Proposal (7) to South American Classification Committee

Combine New World barbets and toucans into a single family, Ramphastidae, as in AOU (1998)

The monophyly of the New World barbets and toucans is probably the most well verified clade in Aves! It has been established by morphology (first proposed in my undergrad thesis in 1982!), DNA hybridization by S&A, and several DNA studies by Lanyon and Co. I think Lester Short was the only one to disagree with this issue, but he agreed with the facts and suggested the be separate for classification. Further, all studies including both Semnornis and Capito or Eubucco have determined that New World barbets are paraphyletic- me proposing the Semnornis is closer to toucans, and recently Barker and Lanyon proposing that Capito and Eubucco are closer to toucans. Either way, Capitonidae is paraphyletic.  Basically, there is no way to recognize Capitonidae with New World barbets alone that does violate the well-established congruence among many independent phylogenetic hypotheses. Now, given that South America does not include any Old-World barbets, it is hard to know exactly what this classification implies for them. One could be most historical by splitting them into at least two new families. But the paraphyly of the new world barbets would still make Capitonidae paraphyletic.  In 1987 (in '82 in the thesis), I suggested that all barbets and toucans be lumped into one family, which I called Ramphastidae because the name is older than Capitonidae. Although strict priority does not apply to family names, I did this to prevent arbitrariness.  Over the years, I have sat on the side lines and seen this very reasonable suggestion be ignored by most major classifications, field guide, and reference works even though repeated studies have found that the toucans and barbets are monophyletic, that the New World taxa are monophyletic, that the toucans are monophyletic, but that NONE of the continental assemblages of barbets is monophyletic.  Finally, the AOU in 1998 recognized a single Ramphastidae including the barbets, but they still recognized a single Capitoninae, which would ALSO be paraphyletic.  So, I don't know what you all think, but I think we should lump all barbets and toucans into one family! At the very least, we are required to lump all the New World barbets and toucans in this checklist into one family. I chose Ramphastidae. Call it anything!  But splitting them as suggested here would be rejecting one of the single BEST supported, most agreed upon results in avian systematics in recent decades!  Let's compare that to recognizing the Galbuliformes as separate from Piciformes. All though this split has been recognized, there is far from agreement about it. Morphological phylogenetic papers have put them within Piciformes (Cracraft, Raikow). Criticism of this by Olson was completely unanalytical. Sibley and Ahlquist put them in outer space, but didn't really conclude what they could be related to. So, this situation is not universally accepted, has not been repeatedly supported. No relationships for the Galbuliformes have been proposed, as far as I know, other than in Piciformes or" we don’t know." I am all for progress, but we don’t know isn't progress because only by showing that the group is closer to something else than to Piciformes can you claim that the obvious similarities they share in the hind limb are convergent and not due to shared ancestry.  So, I wouldn’t necessarily vote against the recognition of Galbuliformes, but in comparison to the toucan/barbet situation, this idea is half-baked and poorly supported.  Of course, the earnest discussions have never really started over this checklist's mission. Perhaps we will have votes for maintaining the Capitonidae for the status quo. But that doesn't explain the non-traditional Galbuliformes.  I would LOVE to hear any comments on these thoughts! 

Rick Prum

_________________________________________________________________________________

Remsen comments: ”Semnornis is the "problem" taxon, right, so another possibility would be to create a third family for that genus until the relationships among the 4 monophyletic groups (assuming the two Semnornis are sisters) is resolved. (That's what I recommended to H-M group). That way, 4 certain monophyletic lineages would be given taxonomic recognition.

Prum comments: “Semnornis is not the only problem taxon. The others are Trachyphonus- the ground barbets, and Caloramphus - the little brown cooperative Asian barbets. Neither Barker nor Lanyon (Mol. Phil. Evol. 15:215-234) nor I found any support for Sibley's two Old World Families. The problem is that Sibley didn't include the problematic taxa Semnornis and Trachyphonus, nor the Asian Caloramphus, which is very different from Megalaima and Psilopogon (the members of Megalaimidae). Barker and Lanyon lacked the latter two.  So, there is essentially no support for the monophyly of any proposed concept of Lybiidae or Megalaimidae including Caloramphus.  And the only critical tests come from my work and Barker and Lanyon, and Lybiidae flunks both tests. There are no agreed-upon clades within Old World barbets.  So the solution presented here represents the following notion: There are at least four families of barbets: Lybiidae, Megalaimidae (including Caloramphus?), Capitonidae, Semnornithidae, Ramphastidae.  There is no support for the monophyly of the first two, and POSTIVE evidence in the only studies with enough sampling that they are NOT monophyletic. The third family is ok. The fourth is ok.  It should be a separate family given that it is either the sister group to Ramphastidae or the sister group to Cap. + Ramph. (Barker and Lanyon).  Given the volumes of data and analysis, I would happily concede that Barker’s hypothesis is better supported than mine. So, if you split up the barbets into multiple families, then you need to recognize Semnornithidae.  But, why on earth would we want to recognize four families of barbets, two of which have no support when a completely rational alternative option exists that is congruent with all data? Will we soon recognize five families of hummingbirds once we get a better idea of their phylogeny? Or 17 families of tanagers? etc.?  Believe me, I am all for saving the Linnean system, which many phylogeneticists are abandoning or simply ignoring. But to make it work we must make rational changes to the system to reflect history. I think proliferation of subfamilies within a family is much more tolerable than proliferation of uncertain families.  I say, we go with the AOU solution. -- note that AOU'98 used subfamily taxa for all three lineages above.  I stand corrected! And about this, I am pleased to be wrong! -- (just to make sure you know ... H&M does have Old World barbets in two separate families, the Sibley familiesSee above. This alternative is NOT supported by any data, and is contradicted by BOTH morphology and DNA sequence data.  Thanks for more information!

