Proposal (460) to South American Classification Committee


Revise generic boundaries in the Buteo group


Effect on South American CL: This would revise generic boundaries extensively in Buteo and Leucopternis.


Background & New Information:  For several years, we’ve had plenty of indication that the current boundaries of the genera Buteo, Leucopternis, and relatives in our current classification are a mess.  Raposo do Amaral et al. (2009) have produced a comprehensive phylogeny of buteonine hawks, and their data will form the primary basis for this proposal.  Findings from earlier papers (see Notes below) are largely consistent with Raposo do Amaral et al. (2009) and will not be discussed further.  Two of the relevant Notes from our SACC classification are:


12a. Genetic data (Raposo et al. 2006, Lerner et al. 2008, Raposo do Amaral et al. 2009) indicate that the genus Leucopternis is polyphyletic; as currently defined, Leucopternis includes at least three distinct groups that are not each others' closest relatives: (1) L. melanops, L. kuhli, and L. semiplumbeus; (2) L. albicollis, L. occidentalis, and L. polionotus; (3) L. plumbeus, L. schistaceus, L. lacernulatus, which are intermingled within a group with Buteogallus and Harpyhaliaetus; and (4) L. princeps, whose placement is uncertain.  Raposo do Amaral et al. (2009) recommended placing princeps in a monotypic genus Morphnarchus, plumbeus in a new monotypic genus Cryptoleucopteryx, schistaceus in Buteogallus, and lacernulatus in a new monotypic genus Amadonastur.  SACC proposals needed.


18. Genetic data (Riesing et al. 2003) indicate that Geranoaetus is the sister taxon to Buteo polyosoma/B. poecilochrous and that maintenance of a monotypic genus is not warranted; it had been placed in Buteo formerly (e.g., Wetmore 1933, Hellmayr & Conover 1949, Friedmann 1950), but recent authors have generally followed Amadon (1963), who suggested that it might be closer to Buteogallus or Leucopternis than to Buteo. Clark (2006) disputed Amadon's rationale for maintaining it is a genus separate from Buteo. SACC proposal to merge Geranoaetus into Buteo did not pass. New genetic data (Lerner et al. 2008) provide even stronger evidence for merger of Geranoaetus, at least as currently defined, because it is the sister species to B. polyosoma. SACC proposal to merge Geranoaetus into Buteo did not pass.  Raposo do Amaral et al. (2009) further confirmed that Geranoaetus is the sister to Buteo polyosoma sensu lato.  SACC Proposal needed.



Raposo do Amaral et al.’s (2009) taxon sampling (105 specimens, 54 species) and gene sampling (6000 bp of 9 genes, mitochondrial and nuclear) is exemplary.  I doubt that anyone will produce a better data set anytime soon.  This proposal deals only with their Group H, whose monophyly has excellent support; the relevant portion of their tree (from their Fig. 3) is pasted in here (sorry for the poor resolution of the screen grab):




The Buteo group itself (Group G) is strongly supported as a monophyletic group as is sister relationship to the Buteogallus group (Group H).  It includes everything in our current classification in Buteo plus Parabuteo, Geranoaetus, and most Leucopternis (minus the 4 species that are part of Group H; see Proposal 459).


Analysis and Recommendation: Virtually every critical node in Group G’s tree has strong support.  Our current Buteo and Leucopternis are both polyphyletic, and so changes must be made.  So, the only point of real discussion is the subjective exercise of how broadly to delimit the genera.  Raposo do Amaral et al. have defined these very narrowly.  However, one option would be to expand Buteo to include all species in Group G.  Even the two outliers, “Leucopternisprinceps (placed by Raposo do Amaral et al. in a resurrected monotypic genus Morphnarchus) and Buteo magnirostris (placed by Raposo do Amaral et al. in a resurrected monotypic genus Rupornis), don’t really stand out (to me anyway) as beyond the range of variation encompassed even with a narrow Buteo).  In my subjective view, there is so much variation among the species in narrowly defined Buteo in terms of color, shape, behavior, and size that adding the other 6 genera recognized by Raposo do Amaral et al. does not really add to its phenotypic heterogeneity.  Narrow Buteo includes heavy, robust species of open country such as B. regalis and small chunky species of forest such as B. platypterus.  It includes B. nitidus, long placed in a monotypic genus Asturina.  It includes blackish species such as B. albonotatus, gray-plumaged species such as B. nitidus, and relatively pale species (ventrally) such as B. brachyurus.  It also includes all Old World Buteo sampled so far.  More importantly, Raposo do Amaral’s et al.’s Buteo is not unequivocally monophyletic if nitidus is included, and in fact one could make a case for resurrection of Asturina based on weak support for the node that links it to core Buteo.


