Molecular Phylogenetics and Evolution 35 (2005) 459–469 www.elsevier.com/locate/ympev A phylogenetic framework for the terns (Sternini) inferred from mtDNA sequences: implications for taxonomy and plumage evolution Eli S. Bridgea,¤, Andrew W. Jonesb, Allan J. Bakerc a b Department of Biology, University of Memphis, 3700 Walker Ave., Ellington Hall 103, Memphis, TN 38152, USA Department of Ecology, Evolution, and Behavior, Bell Museum of Natural History, University of Minnesota, 100 Ecology Bldg., 1987 Upper Buford Circle, Saint Paul, MN 55108, USA c Centre for Biodiversity and Conservation Biology, Department of Natural History, Royal Ontario Museum, 100 Queens Park, Toronto, Ont., Canada M5S 2C6 Received 5 July 2004; revised 3 December 2004 Available online 25 January 2005 Abstract We sequenced 2800 bp of mitochondrial DNA from each of 33 species and 2 subspecies (35 taxa) of terns (Sternini), and employed Bayesian methods to derive a phylogeny with good branch support based on posterior probabilities. The resulting tree conWrmed many of the generally accepted taxonomic groups, and led us to suggest a revision of the terns that recognizes 12 genera, 11 of which correspond to a distinct clade on the tree or a highly divergent species (1 genus was not represented in the phylogeny). As an example of how the molecular phylogeny reXects similarities in morphology and behavior among the terns, we used the phylogeny to examine the evolution of the breeding (alternate) head plumage patterns among the terns to test the hypothesis that this character is phylogenetically informative. The three basic types of head plumage (white crown, black cap, and black cap with a white blaze on the forehead) were highly conserved within clades, with notable exceptions in two white-crowned species that evolved independently among the black-capped terns. Based on the appearance of the close relatives of these exceptional species, their white crowns appear to be due to the retention of either winter (basic) plumage characteristics or perhaps juvenile characteristics when the birds molt into their breeding plumage. Examination of the evolutionary history of head plumage indicated that the white-crowned species such as the noddies (Anous) and the white tern (Gygis alba) are probably most representative of ancestral terns.  2004 Elsevier Inc. All rights reserved. Keywords: Tern; Noddy; Charadriiformes; Plumages; Phylogeny; Taxonomy; Evolutionary history; Sternini; Sternidae; Sterninae 1. Introduction The terns (Charadriiformes: Laridae: Sternini) are a distinctive group of seabirds that occupy aquatic environments the world over and demonstrate an interesting array of variations on a life history centered around aquatic foraging and colonial nesting. Among the terns is the Arctic tern (Sterna paradisaea), which migrates ¤ Corresponding author. Fax: +1 901 678 4767. E-mail address: ebridge@memphis.edu (E.S. Bridge). 1055-7903/$ - see front matter  2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2004.12.010 farther than any other animal, as well as several species with entirely sedentary life histories. The terns also demonstrate a diverse array of nesting habits, social behaviors, and molting patterns. However, understanding of the evolutionary history of these variable life history characteristics and our capacity to use the terns in comparative studies are limited by the lack of a well-supported systematic analysis of the evolutionary relationships among these birds. According to Sibley and Monroe (1990), the terns comprise a tribe, Sternini, of 45 species in 7 genera, with 460 E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 the majority of the terns (32 species) classiWed under the genus Sterna (Table 1). Other classiWcation schemes recognize the terns as a subfamily, Sterninae (e.g., American Ornithologist’s Union, 1998, Higgins and Davies, 1996). The widely accepted classiWcation system for the terns appears to have been inXuenced largely by Moynihan’s (1959) taxonomic revision of the Laridae. Moynihan (1959) used his knowledge of general morphology and behavior to classify the terns as shown in Table 1, with three major groupings worthy of generic status: the noddies (Anous), the Inca tern (Larosterna), and the blackcapped terns (Sterna). In contrast to Moynihan’s extremely “lumped” revision of the terns, a more recent classiWcation by Gochfeld and Burger (1996) divided the terns among 10 genera (Table 1). The groups recognized by both of these classiWcation schemes are based largely on speculative criteria such as general appearance and behavior. Additionally, the utility of these morphologyand-behavior-based classiWcations for furthering our understanding of the evolution of life-history traits is limited because any inference about behavior or morphology from such schemes is circular. Previous studies of evolutionary relationships among the terns are generally lacking in either their comprehensiveness or analytical rigor. In large-scale cladistic studies of the Charadriiformes only 12 species of terns were included and their relationships were unresolved (Chu, 1995; Strauch, 1978). Similarly, Sibley and Ahlquist’s (1990) DNA–DNA hybridization study included only four tern species, and Hackett’s (1989) sequential electrophoresis analysis included 14 tern species. Thus, these studies had relatively poor representation of the 45 extant species of terns. The most comprehensive assessment of tern relationships that employed systematic methodology is Schnell’s (1970a,b) phenetic study of the Laridae, which included 42 tern species. However, the results from this study are diYcult to interpret in terms of phylogenetic relationships. Schnell (1970a,b) summarized the results of various analytical techniques applied to diVerent morphological data sets in 14 phenograms, all of which show fundamentally diVerent topologies. Notably, neither Hackett’s (1989) nor Schnell’s (1970a,b) studies were speciWcally attempting to construct a phylogeny of the terns. In this paper we present the Wrst