Proposal (1029) to South American Classification Committee

Revise the taxonomy of Oceanites species

Current SACC notes read:

2a. Jaramillo (2003) suggested that the subspecies exasperatus might be a separate species from nominate Oceanites oceanicus.

3. A new species, Oceanites pincoyae, “Pincoya Storm-Petrel”, has been described from Chilean waters (Harrison et al. 2013). Recognized by del Hoyo and Collar (2014). SACC proposal passed to recognize pincoyae.

3a. Called "White-vented Storm-Petrel" in Meyer de Schauensee (1970), Hilty and Brown (1986), Sibley and Monroe (1990), Schulenberg et al. (2007), and elsewhere. SACC proposal to change English name to White-vented Storm-Petrel did not pass.

Effect on SACC: This proposal would split galapagoensis from gracilis, in addition to exasperatus, and chilensis from oceanicus. Furthermore, it recognize the taxon Oceanites barrosi sp. nov. as a valid species. The genus Oceanites, currently with three species, is expanded to a total of seven species following Norambuena et al. (2024).

Background & New information: The current classification considers three species within Oceanites: Oceanites oceanicus, O. gracilis and O. pincoyae. The species O. oceanicus sensu lato is considered to comprise three subspecies (Clements et al. 2023): nominate O. o. oceanicus (Kuhl, 1820); O. o. exasperatus Mathews, 1912; and O. o. chilensis Murphy, 1936. The species O. gracilis sensu lato has two subspecies, nominate O. g. gracilis (Elliot, 1859) and O. g. galapagoensis Lowe, 1921, while O. pincoyae is a recently described monotypic species (Harrison et al. 2013; Remsen et al. 2023).

Previous studies of the systematics of the genus only included a partial representation of Oceanites. The first phylogeny of Procellariiformes based on mitochondrial DNA Cytb (Nunn & Stanley 1998) considered Oceanitidae as a subfamily of Hydrobatidae despite the evident paraphyly of these clades. This study only included samples of O. oceanicus (no subspecies specified), which was recovered as a sister to a clade including Pelagodroma marina (Latham, 1790), Garrodia nereis (Gould, 1841), Fregetta tropica (Gould, 1844), and F. grallaria (Vieillot, 1818). Later, Robertson et al. (2011) incorporated a sequence of O. o. exasperatus into a new phylogeny of Oceanitidae, which showed a close relationship with an unspecified taxon O. oceanicus based on Cytb (Nunn & Stanley 1998), and in a phylogeny based on the 7th intron of b-fibrinogen, O. o. exasperatus was sister to a clade that includes Fregetta and Pelagodroma (see Robertson et al. 2011). Robertson et al. (2011) also generated O. g. gracilis and O. g. galapagoensis sequences, but due to their short sequences (only 132 bp) these were not included in the phylogeny. Other studies have only shown the relationships of O. oceanicus (based on Cytb) with the other species of the family (e.g., Hackett et al. 2008; Cibois et al. 2015; Prum et al. 2015; Robertson et al. 2016; Reddy et al. 2017) or used O. gracilis (based on ND1) as an outgroup of a phylogeny looking at the evolution of Oceanodroma (Sausner et al. 2016).

Phylogenetic analysis:

According to a new phylogenetic hypothesis based on new sequence data of the mitochondrial gene Cytb and linear morphological measurements of all species and five subspecies-level taxa in Oceanites (Norambuena et al. 2024), the genus Oceanites is monophyletic and composed of three well-supported clades (posterior probability > 0.95): (1) chilensis; (2) exasperatus; (3) gracilis, pincoyae, and barrosi sp. nov.; and (4) oceanicus and galapagoensis.

Phylogenetic hypothesis of the genus Oceanites based on BEAST from Cytb gene. Numbers on nodes are posterior probability values from the Bayesian analysis. Outgroups are not shown.

This tree also shows that gracilis, galapagoensis, oceanicus, chilensis, pincoyae, and exasperatus are each monophyletic. In addition, the populations of Oceanites present in central Chile (barrosi sp. nov.) were sister to O. pincoyae but with high divergence (5%). Several of the taxa presently considered polytypic species are shown to be paraphyletic. The taxon chilensis appears as a basal clade to the other Oceanites species. Samples from the Andes of central Chile, form a clade together with pincoyae and are phylogenetically distant from the samples of chilensis from their main distribution (close to its Terra Typica - Cape Horn, Magallanes, Chile).

According to a time-calibrated tree (Norambuena et al. 2024), the split between Oceanites genera and the other genera in Oceanitidae is estimated at ~32.7 Mya (40.7–22.4 Mya; 95% HPD), and the oldest divergence within Oceanites (the split between O. chilensis and other Oceanites) dated to the late Oligocene, around c. 21.3 Mya (29.3–13.3 Mya; 95% HPD). The most recent split was between O. pincoyae and O. barrosi sp. nov. dated to the Late Miocene, around c. 6.7 Mya (10.7–2.6 Mya; 95% HPD). Phylogenetic and morphological data support O. pincoyae as a distinctive species-level taxon.

Morphological analysis:

For the five measurements: ‘wing length’, ‘tail length’, ‘tarsus length’, ‘mid-toe claw’, and ‘culmen’ the Bartlett’s test of sphericity and KMO measures were p<0.001 and >0.78, respectively. PCs 1 and 2 presented the highest eigenvalues (>0.8) and explained 75.6% of the total variation. PC1 correlated positively with ‘wing length’, ‘tail length’, ‘tarsus length’, and ‘mid-toe claw’ and can be interpreted as a component reflecting overall size; PC2 correlated positively with ‘culmen’. Scatterplots of PCs showed a gradual variation between Oceanites species in the PC1 axis, with only marked differences between gracilis and the oceanicus complex, and overlapping between pincoyae-chilensis and exasperatus, respectively (Fig. 4). PC2, or the culmen measures, do not allow for separation of the populations. The LDA based on PCA results resulted in a 77.3% correct classification of the assigned species.

Distribution of average scores between PC1 and PC2 axes of morphological variation between species/subspecies of Oceanites genera. Ellipses represent 75% of the variation.

