Proposal (175) to South American Classification Committee

Change genera and linear sequence within the terns (Sterninae)

Effect on South American CL: This would make changes in the genera recognized and their placement within the terns reflect recent genetic data.

Background: Our current sequence of genera in the Sterninae is a conventional one, with the genus Sterna encompassing a high amount of the morphological variation in the subfamily. The list begins with the more gull-like members, and ends with the more oddball terns. I don't think there is much controversy in the arrangement of the terns; it just has not been looked at in many years. 

Sibley and Monroe (1990) classified the terns in 7 genera, including a very broad Sterna that includes the "brown-winged terns" (Sooty/Bridled and relatives) as well as the "little" terns, and "crested terns" (Royal-Elegant-Sandwich etc.). The more recent arrangement of Burger and Gochfeld (1996) classified the terns in 10 genera, separating the "crested terns" (Thalassoica), the Caspian Tern, and Gull-billed Tern from Sterna. They did not separate the "little" or "brown-winged" terns as different genera from Sterna, but they did imply that these are good clades within Sterna. Both of these arrangements agree on the separation of the noddies from the more typical terns and give them a basal position. The noddies include Anous, Procelsterna, and Gygis. They also agreed that the "marsh terns" Chlidonias are a valid genus. 

Our current sequence is closer to that of Sibley and Monroe than Burger and Gochfeld, with the main difference from S & M being the placement of the Inca Tern in our linear sequence:

Sterna nilotica Gull-billed Tern 
Sterna caspia Caspian Tern
Sterna elegans Elegant Tern 
Sterna sandvicensis Sandwich Tern 
Sterna maxima Royal Tern 
Sterna dougallii Roseate Tern 
Sterna hirundinacea South American Tern 
Sterna hirundo Common Tern 
Sterna paradisaea Arctic Tern
Sterna vittata Antarctic Tern 
Sterna forsteri Forster's Tern 
Sterna trudeaui Snowy-crowned Tern 
Sterna antillarum Least Tern 
Sterna superciliaris Yellow-billed Tern
Sterna lorata Peruvian Tern 
Sterna anaethetus Bridled Tern
Sterna fuscata Sooty Tern 
Chlidonias niger Black Tern 
Phaetusa simplex Large-billed Tern 
Anous stolidus Brown Noddy
Anous minutus Black Noddy 
Procelsterna albivitta Gray Noddy
Gygis alba White Tern 
Larosterna inca Inca Tern

New information: A recent set of work (Bridge et al 2005) using mtDNA sequence data (2008 bp) has looked at 33 species of terns and an additional two taxa (subspecies). The resulting phylogeny confirmed previous hypotheses of relationships within the group, and they suggested a revision of the terns which recognizes 12 genera. They showed that head ornamentation patterns in alternate plumages, presumably important in courtship, map very closely on to their phylogeny, adding strength to the results. In other words, the phylogeny makes good sense, and the division of the genus Sterna seems worthwhile as the new arrangement gives much more information on tern relationships.

Posterior probabilities support is high for most nodes, the weakly supported nodes being the placement of Phaetusa simplex, the placement of the Sterna trudeaui/forsteri clade, and placement of Chlidonias hybridus. Many of the clades defined by the Bayesian tree corresponded well with groups described in Gochfeld and Burger (1996). These groups are the noddies, "brown-winged terns", the "little", the marsh terns (Chlidonias), the crested terns, and the typical terns (Caspian, Inca, Large-billed, Gull-billed). S. caspia and S. nilotica form another distinctive clade, but these species were each placed in monotypic genera by Gochfeld and Burger (1996). Finally, two species, P. simplex and L. inca, do not appear to belong to any of these morphologically conservative clades.

The molecular phylogeny indicates that previous classification schemes are flawed because they included paraphyletic genera. The most general shortcoming is the failure to recognize the "little" terns (S. albifrons and allies) and the brown-winged terns (S. fuscata and allies) as groups distinct from the typical black-capped terns, which causes taxonomy to conflict with monophyletic groups. Two possible naming systems can resolve this. The first, and more conservative, recognizes only three genera: Anous, Gygis, and Sterna. This revision would leave Anous and Gygis unchanged but would group all other terns under the genus Sterna (including Chlidonias). The second is the suggested classification which modifies that of Gochfeld and Burger (1996) to include two additional genera in recognition of the distinct clades formed by the brown-winged and small terns, bringing the number of genera among the terns up to 12. The latter option is preferred as it better reflects the structure of the molecular phylogeny, and also matches well with morphological data (size, head ornamentation pattern). Bridge et al (2005) suggested resurrecting the genera Onychoprion, which Wagler (1832) created in his synonymous description of S. fuscata, and Sternula, which Gould (1843) put forth in the original description of S. nereis, to distinguish the brown-winged clade and the little terns, respectively. Bridge et al. (2005) stated that designation of several monospecific genera (i.e., Phaetusa, Larosterna, Gelochelidon, and Hydroprogne) used by Gochfeld and Burger (1996) is warranted both to maintain some degree of continuity with currently used naming systems and to designate these four species as being morphologically unique and highly divergent among the terns. They were unable to offer empirically based taxonomic recommendations regarding Procelsterna because no tissues were available, but considering its distinctive plumage, they suspect that it should retain its own generic status.

