Proposal (771) to South American Classification Committee
Break up Megascops guatemalae into several species (II)
A. Recognize M. centralis as a separate species
B. Recognize M. roraimae group as a separate species
Effect on AOU SACC classification: The widespread species Megascops guatemalae would be broken up into several species-level taxa such that M. guatemalae (and potentially a separate M. vermiculatus) would be restricted to Middle America, thus extralimital to SACC’s region, and M. centralis, and M. roraimae (the latter containing napensis and pallida) present within SACC territory.
Background: Way back in 2001, SACC Proposal #12 suggested the splitting up of Megascops guatemalensis, a species that spanned Middle and South America from Mexico to Bolivia and the Guianan Shield. In the proposal, Mark Robbins laid out the situation as it was in the early 2000s regarding the complex, with little more than a few published sound recordings, the bizarre Marshall et al. (1991) cassette tape review/publication, and some field guides that had jumped on board with splitting the complex without any formal study as reference. The proposal was soundly rejected by the committee, and rightly so. However, in the past few years, a smattering of new publications, and a fair number of sound recordings (easily accessible online) have come out that provide some of the missing pieces to this situation, and I think it can be viewed again in slightly better light.
New information: Two new publications, particularly, have shed light on the taxonomy of several members of Megascops, and taken together, they may finally be sufficient to allow SACC to act on the M. guatemalae complex. The first was Dantas et al. (2016), who provided a phylogenetic analysis of the genus (albeit incomplete), and Krabbe (2017), who reviewed the genus, assembling and analyzing vocalizations. In the case of the M. guatemalae complex, Dantas et al. (2016) provided a phylogenetic structure that was not well resolved, but showed that the southern taxa form one clade apart from the Mexican taxa. Dantas et al. said of this:
“The position of another well-supported group within this sub-clade, M. guatemalae/vermiculatus, is unresolved (Fig. 1); in contrast, internal relationships in this clade are all well supported and within-species divergences are relatively high. For example, two samples of M. guatemalae (M. g. guatemalae and M. g. thompsoni) from Mexico are 1.8% divergent from one another (uncorrected mitochondrial p-distance) and are monophyletic with respect to the M. vermiculatus group. Within M. vermiculatus, the Tepuian (roraimae) and Andean (napensis) populations are sisters, with the Panamanian sample (vermiculatus) sister to these two. Pairwise uncorrected genetic distances between M. guatemalae and M. vermiculatus samples ranged from 8.4% to 9.4%, with distances within M. vermiculatus ranging from 3.5% (roraimae–napensis) to 6.3% (vermiculatus–napensis).”
The portion of the tree in question is here:
I must point out, based on the locality of the “M. vermiculatus vermiculatus” specimen used here, it is very likely that specimen actually represents Hekstra’s (1982) taxon centralis--true vermiculatus is from Costa Rica and into the western 1/3 of Panama’s Caribbean slope (using Xeno-canto sound recordings as a basis for identification). Thus, true vermiculatus is not included in this tree, which muddies the conclusions a bit. However, Krabbe (2017) provided a sound analysis that should help bridge some of the deficiencies of the tree. In his paper, Krabbe stated:
“Although slight vocal differences of M. [guatemalae] vermiculatus in pitch and pace from M. g. guatemalae and M. guatemalae hastatus are suggested, sample sizes are small and no recording of the N Nicaraguan form M. guatemalae dacrysistactus was included in the comparison. Additional material is needed to assess the rank of vermiculatus. The data indicate distinct vocal differences in all measurements between M. [guatemalae] centralis (Hekstra) and geographically adjacent forms in the clade (vermiculatus, pallidus, napensis), strongly supporting the suggested species rank of M. centralis (Ridgely & Greenfield 2001, Gill and Donsker 2016). M. [guatemalae] pallidus (Hekstra) has been overlooked by most authors and deserves mention. The similarity in plumage pattern, tarsal feathering, size and proportions to Mexican guatemalae of a specimen (AMNH120332) from the Paria Peninsula, coastal mountains of Venezuela, led Chapman (1931) to conclude that it was specifically distinct from roraimae. Hekstra (1982b) named birds from the coastal mountains of Venezuela pallidus (including in that form AMNH476699 [type], FMNH91892, FMNH91893, and four additional specimens in AMNH , British Museum and Frankfurt Museum) and also described them as being very similar in plumage and tarsal feathering to guatemalae. It would appear that pallidus is a valid taxon. Birds from the coastal mountains vocalize similarly to birds from the eastern slope of the Serranía de Perijá, and usually differ in both pitch and change of pitch from M. [guatemalae] roraimae. However, some recordings of the two are indistinguishable suggesting that despite the morphological differences, pallidus may have been correctly referred to roraimae by König et al. (1999).
“Analyses of the material indicate no consistent vocal difference between M. [guatemalae] napensis and M. [guatemalae] roraimae. Differences between them in plumage and slightly in proportions and tarsal feathering (Chapman 1931) support the validity of the taxon napensis, but their similar vocalizations suggest that napensis was correctly ranked as a subspecies of roraimae by Ridgely and Greenfield (2001).”