Oceanites barrosi sp. nov.:

Based on the phylogenetic hypothesis and morphological analyses and following the general lineage species concept, we propose the recognition of the Oceanites population of central Chile as a new taxon following the main points presented in Norambuena et al. (2024; see extensive photographic material therein): (1) Oceanites barrosi wing is, on average, larger than in O. chilensis but smaller than in O. pincoyae. At the same time, its tail and tarsus measurements are smaller than in O. chilensis and larger than in O. pincoyae. Noticeably smaller than O. exasperatus and somewhat smaller than O. oceanicus mainly in wing and tail length. (2) Restricted white tips on the belly, never as extensive as in O. galapagoensis, O. gracilis, or O. pincoyae, but typically not dark-bellied like O. chilensis, O. oceanites, and O. exasperatus. (3) Bold double pale line on underwing due to pale tipping on underwing coverts. Underwings are dark in O. exasperatus and O. oceanicus, and pale tipping not as bold in O. chilensis. (4) Square-cut tail with conspicuous white, rectangularly shaped rump patch. (5) In-flight, protruding feet with yellow webs. (6) Well-differentiated genetically (sister species of pincoyae, both with 5% of genetic distance). (7) High Andean breeding distribution in central Andes of Chile above treeline (see Barros 2017).

Summary statistics of morphological data of each taxon within Oceanites. Data are presented as mean ± standard deviation.

Linear sequencing and English names:

This new phylogenetic hypothesis suggests a new linear sequencing within the genus Oceanites. Following the criteria of Remsen et al. (2023); this should be as follows:

Oceanites chilensis (Mathews 1934) – Fuegian Storm-Petrel

Oceanites exasperatus (Mathews 1912) – Antarctic Storm-Petrel

Oceanites gracilis (Elliot 1859) – Graceful Storm-Petrel

Oceanites pincoyae (Harrison et al. 2013) – Pincoya Storm-Petrel

Oceanites barrosi (Norambuena et al. 2024) – Andean Storm-Petrel – Golondrina de mar andina (Chilean name)

Oceanites galapagoensis (Lowe 1921) – Lava Storm-Petrel

Oceanites oceanicus (Kuhl 1820) – Subantarctic Storm-Petrel

Map of breeding distributions of Oceanites species in the New World. Note that barrosi and likely pincoyae are nesting inland in mountain habitats, and similarly gracilis is inland in the desert. The rest are more marine in their nesting sites.

Discussion:

The elephant in the room here is that all of these taxa vary only slightly in plumage, with pincoyae the most extreme. They vary in how much pale they have on the belly, how much pale on the underwing, and how bold the greater covert bar is on the upperwing. Conservative plumage is the norm in storm-petrels, and many Procellariformes, and this creates a huge problem for taxonomists and field observers. We are visually oriented, so we cannot “see” these species, these seabirds end up being quite a conundrum to us as taxonomists. But to back track a bit, consider that Leach’s (Hydrobates leucorhous) and European Storm-Petrel (Hydrobates pelagicus) are often difficult to separate in the field from Oceanites exasperatus, as they share nearly the same plumage pattern of brownish body, white rump, and a pale bar on the upperwing. The similarities are striking, yet these birds are in different *** families ***!! We need to stress the importance of this point to put into context the plumage similarities of taxa within Oceanites. Similarity in plumage is just the norm in storm-petrels.

With regards to the question of biological species, we do have ample evidence that banded Oceanites return to their colonies and are extremely philopatric (Bretagnolle 1989), as such geographic structuring of populations is expected. The birds themselves are known to recognize their sex and species by vocalizations. However, very little data is available regarding this point in the literature. Bretagnolle (1989) studied voice of birds from Adelie Land (exasperatus) and from Kerguelen Island (oceanicus). He found consistent and statistically significant differences in their voice. These two taxa differ primarily on size, and breeding latitude, but not in plumage. One is Antarctic, and one is subantarctic in breeding region. In addition, he informally notes that other recordings of oceanicus and exasperatus from additional locations, South Georgia, Crozet Is., and South Sandwich Islands also show geographic differences. He concludes that “Geographic variation in vocalizations thus corroborates and parallels the taxonomic conclusions established from morphological characteristics”. Thus, for two of our most cryptic species, oceanicus and exasperatus, we have solid evidence of vocal differences. Bretagnolle and Robisson (1991) created digitized voices to understand the parameters most important for species recognition in Wilson’s Storm-Petrel, although they worked in Adelie Land and Kerguelen, they considered these two populations the same species and as such worked to understand the underlying parameters in species recognition such as the duration of syllables vs. silence between syllables. They did not directly do playback tests to members of each population, which we treat as species. However, based on the significant differences in voice in parameters that they consider as isolating mechanisms they note “One may then expect the Kerguelen Islands birds to be recognized by the Antarctic birds.” Unfortunately, as multiple of our taxa do not have their nests known yet (galapagoensis, barrosi and pincoyae), and most others are extremely difficult to visit, voice is either not known, or has not been analyzed for remaining populations.

Zidat et al. (2017) may have unlocked a key to understanding reproductive isolation in cryptic Procellariformes. They conducted a study on Calonectris shearwaters, and not only looked at their morphology, and genes, but the chemical signal of their uropygial glands. They state, “We also found that chemical labels remain distinct in sympatry, suggesting their divergence is not purely due to environmental effects.” The conclusion here is something that many seabird researchers have postulated for some time, that many Procellariformes may smell their species. Other examples exist of Procellariformes recognizing their mates, and homing to their islands based on smell. We point this out, as it helps to explain why taxa that are well differentiated genetically, such as our Oceanites, or the Hydrobates castro complex for example, are so conservative in plumage. It is expected that the barriers to reproductive isolation will involve smell, and voice.

Ecological differences are worth pointing out. The main one is not only the latitude of breeding, and adjacent water types, warm, cool, cold or Antarctic which are ecologically radically different habitats, but also where they breed. The species oceanicus, exasperatus and chilensis are on islets, or isolated Antarctic coastline, but definitely strongly marine in where they breed. The species galapagoensis has no known nest yet, and they live in the well researched Galapagos archipelago. It is hypothesized that they are nesting on old lava flows upslope on the geologically newer islands (Isabela, Fernandina). The species gracilis is breeding mainly in the desert of Chile and likely Peru, some of them well inland. The new species barrosi is clearly breeding high up in the central Andes of Chile based on ample evidence (Barros 2017). Current searches for pincoyae have clarified that it is likely moving up slope along river valleys, so it is also inland and likely in the lower southern Andes. This explains the failure of finding them in islands in Reloncavi. Our evidence, and measurements show that exasperatus is the expected and thus far only known species that migrates to the northern hemisphere in the non-breeding season. The others are likely resident or short-distance migrants.

