Proposal (86) to South American Classification Committee
Treat the taxon melanolaemus as a subspecies of Atlapetes rufinucha
Effect on South American CL: This proposal would lump two taxa that we treat as separate species into a single, more traditional Atlapetes rufinucha.
Background: The taxon melanolaemus has traditionally been treated as a subspecies of Atlapetes rufinucha; they are evidently parapatric, with the barrier, if it exists, probably being the dry upper Madidi valley in northern depto. La Paz, Bolivia. Melanolaemus looks like nominate rufinucha except that its breast is markedly clouded to varying degrees with irregular blackish markings; also, its head is more extensively black, connecting the black face to the black malar streak and lacking the pale marks in the loral area of nominate rufinucha.
New information: Garca-Moreno & Fjelds (1999) found that the "species" Atlapetes rufinucha was polyphyletic, with the northern form, A. latinuchus, more closely related to nominate A. schistaceus than to A. rufinucha. This was based on a very small (275 bp) portion of the cytochrome-b gene, and the bootstrap values are unimpressive (and I doubt that these results would be publishable in 2003). Nevertheless, their resulting tree shows sufficient geographic structure, with adjacent Atlapetes usually appearing as sisters, so I suspect that many of the results will hold up with larger data sets. Of particular relevance to this proposal is that the 2nd-highest bootstrap value in their tree (72%) was between nominate rufinucha and Atlapetes fulviceps, the parapatric species that borders the southern distribution of A. rufinucha; also, melanolaemus clusters with a set of parapatric taxa from southern Peru, A. melanopis, A. canigenis, and A. forbesi. The low bootstrap values for this "southern Peru" group mean that it should not yet be taken seriously, despite its geographic "sense"; regardless, if the rufinucha-fulviceps node is solid, then that would make traditional rufinucha a paraphyletic species without the removal of melanolaemus. For that reason, and the lack of any signs of intergradation despite their presumed parapatry, beyond a single suspicious specimen, Garca-Moreno & Fjelds (1999) elevated melanolaemus to species rank, and this was followed by Dickinson (2003). No comparative data on vocalizations have been published.
Analysis: The data for species rank of melanolaemus are fairly weak. The genetic data are intriguing but weak by current standards. The contact zone, if there is one, has not been studied, and that single specimen (at LSUMZ) may indicate some gene flow between them, although it is ca. 100 km south of the potential contact zone, and no other specimens from the area show any phenotypic sign of intermediacy. [As an aside, I personally do not object strongly to paraphyletic species ... I know this is heresy among strict phylogeneticists, but in my opinion, phylogenetic analyses are not intended for population-level processes in which problems with lineage-sorting and gene trees are inevitable; Willi Hennig himself explicitly noted that cladistic analysis was inappropriate at the population level for this reason; it seems likely to me that historical gene flow among adjacent populations requires that he label "monophyletic" at the population level refers only to the genes analyzed and to the current time period.]
An argument in favor of retaining species rank for melanolaemus could be constructed as follows. The genetic data, weak as they are, are nonetheless consistent with species rank. The geography of phenotype distribution is also consistent with species rank; these two taxa may be parapatric with no sign of substantial gene flow, or if separated at all, it is by a narrow dry valley. Melanolaemus, by virtue of its more extensively black face, shows a fundamental difference in head pattern from nominate rufinucha that might affect mate selection; these pattern differences are actually greater than those between the head patterns of nominate rufinucha and the latinuchus species group further north. Parapatric Atlapetes canigenis also shows this "expansion" of black in the facial area, suspiciously similar to that of melanolaemus. As emphasized by Garca-Moreno & Fjelds (1999), minor differences in Atlapetes brush-finches seem to "matter" in that there are no known cases of zones of intergradation or hybridization among parapatric taxa. Even In concert, these arguments are fairly weak, but perhaps stronger than the arguments for the traditional classification.
Recommendation: I tentatively vote NO on this one (i.e., stick with current classification). Although the evidence is weak for species rank of melanolaemus, I think that the "burden of proof" in this falls on the case for considering it conspecific with Atlapetes rufinucha.
DICKINSON, E. C. (ed.). 2003. The Howard and Moore complete checklist of the birds of the World, Revised and enlarged 3rd Edition. Christopher Helm, London, 1040 pp.
GARCêA-MORENO, J., AND J. FJELDS. 1999. Re-evaluation of species limits in the genus Atlapetes based on mtDNA sequence data. Ibis 141: 199-207.
Van Remsen, December 2003
P.S.: If this proposal doesn't pass, then I'll do a proposal on the English name of melanolaemus, for which there are two competing choices.
Comments from Stiles: "[NO]. I agree that the evidence is not overwhelming, but it is a shade better than the evidence for the contrary - and solid contrary evidence should be forthcoming to change the "status quo". NO to both (maintain species status, at least for now)."
Comments from Zimmer: "NO, although the evidence for separate species status is somewhat weak."
Comments from Robbins: "[NO] Although the evidence certainly isn't compelling for treating melanolaemus as a species, the data for considering it as a subspecies is even less. Hence, I vote for following the "status quo". A "no" vote.
Comments from Stotz: "YES, for lumping all of these into rufinucha. I have to say that I hardly consider the splitting of these species as the "status quo." To me the status quo is the broad rufinucha, which we had until 1999. Van is correct to point to the short piece of DNA used in the Garcia-Moreno and Fjelds study and to note the weak support for the taxa that they suggest. There are only two branches with over 50% bootstrap support. They support a northern clade, and a sister relationship between rufinucha and fulviceps. I am willing to overlook this because of the shortness of the segment of DNA that was studied. My personal feeling is that we would be better off with the original 4 species (schistaceus, rufinucha, rufigenis and tricolor) with the recognition that there are problems that need to be solved, but as that is not currently on the table, I don't think placing terborghi, melanolaemus, and rufinucha into a single species conflicts with any of the results of the 1999 in a significant way. Terborghi and melanolaemus occupy basically adjacent areas to canigenis, but only if you believe the poorly supported results and believe that species have to be monophyletic is that a problem. Finally, I have to say that it seems strange to me that we completely follow the novel arrangement suggested based on very weak data for these brush-finches, while Poospiza whitii and Hyloctistes virgata are not split."
Comments from Schulenberg: "YES. I think that someday we should vote on the split of Atlapetes rufinucha rather than take the new arrangement for granted. The split may fit with our world view, but in concrete terms has poor support."
Comments from Jaramillo: "NO. Somewhat reluctantly, because of the poor data involved. However, there is a paper that exists and proposes a phylogenetic organization that makes sense in terms of biogeography, and lack of intermediate specimens, etc. Although poor, this piece of work does shift the burden of proof in my mind. Until someone performs a stronger analysis that refutes this, we should keep this taxon as a separate species."
Comments from Nores: "NO, yo voto en contra de considerar a Atlapetes melanolaemus como una subespecie de A. rufinucha. Pienso que las caractersticas morfolgicas, genticas y su distribucin paraptrica sin intergradacin tienen suficiente peso como para asignar nivel de especie a melanolaemus. No obstante, el razonamiento de Stotz es tambin convincente."