hypothesis of phylogenetic relationships among the terns using dense taxon sampling and current methods of tree-building with DNA sequence data. To demonstrate the utility of the phylogeny in understanding how characteristics related to behavior and morphology are distributed among the terns, we examined how the three distinct forms of head plumage found in terns relate to the phylogenetic relationship deWned by the tree. The majority of terns have a distinctive black cap that often contrasts markedly with gray and white body plumage. A few terns have a similar black cap but bear a white blaze on the forehead that extends from the base of the bill to just posterior to the eyes (see Fig. 1). A third type of head plumage is that of the noddies (Anous) and the white tern (Gigys alba), wherein the crown is entirely white. We used our mtDNA phylogeny to evaluate whether these head plumage-based groups correspond to groups of closely related species and to test whether head plumage is a phylogenetically conserved character. 2. Methods 2.1. Taxon sampling and DNA sequencing Species names used throughout this paper follow Sibley and Monroe (1990). Among the taxa included in our study was the “Cayenne” tern (Sterna sandvicensis eurygnatha), which is widely recognized as a South American subspecies of S. sandvicensis (Gochfeld and Burger, 1996; Sibley and Monroe, 1990) and often hybridizes with the North American S. s. acuXavida subspecies (Hayes, 2004). However, S. s. eurygnatha is morphologically distinct from other S. sandvicensis subspecies (Junge and Voous, 1955) and accordingly is given species status by some authors (e.g., Harrison, 1983). We sequenced mtDNA from tissue samples of 35 tern species or subspecies and from 1 gull species (Larus delawarensis), which served as an outgroup, such that the total number of taxa was 36 counting the two subspecies of S. sandvicensis (see Table 2). We chose to use sequences from a gull, L. delawarensis, to root the tree based on the close phylogenetic relationship between the gulls (Larini) and the terns (Paton et al., 2003). Recognized species not included in the phylogeny are noted in Table 1. Most of the tissue samples used for DNA sequencing came from the tissue holdings at the University of Minnesota Bell Museum of Natural History and the Royal Ontario Museum. These samples were supplemented with donations from the University of Michigan Museum of Zoology, the Louisiana State University Museum of Natural Sciences, the Field Museum of Natural History, the South Australian Museum, and the Zoological Museum University of Copenhagen. Many of the samples from the Royal Ontario Museum lacked museum vouchers, but we were often able to guard against errors in identiWcation and record keeping by sequencing at least one of the targeted DNA regions from a second member of the same species and comparing these sequences to conWrm the identity of the Wrst (Table 2). Genomic DNA was isolated using either a DNeasy Tissue Kit (Qiagen, Valencia, CA) or by following a variation on the phenol–chloroform extraction protocol of Hillis et al. (1996). Mitochondrial DNA from part of the cytochrome b (cyt b) gene, the entire NADH 2 (ND2) 461 E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 Table 1 ClassiWcations of the terns from past studies and the recommended classiWcation based on the mtDNA phylogeny Moynihan (1959) Sibley and Monroe (1990) Gochfeld and Burger (1996) Suggested Anous Dark noddies Anous stolidus A. minutus A. tenuirostris Intermediate noddies A. cerulea A. albivitta White noddies A. alba Larosterna Inca tern Larosterna inca Sterna Little terns Sterna albifrons S. superciliaris S. nereis S. lorata Gull-billed tern S. nilotica Large-billed tern S. simplex Marsh terns S. niger S. leucoptera S. hybrida Crested terns S. caspia S. maxima S. bergii S. sandvicensis S. elegans S. bernsteini S. eurygnatha Typical black-capped terns Sterna dougallii S. sumatrana S. hirundinacea S. vittata S. virgata S. paradisaea S. aleutica S. striata S. forsteri S. trudeaui S. repressa S. balaenarum S. lunata S. anaethetus S. fuscata S. acuticauda S. aurantia S. albostriatus S. hirundo Anous Anous stolidus Anous minutus Anous tenuirostris Procelsterna Procelsterna cerulea P. albivitta Gygis Gygis alba Gygis microrhyncha Phaetusa Phaetusa simplex Larosterna Larosterna inca Chlidonias Chlidonias albostriatus C. hybridus C. leucopterus C. niger Sterna Sterna nilotica S. caspia S. aurantia S. maxima S. elegans S. bengalensis S. bergii S. bernsteini S. sandvicensis S. dougallii S. striata S. sumatrana S. hirundinacea S. hirundo S. paradisaea S. vittata S. virgata S. forsteri S. trudeaui S. albifrons S. saundersi S. antillarum S. superciliaris S. lorata S. nereis S. balaenarum S. repressa S. acuticauda S. aleutica S. lunata S. anaethetus S. fuscata Noddies Anous Anous stolidus A. minutus A. tenuirostris Procelsterna Procelsterna cerulea P. albivatta Gygis Gygis alba Atypical black-capped terns Phaetusa Phaetusa simplex Gelochelidon Gelochelidon nilotica Hydroprogne Hydroprogne caspia Inca tern Larosterna Larosterna inca Marsh terns Chlidonias Chlidonias niger C. leucopterus C. hybridus Typical black-capped terns Sterna Sterna aurantia S. dougallii S. striata S. sumatrana S. hirundinacea S. vittata S. virgata S. paradisaea S. aleutica S. hirundo S. forsteri S. repressa S. acuticauda S. albostriatus S. trudeaui Small terns S. albifrons S. saundersi S. superciliaris S. nereis S. lorata S. balaenarum Brown-winged terns S. fuscata S. lunata S. anaethetus Crested terns Thalasseus Thalasseus maximus T. bergii T. sandvicensis T. elegans T. bernsteini T. bengalensis Noddies Anous Anous stolidus A. minutus A. tenuirostris Procelsterna Procelsterna ceruleaa P. albivittab,a Gygis Gygis alba Gygis microrhynchaa,b Brown-winged terns Onychoprion Onychoprion fuscata O. lunata O. anaethetus O. aleutica Little terns Sternula Sternula albifrons S. antillarum S. superciliaris S. nereis S. lorataa S. saundersia S. balaenaruma Atypical black-capped terns Phaetusa Phaetusa simplex