Conclusion:

Based on the phylogenetic hypothesis, and morphological analyses (Norambuena et al. 2024), we suggest elevating to species status the taxa galapagoensis, chilensis, and exasperatus, and consider the new taxon barrosi as a valid species, thus recognizing a total of seven species within the genus Oceanites. The cryptic nature of these well differentiated species is troubling. However, the species sort out ecologically based on oceanographic parameters, migratory behavior, as well as extreme differences in nesting localities. It would be inconceivable that an island nesting chilensis could magically learn to migrate to the High Andes like barrosi for example. The little vocal data available confirms a significant difference between oceanicus and exasperatus which are two of the most similar species in plumage.

1. We recommend a yes vote to accept barrosi as a new species.

2. We recommend a yes vote to elevate chilensis, exasperatus and galapagoensis as species.

3. We have recommended a suite of English Names, which perhaps may require a separate proposal.

References:

Barros, R. (2017) ¿Por qué aparecen golondrinas de mar en la cordillera de Chile central? La Chiricoca, 22, 4–18.

Bretagnolle, V. (1989). Calls of Wilson’s Storm Petrel: functions, individual and sexual recognitions, and geographic variation. Behaviour 111: 98–112.

Bretagnolle, V. and Robisson, P. (1991). Species-specific recognition in birds: an experimental investigation of Wilson’s Storm-Petrel (Procellariiformes, Hydrobatidae) by means of digitalized signals. Canadian Journal of Zoology. 69(6): 1669–1673.

Cibois, A., Thibault, J.-C., LeCroy, M., & Bretagnolle, V. (2015) Molecular analysis of a storm petrel specimen from the Marquesas Islands, with comments on specimens of Fregetta lineata and F. guttata. Bulletin of the British Ornithologists’ Club, 135, 240–246.

Clements, J.F., Rasmussen, P.C., Schulenberg, T.S., Iliff, M.J., Fredericks, T.A., Gerbracht, J.A., Lepage, D., Spencer, A., Billerman, S.M., Sullivan, B.L., & Wood, C.L. (2023) The eBird/Clements Checklist of Birds of the World: v2023.

Hackett, S.J., Kimball, R.T., Reddy, S., Bowie, R.C.K., Braun, E.L., Braun, M.J., Chojnowski, J.L., Cox, W.A., Han, K., Harshman, J., Huddleston, C.J., et al. (2008) A phylogenomic study of birds reveals their evolutionary history. Science, 320, 1763–1768.

Harrison, P., Sallaberry, M., Gaskin, C.P., Baird, K.A., Jaramillo, A., Metz, S.M., Pearman, M., O'Keeffe, M., Dowdall, J., Enright, S., Fahy, K., Gilligan, J., & Lillie, G. (2013) A new storm-petrel species from Chile. Auk, 130(1), 180–191.

Norambuena H.V., R. Barros, A. Jaramillo, F. Medrano, C. Gaskin, T. King, K. Baird & C.E. Hernández (2024) Resolving the conflictive phylogenetic relationships of Oceanites (Oceanitidae: Procellariiformes) with the description of a new species. Zootaxa

Nunn, G.B., & Stanley, S.E. (1998) Body size effects and rates of cytochrome b evolution in tube-nosed seabirds. Molecular Biology and Evolution, 15, 1360–1371.

Prum, R.O., Berv, J.S., Dornburg, A., Field, D.J., Townsend, J.P., Lemmon, E.M., & Lemmon, A.R. (2015) A comprehensive phylogeny of birds (Aves) using targeted next-generation DNA sequencing. Nature, 526, 569–573.

Reddy, S., Kimball, R.T., Pandey, A., Hosner, P.A., Braun, M.J., Hackett, S.J., Han, K., Harshman, J., Huddleston, C.J., Kingston, S., Marks, B.D., et al. (2017) Why do phylogenomic data sets yield conflicting trees? Data type influences the avian tree of life more than taxon sampling. Systematic Biology, 66, 857–879.

Remsen Jr., J.V., Areta, J.I., Bonaccorso, E., Claramunt, S., Del-Rio, G., Jaramillo, A., Lane, D.F., Robbins, M.B., Stiles, F.G., & Zimmer, K.J. (2023) A Classification of the Bird Species of South America. American Ornithological Society.

Robertson, B.C., Stephenson, B.M., & Goldstein, S.J. (2011) When rediscovery is not enough: taxonomic uncertainty hinders conservation of a critically endangered bird. Molecular Phylogenetics and Evolution, 61, 949–952.

Robertson, B.C., Stephenson, B.M., Ronconi, R.A., Goldstien, S.J., Shepherd, L., Tennyson, A., Carlile, N., & Ryan, P.G. (2016) Phylogenetic affinities of the Fregetta storm-petrels are not black and white. Molecular Phylogenetics and Evolution, 97, 170–176.

Sausner, J., Torres-Mura, J.C., Robertson, J., & Hertel, F. (2016) Ecomorphological differences in foraging and pattering behavior among storm-petrels in the eastern Pacific Ocean. Auk, 133, 397–414.

Zidat, T., G. Dell’Ariccia, M. Gabirot, P. Sourrouille, B. Buatois, A. Celerier, F. Bonadonna, P Crochet. 2017. Reproductive isolation maintains distinct genotypes, phenotypes and chemical signatures in mixed colonies of the two European Calonectris shearwaters (Procellariiformes: Procellariidae), Zoological Journal of the Linnean Society, Volume 181, Issue 3, Pages 711–726, https://doi.org/10.1093/zoolinnean/zlx002

Heraldo V. Norambuena, Rodrigo Barros, Fernando Medrano & Alvaro Jaramillo, July 2024

Note on voting for Remsen: Let’s divide the voting into the three parts in the Recommendations, i.e.

