- Email: [email protected]
Phylogenetic affinities of the Fregetta storm-petrels are not black and white
Molecular Phylogenetics and Evolution 97 (2016) 170–176
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Phylogenetic affinities of the Fregetta storm-petrels are not black and white q Bruce C. Robertson a,⇑, Brent M. Stephenson b, Robert A. Ronconi c, Sharyn J. Goldstien d, Lara Shepherd e,f, Alan Tennyson e, Nicholas Carlile g, Peter G. Ryan h a
Allan Wilson Centre for Molecular Ecology and Evolution, Department of Zoology, University of Otago, PO Box 56, Dunedin 9054, New Zealand Eco-Vista: Photography & Research Ltd, P.O. Box 8291, Havelock North 4157, New Zealand Department of Biology, Dalhousie University, Halifax, Nova Scotia, Canada d School of Biological Sciences, University of Canterbury, PB 4800, Christchurch 8140, New Zealand e Museum of New Zealand Te Papa Tongarewa, P.O. Box 467, Wellington, New Zealand f School of Biological Sciences, Victoria University of Wellington, Wellington, New Zealand g Office of Environment & Heritage, Dept. of Premier & Cabinet, P.O. Box 1967, Hurstville, NSW 2220, Australia h Percy FitzPatrick Institute of African Ornithology, DST-NRF Centre of Excellence, University of Cape Town, Rondebosch 7701, South Africa b c
a r t i c l e
i n f o
Article history: Received 15 April 2015 Revised 20 December 2015 Accepted 10 January 2016 Available online 20 January 2016 Keywords: Phylogeny Fregetta storm-petrels Cytochrome b gene b-fibrinogen gene Procellariiformes South Atlantic Ocean
a b s t r a c t The Fregetta storm-petrels generally are regarded to comprise two species: black-bellied storm-petrels F. tropica (monotypic) breed at Antarctic and sub-Antarctic islands (46–63°S), and white-bellied stormpetrels F. grallaria breed at south temperate islands (28–37°S), with four recognized subspecies. Confusion surrounds the status of birds at Gough Island (40°S), central South Atlantic, which have been attributed usually to a white-bellied form of black-bellied storm-petrel F. t. melanoleuca. We use cytochrome b and nuclear b-fibrinogen gene sequences to show that F. t. melanoleuca are present during the breeding season at Gough and islands in the nearby Tristan da Cunha archipelago (37°S), exhibiting limited divergence from F. t. tropica. We also show that there is greater diversity among F. grallaria populations, with eastern South Pacific F. g. segethi and F. g. titan differing by c. 0.011, and both differing from western South Pacific nominate F. g. grallaria by c. 0.059. The Tristan archipelago supports a population of F. grallaria closely allied to the nominate form, as well as a distinct form identified as F. g. leucogaster. Further research is needed to assess how F. grallaria and F. tropica segregate in sympatry at Tristan and Gough, and why this is the only location where both species have white-bellies. Ó 2016 Elsevier Inc. All rights reserved.
1. Introduction The storm-petrels of the Southern Hemisphere genus Fregetta Bonaparte (sub-family Oceanitinae) traditionally contains two species (e.g. Jouanin and Mougin, 1979;): the black-bellied stormpetrel, F. tropica (Gould, 1844) that breeds in the Antarctic and sub-Antarctic; and the temperate-breeding white-bellied stormpetrel, F. grallaria (Vieillot, 1818). The taxonomy of the Fregetta storm-petrels has been debated vigorously over the last 100 years (e.g. Salvadori, 1908; Mathews, 1933; Murphy and Snyder, 1952), with various sub-specific designations proposed and revised (Mathews, 1933; Murphy and Snyder, 1952), and confusion remains as to whether a white-bellied form of F. tropica exists on q
This paper was edited by the Associate Editor Edward Louis Braun.
⇑ Corresponding author.
E-mail address: [email protected] (B.C. Robertson). http://dx.doi.org/10.1016/j.ympev.2016.01.004 1055-7903/Ó 2016 Elsevier Inc. All rights reserved.
Gough Island (40°S, 10°W) in the mid-South Atlantic (Clancey, 1981; Brooke, 2004; Ryan, 2007). Robertson et al. (2011) showed using phylogenetic analyses that the New Zealand storm-petrel (Oceanites maorianus) is not an Oceanites, but is better placed with F. grallaria and F. tropica in the genus Fregetta, and suggested the name Fregetta maoriana. Recently, Cibois et al. (2015) clarified the identity of an unusual ‘pealea’ specimen (American Museum of Natural History 194110) collected off the Marquesas Islands, concluding that it was closely related to the F. grallaria clade but that its exact placement was uncertain. They also suggested that F. grallaria was not monophyletic, based on a short cytochrome b sequence from a Natural History Museum, London, specimen (1953.55.101) collected on Gough Island and assigned to F. g. leucogaster. Fregetta taxonomy has been based on various morphological characters (Marchant and Higgins, 1990; Stephenson et al., 2008) including chin and throat feather coloration (bases white in
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
F. tropica, not in F. grallaria; but see Stephenson et al., 2008); toe projection beyond the tail tip in flight (absent in F. grallaria; but see Howell, 2010); tarsal scale pattern (holothecal in F. tropica, scutellated in F. grallaria); morphometrics (bill, tarsus, toes longer in F. tropica; wings relatively longer in F. grallaria); and toe morphology and symmetry (Marchant and Higgins, 1990; Stephenson et al., 2008). Various combinations of these characters have been used in revisions of Fregetta taxonomy (Salvadori, 1908; Kinghorn and Cayley, 1922; Mathews, 1932a, 1933; Murphy and Snyder, 1952). However, many of the early taxonomic revisions were based on the examination of single specimens of unknown breeding provenance (Murphy, 1936; Murphy and Snyder, 1952). Recently several researchers have called for a revision of F. grallaria, suggesting that some subspecies might warrant full species status due to plumage and morphological differences (e.g. Stephenson et al., 2008; Howell, 2014). Currently, four subspecies are recognized for F. grallaria (nominate grallaria, titan, segethi and leucogaster) and one or two subspecies for F. tropica (nominate tropica and melanoleuca), depending on the status of the Gough population (cf. Marchant and Higgins, 1990; Brooke, 2004; Onley and Scofield, 2007; Dickinson and Remsen, 2013). F. g. grallaria breeds on the Lord Howe islands (32°S, 159°E) in the Tasman Sea, and the
171
Kermadec Islands (29°S, 178°W) north of New Zealand (Fig. 1). The nominate form is unusual in being polymorphic, with some blackbellied forms at Lord Howe islands (Fig. 2; Marchant and Higgins, 1990) and a less extensive degree of polymorphism at the Kermadec Islands (Tennyson and Taylor, 1990). F. g. titan of Rapa Island (28°S, 144°W) in the Austral Islands, French Polynesia, was defined on size, being larger than F. g. segethi from the Juan Fernandez Islands (34°S, 81°W) in the eastern South Pacific (Murphy, 1924, 1928). Finally, F. g. leucogaster breeds at the Tristan da Cunha group (37°S, 12°W) and perhaps Gough Island (Brooke, 2004) in the central South Atlantic, and a tiny relict population breeds at St Paul Island (39°S, 77°E) in the southern Indian Ocean (Tollu, 1984; Worthy and Jouventin, 1999). This form probably also bred at nearby Amsterdam Island (38°S, 77°E) prior to the arrival of rats and other introduced mammals (Worthy and Jouventin, 1999). F. t. tropica breeds in three regions further south than F. grallaria (Brooke, 2004): at subAntarctic islands south of New Zealand (Auckland, Antipodes and Campbell, 49–52°S) and in the southwest Indian Ocean (Kerguelen, Crozets and the Prince Edwards, 46–50°S), and at peri-Antarctic islands in the South Atlantic and Scotia Sea area (South Georgia, South Orkneys, South Shetlands, Bouvet and probably the South Sandwich Islands, 54–63°S).
Fig. 1. Sampling locations of the New Zealand storm-petrel Fregetta maoriana (pentagon), white-bellied storm-petrels F. grallaria (circle) and black-bellied storm-petrels F. tropica (square) used in the analysis. Open symbols are breeding locations not sampled in this study.
172
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
Fig. 2. Plumage variation of Fregetta storm-petrels, including the ‘‘white-bellied” form of F. tropica (F. t. melanoleuca) found in sympatry with F. grallaria (F. g. leucogaster) on Inaccessible Island, Tristan da Cunha (dorsal and ventral views of both specimens; both captured on the island’s plateau on 22 September 2011). Other specimens shown are F. maoriana (photo by BMS); F. t. tropica (Canterbury Museum, AV502, Fairchild’s Garden, Auckland Islands 1 February1929, photo by P. Scofield); F. g. segethi (Isla Santa Clara, 2003, photo by P. Scofield); and a dusky-bellied form of F. g. grallaria (Lord Howe group, photo by NC).
