Proposal (#235) to South American Classification Committee

Change classification of the Formicariidae

Background: This proposal would revise the family-level classification of the Formicariidae, which as currently defined is not a monophyletic group.

Currently, the Formicariidae includes species in the following genera:


New Information: Several recent studies (Irestedt et al. 2002, Chesser 2004, Rice 2005a, 2005b) have assessed the relationships among these genera and genera of other suboscine families based on mitochondrial (cytochrome b, ND2, ND3, COI) and nuclear (cmyc, RAG-1, myoglobin intron 2, Beta-fibrinogen intron 7) sequence data, employing different taxon-sampling strategies. Based on the conclusions of these studies, which are consistent with those of a forthcoming publication by R. Moyle et al. using nuclear RAG-1 and RAG-2 data (which was partially presented at the AOU meeting in 2004), it is clear that there are three distinct lineages of "Formicariidae":

(1) Formicarius and Chamaeza
(2) Grallaria, Myrmothera, Hylopezus, and Grallaricula
(3) Pittasoma

The affinities of these three groups are not yet well-established, except for the strongly supported sister relationship between Pittasoma and the genus Conopophaga, currently placed in its own monogeneric family (Conopophagidae). Despite the lack of strong support for "deep" relationships, a clade formed by any of the possible combinations of groups (1), (2), and (3) to the exclusion of birds presently included in other families has not been recovered in any analysis conducted so far.

Analysis: Taken together, these data strongly suggest that Formicariidae is not monophyletic, and that it comprises three phylogenetically distinct groups. Clearly, this calls for a change in classification at the family level to be consistent with phylogenetic relationships.

Based on the results of the phylogenetic analyses, Irestedt et al. (2002) proposed to limit membership in the Formicariidae only to Formicarius and Chamaeza. I believe this is a sensible option. Rice (2005b) presented a somewhat different alternative, in which he suggested that the Formicariidae could include not only Formicarius and Chamaeza, but also the tapaculos (Rhinocryptidae), as these two seemed to be sister clades based on some molecular work. This is only a matter of taste, but it seems to me that this would result in a highly heterogeneous family, although I could easily be convinced that Rhynocryptids are already quite heterogeneous to begin with, and that Formicarius and Chamaeza do not really add much variation. More importantly, however, I don't think this is the best alternative because support for the ((Formicarius, Chamaeza), Rhinocryptidae) arrangement is not strong. In fact, ongoing work based on c. 4000 bp of RAG sequence by Moyle et al. shows that Formicarius and Chamaeza are sister to the Furnariidae (i.e. Furnariidae + Dendrocolaptidae) with strong support, not to the Rhinocryptidae.

Irestedt et al. also proposed to erect a new family (Grallariidae) for the clade formed by Grallaria, Myrmothera, Hylopezus, and Grallaricula. I cannot think of a better alternative regarding the placement of these taxa.

In addition, we need to decide on what to do with Pittasoma. There are two alternatives: one is to merge this genus with Conopophaga in a single family (Conopophagidae), as advocated by Rice (2005b) and endorsed by Krabbe and Schulenberg (2003, HBW). The other alternative would be to place Pittasoma in a new monogeneric family, Pittasomidae (?). Considering the morphological uniformity of the species of Conopophaga and how distinct these are from Pittasoma (as evidenced by traditional, morphology-based taxonomy), at first glance it would seem that the option of placing them in different families represents the best possible course. However, Rice (2005a) presented information on several different traits (morphology, natural history, vocalizations) that support the close relationship between Pittasoma and Conopophaga. Whether one wants to recognize one or two families is open for discussion - all the traits discussed by Rice only help strengthen the support for this clade by providing synapomorphies that complement the mitochondrial and nuclear data, but do not resolve the issue of the taxonomic rank that should be given to its two constituent lineages. An additional argument that one could put forward in favor of Rice's proposal is that by placing Pittasoma in Conopophagidae, we avoid introducing new family names, which could be said to be conservative.

To summarize, there are five subproposals here:

(a) new circumscription of Formicariidae: Formicarius and Chamaeza
(b) new circumscription of Formicariidae: Formicarius, Chamaeza, and the tapaculos.
(c) accept family Grallariidae
(d) place Pittasoma in Conopophagidae
(e) erect family Pittasomidae

Recommendation: (a) YES, (b) NO, (c) YES, (d) YES, (e) NO.

Literature Cited: See SACC website.


