Proposal (741) to South American Classification Committee

Split Zimmerius vilissimus into two or three species

Effect on SACC: Elevates Zimmerius improbus and Zimmerius petersi to species rank.

Background: Paltry Tyrannulet (Zimmerius vilissimus) is a common and familiar bird of Central America and northwestern South America. It occurs in several disjunct populations that correspond to recognized subspecies as well as some populations that may additional taxonomic recognition. There are five named taxa that NACC and SACC currently include within Zimmerius vilissimus:

· Zimmerius v. vilissimus - Paltry Tyrannulet, southeastern Mexico, southern Belize, much of Guatemala, and possibly northwestern Honduras

· Zimmerius v. parvus - Mistletoe Tyrannulet, most of northern Honduras, eastern Nicaragua, Costa Rica, Panama, and a small part of Colombia bordering Panama.

· Zimmerius z. petersi - Venezuelan Tyrannulet, coastal mountains of north-central Venezuela

· Zimmerius z. improbus - Mountain Tyrannulet or Specious Tyrannulet, Eastern Cordillera of Colombia and Mérida Cordillera of Venezuela

· Zimmerius z. tamae - Mountain Tyrannulet or Specious Tyrannulet, Tamá Paramo of Venezuela, Santa Marta Range of Colombia, Serranía de Perijá of Colombia and Venezuela

This map is based on the range map from the Neotropical Birds Online default account (http://neotropical.birds.cornell.edu/portal/species/overview?p_p_spp=436521 - I've just submitted an updated account for this species), and has a few errors: the Z. i. vilissimus population in southern Guatemala extends into central El Salvador, and Z. i. petersi is separated from Z. i. improbus by an extensive low elevation region near Barquisimetro, Venezuela. The southernmost Z. i. tamae population is restricted to the Tamá Paramo in western Venezuela, with Z. i. improbus populations to both the east and west (though references disagree on the exact range of these two taxa, particularly the Norte de Santander, Colombia, population).

SACC is perhaps the last group maintaining this complex as a single species. Taxonomic arrangements accepted by various references include:

· Howard and Moore - Dickinson & Christidis (2014): Z. vilissimus (Paltry Tyrannulet), Z. improbus (Mountain Tyrannulet), and Z. petersi (Venezuelan Tyrannulet)

· Handbook of Birds of the World - Fitzpatrick et al. (2004): Z. vilissimus (Paltry Tyrannulet), Z. improbus (Venezuelan Tyrannulet)

· IOC World Bird List - Gill and Donsker (2017): Z. vilissimus (Paltry Tyrannulet), Z. parvus (Mistletoe Tyrannulet), Z. improbus (Specious Tyrannulet), and Z. petersi (Venezuelan Tyrannulet)

SACC has previously visited this species, in Proposal 441 (http://www.museum.lsu.edu/~Remsen/SACCprop441.htm). Proposal 441 was based on a Donegan et al. (2010) qualitative review of calls from xeno-canto as well as references to plumage differences. Without a rigorous review of the calls, the committee did not accept the proposal. Several voting members preferred to hold off until genetic data were available.

New information: Rheindt et al. (2013) undertook a molecular analysis of the entire genus Zimmerius, with denser taxon-sampling within the Z. vilissimus complex. Their samples came from museum vouchers, and they obtained sequence from both mitochondrial DNA (ND2 plus some adjacent tRNA-Met) and nuclear DNA (FiB5). Their results are robust: they recovered the same topology among named taxa using maximum parsimony, maximum likelihood, and Bayesian approaches, and when analyzing each gene separately as well as with a concatenated dataset.

Their Figure 3, shown below, is based on the concatenated dataset. They have highlighted in red the taxa currently included in Zimmerius vilissimus. There are at least three non-sister taxa within Zimmerius vilissimus. Only one Z. v. petersi sample was available, but it is a distinct lineage that is sister to Z. gracilipes. Secondly, Z. v. improbus is a separate clade that is sister to five other South American taxa (with strong support for improbus monophyly, but without significant support for this arrangement). Finally, Z. v. parvus and Z. v. vilissimus are part of an unresolved polytomy together with Z. albigularis.

Figure 3. Bayesian tree topology of the concatenated dataset (outgroup not shown). Nomenclature is based on the taxonomic recommendations of the present study. The green vertical bar marks the former Zimmerius chrysops (sensu Remsen et al., 2012). Taxa formerly subsumed under Zimmerius vilissimus are given in red. Branch support is given in the order: parsimony bootstrap/maximum likelihood bootstrap/Bayesian posterior probabilities. Only significant branch support is given, here defined as > 90 (for Bayesian) or > 85 (for parsimony/likelihood). A bold ‘100’ indicates maximum branch support for all three analytical modes. Where only one nonbold number is given, it refers to Bayesian support and implies that likelihood and parsimony support were not significant.

The structure within the Z. i. improbus / tamae group was unclear; samples instead fell into three groups: (1) Norte de Santander, Colombia & Táchira, Venezuela: 4 samples; (2) Santa Marta Range of Colombia: 2 samples; (3) Serranía de Perijá of Colombia: 2 samples. This group needs additional clarification with molecular, song, and plumage analysis.

The authors suggest that four taxa be recognized: Z. vilissimus, Z. parvus, Z. petersi, and Z. improbus. The Z. vilissimus-parvus split should be considered by the NACC; the remaining splits directly impact South America.

Recommendation: Based on the Rheindt et al. (2013) molecular analysis, I suggest a vote structured as follows:

PART A:

NO: Maintain the status quo, with five named subspecies within Zimmerius vilissimus.