A YES vote on this proposal would be for adopting Raposo do Amaral et al. as is, namely recognize 7 genera (in the following sequence) for the group instead of 4, as follows:


Morphnarchus (monotypic; formerly L. princeps)

Rupornis (monotypic; formerly B. magnirostris)

Parabuteo (unicinctus plus leucorrhous)

Geranoaetus (to also include B. polyosoma and B. albicaudatus, a combination we’ve voted down previously)

Pseudastur (resurrected for L. albicollis/occidentalis; this might actually become monotypic given that albicollis is evidently paraphyletic with respect to occidentalis in their tree)

Leucopternis (melanops, kuhli, semiplumbeus)

Buteo (nitidus, platypterus, albigula, brachyurus, swainsoni, galapagoensis, albonotatus, ventralis)


A NO vote would be to broaden generic boundaries, from small tweaks to as much as including everything in Buteo.  If this proposal fails, I’ll write additional proposals to take into account broader generic limits.  I do not have a strong recommendation, but I am going to vote NO, given the points made about on the heterogeneity of narrow Buteo.  Because delimiting genera is a subjective exercise as long as each is monophyletic, I hope to hear others opinions, and my NO vote is not a firm one.


Literature Cited:

RAPOSO DO AMARAL, F., F. H. SHELDON, A. GAMAUF, E. HARING, M. RIESING, L. F. SILVEIRA, AND A. WAJNTAL.  2009.  Patterns and processes of diversification in a widespread and ecologically diverse avian group, the buteonine hawks (Aves, Accipitridae). Molecular Phylogenetics and Evolution 53: 703-715.


Van Remsen, August 2010





Comments from Bret Whitney: “I like Rupornis for Roadside Hawk.  It’s a bizarre bird, frankly, a highly successful New World raptor not quite like anything else, intermediate morphologically and behaviorally between Accipiter and small Buteo with distinctive vocalization patterns and behaviors (nothing else is even close to as consistently vocal as Rupornis, for instance).  To lump it into Buteo just obscures the picture, in my opinion.  Yeah, I know, having such different types as Ferrug and Broad-winged in one genus is also “uncomfortable” -- but I guess the best path forward is to split out only what seems clearly divergent when ornithologists can agree that there’s good concordance among various datasets.”


Comments from Robbins: “NO, to be consistent with my comments under proposal 459.  I would also add vocalizations to Van’s comments about variation in Buteo; it is considerable among even closely allied species.”


Comments solicited from Fabio Raposo: “Thanks very much to Van and the committee for requesting comments and letting us be part of this discussion.  A few important points on this one (for more general comments on the buteonine proposals, please see SACC #459):


Yes, there is a lot of heterogeneity in Buteo, even when narrowly defined as we did.  However, the other monophyletic groups treated in our paper as genera are far from heterogeneous, since they have clear identities as groups (see below, we explicitly indicate why we strongly support that point of view). Exactly for this reason, we don’t understand why an enormous, uninformative and undiagnosable Buteo would be a better option - why add more heterogeneity to Buteo, when we can indicate specific characteristics of each of those clades? Furthermore, our classification is "testable" - those with access to a good bird collection will clearly see those groups/characters in 10 minutes with the birds on a table (this is something that we did over and over again before closing this paper), while a catchall Buteo has no concrete identity as a group. A huge, undiagnosable, species-hungry Buteo (by the way, rejected by the committee twice when dealing with Geranoaetus - see SACC #282 and SACC #387) would be unnecessarily in the top 10 largest avian genera ever.


genus Leucopternis (L. semiplumbeus, L. kuhli and L. melanops): generally speaking, small, black and white Neotropical buteonine hawks, found exclusively in forests. Plain white underparts, narrow black streaks or solid gray on the sides of the neck; short wings (primaries extending to less than half of tail); plain black tails with one medial white band in adults (and two white bands in young birds).


genus Pseudastur (P. albicollis, P. occidentalis and P. polionotus): generally speaking, large, black and white Neotropical buteonine hawks - found exclusively in forests. White underparts, plain white tails with a single black band, and secondaries and tertials with conspicuous white tips.


-- genus Geranoaetus (G. melanoleucus, G. polyosoma and G. albicaudatus): generally speaking, large, dark headed open vegetation buteonine hawks - mostly Neotropical. Densely feathered faces, solid black or gray sides of the neck, white underparts with flanks (and sometimes belly) finely streaked, long primaries that reach the tip of the tail; densely barred tails with narrow blackish bars and a subterminal large black band, or entirely solid black.


-- genus Parabuteo (P. unicinctus and P. leucorrhous): Plumage is strikingly similar between those two species - overall black or dark brown coloration, white rump and undertail coverts, and mostly rufous or light brown thighs.


On the other hand, Rupornis magnirostris (as well indicated by Bret) and Morphnarchus princeps are very divergent species, isolated from the rest of the tree by lots of evolutionary time and autapomorphic characteristics - and that’s why monotypic genera fit best for those two.  Finally, a correction to the proposal: Pseudastur also includes L. polionotus, which is clearly a good species, sister to the Amazonian subspecies of L. albicollis. So even if the trans-Andean L. albicollis subspecies are lumped with L. occidentalis, there still would be three species (L. polionotus, L. albicollis from Amazonia, and trans-Andean L. albicollis+L. occidentalis) - so nothing to worry about this genus becoming monotypic.