Of interest to the committee, but a bit tangential to this proposal is that they cannot conclusively address the controversy regarding whether to designate S. sandvicensis sandvicensis and S. s. eurygnatha as different species. The small (0.29%) genetic divergence suggests that these two taxa should be regarded as subspecies; however, they may also constitute two species that have diverged quite recently. The decision to split these taxa into two species requires further research with many vouchered samples from throughout their ranges, particularly in the Caribbean where the two subspecies commonly hybridize (Hayes, 2004).

The suggested sequence is as follows with areas of most uncertain placement (weak nodes in phylogeny) highlighted in red:

Anous stolidus Brown Noddy
Anous minutus Black Noddy 
Procelsterna albivitta Gray Noddy
Gygis alba White Tern
Onychoprion fuscata Sooty Tern
Onychoprion anaethetus Bridled Tern
Sternula antillarum Least Tern 
Sternula superciliaris Yellow-billed Tern
Sternula lorata Peruvian Tern 
Phaetusa simplex Large-billed Tern
Gelochelidon nilotica Gull-billed Tern 
Hydroprogne caspia Caspian Tern
Larosterna inca Inca Tern
Chlidonias niger Black Tern
Sterna hirundo Common Tern 
Sterna dougallii Roseate Tern 
Sterna paradisaea Arctic Tern
Sterna hirundinacea South American Tern 
Sterna vittata Antarctic Tern
Sterna forsteri Forster's Tern 
Sterna trudeaui Snowy-crowned Tern 
Thalasseus sandvicensis Sandwich Tern 
Thalasseus elegans Elegant Tern
Thalasseus maximus Royal Tern

Analysis and Proposal: This phylogeny appears to be quite solid, and for the most part it matches quite well with published taxonomies, or at least proposed clades, of this group. The main problem in the taxonomy of the terns is that of how broad to make the genus Sterna. The present data set strongly supports that the "brown-winged terns" (Onychoprion) and the "little terns" (Sternula) are not only good clades in themselves, but are quite distantly related to Sterna, and as such they should be separated from Sterna. The placement of Phaetusa is not well resolved, nor is the Forster's/Snowy-crowned clade, which shows some affinities with the crested terns (Thalasseus) in some of their trees. However, based on size, morphology and even voice it seems a good decision to leave those in Sterna. The Gull-billed and Caspian terns are shown to be sister species, although not that closely (genetic distance) related to each other, one could argue either way to lump them both under one genus or keep them separate. Keeping them separate, as is done here, is a more traditional approach.

Finally, the multi-species genera in this new phylogeny show many morphological similarities, and nothing is really "way out there" in these results. They make a great deal of sense. In addition, vocal data (from my personal experience) really matches well to this phylogeny. For example, vocally Caspian tern is very much unlike Thalasseus vocally, although it resembles the Royal Tern outwardly, and in tern the Thalasseus I know all sound pretty similar. Vocally, the Gull-billed Tern has some harsh sounds that are supportive of a relationship with Caspian. New World Sternula, are vocally more similar to each other than to other terns, and similarly Onychoprion are quite different to other terns and more similar to each other (Aleutian being a bit of an exception, although the relationships are there when you listen to a full range of vocalizations within this group).

Recommendation: Because our linear sequence and classification should reflect phylogenetic data, and because the data appear solid, I will vote YES on this new re-arrangement of the terns. Whatever problems there might be with this sequence, it is grounded in phylogenetic hypotheses and data and is certainly closer to the true phylogeny of the terns than any other sequence currently in use.

References:

Bridge, E.S., Jones, A. W., Baker, A.J. 2005. A phylogenetic framework for the terns (Sternini) inferred from mtDNA sequences: implications for taxonomy and plumage evolution. Molecular Phylogenetics and Evolution 35: 459-469.