The Fregetta storm-petrels at Tristan da Cunha and Gough Island have caused the most taxonomic debate (Table S1). Whitebellied birds from these islands have been treated as either a full species, or more recently a subspecies of F. grallaria, a white-bellied F. tropica, or both. Gould (1844) named Thalassidroma leucogaster based on a specimen collected at sea mid-way between Tristan and the Cape of Good Hope. Salvadori (1908) named Fregetta melanoleuca based on a specimen apparently from Gough Island (Bourne in Palmer, 1962; Jouanin and Mougin, 1979). Additionally Mathews (1932b) suggested that a distinct grallaria subspecies existed at the Tristan archipelago and described Fregettornis grallaria tristanensis (he considered that the nesting grounds of leucogaster were unknown; Mathews (1933)). Soon afterwards Mathews (1936) considered that his new taxon should be more appropriately referred to Fregettornis grallaria aquerea (Kuhl, 1820), but a year later referred this taxon to the genus Fregetta (Mathews, 1937a, 1937b), and by 1943 had adopted the generic name Cymodroma (Mathews and Hallstrom, 1943). Murphy and Snyder (1952) proposed that the name leucogaster took precedence over tristanensis for ‘‘the birds from the area”, hence Fregetta g. leucogaster. This taxonomy has been adopted by most subsequent workers. However, Hagen (1952), citing Murphy (1936) and others, suggested that tristanensis was synonymous with melanoleuca, hence F. g. melanoleuca, which was perpetuated by Elliott (1957) and Swales (1965). However, Bourne considered melanoleuca to be a white-bellied form of F. tropica from Gough Island, hence F. t. melanoleuca (Palmer, 1962;
Jouanin and Mougin, 1979). While many recent workers have adopted Bourne’s conclusion, the relationships of these birds continues to be debated (e.g. Clancey, 1981; Carboneras, 1992; Bourne, 2000; Brooke, 2004). For example, Swales (1965), Fraser et al. (1988) and Ryan (2007) recognized only F. grallaria on Gough Island and the Tristan da Cunha group as a resident, but Fraser et al. (1988) noted that some of the birds on Inaccessible Island had elevated nostrils more characteristic of F. tropica. Based on an examination of skins in the Natural History Museum, London, Howell (2010) suggested that the small, long-legged birds on Gough were F. tropica whereas those from Tristan were F. grallaria. Given the considerable taxonomic uncertainty across the genus Fregetta, and claims that F. grallaria may not be monophyletic (Cibois et al., 2015), we address the phylogenetic affinities of the group using analysis of mitochondrial (cytochrome b) and nuclear (b-fibrinogen) genes from representatives of all sub-species using known provenance specimens. 2. Material and methods 2.1. Taxon sampling Blood and tissue samples were obtained from storm-petrels at breeding localities (Fig. 1; supplementary material Table S2). DNA was extracted following the methods of Robertson et al. (2011). We also incorporated published DNA sequences in our analyses (Accession numbers in Fig. 3 and Table S2).
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
173
Fig. 3. Bayesian inference tree of a (A) partial cytochrome b alignment (961 bp) and (B) an alignment of the seventh intron alignment of the b-fibrinogen gene (871 bp) of storm-petrels. The two numbers at each node denote Bayesian posterior probabilities (top) and Maximum Likelihood bootstrap values for 100 replicates (bottom): values <50% are not shown. Sample names refer to sample information or sequence accession numbers (Table S2, supplementary material).
2.2. Genetic analysis We obtained DNA sequence data for the mitochondrial cytochrome b gene (Sorenson et al., 1999) and the nuclear b-fibrinogen gene (Prychitko and Moore, 1997) following Robertson et al. (2011). We amplified 1149 base pairs (bp) of the cytochrome b gene using PCR primers L14675 and H16064 (Sorenson et al., 1999) and a c. 890 bp fragment of the b-fibrinogen gene containing the entire seventh intron was amplified with primers FIB-BI7U and FIB-BI7L (Prychitko and Moore, 1997) (see Robertson et al. (2011) for further details). PCR fragments were sequenced (cytochrome b: primers L14675, L254A and H16064; b-fibrinogen: FIB-BI7U and FIB-FIB-BI7L) on an ABI 3730xl automated sequencer following Robertson et al. (2011). Sequences were aligned in Geneious (6.1.8) (Biomatters Ltd) and all variable sites were confirmed by visual inspection of the chromatograms. Sequences were translated to detect stop codons and reading frame errors for comparison among partial-sequence data sets. To confirm no biases were present in the data, pairwise transition and transversion differences were calculated separately and plotted against pairwise distance to test for substitution saturation within the data, and Pattern Homogeneity (Disparity Index) with Monte-Carlo testing (Kumar and Gadagkar, 2001) was conducted in MEGA version 6.0.5 (Tamura et al., 2013) to confirm base composition consistency among taxa. Substitution saturation was not evident (Fig. S1) and pattern homogeneity tests showed no significant disparity in nucleotide composition, confirming sequence evolution across the storm-petrels to be homogenous and suitable for phylogenetic estimates (see also Fig. S2). 2.3. Phylogenetic analysis Phylogenetic analysis of alignments was performed using maximum likelihood (ML) and Bayesian methods with penguin species (Aptenodytes patagonicus, Accession No. AF076044 – cytochrome b; or Eudyptula minor AY695222 – b-fibrinogen) as outgroups (see
Nunn and Stanley, 1998; Kennedy and Page, 2002; Penhallurick and Wink, 2004). DNA sequences are accessioned in GenBank (KU558992–KU559009). For all datasets, we determined the best fit evolutionary model using Modeltest (Posada and Crandall, 1998) following Akaike’s Information Criterion (AIC) implemented in Geneious. For the cytochrome b alignment the evolutionary model GTR + G + I was used, while for the b-fibrinogen sequence a GTR model was used. Maximum likelihood analysis was performed with 100 bootstrap replicates using PAUP⁄4.0b10 (Swofford, 1998) in Geneious. The genetic distance among taxa was also determined using maximum likelihood parameters in Mega 6.0.5 (Tamura et al., 2013) following the recommendations of Fregin et al. (2012); the ML distances were compared to p-distance values and no evidence of overestimation was noted for the ML distances (Fig. S1). Bayesian analysis was conducted for three independent runs, sampling every 100 generations over 1 million generations and four chains with a 25% burn-in period in MrBayes (Ronquist and Huelsenbeck, 2003), using likelihood parameters determined by Modeltest: 4by4 nucleotide model with substitution type 6 (GTR) for all sequence alignments and among-site rate variation estimation for invariable sites and the gamma distribution. Stationarity was observed after 1 million generations. MCMC convergence was assessed by evaluating the standard deviation of the split frequencies and ESS (effective sample size) values in MrBayes and examining likelihood plots in Tracer v.1.5 (Rambaut and Drummond, 2007). 3. Results We used 961 bp of cytochrome b sequence and 871 bp of the nuclear b-fibrinogen gene to explore the affinities of Fregetta storm-petrels. We also used 557 bp of the cytochrome b to examine the placement of F. grallaria titan and the museum specimen AMNH 194110 investigated by (Cibois et al., 2015) using existing sequences (accession Nos. KP857579 and KP857580). For both
174
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
genes, all putative Fregetta storm-petrels formed a monophyletic clade distinct from other Oceanitinae taxa, which was distinct from the Hydrobatinae taxa, as noted by Robertson et al. (2011) (Figs. 3 and S3). As also noted in that study, there is support for three main clades within Fregetta, corresponding to F. tropica, F. grallaria and F. maoriana (Fig. 3) with each species showing 0.083–0.089 (range) sequence divergence from the others for the 961 bp of cytochrome b gene (Table S3). Relationships inferred using the nuclear b-fibrinogen gene were similar to cytochrome b, albeit with lower resolution (Fig. 3). Fregetta storm-petrels with white bellies sampled on islands of the Tristan da Cunha group included some individuals that clade with F. tropica and others that clade with F. grallaria (Figs. 3 and S3). There is thus a ‘‘white-bellied” F. tropica, presumably F. t. melanoleuca, which occurs not only at Gough Island, but also at the Tristan archipelago. However, there is no definitive evidence that F. grallaria occurs at Gough Island; sequence data from the British Museum specimen reported in Cibois et al. (2015) is identical to F. t. melanoleuca. Sequence divergence (range; 961 bp cytochrome b gene) between F. t. melanoleuca and the sub-Antarctic F. tropica (0.008–0.009) is minor compared to that between F. grallaria subspecies (Fig. 3, Table S3). For example, nominate F. grallaria from Lord Howe Island and F. g. segethi from the Juan Fernandez Islands diverge by 0.06–0.065 (Table S3). However, the difference between F. t. melanoleuca and nominate F. t. tropica (0.008) is greater than divergence within F. t. tropica from the New Zealand (Antipodes Island) and Indian Ocean (Marion Island) sub-Antarctic (0.002, Table S3). For the 961 bp cytb fragment, all F. grallaria storm-petrels were grouped in a well-supported clade that showed more divergence than the F. tropica clade (Fig. 3). Within F. grallaria there was good support for two subgroups (Fig. 3, Table S3). Nominate F. g. grallaria from Lord Howe grouped with F. grallaria from Tristan da Cunha, whereas taxa from the central and eastern South Pacific grouped together (Fig. 3). These placements are supported by strong sequence divergence between Juan Fernandez F. g. segethi and Lord Howe F. grallaria (0.06–0.065), and low divergence within groups (e.g. Lord Howe grallaria vs Tristan da Cunha grallaria: 0.002). Unexpectedly a second lineage of F. grallaria was observed among F. grallaria storm-petrels from the Tristan da Cunha group (Figs. 3 and S3). This taxon is distinct from the Lord Howe-like birds (0.028), being more divergent than F. g. titan is to F. g. segethi (0.009–0.013) (Table S4). Subsequent sampling in 2011 confirmed the presence of multiple individuals from both F. grallaria haplotypes on Inaccessible, both on the island plateau and at the coast. To be conservative, we consider the F. grallaria taxon on Inaccessible most divergent from the Lord Howe birds to be F. g. leucogaster (Gould, 1844), although the holotype of F. g. leucogaster needs reexamination. Based on the 557 bp cytochrome b sequence, F. g. titan and the ‘pealea’ specimen (AMNH 194110) are placed together in a clade with F. g. segethi (Fig. S3) and each lineage shows limited divergence from the other (0.009–0.013, Table S4). 