Gelochelidon Gelochelidon nilotica Hydroprogne Hydroprogne caspia Inca tern Larosterna Larosterna inca Marsh terns Chlidonias Chlidonias niger C. leucopterus C. hybrida C. albostriatus Typical black-capped terns Sterna Sterna dougallii S. striata S. sumatrana S. hirundinacea S. vittata S. paradisaea S. hirundo S. forsterib S. trudeauib S. acuticaudaa S. aurantiaa S. repressaa S. virgataa Crested terns Thalasseus Thalasseus maximus T. bergii (continued on next page)) 462 E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 Table 1 (continued) Moynihan (1959) Sibley and Monroe (1990) Gochfeld and Burger (1996) Suggested T. sandvicensis sandvicensis T. s. eurygnatha T. elegans T. bengalensis T. bernsteinia Informal groups are designated by non-italic type. a Not included in mtDNA tree; group membership is speculative. b Group membership only weakly supported by mtDNA tree; perhaps a crested tern. Fig. 1. Mitochondrial DNA phylogeny of the terns. Branch lengths indicate divergence times according to the scale below the tree. Posterior probabilities are listed only for branches with values less than 1. Dorsal views of the heads of all species in the tree are shown, and branch shading illustrates the evolutionary history of head plumage types. Text and brackets to the right of the Wgure indicate the recommended genus-level revision of the naming system (note that the genus Procelsterna is missing because we lacked tissue for it and could not include it in the phylogeny). A color version of this Wgure given in Appendix A. gene, and part of the 12S ribosomal subunit (12S) were ampliWed by polymerase chain reaction (PCR; Saiki et al., 1988). Primers used in association with each mtDNA region were as follows: ND2: L5215 (Hackett, 1996), H1064 (Drovetski et al., 2004), metL (5⬘-AAGCTAT CGGGCCCATACCCG-3⬘; O. Haddrath, unpublished), 463 E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 Table 2 Museum specimens used in this study Specimen VeriWcation Ring-billed Gull (Larus delawarensis); Washington, USA; BMNH [5027] Black Noddy (Anous minutus); Hawaii, USA; UMMZ 233348 [T-748] Brown Noddy (A. stolidus); Hawaii, USA; BMNH 44974 [X8367] Lesser Noddy (A. tenuirostris); Ascension Island; ZMUC 113341 [C1067] White tern (Gygis alba); Hawaii, USA; LSUMNS B-35109 [DLD7544] Large-billed tern (Phaetusa simplex); Rio Grande do Sul, Brazil; ROM [L50140] Inca tern (Larosterna inca); Captive; UMMZ 234198 [T-971] Black tern (Chlidonias niger); Kargopol’skiy rayon, Russia; BMNH 44291 [AWJ063] Black-fronted tern (C. albostriatus); New Zealand; ROM [BFT001] Whiskered tern (C. hybridus); W. Australia; ROM [AJD6149] White-winged tern (C. leucopterus); W. Australia; ROM [SSB030] Aleutian tern (Sterna aleutica); Alaska, USA; BMNH 42083 [JK9404] Amazon tern (S. supercilliaris); Para, Brazil; ROM [G29502] Antarctic tern (S. vittata); Antarctica; LSUMNS B-9899 Arctic tern (S. paradisaea); Mezenskiy rayon, Russia; BMNH 44530 [AWJ101] Black-naped tern (S. sumatrana); Micronesia; FMNH 346067 Bridled tern (S. anaethetus); W. Australia; ROM [AJB5615] Caspian tern (S. caspia); Minnesota, USA; BMNH 42160 [JK9424] Cayenne tern (S. sandvicensis eurygnatha); Para, Brazil; ROM [G12327] Common tern (S. hirundo); Kargopol’skiy rayon, Russia; BMNH 44288 [AWJ060] Crested tern (S. bergii); New S. Wales; ROM [AJB5621] Elegant tern (S. elegans); Baja California Sur, Mexico; LSUMNS B-5788 Fairy tern (S. nereis); S. Australia; SAM ABTC2326 Forster’s tern (S. forsteri); Minnesota, USA; BMNH 44052 [X8202] Gray-backed tern (S. lunata); Hawaii, USA; BMNH 44973 [X8536] Gull-Billed tern (S. nilotica); Rio Negro, Argentina; ROM [G3] Least tern (S. antillarum); Louisiana, USA; LSUMNS B-8423 [DLD2137] Lesser crested tern (S. bengalensis); W. Australia; ROM [AJB6104] Little tern (S. albifrons); W. Australia; ROM [AJB6071] Roseate tern (S. dougallii); Massachusetts, USA; BMNH 44190 [X8594] Royal tern (S. maxima); Rio Grande do Sul, Brazil; ROM [NO7420] Sandwich tern (S. sandvicensis acuXavida); Louisiana, USA; LSUMNS B-8458 [SJH21] Sooty tern (S. fuscata); New S. Wales, Australia; ROM [AJB5625] South American tern (S. hirundinacea); Buenos Aires, Argentina; ROM [Sth001] Snowy-crowned tern (S. trudeaui); Rio Grande do Sul, Brazil; ROM [J14824] White-fronted tern (S. striata); New Zealand; ROM [WFT001] Voucher cyt b Voucher None 12S, cyt b, ND2 None 12S, cyt b, ND2 12S, cyt b, ND2 None 12S, cyt b, ND2 None Voucher 12S, cyt b, ND2 12S, cyt b, ND2 12S, cyt b, ND2, voucher 12S, cyt b, ND2 None 12S, cyt b, ND2, voucher 12S, cyt b, ND2 12S, cyt b, ND2, voucher 12S, cyt b, ND2 None 12S, cyt b voucher voucher None 12S, cyt b, ND2 12S, cyt b, ND2 12S, cyt b, ND2 voucher 12S, cyt b 12S, cyt b, ND2 12S, ND2, voucher None None 12S, cyt b, ND2 Common and scientiWc names from Sibley and Monroe (1990) are followed by the specimen’s locality, its museum of origin, and the accession and Weld numbers if available (Weld numbers are in brackets). Sequences or vouchers used to validate species identity are listed in the veriWcation column. Museum abbreviations are as follows: BMNH, University of Minnesota Bell Museum of Natural History; ROM, Royal Ontario Museum; UMMZ, University of Michigan Museum of Zoology; LSUMNS, Louisiana State University Museum of Natural Sciences; FMNH, Field Museum of Natural History; SAM, South Australian Museum; and ZMUC, Zoological Museum University of Copenhagen. and ASN (5⬘-GATCRAGGCCCATCTGTCTAG-3⬘; O. Haddrath, unpublished); 12S: L1537 (5⬘-CAATCTT GTGCCAGCCACCGCGG-3⬘; O. Haddrath, unpublished) and 12Send (5⬘-GTGCACCTTCCGGTACACT TACC-3⬘; O. Haddrath, unpublished); cyt b: B52 (5⬘-GNAAATCYCACCCNCTWCTHAAAAT-3⬘; O. Haddrath, unpublished) and B6 (T. Burt, pers. comm.). The thermal cycles used for PCR ampliWcation are described in Buehler and Baker (2003) and in Drovetski et al. (2004). PCR products were cleaned using a Qiaquick PCR PuriWcation Kit (Qiagen Valencia, CA). With the exception of two ND2 sequences, all sequencing was performed by the University of Minnesota Advanced Genetic Analysis Center on ABI 377 automated