C. Daniel Cadena (in consultation with Nate Rice and Rob Moyle),


Comments from Robbins: "Cadena summarizes data from various studies and presents (with input from Rice and Moyle) logical arguments for all five subproposals. Thus, I vote as follows:

a) yes
b) no
c) yes
d) yes
e) no

Comments from Stiles: "The question here is, are the published data sufficient to justify the splits proposed? The Moyle et al. data sound good but as they are unpublished, hence unreviewed (?), they should not be taken as evidence at this point. As a general rule, I feel that the data supporting arguments for taxonomic changes should at least be accepted for publication Iis this the situation with this MS?); a manuscript only submitted, or only a first draft, may suffer major changes before acceptance, including incorporation of additional data and possible changes of some or all conclusions.

Aside from the Moyle et al. data, concordant results from two independent studies using several genes plus morphology do appear sufficient to justify splitting Formicariidae into at least two families: Formicariidae sensu stricto (Formicarius, Chamaeza) and Grallariidae (Grallaria, Grallaricula, Hylopezus, Myrmothera). The Pittasoma + Conopophaga clade also seems well established, the main question being whether to put Pittasoma in its own family or subsume it into Conopophagidae. If only because I find tiny families unedifying if reasonable alternatives exist for combining them to better indicate relationships, I favor transferring Pittasoma to the Conopophagidae, at least for now. Hence, YES to all (a,b,c) of the "sub-proposals").

Comments from Zimmer: "YES. My votes on the various options of this proposal are as follows: a) yes; b) no; c) yes; d) no; e) yes.

I think that restricting Formicariidae to Formicarius and Chamaeza makes the most sense. Rhinocryptidae is so heterogeneous that its inclusion within Formicariidae is not only uninformative, but it would obscure relationships. Grallariidae seems like a natural grouping. The real dilemma is what to do with Pittasoma. The evidence for its relationship with Conopophaga appears strong, not only from a molecular standpoint, but also from a morphological and vocal standpoint. The question is at what level do we treat the similarities and differences between the two genera. In spite of certain similarities in plumage pattern and some vocal characters, I'm more impressed by the differences between the two groups. The difference in size and build is obvious. The two species of Pittasoma average 96-110 g in weight. The various gnateaters range from 20-28 g, except for the outsized melanogaster, which is still only 36-43 g, less than half of the size of any Pittasoma. The two species of Pittasoma lack the distinctive white or silvery postocular tufts that characterize all but one (melanops) of the species of Conopophaga. These postocular tufts are a conspicuous feature of the gnateaters, and are used prominently in display and territorial interactions with conspecifics. The tufts are arguably most developed and prominent in C. melanogaster, which, in its size, more terrestrial habits, vocalizations, and lowland Amazonian distribution, would otherwise seem to be the gnateater that is a possible bridge to Pittasoma. Pittasoma does not produce a mechanical wing-whirring sound, nor an accompanying chatter call, both of which are prominent features of all Conopophaga species except melanops. Both the wing-whirring and the chatter are regular features of male-female chases in the gnateaters, and nothing similar is seen in either species of Pittasoma. The two species of Pittasoma both have songs that involve incredibly long series (often lasting minutes rather than seconds) of well separated whistled notes (not linked together in a rattle), and arresting alarm calls that recall squirrels (Sciurus) more than birds. None of the gnateaters has a similar song, and only melanogaster has calls that even remotely resemble those of Pittasoma. Ecologically, the two groups are even more dissimilar, with Pittasoma being truly terrestrial birds that regularly follow army ant swarms, whereas Conopophaga are understory birds that descend to the ground but are not terrestrial, and they rarely attend ant swarms. The two Pittasoma species and the various gnateaters, treated separately, comprise two very distinct, and internally cohesive, uniform groups. Combined, they become much more heterogeneous, from a vocal, morphological, and ecological standpoint. The question becomes one of whether we treat the differences at the generic or familial level. I would favor treating them as being in separate families."

Comments from Jaramillo: "Great summary by Daniel of various results recently published on this subject. I agree with the recommendations, and do think that erecting a Pittasomidae seems unwarranted. The expanded Conopophagidae just takes a bit of getting used to, but otherwise it makes sense to me.

(a) YES- new circumscription of Formicariidae: Formicarius and Chamaeza
(b) NO- new circumscription of Formicariidae: Formicarius, Chamaeza, and the tapaculos.
(c) YES- accept family Grallariidae
(d) YES- place Pittasoma in Conopophagidae
(e) NO - erect family Pittasomidae"

Comments from Stotz on 235e: "YES. I find it hard to swallow Pittasoma and the gnateaters as a single family.  They constitute two very dissimilar groups (though not very species-rich).  I think we are better served by a separate family for Pittasoma."