable taxa for some highest-order ranks were not found, and in these few instances new taxa were proposed, e.g. Terrestrornithes, and hyphenates of several others.

Among several frequently cited, 19th-century authors of higher-order taxa, two –Rafinesque (1815) and Leach (1820) – were disqualified following the adjudication of most modern systematists (Bock, 1994). We avoided group names of strongly militaristic overtones, e.g. the ‘brigades’ and ‘legions’ of Gadow (1893). We found the compendia by Lambrecht (1933), Wetmore (1930, 1960), Mayr (1958), Storer (1960a, 1971b), Brodkorb (1963, 1964, 1967, 1971b, 1978), Sibley et al. (1988, 1990) and Sibley & Ahlquist (1972, 1990) to be critical for ascertainment of taxonomic authorships. Many taxa named by Stresemann (1959) or delimited by Verheyen (1961) also were adopted. Full citations of references for higher-order taxa were not included herein for the sake of brevity. For taxa of rank greater than ordinal, we adopted, where possible, the system of suffixes proposed by Sibley et al. (1988, 1990) and Sibley & Ahlquist (1990). Given the dubious comparability of taxonomic ranks of higher-order names, we elected to forego annotation of taxa of supraordinal rank with the conventional specification of ‘new rank’ in this proposal.

The higher-order names given in bold type reflect inferred groupings, and although beyond the ranks of pervue by the ICZN, we provide diagnostic and supportive characters for these taxa (Table 2; Livezey & Zusi, 2006), whether new or conserved from historical works. The latter basis was preferred for naming higher-order groups, with inexactitude of content conceded as in such names used by Clarke & Norell (2002). For example, the content of an established name (e.g. Ornithurae) implicitly is defined herein (i.e. sensu present study). Taxonomy bearing on outgroup taxa (i.e. those preceeding Neornithes) are considered especially tentative. A minor point of contention is the position of Lithornis (Houde, 1988), a relative to palaeognathous Neornithes, inferred to be the sister-group of Tinamidae by Clarke & Norell (2002) and Clarke (2004), but inferred to be the sister-group of Neornithes by Clarke & Chiappe (2001), Leonard et al. (2005) and the present analysis (Fig. 12). Should a sister-group relationship between Lithornis and modern palaeognathous taxa be favoured, Panpalaeognathae Gauthier and de Queiroz, 2001 is available for the clade comprising both groups. Use of the traditional, higher-order taxon Carinatae Merrem, 1813, was precluded by provisional monophyly of Hesperornis and Ichthyornis in the present study (see also Rees & Lindgren, 2005), avoiding as well the implication of the name with respect to secondary obsolesence or loss of the carina sterni among Neornithes (Livezey, 2003a).

Supraordinal names proposed herein were intended to follow the convention of seniority of taxa and (to a lesser degree) included type family, as is typical of lower-scale taxa. Three important higher-order synonyms are: NeoavesSibley et al., 1988, senior to PlethornithesGroth & Barrowclough, 1999 (availability questionable in present context), and distinct from EoavesSibley et al., 1988. There are also, two taxa –‘Cracrafti’ and ‘Conglomerati’– informally proposed as alternatives by Slack et al. (2006a). Unused herein is the potentially useful higher-order name Euornithes Sereno, 1999. Gaviomorphae replaced Colymbimorphae (Gadow, 1893) by revision of the former type genus Colymbus. We also replaced the senior, unfamiliar Dypsporomorphae Ogilvie-Grant, 1898, with the more familiar derivation Pelecanimorphae. Similar reasoning led to the suppression of Aetomorphae (Huxley, 1864) by Falconimorphae (Seebohm, 1890), the latter derived from an included ordinal taxon, but junior to the less representative alternative of Strigimorphae (Wagler, 1830). Uniform emendation of superorders was not imposed herein for non-Neornithes. The importance of dichotomy among higher-order names of comparable rank for comparability with the phylogenetic tree resulted in redundancy of supraordinal names for some clades, e.g. Subdivision Dendrornithes comprises a single Section Raptores, which in turn comprises a single Superorder Raptoromorphae. Antiquity of historical, higher-order taxa often resulted in minor differences in content – e.g. AnomalogonatesGarrod, 1874 optimally should exclude the Cuculiformes for consistency with the myological diagnosis implied by the name.

