Proposal (250) to South American Classification Committee

Revise genera and linear sequence within the gulls (Larinae)

Effect on South American CL: This would make changes in the genera recognized and their placement within the gulls reflecting recent molecular data.

Background: Our current sequence of genera in the Larinae is a conventional one, with the genus Larus encompassing a high amount of the morphological variation in the subfamily. The list begins with the somewhat atypical members of the larger white-headed gull group, continues with the typical white-headed gulls, followed by an assortment of smaller black-headed gulls, and ends with the oddities (kittiwake, Sabine's and Swallow-tailed). The systematics of gulls has been very, very controversial as a whole, particularly in the large white-headed gull group. Fortunately, we don't have many of those to deal with in South America. 

Dwight (1925) split the Laridae family into two large groups: (1) the Larae comprises large species with a white head in breeding plumage (except Larus ichythyaetus, L. fuliginosus, L. hemprichii, L. leucophthalmus) and (2) the Xemae small species often a dark hooded in breeding plumage. Within each group, some species were assigned to monospecific genera (Gabianus pacificus, Leucophaeus scoresbii, Pagophila eburnea for Larae; Xema sabini, Rhodostethia rosea and Creagrus furcatus for Xemae) or to a genus with two species in the case of the Kittiwakes (Rissa). Similarly, Moynihan (1959) on the basis of behavior, vocalizations, and plumage also proposed to divide gulls into two large groups assigned to the subgenus Larus and Xema. The main divergence with Dwight concerned the placement of species with a dark hood or a dark plumage that were further divided into 'masked' (the small black-headed ones, Brown-hooded and Andean for example) and 'primitive' hooded species (Laughing, Franklin's etc.) groups. On the basis of osteological and integumentary characters, Chu (1998) assigned gull species to 'Sternines' that comprised the 'masked' species group as well as Rissa, Xema, Pagophila, Rhodostethia and to 'Larines' corresponding to Dwight's Larae plus Moynihan's "primitive" hooded species.

Our current linear sequence is:

Larinae

Leucophaeus scoresbii Dolphin Gull 

Larus belcheri Belcher's Gull 

Larus atlanticus Olrog's Gull 

Larus modestus Gray Gull

Larus delawarensis Ring-billed Gull 

Larus dominicanus Kelp Gull 

Larus argentatus Herring Gull 

Larus fuscus Lesser Black-backed Gull 

Larus cirrocephalus Gray-hooded Gull 

Larus maculipennis Brown-hooded Gull 

Larus serranus Andean Gull 

Larus fuliginosus Lava Gull 

Larus atricilla Laughing Gull 

Larus pipixcan Franklin's Gull 

Larus minutus Little Gull 

Rissa tridactyla Black-legged Kittiwake 

Xema sabini Sabine's Gull 

Creagrus furcatus Swallow-tailed Gull

New information: Crochet et al., (2000) provided a molecular phylogeny based on mtDNA sequences that supported the hypothesis that the "hooded" species belonged to two basal lineages as suggested by Chu (1998). In addition, they found several species-groups that were strongly supported by mtDNA data but were mostly incongruent with previous hypothesis on gulls' relationships due to numerous instances of convergence between unrelated species or quick divergence of closely related species. Several species were missing in the data set of Crochet et al. (2000). The results of Crochet et al. (2000) demonstrated that the current nomenclature of the Laridae (using the genus name Larus for most species but a few morphologically divergent ones, as in Burger and Gochfeld, 1996) is inadequate, as Larus as currently used is not a monophyletic clade. Pons et al (2005) continued this line of work by proposing a phylogeny based on a mtDNA composite segment (parts of cytochrome b and control region) and included all gull species recognized by Burger and Gochfeld (1996). Their results show that gulls constitute a monophyletic clade, separate from the terns, however the genus Larus, as currently recognized, is not monophyletic. The major clades in Pons et al (2005) receive strong support using various analyses (Maximum Parsimony, Maximum Likelihood, Bayesian), although the basal notes (relationships among species groups) are less well supported. 

To match nomenclature with phylogeny, three options are possible.

1) The least preferred option as suggested by Moynihan (1959) and Chu (1998), is to place all gulls in the genus Larus. This is the least taxonomically informative treatment. 

2) To assign a generic rank to each of the two main gull clades (Larus and Xema). This is only weakly supported by the molecular study, and conceals much of the diversity within the group and would remove some well-established genera. 

3) To revalidate several genus names which are no longer in use. This would give generic rank to each of the main species groups supported by the genetic results, and often by behavioral and morphological data. 

The gulls would then be divided into the following genera (bold is one which is found within our region): 

1-Rissa (R. tridactyla, R. brevirostris); 

2-Creagrus furcatus; 

3-Hydrocoloeus Kaup, 1829 (H. minutus, H. roseus)

4-Pagophila eburnea; 

5-Xema sabini

6-Chroicocephalus Eyton, 1836 for the "masked" species; 

7-Leucophaeus Bruch, 1853 for the 'hooded' species group; 

8-Ichthyaetus Kaup, 1829 for the 'black-headed' species group; 

9-Larus for the ="white-headed" species 

10-Saundersilarus saundersi.

