Proposal (591) to South American Classification Committee

Revise the classification of the Pipridae

Effect on SACC: If adopted, this would (A) recognize the newly named genus Cryptopipo for Xenopipo holochlora, (B) revise the linear sequence of genera in the family, and (C) recognize two subfamilies within the family.

Background: Our current classification of the Pipridae reflects substantial new increments of data from recent research. Our note summarizes the situation as follows:

1. Sequence of genera and composition of the family follow Prum (1990a, 1992). Genetic data (Tello et al. 2009, McKay et al. 2010) confirm the monophyly of the Pipridae as constituted above, but also provide evidence that the existing linear sequence of genera conflicts in several ways with phylogenetic data. SACC proposal passed to change linear sequence of genera. Ohlson et al. (2013) proposed recognizing two subfamilies, Neopelminae and Piprinae, and additional tribes within Piprinae. SACC proposal needed; their results also indicate that modifications to the linear sequence of genera are also needed. SACC proposal needed.

New information: Ohlson et al. (2013) investigated relationships within the family using DNA sequence data from three nuclear introns and one mitochondrial gene (ND2). They sampled all genera and most species. I have pasted in a screen grab of their tree below:

Their results are largely consistent with those of previous studies except for the polyphyly of Chloropipo, members of which are in three parts of the tree. Because Ohlson et al. was the first study to include all five species in broadly defined Xenopipo, these results do not conflict with previous studies. McKay et al. (2010) included only unicolor, and Tello et al. (2009) included only atronitens and uniformis, which they found to be sisters, as in Ohlson et al.

Support for almost all nodes is very strong. Their results add further support to changes previously adopted by SACC in terms of breaking up Pipra into several genera.

To reconcile classification with their phylogeny, they recommended the following changes at the genus level: (1) Xenopipo reduced to type species atronitens and its sister species X. uniformis; (2) Chloropipo is resurrected for the two Andean species, flavicapilla and unicolor; and (3) a new genus is described for holochlora, which is sister to Lepidothrix and thus not closely related to Xenopipo or Chloropipo.

This proposal is in three parts:

A. Recognize Cryptopipo. I recommend a YES on this. The genetic data show clearly that holochlora is more closely related to Lepidothrix than to Xenopipo or Chloropipo, and to include it within Lepidothrix is the only other option given the data. That option was considered and rejected by Ohlson et al. because: “it differs in so many aspects of morphology and behavior that we are reluctant to include it in that genus”. I agree with that statement. A NO vote would thus favor either inclusion of holochlora in Lepidothrix or retention of broadly defined Xenopipo.

B1. Revise linear sequence of genera.

Our current sequence is:

Neopelma

Tyranneutes

Ilicura

Masius

Corapipo

Antilophia

Chiroxiphia

Xenopipo

Machaeropterus

Dixiphia

Ceratopipra

Manacus

Heterocercus

Pipra

Lepidothrix

To alter the sequence to conform to the Ohlson et al. tree, i.e., the addition of Chloropipo and Cryptopipo, the following changes are needed in red:

Neopelma

Tyranneutes

Chloropipo

Ilicura

Masius

Corapipo

Antilophia

Chiroxiphia

Xenopipo

Machaeropterus

Dixiphia

Ceratopipra

Manacus

Heterocercus

Pipra

Cryptopipo

Lepidothrix

B2. Further revise linear sequence of genera.

However, some additional changes are advisable if we accept the Ohlson et al. tree as the best available data and use on of the conventions for sequencing genera, i.e., for sister taxa, least-diverse group first. For example, within the Piprini in the figure, support is strong for Cryptopipo + Lepidothrix as sister to the more diverse lineage (Heterocercus + Manacus + Pipra + Machaeropterus + Dixiphia + Ceratopipra). Within the latter group, support varies from so-so to strong for the following relationship: Heterocercus (Manacus + ((Pipra + (((Machaeropterus + ((((Dixiphia + Ceratopipra)))))))))). I have mixed feelings on whether an overhaul of the sequence is warranted. On the other hand, these are the best available data, and our previous rearrangement of the sequence hardly leaves our current one with much of an historical legacy. Therefore, I lean towards going all the way on the rearrangement at this point – as long as we’re going to make some changes, might as well make all those indicated by the data, i.e.:

Tyranneutes

Neopelma

Chloropipo

Antilophia

Chiroxiphia

Ilicura

Masius

Corapipo

Xenopipo

Cryptopipo

Lepidothrix

Heterocercus

Manacus

Pipra

Machaeropterus

Dixiphia

Ceratopipra

This is exactly the sequence recommended by Ohlson et al. (2013) except for the flip-flop of the groups (Antilophia + Chiroxiphia) and (Ilicura + ((Masius + Corapipo))).