Further study will probably subdivide Superorder Passerimorphae so as to comprise the Superorder CoracomorphaeHuxley, 1867 and Superorder Passerimorphae (Linnaeus, 1758), the latter to comprise Piciformes and Passeriformes (cf. Manegold, 2005). Within the Passeriformes, the most suspicious anomaly in the present analysis was that of Menura; broader samples may justify its transfer to the Passerida, and thereby the first subordinal taxon instead may comprise the Acanthisittidae (Barker et al., 2002, 2004; Ericson et al., 2002a, b), representatives of which were not available for analysis here.

Principally because of limitations on available specimens, delegation of ordinal rank to the extinct elephant-birds (Aepyornithiformes) was favoured marginally over inclusion at lower rank within the Struthioniformes. A detailed classification of Order Anseriformes, including fossil taxa, was presented by Livezey (1997b), and a preliminary classification of the traditionally delimited Gruiformes, significantly revised by the present analysis relative to that proposed by Livezey (1998b), which tentatively recognized monophyly of the traditional order. (Sub)fossil taxa are plausible candidates for inclusion as sequential sister-groups of Galloanseromorphae (Diatrymidae, Gastornithidae and Dromornithidae) or membership within the Galliformes (Sylviornithidae) or Anseriformes (e.g. Mourer-Chauviré & Balouet, 2005) and are included based on published description (e.g. Cracraft, 1968; Livezey, 1997a) and cursory examinations. Material essential for rigorous diagnosis is rare or lacking, but we considered provisional hypotheses to indicate groupings likely but as yet undemonstrated by formal analysis preferable in such cases to no inference presented at all.

Subclass Avialae Gauthier, 1986

[Infraclass Alvarezsauria (Bonaparte, 1991)]

Infraclass Aves Linnaeus, 1758

Parvclass Palaeoaves; new name

Superorder ArchaeornithesGadow, 1893

Order Archaeopterygiformes Fürbringer, 1888

Family Archaeopterygidae Huxley, 1872

Order Confuciusornithiformes (Chiappe et al., 1999)

Family Confuciusornithidae Hou et al., 1995

Superorder Euenantiornithes Walker, 1981; incertae sedis

Order Rahonaviformes; new name

Family Rahonavidae; new name

Order Apsaraviformes; new name

Family Apsaravidae; new name

Parvclass Ornithurae Haeckel, 1866

Superorder Odontoholomorphae (Stejneger, 1885)

Order Hesperornithiformes (Fürbringer, 1888)

Family Hesperornithidae Marsh, 1872

Order Ichthyornithiformes (Marsh, 1873)

Family Ichthyornithidae (Marsh, 1873)

Parvclass EoavesSibley et al., 1988; incertae sedis

Order Lithornithiformes Houde, 1988

Family Lithornithidae Houde, 1988

Parvclass NeornithesGadow, 1893

Cohort Palaeognathae Pycraft, 1900

Subcohort Crypturi Goodchild, 1891

Superorder Dromaeomorphae (Huxley, 1867)

Order Tinamiformes (Huxley, 1872)

Family Tinamidae Gray, 1840

Subcohort Ratitae Merrem, 1813

[Superorder Apterygimorphae; incertae sedis]

Order Apterygiformes (Haeckel, 1866)

Family Apterygidae Gray, 1840

Order Dinornithiformes (Gadow, 1893)

[Family Anomalopterygidae (Archey, 1941)]

[Family Dinornithidae (Owen, 1843)]

Superorder Casuariimorphae; new taxon

Order Casuariiformes (Forbes, 1884)

Family Casuariidae Kaup, 1847

Family Dromaiidae Richmond, 1908

Superorder Struthionimorphae; new taxon

Order Aepyornithiformes (Newton, 1884)

Family Aepyornithidae Bonaparte, 1853

Order Struthioniformes (Latham, 1790)