Part 1. Treatment of barrosi as a species

Part 2. Elevation of three subspecies to species rank

Part 3. English names

_________________________________________________________________________________________________________

Comments from Steve Howell: “Comments on Andean Storm Petrel Oceanites barrosi (Norambuena et al. 2024. Resolving the conflictive phylogenetic relationships of Oceanites (Oceanitidae: Procellariiformes) with the description of a new species. Zootaxa 5486(4):451–475)

“First off, I really like the English name for the proposed new taxon, and clearly there is more to Oceanites taxonomy than has met the human eye. The authors are to be commended for attempting to shed more light on this challenging group of cryptic taxa, whose traditional relationships have undoubtedly been muddled by the human visual preoccupation of using plumage patterns for taxonomy. For groups such as storm petrels, morphology (adaptations to oceanic habitats and feeding) plus vocalizations (for species recognition at night) are surely of more taxonomic import than plumage patterns. And given that these birds perceive the world through their nostrils as much as—or more than—in any other manner, it would be great to develop a way to measure olfactory factors.

“To paraphrase Storrs Olson, however, a more accurate title for this paper might have been “Towards a less imperfect understanding of relationships within the genus Oceanites,” and I remain unconvinced about the new interpretation of ‘species’ limits and whether another new taxon, let alone ‘species,’ warrants description.

“To some extent, though, this all depends on which species concept one follows. The authors state that they follow the “General Lineage Species Concept,” which—based on this paper—I imagine would result in 50,000+ (or 100,000+?) bird species (= lineages) worldwide. Not that this is wrong, it’s just a different reality from most current avian taxonomies; e.g., think of the 35+ “presumptive species” of Gray-breasted Wood Wren proposed by Cadena et al. (2019; Biological Journal of the Linnean Society 126:487–506), and so on and on...

“Beyond species concepts and philosophy, I have a few comments on a paper that apparently was not peer-reviewed (at least, no reviewers are thanked or listed in the acknowledgements). Some comments no doubt stem from my own ignorance of molecular methods, and for this I apologize in advance. In no specific sequence, some thoughts are:

“Sample Size

I don’t know whether using only mitochondrial DNA from a single gene, and with samples of only 1 or 2 individuals, reveals anything taxonomically meaningful, but my sense from reading a lot of literature (and NACC and SACC comments) is that gene trees (esp. using only mtDNA) are not necessarily reliable for inferring species-level relationships. Thus, it seems a little premature to propose a wholesale revision of species limits and describe a new species based on only one mitochondrial gene (which might prove to be biologically and taxonomically uninformative?), along with molecular samples (for 4 of the 7 populations) that comprise only 1 to 3 individuals, and with morphological data (for 3 of the 7 populations) from only 5 individuals. Whether it matters technically that the type specimen was not one of the molecular samples I have no idea; the assumption that it is the same taxon is reasonable.

“For meaningful morphometric analyses, appreciably larger samples should be used; moreover, Table S3 gives only mean and SD, rather than ranges, thus confusing statistical significance with biological significance. Consequently, all the fancy stats, principal component analyses, and figures for morphological differences should be viewed with caution—and might look quite different with useful sample sizes.

“In this regard, on page 459 the authors disingenuously state “Only three individuals of pincoyae were assigned to chilensis” in their morphological analysis, but it appears they had measurements for only 12 pincoyae (and only 5 chilensis!) which means that 25% were ‘misclassified” = a fairly high proportion. I’m also unsure what “natural wing length” is? Wing chord or flattened and straightened wing are the two usual metrics, and they can vary by a few mm between people measuring them and between live birds and old specimens.

“Taxon Identification

The authors state that “Genetic samples from O. g. galapagoensis, O. o. oceanicus, and O. o. exasperatus were obtained from previous studies and GenBank (Table 1).” This assumes specimens and samples were correctly identified, which is not always true for storm petrels, cf. fallacious claims made of a close relationship between Black and Markham’s Storm Petrels based on genetic analysis (Wallace et al. 2017, Molecular Phyl. & Evol. 107:39–47; cf. Howell & Zufelt 2019:344, Oceanic Birds of the World) when both samples were from Black Storm Petrels and the two taxa are, realistically, better placed in different genera! But let’s assume for now the Oceanites samples used here were correctly identified. But...

“I see a genuine problem (an interesting one, though) concerning the (nominate) taxon oceanicus. All (?) authors, including of this paper, state that oceanicus breeds on subantarctic [emphasis mine] islands—i.e., places like South Georgia, which has been designated by some as the type locality of oceanicus. However, the 5 specimens of “oceanicus” used in this paper are from the Falklands, which are not true subantarctic islands and are biogeographically more similar to Cape Horn and the Fuegian zone of southern Chile, cf. author’s Figure S1. Based on morphology, plumage, and biogeography, other authors (Howell & Zufelt 2019, op. cit., Howell in press, Marine Ornithology) have provisionally suggested Falkland-breeding Oceanites may even be referrable to a broadly defined “chilensis.”

“This is not to say, however, that Oceanites breeding in the Falklands aren’t oceanicus that travel to forage in cold subantarctic waters—oceanic birds are adapted to marine habitats and use the nearest suitable land for nesting—but using birds from South Georgia or another truly subantarctic location would have been better—indeed, such birds might prove different again from Falkland birds, who knows? For example, some Falkland seabirds are endemic taxa or potentially separate species, distinct from populations/taxa at both Cape Horn and South Georgia/true subantarctic islands, e.g., “Common Diving Petrels.” Moreover, oceanicus breeding on different subantarctic islands (e.g., South Georgia, Crozet, Kerguelen) vary in measurements—in some cases as much as taxa recognized in the present study—and with appreciably larger samples (Marchant & Higgins 1990; HANZAB vol. 1).

“Thus, I argue that true (nominate) oceanicus was not examined in this paper, and that the taxonomic ID of Falkland breeders remains to be elucidated. Also, samples from various true subantarctic islands should be included in any comprehensive study of the complex.

“Biogeography

Nice things the paper reveals include the ‘closeness’ of mainland-breeding “pincoyae” and “barrosi” and gracilis, which makes biogeographic sense, although (pending larger samples and more thorough analysis) it may be that “pincoyae” is simply a southern subspecies/variety of “barrosi,” one in which a ‘white morph’ is frequent (as with dark morphs being frequent in southern-breeding Leach’s Storm Petrels Hydrobates leucorhous). If that proves to be the case, the wide plumage variation in presumed “pincoyae” (inexplicably overlooked by Harrison et al. 2013, cf. Howell & Schmitt 2016, Dutch Birding 38(6):384–388) would be explained (cf. Flood et al. 2024, Marine Ornithology 52:165–171; Howell in press, op. cit.); and then barrosi would become a junior synonym of pincoyae... The English name Andean Storm Petrel for both taxa combined, however, would still be appropriate.

