Proposal (773) to South American Classification Committee

 

Split Glaucidium tucumanum from G. brasilianum

 

Effect on AOU SACC classification: Adjust the species level taxonomy of a population currently recognized as a subspecies under Glaucidium brasilianum.

 

Background: From the SACC footnote under the name Glaucidium brasilianum:

 

The subspecies tucumanum was treated as a separate species from Glaucidium brasilianum by Heidrich et al. (1995b), Wink and Heidrich (1999), and Wink et al. (2008) based on genetic data and slight vocal differences. König et al. (1999) followed this treatment, also noting differences in habitat and plumage, as did Marks et al. (1999). Proposal badly needed.

 

All this goes back to a paper by Heidrich et al. (1995) that provided a phylogenetic and voice analysis of American Glaucidium. One of the outcomes was the discovery that the sample used for G. b. tucumanum was not monophyletic with respect to the rest of G. brasilianum, but rather embedded within a clade containing G. bolivianum and G. hardyi. The authors noted that the crown of a specimen identified as “tucumanum” appeared more spotted, and that recordings identified as that taxon were minutely different from other populations of G. brasilianum. Although the authors suggested that further taxonomic assessment was necessary, they did not recommend splitting tucumanum from brasilianum in the paper. König et al. (1999) then went the extra step, apparently convinced by Heidrich et al. (1995), and split G. tucumanum, a taxonomy also used by Marks et al. (1999) and König and Weick (2008). Finally, the recent taxonomy of HBW Alive/Birdlife International has retained this split, but has increased the size of the distribution of G. tucumanum well into the lowlands of the Chaco (see below).

 

So, let’s look at the situation a bit more carefully: Chapman (1922) described G. b. tucumanum based on specimens from Salta (!), Argentina, using AMNH material. His characters outlining the taxon were:

 

“Resembling the black and white-barred tail phase of Glaucidium brasilianum brasilianum but upperparts, wings and streaks below fuscous with (in one specimen) a barely perceptible tinge of brown; the crown with small, inconspicuous whitish spots or shaft-streaks; broken nuchal band, white; back with practically no white markings. d1; Wing, 90; tail, 65; tarsus, 16 mm. 9: Wing, 95; tail, 65; tarsus, 16 mm.

 

He also stated:

 

While I have no doubt of the distinctness of the form here described, I do not know whether it should be accorded specific or subspecific rank.

 

The specimen from which the “tucumanum sample was taken by Heidrich et al. (1995) was from “El Tala, Salta, Argentina (SMNS 63625) [SMNS = Staatliches Museum für Naturkunde Stuttgart, Stuttgart, Deutschland]; I presume that this means the tissue specimen has a voucher skin. Heidrich et al. (1995) reported a large brown morph and a small gray morph, for which sonograms were also provided, from this locality, the latter is what was considered tucumanum by the authors, implying that two taxa of Glaucidium are syntopic in the Tucuman yungas, and therefore one must be a species separate from G. brasilianum. Another paper by Wink et al. (2008) provided another phylogenetic tree for the Glaucidium, and reaffirmed the species status of G. tucumanum.

 

Analysis: Given the anomalous result (that tucumanum was not part of the brasilianum clade) in the phylogenetic analysis of Heidrich et al. (1995), my first inclination would be to question the identity of the samples. Note that G. bolivianum is known from the yungas of Argentina (despite the name, the type locality is in Jujuy!), which gives me reason to believe that the sample used for “tucumanum” was, in fact, a misidentified G. bolivianum! Even if the tissue sample wasn’t mislabeled, and the specimen used was not of G. bolivianum, then one must compare it to the holotype of G. b. tucumanum to be sure they represent the same taxon. Furthermore, unless the vocalizations reported by Heidrich et al. (1995) were reliably vouchered to birds (and sample sizes are large enough), it would seem premature to suggest that tucumanum has an identifiably different voice from other G. brasilianum populations. Again, according to Heidrich et al. (1995), the differences in voice were not enough for them to propose species status by that character alone. Here are the trees published in Heidrich et al. (1995):

 

 

 

By contrast, the tree presented in Wink et al. (2008) disagrees with the tree topologies of Heidrich et al. (1995) in that it does not show tucumanum embedded within a clade with bolivianum/hardyi, but rather, the sample labeled “tucumanum” is sister to one labeled “brasilianum,” and is embedded in a clade that is largely comprised with other samples of the brasilianum group (with an anomalus G. griseiceps also included… again, a misidentified sample?). To me, this implies that the tucumanum sample used by Heidrich et al. (1995) was almost surely a misidentified bolivianum (and perhaps this accounted for the “two morphs” reported in that paper?). Strangely, the disagreement of the trees was not acknowledged by Wink et al. (2008), in fact, they still cited Heidrich et al. (1995) as the basis for considering tucumanum a separate species from the remainder of the brasilianum group (although only mentioning “voice and size” as the evidence, nothing about phylogenetic placement). Conveniently, Wink et al. (2008) do not provide the sources of their samples, so we cannot know if they are the same as those used by Heidrich et al. (1995) or not. See the screengrabs of the Wink et al. (2008) tree below (first the larger figure, then a blow up of the relevant clade. Note weak node support within the G. brasilianum clade):

 

 

Furthermore, König and Weick (2008) absorbed G. b. pallens (type locality: western Paraguay) into G. tucumanum, and suggested the name “Chaco Pygmy-Owl” for this species (never mind the fact that tucumanum, sensu stricto, is from arid Andean slopes, not the Chaco), a taxonomic move that has not been followed by other authors.