Gochfeld, M., Burger, J., 1996. Family Sternidae (Terns). In: del Hoyo, J., Elliot, A., Sargatal, J. (Eds.), Handbook of the Birds of the World, Vol. 3. Lynx Edicions, Barcelona, Spain, pp. 624-667.

Gould, P.J., 1843. Descriptions of thirty new species of birds from Australia. Proc. Zool. Soc. Lond., Part 10, No. 117, 131-164.

Hayes, F.E., 2004. Variability and interbreeding of Sandwich Terns and Cayenne Terns in the Virgin Islands, with comments on their systematic relationship. N. Am. Birds 57, 566-572.

Sibley, C.G., Monroe, B.L.J., 1990. Distribution and Taxonomy of Birds of the World. Yale University Press, New Haven, CT.

Wagler, J.G., 1832. Onychoprion serrata. Isis von Oken 25, 277.

Alvaro Jaramillo, May 2005

Comments from Remsen: "YES. Alvaro's proposal covers the situation well, and I see no reason not to adopt the proposed classification. I asked Eli Bridge for comments, and he thought the proposal was excellent. I also asked co-author Andy Jones and received the following:

'Alvaro's summary is well-written and accurate; the only comment I have is that the Phaetusa branch occasionally moves around because it is a fairly long branch and sometimes has long-branch attraction issues. It didn't consistently attach elsewhere in the tree, so we are fairly sure that it is in the correct placement, but cannot have a higher posterior probability value because of the multiple substitutions in its history.'

Comments from Zimmer: "YES. Alvaro has provided a good review and a sound rationale for the proposed changes. I would agree with his comments regarding vocal characters fitting the new arrangement better."

Comments from Michael Gochfeld & Joanna Burger:

"Gochfeld and Burger comments on Alvaro Jaramillo's Proposed List for South American Terns. (June 21, 2005)

"SUMMARY: This is a good and justifiable list if the Bridge et al. work can be replicated. There is always a question about major revisions based on one character set. But maybe this is the definitive set. Placing Phaetusa and particularly Larosterna right in the middle of the list seems awkward. The addition of the two genera for the "little terns" and "Sooty tern group)" is an excellent modification to the genera we listed in the Handbook of Birds of the World (Gochfeld and Burger 1996), although Onychoprion is an awful name for a genus. It remains to be validated that certain little known species are correctly assigned (this doesn't much impact the South American list). Although we tend to be splitters, the evidence we have reviewed supports considering eurygnatha part of sandvicensis, although it may be a color variant rather than a subspecies. Parenthetically, we have seen birds with black bills and yellow tips mated to birds with all orange bills, both in Puerto Rico (Culebra) and in Argentina (Peninsula Valdez).

"COMMENTARY
We are pleased to comment on the listing of the terns for South America, and are sorry it took so long to get to this. We have had some family and work priorities. We also appreciate the reference to the Bridge et al. paper in May 2005, which we had not yet seen. It is very gratifying to see their results.

"Many of the clades defined by the Bayesian tree corresponded well with informal groups described in Gochfeld and Burger (1996). These groups are the noddies (Anous, Gygis, and Procelsterna), the brown-winged terns (four species of Sterna), the small terns (four species of Sterna), the marsh terns (Chlidonias), the crested terns (Thalasseus in Gochfeld and Burger (1996)), and the typical terns (several Sterna species). S. caspia and S. nilotica form another distinctive clade but these species were each placed in monotypic genera by Gochfeld and Burger (1996). Finally, two species, P. simplex and L. inca, do not appear to belong to any of these morphologically conservative clades.

"Although we are skeptical (or guardedly optimistic) about molecular phylogeny as the sole basis for phylogeny, in this case it seems to confirm and clarify at least the clades (we remember that each time Sibley changed technologies (egg whites to serum to DNA) some "relationships" changed, and he anchored his with traditional characters. .Hopefully further studies of terns will be done and will come out with the same groupings. We also considered Sternidae and Laridae separate at the family level.

"We have had personal field experience with almost all the terns, excluding the two species of Procelsterna, Thalasseus zimmermani (or bernsteini), and the Kerguelen Tern.  We have seen most of the other species on their breeding grounds.

"In preparing that chapter, we discussed repeatedly whether the little terns (Sternula) should be recognized as a separate genus. We favored it because they are such a coherent sub-group. Probably the editors encouraged us not to be too radical in a handbook, and it had been decades since Sternula was in use. We did not consider separating the Sooty, Gray-backed, and Bridled Tern, partly because the alternative genus name is unpronounceable and non-euphonious. But now that we no longer work on these species, we applaud their separation. So, adding these two genera to Gochfeld & Burger (1996) makes sense.