“Thus, we would have chilensis adapted to cold Fuegian waters, “pincoyae/barrosi” adapted to cooler southern Humboldt Current waters, and gracilis adapted to warmer northern Humboldt Current waters.

“I also find intriguing the suggestion of galapagoensis (n = 2) and (nominate) “oceanicus” (n = 5) being sister taxa. However, a) true oceanicus was not necessarily sampled, and b) this purported close relationship may prove spurious when more genes are sampled.

“On p. 463, the authors state: “The results support the validity of the taxon O. (oceanicus) exasperatus, and its plumage similarity with O. (oceanicus) oceanicus sustains the hypothesis that exasperatus is a sibling or cryptic species...” but sibling to what?

“Additionally on p. 463, the authors state “Given our phylogeny and that a sample from North America falls within the exasperatus clade, it is likely that only exasperatus migrates to the northern hemisphere while all other forms are resident in the southern hemisphere...” Is there any evidence that true (subantarctic) oceanicus (e.g., from South Georgia) remains in the southern hemisphere? Perhaps it does, but a single sample from North America is tenuous support for such a sweeping statement. Either way, this would be an interesting thing to establish, along with how the wing molt strategy of true oceanicus might differ from exasperatus if the latter is the only transequatorial migrant Oceanites (cf. Howell in press, op. cit.).

“Summary

Until a more thorough analysis of the Oceanites complex is conducted, with appropriate samples and potentially more informative molecular analysis (and ideally song samples, and playback—dream on!), this study is tantalizing but inconclusive.

“To my mind, recognition of “barrosi” and “pincoyae” as (biological) species still awaits critical and unbiased study; as noted by Howell & Schmitt (2016, op. cit.), most of the “unique” specific features attributed to “pincoyae” (behavior, foraging ecology, molt timing) are either spurious or uncertain; even plumage characters of presumed pincoyae may overlap completely with presumed chilensis (now = presumed “barrosi”?) (Flood et al. 2024, op. cit.).

“To those with historical knowledge of storm-petrel taxonomy and nomenclature, the saga of naming different populations of Oceanites in central and southern Chile, based on limited samples, brings to mind the naming of various different taxa of Leach’s Storm Petrels in western North America in the last century.

“Indeed, I wonder whether a study using the exact same methods and sample sizes as used by Norambuena et al. for Oceanites, but conducted on different populations of Pacific coast Leach’s Storm Petrels from Alaska to Mexico (not including Townsend’s and Ainley’s) might not result in one or more new ‘species’ of Leach’s being recognized? Again, who’s to say what’s ‘right’ or ‘wrong,’ although multiple species of Leach’s doesn’t accord with what we know of the biology (e.g., songs from Alaska to Mexico sound similar), despite a wide latitudinal range. Conversely, song playback tests for Band-rumped Storm Petrels reveal good morphological, ecological, and biological species that myopic molecular analyses failed to reveal; cf. Howell 2021; North American Birds 72(1):16–25).

“In conclusion, real-time biology and molecular sampling do not always align, and there is still much to learn about storm petrels!”

Comments from Areta:

“A) YES to O. barrosi. Although the type specimen from Río Blanco was not sequenced, the three birds sequenced came from close enough to the type locality (e.g., Río Colorado is some 30km from Río Blanco) as to make it sound very reasonable that the authors sampled the same breeding population. As far as we know, there are not two different taxa breeding in the same or nearby burrows, which seems to be the main risk of not having sequenced the type. Thus, it is a judgment call that I am willing to accept at this point, while hoping for this holotype to be sequenced sooner than later. The 5% divergence reported in Cyt-b is reasonably large in relation to intra-taxon variation (see for example the rather homogeneous genetics of the Antarctic exasperatus). In terms of divergence, barrosi sits on the lower end of differentiation among the species recognised by Norambuena et al. (2024), and by this yardstick (and in the absence of vocal data) there is room to discuss whether barrosi would better be considered a subspecies of pincoyae or a separate species. Dealing with cryptic species that have very similar plumages and morphology of course will cause headaches when trying to identify the taxa in the field. I wonder whether some of the questionings of Howell & Schmitt (2016; Dutch Birding, discussed here https://www.museum.lsu.edu/~Remsen/SACCprop721.htm) regarding the unexplained variation in individuals of Oceanites "oceanicus" (and their skepticism towards pincoyae as a separate species) could be explained by the existence of barrosi. There are some unknowns here that make the case less clearcut than the 3-way split discussed in B, however, the data collectively tips the scale towards recognition of barrosi as a species, and so I vote to accept it. It also remarkable that there is at least one specimen from Argentina (Las Cuevas, Mendoza) that needs proper assessment (Mark Pearman is working on this), but which may prove to belong to barrosi (see map of Oceanites oceanicus in Pearman and Areta 2020 where such a specimen has been indicated by a circle). Barros (2017) discussed and showed photographs of this specimen (http://www.lachiricoca.cl/wp-content/uploads/2018/03/CHIRICOCA-22-Golondrinas.pdf)".

“B) YES to the recognition of exasperatus, chilensis, and galapagoensis as separate species. The three have long been recognised as distinct taxa with drastically different breeding areas, and the genetic evidence in Norambuena et al. 2024 indicates deep splits. Although no doubt future genomic analyses will provide further clarifications, the mitochondrial data is enough for me to push these taxa to species level.

“C) Hmmm. Though I am not a voter, I think I prefer to call barrosi something like "Aconcagua Storm-Petrel". The Andes harbour several species of breeding storm-petrels, so it does not seem to be particularly informative to call it Andean Storm-Petrel. As for gracilis, the scientific name seems to refer to it being gracile, which is not the same as graceful (also: is there a need to change the name Elliot´s?) . Why "Lava" instead of the more prosaic "Galapagos"? Yes, there is a Lava Gull and a Lava Heron, but... “

Comments solicited from Hadoram Shirihai: “I saw your [Steve Howell] comments now on the SACC, with which I agree 100% on every point!

