Proposal (818) to South American Classification Committee
Split Pyrocephalus rubinus into multiple species
The Vermilion Flycatcher is a widespread, common species that forms a monotypic genus. It is not a species that has stood out taxonomically, other than it often gets called out as unusual for a tyrannid because the male is so brightly and distinctively colored. What has also caught the attention of some is that while male plumage of various geographical forms is similar, the plumages of females are not, with some being quite distinctive … specifically, those from the Galapagos islands.
Pyrocephalus rubinus is very widespread and it shows substantial geographic variation with 12 traditionally recognized subspecies (see distribution map below), much of it based on differences in female plumage. Although no suggestions to separate the species into multiples has been made in the past, it is worthwhile to note that a largely ignored paper by DeBenedictis (1966) notes the radically different voice of one of the Galapagos populations. DeBenedictis described the aerial display and vocalization of one population (Isabela Island) in the Galapagos and confirmed that it is fundamentally distinct from mainland populations. Rather than a rising series of notes as on the mainland, the Galapagos population gives a single repeated note. Based on this paper, one would think that the single species status of Pyrocephalus rubinus would have been called into question, but as mentioned above this note has largely been ignored, although recent authors have suggested that P. rubinus is more than one species (Ellison et al., 2009; Farnsworth and Lebbin, 2004.). Recordings of the Galapagos birds have not been widely available. In his first trip to Galapagos, A. Jaramillo was able to obtain poor recordings of the Isabela population of the Vermilion Flycatcher and confirmed the description of the vocal display as noted by DeBenedictis (1966).
Carmi et al. (2016) took a fresh look at Pyrocephalus, with molecular datasets in order to clarify the relationships of taxa within the genus. A total of 85 individuals was sampled, from 10 of the 12 named subspecies in Dickinson and Christidis (2014). Two mitochondrial protein-coding genes (ND2 and Cyt b) were extracted as well as two nuclear loci (ODC and FGB5).
The mitochondrial DNA tree shows that Pyrocephalus is monophyletic and is separated by a very deep branch from closest relatives. Seven clades show up in the data, including three from the Galapagos Islands. These clades from two sister groups, one of the three clades from the Galapagos, and the other of the remaining four clades from the continent. In the Galapagos, one clade corresponds to the subspecies dubius from San Cristobal Island, the geologically oldest island in the archipelago with a member of Pyrocephalus. The other two correspond to nanus, one clade from the older northern islands (Pinta, Marchena, Santiago, Rábida, Pinzón, and Santa Cruz) and the other from the younger southern and western islands (Fernandina, Isabela, and Floreana). The continental clades separate into two groups: a South American group and a North American group. The South American clade further separates into the austral migratory rubinus group, the populations along Western South America, and those in northern South America.
The nuclear allele networks show a different pattern. In the ODC network, one allele (d) was found throughout continental populations. Of the five alleles that differed from allele d by more than one substitution, three were found in Austral rubinus. The possible root of the allele network was closer to two of these rubinus alleles. A total of seven alleles were unique to rubinus. One allele was unique to nanus from the Galapagos. The highest allele diversity was found in rubinus. In the FGB5 network there were three groups of alleles, one of several alleles found in the North American clade, one of birds from the Western South American clade, and one of rubinus the Austral migratory clade. Structure was evident in this dataset; alleles from rubinus were all characterized by a 3-bp deletion, North American birds by a 14-bp deletion. Again, rubinus showed the highest allele diversity of any group. No nanus data was available for this gene.
Galapagos – the separation between the Galapagos group and the mainland group is estimated to be roughly a million years ago. These birds are smaller than mainland birds, with visibly different female coloration. Males are different as well, with a more restricted red cap than mainland birds. Structurally the Galapagos birds have very short, weak tails, and short, rounded wings. As noted above, nanus is vocally quite different from mainland birds. Their preference for open forest and forest edge is a habitat quite different from the more open country taken by the continental populations. Osteological differences have also been noted and used to suggest species status for Galapagos birds (Steadman 1986). In summary, various independent lines of evidence can be used to conclude that there is a different species on Galapagos than the mainland. What is novel is that the genetic data also clarifies the distinctness of the San Cristobal population dubius. Unfortunately, no vocal data is available for dubius, and it may in fact be extinct now. The branch length separating dubius from the other Galapagos populations is quite old, suggesting the split is older than any division seen in the mainland clades. The geographic pattern also fits a general one seen in the Galapagos, with the old branch (dubius) restricted to the older eastern islands, in this case San Cristobal. More work is needed to understand if more than one species is present in nanus, but certainly dubius appears to be a good species.