sequencers. ND2 sequences from S. trudeaui and S. bergii were pieced together from parts of the gene sequenced manually using a Thermo Sequenase Cycle Sequencing Kit (Amersham–Pharmacia Biotech, Amersham, UK), from partial sequences produced by the Advanced Genetic Analysis Center, and from sequence generated by a Licor 4200 long-read DNA sequencer at the Royal Ontario Museum. Sequences were edited and aligned using Sequencher v4.1.2 (Gene Codes, Ann Arbor, MI). We were able to align the 12S sequence without adding gaps for all but one taxon, Sterna anaethetus, which required a 1-bp gap. Examination of amino acid sequences conWrmed the mitochondrial origin of the protein-coding genes. The Wnal edited data set included 2800 bp for each taxon: 1050 bp for cyt b, 1041 bp for ND2, and 709 bp for 12S. There were few missing bases for most taxa with the exception of Anous minutus, for which we lacked over 500 bp from the cyt b gene. All sequences used for phylogenetic inference are deposited in GenBank (Accession Nos. AY631284–AY631391). 464 E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 2.2. Phylogenetic and evolutionary analyses We generated the Wnal mtDNA phylogeny using Bayesian inference with the program MrBayes v3.0 (Huelsenbeck and Ronquist, 2001). We chose this method of analysis because it allowed us to use a partitioned likelihood model wherein all parameter values were generated separately for each DNA region. For each partition, we speciWed a general time reversible model with empirical base frequencies and with rate variation among sites modeled as a gamma distribution. The Markov chain Monte Carlo (MCMC) search was run with four chains that were incrementally “heated” according to the default values of the program to ensure an adequate search of the tree space. The chain ran for 2,000,000 generations with trees sampled every 2000 generations. A graph of ¡log likelihood vs. generation (not shown) revealed that the ¡log likelihood leveled oV after approximately 30,000 generations; thus, we discarded trees from the Wrst 100,000 generations as a conservative “burn-in.” A phylogeny was constructed from the remaining 951 trees by compiling a majority-rule consensus tree in PAUP* v4.0b2a (SwoVord, 1999). Because the trees from every 2000th generation were a sample from the posterior distribution of most likely trees (Tierney, 1994), the probability of each node can be estimated based on the proportion of trees in the sample that support the node. In addition to the Bayesian phylogenetic analysis, we performed heuristic searches using maximum parsimony (MP) and maximum likelihood (ML) in PAUP to determine to the extent to which these diVerent methodologies agree in their resulting tree topologies, and we calculated ML bootstrap values for the Bayesian tree topology with 100 bootstrap replications. We consulted Gochfeld and Burger (1996) in scoring winter plumage characters associated with each species and subspecies. Head plumage characteristics were mapped onto the phylogeny using a simple parsimony model in the Mesquite software package (Maddison and Maddison, 2003). 2.3. Estimating divergence times A likelihood ratio test indicated that the full data set (with outgroup removed) did not adhere to a model of evolution with a molecular clock enforced (df D 33,
2 D 167.87, p < 0.001). Similarly, likelihood ratio tests examining each DNA region separately also showed that only sequence data from 12S appeared to be clocklike (12S: df D 33,
2 D 29.80, p D 0.63; ND2: df D 33,
2 D 217.12, p < 0.001; cyt b: df D 30,
2 D 43.66, p D 0.10). Thus, we generated divergence times by applying Sanderson’s (1997) non-parametric rate smoothing method in the program r8s v1.50 (Sanderson, 2002) to the likelihood branch lengths generated in the Bayesian analysis from the entire data set. We calibrated divergence times by assigning dates to two speciation events. The Wrst calibration point was the gull–tern split estimated by Paton et al. (2003) to have occurred 24.4 million years before present (MYBP). The second was the divergence of Chlidonias niger and C. leucopterus. Howard (1946) found C. niger fossils in Oregon dating to approximately 2 MYBP, suggesting that the latter date corresponds roughly with C. niger’s colonization of the Northwestern United States. Because all Chlidonias terns other than C. niger are restricted to the old world, we assigned a date of 2 MYBP as a minimum divergence of this species from its sister species, C. leucopterus. 3. Results 3.1. Tree topologies and branch support Bayesian analysis produced a generally well-supported tree with several distinct clades of species and with only three poorly supported nodes based on posterior probabilities (Fig. 1; a color version of this tree is given in the supplementary material available online, as described in Appendix A). The three weak nodes involved the positions of Phaetusa simplex, C. hybridus, and the clade formed by S. forsteri and S. trudeaui. Among the trees sampled in the Bayesian analysis, P. simplex occasionally formed a clade with S. caspia and S. nilotica or grouped as a distant sister species to the small terns (Sternula in Fig. 1). A similar situation describes the poor posterior probability support associated with the S. fosteri–S. trudeaui clade, which either grouped in a basal position with the crested terns (Thalaseus in Fig. 1), was positioned as a sister group to both crested terns and black-capped terns (Sterna in Fig. 1), or was grouped in a basal position with the blackcapped terns as shown in Fig. 1. The third branch with a low posterior probability is the one grouping C. niger, C. leucopterus, and C. albostriatus. The uncertainty of this node is due to the fact that C. albostriatus was the most basal of the Chlidonias terns in roughly one-third of the trees sampled from the MCMC chains, as opposed to C. hybridus being the most basal Chlidonias tern as shown in Fig. 1. Both of the most optimal ML and MP trees diVered from the Bayesian tree in that the S. forsteri–S. trudeaui