"There are nuances unrelated to South America. Sterna aleutica, for example, doesn't strike us as Sooty-like, nor does S. albostriata, which nests on rocky islands in New Zealand's braided rivers, act marsh tern like at all. Jaramillo mentions voice as a clue, and aleutica does have a multinote call which could be construed as similar to fuscata. But actually, Gull-billed Terns (at least in NJ) have a multi-note call which is somewhat reminiscent of Sooty Tern. So much as we like tern voices, it is not likely that they have phylogenetic implications.

"We actually find it surprising that Gelochelidon and Hydroprogne are sisters. Nothing in our experience would suggest it. It's hard to believe that their relationship to the typical Sterna is further removed than is Larosterna, and we would have expected Hydroprogne to be closer to the "crested terns" (Thalasseus). But of course we are impressed by the behavior, even though that is ecologically modifiable and can be convergent. It is gratifying that neither of these is particularly gull-like.

"We have no doubt that S. forsteri belongs with the typical terns. If we mentioned (G&B) that there was a question about the relationship, we shouldn't have implied that we questioned it. Similarly, it's great to see how tight the mitochondrial data support S. dougallii as a typical tern.

"Although we tend to be splitters, the evidence we have reviewed supports considering eurygnatha part of sandvicensis, although it may be a color variant rather than a subspecies. In the Caribbean there is continuous variation in bill color, so it can't be a morph. And we have seen birds with black bills and yellow tips mated to birds with all orange bills, both in Puerto Rico (Culebra) and in Argentina (Peninsula Valdez). In both places we could ascertain that these were nesting birds, although the results of the mating were not followed.

"It is unfortunate that Bridge et al had no material from Procelsterna and we have no personal experience with it. We followed others in recognizing two species, which could also be well-marked latitudinal representatives.

"Turning now to the sequence of the species list and the cladogram as well as the comments of Alvaro Jaramillo (whose book on Chilean birds were have been studying),

"It is interesting and novel to see the Noddies appear first on the list. The other sequence follows from the cladogram. Although we are hesitant to rely on a single character to turn the world upside down, we don't have a really different suggestion. However, it seems awkward to embed Phaetusa and particularly Larosterna in the middle of this sequence.

"It's not that we don't believe the data, but Bridge et al, talk a lot about plumage evolution, and Larosterna is pretty far-out for a tern and Jaramillo and Bridge note this.

Comments from Robbins: "YES, to changing the linear sequence of the Sterninae to reflect the new molecular data."

Comments from Stotz: "YES, but with hesitations and reservations and not complete agreement. In general, this looks good. Two problems concern me. One is that I have to agree with Berger and Gochfeld that Larosterna in the middle of things seems like a problem. The second is that the resurrection of Thalasseus is not required by the data, and is not discussed as far as I can tell by Bridge et al at all. Sterna with Thalasseus included would still be monophyletic, and it would mean that Bridge et al.'s concern about the placement of forsteri and trudeaui would be moot, at least with respect to generic placement. I would be interested in the comments of a molecular sort about this paper and the taxonomic conclusions."

Comments from Stiles: "YES. (I was never all that convinced by a broad Sterna anyway, and am glad to see several oddballs recognized as such. I am not unhappy with Thalasseus as a genus - it seems to be a coherent group set apart from the typical terns, such that it would probably merit at least subgenus status in any case."

Comments from Jaramillo: "YES - Great to get these comments from Gochfeld and Burger. I do think that having Larosterna and the oddballs in the middle of the phylogeny looks odd, uncomfortable, but in the end I am comfortable with it. This begs the question, why is Larosterna so ornately ornamented; seems like an interesting behavioral ecology project for someone. Thalasseus could be included in Sterna, and in effect this is a matter of personal opinion. Since the data are clear that it is a monophyletic group, and there are consistent morphological and vocal differences between Thalasseus and Sterna, I think that dividing them up gives each genus a larger "information content" than keeping them together."

Comments from Nores: "YES. La nueva secuencia soportada por análisis mitocondriales aparece en general como bien fundada. Sin embargo, como señalan Gochfeld y Burger no estoy convencido que la filogenia molecular de por sí sola sea suficiente para construir filogenias. Me parece casi imposible que Sternula y Sterna que son tan similares entre sí, puedan estar separados por cosas tan diferentes como son Larosterna inca, Phaetusa simples y Gelochelidon nilotica. Tampoco estoy muy de acuerdo en separar Thalasseus de Sterna, pero en caso de hacerlo Thalasseus tendría que ir antes de Sterna en la secuencia."