4. Discussion Our phylogenetic analyses show that there are three main groupings among Fregetta storm-petrels, which are consistent with the traditional species designates of F. tropica and F. grallaria (e.g. Jouanin and Mougin, 1979;), and the recently proposed F. maoriana (Robertson et al., 2011). The apparent paraphyly among F. grallaria reported by Cibois et al. (2015) resulted from the misidentification of a F. t. melanoleuca specimen from Gough Island. As noted for other seabirds (e.g. Friesen et al., 2007), there was greater diversity among the tropically/subtropically distributed populations of F. grallaria compared with the more temperate/polar distribution of
F. tropica (Fig. 1). For example, F. t. tropica from the populations on the Antipodes and Marion Island (8600 km apart) showed cytochrome b sequence divergence of only 0.002, while F. t. melanoleuca from Inaccessible Island showed c. 0.01 divergence to F. t. tropica from Marion Island, 4200 km away (Table S3). On the other hand, among F. grallaria, birds from the central and eastern South Pacific (subspecies titan and segethi) differed by c. 0.011 (Table S4), which places them in the divergence range of subspecies if melanoleuca is accepted as a valid subspecies of tropica. However, nominate grallaria of the western South Pacific differed from the other two Pacific forms by c. 0.062, more typical of species-level differences in the storm-petrels (e.g. Robertson et al., 2011). We confirm the existence of the much-debated white-bellied form of F. tropica, not only on Gough Island, but also on the uninhabited islands of the Tristan da Cunha group (Fig. 1). Speculation on the existence of this white-bellied form of F. tropica has long caused confusion in the taxonomy of the Fregetta storm-petrels. For example, Clancey (1981, p. 195) commented on Bourne’s proposal that a white-bellied form of F. tropica breeds on Gough Island saying that ‘‘the occurrence of a breeding population of F. tropica on Gough Island alongside that of F. grallaria remains highly equivocal” and Carboneras (1992) considered that this white-bellied subspecies of F. tropica ‘‘may be invalid or may be referable to F. grallaria”. Measurements of skins (n = 4; Howell, 2010) and our sequence data (n = 3) suggest that only F. tropica occurs on Gough Island, however the small numbers of birds sampled mean that we cannot state conclusively that F. grallaria does not also breed on this island. We also found that the islands of the mid-Atlantic are home to two distinct lineages of F. grallaria. Until now it was thought that only F. g. leucogaster was present on the islands of the midAtlantic (Brooke, 2004). However, some adults sampled on Inaccessible Island (Tristan da Cunha group) were closely allied (0.002 difference) to the nominate form F. g. grallaria sampled as fledglings from Lord Howe Island (Fig. 1), whereas other Inaccessible birds, which we propose be treated as F. g. leucogaster, grouped more distantly with nominate F. g. grallaria (differing by 0.023). The existence of birds on the Tristan archipelago allied to the nominate F. g. grallaria on Lord Howe Island is surprising. Our finding is not a methodological artifact, as DNA sequences for the Lord Howe Island birds were generated independently in a separate facility (Victoria University of Wellington) to where all Tristan samples were sequenced (University of Otago). The most parsimonious explanation for the nominate race on Tristan is recent movement between these islands; nominate F. g. grallaria from the western South Pacific are very unlikely to be non-breeding visitors to Tristan. Lord Howe and the Kermadec Islands are more than 12,000 km from Tristan, and individuals were sampled on the Plateau of Inaccessible Island, well inland, during early summer (November) when birds start to display at their Pacific breeding islands (Marchant and Higgins, 1990). Little is known of the at-sea distribution of Fregetta (Marchant and Higgins, 1990; Shirihai, 2008; Spear and Ainley, 2007). F. maoriana has only been recorded from New Zealand waters (Gaskin and Baird, 2005), but similar birds have been sighted recently off the east coast of Australia and New Caledonia (Gaskin et al., 2011; Collins, 2013). Also, measurements suggest that the type specimen of nominate F. grallaria collected off Australia (Vieillot, 1818) is an eastern Pacific form from Juan Fernandez Island (F. g. segethi; see Murphy, 1924). Analysis of the holotypes of these taxa would resolve this point, but this is beyond the scope of the present study. Support for eastern Pacific F. grallaria dispersing into the western Pacific is found in the strong phylogenetic affinities of a beach-wreck F. grallaria collected on the Queensland coast (Fraser Island, Queensland Museum, O31230) grouping with
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
F. g. titan (cytochrome b; Robertson unpublished data), which breeds on Rapa Island in the central Pacific and was only known to occur in more eastern Pacific waters (Brooke, 2004; Spear and Ainley, 2007). Recently, the contentious pealea specimen AMNH 194110 was placed in F. grallaria allied with F. g. titan, but F. g. segethi was not included in that analysis (Cibois et al., 2015). Our findings suggest that this specimen may not be closely related to F. g. segethi, or any other currently recognized taxon. While further work is required on the phylogenetic affinities of AMNH 194110, our findings, along with those of Cibois et al. (2015), indicate that a fifth subspecies of F. grallaria should be recognized. AMNH 194110 is the holotype of Mathews’ (1933) Fregettornis guttata and we recommend that this name be reinstated as Fregetta grallaria guttata (Mathews, 1933), using our current understanding of the relationships of this taxon and the current terminology of this group. This taxon is represented by the sole specimen collected off the Marquesas Islands. Another ‘streaked’ Fregetta specimen, the holotype of Thalassidroma lineata Peale, 1848, has sometimes been linked with specimen AMNH 194110 (see Cibois et al., 2015), but Thalassidroma lineata Peale, 1848, is normally considered a junior synonym of Fregetta tropica (see Gill et al., 2010). Clearly more work is needed to understand the distribution, abundance and relationships among Fregetta taxa. In particular, further research is needed to assess how F. grallaria and F. tropica segregate in sympatry at Tristan, and why both forms have white-bellies at the only location where the two species cooccur. The presence of two taxa of F. grallaria at Tristan merely adds to the intrigue posed by these birds. Unfortunately, Fregetta storm petrels are becoming increasingly rare on Gough Island, apparently due to mouse predation (Cuthbert et al., 2013). The few samples collected since 2000 have all been F. t. melanoleuca. If, as seems likely (Marchant and Higgins, 1990), F. grallaria breeds later than F. t. melanoleuca, any F. grallaria that might breed on Gough are more likely to be targeted by mice, which become increasingly serious predators of seabird chicks as other food sources dwindle in autumn (Cuthbert et al., 2013). Acknowledgments We thank Gary Graves, Martim Melo, Graham Parker, Robert Prys-Jones, Kalinka Rexer-Huber, Sarah Jacob, Fiona McDuie, Paul Scofield, Martin Stervander and Ross Wanless for providing samples, sample information or assistance in the field; Fiona Robertson and Tasman Gillies for assistance in the laboratory; Fiona Robertson for assistance with manuscript preparation; and Edward Braun and one anonymous referee for helpful discussions and comments. BCR acknowledges financial support from University of Otago (PBRF funding). Sampling at Tristan da Cunha and Gough Island was conducted with the permission of the Tristan Administrator and Island Council, through Tristan’s Conservation Department; logistic support was provided by the South African Department of Environmental Affairs and Ovenstones Fishing. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ympev.2016.01. 004. References Bourne, W.R.P., 2000. The South Indo-Atlantic Fregatta storm-petrels. Sea Swallow 49, 54–56. Brooke, M., 2004. Albatrosses and Petrels across the World. Oxford University Press, Oxford.
