Puffinus, Brisson, 1760

Salvagni, Thamara & Carlos, Caio J., 2021, Phylogenetic relationshipsof the shearwater taxaPuffinus (sensu lato) and Calonectris (Aves: Procellariiformes: Procellariidae) based on osteological characters, Zootaxa 4963 (2), pp. 243-292 : 272-273

Puffinus (sensu lato)

The final hypothesis found all other members of shearwaters to form a single, monophyletic grouping: Puffinus (sensu lato). We did not recover the split of this clade into two subgroups (i.e. genera Ardenna and Puffinus ) as suggested by molecular-based hypotheses (e.g. Austin 1996; Heidrich et al. 1998; Austin et al. 2004; Penhallurick & Wink 2004; Pyle et al. 2011). The Puffinus (sensu lato) clade, as recovered herein, is relatively well-supported (RBS = 76) and established by eight unambiguous synapomorphies, plus six optimized by either ACCTRAN or DELTRAN. Some of these characters include the distinctly curved pars intermedia of the mandible, the dorsoventrally enlarged spina externa of the sternum and the absence of fossa renalis (characters 50, 56, and 90, respectively). Despite minor conflicts, the present hypothesis recalls early morphological studies ( Kuroda 1954; Wragg 1985), although the results cannot be directly compared as they did not use the same phylogenetic approach as we have; however, we can nevertheless discuss the congruence of these different studies.

The need for recognition of two genera among Puffinus (sensu lato) has been suggested in many phylogenetic hypotheses based on molecular data, all using the same marker (e.g. Austin 1996; Heidrich et al. 1998; Austin et al. 2004; Penhallurick & Wink 2004). However, monophyly of the currently accepted Ardenna is not supported in the present analysis. Instead, its members form successive sister taxa to Puffinus (sensu stricto). The topology of the final hypothesis differs significantly from the molecular hypotheses, wherein Ardenna and Puffinus (sensu stricto) are shown to be two different clades, sometimes not even sister genera (e.g. Heidrich et al. 1998; Nunn & Stanley 1998; Ramirez et al. 2010; Pyle et al. 2011). Alternatively, the morphological, pre-cladistic hypotheses of Kuroda (1954) and Wragg (1985) suggested that a single “ancestral” taxon has sequentially given rise to each phylogenetic subgroup within the Puffinus (sensu lato) clade. The final hypothesis, ( Ardenna pacifica ( A. bulleri ( A. creatopus ( A. carneipes ( A. gravis (( A. grisea + A. tenuirostris ) ( Puffinus nativitatis ( P. puffinus ( P. mauretanicus ( P. opisthomelas ( P. gavia ( P. huttoni ( P. yelkouan ( P. lherminieri + P. assimilis )))))))))))))), appears to match this perception.

Kuroda (1954), who studied shearwaters based on different types of data, including external morphology, osteology, and behaviour, proposed a classification with five monophyletic subgroups within Puffinus (sensu lato). According to him, A. pacifica and A. bulleri are closely related, classified as the ‘ Thyellodroma subgroup’, the most “ancestral” group among all Puffinus (sensu lato). Wragg (1985), using only morphological characters, found the same result. Although A. pacifica and A. bulleri formed a clade in the equally weighted tree, they did not group together in the implied weighting tree, as A. pacifica appears to be the sister taxon to all other species, followed by A. bulleri . A similartopology, regarding both species, was presented in the molecular-based hypothesis by Heidrich et al. (1998). All other molecular studies support the grouping of ( A. pacifica + A. bulleri ), which is typically regarded as sister to a clade comprising all other species of Ardenna (e.g. Penhallurick & Wink 2004; Pyle et al. 2011; Welch et al. 2014).

The next species, sister taxon to the remaining species of Puffinus (sensu lato), is A. creatopus , followed by A. carneipes . Kuroda (1954) proposed that these species be combined in the ‘ Hemipuffinus subgroup’, although, in the present hypothesis, they are sequentially arranged. Wragg (1985) did not discuss this subgroup, as his proposal showed a polytomy regarding the position of A. creatopus , A. carneipes , and the clade formed by the remaining species of Puffinus (sensu lato). Molecular-based hypotheses typically consider A. creatopus and A. carneipes as sister taxa; however, the subgroup relationship among the other Ardenna species is sometimes conflicting. Most studies indicate A. gravis as sister taxon to the ‘ Hemipuffinus subgroup’ (e.g. Austin et al. 2004; Penhallurick & Wink 2004; Kawakami et al. 2018), but the grouping ( A. grisea ( A. creatopus + A. carneipes )) was suggested by Tennyson & Shepherd (2017), and Ramirez et al. (2010) was unable to define the position of ‘ Hemipuffinus ’ within the genus due to a polytomy. Although Heidrich et al. (1998) did not include A. carneipes in their molecular analyses, which could have changed their resulting trees, the topology they presented for the relationship ( A. pacifica ( A. bulleri + A. creatopus )) is similar to the present hypothesis.

The last three Ardenna species , following Kuroda (1954), were divided into two other subgroups: ‘ Ardenna ’ for A. gravis ; and ‘ Neonectris ’ for A. grisea and A. tenuirostris ( Puffinus nativitatis was also placed in the latter). According to Kuroda (1954), ‘ Ardenna ’ is closely related to the group formed by ‘ Neonectris ’ and all other Puffinus (sensu stricto).