clade was sister to the crested terns (Thalasseus in Fig 1) and the other black-capped terns (Sterna in Fig. 1). In addition the MP tree grouped L. inca with S. caspia and S. nilotica and placed P. simplex in a clade with the little terns (Sternula in Fig 1). ML bootstrap support for the Bayesian tree topology was poorer than the posterior probabilities (Fig. 2). However, we stress that because the ML model of evolution was not partitioned as was the Bayesian model, the bootstrapping analysis did not account for diVerent levels of homoplasy in each gene E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 465 taxa, pairwise sequence divergence ranged from 0.29% (S. sandvicensis vs. S. eurygnatha) to 16% (Anous minutus vs. S. supercilliaris; Table 3), which corresponded to divergence times of approximately 300,000 years ago and 15.7 MYBP, respectively, assuming that our divergence-time estimates are accurate. Based on the estimate of the gull–tern split at 2.4. MYBP, most of the speciation that gave rise to the current assemblage of tern species occurred within the last 10 million years. The mtDNA tree makes evident the presence of several highly divergent taxa, including Gygis alba, P. simplex, and L. inca, all of which split from their closest relatives more than 8 million years ago (Fig. 1) and bear several morphologically distinct features (e.g., almost entirely white plumage in G. alba, extremely large bill in P. simplex, and moustache ornament in L. inca). 4. Discussion 4.1. Taxonomic implications Fig. 2. Unsmoothed phylogram of the Bayesian tree topology showing ML distances (scale shown below the tree) and ML bootstrap support indices (number of supported nodes in 100 replications). Nodes that lack numbers had bootstrap scores of 100. region, making it less suitable for our data. Therefore, we emphasize the results from the Bayesian support indices over the ML bootstrapping. Many of the clades deWned 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 (Thallasseus 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. 3.2. Sequence divergence A matrix of percent sequence divergence is presented in Table 3. Sequence divergence between the terns and the gull outgroup averaged 12.8%. Among the ingroup ClassiWcation schemes for the terns range from the conservative revision by Moynihan (1959) that recognized only three genera to the recent classiWcation by Gochfeld and Burger (1996) with 10 genera (Table 1). The widely accepted checklist by Sibley and Monroe (1990) falls between these two extremes, recognizing seven genera among 45 species (Table 1). Our phylogeny indicates that all of these classiWcation schemes are Xawed because they include paraphyletic genera. The most general shortcoming is the failure to recognize the small terns (S. albifrons and allies) and the brownwinged terns (S. fuscata and allies) as groups distinct from the typical black-capped terns, which causes taxonomy to conXict with monophyletic groups. In developing a classiWcation scheme that corresponds with phylogenetic relationships, we see two possible naming systems. The Wrst, and more conservative, resembles Moynihan’s view of the terns, recognizing 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). Our alternative classiWcation scheme would modify that of Gochfeld and Burger (1996; see Table 1) 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 (Table 1; Fig. 1). There are no objective methods for choosing between these scenarios, as each allows for monophyletic genera. However, we favor the latter classiWcation scheme because it is more illustrative of the structure of the phylogeny and more informative regarding the ecology, plumage, and natural history of the species comprising each of the major clades. Thus, in addition to the genera 466 Table 3 Pairwise percent genetic divergences among the terns and a gull (Larus delawarensis) outgroup 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 13.5 13.1 15.1 13.1 12.2 11.9 12.1 13.1 13.8 13.5 14.1 13.7 13.2 13.9 12.8 13.1 13.6 13.4 13.1 13.0 12.6 12.8 12.9 13.3 12.8 14.3 12.8 13.0 12.9 13.0 13.4 13.1 13.6 13.9 12.9 13.3 14.4 14.9 13.6 14.1 14.2 14.0 13.1 13.9 14.8 13.8 13.8 13.1 13.1 13.0 13.6 14.2 14.1 13.6 13.7 13.9 14.2 13.6 13.4 14.6 14.3 13.0 13.3 13.6 12.9 13.5 12.9 12.8 13.1 12.9 13.8 12.7 13.1 12.9 12.6 13.3 12.9 13.0 13.6 12.6 12.5 13.0 13.3 12.8 13.2 13.5 13.4 14.6 14.9 16.0 14.6 15.0 15.5 15.5 14.4 15.3 15.9 15.3 14.8 14.3 14.0 14.2 14.7 15.5 14.7 15.3 15.1 15.2 15.3 14.5 15.0 15.1 14.5 7.8 9.1 11.0 7.7 8.2 7.8 7.9 7.9 8.9 11.4 7.9 11.2 7.5 7.3 11.2 8.1 8.0 8.3 8.5 7.6 8.0 10.8 10.9 10.8 10.6 10.2 7.7 9.1 10.7 7.7 8.2 7.1 7.1 7.5 8.2 11.5 7.0 10.4 7.3 7.2 10.0 7.4 7.3 8.1 8.7 6.8 7.0 10.6 10.5 10.2 10.9 9.6 7.5 8.9 10.6 7.8 8.1 7.4 7.3 7.6 7.9 11.1 7.4 10.4 7.1 6.7 10.0 7.3 7.3 8.1 8.5 6.8 7.3 10.5 10.3 10.4 10.7 10.3 7.3 8.4 10.6 7.3 7.7 7.0 7.1 7.2 8.1 11.1 6.9 10.5 6.7 6.7 10.0 7.2 7.1 8.0 8.4 6.7 7.0 10.4 10.6 10.2 10.5 9.7 13.4 13.5 13.9 13.4 13.8 13.1 13.0 12.9 12.7 13.7 12.7 12.6 13.3 13.3 12.0 12.9 12.8 13.6 12.6 13.0 12.8 13.7 12.1 12.5 13.7 13.2 14.4 13.9 13.1 12.8 12.9 13.9 8.8 13.3 13.2 11.8 11.7 11.9 12.2 11.1 15.3 8.9 8.7 8.1 8.3 12.7 14.5 12.8 12.1 11.8 12.0 14.4 4.3 4.0 2.1 14.2 4.5 4.5 14.4 4.9 15.0 9.0 9.6 11.3 9.0 9.4 8.5 8.4 8.9 8.6 12.0 8.1 10.9 8.5 8.9 10.8 8.6 8.1 9.2 8.5 8.2 8.3 11.3 10.6 10.5 11.4 11.0 10.7 10.8 11.3 10.1 10.7 10.6 10.7 10.5 10.2 11.7 10.5 11.0 10.6 10.3 10.3 10.0 10.5 10.6 9.5 10.6 10.3 11.0 10.6 10.6 11.2 11.3 11.5 5.2 10.7 11.2 11.1 11.1 10.9 10.3 3.3 10.7 11.4 10.8 10.8 10.5 10.7 11.1 11.0 10.5 10.8 10.7 5.0 10.9 11.1 11.3 11.3 11.7 10.6 11.1 10.3 10.4 11.1 10.1 11.7 10.1 7.4 11.0 10.6 7.3 10.7 10.5 11.4 9.7 10.0 10.1 11.5 7.2 11.1 11.1 11.6 11.0 11.3 10.6 10.6 10.8 10.3 11.9 10.3 3.9 10.9 10.3 6.2 10.5 10.7 11.3 9.8 10.2 10.2 11.9 11.5 11.7 3.8 11.5 11.9 11.3 11.2 11.2 10.7 5.3 11.0 11.9 11.2 11.3 11.8 11.0 11.1 11.4 11.2 11.0 11.1 6.3 8.3 10.8 6.3 6.6 2.4 2.4 6.2 8.2 11.6 1.2 10.3 6.2 5.7 10.2 5.8 2.3 6.6 8.3 1.9 6.3 8.5 4.8 6.4 6.3 7.9 9.6 8.3 8.2 4.9 10.9 11.0 11.4 11.0 11.0 6.2 8.1 3.6 6.4 6.1 6.5 8.6 3.3 7.1 6.4 2.6 8.6 7.1 1.0 6.0 2.7 8.6 7.1 1.0 5.8 6.1 8.2 4.7 6.3 5.9 8.1 7.1 9.6 8.7 8.1 11.6 11.3 11.7 11.7 11.1 1.7 7.9 6.7 2.4 5.8 10.3 10.0 11.6 10.7 10.8 6.1 8.5 4.8 6.3 6.1 5.5 8.2 4.6 6.0 5.8 10.0 9.7 11.3 10.2 10.2 5.4 8.1 6.5 5.8 2.5 8.7 6.9 6.5 8.8 8.0 7.8 Numbers in the left most column correspond to both species names and numbers in subsequent column headings. 