175
Carboneras, C., 1992. Family Hydrobatidae (Storm-Petrels). In: del Hoyo, J., Elliott, A., Sargatal, J. (Eds.), Handbook of the Birds of the World, vol. 1. Lynx Edicions, Barcelona, pp. 258–271. Cibois, A., Thibault, J., LeCroy, M., Bretagnolle, V., 2015. Molecular analysis of a storm-petrel specimen from the Marquesas Islands, with comments on the lineata and guttata specimens. Bull. Br. Orn. Club 135, 240–246. Clancey, P.A., 1981. On birds from Gough Island, Central South Atlantic. Durban Mus. Novitat. 12, 187–200. Collins, C., 2013. Unravelling the mystery of the ‘‘New Caledonian storm-petrel”. Birding World 26, 256–259. Cuthbert, R.J., Louw, H., Lurling, J., Parker, G., Rexer-Huber, K., Sommer, E., Visser, P., Ryan, P.G., 2013. Low burrow occupancy and breeding success of burrowing petrels at Gough Island: a consequence of mouse predation. Bird Conserv. Int. 23, 113–124. Dickinson, E.C., Remsen, J.V., Jr. (Eds.), 2013. The Howard and Moore complete checklist of the birds of the world, . fourth ed.. Non-Passerines fourth ed., vol. 1 Aves Press, Eastbourne. Elliott, H.F.I., 1957. A contribution to the ornithology of the Tristan da Cunha Group. Ibis 99, 545–586. Fraser, M.W., Ryan, P.G., Watkins, B.P., 1988. The seabirds of Inaccessible Island, South Atlantic Ocean. Cormorant 16, 7–33. Fregin, S., Haase, M., Olsson, U., Alström, P., 2012. Pitfalls in comparisons of genetic distances: a case study of the avian family Acrocephalidae. Mol. Phylogenet. Evol. 62, 319–328. Friesen, V.L., Burg, T.M., McCoy, K.D., 2007. Mechanisms of population differentiation in seabirds. Mol. Ecol. 16, 1765–1785. Gaskin, C., Baird, K., 2005. Observations of black and white storm petrels in the Hauraki Gulf, November 2003-June 2005: were they of New Zealand storm petrels? Notornis 52, 181–194. Gaskin, C.P., Fitzgerald, N., Cameron, E.K., Heiss-Dunlop, S., 2011. Does the New Zealand storm petrel (Pealeornis maoriana) breed in northern New Zealand? Notornis 58, 104–112. Gill, B.J., Bell, B.D., Chambers, G.K., Medway, D.G., Palma, R.L., Scofield, R.P., Tennyson, A.J.D., Worthy, T.H., 2010. Checklist of the Birds in New Zealand, Norfolk and Macquarie Islands, and the Ross Dependency Antarctica, fourth ed. Te Papa Press in association with the Ornithological Society of New Zealand, Wellington. Gould, J., 1844. On the family Procellaridae, with descriptions of ten new species. Ann. Mag. Nat. History, London 13, 360–368. Hagen, Y., 1952. Birds of Tristan da Cunha. Res. Norweg. Sci. Exped. Tristan da Cunha 1937–38 20, 1–248, Oslo. Howell, S.N.G., 2010. Identification and taxonomy of white-bellied storm petrels, with comments on WP report in August 1986. Dutch Bird. 32, 36–42. Howell, S.N.G., 2014. Titan storm petrel. Dutch Bird. 36, 162–165. Jouanin, C., Mougin, J.-L., 1979. Order Procellariiformes. In: Mayr, E., Cottrell, G.W. (Eds.), . second ed., Check-List of Birds of the World second ed., vol. 1 Museum of Comparative Zoology, Cambridge, MA, pp. 48–121. Kinghorn, J.R., Cayley, N.W., 1922. On the status of several species belonging to the two genera, Fregetta Bp. and Fregettornis Mathews. Emu 22, 81–97. Kennedy, M., Page, R., 2002. Seabird supertrees: combining partial estimates of 262 procellariiform phylogeny. Auk 119, 88–108. Kumar, S., Gadagkar, S.R., 2001. Disparity index: a simple statistic to measure and test the homogeneity of substitution patterns between molecular sequences. Genetics 158, 1321–1327. Marchant, S., Higgins, P.J., 1990. Handbook of Australian, New Zealand and Antarctic Birds, vol. 1. Oxford University Press, Oxford. Mathews, G.M., 1932a. Pealeornis, gen. nov. Bull. Br. Orn. Club 52, 132. Mathews, G.M., 1932b. Fregettornis grallaria tristanensis, subsp. nov. Bull. Br. Orn. Club 52, 123–124. Mathews, G.M., 1933. On Fregetta Bonaparte and allied genera. Novitates Zool. 39, 34–54. Mathews, G.M., 1936. Remarks on procellarian and puffinine petrels. Emu 36, 91– 98. Mathews, G.M., 1937a. Key (dichotomous antithesis) of the storm-petrels. Emu 37, 136–143. Mathews, G.M., 1937b. Notes on storm-petrels, with a description of a new species and genus. Bull. Br. Orn. Club 57, 144–147. Mathews, G.M., Hallstrom, E.J.L., 1943. Notes on the Order Procellariiformes. Verity Hewitt Bookshop, Canberra. Murphy, R.C., 1924. Birds collected during the Whitney South Sea Expedition II. Am. Mus. Novitat. 124, 1–13. Murphy, R.C., 1928. Birds collected during the Whitney South Sea expedition. IV. Am. Mus. Novitat. 322, 1–5. Murphy, R.C., 1936. Oceanic birds of South America. The American Museum of Natural History, New York. Murphy, R.C., Snyder, J.P., 1952. The ‘‘Pealea” phenomenon and other notes on storm-petrels. Am. Mus. Novitat. 1596, 1–16. Nunn, G.B., Stanley, S.E., 1998. Body size effects and rates of cytochrome b evolution in tube-nosed seabirds. Mol. Biol. Evol. 15, 1360–1371. Onley, D., Scofield, P., 2007. Albatrosses, Petrels and Shearwaters of the World. Princeton University Press, Princeton. Palmer, R.S. (Ed.), 1962. Handbook of North American Birds. Loons through Flamingos, vol. 1. Yale University Press, New Haven. Penhallurick, J., Wink, M., 2004. Analysis of the taxonomy and nomenclature of the Procellariiformes based on complete nucleotide sequences of the mitochondrial cytochrome b gene. Emu 104, 125–147.
176
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14, 817–818. Prychitko, T.M., Moore, W.S., 1997. The utility of DNA sequences of an intron from the beta-fibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Picidae). Mol. Phylogenet. Evol. 8, 193–204. Rambaut, A., Drummond, A.J., 2007. Tracer v 1.5. Robertson, B.C., Stephenson, B.M., Goldstien, S.J., 2011. When rediscovery is not enough: taxonomic uncertainty hinders conservation of a critically endangered bird. Molec. Phylogenet. Evol. 61, 949–952. Ronquist, F., Huelsenbeck, J.P., 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572–1574. Ryan, P.G. (Ed.), 2007. A Field Guide to the Animals and Plants of Tristan da Cunha and Gough Island. Pisces Publications, Newbury. Salvadori, T., 1908. Description of an apparently new species of petrel (Fregetta). Bull. Br. Orn. Club 21, 78–80. Shirihai, H., 2008. The Complete Guide to Antarctic Wildlife, second ed. Princeton University Press, Princeton. Sorenson, M.D., Ast, J.C., Dimcheff, D.E., Yuri, T., Mindell, D.P., 1999. Primers for a PCR-based approach to mitochondrial genome sequencing in birds and other vertebrates. Mol. Phylogenet. Evol. 12, 105–114.
Spear, L.B., Ainley, D.G., 2007. Storm-petrels of the eastern Pacific Ocean: species assembly and diversity along marine habitat gradients. Ornithol. Monogr., vol. 62. The American Ornithologists’ Union, Washington, D.C.. Stephenson, B.M., Gaskin, C.P., Griffiths, R., Jamieson, H., Baird, K.A., Palma, R.L., Imber, M.J., 2008. The New Zealand storm-petrel (Pealeornis maoriana Mathews, 1932): first live capture and species assessment of an enigmatic seabird. Notornis 55, 191–206. Swales, M., 1965. The sea-birds of Gough Island. Ibis 107, 215–229. Swofford, D.L., 1998. PAUP⁄. Phylogenetic analysis using parsimony (and other methods). Version 4b10. Sinauer Associates, Sunderland, MA. Tamura, K., Stecher, G., Peterson, D., Filipski, A., Kumar, S., 2013. MEGA6: molecular evolutionary genetics analysis version 6.0. Mol. Biol. Evol. 30, 2725–2729. Tennyson, A., Taylor, G., 1990. Curtis Island. OSNZ News 57, 10. Tollu, B., 1984. La Quille (île Saint-Paul, Océan Indien), sanctuare de populations relicts. Oiseau 54, 79–85. Vieillot, L.J.P., 1818. Le Pétrel Échasse, Procellaria grallaria, Vieill. Nouveau Dictionnaire D’Histoire Naturelle. Nouv. ed, 25, 418. Worthy, T.H., Jouventin, P., 1999. The fossil avifauna of Amsterdam Island, Indian Ocean. Smithson. Contrib. Paleobiol. 89, 39–65.
