Proposal (116) to South American Classification Committee

Change linear sequence of genera in Columbidae

Effect on South American CL: This would alter our linear sequence of species in Columbidae to reflect the results of molecular phylogenetics.

Background: Our current sequence of New World genera in the Columbidae is the traditional one (e.g., Meyer de Schauensee 1970) that starts with Columba and ends with Geotrygon. The ground doves (Columbina, Claravis, Uropelia, Metriopelia) are placed roughly in the middle of the sequence.

I cannot find any published rationale to support this sequence.

New information: Regardless of whatever rationale was behind the traditional sequence, Johnson (2004), using mitochondrial (cytochrome b and ATPase8) and nuclear genes (FIB7), found that two genera of American ground doves (Columbina [including Scardafella] and Metriopelia) were basal to all other pigeon genera sampled, including a variety of Old World and New World genera, the latter including Patagioenas, Geotrygon, and Leptotila. Support for this branching pattern was very strong (> 90% bootstrap and Bayesian probability). Although this result is fairly shocking, the support for the deep nodes throughout the tree is excellent.

I asked Kevin Johnson how confident that he was in the results and whether ground doves are basal and should be listed first. Kevin wrote: "Yes, with respect to all other NW taxa, I am very confident that the NW Ground doves are basal. Placing them at the front of a linear sequence for NW taxa is quite reasonable. My only reservation would be if someone asked me to construct a linear sequence for ALL Columbiformes, because the basal relationships among some of the major lineages worldwide are still not resolved to my satisfaction."

Analysis: Our classification obviously should attempt to reflect phylogeny wherever possible, so we need to change our sequence to place the ground doves at one end of the sequence. A basal phylogenetic position of the ground doves, by convention, requires that they be moved to the beginning of the sequence. To do this with minimum disturbance to current sequence, the Columbidae sequence should be as follows:

Columbina
Claravis
Metriopelia
Uropelia

then

Columba .... [etc.]

Genetic support for inclusion of Claravis in the clade is in Johnson & Clayton (2000). Kevin Johnson has unpublished data that also places Uropelia in this clade. The close relationship of these two genera to Columbina and Metriopelia has been suspected for a long time, so I think we're safe on many counts for treating them as a monophyletic group.

Recommendation: I vote YES on this because it represents one of the few cases in which we have a good data-set to support the sequence of genera in any family in our classification.

Literature Cited:

JOHNSON, K. P. 2004. Deletion bias in avian introns over evolutionary timescales. Molecular Biology & Evolution 21: 599-602.

JOHNSON, K. P., AND D. H. CLAYTON. 2000a. Nuclear and mitochondrial genes contain similar phylogenetic signal for pigeons and doves (Aves: Columbiformes). Molecular Phylogenetics & Evolution 14: 141-151.

Van Remsen, April 2004 (in consultation with Kevin Johnson)

Comments from Jaramillo: "YES Data are strong, and the new linear sequence reflects this new phylogenetic information."

Comments from Stiles: "YES, for the same reasons as in the preceding."

Comments from Zimmer: "YES. The evidence seems strong."

Comments from Nores: "Sí; el hecho de que la posición basal de Columbina y Metriopelia y la inclusión de Claravis y Uropelia en este clade esté basado en dos tipos coincidentes de análisis (nuclear y mitocondrial), dan un buen soporte para la nueva secuencia."