Family Struthionidae Vigors, 1825

Family Rheidae (Bonaparte, 1853)

Cohort Neognathae Pycraft, 1900

Subcohort GalloanseraeSibley & Ahlquist, 1990

Superorder Galloanserimorphae (Sibley et al., 1988)

Order Galliformes (Temminck, 1820)

Suborder Craci Sibley et al., 1988; incertae sedis

Superfamily Megapodioidea (Lesson, 1831)

Family Megapodiidae Lesson, 1831

[Family Sylviornithidae Mourer-Chauviré & Balouet, 2005]

Superfamily Cracoidea (Vigors, 1825)

Family Cracidae Vigors, 1825

Suborder Phasiani (Vigors, 1825)

Superfamily Meleagridoidea (Gray, 1840)

Family Meleagrididae Gray, 1840

Superfamily Phasianoidea (Vigors, 1825); sedis mutabilis

Family Phasianidae (Vigors, 1825); sedis mutabilis

(Sub)Family Tetraonidae Vigors, 1825

Subfamily Perdicinae (Bonaparte, 1838)

Subfamily Odontophorinae Gould, 1844

Subfamily Phasianinae Vigors, 1825

Subfamily Numidinae Reichenbach, 1850

[Order Dromornithiformes Fürbringer, 1888]

Family Dromornithidae Vigors, 1825

[Order Diatrymiformes (Shufeldt, 1913)]

Family Diatrymidae Shufeldt, 1913

Order Anseriformes (Wagler, 1831)

Suborder Anhimae Wetmore & Miller, 1926

Family Anhimidae Stejneger, 1885

Suborder Anseres Wagler, 1831

Superfamily Anseranatoidea (Sclater, 1880)

Family Anseranatidae Sclater, 1880

Superfamily Anatoidea (Vigors, 1825)

[Family Presbyornithidae Wetmore, 1926]

Family Anatidae (Vigors, 1825)

Subfamily Anserinae Vigors, 1825

Subfamily Anatinae (Vigors, 1825)

Subcohort NeoavesSibley et al., 1988

Division Natatores Baird, 1858

Subdivision Pygopodo-tubinares; new taxon

Superorder Gaviomorphae; new taxon

Order Gaviiformes Wetmore & Miller, 1926

Family Gaviidae Allen, 1897

Order Podicipediformes (Fürbringer, 1888)

Family Podicipedidae Bonaparte, 1831

Superorder Procellariimorphae (Fürbringer, 1888)

Order Sphenisciformes Sharpe, 1891

Family Spheniscidae Bonaparte, 1831

Order Procellariiformes Fürbringer, 1888

Suborder Pelecanoidi (Gray, 1871)

Family Pelecanoididae Gray, 1871

Suborder Procellarae (Gadow, 1893)

Superfamily Oceanitoidea (Huxley, 1868)

Family Oceanitidae Forbes, 1882

Superfamily Procellarioidea (Fürbringer, 1888)

Family Procellariidae Vigors, 1825

Subfamily Procellariinae (Vigors, 1825)

Subfamily Pachyptilinae (Oliver, 1930)

Family Diomedeidae Gray, 1840

Subdivision Stegano-grallatores; new taxon

Superorder PelecanimorphaeHuxley, 1867

[Order Odontopterygiformes (Spulski, 1910)]

Family Odontopterygidae Lambrecht, 1933

Order Balaenicipitiformes (Sclater, 1924)

Suborder Balaenicipites (Sclater, 1924)

Family Balaenicipitidae (Sclater, 1924)

Order Pelecaniformes Sharpe, 1891

Suborder Phaethontes (Sharpe, 1891)

Family Phaethontidae Brandt, 1831

Suborder Steganopodes (Chandler, 1916)

Infraorder Fregatides (Sharpe, 1891)

Superfamily Fregatoidea (Garrod, 1874)

Family Fregatidae Garrod, 1874

Infraorder Pelecanides (Sharpe, 1891)

Parvorder Pelecanida (Sharpe, 1891); new rank

Family Pelecanidae Vigors, 1825

Parvorder Sulida (Reichenbach, 1849); new rank

Superfamily Suloidea (Reichenbach, 1849)