“I can add that in relation to point 6 below, between 2013 and 2019 I did quite an extensive survey, year round, of much of pelagic waters and islands off Chile, the Humboldt Current and beyond. Some of my works you know. The unknown large birds that I am reporting in point 6, between Nov and Mar, are from off c. Chile, including pelagic waters off Valparaiso. The vast majority are of such examples. Many of them have variable degrees of pale ventral area - BUT they are large and by far so, from the measurements given for the new species barrosi - see measurements in point 6! I include for you screenshots from the pages of one of the unpublished reports of the work. So, based on the data they give in the article for barrosi sp. nov. (the Andean Storm-Petrel), such birds as the latter from the Andean Mts east of Santiago, same season, by and large do NOT exist in the ocean below. They do NOT!!! Or their measurements in the paper are completely wrong.

“Amazingly, we think the same also about the Antarctic and Sub-Antarctic; this is what I wrote to Alvaro on FB yesterday:

"one more comment, for birders going to or already have visited South Georgia and will use this paper to try to guess what kind of 'Wilson's Storm-Petrel' they are seeing: this new paper suggests that 'oceanicus' is the form of the subantarctic islands including South Georgia, Falklands, etc. But my data are clearly showing that South Georgia and the Falklands are very different in biometrics, and with the former much closer to the Antarctic birds. So, if one is going on a voyage to South Georgia, he or she cannot just automatically 'tick' it as 'Subantarctic Storm-Petrel', as this paper suggests, as it is NOT, or not sure yet! Secondly, the type locality of the oceanicus is unclear, so this needs to be sorted out, too."

“Last point, based on taxonomic thinking we advocate for the gadflies, and based on what we know so far about Oceanites spp., and if I take your idea that maybe plumage variation is the least important, but not size differences, I would only describe these four taxa as three ecologically distinctive conservation units, as: gracilis (1), pincoyae-barrosi (2), and galapagoensis (3). But I will have a better view once Vincent and I can process much of the data collected over the years.”

“The main obstacles with the recognition of the new Oceanites barrosi sp. nov. Andean Storm-Petrel (Norambuena et al 2024, Zootaxa 5486 (4)) are:

“1. The selected type specimen LACM 25182), and other similar pale-bellied birds that I handled inland, near and east of Santiago, have both measurements and plumage features strongly approaching gracilis (Elliot’s Storm-Petrel). The type is by far more white-bellied, like gracilis and pincoyae, while the other birds shown in photos from traps in the same area are much darker, even without any pale ventral area.

“2. The paper lacks genetic check/comparison of the type specimen LACM 25182, and the birds sampled for barrosi in the tree in Figure 2 not shown in images too. I see the danger that both sources are of different taxa?! 

“3. So, I am left with no other choice to question the description of barrosi, and to ask how I should refer to the birds inland (same area of the type locality), near and east of Santiago with wings 126-128 mm and plumage like gracilis too?

“4. The work selected two paratypes, the ones from the Santiago museum, but they overlooked (or avoided) that in the same cabinet in the museum there are two specimens from the same area inland that are small and very much gracilis-like in plumage.

“5. With Fig S2 (in page 471, Calibrated phylogeny of Oceanites), the work compared the dated splits with other tubenose pairs. Here for comparison, they calculated 1.56 MYA between the Northern Giant Petrel M. halli and Southern Giant Petrel M. giganteus. However, now several studies have found that the divergence of these have occurred only approximately 0.2 MYA. That is about 90% discrepancy using the same marker, Cyt-b. I am unsure how they got the 1.56 MYA for the latter? For the split of pincoyae versus barrosi, they give 6.69 MYA, but if a comparative 90% error might apply, the true divergence might be down to about 0.5 MYA, or less. And yet, this is just by Cyt-b, but if you used nuclear markers, it will probably will not be more than a few thousand years, just as with many other tubenose pairs. So, I suspect that the work have by over 90% miscalculated (overblown) the dated splits, which have mislead them altogether!

“6. The paper gives barrosi (in TABLE S3, form birds trapped in Andean mountains east of Santiago) as having an average wing 135.8 mm and tail 61.3 mm. Such birds, however, I failed to find along the Humboldt Current, including pelagic waters off Valparaiso, from 19 November to 16 March, i.e. in the same waters, same dates, virtually only large-sized Oceanites [oceanicus] complex birds with average wing 148.7 mm and tail 65.1 mm, i.e. by far larger than the new barrosi sp. Nov. (I have samples of such birds, and I am unsure of their origin, pending analysis).

“7. My suspicion is that because they miscalculated the dated splits, they concluded that pincoyae and barrosi are different species, and also both from gracilis, but in fact all three represent variation of same thing, gracilis! Their tree (Figure 2) is confirming that, but it just overblown the divergence between them!”

Response from Heraldo V. Norambuena: “It is interesting to see the interest that our work with Oceanites has generated, and we hope that it will be the starting point for a more in-depth analysis of the evolution and speciation processes of this fascinating group. Some points to clarify/answer:

1) As clearly indicated in our article (Norambuena et al. 2024) and the proposal, we define species based on our phylogenetic hypothesis and morphological analysis, not just on morphology. For those more comfortable with the concept of morpho-species, it is expected that our results will not please them. If we were to base ourselves on morpho-species, we would have 8,000-9,000? described species (as in the period before Sibley & Ahlquist 1990), and multiple proposals submitted to the SACC would not have been validated. We agree that vocalizations are essential for species delimitation. Indeed, we have described some vocalizations for Oceanites gracilis (Barros et al. 2020), and we have acoustic monitoring with ARUs in progress with Markham’s and Ringed Storm-petrels. Still, we must be realistic because it is a significant challenge for seabirds to generate this data. First, we need to find the breeding sites of barrosi and pincoyae, we are close, just as we have already done with several Atacama Desert storm-petrels (e.g. Schmitt et al. 2015, Barros et al. 2019, 2020, Medrano et al. 2019, Norambuena et al. 2021). Our species delimitation is not based on Figure S2 and its divergence times but on Figure 2, which shows a topology and branch lengths representing the cyt-b-based genetic divergence for each clade. Mr. Shirihai is, therefore, misinterpreting the phylogenetic trees. The monophyly of gracilis concerning pincoyae and barrosi is quite evident in Fig. 2, so the criterion of considering this large clade as gracilis alone does not make sense with phylogenetic/evolutionary species concepts (see de Queiroz 1998, 1999, 2007). I would agree with this point if the phylogeny showed a clade with no taxonomic or geographic correspondence and slight genetic divergence (as occurs with the clade of exasperatus in our Fig. 2), but that is not the case. It should be noted that a multispecies approach such as ours provides greater clarity of the process of cladogenesis or speciation than one based on population genetics of a single taxon (as Mr. Howell points out concerning Leach’s Storm Petrels; there is an analysis scale error!). We decided to calibrate the phylogeny to explain the biogeographic process behind the speciation of the group to understand better certain relationships such as galapagoensis/gracilis, which, according to our analysis in BioGeoBEARS is due to a jump-dispersal event, due to a long-distance colonization event in biogeographic contexts quite different from what we see today (marine introgressions before the Andean uplift were quite frequent in Chile and Argentina, you can see a review here Hoorn et al. 2022 https://doi.org/10.1093/botlinnean/boab098). As Howell points out, it is essential not to forget the biogeography behind these processes, not only the current but also the historical ones.