As recovered herein, A. gravis is sister to (( A. grisea + A. tenuirostris ) + ( Puffinus (sensu stricto ))). Kuroda (1954) placed P. nativitatis in his ‘ Neonectris ’ subgroup based on external morphology and behaviour. Afterwards, osteological analysis by Wragg (1985) led to the hypothesis that P. nativitatis was sister to all other Puffinus (sensu stricto) species, making ‘ Neonectris ’ a polyphyletic group. Wragg’s (1985) hypothesis is now supported by molecular-based studies (e.g. Austin 1996; Heidrich et al. 1998; Tennyson & Shepherd 2017), in which Puffinus (sensu lato) is split into two genera, and P. nativitatis is part of Puffinus (sensu stricto). Nevertheless, the present analysis found ( A. grisea + A. tenuirostris ) to be monophyletic, supported by two unambiguous synapomorphies, plus four according to either ACCTRAN or DELTRAN. Furthermore, this group appears to be the sister clade to Puffinus (sensu stricto). This relationship was previously suggested by Wragg (1985); however, it is not present in most molecular hypotheses. Austin (1996) and Heidrich et al. (1998) placed A. tenuirostris as sister taxon to ( A. gravis + A. grisea ), whereas other hypotheses propose ( A. tenuirostris ( A. grisea ( A. gravis ))) (e.g. Penhallurick & Wink 2004; Pyle et al. 2011; Welch et al. 2014). The conflicting placement of these species has also produced some unresolved relationships within the genus (i.e. Austin et al. 2004; Ramirez et al. 2010; Tennyson & Shepherd 2017; Kawakami et al. 2018).

Molecular-based studies have consistently presented a monophyletic Puffinus (sensu stricto), regardless of its position in relation to Ardenna and Calonectris (e.g. Nunn & Stanley 1998; Austin et al. 2004; Pyle et al. 2011; Tennyson & Shepherd 2017). The monophyly of Puffinus (sensu stricto) in the present analysis is congruent with these hypotheses (RBS = 82), although its members are placed in the same larger Puffinus (sensu lato) clade as Ardenna species in both equal and implied weight analyses. In the present hypothesis, the monophyly of Puffinus (sensu stricto) was supported by 12 synapomorphies from discrete characters (i.e., nine unambiguous, plus three according to eitherACCTRAN or DELTRAN), as well as 20 from continuous characters. Among other characters discussed by Kuroda (1954), these synapomorphies include the cranially positioned apex carinae of the sternum and the slender, pointed crista cnemialis cranialis of tibiotarsus (characters 62 and 98, respectively). Puffinus nativitatis , according to the present analysis, is sister to all other Puffinus (sensu stricto) species, congruent with both morphological (e.g. Wragg 1985) and molecular-based hypotheses (e.g. Austin 1996; Nunn & Stanley 1998; Penhallurick & Wink 2004). It should be noted, however, that P. subalaris , which has been found to group with P. nativitatis as sister taxon to the remaining Puffinus (sensu stricto) ( Austin et al. 2004; Pyle et al. 2011), was not included in these analyses.

In the present analyses, six of nine included species of Puffinus (sensu stricto) are from the group identified as “Manx-type shearwaters”, and two others are part of the “Little/Audubon’s complex” ( Onley & Scofield 2007). The hypothesis for the relationship of these eight species presented herein, ( Puffinus puffinus ( P. mauretanicus ( P. opisthomelas ( P. gavia ( P. huttoni ( P. yelkouan ( P. lherminieri + P. assimilis ))))))), is highly congruent with earlier morphological studies ( Kuroda 1954; Wragg 1985).

Hypotheses based on molecular data have consistently shown some phylogenetic associations among these species; however, their placement within the genus is unclear. In most studies, ( P. gavia + P. huttoni ) is the sister group to the remaining species (e.g. Austin et al. 2004; Pyle et al. 2011; Tennyson & Shepherd 2017), yet some conflicts can be found (e.g. Heidrich et al. 1998). In the implied-weight hypothesis herein, P. puffinus occupies this position, sequentially followed by P. mauretanicus . In the molecular-based analyses, P. mauretanicus is sister to P. yelkouan (e.g. Austin 1996; Pyle et al. 2011; Kawakami et al. 2018), but in the present hypothesis, P. yelkouan is sister to ( P. lherminieri + P. assimilis ), whereas P. mauretanicus appears to fall in between P. puffinus and P. opisthomelas . Heidrich et al. (1998) and Austin et al. (2004) have, alternatively, suggested a monophyletic ( P. puffinus ( P. mauretanicus + P. yelkouan )). Regarding P. opisthomelas , the species is, herein, placed between P. mauretanicus and P. gavia , but in some molecular hypotheses it is found among species from the ‘Little-Audubon’s complex’ (e.g. Austin et al. 2004; Pyle et al. 2011; Tennyson & Shepherd 2017). The only two members of this group included in the present analyses, P. lherminieri and P. assimilis , are sister taxa, grouped in the least inclusive clade of the final hypothesis.