Proposal (1046) to South American Classification Committee

Restore the monotypic genus Pterocnemia for Rhea pinnata

Effect on South American CL: This proposal would restore one genus to the checklist.

Background: SACC Proposal 348 (2008), submitted by Manuel Nores, suggested that Pterocnemia should be merged into Rhea, based primarily on the argument that the two rheas appear morphologically similar. Comments from José Fernando Pacheco included a reference to a 2007 publication by Delsuc et al. that documented a brood of hybrid chicks produced in captivity. The proposal passed on a 6 to 2 vote, although several members noted hesitancy. Some of the members voting in favor of the proposal suggested that retaining two genera was not ‘informative’, but there were also questions raised about whether there were other pertinent studies available that would help answer the question.

New Information: Some available information is not so new. For example, Haddrath and Baker (2001) analyzed mitochondrial DNA of a variety of living and extinct ratites, and they estimated that the two rheas diverged about 13.7 million years ago (see their Table 2). However, molecular systematics has advanced rapidly since then, and perhaps the most detailed analysis currently available, for Aves in general, is Stiller et al. (2024). This study did include both rheas, and their phylogeny (Extended Data Fig. 2) indicates that the two taxa diverged about 20 MYA, or perhaps slightly earlier. This rather ancient divergence is roughly similar in date, based on the Stiller et al. phylogeny, to that of Ibidorhyncha and Himantopus, or between Nectariniidae and Thraupidae, to give just two examples.

Recommendation: Molecular data indicates that the two rheas are not closely related, and their divergence is much older than the vast majority of congeneric bird species, so I would recommend voting ‘Yes’. This is one of many examples of molecular data indicating that ideas of relationships based of phenotypic similarities or differences may not reflect the actual evolutionary history of the taxa. It has perhaps been more frequent to merge taxa that are morphologically divergent when evidence of close relationship has become available (e.g. Hawaiian honeycreepers, the Vangidae, etc.), but splitting taxa that retain ancestral similarities (despite extensive genetic differences) may be less common. Nonetheless, both are part of the emerging pattern of morphology and genetics telling different evolutionary stories, a pattern that probably requires more attention and study. I suggest that lumping two species that are so genetically distinct into the same genus is misleading and obscures what may be interesting questions for further research. This proposal would not require description and naming of new taxa, since Pterocnemia appears to be a valid name and has been used in the checklist relatively recently, so this seems like an ‘easy fix’.

References:

Haddrath, O., & A.J. Baker. 2001. Complete mitochondrial DNA genome sequences of extinct birds: ratite phylogenetics and the vicariance biogeography hypothesis. Proc. R. Soc. Lond. B 268:939-945.

Stiller, J., and 51 other authors. 2024. Complexity of avian evolution revealed by family-level genomes. Nature 629:851-860, and supplementary information available online (https://doi.org/10.1038/s41586-024-07323-1).

D. James Mountjoy, April 2025

Addendum from Remsen: Here is a screen shot of the relevant portion of Extended Data Figure 2, which I encourage voters to peruse to look at comparative branch lengths throughout the nonpasserine tree:

Here are also some photos from Macaulay for comparisons – see photos for credits:

Voting Chart: https://www.museum.lsu.edu/~Remsen/SACCPropChart1044+.htm

Comments from Stiles: “YES. Looking at the phylogeny, Pterocnemia definitely should be split from Rhea (as the proposal stated, an ‘easy fix’.)”

Comments from Claramunt: “NO. I am reluctant to split two sister species into two different genera in general, and in this case, I don’t see compelling evidence of high levels of divergence. First, note that the Rhea node in Stiller et al, was calibrated: a minimum age 20.1 Ma was enforced on that node based on the phylogenetic position of an old fossil rheid that is in the R. pennata lineage (see Stiller et al. supplemental materials). So, the age has more to do with this calibration restriction than the genomic divergence between the two rheas, which is not unusually high by any means. See for example Fig. 1b in Haddrath and Baker: genomic (mtDNA) divergence between the two rheas is greater than the divergence between two moas but less than the divergence between emu and cassowaries (see also their Fig 2).

“I don’t find the age argument valid either. It’s easy to find younger genera among birds, but when focusing on palaeognaths, several other genera have comparable ages: look for example at the age of Crypturellus, Nothoprocta, and Nothocercus in the Stiller et al. tree, all in the early Miocene too.”

“The two rheas are very similar in all respects. I have seen misidentified specimens in museums! I fail to see any compelling evidence in the proposal for why we need to split these two species into different genera.”

