Proposal (647) to South American Classification Committee


Spilt Ardenna from Puffinus


Background:  Our current SACC footnote is as follows:


23a. Based on genetic data, Penhallurick & Wink (2004) proposed that Puffinus be subdivided into two genera: (1) Ardenna for creatopus/carneipes, gravis, griseus, tenuirostris, pacifica, and bulleri; and (2) Puffinus for all other taxa. However, see Rheindt & Austin (2005) on use of genetic distances for assigning taxon rank.  SACC proposal to split Ardenna from Puffinus did not pass.  Other genetic data (Austin et al. 2004, Pyle et al. 2011) have confirmed these findings, and Dickinson & Remsen (2013) and del Hoyo & Collar (2014) adopted this classification.  Proposal badly needed.


Note that we previously rejected a proposal to do this based on Penhallurick & Wink (2004), who found a deep split within Puffinus.  Objections were that this was based on only 1 gene, and that the two proposed genera were diagnosed mainly by body size.  However, Jaramillo, who knows these pelagics better than anyone else on SACC, pointed out biogeographic themes that separated the two genera.


New information:  Shortly after publication of the Penhallurick-Wink paper, Austin et al. (2004) published a phylogeny, also based exclusively on cytochrome b.  Not surprisingly, therefore, they also found the same two divisions in Puffinus.  Both studies found that Calonectris was the most closely related genus to Puffinus but were unable to resolve relationships among Calonectris and the two groups of Puffinus.

         Pyle et al. (2011) also used cytochrome b to produce a phylogeny of shearwaters (the focus of which was to investigate the specimen that they showed was a separate species, Puffinus bryani), and their tree was, no surprise, similar to the previous two.  Calonectris is sister to true Puffinus, but the support for that node is weak.


Analysis and recommendation:  If the support for a Calonectris-Puffinus sensu stricto were solid, then there would be no controversy in recognizing Ardenna as separate from Puffinus.  As it stands, the decision is subjective.  The three published trees are all based on the same mitochondrial gene.  Nonetheless, I favor a YES on this one – we adopted the split in Dickinson & Remsen (2013), and HBW also did this independently.  Our rationale was that the degree of genetic divergence, at least in cyt b, is fairly deep, and regardless of sister relationships, if we recognized Calonectris, then we also ought to recognize Ardenna as an equally divergent group.  As noted long ago by Alvaro, the two groups of shearwaters differ in general biogeography.  Except for Wedge-tailed (pacifica), the Ardenna shearwaters (Buller’s bulleri, Slender-billed tenuirostris, Sooty grisea, Great gravis, Pale-footed carneipes, Pink-footed creatopus) breed in cold-water islands in the Southern Hemisphere, mainly in the Pacific), whereas Puffinus shearwaters are more widespread, especially at tropical latitudes, and several species breed in the Atlantic and Mediterranean (e.g. Manx puffinus and Audubon’s lherminieri).  True Puffinus are largely black or dark brownish above and white below, whereas species in the Ardenna group show much greater variation in plumage (and none are truly solid blackish above).  These are all weak differences, with overlap.  (I predict that seabird biologists will find stronger, non-overlapping differences.)

         Nonetheless, Peters (1931) recognized the two groups as subgenera, but placed Wedge-tailed in its own subgenus Thyellodroma.  Oberholser (1917; evaluated all the shearwaters on morphological grounds, including the four genera mentioned above plus Hemipuffinus (for carneipes and creatopus), Alphapuffinus (for assimilis, lherminieri, persicus), and Neonectris (for tenuirostris and griseus).  He concluded that Calonectris, Ardenna, Thyellodroma, and Puffinus should all be recognized based on morphological differences (but not the other three genera).  This classification would match that in the proposal and in the classification of Penhallurick and Wink except for inclusion of Thyellodroma in Ardenna.  Oberholser’s Thyellodroma does form a monophyletic group (pacifica and bulleri) that is sister to the rest of Ardenna (and might be the topic of some future proposal if a new phylogeny emerges with many additional loci that would allow a broader view of degrees of divergence).

         A YES vote, to recognize Ardenna, would produce the follow changes in our classification.  Let’s call this subproposal A.


Ardenna pacifica Wedge-tailed Shearwater
Ardenna bulleri
Buller's Shearwater
Ardenna grisea
Sooty Shearwater
Ardenna tenuirostris
Short-tailed Shearwater
Ardenna gravis
Great Shearwater
Ardenna creatopus
Pink-footed Shearwater
Ardenna carneipes
Flesh-footed Shearwater

         Note the changes to the variable endings of A. pacifica and A. grisea.


         A minor bookkeeping change in linear sequence is needed regardless of the vote on A.  Let’s call this subproposal B.  Following the convention of linear sequencing in which the least diverse branch is listed first, the sequence should be (using the topology in Pyle et al., in which all the nodes have reasonable support), with indentations used to signal nodes:











         I recommend a YES on this minor sequence change because it makes our sequence reflect the best data on relationships.


         Finally, one more extremely minor change in our classification is needed to make our sequence conform to the topology of the tree in Pyle et al. (2011).  The reason that we ranked P. subalaris as a separate species is that Austin et al. (2004), and now Pyle et al. (2011) found that subalaris falls outside the large group of species that contains P. lherminieri (and P. assimilis).  Therefore, following conventions of linear sequencing, subalaris should precede P. lherminieri and P. assimilis.  Let’s call this subproposal C, and I recommend a YES on this as a matter of bookkeeping.


Literature Cited

AUSTIN, J. J., V. BRETAGNOLLE, AND E. PASQUET. 2004. A global molecular phylogeny of the small Puffinus shearwaters and implications for systematics of the Little-Audubon Shearwater complex. Auk 121: 847-864.

OBERHOLSER, H. C.  1917.  Notes on the genus Puffinus Brisson.  Auk 34: 471-475.

PENHALLURICK, J, AND M. WINK.  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.

PYLE, P., A. J. WELCH, AND R. C. FLEISCHER.  2011.  A new species of shearwater (Puffinus) recorded from Midway Atoll, northwestern Hawaiian Islands.  Condor 113: 518-527.


Van Remsen, September 2014




Comments from Robbins: “Given that this is totally subjective I could go either way.”


Comments from Stiles:

“A. YES.  Given that the splits between Ardenna-Puffinus and Puffinus-Calonectris are of similar age, it seems reasonable to recognize Ardenna, even though only a single gene was sequenced.

“B. YES, on present evidence (although one hopes that more of the genotypes may son be sequenced).

“C. YES, for keeping the books consistent.”


Comments from Nores:

“A: YES, although not very convinced (among other things, only a single gene was sequenced).


“C: YES”


Comments from Pacheco: “A – YES; According to the proposal, since the three studies demonstrated the deep level of genetic divergence, despite based only on cytochrome b


C – YES”


Comments from Jaramillo: “A – YES. As noted in the proposal, there are also biogeographical considerations. Also Ardenna are bulkier, heftier birds than Puffinus. The Thyellodroma subgenus is indeed structurally different, with the wedge tail, and the more bowed-wing flight style but more data is needed on them. Currently Ardenna is diverse enough between the bulkier gravis-carneipes-creatopus group vs. grisea-tenuirostris that Thyellodroma fits ok in this relatively diverse genus. B – YES. C – YES, move order of subalaris.