Proposal (172) to South American Classification Committee
Change linear sequence of species in Geositta
Effect on SACC: This proposal would change the linear sequence within Geositta (Furnariidae) to conform to recent phylogenetic results.
Background: Our current sequence is the one used by Remsen (2003), which differed from traditional sequences (e.g., Meyer de Schauensee 1970) only in placing G. tenuirostris close to G. cunicularia based on plumage similarities, and by including poeciloptera in the genus, but at the beginning of the sequence to reflect its former placement in the monotypic genus Geobates:
New data: Cheviron et al. (2005) analyzed about 850 bp of two mitochondrial genes to produce a phylogenetic hypothesis for the genus. Their nodes with strong bootstrap support (> 75%) were as follows:
1. G. cunicularia and G. tenuirostris are sister species, and G. peruviana is the sister to these two.
2. G. antarctica, G. isabellina, G. saxicolina, and G. maritima form a strongly supported group.
3. The above four species along with G. poeciloptera, G. crassirostris, G. punensis, and G. rufipennis also together form a strongly supported group.
4. G. punensis and G. rufipennis are sisters (borderline bootstrap support, 75%).
These findings conflict with the sequence in Remsen (2003) in several ways, most notably in placement of G. cunicularia and G. antarctica in different groups, contrary to traditional view that plumage similarities suggest that they are closely related if not sister species (e.g., Vaurie 1980).
To incorporate the above findings into our current sequence with minimum disturbance, the sequence is as follows (with moves highlighted in red)
Geositta peruviana *
Geositta poeciloptera **
* moved forward to indicate basal position within first group; this also highlights that it is not the sister to the superficially similar G. maritima.
** not placed at the end to emphasize that it is indeed embedded within one of the groups of Geositta.
Cheviron et al. (2005) also proposed a linear sequence in their appendix 2 (done by F. Vuilleumier, fide Z. Cheviron). It differs from the above one in:
(a) placing peruviana after cunicularia + tenuirostris. This, however, does not follow the convention of placing basal taxa first.
(b) punensis + rufipennis are placed last. The difference is an arbitrary branch movement within a 3-way polytomy, so this one is entirely subjective. I like mine better because (1) it places last the two oddballs, for which there is no support for a sister relationship to any other species; (2) it is closer to the traditional sequence, which has crassirostris last, thus creating less havoc.
Recommendation: As with any such data-set, additional genetic data might produce a slightly different structure (as well as resolve the polytomies within the two major groups). Nonetheless, these new data, which include all species in the genus, are certainly superior to the qualitative assessments of relationships based on morphological similarities that guided the traditional linear sequence. I see no reason not to modify our current sequence to reflect these new data and recommend a YES on this one.
Van Remsen, April 2005 (with input from Zac Cheviron)
Comments from Robbins: "[YES]. Arguments for the new arrangement based on the Cheviron et al. data seem logical. I vote "yes."
Comments from Stiles: "YES, at least tentatively. 850 bp of two mitochondrial genes is not a huge sample, but the results are probably an improvement on previous arrangements."
Comments from Jaramillo: "YES - The genetic analysis is certainly the best dataset we have thus far. It makes sense in various ways: antarctica never seemed all that similar to cunicularia to me. Vocalizations will be interesting to analyze once we have a good sample, although these species are seldom recorded unfortunately. I have good evidence for more species-level taxa, but unpublished at the moment. I hope that this next trip to Chile will get me across the threshold for at least one of these publications; stay tuned."
Comments from Pacheco: "[Yes] É forćoso reconhecer que os dados provenientes desta análise genética sčo, no momento, os melhores disponíveis e devem ser considerados."
Comments from Nores: "YES. En general, salvo la posición de Geositta antarctica. A pesar de que los estudios mitocondriales de Cheviron et al. la agrupan con G. isabellina, G. saxicolina, and G. maritima, desde el punto de vista morfológico y sobre todo biogeográfico, es mucho más lógico la agrupación con G. cunicularia. No parece factible que una especie de llanura del sur del continente esté relacionada filogenéticamente con otras muy lejanas geográficamente; dos de alta montaĖa: G. isabellina, G. saxicolina y la tercera (G. maritima), restringida al desierto de Perú y norte de Chile. Lo lógico es que derive de G. cunicularia con la cual se superponen en distribución. Tampoco me gusta mucho separar G. rufipennis de G. isabellina que son las más parecidas entre si, pero esto puede ser sólo apariencia."