Proposal (47) to South American Check-list Committee:

Separation from Columba of New World pigeons as Patagioenas

Johnson and Clayton (2000) and Johnson et al. (2001) have shown that Old World species of the pigeon genus Columba are more closely related to (are sister species of) the large genus Streptopelia than to New World species currently considered as Columba. They recommend that the New World species all be placed into the genus Patagioenas Reichenbach, 1853.

The recommendation by Johnson et al. (2001) is based on study of over 3600 base pairs of nuclear and mitochondrial gene sequences. Among the differences between Streptopelia-OW Columba and NW "Columba" is a 166 bp deletion in the OW groups. This study builds on an earlier, broader, study in the Columbidae (Johnson and Clayton 2000) that had the same results relative to Columba, i.e., that the NW and OW taxa were not monophyletic.

This is not entirely a surprise. Reichenbach divided New World (North American) pigeons into at least three genera (on the same page), distinct from Columba, which were recognized by Ridgway (1916) on morphological grounds. Crumley and Irwin (1944) showed serological differences between NW and OW "Columba", and Johnston (1962) considered morphological and behavioral differences between them.

Despite suggestions of generic differences, most authors continued to treat NA pigeons in Columba, because it was difficult to draw lines. Others used all three generic names for NA birds. Baird (in Baird, Cassin and Lawrence 1858) and Johnston (1962) left fasciata and relatives in Columba but recognized another subgenus or genus for other NA species treated. Oberholser (1974) put band-tails and red-bills in a Reichenbach genus, but did not discuss other NA species. I cannot find another use of Patagioenas for all NA birds as distinct from Columba.

I am accepting Baird (1858) as first revisor, choosing Patagioenas as the name for all NA columbiform (ss) pigeons even though he did not include Band-tails.

I recommend accepting the suggestion of Johnson et al. (2001) that all North American pigeons currently in Columba be transferred to Patagioenas. It is possible that Patagioenas will be split further in the future, and some South American pigeons may be distinct enough for generic status.

Literature Cited:

BAIRD, S., J. CASSIN, and G. N. LAWRENCE. 1858. Reports of explorations and surveys, to ascertain the most practicable and economical route for a railroad from the Mississippi River to the Pacific Ocean. Vol. IX. Part II. Birds. A. O. P. Nicholson, Washington, D.C.

JOHNSON, K. P., S. DE CORT, K. DINWOODEY, A. C. MATEMAN, C. TEN CATE, C. M. LESSELLS, and D. H. CLAYTON. 2001. A molecular phylogeny of the dove genera Streptopelia and Columba. Auk 118:874-887.

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

JOHNSTON, R. F. 1962. The taxonomy of pigeons. Condor 64:69-74.

REICHENBACH, L. 1851 (1852). Handbuch der speciellen Ornithologie. Expedition der Vollstndigsten Naturgeschichte, Dresden.

Oberholser. 1974. Texas Birdlife.

Crumley and Irwin. 1944,. Amer. Nat. 78.

Ridgway. 1916.

Richard C. Banks, July 2003

______________________________________________________________________________________

Comments from Jaramillo: "YES The data is clear on this one. The DNA data, morphology, and behaviour all point in the same direction, I think that the polyphyly of Columba is a solid and correct conclusion. We used Patagioenas in the Birds of Chile book, so I am in definite agreement that dividing Columba is a good idea. The issue of further division of Patagioenas is something that we need not consider at this point, as all potential new genera are all subsets of Patagioenas. In other words, Patagioenas appears to be a monophyletic group. When more data is available on relationships within Patagioenas, then we can consider that aspect."

Comments from Schulenberg: " My vote: YES. The genetic data seem pretty clear. I do note Johnson et al.'s caveat (2001: 884) that "Further analysis is needed to determine if all New and Old World 'Columba' species cluster by biogeographic distribution, but we suspect that will be the case". Those would be my suspicions as well, I guess."

Comments from Robbins: "I vote "yes" for transferring New World pigeons currently in Columba to the genus Patagioenas. "

Comments from Stiles: "Patagioenas: YES, the data seem solid and convincing. In general, I find the use of genetic data more convincing above the species level (depending upon the kind of data, which and how many genes, etc.) than at or below it - especially, in the latter case, when other data are lacking or unconvincing. This doesn’t mean that I reject genetic data, but I do reject splits at the species level based solely upon such data. If one subscribes to some variant of the biological species concept, as I do, then genetic data don't cut it by themselves. Mutations are random events that over long enough time periods approach constant rates, but over the short periods over which much speciation may occur, this constancy breaks down - there is no way to predict when and where lightning will strike next. Especially as there is no set rule to how much genetic change is required for speciation - isolating mechanisms can develop quickly with rather little change in the whole genome under conditions of, say, strong sexual selection but only very slowly in other contexts, where gradual accumulation of small changes might have to occur. Given that the genes people sequence probably have little or no direct role in producing isolating mechanisms as such (and given our general ignorance of how genes produce many particular aspects of the phenotype), no neat correspondence between degree of genetic differentiation and status as species or subspecies is to be expected. The best that genetic data can gives us is an estimate of when two lineages split (within the limits set by saturation, etc.) and this information is most useful above the species level, in defining limits for genera, families etc. It can tell us nothing direct about whether the particular changes that might produce isolating mechanisms have occurred, or are likely to."

Comments from Silva: "YES. The data presented are strong enough to support this change."

Comments from Nores: "[YES.] Si estoy de acuerdo. Resultan muy convincentes las razones dadas por Johnson et al , sobre todo en el hecho que genéticamente Columba del Viejo Mundo están mas relacionadas con Streptopelia que con Columba del Nuevo Mundo."