C. Add subfamilies. Ohlson et al. (2013) found the same pattern as Tello et al. (2009) and McKay et al. (2010), namely a deep division within the family, with Tyranneutes and Neopelma forming one lineage, and the rest of the genera in the other. I support recognition of this major division with subfamily rank for the two lineages.

Literature Cited:

McKAY, B. D., F. K. BARKER, H. L. MAYS JR., S. M. DOUCET, AND G. E. HILL. 2010. A molecular phylogenetic hypothesis for the manakins (Aves: Pipridae). Molecular Phylogenetics and Evolution 55: 733-737.

OHLSON, J., J. FJELDSÅ, AND P. G. P ERICSON. 2013. Molecular phylogeny of the manakins (Aves: Passeriformes: Pipridae), with a new classification and the description of a new genus. Molecular Phylogenetics and Evolution 69: 796–804.

TELLO, J. G., MOYLE, R. G., D. J. MARCHESE, AND J. CRACRAFT. 2009. Phylogeny and phylogenetic classification of the tyrant flycatchers, cotingas, manakins, and their allies (Aves: Tyrannides). Cladistics 25: 1-39.

Van Remsen, October 2013

__________________________________________________________________________________________________________________

Comments from Stiles: “YES. The Ohlson et al. paper provides the most thorough and comprehensive genetic data set so far, and makes the description of Cryptopipra for holochlora necessary, and recognition of two subfamilies desirable. Their data do make me wonder if Corapipo and Masius ought to be merged? It looks like Masius makes Corapipo paraphyletic?”

Comments solicited from Jan Ohlson: “"I thank Van Remsen for so quickly bringing our proposals for changes of Pipridae classification to the SACC board. As might be expected, I fully endorse the proposals put forward by Van, and I am also strongly in favor of the alternative B2 regarding sequence of genera, as this best reflects the phylogenetic tree given the conventions accepted by the SACC".

Comments from Pacheco: “YES. Eu concordo com as recomendações – incluindo as adaptações propostas por Remsen – derivadas de dois amplos estudos.”

Comments from Zimmer: “YES on parts A, B and C, based largely on the data of Ohlson et al. (2013).”

Comments from Pérez-Emán: “YES to A and B2. The new phylogenetic study on Pipridae by Ohlson et al. (2013) found a paraphyletic Chloropipo (or Xenopipo) clearly requiring a new generic name for C. holochlora (Cryptopipo). This study also provides the basis for updating linear sequence in the family, considering the most comprehensive phylogenetic hypothesis currently known. I would vote NO for C, as I consider this taxonomic level should consider not only molecular data but also other characters that unambiguously show support for these taxonomic categories.”

Comments from Cadena: “591A. YES. Regarding Gary's comment on eventually lumping Corapipo and Masius, we should note that the relevant node is rather poorly supported (only 0.74 posterior probability and 48% bootstrap). This lack of resolution and the plumage differences between genera suggests we should be conservative for now.

“591C. NO. Not because I think the phylogenetic split is not strong, but rather based on a more philosophical point of view. As I have commented on other proposals before, I think we need to be consistent in our classification across the board. If we are going to recognize subfamilies (i.e. well supported clades within families), then we need to do this across all families, not haphazardly in those families for which someone happens to present a proposal. Do we want to do this for all families? This would be a lot of work, and I don't think we should. If we do, then where do we stop once we are done? If we recognize subfamilies, then why not tribes, suborders, etc..? I say we should stick to major taxonomic ranks; sure, they are arbitrary, but they are arguably more manageable in terms of number and in the degree of consistency that our taxonomy already has (e.g., we recognize families, but not subfamilies, in every order).”