 

König et al. (1999), and König and Weick (2008) have been rather reckless, to be kind, in some of the novel taxonomies they have presented, this being an excellent example. An apparent misidentified sample, published without being questioned, has provided the impetus for the splitting of a taxon from G. brasilianum, and even though follow-up publications authored by the same lab suggest that the initial result was, indeed, erroneous, they have not amended their taxonomy accordingly (presumably because it agreed with “field observations” that may or may not actually involve the taxon in question). That their taxonomy has influenced HBW Alive/Birdlife International is rather alarming. HBW Alive/Birdlife International (https://www.hbw.com/species/tucuman-pygmy-owl-glaucidium-tucumanum) have taken a strange hybrid route in their taxonomy of the form, accepting G. tucumanum as a species separate from G. brasilianum. However, in their range map, they include the Chaco, but they *exclude pallens* from tucumanum! This scenario has invited quite some confusion on the HBW website and other websites (e.g., Xeno-canto.org) about what names to put on birds from Bolivia, Paraguay, and Argentina.

 

Recommendation: Suffice it to say, I am rather unimpressed by the evidence presented by Heidrich et al. (1995), König et al. (1999), Wink et al. (2008), and König and Weick (2008), in support of species status for G. tucumanum. To me, this taxonomy is a series of errors that some (e.g., HBW Alive) have taken on faith without really reviewing the evidence. If the steps outlined above were taken to confirm that the tissues, specimens, and voice of König et al.’s “G. tucumanum” truly refer to that taxon, and that the two published phylogenies—that disagree with one another in topology—really do support species status for G. tucumanum, these data have yet to be convincingly published. Until this is done, I must recommend a strong NO vote on this one.

 

Literature cited:

Chapman, F. M. 1922. Descriptions of apparently new birds from Colombia, Ecuador, and Argentina. American Museum Novitates 31: 1-8.

Heidrich, P., C. König, and M. Wink. 1995. Bioakustik, Taxonomie und molekulare Systematik amerikanischer Sperlingskäuz (Strigidae: Glaucidium spp.). Stuttgarter Beiträge zur Naturkunde, Ser. A, 534: 1-47.

König, C., F. Weick, and J.-H. Becking. 1999. Owls, a guide to the owls of the world. Yale Press, New Haven.

König, C, and F. Weick. 2008. Owls of the World (second edition). Yale Press, New Haven.

Marks, J. S., R. J. Cannings, and H. MIkkola. 1999. Family Strigidae (typical owls). Pp. 76-243 in Handbook of birds of the world. Lynx Edicions, Barcelona.

Wink, M., P. Heidrich, H. Sauer-Gürth, A.-A. Elsayed, and G. Gonzalez. 2008. Molecular phylogeny and systematics of owls (Strigiformes). Pp. 42-63 in König, C, and F. Weick. 2008. Owls of the world (second edition). Yale Press, New Haven.

 

Dan Lane, January 2018

 

__________________________________________________________

 

Comments by Robbins: "NO, for reasons underscored by Dan Lane."

 

Comments from Stiles: "NO. I fully agree with Dan that there is too much uncertainty regarding specimen identifications and disagreements in genetic data to accept tucumanum as a valid species as things stand."

 

Comments from Areta: "NO. As pointed out by Dan, evidence is meager and inconsistent. The phylogenetic placement of tucumanum as sister to bolivianum and hardyi in Heidrich et al. (1995) cannot be attributed to the sampling of a bolivianum from the series of Calamuchita specimens that MAY be the source of the sample: there is no bolivianum in the Sierras de Córdoba. However, it may be the case that they have sampled another bolivianum from somewhere else. Without more knowledge on the provenance of this sample, there is little hope to understand what the Heidrich et al. (1995) tree shows and so the placement of tucumanum must be regarded as dubious in this work. The sample of tucumanum from Wink et al. (2008) [and the 2009 Ardea paper] comes from IPMB 6310, but I am unable to find where that voucher is from, except from knowing that it came from Argentina. Placement of tucumanum as sister to brasilianum from Argentina is more appealing than the result of Heidrich et al. (1995). However, without knowing where this sample came from, we cannot be sure that it does pertain to tucumanum (perhaps I am missing something on the collecting locality of IPMB 6310?).

         "The type of tucumanum comes from Rosario de Lerma, Salta, Argentina, at 4800ft (1600m) asl (Chapman 1922). Based on other specimens collected by Miller and Moyle within two days of the date of collection of this specimen (10 Jan 1916), it must have been collected in some fairly open interfase between the humid Yungas and the drier woodland that contacts with the Andean foothill. The area of collection is not an arid slope (it becomes arid above ca. 2500m), but rather an area with rainforest and somewhat drier forest segments at the base of the mountains. Many species found in this interface extend to the Sierran Chaco, while others do not. Around Salta city in the Lerma Valley and its surroundings (i.e., some 30-40 km N of where the type of tucumanum was collected), I have always heard typical brasilianum vocalizations and my recordings from this area show that there is no obvious way to distinguish vocalizations of tucumanum from brasilianum. At higher elevation in wetter places in the Lerma Valley, bolivianum is very local. The differences in vocalizations between bolivianum and brasilianum are obvious and immediate, unlike those between tucumanum and brasilianum that seem undistinguishable. I would also argue that in comparison to other Glaucidium species comparisons, the pair tucumanum-brasilianum would not meet the standard level of differentiation between good species.

         "Although formal vocal analyses are still lacking and genetic data is scarce, available data is not enough to favor recognizing Glaucidium tucumanum as a separate species from Glaucidium brasilianum."

 

Comments from Zimmer: “NO.  Vocal differences between tucumanum and brasilianum appear to be minimal at best, which, combined with the uncertainty regarding specimen provenance and inconsistencies in the genetic data sets, is enough to make this one a non-starter in my opinion.”