Proposal (283) to South American Classification Committee
Change linear sequence in Hemispingus
Effect on SACC: This would change the species sequence slightly within Hemispingus.
Background & New information: Our current linear sequence is as follows:
No explicit rationale, to my knowledge, has been published to support this sequence, although it is clearly based on qualitative similarities in plumage and size.
Garca-Moreno et al. (2001) published a phylogenetic hypothesis based on mtDNA sequences of 9 of the 12 species above, and also analyzed 20 morphological characters (plumage patterns and leg and bill size and shape). The gene sampling was poor (only 310 bp of ND2), and some of the morphological "characters" represent arbitrary breaks on continuous variation (e.g., in bill curvature, wing-tail ratio, wing-tarsus ratio). Their summary hypothesis that combined these datasets was conservative with respect to interpretation of the patterns and largely conformed to existing understanding of relationships. The only real difference from the current sequence is that superciliaris clustered with verticalis + xanthophthalmus, although that node did not have adequate bootstrap support. Although they had no tissue, they assumed that reyi was the sister species of superciliaris (as is also implied in the traditional linear sequence). Their summary hypothesis (Fig. 4) places these latter two as sisters and together as sister to verticalis +xanthophthalmus in their "warbler-like" group.
Garca-Moreno and Fjelds (2003) added samples of H. parodii and Cnemoscopus rubrirostris to the same genetic dataset above and performed additional analyses. They confirmed that parodii and calophrys are sister taxa. Also, all their new analyses placed rufosuperciliaris as basal in the tree, which is also consistent with the divergent plumage and morphology of this species (goeringi not sampled but considered the likely sister to rufosuperciliaris). Otherwise the branching pattern of their preferred hypothesis remained largely unchanged except that they moved trifasciatus next to frontalis + melanotis. Their genetic data, however, do not support this adequately -- the critical nodes do not have acceptable levels of bootstrap support. Therefore, their preferred tree relies on a single plumage character, ochraceous tones, to place trifasciatus as basal to piurae + frontalis + melanotis.
Analysis and Recommendation:
Although easy to find fault with the small genetic data set, I recommend following it to some extent because it is an explicit, testable hypothesis based on a quantitative analysis (and it also not counterintuitive). Thus, it is more scientific than any previous classification, for which no supporting rationale is available to my knowledge. However, concerning the relationship of trifasciatus (formerly placed in a monotypic genus) to frontalis + melanotis, given that only ca. 300 bps went into the analysis, and that no well-supported nodes support their novel arrangement, I recommend not incorporating that result into the classification until corroborated by additional analyses. Otherwise, to change our sequence to fit their hypothesis, we have to move superciliaris and reyi as follows, and rufosuperciliaris and goeringi to a basal position:
GARCêA-MORENO, J., AND J. FJELDS. 2003. Phylogenetic relationships among Hemispingus tanagers. Ornitologia Neotropical 14: 363-370.
GARCêA-MORENO, J., J. OHLSON, AND J. FJELDS. 2001. MtDNA sequences support monophyly of Hemispingus tanagers. Molecular Phylogenetics & Evolution 21: 424-435.
Van Remsen, June 2007
Comments from Stiles: "YES. The change is minor and accords with the evidence available (which ain't much, but is better than what precedes it, which is essentially zero."
Comments from Cadena: "NO. I agree with Van in that this new arrangement is perhaps more scientific than the traditional one in that it is based on a formal analysis of some sort of data and not purely on speculation. However, the data and analyses used to support this change are weak, so I see no reason to change. From proposal 284 we know there is a forthcoming publication by Kevin Burns with much better data, so why not wait? Even for somewhat trivial matters like linear sequences within a genus, I'd rather wait a few months than make a change that will later have to be reversed."
Comments from Robbins: "NO. After reading Daniel's comments, I vote "no", with the anticipation that we will soon see a better data set from Ken Burns."
Comments from Zimmer: "NO. Given the promise of an impending publication based on more rigorous analysis, I'd also prefer to wait."
Comments from Nores: "NO. Pienso que si hay un trabajo de Burns prximo a aparecer con ms datos, sera apresurado aceptar ahora cambios en la secuencia. Tambin habra que ver si H. auricularis es una especie vlida como sealado por Garca-Moreno et al. (2001)."
Comments from Stotz: "NO. This is probably a correct change, but I think we should await the Burns information before making a change of this sort."
Comments from Jaramillo: "YES - This is a better arrangement, based on actual data! The Burns paper will be welcome and certainly there can be shifts made when that comes out. But a bird in the hand is worth two in the bush, until that is published and you know how that can be ... we have the present available data and linear sequence suggested and I think it is ok to go with it."
Comments from Pacheco: "Tal qual os demais, comeando pelo Daniel, eu tambm considero oportuno aguardar a anlise mais ampla de Kevin Burns."