Proposal (461a, b) to South American Classification Committee

Remove Busarellus from buteonine genera in linear sequence of Accipitridae and rearrange linear sequence on non-buteonine genera

Part a: Move Busarellus to follow Ictinia.

Effect on South American CL: This proposal would overhaul our linear sequence of accipitrid genera.

Background & New Information & Analysis: The genus Busarellus has traditionally placed within the buteonine genera of the Accipitridae, but this is clearly not correct. The relevant Note from our SACC classification is:

17a. Olson (1982) found morphological evidence that Busarellus may be more closely related to a group of largely Old World genera (Milvus, Haliastur, Haliaetus, Ichthyophaga) than to the New World genera with which is traditionally associated in linear sequences (e.g., Friedmann 1950, Meyer de Schauensee 1970), and this is reflected in the linear sequence of the AOU (1998). Genetic data (Griffiths et al. 2007, Raposo do Amaral et al. 2009) also indicate that it is not closely related to any buteonine genera, where traditionally place, but rather closer to kites. SACC proposal needed.

Olson (1982) used toe morphology to predict that Busarellus is not closely related to buteonine hawks but rather to Old World kites. Raposo do Amaral et al.’s (2009) phylogeny (see proposals 459, 2460) clearly indicated that Busarellus is not within the buteonines; because their analysis targeted that group, only a few outgroup taxa were used. Nevertheless, a well-supported node group Busarellus with Geranospiza and Rostrhamus sociabilis to the exclusion of all the buteonines as well as Ictinia, Old World Butastur; and Haliaeetus; sampling did not include the taxa mentioned by Olson, but regardless, Busarellus cannot be considered buteonine. The relevant portion of their tree (from their Fig. 3) is pasted in here:

Therefore, at present, there is no evidence that Busarellus is a buteonine hawk. Although which non-buteonines are actually its closest relatives cannot be determined without additional taxon sampling, it is clear that it must be moved elsewhere. Griffiths et al. (2007) sampled extensively outside the buteonines but sequenced only one (nuclear) gene (RAG-1). Their maximum-likelihood tree places Busarellus with Butastur and Ictinia; however, support for the structure of this whole region of the tree is weak to nonexistent, and that further sampling could show it to be the sister of any of the genera in the tree from Rostrhamus to Geranospiza, and possibly even others more basal in the group:

Thus, we could place it anywhere in the current sequence of non-elanine kite genera:

Rostrhamus sociabilis Snail Kite
Helicolestes hamatus Slender-billed Kite
Harpagus bidentatus Double-toothed Kite
Harpagus diodon Rufous-thighed Kite
Ictinia mississippiensis Mississippi Kite
Ictinia plumbea Plumbeous Kite

Note that the RAG-1 tree conflicts strongly with the Raposo do Amaral et al. tree in terms of which genera are closest to Busarellus: Rostrhamus or Geranospiza but definitely not Ictinia in Raposo do Amaral et al., but most likely (statistically) Ictinia or Geranospiza and much less likely Rostrhamus in Griffiths et al. Which one is correct requires additional taxon and gene sampling. If part b (below) is rejected, arbitrarily placing it after Ictinia in our current sequence in hopes that the unsupported structure in RAG-1 tree turns out to be correct would be my recommendation.

We could also place it next to Geranospiza, currently between Accipiter and Leucopternis, but that placement of Geranospiza makes little sense in light of the results of Griffith et al. (2007). This leads to the second part of the proposal: an overhaul of the sequence of genera

Part b: rearrange sequence of genera to follow Griffiths et al. (2007)

Here I propose that we change our sequence so that it corresponds more closely with that of Griffiths et al. (2007), which we already used to move some of the kite genera (e.g., Elanus and Gampsonyx) to the beginning of the sequence. My rationale is that (a) as long as Busarellus must be moved, and (b) the likely close relatives to Busarellus are separated by monophyletic groups such as Accipiter + Circus, we might as well adopt the general structure of the Griffiths et al. tree. This would also place Geranospiza among the other candidate genera for the sisters to Busarellus (instead of between Accipiter and the buteonines). There will always be questions about trees based on a single gene, but the support for the deep nodes in the RAG-1 tree are strong, and the groupings themselves are all sensible. Additional data might require changes to the sequence, but basing our sequence on the RAG-1 tree reflects the best data currently available. Finally, the current linear sequence of groups of genera is based on mostly historical momentum, which is largely a concoction of antiquated ideas about “primitive” and “advanced” hawks – therefore, ANY new sequence based on modern data would have more published support and testable hypotheses than the historical one.