2) Measurements: Regarding this criticism and the exaggerated assertion that there is no bird of such size in the ocean, we can specify that our measurements for barrosi were mainly taken from birds captured with mist nets in the mountain range near their breeding area and skins from land near those areas, in Río Blanco (6; 2 of them are part of the type series), Río Colorado (2) and El Morado (1). We considered the measurements of three specimens from the Museo Nacional de Historia Natural de Santiago: MNHNCL3606 from Cerro Manquehue, Santiago (paratype, female, Nov-1966), MNHNCL3500 from Aconcagua (paratype, male, 15-03-1961) and MNHNCL1957 from Santiago (male, Feb-1944). The specimen MNHNCL3606 measures 143mm in the wing and 58mm in the tail; therefore, O. barrosi individuals can reach the measurements that Mr. Shirihai indicates (we measured ‘natural wing length’). Mr. Shirihai argues that there are gracilis-type specimens in the collection but does not indicate the location or number of the specimens. It should only be noted that there are indeed three specimens labeled as gracilis, two from Mejillones, Antofagasta, 1,400 km away from Santiago (MNHNCL4673 without sex from 19-12-1982 and MNHNCL4674 without sex from 20-12-1982), a third MNHNCL4722 with locality Pudahuel, Chile (without sex, 17-05-1982). Our measurements for that specimen (MNHNCL4722) are 126 mm wing and 55 mm tail, which coincides with what Mr. Shirihai points out, with the difference that the amount of white on the belly does not correspond to gracilis; on the contrary it shows the pattern that we describe for barrosi (we attach a photo of the specimen and a comparison with gracilis captured in northern Chile). For O. gracilis, we included in our analysis three birds captured at their breeding sites on Isla Chungungo and Pampa del Indio Muerto; information on that species is published in Barros et al. 2020 (https://doi.org/10.5253/arde.v108i2.a7). Of all the skins reviewed, we discarded about 50 specimens from our analysis because they did not have a precise location on the labels (many from AMHN), or because they were caught at sea, for example, USNM skins collected in “Pacific Ocean off Peru”. Therefore, our sample at this point is low because we were picky in selecting the skins to consider, and the data are mainly adjusted to data from birds captured near or at the breeding sites. Many data reported by Harrison et al. (2013) were not included because they did not consider Mid-toe+claw a relevant measure. At this point, I would like to clarify that measurements and captures of birds at sea will most likely be misleading because, as we pointed out on page 463 of our paper, the distribution of each taxon at sea must be assessed with the appropriate technology (e.g. geolocators). At certain times of the year chilensis/pincoyae/barrosi/exasperatus? could coincide in some regions of the Pacific (or Humboldt current) (I know it is a problem, but it is a great challenge to solve with science.) Any morphological data discussed should precisely indicate the number of specimens, location, and date. At the ROC, we collaborate with institutions that are rehabilitating birds affected by light pollution in Santiago and northern Chile (birds captured on land require capture permits), so in future works, we will be able to have more information for several species (the data from our publication are available upon request).

3) Holotype: It is an excellent point to sequence the holotype. I recognize it was a mistake not to have done so, but that does not invalidate the pattern evident in the phylogenetic analysis. We will consider this observation and ask for samples from the LACM specimens and the MNHNCL to sequence the entire type series. The description of the species requires us to define a morphological standard, but as in any population, the distribution of the traits will (or should) have a normal distribution, with an average of specimens adjusting to that standard described (Diagnosis) and with specimens that deviate from that average, with a greater or lesser amount of white, and that also applies to sizes. The proposal discusses that point, but I can add that O. barrosi does not have a white patch on its belly, as O. pincoyae and O. gracilis do. Oceanites barrosi have white feathers that may stand out between the belly feathers; this white may be more evident in museum skins because the taxidermy process disarranges these feathers, the "gracilis" close to the type locality including those pointed out by Mr. Shirihai are indeed barrosi.

4) Calibrated phylogeny: On this point, Mr. Shirihai does not cite a source. The only work that explicitly shows the divergence between Macronectes halli vs. M. antarcticus is by Techow et al. (2010), who points out a divergence of 0.5 MYA, not 0.2 MYA, as Mr. Shirihai points out. Therefore, the discrepancy he indicates is miscalculated. In our work, we are critical of the shortcomings of these calibrations, and we suggest that this is a point to be improved with a multilocus or genomic coverage (last lines, first paragraph, page 464). Regarding the result, in addition to the uncited example indicated by Mr. Shirihai, multiple examples with Procellariiformes indicates quite old ages for several clades of this group (e.g., Gangloff et al. 2012, Wallace et al. 2017, Estandía et al. 2021 pre-print). Some studies have shown that cyt-b-based calibrations often estimate older dates than multilocus or genomic ones (we have experience with Neotropical Pipits, see van Els et al. 2019, Norambuena et al. 2018, 2021). Still, those differences do not give variations from 6 MYA to 0 MYA (that is why the tree must be based on Bayesian analysis because it is the most accurate way to have reliable and informative branch lengths); if that were the case, the topology of the phylogeny based on cyt-b or another mitochondrial gene would be closer to a polytomy with very short branch lengths, and that is not the case with our phylogeny with well-resolved clades. SACC members have enough experience to draw their conclusions.

