Proposal (241) to South American Classification Committee
Remove Cathartidae from the Ciconiiformes
Background: The New World vultures have been placed as a family in the Falconiformes through most of their taxonomic history. Morphological and behavioral data (Ligon 1967, König 1982, Rea 1983) that showed intriguing similarities with the storks, Ciconiidae, were generally considered examples of convergence or chance until Sibley & Ahlquist's DNA-DNA hybridization data also supported a stork-cathartid relationship. Harshman's (1994) reanalysis of the DNA-DNA hybridization also confirmed the conclusions of the original analysis. Biochemical data from Jacob (1983) and Wink (1995) were also consistent with a stork-vulture relationship. On the basis of these congruent data sets, the AOU Checklist (1998) removed the Cathartidae to the Ciconiiformes.
New information: Griffiths (1994) pointed out weaknesses in the data used to support a stork-cathartid relationship, and also provided data from syringeal morphology that were consistent with placement in the Falconiformes. Analyses of DNA sequence data have failed to find any close relationship between Ciconiidae and Cathartidae. The most recent of these (Ericson et al. 2006) found that the two families are in different groups of the Neoaves, with > 95% Bayesian support for the two nodes that define each group. The group of orders in which the Cathartidae falls includes such diverse groups as owls, the coraciiform assemblage, parrots, seriemas, trogons, colies, Passeriformes, as well as the Accipitridae and Falconidae (which are not closely related). However, the Cathartidae branch does not associate closely with any other branch in this group. The group in which the Ciconiidae falls includes Procellariiformes, penguins, loons, rails, cranes, bustards, most "Pelecaniformes," cuckoos, and turacos, as well as all the herons and ibises; parallel to the Cathartid, the Ciconiidae do not fall unambiguously within any cluster of orders, including that which contains herons, ibis, and pelicans. Previously, Cracraft et al. (2004) had also failed to find any support for a stork-cathartid relationship, and in fact found some support for a cathartid-accipitrid relationship as well as results fairly consistent with those of Ericson et al. for placement of the Ciconiidae. Fain & Houde (2004) found that Ciconiidae was the sister group to Ardeidae, and that these two grouped with a cluster of groups that included Threskiornithidae, Musophagidae, Cuculidae, Procellariiformes, and Gaviiformes; the Cathartidae branch is part of a large polytomy and does not for certain associate with Accipitridae or Falconidae.
Analysis and Recommendation: I think there are other analyses with similar results albeit weaker taxon-sampling, but for the purpose of this proposal, the above three suffice to show that sequence data that support a Ciconiiformes-Cathartidae relationship are nonexistent, and that regardless of the differences of the branching patterns among the studies, these two groups are not part of the same cluster of orders. Also, support for a relationship to any of the traditional Falconiformes (itself a non-monophyletic group) is weak at best. To maintain the Cathartidae within the Ciconiiformes, despite the earlier support for this relationship, ignores all recent genetic data. Therefore, I strongly recommend removal of Cathartidae from Ciconiiformes.
For now, this proposal asks for YES/NO vote on removal, without regard to where to place Cathartidae. If it passes, I'll then construct a proposal for the alternatives, namely (1) family Incertae Sedis, (2) return to traditional position in Falconiformes, or (3) placement in their own order, Cathartiformes. Concerning the three alternatives, I lean towards the latter because (1) ordinal rank for New World vultures is already in use in many classifications (1100+ hits on GOOGLE yesterday for "Cathartiformes"), (2) it accomplishes the same thing as Incertae Sedis by removing the group from within any existing order, and (3) New World vultures are diagnosable in fossil record as far back as at least the Upper Eocene and thus constitute a lineage as old as most avian bird orders. But that's for a subsequent proposal.
Literature Cited (partial -- See SACC Biblio for rest):
CRACRAFT, J., F. K. BARKER, M. BRAUN, J. HARSHMAN, G. J. DYKE, J. FEINSTEIN, S. STANLEY, A. CIBOIS, P. SCHIKLER, P. BERESFORD, J. GARCÍA-MORENO, M. D. SORENSON, T. YURI, AND D. P. MINDELL. 2004. Phylogenetic relationships among modern birds (Neornithes): toward an avian tree of life. Pp. 468-489 in "Assembling the Tree of Life" (Cracraft, J. and Donoghue, M. J., eds.). Oxford University Press.
ERICSON, P. G. P., C. L. ANDERSON, T. BRITTON, A. ELZANOWSKI, U. S. JOHANSSON, M. KALLERRSJO, J. I. OHLSON, T. J. PARSONS, D. ZUCCON, AND G. MAYR. 2006. Diversification of Neoaves: integration of molecular sequence data and fossils. Biology Letters online: 1-5.
FAIN, M. G., AND P. HOUDE. 2004. Parallel radiations in the primary clades of birds. Evolution 58: 2558-2573.
GRIFFITHS, C. S. 1994. Monophyly of the Falconiformes based on syringeal morphology. Auk 111: 787-805.
LIGON, J. D. 1967. Relationships of the cathartid vultures. Univ. Mich. Mus. Zool., Occas. Pap. 651.
Van Remsen, August 2006
Comments from Stiles: "YES. After reading Griffiths and her inability to duplicate Sibley & Ahlquist's tree, I am more inclined to discount the cathartid-ciconiid relationship. Rather a shame, as it introduced a nice little bit of symmetry - the ciconiid brings the baby, the cathartid takes it away when itęs done ... all in the family (or order), as it were.. I also lean toward Cathartiformes, as a more satisfactory alternative than returning these birds to Falconiformes, to which their relationship is decidedly distant."
Comments from Zimmer: "YES. I too, would support placement in a separate order (Cathartiformes) as opposed to placement in an existing order."
Comments from Stotz: "YES. This is long overdue. It became clear pretty early after Cathartidae was moved into Ciconiiformes that the characters joining it with the storks were primitive characters."
Comments from Pacheco: "YES. Este estudo demonstra que a relaćčo de proximidade com Ciconiidae, tčo patente nos outros estudos, nčo pode ser corroborada. Voto pela adoćčo de Cathartiformes, tal qual em uso igualmente pelo CBRO."