Family Sulidae Reichenbach, 1849

Superfamily Phalacrocoracoidea (Bonaparte, 1854)

Family Phalacrocoracidae (Bonaparte, 1854)

Family Anhingidae Ridgway, 1887

Superorder Ciconiimorphae (Garrod, 1874)

Order Ciconiiformes Garrod, 1874

Suborder Scopi (Bonaparte, 1853)

Family Scopidae (Bonaparte, 1853)

Suborder Ciconiae (Bonaparte, 1874)

Superfamily Ciconioidea (Sundevall, 1836)

Family Ciconiidae Sundevall, 1836

Family Phoenicopteridae Bonaparte, 1838

Superfamily Threskiornithoidea (Richmond, 1917)

Family Threskiornithidae Richmond, 1917

Family Plataleidae (Bonaparte, 1838)

Order Ardeiformes (Wagler, 1831)

Family Cochleariidae Ridgway, 1887

Family Ardeidae Vigors, 1825

Subfamily Botaurinae Bock, 1956

Tribe Botaurini (Bock, 1956)

Tribe Tigriornithini Bock, 1956

Subfamily Ardeinae (Vigors, 1825)

Tribe Nycticoracini Bock, 1956

Tribe Ardeini Bock, 1956

Division Terrestrornithes; new taxon

Subdivision Telmatorae (Lowe, 1931)

Superorder CharadriimorphaeHuxley, 1867

Order Gruiformes (Bonaparte, 1854)

Suborder Cariamae (Wagler, 1830)

Infraorder Otides Sibley et al., 1988

Family Otididae Gray, 1840

Infraorder Cariamides (Fürbringer, 1888)

Superfamily Cariamoidea (Gray, 1853); sedis mutabilis

[Family Bathornithidae Wetmore, 1933]

Family Cariamidae Bonaparte, 1853

[Family Phorusrhacidae (Ameghino, 1899)]

Suborder Eurypygae (Fürbringer, 1888)

Infraorder Eurypygides Sibley et al., 1988

Family Eurypygidae Selby, 1840

Infraorder Rhynochetides Sharpe, 1891

Family Rhynochetidae Newton, 1868

Family Aptornithidae Bonaparte, 1856

Suborder Grues Bonaparte, 1854

Superfamily Psophioidea (Bonaparte, 1831)

Family Psophiidae Bonaparte, 1831

Superfamily Gruoidea (Vigors, 1825)

Family Aramidae Bonaparte, 1854

Family Gruidae Vigors, 1825

Order Turniciformes (Huxley, 1868); incertae sedis

Family Turnicidae (Gray, 1840)

Family Mesitornithidae Wetmore, 1960

Order Ralliformes (Reichenbach, 1854)

Family Heliornithidae Gray, 1841

Family Rallidae (Reichenbach, 1854)

Order Charadriiformes (Fürbringer, 1888)

Suborder Pedionomae (Gadow, 1893)

Family Pedionomidae Gadow, 1893

Suborder Parrae (Gadow, 1893)

Family Jacanidae Stejneger, 1885

Family Rostratulidae Ridgway, 1919

Suborder Limicolae (Beddard, 1898)

Infraorder Dromaides (Sharpe, 1891)

Family Dromadidae Gray, 1840

Infraorder Scolopacides (Strauch, 1978)

Superfamily Thinocoroidea (Gray, 1845)

Family Thinocoridae (Gray, 1845)

Superfamily Scolopacoidea (Vigors, 1825)

Family Scolopacidae Vigors, 1825

Family Phalaropodidae Bonaparte, 1831

Infraorder Charadriides (Huxley, 1867); incertae sedis

Superfamily Charadrioidea (Vigors, 1825)

Family Charadriidae Vigors, 1825

Superfamily Glareoloidea (Brehm, 1831)

Family Glareolidae Brehm, 1831

Subfamily Glareolinae Brehm, 1831

Subfamily Cursoriinae Gray, 1840

Superfamily Burhinoidea (Mathews, 1912)