11.2 3.7 1.1 11.6 6.6 12.2 8.8 1.8 6.5 6.6 4.8 4.6 11.4 11.0 7.8 7.9 11.4 8.1 12.0 9.9 7.7 8.3 8.4 8.3 7.9 11.4 11.3 11.1 11.1 11.7 10.6 5.1 10.6 11.1 11.2 11.2 11.7 11.4 10.8 4.3 4.6 11.0 6.5 11.6 9.0 4.6 6.5 6.5 3.5 11.3 4.5 4.6 11.5 6.8 11.8 9.4 4.6 7.1 7.3 10.3 6.2 6.6 10.7 2.6 11.9 8.4 6.4 0.3 10.3 6.1 6.5 10.7 2.6 11.8 8.5 6.3 10.8 4.0 1.6 11.1 6.3 11.7 8.7 10.3 8.6 8.6 10.4 8.2 11.3 11.6 11.4 11.7 12.1 11.5 10.0 5.7 6.5 10.3 6.6 10.6 11.3 10.7 3.7 10.4 8.0 E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 1 Larus delawarensis 2 Sterna vittata 3 S. trudeaui 4 S. supercilliaris 5 S. sumatrana 6 S. striata 7 S. s. eurygnatha 8 S. sandvicensis 9 S. paradisaea 10 S. nilotica 11 S. nereis 12 S. maxima 13 S. lunata 14 S. hirundinacea 15 S. hirundo 16 S, fuscata 17 S. forsteri 18 S. elegans 19 S. dougalii 20 S. caspia 21 S. bergii 22 S. bengalensis 23 S. antillarum 24 S. anaethetus 25 S aleutica 26 S. albifrons 27 Phaetusa simplex 28 Larosterna inca 29 Gygis alba 30 Chlidonias niger 31 C. leucopterus 32 C. hybridus 33 S. albostriatus 34 Anous tenuirostris 35 A. stolidus 36 A. minutus E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 used in Gochfeld and Burger (1996), we suggest resurrecting the genera Onchycoprion, which Wagler (1832) created in his synonymous description of S. fuscata (see Coues, 1897), and Sternula, which Gould (1843) generated in the original description of S. nereis, to distinguish the brown-winged clade and the small terns, respectively (Table 1, Fig. 1). Designation of several monospeciWc 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. We are unable to oVer empirically based taxonomic recommendations regarding Procelsterna because no tissues were available to us, but considering its distinctive plumage, we suspect that it should retain its own generic status. The mtDNA phylogeny resolves several disputed aspects of tern taxonomy. For instance, recent considerations of S. dougallii have noted that this species bears similarities to both the crested terns and the typical terns, and were unable to assign this species to either of these groups (Gochfeld and Burger, 1996, Gochfeld et al., 1998). The mtDNA phylogeny places S. dougallii squarely among the typical terns. Similarly, the classiWcation of C. albostriatus within either Chlidonias or Sterna has been subject to considerable confusion because like the other Chlidonias terns it has markedly dark plumage and disperses inland for breeding. However, C. albostriatus does not share the distinctive marsh nesting habits of the other Chlidonias species. Our results conWrm that this taxon belongs in Chlidonias and that its plumage reXects its systematic aYnities more strongly than does the absence of marsh nesting. We were unable to obtain tissue samples of two-other dark-plumaged terns, S. acuticauda and S. repressa, for our phylogenetic analysis, but the plumages and behavior of these birds are similar to those of other Chlidonias terns. Notably, aside from its dark plumage S. acuticauda is an inland nesting species associated primarily with freshwater habitats. Hence, it is possible that one or both of these species belong to the genus Chlidonias rather than Sterna. However, the dark plumages of S. acuticauda and S. repressa may also reXect Gloger’s rule in that these South Asian species live in warm and sunny environments. Thus, we refrain from recommending taxonomic shifts for S. acuticauda and S. repressa without further supporting evidence. Unfortunately, a number of other disputed issues regarding tern systematics remain unresolved. For example, although the mtDNA tree conWrms suspicions that S. trudeaui and S. forsteri are sister species (Gochfeld and Burger, 1996; McNicholl et al., 2001; Schnell, 1970b), it is unclear whether these species should be grouped with the crested terns (Thalasseus in Fig. 1) or the typical terns (Sterna in Fig. 1). The mtDNA phylogeny favors grouping S. trudeaui and S. forsteri as sister 467 to the typical terns. However, this determination is based on a node with a posterior probability of 0.56, with the remainder of the posterior distribution favoring placement of the S. trudeau–S. forsteri clade among the crested terns or as sister to both the crested and blackcapped terns. Based on their small size (compared to most of the crested terns), their temperate breeding ecology, and their lack of a distinct crest, S. trudeaui and S. forsteri are outwardly more similar to the typical terns than to the crested terns. Thus, most of the available evidence favors keeping S. trudeaui and S. forsteri as members of Sterna; however, further examination of the taxonomic position of these sister species is needed. Similarly, we cannot conclusively address the controversy regarding whether to designate S. sandvicensis sandvicensis and S. s. eurygnatha as diVerent 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 (see Hayes, 2004). 