Remsen comments: “A couple of points concerning Rick's proposal: -- just for clarification ... the H-M classification does not include any Old-World barbets within Capitonidae and recognizes its sister relation to Ramphastidae. -- additional solutions to the "Semnornis problem" that would maintain family rank for toucans and barbets: (1) elevate the AOU’s Semnornithinae to family rank; (2) place Semnornis as Incertae Sedis (as AOU does with a bunch of tyrannoid genera, largely because of Rick’s studies, to avoid collapsing Tyrannidae, Pipridae, and Cotingidae into 1 family). Is incertae sedis something checklist's do when Rick Prum suggests a change? (This is not a serious question!) 

Comments from Prum: ”If we collapse birds as "different" as New World barbets and toucans into 1 family (as does AOU '98), then it would be difficult to argue that Furnariidae + Dendrocolaptidae should not also be merged into 1 family, as well as other 9-primaried oscine families.  What birds that are so different? What about the variation in size of falcons? Does anyone doubt that the smallest Old World falcons are really falcons? Some of them might even be genus Falco!?  Do we lose any information by expanding our concept of Ramphastidae by including some small ones? And is Semnornis big enough to be a toucan but Capito is not? What about Psilopogon? Looks kind of toucan-like to me.  The problem is not whether the members of various family groups are variable, but whether the groups are monophyletic. The question of ovenbirds and woodcreepers is whether ovenbirds are monophyletic without woodcreepers, not whether they are too different to be combined or too similar to be separated. Families are not comparable levels of diversity. They are just clades with names (hopefully).  Some families are big and some little. Some homogeneous and others variable. We should give up making them parallel or consistent since they never can be. We should just make them monophyletic. SO, I think we ought to mess with tradition only enough to render it historically accurate. 

Comments from Remsen: “Rick et al. -- I'm not getting my point across. The monophyly of the Dendrocolaptidae or Furnariidae is not an issue*. The reason I brought that up is that your (and AOU '98's) broad Ramphastidae encompasses taxa that are less "similar" by traditional, admittedly non-phylogenetic reasoning (as evidenced by placement of barbets in separate family for a century or so), and in my subjective opinion as well, than does the Dendro + Furn group. A century of taxonomic ranking as separate families would suggest to most that the difference between toucans and New world barbets is qualitatively different than your example of small Falco from large Falco. I don't think most of us think of barbets as just smaller toucans.  Given that ranking of higher-level taxa is largely subjective and guided mostly by historical momentum, I think it is worth considering two alternative solutions, namely (1) three families (Capitonidae, Semnornithidae, Ramphastidae) or (2) two plus Incertae Sedis status for Semnornis**... Looming ahead is the same problem with the Schiffornis bunch -- leave them as Incertae Sedis or do away with traditional family-level taxa Tyrannidae, Pipridae, and Cotingidae.

* I'm at home so can't check, but other than one phenotypic character noted by Clench (with incomplete taxon sampling), is there any evidence that the Dendros and Furnariidae are monophyletic with respect to each other?  

** Other than the distinctive bill tip, is there any evidence that Semnornis itself is monophyletic?

Comments from Whitney: "I agree with Rick, Mark, and probably about everyone else that we ought to leave systematics alone unless well-corroborated evidence for monophyly dictates a change. I also feel that "incertae sedis" is a useful place to put things that represent problems, or about which there exists significant doubt for whatever reason. This has two benefits: 1) it goes on hold rather than getting lost somewhere at a higher taxonomic rank, and 2) it is probably more likely to receive special attention, which is always good. I see no disadvantages to calling something or even a small group of taxa "incertae sedis". Is there a disadvantage I'm overlooking? C’mon, Rick, it's not all that many taxa!  Thus, I favor leaving "the big ones" as Ramphastidae and” the little ones" as Capitonidae. We all know what we are talking about, in part because that’s how we grew up, in part because these are unambiguous phenotypic groups (with Semnornis the odd one, "incertae sedis"). That said, and after reading the latest rounds of correspondence, I am still unclear as to whether phylogenetic analysis identified ANY break between these traditional assemblages or not. I'd like to have that answered; sorry if I've missed it along the way. I would be comfortable with Semnornis as ”incertae sedis" for now. And yes, Van, let's wait to verify that Semnornis is indeed monophyletic (if that piece of the puzzle is still missing).  As for Dendro and Furnar... if the tail is entirely rufous with powerful, sharp, strongly curved, well-exposed spines, it's a woodcreeper. If not, it isn’t. I don't think it needs to be any more complicated than that. One of these days, genetic evidence should reveal the split that’s clearly there -- if indeed it is recoverable from the organisms extant today.”

Comments from Stotz: "My feeling on 7 is that I'm not sure what the best treatment is, but I feel like putting everything in Ramphastidae sweeps remaining issues under the rug. Given that the Old World Barbets are now usually split into multiple families, I would think we could survive the same with the new world taxa.

Comments from Jaramillo: "I vote NO on this one, mainly because the option of separating the Ramphastidae, Capitonidae and Semnornithidae as Incertae Sedis is reasonable. I don't like the idea of lumping all into Ramphastidae, it makes the family less meaningful, information is lost in terms of understanding that a ramphastid is. I prefer the option of expanding the families rather than lumping."