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev
Phylogenetic affinities of the Fregetta storm-petrels are not black and white q Bruce C. Robertson a,⇑, Brent M. Stephenson b, Robert A. Ronconi c, Sharyn J. Goldstien d, Lara Shepherd e,f, Alan Tennyson e, Nicholas Carlile g, Peter G. Ryan h a
Allan Wilson Centre for Molecular Ecology and Evolution, Department of Zoology, University of Otago, PO Box 56, Dunedin 9054, New Zealand Eco-Vista: Photography & Research Ltd, P.O. Box 8291, Havelock North 4157, New Zealand Department of Biology, Dalhousie University, Halifax, Nova Scotia, Canada d School of Biological Sciences, University of Canterbury, PB 4800, Christchurch 8140, New Zealand e Museum of New Zealand Te Papa Tongarewa, P.O. Box 467, Wellington, New Zealand f School of Biological Sciences, Victoria University of Wellington, Wellington, New Zealand g Office of Environment & Heritage, Dept. of Premier & Cabinet, P.O. Box 1967, Hurstville, NSW 2220, Australia h Percy FitzPatrick Institute of African Ornithology, DST-NRF Centre of Excellence, University of Cape Town, Rondebosch 7701, South Africa b c
a r t i c l e
i n f o
Article history: Received 15 April 2015 Revised 20 December 2015 Accepted 10 January 2016 Available online 20 January 2016 Keywords: Phylogeny Fregetta storm-petrels Cytochrome b gene b-fibrinogen gene Procellariiformes South Atlantic Ocean
a b s t r a c t The Fregetta storm-petrels generally are regarded to comprise two species: black-bellied storm-petrels F. tropica (monotypic) breed at Antarctic and sub-Antarctic islands (46–63°S), and white-bellied stormpetrels F. grallaria breed at south temperate islands (28–37°S), with four recognized subspecies. Confusion surrounds the status of birds at Gough Island (40°S), central South Atlantic, which have been attributed usually to a white-bellied form of black-bellied storm-petrel F. t. melanoleuca. We use cytochrome b and nuclear b-fibrinogen gene sequences to show that F. t. melanoleuca are present during the breeding season at Gough and islands in the nearby Tristan da Cunha archipelago (37°S), exhibiting limited divergence from F. t. tropica. We also show that there is greater diversity among F. grallaria populations, with eastern South Pacific F. g. segethi and F. g. titan differing by c. 0.011, and both differing from western South Pacific nominate F. g. grallaria by c. 0.059. The Tristan archipelago supports a population of F. grallaria closely allied to the nominate form, as well as a distinct form identified as F. g. leucogaster. Further research is needed to assess how F. grallaria and F. tropica segregate in sympatry at Tristan and Gough, and why this is the only location where both species have white-bellies. Ó 2016 Elsevier Inc. All rights reserved.
1. Introduction The storm-petrels of the Southern Hemisphere genus Fregetta Bonaparte (sub-family Oceanitinae) traditionally contains two species (e.g. Jouanin and Mougin, 1979;): the black-bellied stormpetrel, F. tropica (Gould, 1844) that breeds in the Antarctic and sub-Antarctic; and the temperate-breeding white-bellied stormpetrel, F. grallaria (Vieillot, 1818). The taxonomy of the Fregetta storm-petrels has been debated vigorously over the last 100 years (e.g. Salvadori, 1908; Mathews, 1933; Murphy and Snyder, 1952), with various sub-specific designations proposed and revised (Mathews, 1933; Murphy and Snyder, 1952), and confusion remains as to whether a white-bellied form of F. tropica exists on q
This paper was edited by the Associate Editor Edward Louis Braun.
⇑ Corresponding author.
E-mail address: [email protected] (B.C. Robertson). http://dx.doi.org/10.1016/j.ympev.2016.01.004 1055-7903/Ó 2016 Elsevier Inc. All rights reserved.
Gough Island (40°S, 10°W) in the mid-South Atlantic (Clancey, 1981; Brooke, 2004; Ryan, 2007). Robertson et al. (2011) showed using phylogenetic analyses that the New Zealand storm-petrel (Oceanites maorianus) is not an Oceanites, but is better placed with F. grallaria and F. tropica in the genus Fregetta, and suggested the name Fregetta maoriana. Recently, Cibois et al. (2015) clarified the identity of an unusual ‘pealea’ specimen (American Museum of Natural History 194110) collected off the Marquesas Islands, concluding that it was closely related to the F. grallaria clade but that its exact placement was uncertain. They also suggested that F. grallaria was not monophyletic, based on a short cytochrome b sequence from a Natural History Museum, London, specimen (1953.55.101) collected on Gough Island and assigned to F. g. leucogaster. Fregetta taxonomy has been based on various morphological characters (Marchant and Higgins, 1990; Stephenson et al., 2008) including chin and throat feather coloration (bases white in
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
F. tropica, not in F. grallaria; but see Stephenson et al., 2008); toe projection beyond the tail tip in flight (absent in F. grallaria; but see Howell, 2010); tarsal scale pattern (holothecal in F. tropica, scutellated in F. grallaria); morphometrics (bill, tarsus, toes longer in F. tropica; wings relatively longer in F. grallaria); and toe morphology and symmetry (Marchant and Higgins, 1990; Stephenson et al., 2008). Various combinations of these characters have been used in revisions of Fregetta taxonomy (Salvadori, 1908; Kinghorn and Cayley, 1922; Mathews, 1932a, 1933; Murphy and Snyder, 1952). However, many of the early taxonomic revisions were based on the examination of single specimens of unknown breeding provenance (Murphy, 1936; Murphy and Snyder, 1952). Recently several researchers have called for a revision of F. grallaria, suggesting that some subspecies might warrant full species status due to plumage and morphological differences (e.g. Stephenson et al., 2008; Howell, 2014). Currently, four subspecies are recognized for F. grallaria (nominate grallaria, titan, segethi and leucogaster) and one or two subspecies for F. tropica (nominate tropica and melanoleuca), depending on the status of the Gough population (cf. Marchant and Higgins, 1990; Brooke, 2004; Onley and Scofield, 2007; Dickinson and Remsen, 2013). F. g. grallaria breeds on the Lord Howe islands (32°S, 159°E) in the Tasman Sea, and the
171
Kermadec Islands (29°S, 178°W) north of New Zealand (Fig. 1). The nominate form is unusual in being polymorphic, with some blackbellied forms at Lord Howe islands (Fig. 2; Marchant and Higgins, 1990) and a less extensive degree of polymorphism at the Kermadec Islands (Tennyson and Taylor, 1990). F. g. titan of Rapa Island (28°S, 144°W) in the Austral Islands, French Polynesia, was defined on size, being larger than F. g. segethi from the Juan Fernandez Islands (34°S, 81°W) in the eastern South Pacific (Murphy, 1924, 1928). Finally, F. g. leucogaster breeds at the Tristan da Cunha group (37°S, 12°W) and perhaps Gough Island (Brooke, 2004) in the central South Atlantic, and a tiny relict population breeds at St Paul Island (39°S, 77°E) in the southern Indian Ocean (Tollu, 1984; Worthy and Jouventin, 1999). This form probably also bred at nearby Amsterdam Island (38°S, 77°E) prior to the arrival of rats and other introduced mammals (Worthy and Jouventin, 1999). F. t. tropica breeds in three regions further south than F. grallaria (Brooke, 2004): at subAntarctic islands south of New Zealand (Auckland, Antipodes and Campbell, 49–52°S) and in the southwest Indian Ocean (Kerguelen, Crozets and the Prince Edwards, 46–50°S), and at peri-Antarctic islands in the South Atlantic and Scotia Sea area (South Georgia, South Orkneys, South Shetlands, Bouvet and probably the South Sandwich Islands, 54–63°S).
Fig. 1. Sampling locations of the New Zealand storm-petrel Fregetta maoriana (pentagon), white-bellied storm-petrels F. grallaria (circle) and black-bellied storm-petrels F. tropica (square) used in the analysis. Open symbols are breeding locations not sampled in this study.
172
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
Fig. 2. Plumage variation of Fregetta storm-petrels, including the ‘‘white-bellied” form of F. tropica (F. t. melanoleuca) found in sympatry with F. grallaria (F. g. leucogaster) on Inaccessible Island, Tristan da Cunha (dorsal and ventral views of both specimens; both captured on the island’s plateau on 22 September 2011). Other specimens shown are F. maoriana (photo by BMS); F. t. tropica (Canterbury Museum, AV502, Fairchild’s Garden, Auckland Islands 1 February1929, photo by P. Scofield); F. g. segethi (Isla Santa Clara, 2003, photo by P. Scofield); and a dusky-bellied form of F. g. grallaria (Lord Howe group, photo by NC).