Comments from Remsen: “YES. Without a fixed definition of what constitutes a genus, all arguments in a case such as this are subjective. It’s that subjectivity, however, which makes me favor using the independent variable, time, as a metric, at least in a crude way, to give genus and other higher categories some objective meaning. Most of you have heard my reasoning on this too many times already but … using time bands as the primary way of delimiting higher taxa at least provides some information content to these categories (vs. eclectic views of degree of morphological differentiation). Oscar Johnson, Eamon Corbett, and I have a manuscript showing that with few exceptions, the three most frequently used higher-level categories (order, family, genus) do correspond roughly to broad categories of lineage ages as estimated by time-calibrated trees -- but that’s not published, which is my fault. For now, you’ll just have to take my word for it or begin to pay attention to this when you see trees: taxa traditionally treated as genera are almost all 5-15 million years old (families 15-40 myo; orders 40-65+). Thus, traditional taxonomic ranks defined on perceived degree of differentiation are mostly consistent with lineage age far greater than expected by chance. Thus, the cost of using lineage age would cause few perturbations in traditional ranking schemes. The benefit is that it would facilitate comparisons of relative rates of morphological evolution, which to me is of greater biological importance than trying to do the opposite, namely use morphology to somehow define taxonomic categories. Use of time bands to demarcate higher taxa is classic “typology on a continuum”, but at least the continuum has a quantifiable scale and is independent of the taxonomy. End of sermon.

“As many have concluded, including Santiago above, the degree of morphological difference between the two rheas is probably closer to taxa treated as congeners than those in separate genera, although this was not considered a pressing problem by the many ornithologists who treated them in separate genera from 1871 until relatively recently. Body size and minor plumage differences on their own would not be likely characters to use in delimitation of genera. Degree of tarsal feathering is likewise not a character associated with generic differences; for example, Buteo regalis and B. lagopus are not placed in their own genus, and in Lagopus species, this character changes seasonally. That leaves tarsal scutellation pattern, a character considered to be of much greater importance to previous generations of avian systematists, before the advent of genetic techniques. I assume the rationale for its importance is that the patterns themselves have no known adaptive function (as far as I know) and thus might roughly reflect differences in ages of taxa. Here are close-ups of the tarsal scutellation differences in the two species (Greater on left):

“Lesser Rhea has the upper part of the tarsus feathered, which might be an adaptation for the much colder climate in which it lives. But look at the fairly substantial differences in the pattern of the scutes (which most references have described in misleading ways): transverse in americana and reticulate in pinnata except for the distal few. The number of transverse scutes is given as 8 to 18 by Blake (1977; Manual … ) and Davies (2002; Ratites and Tinamous; Oxford); the individual Lesser in the photo has few transverse scutes, but examining photos of additional individuals shows that others can have twice that many. I am not going to argue that this means that they should be placed in separate genera. The first step in that potential line of reasoning would be to do comparisons within other genera of birds, especially Paleognathae. My intent here is only to try to figure out why previous generations were comfortable placing them in separate genera.

“Santiago pointed out that the 20.1 mya node was enforced on the tree by a fossil, not by degree of genomic difference per se. My response is that if the two lineages can be distinguished morphologically going back to the early Miocene, then it is likely that genomic differences must also be substantial. These two lineages have been evolving separately for a long time.

“Looking through the nonpasserine tree in Stiller et al., I spotted 15 sets of congeners (including Ciccaba-Strix, which we treat as congeneric). Subtracting two sets that include species now treated by us as separate genera (Charadrius and Nyctibius) and one case of paraphyly (Pterocles-Syrrhaptes), that leaves 12 sets. Of those, only 5 have estimated node ages of more than 15 mya: the rheas, three genera of tinamous, and Anhinga. This comparison is strongly biased towards young node ages because not all species in a genus were sampled except for the rheas. The Rhea node is the oldest of the 5; note that 3 of the remaining 4 are Paleognathae. Regardless of taxonomy, this suggests the biologically interesting point of very conservative morphological evolution in those 5 sets.

“The estimated 20 mya split between the rheas is older than that estimated between Chionis and Pluvianellus, Glaucidium and Strix, Aptenodytes and Pygoscelis, Arenaria and Calidris, Odontophorus and Callipepla+Colinus, gulls and terns+skimmers, Circaetus (Old World snake-eagles) and Spizaetus+Aquila+Haliaetus, and others. Note that there are fairly large confidence intervals around all those nodes except the Rhea node; this makes precise comparisons of these estimates unwise. Nonetheless, treating the rheas as congeners is out-of-whack with just about everything else.

“Regardless of whether one thinks lineage age should be taken into consideration in taxonomy, the reason I vote YES on this one is that we (SACC) subsumed a long-standing genus (Pterocnemia) based on a superficial assessment of morphological similarities and a case of hybridization in an aviary, and even then, as noted in the new proposal, there was considerable ambivalence. In contrast, the case for separate genera is based on a comprehensive and sophisticated analysis that shows that treatment as congeners is THE extreme in the taxa sampled. For the tinamous or anhingas, in contrast, no other generic divisions have been used that I can find in modern times (fide Peters checklist synonymies).

“One final point on the often-stated opinion that use of monotypic genera provides no information. Indeed, monotypic genera provide no immediate information on sister taxa without reference to the phylogeny. This cost, however, is offset, in my opinion, by the benefit of calling attention to the point that data suggest that the lineage has been evolving separately for a very long time regardless of degree of morphological differentiation.”