5) Finally, for the Atlantic taxa and populations, the lack of clarity of the 'terra typica' of oceanicus is a problem with no solution. I agree that it is essential to understand well what happens with Malvinas/Falklands – South Georgia – Isla Los Estados, but you will realize that considering our results, we had to propose something and biologically Antarctic vs. Sub-Antarctic makes sense, Bourne (1964) and Murphy & Beck (1918) had already pointed this out in some way, and we cited both in the discussion of our article. However, I agree that it is necessary to obtain samples from South Georgia and other sub-Antarctic islands (I hope that colleagues working on that side will achieve it soon!). But again, oceanicus (or what should be oceanicus) and exasperatus are phylogenetically diagnosable.”

Comments from Manuel Marín:

“Morphometric data. The authors provided morphometric data on 79 individuals. However, 37 of these were from live, mist-netted birds, whereas the remainder were from museum specimens. Comparing measurements from live birds to specimens is problematic because of known shrinkage of museum specimens. Further, within each of those two categories, no indication is given in terms of who measured the specimens, other than “all the information in the database is unpublished and was measured by H.V.N. and R.B.”. If each sample was measured by the same individual, then that should have been mentioned in the Methodology; if more than one person measured specimens within each category, then that also should have been mentioned because individual differences in the way specimens are measured is a known source of error that can be important when measurement differences are close. This may not have affected the results at all, but as is, we have no way of assessing this potential problem. Because no museums are in the Acknowledgments for assistance in providing measurements or for access to the collections, the measurements must have been done during visits to those collections by H.V.N. or R.B. themselves. The Zootaxa review process let the authors down on not clarifying these details (and not requiring the customary courtesy of acknowledgements to curators of the collections used).

“Type specimen: I know the type specimen well and also the other older specimens mentioned in the paper. In 2000, I identified the type as O. gracilis (still on the label in my writing) for a paper on seabird vagrancy. The identification was on the base of size and the white feathers on the belly. Type specimen: LACMNH 25182: (my measurements) Wing 136; tail 59; tarsus 30.2; culmen 10.8.

“At the time I compared it to 11 specimens from Cape Horn for O. o. chilensis. However, I failed myself to notice the Oceanites paper by Murphy (1918; Bull AMNH 37:117-146), in which he discussed age classes and plumage: A) young birds have white feathers on the lower belly but through aging they become dark – pretty much like Cypseloidine swifts. B) and there is a size difference – the young birds are smaller than older ones. Norambuena et al. labeled the type specimen as adult. However, in my notes I wrote that it has DOWNY feathers on the lower body and flanks—and my goodness if you look carefully at the picture, they are there, and they can be seen as you zoom in. Indeed, the type is a young of the season, in my opinion.

“Plumage variation: Although Murphy (1918) is cited (as Murphy & Beck 1918) with respect to plumage variation, the problems that Murphy discussed were not dealt with directly in the paper’s analysis of plumage and morphology. Here are the relevant sections from Murphy:

“In the same paper on page 126, Murphy referred to young individuals that are smaller and have weak unguis. ¨These features, added to those of its juvenal plumage (which seems the case here) give the skin a sufficiently different facies from the well known adult form of Oceanites o. oceanicus to serve an ornithologist who was unacquainted with the characters of a juvenal bird as the type of a new race. ¨ In this respect I agree with Howell & Schmitt (2016) that is premature to refer to O. pincoyae as a full species, but equally applicable to O. barrosi. I have been watch closely for the ¨Pincoya¨ type, and I have photos of classic ¨Pincoya¨ types (e.g., white underwings and white from lower chest to belly) at different locations as far north as central Chile area (which would mean a considerable range extension) where ¨barrosi¨ is supposed to be. The degree of variation as mentioned by Howell and Schmitt (2016) is quite large and needs to be addressed before making taxonomic conclusions. An alternative hypothesis is that needs to be addressed directly is that all might turn out to be different age classes of chilensis. Keep in mind that seabirds are not like most landbirds and can take many years before acquiring true adult plumage and before first breeding and might stay for long time at sea before returning for first breeding.

“Misc.: The idea of nesting in the Andes is highly plausible as they are not that far from the coast (110 -140 km), not much for a seabird, and the high Andes emulate the same climatological environment as nesting far south. However, this still needs to be proven; ss far as I know there is no direct evidence whatsoever, as many (if not all) of the individual birds captured inland relate to some weather anomaly, e.g. see Marín 2002 (‘The occurrence of vagrant seabirds inland in Chile’. Cotinga 17: 62-65).”

Comments from David Ainley (voting for Remsen): “#1 --- NO, for #2 --- NO, and for #3, I suggest that all the renaming hold off until questions 1 and 2 are resolved. On the other hand, I’m not a taxonomist, though I stuck my toe into the pool a couple of times in the past. But I did rely on others to save me from drowning, like Storrs Olson.

“Storm-petrels distributed across an extensive, latitudinal space, e.g. Leach’s, show much plumage and size variation, in somewhat of a cline along NA west coast. However, this has not resulted in a parsing into a series of ‘species.’ Pretty much color morphs. The Atlantic and western Pacific don’t show the same extreme of ocean climate variation as either the eastern North Pacific or the eastern South Pacific, somewhat the product of eastern boundary currents. Storm-petrels along these coasts could well reflect the substantive, regional environmental gradation/variation that exists.

“So, the same phenomenon seen in Hydrobates of the eastern North Pacific could very well apply to Oceanites in the eastern South Pacific, i.e. for Oceanites oceanicus. (BTW, why do dark storm petrels tend to frequent closer-to-coast waters than those having substantial white? I’ve long wondered about that.)

“For the geographic distribution of breeding areas for different Oceanites ‘species’ (as shown in the map sent as part of the proposal) it could well be as much a result of species variation as it is for humans not being able to cover the entire potential breeding space, i.e. the Andes etc. from northern to southern, thus, to sample the ‘in-betweens.’ At sea, these ‘species’ are mostly inseparable, except for a few, e.g. gracilis (once upon a time Elliott’s SP), in part owing to the likely broad overlap of Oceanites-types. So, what’s the point of trying to recognize all these ‘species,’ especially on the basis of the very small sample sizes (1-3???) thus far used for measurements and genetics? Golly, ornithologists from yester-year should well be turning in their graves, given all the effort they put out in their careers to recognize species variation as much as they did, measuring countless specimens, but then along comes someone with one specimen (3?) that somehow represents a new species based on DNA, although it can’t really be separated, by eye, from others in the region?”