Note that the "black-headed," "hooded," and "white-headed" species groups form a monophyletic clade in all analyses, the amount of divergence (genetic, morphological, and behavioral) among them is similar to the divergence among the other genera of gulls, and Pons et al. (2005) prefer to treat them as distinct genera.

Following the suggestions in Pons et al (2005) and Crochet et al (2000) a linear sequence that better reflects phylogeny in the gulls is:

Larinae

Creagrus furcatus Swallow-tailed Gull

Rissa tridactyla Black-legged Kittiwake 

Xema sabini Sabine's Gull 

Chroicocephalus serranus Andean Gull

Chroicocephalus maculipennis Brown-hooded Gull

Chroicocephalus cirrocephalus Gray-hooded Gull 

Hydrocoloeus minutus Little Gull

Leucophaeus scoresbii Dolphin Gull

Leucophaeus modestus Gray Gull

Leucophaeus atricilla Laughing Gull - I assume it remains atricilla, not atricillus? 

Leucophaeus pipixcan Franklin's Gull

Leucophaeus fuliginosus Lava Gull 

Larus belcheri Belcher's Gull 

Larus atlanticus Olrog's Gull 

Larus delawarensis Ring-billed Gull 

Larus dominicanus Kelp Gull 

Larus fuscus Lesser Black-backed Gull

Larus argentatus Herring Gull

Creagrus, Rissa, and Xema are basal, whereas Larus is the newest group.

Of interest to the committee, but tangential to this proposal is that Larus belcheri and atlanticus are found to be less closely related in terms of genetic distance than one would predict from their appearance. In other words, the Pons et al (2005) data supports species status for atlanticus. In addition this and other works strongly suggest that smithsonianus is not closely related to argentatus, something I will try and tackle in a separate proposal (my guess is that the AOU is doing the same?).

Analysis and Proposal: This new phylogeny seems to be quite solid, even though it differs in many ways from the traditional arrangement of the gulls. This is not surprising as many of these species appear to be relatively new (large white headed group) and morphological divergence is often minimal, and convergence in appearance is common. Interestingly often immature plumages match up more closely to the proposed phylogeny than the adult plumages. The main issue is how wide to make Larus. I think that these new works make a good argument for using a wider number of genera that can be applied to the major clades of gulls, rather than having too broad a Larus, which would mask much of the diversity within this group. 

A vocal analysis of the group would be interesting as I am sure that it would match quite well to the proposed phylogeny. Gulls in Leucophaeus sound distinctly different from Chroicocephalus or Larus for example. Although vocally Leucophaeus modestus and atricilla are so similar, that I have a hard time believing they are not sister species.

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 gulls. 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 gulls than any other sequence currently in use. 

References:

Burger, J., Gochfeld, M., 1996. Family Laridae (Gulls). In: del Hoyo, J., Elliot, A., Sargatal, J., (Eds.), Handbook of the birds of the world, vol. 3, Hoatzin to Auks. Lynx Edicions, Barcelona, Spain, pp. 572-623.

Chu, P.C., 1998. A phylogeny of the Gulls (Aves: Larinae) inferred from osteological and integumentary characters. Cladistics 14, 1-43.

Crochet, P.-A., Bonhomme, F., Lebreton, J.-D., 2000. Molecular phylogeny and plumage evolution in gulls (Larini). J. Evol. Biol. 13, 47-57.

Dwight, J., 1925. The gulls (Laridae) of the world: their plumages, moults, variations, relationships and distribution. Bull. Am. Mus. Nat. Hist. 52, 63-401.

Moynihan, M., 1959. A revision of the family Laridae (Aves). American Museum Novitates 1928, 1-42.

PONS, J.-M., A. HASSANIN, AND P.-A. CROCHET. 2005. Phylogenetic relationships within the Laridae (Charadriiformes: Aves) inferred from mitochondrial markers. Molecular Phylogenetics & Evolution 37: 686-699.

Alvaro Jaramillo, December 2006

_______________________________________________________________________________________________

Comments from Stiles: "YES. Although it seems to be only a single gene, there appears to be morphological and vocal evidence (at least the former published) supporting the rearrangement of gull genera as proposed."

Comments from Zimmer: "YES". Although I wouldn't be surprised to see still more revision within this group, I think this is a nice starting point, and the greater recognition of generic-level diversity within the family also fits nicely with a number of vocal, morphological and ecological differences. Larus (in the revised sense) coming at/near the end of the sequence makes a lot more sense to me as well, and would certainly seem to jive with what appears to be a fairly recent radiation of a number of species."

Comments from Pacheco: "YES. Reconheço igualmente esta proposição, baseada em evidências genética, morfológica e vocal, como um ponto de partida após décadas de sequência meramente tradicional."

Comments from Remsen: "YES. Genetic data look solid, and the concordance with various phenotypic characters is reassuring. I went through Pons and created a linear sequence without looking at Alvaro's proposal: the two were identical."