Proposal (337) to South American Classification Committee
Change linear sequence of species in the genus Trogon
This proposal is to change the current linear sequence of the currently recognized species in the genus Trogon to reflect recent phylogenetic data.
Background: Currently, the SACC sequence is:
Although this sequence groups clusters of species traditionally regarded as closely related because of plumage similarities, I have no idea if/what rationale guided the overall sequence.
New data: Moyle (2005) analyzed DNA sequences from 2 genes, mitochondrial ND2 and nuclear RAG-1, for all genera in the family, including single samples for 14 species in Trogon. DaCosta & Klicka (2008) produced a phylogenetic hypothesis for all 17 traditional species in the genus plus numerous additional population/subspecies samples (n = 160 individuals) using 1041 bp of the mitochondrial gene ND2. The two analyses produced highly congruent results (but expected to a degree giving overlap in genes sampled). The resolution of the branching pattern was very good using the usual measures of node support, and I think we can have reasonably high confidence in the branching pattern for construction of a new linear sequence.
Both analyses indicate a major break in the genus in that the "first 4" species (viridis through violaceus) form a sister group to the "last 3" (melanurus, comptus, massena), and that together they form a sister group to the "middle three" (collaris through rufus). Therefore, whatever sequence we chose will "disturb" the overall structure of the current sequence. I recommend that we start with the last 3, then follow with the first 5, and then finish with the middle group (thus minimizing change by bringing the last 3 to the front).
Relationships among the "last 3" differ between the studies of Moyle (more loci) and DaCosta and Klicka (more samples). Nodal support in both studies contained weaknesses for the placement of comptus, and this should perhaps be treated as a polytomy, especially. Staying as close to the traditional sequence as possible, the order should be massena-comptus-melanurus.
For the "first 4," viridis is sister to the other 3, and violaceus is sister to curucui and surrucura, and the latter should be listed last (SE-most taxon).
The sequence of the "middle group" should begin with rufus, because it is "basal" to the collaris + personatus.
Translating all that to a linear sequence, using the usual conventions ("basal" taxa first; for sister taxa, NW-most taxon listed first; for polytomies, stay as close to traditional sequence as possible), the result is:
Some severe problems with species limits in melanurus and violaceus recognized in Bob's and Steve's field guides and confirmed by DaCosta & Klicka may create sequence problems when the current species limits are revised, but until those situations are addressed by separate proposals, there isn't anything we can do about that yet. Therefore, I recommend a YES on this one
DaCOSTA, J. M., AND J. KLICKA. 2008. The Great American Interchange in birds: a phylogenetic perspective with the genus Trogon. Molecular Ecology 17: 1328-1343.
MOYLE, R. G. 2005. Phylogeny and biogeographical history of the Trogoniformes, a pantropical bird order. Biological Journal Linnean Society 84: 725-738.
Van Remsen and Jeff
DaCosta, March 2008
Comments from Zimmer: "YES. The proposed sequence makes sense, given both the genetic data and the standard conventions that we usually follow. As noted, this could get messy with anticipated revisions in species-limits."
Comments from Stotz: "YES. I am a little concerned that we might have to revisit the sequence of Trogon sooner rather than later, but the basic structure outlined here seems to be a significant improvement."
Comments from Stiles: "YES. This is clearly the best sequence on current evidence. We can deal with modifications later, when species limits have been properly addressed."
Comments from Nores: "YES. Estoy de acuerdo con esta nueva secuencia, ya que la anterior tenía poco sustento. Lo que sí, llama la atención que el color de la parte ventral del cuerpo y de la cola tenga tan poca importancia. Aunque hay un grupo de vientre rojo y cola negra (massena, comptus y melanurus), hay otras especies de vientre rojo y cola blanca o blanca y negra que no están relacionadas. Algo similar pasa con los de vientre amarillo."
Comments from Jaramillo: "YES - Until species limit proposals are written and we may have to revisit."