YES: Split Zimmerius vilissimus into two species:

Zimmerius vilissimus (including vilissimus and parvus) - Paltry Tyrannulet

Zimmerius improbus (including tamae and petersi) - Venezuelan Tyrannulet

PART B:

NO: If part A is accepted, then this would maintain two species: Zimmerius vilissimus and Zimmerius improbus.

YES: Split Zimmerius vilissimus into three species:

Zimmerius vilissimus (includes Z. v. parvus) - Paltry Tyrannulet

Zimmerius improbus (includes Z. i. tamae) - Mountain or Specious Tyrannulet

Zimmerius petersi (monotypic) - Venezuelan Tyrannulet

My recommendation is YES for both parts.

· Part A is straightforward: vilissimus and improbus are well-sampled in this phylogeny, and are clearly not sister taxa. Their calls and plumage differ, and they are quite disjunct. Their altitudinal ranges differ (all elevations in Central America; mid-elevations in South America).

· Part B recognizes three non-sister taxa as full species. Rheindt et al. (2013) analyzed the petersi-improbus split; forcing those two taxa to be sister taxa was rejected by a Shimodaira-Hasegawa test. There are strong plumage differences between improbus and petersi, and Rheindt et al. mentions that there are call differences, though details have not been published. The English name for Z. improbus is unclear: both Mountain and Specious Tyrannulet are already in use.

References:

Donegan, T., Salaman, P., Caro, D. & McMullan, M. 2010. Revision of the status of bird species occurring in Colombia 2010. Conservación Colombiana 13: 25-54.

Rheindt, F. E., A. M Cuervo, and R. T. Brumfield. 2013. Rampant polyphyly indicates cryptic diversity in a clade of Neotropical flycatchers (Aves: Tyrannidae). Biological Journal of the Linnean Society 108:889-900.

Andrew W. Jones, February 2017

Note from Remsen: If this proposal passes, then a separate proposal needed for English names. SACC policy on English names is to avoid using the "parent" species English name, i.e. Paltry Tyrannulet, for one of the daughter species if avoidable.

___________________________________________________________

Comments from Stotz: “A YES. B YES. I would rather have more explicit description of the vocal differences here, but the fact that these taxa are all over the tree indicates to me that they should be fully split. There are a couple of English name issues. If we split up vilissimus, normally we would not use Paltry Tyrannulet for vilissimus, but would create a new name. I don’t have a good alternative, but would be open to any ideas anybody else has (maybe Tom Schulenberg has an idea). For improbus, I would say Mountain. I have no idea what the theory behind Specious Tyrannulet as an English name is.”

Comments from Stiles: “YES to A and B. The three S.A. taxa are nowhere near being sisters and differ considerably in plumage. Refer the vilissimus-parvus split to NACC.”

Comments from Zimmer: “YES to A and B”. It’s pretty clear from looking at the tree that the three South American taxa are not sisters, and the three-way split makes more sense than any other option. My memory of petersi is that it was vocally and morphologically a pretty different beast from the Central American populations that I was much more familiar with. I can’t speak to differences between improbus/tamae and the others, but the tree tells me enough. I would argue that the four-species split favored by Rheindt et al (2013) is the even better path, but the split of parvus and vilissimus is out of our jurisdiction. As for English names: I think “Mistletoe Tyrannulet” for parvus and “Venezuelan Tyrannulet” for petersi should be baked in, and I have a definite preference for “Mountain Tyrannulet” over “Specious Tyrannulet” for improbus/tamae. If NACC does, indeed, end up splitting vilissimus and parvus, then I would think that a new name would be desirable for vilissimus. If NACC doesn’t split those two, then I would actually suggest adopting “Mistletoe” as the modifier for the combined vilissimus/parvus.”

Comments from Jaramillo: “YES – to A and B. Having recently returned from the Perijá Mountains after a visit to Guatemala, it is hard to fathom that improbus and vilissimus are considered sympatric. Not only are they quite different in plumage, the size and structure differences are easily visible. So, it is heartening to see the molecular data bear this out.”

Comments from Areta: "YES to A and B. I wish there was more information available for petersi, but how little data is available on genetics and vocalizations is consistent with species level differentiation for Zimmerius petersi."

Comments from Robbins: "YES to both A and B, as the genetic data demonstrate that the three pertinent taxa to our committee aren’t even sister taxa.?"

Comments from Claramunt:

"A YES. Z. improbus seems a well differentiated clade and separated from the trans-Andean clade.

"B YES. Z. petersi falls in a clade with bolivianus and gracilipes only in mtDNA but Bf5 strongly suggests affinities with improbus, which makes more biogeographical sense (note that the proposal is misleading in saying that analyses of the two genes produced the same topology). So, there’s gene-tree incongruence here. But given the levels of gene divergence, the conflict seems an issue lineage sorting in ancestral populations, not ongoing gene flow. In addition, there are plumage differences (although I wouldn’t say they are “strong”) and songs seem also to differ drastically (two notes versus one note, if I’m interpreting well what is available on-line).

"Finally, note that two of the basal nodes in the mtDNA tree are not strongly supported and one of those conflicts with one strongly supported BF5 node. So, be cautious in reading the mtDNA tree. Some nodes may reflect just the stochasticity of gene genealogies rather than species relationships. For example, I would not be surprised if BF5 turns to be right in suggesting that petersi is sister to improbus."

Comments from Pacheco: "YES to A and B. In both cases, the data now available indicate that those involved taxa are not sisters."