The RAG-1 tree:

and might be better visualized from their more schematic version:

Adopting the Griffiths et al. tree would involve one major rearrangement: bringing forward in the sequence the eagle genera Morphnus, Harpia, and Spizaetus, thus leaving the buteonines last (see proposals 459 and 460 for the proposed realignments in those genera). The only other “major” change would be the above-mentioned move of Geranospiza forward to the kites.

Our current sequence is as follows (with Busarellus and putative relatives in red):

Elanus
Gampsonyx
Chondrohierax
Leptodon
Elanoides
Rostrhamus
Helicolestes
Harpagus
Ictinia
Circus
Accipiter
Geranospiza
Leucopternis
Buteogallus
Harpyhaliaetus
Busarellus
Geranoaetus
Parabuteo
Buteo
Morphnus
Harpia
Spizaetus

A sequence consistent with Griffiths et al. that minimizes the number of changes is (note that proposals 459 and 460 would affect the buteonine genera Leucopternis through Buteo, but the group itself would stay intact); taxa whose position is changed are shown in blue:

Elanus
Gampsonyx
Chondrohierax
Leptodon
Elanoides
Morphnus
Harpia
Harpagus
Circus
Accipiter
Rostrhamus
Helicolestes
Busarellus
Ictinia
Geranospiza
Leucopternis
Buteogallus
Harpyhaliaetus
Geranoaetus
Parabuteo
Buteo

Thus, the beginning and end of the sequence stays relatively stable. The harpy eagles and booted eagles are brought forward, as is Harpagus slightly and Circus + Accipiter slightly. The milvine kites, which would include Rostrhamus through Geranospiza, are all now together, although whether they form a monophyletic group is questionable from the RAG-1 tree. The main accomplishment is bringing Busarellus out of the buteonines, placing it among its potential relatives. The proposed new sequence also rescues the harpy eagle types from any association with the monophyletic buteonines. I considered moving Harpagus and Rostrhamus (with unproven close relative Helicolestes) slightly to match the RAG-1 branching pattern, but there is little or no support for the nodes on that section of the tree. Tweaks welcomed.

Recommendation: The proposed sequence is consistent with the phylogenetic hypotheses of Griffiths et al.’s RAG-1 tree, and extricates Busarellus from the buteonines. Pending tweaks, I recommend a YES vote to incorporate the best phylogenetic data available into our sequence. The vote is broken into two parts:

a) = to move Busarellus out of buteonines
b) = to use proposed sequence above.

Literature Cited:

GRIFFITHS, C. S., G. F. BARROWCLOUGH, J. G. GROTH, AND L. MERTZ. 2007. Phylogeny, diversity and classification of the Accipitridae based on DNA sequences of the RAG-1 exon. J. Avian Biology 38: 587-602.

OLSON, S. L. 1982. The distribution of fused phalanges of the inner toe in the Accipitridae. Bull. British Ornithol. Club 102: 8-12.

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 solicited from Fabio Raposo: “Two important references may make the difference here. Based on a representative taxonomic sampling of the family Accipitridae, Lerner & Mindell 2005 recovered species of Haliaeetus, Ichthyophaga, Haliastur and Milvus as the closest relatives of the "core" buteonine hawks + Ictinia, Geranospiza and Rostrhamus (Fig. 1 of that paper), based on two mitochondrial markers.  A smaller taxonomic sampling of the mitochondrial dataset plus one nuclear intron provides a similar picture with even better support (Fig. 2). "Kites" seem to represent an artificial group, since kite genera appear all over their accipitrid tree. So, it is possible that Geranospiza, Rostrhamus, Ictinia and Butastur are indeed better thought as buteonine hawks - and Busarellus and Butastur also seem to be part of this clade (see Lerner et al. 2008, besides Amaral et al. 2009).

Lerner, H.R., Mindell, D.P., 2005. Phylogeny of eagles, Old World vultures, and other Accipitridae based on nuclear and mitochondrial DNA. Molecular Phylogenetics and Evolution 37, 327–346.

Lerner, H.R.L., Klaver, M.C., Mindell, D.P., 2008. Molecular phylogenetics of the buteonine birds of prey (Aves, Accipitridae). Auk 125, 304–315.