The Fregetta storm-petrels at Tristan da Cunha and Gough Island have caused the most taxonomic debate (Table S1). Whitebellied birds from these islands have been treated as either a full species, or more recently a subspecies of F. grallaria, a white-bellied F. tropica, or both. Gould (1844) named Thalassidroma leucogaster based on a specimen collected at sea mid-way between Tristan and the Cape of Good Hope. Salvadori (1908) named Fregetta melanoleuca based on a specimen apparently from Gough Island (Bourne in Palmer, 1962; Jouanin and Mougin, 1979). Additionally Mathews (1932b) suggested that a distinct grallaria subspecies existed at the Tristan archipelago and described Fregettornis grallaria tristanensis (he considered that the nesting grounds of leucogaster were unknown; Mathews (1933)). Soon afterwards Mathews (1936) considered that his new taxon should be more appropriately referred to Fregettornis grallaria aquerea (Kuhl, 1820), but a year later referred this taxon to the genus Fregetta (Mathews, 1937a, 1937b), and by 1943 had adopted the generic name Cymodroma (Mathews and Hallstrom, 1943). Murphy and Snyder (1952) proposed that the name leucogaster took precedence over tristanensis for ‘‘the birds from the area”, hence Fregetta g. leucogaster. This taxonomy has been adopted by most subsequent workers. However, Hagen (1952), citing Murphy (1936) and others, suggested that tristanensis was synonymous with melanoleuca, hence F. g. melanoleuca, which was perpetuated by Elliott (1957) and Swales (1965). However, Bourne considered melanoleuca to be a white-bellied form of F. tropica from Gough Island, hence F. t. melanoleuca (Palmer, 1962;
Jouanin and Mougin, 1979). While many recent workers have adopted Bourne’s conclusion, the relationships of these birds continues to be debated (e.g. Clancey, 1981; Carboneras, 1992; Bourne, 2000; Brooke, 2004). For example, Swales (1965), Fraser et al. (1988) and Ryan (2007) recognized only F. grallaria on Gough Island and the Tristan da Cunha group as a resident, but Fraser et al. (1988) noted that some of the birds on Inaccessible Island had elevated nostrils more characteristic of F. tropica. Based on an examination of skins in the Natural History Museum, London, Howell (2010) suggested that the small, long-legged birds on Gough were F. tropica whereas those from Tristan were F. grallaria. Given the considerable taxonomic uncertainty across the genus Fregetta, and claims that F. grallaria may not be monophyletic (Cibois et al., 2015), we address the phylogenetic affinities of the group using analysis of mitochondrial (cytochrome b) and nuclear (b-fibrinogen) genes from representatives of all sub-species using known provenance specimens. 2. Material and methods 2.1. Taxon sampling Blood and tissue samples were obtained from storm-petrels at breeding localities (Fig. 1; supplementary material Table S2). DNA was extracted following the methods of Robertson et al. (2011). We also incorporated published DNA sequences in our analyses (Accession numbers in Fig. 3 and Table S2).
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
173
Fig. 3. Bayesian inference tree of a (A) partial cytochrome b alignment (961 bp) and (B) an alignment of the seventh intron alignment of the b-fibrinogen gene (871 bp) of storm-petrels. The two numbers at each node denote Bayesian posterior probabilities (top) and Maximum Likelihood bootstrap values for 100 replicates (bottom): values <50% are not shown. Sample names refer to sample information or sequence accession numbers (Table S2, supplementary material).
2.2. Genetic analysis We obtained DNA sequence data for the mitochondrial cytochrome b gene (Sorenson et al., 1999) and the nuclear b-fibrinogen gene (Prychitko and Moore, 1997) following Robertson et al. (2011). We amplified 1149 base pairs (bp) of the cytochrome b gene using PCR primers L14675 and H16064 (Sorenson et al., 1999) and a c. 890 bp fragment of the b-fibrinogen gene containing the entire seventh intron was amplified with primers FIB-BI7U and FIB-BI7L (Prychitko and Moore, 1997) (see Robertson et al. (2011) for further details). PCR fragments were sequenced (cytochrome b: primers L14675, L254A and H16064; b-fibrinogen: FIB-BI7U and FIB-FIB-BI7L) on an ABI 3730xl automated sequencer following Robertson et al. (2011). Sequences were aligned in Geneious (6.1.8) (Biomatters Ltd) and all variable sites were confirmed by visual inspection of the chromatograms. Sequences were translated to detect stop codons and reading frame errors for comparison among partial-sequence data sets. To confirm no biases were present in the data, pairwise transition and transversion differences were calculated separately and plotted against pairwise distance to test for substitution saturation within the data, and Pattern Homogeneity (Disparity Index) with Monte-Carlo testing (Kumar and Gadagkar, 2001) was conducted in MEGA version 6.0.5 (Tamura et al., 2013) to confirm base composition consistency among taxa. Substitution saturation was not evident (Fig. S1) and pattern homogeneity tests showed no significant disparity in nucleotide composition, confirming sequence evolution across the storm-petrels to be homogenous and suitable for phylogenetic estimates (see also Fig. S2). 2.3. Phylogenetic analysis Phylogenetic analysis of alignments was performed using maximum likelihood (ML) and Bayesian methods with penguin species (Aptenodytes patagonicus, Accession No. AF076044 – cytochrome b; or Eudyptula minor AY695222 – b-fibrinogen) as outgroups (see
Nunn and Stanley, 1998; Kennedy and Page, 2002; Penhallurick and Wink, 2004). DNA sequences are accessioned in GenBank (KU558992–KU559009). For all datasets, we determined the best fit evolutionary model using Modeltest (Posada and Crandall, 1998) following Akaike’s Information Criterion (AIC) implemented in Geneious. For the cytochrome b alignment the evolutionary model GTR + G + I was used, while for the b-fibrinogen sequence a GTR model was used. Maximum likelihood analysis was performed with 100 bootstrap replicates using PAUP⁄4.0b10 (Swofford, 1998) in Geneious. The genetic distance among taxa was also determined using maximum likelihood parameters in Mega 6.0.5 (Tamura et al., 2013) following the recommendations of Fregin et al. (2012); the ML distances were compared to p-distance values and no evidence of overestimation was noted for the ML distances (Fig. S1). Bayesian analysis was conducted for three independent runs, sampling every 100 generations over 1 million generations and four chains with a 25% burn-in period in MrBayes (Ronquist and Huelsenbeck, 2003), using likelihood parameters determined by Modeltest: 4by4 nucleotide model with substitution type 6 (GTR) for all sequence alignments and among-site rate variation estimation for invariable sites and the gamma distribution. Stationarity was observed after 1 million generations. MCMC convergence was assessed by evaluating the standard deviation of the split frequencies and ESS (effective sample size) values in MrBayes and examining likelihood plots in Tracer v.1.5 (Rambaut and Drummond, 2007). 3. Results We used 961 bp of cytochrome b sequence and 871 bp of the nuclear b-fibrinogen gene to explore the affinities of Fregetta storm-petrels. We also used 557 bp of the cytochrome b to examine the placement of F. grallaria titan and the museum specimen AMNH 194110 investigated by (Cibois et al., 2015) using existing sequences (accession Nos. KP857579 and KP857580). For both
174
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
genes, all putative Fregetta storm-petrels formed a monophyletic clade distinct from other Oceanitinae taxa, which was distinct from the Hydrobatinae taxa, as noted by Robertson et al. (2011) (Figs. 3 and S3). As also noted in that study, there is support for three main clades within Fregetta, corresponding to F. tropica, F. grallaria and F. maoriana (Fig. 3) with each species showing 0.083–0.089 (range) sequence divergence from the others for the 961 bp of cytochrome b gene (Table S3). Relationships inferred using the nuclear b-fibrinogen gene were similar to cytochrome b, albeit with lower resolution (Fig. 3). Fregetta storm-petrels with white bellies sampled on islands of the Tristan da Cunha group included some individuals that clade with F. tropica and others that clade with F. grallaria (Figs. 3 and S3). There is thus a ‘‘white-bellied” F. tropica, presumably F. t. melanoleuca, which occurs not only at Gough Island, but also at the Tristan archipelago. However, there is no definitive evidence that F. grallaria occurs at Gough Island; sequence data from the British Museum specimen reported in Cibois et al. (2015) is identical to F. t. melanoleuca. Sequence divergence (range; 961 bp cytochrome b gene) between F. t. melanoleuca and the sub-Antarctic F. tropica (0.008–0.009) is minor compared to that between F. grallaria subspecies (Fig. 3, Table S3). For example, nominate F. grallaria from Lord Howe Island and F. g. segethi from the Juan Fernandez Islands diverge by 0.06–0.065 (Table S3). However, the difference between F. t. melanoleuca and nominate F. t. tropica (0.008) is greater than divergence within F. t. tropica from the New Zealand (Antipodes Island) and Indian Ocean (Marion Island) sub-Antarctic (0.002, Table S3). For the 961 bp cytb fragment, all F. grallaria storm-petrels were grouped in a well-supported clade that showed more divergence than the F. tropica clade (Fig. 3). Within F. grallaria there was good support for two subgroups (Fig. 3, Table S3). Nominate F. g. grallaria from Lord Howe grouped with F. grallaria from Tristan da Cunha, whereas taxa from the central and eastern South Pacific grouped together (Fig. 3). These placements are supported by strong sequence divergence between Juan Fernandez F. g. segethi and Lord Howe F. grallaria (0.06–0.065), and low divergence within groups (e.g. Lord Howe grallaria vs Tristan da Cunha grallaria: 0.002). Unexpectedly a second lineage of F. grallaria was observed among F. grallaria storm-petrels from the Tristan da Cunha group (Figs. 3 and S3). This taxon is distinct from the Lord Howe-like birds (0.028), being more divergent than F. g. titan is to F. g. segethi (0.009–0.013) (Table S4). Subsequent sampling in 2011 confirmed the presence of multiple individuals from both F. grallaria haplotypes on Inaccessible, both on the island plateau and at the coast. To be conservative, we consider the F. grallaria taxon on Inaccessible most divergent from the Lord Howe birds to be F. g. leucogaster (Gould, 1844), although the holotype of F. g. leucogaster needs reexamination. Based on the 557 bp cytochrome b sequence, F. g. titan and the ‘pealea’ specimen (AMNH 194110) are placed together in a clade with F. g. segethi (Fig. S3) and each lineage shows limited divergence from the other (0.009–0.013, Table S4). 