4.2. Divergence time uncertainties In calibrating our tree according to the 24.4 MYBP estimate of the tern–gull split by Paton et al. (2003), we observed a low rate of sequence divergence: roughly 0.5% per million years. In their phylogenetic study of the gulls, Crochet et al. (2000) calibrated sequence divergence against the DNA–DNA hybridization data of Sibley and Ahlquist (1990) and concluded that the gull–tern split occurred 13.5 MYBP. Using this estimate as a calibration gives a more typical divergence rate of roughly 1% per million years and indicates that the divergence times in Fig 1 are overestimates. However, there is also some indication that we have underestimated divergence times in that Olson and Rasmussen (2001) describe what is probably an early Pliocene (3.7–4.8 MYBP) bone fragment from Northeastern United States, which closely resembles modern S. maxima. This specimen could represent an ancestral member of the crested tern group that preceded several speciation events during the last 5 million years as indicated by our divergence estimates. Alternatively, this fossil could indicate that S. maxima was present in North America long before our estimate of its origin at 1.0 MYBP, suggesting that our estimated divergence times are too recent. The incorporation of the earliest C. niger fossils in North America (Howard, 1946) as a calibration point may be seen as problematic because C. niger could have arisen in Eurasia long before its colonization of North America. However, omission of this calibration point dates the C. niger–C. leucotperus split at 1.6 MYBP, 468 E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 which is probably inaccurate because it follows the deposition of C. niger fossils in North America. Thus, inclusion of this calibration point serves to improve the accuracy of our dating scheme, although it almost certainly estimates a minimum divergence time for C. niger and C. leucotperus rather than an actual divergence time. We direct those wishing to examine the eVects of diVerent calibrations of our tree to Appendix A in the online version of this article, which contains a Supplementary data Wle that includes trees calibrated with and without the C. niger fossil data and according to the gull–ternsplit estimates of both 24.4 MYBP and 13.5 MYBP. 4.3. Correspondence between phylogeny and general morphology Examination of interspeciWc variation in the general appearance of terns in light of our mtDNA phylogeny reveals several interesting insights into plumage evolution in this group. In particular, the plumage on the forehead, crown, and nape of terns in breeding (alternate) plumage carries a strong phylogenetic signal. The most parsimonious reconstruction of ancestral head plumage patterns indicates that the ancestor to all terns probably had a mostly white head similar to many of the “whiteheaded” gull species (Crochet et al., 2000; Moynihan, 1959; Fig. 1). A black cap with a white blaze on the forehead is present in the brown-winged terns as well as the small terns, and it appears to be symplesiomorphic in these two groups (Fig. 1). The majority of the terns have a full black cap, with two notable exceptions. Both S. sumatrana and S. trudeaui stand out among their allies in that the black caps associated with their breeding plumages are much reduced—almost absent in S. trudeaui, which bears only an elongated black eye patch on an otherwise white head (Fig. 1). The breeding plumage of S. trudeaui bears a striking resemblance to the winter plumage of its sister taxon, S. forsteri. Similarly, the other nearly white-headed species, S. sumatrana, is sister to a clade formed by the roseate tern (S. dougallii) and the white-fronted tern (S. striata), and the unusual breeding plumage of S. sumatrana resembles that of its close relatives in two ways. First, the winter plumage of S. dougallii is very similar to the breeding plumage of S. sumatrana. Second, the black cap of S. striata does not extend anteriorly all the way to the base of the bill, such that it resembles an intermediate between S. sumatrana and S. dougallii (Fig. 1). Thus, although it is likely that both S. trudeaui and S. sumatrana replace their crown plumage as part of their partial pre-breeding molt, they appear to have retained the white-crown that characterizes the winter plumages of their relatives. However, as molt in these species is poorly documented, it is also possible that S. trudeaui and S. sumatrana forgo the prebreeding molt of their head plumage, which would also give rise to their unusual breeding plumages. Voelker (1996) suggested that the annual cycling between black caps (breeding plumage) and mottled or white head (winter plumage) in most terns is an adaptation associated with social signaling that allows nonbreeding wintering birds to avoid conXicts with breeding congeners resident on the non-breeder’s wintering areas. Following on this line of reasoning, the reduction of the black cap in S. trudeaui and S. sumatrana may be a characteristic retained from the winter or sub-adult plumages that was favored by evolution because of reduced aggression from black-headed congeners. Alternatively, the white heads may serve to improve recognition of conspeciWcs as both of these species nest in areas populated by black-capped species. Morphological and behavioral features have been key to prior classiWcations of the terns, and the high degree to which such prior classiWcations correspond with the topology of our tree demonstrates that many of these characters, particularly plumage patterns, generally agree with phylogenetic relationships inferred from mtDNA sequences. Although this study indicates that general plumage characteristics can provide good evidence of taxon relationships, many have concluded that plumage characteristics are too labile for use in avian systematics because of the numerous potential inXuences on the evolution of plumage coloration, such as sexual selection, species recognition, predator avoidance, and thermoregulatory considerations (reviewed in Omland and Lanyon, 2000). In their phylogenetic study of the gulls, Crochet et al. (2000) determined that the black cap in gulls has no value in determining species relationships. Furthermore, based on the prevalence of black caps in the terns, skimmers (Rynchopinae) and skuas (Stercorarinae), they speculated that the black cap represents a common ancestral state within gulls and the other charadriiform families. However the basal position of Anous and Gygis in our phylogeny contradicts this view with respect to terns, as the parsimony-based reconstruction of plumage states for terns indicates that a white crown represented the ancestral state. However, likelihoodbased reconstruction indicates a probability of 0.81 