Family Burhinidae Mathews, 1912

Superfamily Haematopoidea (Bonaparte, 1838)

Family Haematopidae Bonaparte, 1838

Subfamily Haematopodinae (Bonaparte, 1838)

Subfamily Ibidorhynchinae Bonaparte, 1856

Family Recurvirostridae (Bonaparte, 1831)

Subfamily Recurvirostrinae Bonaparte, 1831

Subfamily Himantopodinae Reichenbach, 1849

Tribe Himantopodini Sibley et al., 1988

Tribe Cladorhynchini; new taxon

Suborder Lari Sharpe, 1891; incertae sedis

Infraorder Chionidides Sharpe, 1891

Family Chionididae Lesson, 1828

Infraorder Alcides (Sharpe, 1891)

Family Alcidae (Vigors, 1825)

Infraorder Larides (Sharpe, 1891)

Superfamily Laroidea (Bonaparte, 1831)

Family Stercorariidae Gray, 1870

Family Laridae (Bonaparte, 1831)

Subfamily Larinae Bonaparte, 1831

Subfamily Sterninae Bonaparte, 1838

Superfamily Rynchopoidea (Bonaparte, 1838)

Family Rynchopidae (Bonaparte, 1838)

Subdivision Dendrornithes (Verheyen, 1961)

Section Raptores Baird, 1858

Superorder Falconimorphae (Seebohm, 1890)

Order Falconiformes Seebohm, 1890

[Suborder Teratornithi (Miller, 1909)]

Suborder Cathartae (Coues, 1824)

Family Cathartidae (Lafresnaye, 1839)

Suborder Accipitres (Vieillot, 1816)

Infraorder Serpentariides (Seebohm, 1890)

Family Sagittariidae Finsch & Hartlaub, 1870

Infraorder Falconides (Sharpe, 1874)

Superfamily Falconoidea (Vigors, 1824)

Family Falconidae Vigors, 1824

Subfamily Falconinae Vigors, 1824

Subfamily Polyborinae Lafresnaye, 1839

Family Pandionidae Sclater & Salvin, 1873

Superfamily Accipitroidea (Vieillot, 1816)

Family Accipitridae (Vieillot, 1816)

Subfamily Accipitrinae (Vieillot, 1816)

Subfamily Gypaetinae (Vieillot, 1816)

Order Strigiformes (Wagler, 1830)

Family Tytonidae (Mathews, 1912)

Subfamily Tytoninae Mathews, 1912

Subfamily Phodilinae Beddard, 1898

Family Strigidae (Gray, 1840)

Section AnomalogonatesGarrod, 1874

Subsection CoccygesHuxley, 1867; incertae sedis

Superorder CuculimorphaeSibley et al., 1988

Order Opisthocomiformes (L’Herminer, 1837)

Family Opisthocomidae Swainson, 1837

Order Cuculiformes (Wagler, 1830)

Suborder Musophagi Seebohm, 1890

Family Musophagidae Bonaparte, 1831

Suborder Cuculi Wagler, 1830

Family Cuculidae Vigors, 1825; sedis mutabilis

Subfamily Neomorphinae Shelley, 1891

Subfamily Centropodinae Horsfield, 1823

Subfamily Crotophaginae Swainson, 1837

Subfamily Cuculinae (Vigors, 1825)

Subfamily Phaenicophacinae (Horsfield, 1822)

Superorder Psittacimorphae (Huxley, 1867); incertae sedis

Order Psittaciformes (Wagler, 1830); sedis mutabilis

Family Nestoridae (Bonaparte, 1850)

Family Psittacidae (Illiger, 1811)

Family Cacatuidae Gray, 1840

Family Loriinidae Selby, 1836

Order Columbiformes (Garrod, 1874)

Suborder Pterocletes (Boucard, 1876)

Family Pteroclidae Bonaparte, 1831

Suborder Columbae (Latham, 1790)

Family Columbidae (Illiger, 1811)

Subfamily Columbinae (Illiger, 1811)