4. Discussion Our phylogenetic analyses show that there are three main groupings among Fregetta storm-petrels, which are consistent with the traditional species designates of F. tropica and F. grallaria (e.g. Jouanin and Mougin, 1979;), and the recently proposed F. maoriana (Robertson et al., 2011). The apparent paraphyly among F. grallaria reported by Cibois et al. (2015) resulted from the misidentification of a F. t. melanoleuca specimen from Gough Island. As noted for other seabirds (e.g. Friesen et al., 2007), there was greater diversity among the tropically/subtropically distributed populations of F. grallaria compared with the more temperate/polar distribution of
F. tropica (Fig. 1). For example, F. t. tropica from the populations on the Antipodes and Marion Island (8600 km apart) showed cytochrome b sequence divergence of only 0.002, while F. t. melanoleuca from Inaccessible Island showed c. 0.01 divergence to F. t. tropica from Marion Island, 4200 km away (Table S3). On the other hand, among F. grallaria, birds from the central and eastern South Pacific (subspecies titan and segethi) differed by c. 0.011 (Table S4), which places them in the divergence range of subspecies if melanoleuca is accepted as a valid subspecies of tropica. However, nominate grallaria of the western South Pacific differed from the other two Pacific forms by c. 0.062, more typical of species-level differences in the storm-petrels (e.g. Robertson et al., 2011). We confirm the existence of the much-debated white-bellied form of F. tropica, not only on Gough Island, but also on the uninhabited islands of the Tristan da Cunha group (Fig. 1). Speculation on the existence of this white-bellied form of F. tropica has long caused confusion in the taxonomy of the Fregetta storm-petrels. For example, Clancey (1981, p. 195) commented on Bourne’s proposal that a white-bellied form of F. tropica breeds on Gough Island saying that ‘‘the occurrence of a breeding population of F. tropica on Gough Island alongside that of F. grallaria remains highly equivocal” and Carboneras (1992) considered that this white-bellied subspecies of F. tropica ‘‘may be invalid or may be referable to F. grallaria”. Measurements of skins (n = 4; Howell, 2010) and our sequence data (n = 3) suggest that only F. tropica occurs on Gough Island, however the small numbers of birds sampled mean that we cannot state conclusively that F. grallaria does not also breed on this island. We also found that the islands of the mid-Atlantic are home to two distinct lineages of F. grallaria. Until now it was thought that only F. g. leucogaster was present on the islands of the midAtlantic (Brooke, 2004). However, some adults sampled on Inaccessible Island (Tristan da Cunha group) were closely allied (0.002 difference) to the nominate form F. g. grallaria sampled as fledglings from Lord Howe Island (Fig. 1), whereas other Inaccessible birds, which we propose be treated as F. g. leucogaster, grouped more distantly with nominate F. g. grallaria (differing by 0.023). The existence of birds on the Tristan archipelago allied to the nominate F. g. grallaria on Lord Howe Island is surprising. Our finding is not a methodological artifact, as DNA sequences for the Lord Howe Island birds were generated independently in a separate facility (Victoria University of Wellington) to where all Tristan samples were sequenced (University of Otago). The most parsimonious explanation for the nominate race on Tristan is recent movement between these islands; nominate F. g. grallaria from the western South Pacific are very unlikely to be non-breeding visitors to Tristan. Lord Howe and the Kermadec Islands are more than 12,000 km from Tristan, and individuals were sampled on the Plateau of Inaccessible Island, well inland, during early summer (November) when birds start to display at their Pacific breeding islands (Marchant and Higgins, 1990). Little is known of the at-sea distribution of Fregetta (Marchant and Higgins, 1990; Shirihai, 2008; Spear and Ainley, 2007). F. maoriana has only been recorded from New Zealand waters (Gaskin and Baird, 2005), but similar birds have been sighted recently off the east coast of Australia and New Caledonia (Gaskin et al., 2011; Collins, 2013). Also, measurements suggest that the type specimen of nominate F. grallaria collected off Australia (Vieillot, 1818) is an eastern Pacific form from Juan Fernandez Island (F. g. segethi; see Murphy, 1924). Analysis of the holotypes of these taxa would resolve this point, but this is beyond the scope of the present study. Support for eastern Pacific F. grallaria dispersing into the western Pacific is found in the strong phylogenetic affinities of a beach-wreck F. grallaria collected on the Queensland coast (Fraser Island, Queensland Museum, O31230) grouping with
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
F. g. titan (cytochrome b; Robertson unpublished data), which breeds on Rapa Island in the central Pacific and was only known to occur in more eastern Pacific waters (Brooke, 2004; Spear and Ainley, 2007). Recently, the contentious pealea specimen AMNH 194110 was placed in F. grallaria allied with F. g. titan, but F. g. segethi was not included in that analysis (Cibois et al., 2015). Our findings suggest that this specimen may not be closely related to F. g. segethi, or any other currently recognized taxon. While further work is required on the phylogenetic affinities of AMNH 194110, our findings, along with those of Cibois et al. (2015), indicate that a fifth subspecies of F. grallaria should be recognized. AMNH 194110 is the holotype of Mathews’ (1933) Fregettornis guttata and we recommend that this name be reinstated as Fregetta grallaria guttata (Mathews, 1933), using our current understanding of the relationships of this taxon and the current terminology of this group. This taxon is represented by the sole specimen collected off the Marquesas Islands. Another ‘streaked’ Fregetta specimen, the holotype of Thalassidroma lineata Peale, 1848, has sometimes been linked with specimen AMNH 194110 (see Cibois et al., 2015), but Thalassidroma lineata Peale, 1848, is normally considered a junior synonym of Fregetta tropica (see Gill et al., 2010). Clearly more work is needed to understand the distribution, abundance and relationships among Fregetta taxa. In particular, further research is needed to assess how F. grallaria and F. tropica segregate in sympatry at Tristan, and why both forms have white-bellies at the only location where the two species cooccur. The presence of two taxa of F. grallaria at Tristan merely adds to the intrigue posed by these birds. Unfortunately, Fregetta storm petrels are becoming increasingly rare on Gough Island, apparently due to mouse predation (Cuthbert et al., 2013). The few samples collected since 2000 have all been F. t. melanoleuca. If, as seems likely (Marchant and Higgins, 1990), F. grallaria breeds later than F. t. melanoleuca, any F. grallaria that might breed on Gough are more likely to be targeted by mice, which become increasingly serious predators of seabird chicks as other food sources dwindle in autumn (Cuthbert et al., 2013). Acknowledgments We thank Gary Graves, Martim Melo, Graham Parker, Robert Prys-Jones, Kalinka Rexer-Huber, Sarah Jacob, Fiona McDuie, Paul Scofield, Martin Stervander and Ross Wanless for providing samples, sample information or assistance in the field; Fiona Robertson and Tasman Gillies for assistance in the laboratory; Fiona Robertson for assistance with manuscript preparation; and Edward Braun and one anonymous referee for helpful discussions and comments. BCR acknowledges financial support from University of Otago (PBRF funding). Sampling at Tristan da Cunha and Gough Island was conducted with the permission of the Tristan Administrator and Island Council, through Tristan’s Conservation Department; logistic support was provided by the South African Department of Environmental Affairs and Ovenstones Fishing. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ympev.2016.01. 004. References Bourne, W.R.P., 2000. The South Indo-Atlantic Fregatta storm-petrels. Sea Swallow 49, 54–56. Brooke, M., 2004. Albatrosses and Petrels across the World. Oxford University Press, Oxford.