for the white-crown ancestral state. A forthcoming phylogeny of the gulls with improved resolution (Crochet, pers. comm.) may help resolve this issue and determine whether the black caps common among the charadriiform seabirds are an example of widespread convergent evolution. Acknowledgments We are greatly indebted to the museums listed in Table 2 that provided tissues for use in this study. In particular, we would like to thank D. Willard, R. BrumWeld, S. Donnellan, D. Dittmann, D. Mindell, and J. Hinshaw for administering and processing tissue loans. We also E.S. Bridge et al. / Molecular Phylogenetics and Evolution 35 (2005) 459–469 thank R. Zink for his support of this project, the use of his lab space, and for access to specimens in the Bell Museum ornithology collection. Much-needed technical advice and aid in DNA sequencing were provided by both O. Haddrath at the Royal Ontario Museum and M. Westberg at the Zink lab. Materials costs for this project were paid for by grants to E.S.B. from the Dayton-Wilkie Natural History Fund. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.ympev.2004.12.010. Nexus tree Wle containing unscaled branch lengths for the terns phylogeny as well as trees with smoothed branch lengths illustrating alternative divergence times based on diVerent calibration schemes. References American Ornithologist’s Union, 1998. Check-list of North American Birds. American Ornithologists’ Union, Lawrence, KS. Buehler, D.M., Baker, A.J., 2003. Characterization of the red knot (Calidris canutus) mitochondrial control region. Genome 46, 565–572. Chu, P.C., 1995. Phylogenetic reanalysis of Strauch’s osteological data set for the Charadriiformes. Condor 97, 174–196. Coues, E., 1897. Onychoprion not Haliplana. Auk 14, 314. Crochet, P.A., Bonhomme, F., Lebreton, J.D., 2000. Molecular phylogeny and plumage evolution in gulls (Larini). J. Evol. Biol. 13, 47–57. Drovetski, S.V., Zink, R.M., Fadeev, I.V., Nesterov, E.V., Koblik, E.A., Red’kin, Y.A., Rohwer, S., 2004. Mitochondrial phylogeny of Locustella and related genera. J. Avian Biol. 35, 105–110. 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. Gochfeld, M., Burger, J., Nisbet, I.C.T., 1998. Roseate Tern (Sterna dougallii). In: Poole, A., Gill, F. (Eds.), The Birds of North America, No. 370. The Birds of North America Inc, Philadelphia, PA, pp. 1– 32. Gould, P.J., 1843. Descriptions of thirty new species of birds from Australia. Proc. Zool. Soc. Lond., Part 10, No. 117, 131–164. Hackett, S.J., 1989. EVects of varied electrophoretic conditions on detection of evolutionary patterns in the Laridae. Condor 91, 73– 90. Hackett, S.J., 1996. Molecular phylogenetics and biogeography of tanagers in the genus Ramphocelus Aves. Mol. Phylogenet. Evol. 5, 368–382. Harrison, P., 1983. Seabirds: An IdentiWcation Guide. Houghton MiZin Company, Boston, MA. 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. 469 Higgins, P.J., Davies, S.J.J.F., 1996. Handbook of Australian, New Zealand & Antarctic Birds, Vol. 3. Snipe to Pigeons. Oxford University Press, Melbourne, Australia. Hillis, D.M., Mable, B.K., Larson, A., Davis, S.K., Zimmer, E.A., 1996. Nucleic acids IV: sequencing and cloning. In: Hillis, D.M., Moritz, C., Mable, B.K. (Eds.), Molecular Systematics, second ed. Sinauer Associates, Sunderland, MA, pp. 321–381. Howard, H., 1946. A review of the Pleistocene birds of Fossil Lake, Oregon. Contrib. Paleo. 8, Carnegie Inst., Wash. Publ. 551, pp. 141– 195. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. Junge, G.C.A., Voous, K.H., 1955. The distribution and relationship of Sterna eurygnatha Saunders. Ardea 43, 226–247. Maddison, W.P., Maddison, D.R., 2003. Mesquite: a modular system for evolutionary analysis Version 1.0. Available from: <http://mesquiteproject.org>. McNicholl, M.K., Lowther, P.E., Hall, J.A., 2001. Forster’s Tern (Sterna forsteri). In: Poole, A., Gill, F. (Eds.), The Birds of North America, No. 595. The Birds of North America Inc, Philadelphia, PA. Moynihan, M., 1959. A revision of the family Laridae. Am. Mus. Novit. 1928, 1–42. Olson, S., Rasmussen, P.C., 2001. Miocene and Pliocene birds from the Lee Creek Mine, North Carolina. Smithsonian Contrib. Paleobiol. 90, 233–365. Omland, K.E., Lanyon, S.M., 2000. Reconstructing plumage evolution in orioles (Icterus): repeated convergence and reversal patterns. Evolution 54, 2119–2133. Paton, T.A., Baker, A.J., Groth, J.G., Barrowclough, G.F., 2003. RAG1 sequences resolve phylogenetic relationships within Charadriiform birds. Mol. Phylogenet. Evol. 29, 268–278. Saiki, R.K., Gelfand, D.H., StoVel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B., Erlich, H.A., 1988. Primer-directed enzymatic ampliWcation of DNA with at thermostable DNA polymerase. Science 239, 487–491. Sanderson, M.J., 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Mol. Biol. Evol. 14, 1218–1231. Sanderson, M.J., 2002. Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach. Mol. Biol. Evol. 19, 101–109. Schnell, G.D., 1970a. A phenetic study of the suborder Lari (Aves). I. Methods and results of principal components analysis. Syst. Zool. 19, 35–57. Schnell, G.D., 1970b. A phenetic study of the suborder Lari (Aves) II. Phenograms. Syst. Zool. 19, 264–302. Sibley, C.G., Ahlquist, J.E., 1990. Phylogeny and ClassiWcation of Birds. Yale University Press, New Haven, CT. Sibley, C.G., Monroe, B.L.J., 1990. Distribution and Taxonomy of Birds of the World. Yale University Press, New Haven, CT. Strauch, J.R., 1978. The phylogeny of the Charadriiformes (Aves): a new estimate using the method of character compatibility analysis. Trans. Zool. Soc. Lond. 34, 263–345. SwoVord, D.L., 1999. PAUP*: Phylogenetic Analysis Using Parsimony (* and other methods), version 4.0b2a. Sinauer, Sunderland, MA. Tierney, L., 1994. Markov chains for exploring posterior distributions (with discussion). Ann. Stat. 22, 1701–1762. Voelker, G., 1996. An hypothesis for seasonal color change in the genus Sterna. J. Avian Biol. 27, 257–259. Wagler, J.G., 1832. Onychoprion serrata. Isis von Oken 25, 277.