Subfamily Didunculinae Gray, 1848

Subfamily Gourinae Gray, 1840

Family Raphidae Wetmore, 1930

Subsection Incessores Baird, 1858

Superorder CypselomorphaeHuxley, 1867

Order Caprimulgiformes (Ridgway, 1891)

Suborder Aegotheli Sibley et al., 1988

Family Aegothelidae (Bonaparte, 1853)

Suborder Caprimulgi Ridgway, 1881; sedis mutabilis

Family Caprimulgidae Vigors, 1825

Family Nyctibiidae (Chenu & Des Murs, 1851)

Family Podargidae (Gray, 1840)

Family Steatornithidae (Gray, 1846)

Order Apodiformes Peters, 1940

Suborder Hemiprocni; new taxon

Family Hemiprocnidae Oberholser, 1906

Suborder Apodi (Peters, 1940)

Family Apodidae (Hartert, 1897)

Subfamily Cypselinae Bonaparte, 1838

Subfamily Apodinae Hartert, 1897

Family Trochilidae Vigors, 1825

Subsection Trogones; new name

Superorder Trogonomorphae; new taxon

[Order Sandcoleiformes Houde & Olson, 1992]

Family Sandcoleidae Houde & Olson, 1992

Order Coliiformes (Murie, 1872)

Family Coliidae (Swainson, 1836)

Order Trogoniformes Wetmore & Miller, 1926

Family Trogonidae Lesson, 1828

Subsection Pico-clamatores; new name

Superorder PasserimorphaeSibley et al., 1988; sedis mutabilis

Order Coraciiformes Forbes, 1884

Suborder Bucerotes Fürbringer, 1888

Infraorder Upupides (Seebohm, 1890)

Family Upupidae Bonaparte, 1831

Family Phoeniculidae Sclater, 1924

Infraorder Bucerotides (Fürbringer, 1888)

Family Bucerotidae (Gray, 1847)

Suborder Halcyones (Forbes, 1884)

Superfamily Motmotoidea (Gray, 1840)

Family Motmotidae Gray, 1840

Superfamily Alcedinoidea (Stejneger, 1885)

Family Todidae Vigors, 1825

Family Alcedinidae (Bonaparte, 1831)

Subfamily Alcedininae Bonaparte, 1831

Subfamily Halcyoninae (Vigors, 1825)

Suborder Coracii (Forbes, 1884)

Infraorder Meropides (Fürbringer, 1888)

Family Meropidae Vigors, 1825

Infraorder Coraciides (Wetmore & Miller, 1926)

Superfamily Coracioidea (Vigors, 1825)

Family Coraciidae Vigors, 1825

Superfamily Leptosomatoidea (Bonaparte, 1850)

Family Leptosomatidae Bonaparte, 1850

Family Brachypteraciidae (Sharpe, 1892)

Order Piciformes (Meyer & Wolf, 1810)

Suborder Galbulae (Fürbringer, 1888)

Family Galbulidae Bonaparte, 1831

Family Bucconidae Boie, 1826

Suborder Pici (Meyer & Wolf, 1810)

Superfamily Capitonoidea (Bonaparte, 1840)

Family Capitonidae Bonaparte, 1840

Family Rhamphastidae Vigors, 1825

Superfamily Picoidea (Vigors, 1825)

Family Indicatoridae Swainson, 1837

Family Picidae Vigors, 1825

Subfamily Jynginae Bonaparte, 1838

Subfamily Picinae Bonaparte, 1838

Order Passeriformes (Linnaeus, 1758)

Suborder Menurae (Sharpe, 1891)

Family Menuridae (Lesson, 1828)

Suborder Passeres Linnaeus, 1758

Infraorder Tyrannides Sibley et al., 1988

Family Tyrannidae Vigors, 1825

Family Pittidae Swainson, 1831

Infraorder Passerides (Linnaeus, 1758)

Parvorder Corvida Sibley et al., 1988

Family Ptilinorhynchidae Gray, 1841

Family Corvidae Vigors, 1825

Parvorder Passerida Sibley et al., 1988

Family Bombycillidae Swainson, 1831

Family Paridae Vigors, 1825

Family Passeridae Illiger, 1811

REFERENCES