Thus, Krabbe (2017) provided persuasive evidence that both of Hekstra’s (1982) taxa, centralis and pallidus, were valid (they have ignored by many authorities, including SACC, as per Proposal #12). Furthermore, by voice, Krabbe (2017) provided strong evidence that M. centralis (type locality in Darién, Panama) is deserving of species status as it is the most different singer of all the M. guatemalensis complex. Using vocalizations to establish its distribution, the points on the Xeno-canto map here (https://www.xeno-canto.org/species/Megascops-centralis) do an adequate job showing that it extends from the Canal Zone of Panama east into the lower Magdalena valley of northern Colombia and south along the Pacific slope of the Andes to southern Ecuador. Given that M. vermiculatus is separated from South American taxa by the very distinctive M. centralis, and then the Andes, we could provisionally consider it specifically distinct from them, and let the NACC sort out the status of vermiculatus with respect to M. guatemalae. Based on Krabbe’s (2017) comments above, and my own impression of voice from listening to recordings available on Xeno-canto, it seems that the best move for the remainder of the South American M. guatemalae complex (including roraimae, napensis, and pallida) is to retain them all in a polytypic M. roraimae. Recordings available here:
Analysis and recommendation: There are two basic changes to be made here, so I will split them into two votes:
A) Based on the above voice evidence, I think that the separation of Megascops centralis from the remainder of the M. guatemalae complex is a given. It is a very distinctive bird compared to the remainder of the group. If this is done, the English name Choco Screech-Owl has been applied to this taxon, and is appropriate. I recommend a YES vote on this.
B) Although the evidence is less compelling, separate the remaining cis-Andean taxa of the complex (roraimae, napensis, and pallida) from the Middle American forms (guatemalae, hastatus, thompsoni, tomlini, cassini, fuscus, dacrysistactus, and vermiculatus), as this makes sense from a biogeographical standpoint. Whether to separate vermiculatus from the remainder of Middle American guatemalae is beyond the jurisdiction of SACC. M. vermiculatus does sound rather more like the roraimae group than guatemalae does to my ear, but with M. centralis intervening, I feel comfortable not considering these disjunct taxa conspecific. Thus, I weakly suggest a YES for this vote as well. If this passes, then the name with priority is M. roraimae, and the English name Foothill Screech-Owl seems appropriate for the cis-Andean taxa, as all are found in foothill elevations.
Dantas, S. M., J. D. Weckstein, J. M. Bates, N. K. Krabbe, C. D. Cadena, M. B. Robbins, E. Valderrama, and A. Aleixo. 2016. Molecular systematics of the new world screech-owls (Megascops: Aves, Strigidae): biogeographic and taxonomic implications. Molecular Phylogenetics and Evolution 94:626-634.
Hekstra, G. P. 1982. Description of twenty-four new subspecies of American Otus (Aves, Strigidae). Univ. Amsterdam Bull. Zool. Museum 9:49-63.
Krabbe, N. K. 2017. A new species of Megascops (Strigidae) from the Sierra Nevada de Santa Marta, Colombia, with notes on voices of New World screech-owls. Ornitología Colombiana 16: 1-27.
Marshall, J. T., R. A. Behrstock, and C. König. 1991. Special review: Voices of the New World Owls (Strigiformes: Tytonidae, Strigidae). The Wilson Bulletin 103:311-338.
Daniel Lane, January 2018
Note from Niels Krabbe: "I had left the specific information for a future paper on the rank of vermiculatus, but as Dan has let out the cat, I should inform that the sample labeled as vermiculatus in the Dantas et al. 2016 paper is indeed centralis. I had a sample from a bird Tom Schulenberg and I collected after tape-recording it in El Oro, Ecuador sequenced (GenBank KU521767), and (after removal of primer remains) it matches the bird from the canal zone in Panama (KT799255) perfectly, leaving no doubt they are the same taxon."
Comments from Stiles: "A. YES, no problem here, given the definite vocal distinction and genetic evidence. B. YES; Foothill Screech-Owl is a good name, and the biogeographical situation also supports considering these as subspecies of roraimae rather than vermiculatus, which would appear to be a separate, monotypic species (though that is NACC’s decision)."
Comments from Robbins: "A. YES. As I pointed out in Proposal 12, the voice (now readily available to all via online sources) of centralis is quite unique and it certainly deserves species status."
Comments from Areta: "A. YES. I had a hard time with this one, until I asked Niels for more information to help dispel my doubts. My central objections were as follows: "It is frequently the case that Megascops species have short and long songs, and a plethora of little-studied and seldom-recorded calls. Although vocal differences are certainly suggestive, I am not convinced that available recordings of vermiculatus are of vocalizations homologous to those of centralis. What if those recordings by centralis are of the short song and most recordings of vermiculatus of the long song? Why would these two taxa lack any short and long song differentiation? Are they closely related to each other? Note also that species in the same clade as "centralis", M. guatemalae and roraimae/napensis, have the two song-types. If Dantas et al. (2016) erred in identifying the sample of centralis as vermiculatus as suggested by Dan, then we would be relying upon an apparently incomplete set of vocalizations, and without DNA backup, to make this split." Niels indicated that centralis is a well-known species that never gives a longer song, which convinced me that the comparisons between vermiculatus and centralis are good to support the split.
"B. A hesitant YES. It seems that vocalizations of roraimae differ to some extent from those of vermiculatus and napensis. However, more thorough vocal analyses of napensis, vermiculatus and roraimae and genetic data from vermiculatus would be welcome to better understand species limits here."
Comments from Zimmer: “YES. This one is a slam-dunk in my opinion. The vocal distinctions between centralis and all taxa in the complex to the north/west and south/east are diagnosable, substantial, and constant, and do not reflect any inadequacies of sampling bias as questioned by Nacho. This one is overdue. B) Recognize M. roraimae group as a separate species. YES, on the biogeographic grounds articulated by Dan in the Proposal – the range of vermiculatus/guatemalae is separated from that of the roraimae-group by the Andes and by the range of the vocally very different centralis.”