175
Carboneras, C., 1992. Family Hydrobatidae (Storm-Petrels). In: del Hoyo, J., Elliott, A., Sargatal, J. (Eds.), Handbook of the Birds of the World, vol. 1. Lynx Edicions, Barcelona, pp. 258–271. Cibois, A., Thibault, J., LeCroy, M., Bretagnolle, V., 2015. Molecular analysis of a storm-petrel specimen from the Marquesas Islands, with comments on the lineata and guttata specimens. Bull. Br. Orn. Club 135, 240–246. Clancey, P.A., 1981. On birds from Gough Island, Central South Atlantic. Durban Mus. Novitat. 12, 187–200. Collins, C., 2013. Unravelling the mystery of the ‘‘New Caledonian storm-petrel”. Birding World 26, 256–259. Cuthbert, R.J., Louw, H., Lurling, J., Parker, G., Rexer-Huber, K., Sommer, E., Visser, P., Ryan, P.G., 2013. Low burrow occupancy and breeding success of burrowing petrels at Gough Island: a consequence of mouse predation. Bird Conserv. Int. 23, 113–124. Dickinson, E.C., Remsen, J.V., Jr. (Eds.), 2013. The Howard and Moore complete checklist of the birds of the world, . fourth ed.. Non-Passerines fourth ed., vol. 1 Aves Press, Eastbourne. Elliott, H.F.I., 1957. A contribution to the ornithology of the Tristan da Cunha Group. Ibis 99, 545–586. Fraser, M.W., Ryan, P.G., Watkins, B.P., 1988. The seabirds of Inaccessible Island, South Atlantic Ocean. Cormorant 16, 7–33. Fregin, S., Haase, M., Olsson, U., Alström, P., 2012. Pitfalls in comparisons of genetic distances: a case study of the avian family Acrocephalidae. Mol. Phylogenet. Evol. 62, 319–328. Friesen, V.L., Burg, T.M., McCoy, K.D., 2007. Mechanisms of population differentiation in seabirds. Mol. Ecol. 16, 1765–1785. Gaskin, C., Baird, K., 2005. Observations of black and white storm petrels in the Hauraki Gulf, November 2003-June 2005: were they of New Zealand storm petrels? Notornis 52, 181–194. Gaskin, C.P., Fitzgerald, N., Cameron, E.K., Heiss-Dunlop, S., 2011. Does the New Zealand storm petrel (Pealeornis maoriana) breed in northern New Zealand? Notornis 58, 104–112. Gill, B.J., Bell, B.D., Chambers, G.K., Medway, D.G., Palma, R.L., Scofield, R.P., Tennyson, A.J.D., Worthy, T.H., 2010. Checklist of the Birds in New Zealand, Norfolk and Macquarie Islands, and the Ross Dependency Antarctica, fourth ed. Te Papa Press in association with the Ornithological Society of New Zealand, Wellington. Gould, J., 1844. On the family Procellaridae, with descriptions of ten new species. Ann. Mag. Nat. History, London 13, 360–368. Hagen, Y., 1952. Birds of Tristan da Cunha. Res. Norweg. Sci. Exped. Tristan da Cunha 1937–38 20, 1–248, Oslo. Howell, S.N.G., 2010. Identification and taxonomy of white-bellied storm petrels, with comments on WP report in August 1986. Dutch Bird. 32, 36–42. Howell, S.N.G., 2014. Titan storm petrel. Dutch Bird. 36, 162–165. Jouanin, C., Mougin, J.-L., 1979. Order Procellariiformes. In: Mayr, E., Cottrell, G.W. (Eds.), . second ed., Check-List of Birds of the World second ed., vol. 1 Museum of Comparative Zoology, Cambridge, MA, pp. 48–121. Kinghorn, J.R., Cayley, N.W., 1922. On the status of several species belonging to the two genera, Fregetta Bp. and Fregettornis Mathews. Emu 22, 81–97. Kennedy, M., Page, R., 2002. Seabird supertrees: combining partial estimates of 262 procellariiform phylogeny. Auk 119, 88–108. Kumar, S., Gadagkar, S.R., 2001. Disparity index: a simple statistic to measure and test the homogeneity of substitution patterns between molecular sequences. Genetics 158, 1321–1327. Marchant, S., Higgins, P.J., 1990. Handbook of Australian, New Zealand and Antarctic Birds, vol. 1. Oxford University Press, Oxford. Mathews, G.M., 1932a. Pealeornis, gen. nov. Bull. Br. Orn. Club 52, 132. Mathews, G.M., 1932b. Fregettornis grallaria tristanensis, subsp. nov. Bull. Br. Orn. Club 52, 123–124. Mathews, G.M., 1933. On Fregetta Bonaparte and allied genera. Novitates Zool. 39, 34–54. Mathews, G.M., 1936. Remarks on procellarian and puffinine petrels. Emu 36, 91– 98. Mathews, G.M., 1937a. Key (dichotomous antithesis) of the storm-petrels. Emu 37, 136–143. Mathews, G.M., 1937b. Notes on storm-petrels, with a description of a new species and genus. Bull. Br. Orn. Club 57, 144–147. Mathews, G.M., Hallstrom, E.J.L., 1943. Notes on the Order Procellariiformes. Verity Hewitt Bookshop, Canberra. Murphy, R.C., 1924. Birds collected during the Whitney South Sea Expedition II. Am. Mus. Novitat. 124, 1–13. Murphy, R.C., 1928. Birds collected during the Whitney South Sea expedition. IV. Am. Mus. Novitat. 322, 1–5. Murphy, R.C., 1936. Oceanic birds of South America. The American Museum of Natural History, New York. Murphy, R.C., Snyder, J.P., 1952. The ‘‘Pealea” phenomenon and other notes on storm-petrels. Am. Mus. Novitat. 1596, 1–16. Nunn, G.B., Stanley, S.E., 1998. Body size effects and rates of cytochrome b evolution in tube-nosed seabirds. Mol. Biol. Evol. 15, 1360–1371. Onley, D., Scofield, P., 2007. Albatrosses, Petrels and Shearwaters of the World. Princeton University Press, Princeton. Palmer, R.S. (Ed.), 1962. Handbook of North American Birds. Loons through Flamingos, vol. 1. Yale University Press, New Haven. Penhallurick, J., Wink, M., 2004. Analysis of the taxonomy and nomenclature of the Procellariiformes based on complete nucleotide sequences of the mitochondrial cytochrome b gene. Emu 104, 125–147.
176
B.C. Robertson et al. / Molecular Phylogenetics and Evolution 97 (2016) 170–176
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14, 817–818. Prychitko, T.M., Moore, W.S., 1997. The utility of DNA sequences of an intron from the beta-fibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Picidae). Mol. Phylogenet. Evol. 8, 193–204. Rambaut, A., Drummond, A.J., 2007. Tracer v 1.5
Spear, L.B., Ainley, D.G., 2007. Storm-petrels of the eastern Pacific Ocean: species assembly and diversity along marine habitat gradients. Ornithol. Monogr., vol. 62. The American Ornithologists’ Union, Washington, D.C.. Stephenson, B.M., Gaskin, C.P., Griffiths, R., Jamieson, H., Baird, K.A., Palma, R.L., Imber, M.J., 2008. The New Zealand storm-petrel (Pealeornis maoriana Mathews, 1932): first live capture and species assessment of an enigmatic seabird. Notornis 55, 191–206. Swales, M., 1965. The sea-birds of Gough Island. Ibis 107, 215–229. Swofford, D.L., 1998. PAUP⁄. Phylogenetic analysis using parsimony (and other methods). Version 4b10. Sinauer Associates, Sunderland, MA. Tamura, K., Stecher, G., Peterson, D., Filipski, A., Kumar, S., 2013. MEGA6: molecular evolutionary genetics analysis version 6.0. Mol. Biol. Evol. 30, 2725–2729. Tennyson, A., Taylor, G., 1990. Curtis Island. OSNZ News 57, 10. Tollu, B., 1984. La Quille (île Saint-Paul, Océan Indien), sanctuare de populations relicts. Oiseau 54, 79–85. Vieillot, L.J.P., 1818. Le Pétrel Échasse, Procellaria grallaria, Vieill. Nouveau Dictionnaire D’Histoire Naturelle. Nouv. ed, 25, 418. Worthy, T.H., Jouventin, P., 1999. The fossil avifauna of Amsterdam Island, Indian Ocean. Smithson. Contrib. Paleobiol. 89, 39–65.