Austral migrant rubinus – There are multiple clades within the mainland Vermilion Flycatchers. Perhaps there are multiple species level questions to be resolved although nothing obvious. However, multiple lines of data clarify that the southern migratory rubinus deserves species status. What is confusing is that the mtDNA data suggest that it is nested within the mainland group. The nuclear data show a different pattern where the distinctness of rubinus is perhaps clearer. There are various reasons why the mtDNA results may be incorrectly showing the relationship of rubinus, and on this Carmi et al. (2016) do not elaborate. The mtDNA data show a sister relationship with the northern South American group, where rubinus winters. It is not impossible that historical hybridization with that population may be reflected in the current mtDNA results and that this may not be its true history? There is no current evidence for hybridization, and the breeding ranges of rubinus and saturatus do not come close.
The important point is that the mtDNA do show rubinus to be a separate clade within the mainland populations. Nuclear DNA further supports distinctness of this group. But more importantly, the birds themselves show a clear biological difference, vocalizations. This section was taken out of the Carmi et al. (2016) paper by the editors. Although sample sizes were low, playback experiments I have conducted are clear: rubinus does not respond to a northern song and vice versa. I have more data currently, all of it unpublished, and the same pattern remains. Furthermore, experiments playing voice of mexicanus to cocachacrae invoke a response, whereas rubinus is ignored by cocachacrae. The deleted text in the paper is the following:
“Males from Belize (subspecies blatteus) were more likely to respond to song from Arizona males (subspecies flammeus) than to song from Uruguay males (subspecies rubinus; Wilcoxon signed-rank test W=0, p<0.5, n=10). Males from Uruguay were more likely to respond to song from Uruguay males than to song from Arizona males [Wilcoxon signed-rank test W=0, p<0.5, nr=6 (n=9)]. No male from Belize responded to songs from Uruguay, and similarly no Uruguayan male responded to songs from Arizona .”
The general nature and pattern of the song is similar in all mainland Vermilion Flycatchers: a short, rising, and terminally accented trill. The North American, coastal South American, and northern South American birds have similar songs. Compared to rubinus they are lower pitched, are delivered more slowly, and the terminal note is clearly lower pitched than the pitch at the crescendo of the trill. Here is a typical example: =
On the other hand, rubinus is higher pitched, rises quickly and the final note is high pitched, similar to the frequency of the end of the crescendo. This gives the voice an upwardly accented nature, quite different from other mainland Vermilion Flycatchers. All can perform bill snaps during the vocal displays. I have not looked at differences in the call notes. As noted above, birds of the different song types (rubinus and non-rubinus) ignore each other’s voice. Given this clear biological response in a suboscine, rubinus acts like a good species.
Based on molecular data, as well as biological (voice) data, we suggest dividing up the Vermilion Flycatcher into four species: Pyrocephalus rubinus, Pyrocephalus obscurus, Pyrocephalus nanus, and Pyrocephalus dubius. Note that rubinus, nanus and dubius would be monotypic. However, P. obscurus would include: obscurus, piurae, ardens, cocachacrae, saturatus, mexicanus, blatteus, flammeus and pinicola.
This is a tough issue as the Vermilion Flycatcher is one of the most widespread and best known of the Tyrannidae. Although it may be troubling for some to retain this name for obscurus, for reasons that have been discussed by this committee elsewhere, in my opinion the argument for keeping the name is persuasive. Essentially every English speaker who watches birds in the Americas knows the Vermilion Flycatcher, changing this name to something else like Northern Vermilion-Flycatcher is adding complexity to an issue that in the end will create very little confusion for most people in English-speaking countries. It really is a non-issue for 99% of the user group of English Names to keep Vermilion Flycatcher even though it now refers to a subset of what that name used to mean.
I am not keen on adding a modifier to Vermilion Flycatcher for the various forms and prefer distinct and evocative names. The easiest of which is to call the possibly extinct Pyrocephalus dubius the San Cristobal Flycatcher.
For Pyrocephalus nanus, the name Galapagos Flycatcher is already taken. In the Galapagos this species is well known, although it is declining at a precipitous rate. It has become a conservation concern, and I think to respect what the locals call it, an evocative name would be Brujo Flycatcher. Locally it is invariably called “pajaro brujo,” the witch bird. As so many tyrannids have such forgettable names, why not call the most colorful passerine of the Galapagos by a colorful name?
Finally, Pyrocephalus rubinus can be given many names. Perhaps coming up with one that highlights its migratory tendency, being the only firm migrant within Pyrocephalus is appealing. But I could not think of any good name that works. I have seen the name “Scarlet Flycatcher” being used, such as on Xeno-canto. I don’t know if this is a name they just pulled out of their cloaca or if it has some historical context? In any case, my preference would be Ruby Flycatcher to match with the scientific name. Male rubinus are darker than obscurus, a darker red below and darker brown above. But I don’t think that the color differences are enough that one could make an argument of ruby or scarlet versus what vermillion means, essentially, they all suggest a red coloration.
Carmi, O., Witt, C.C., Jaramillo, A. & J. P. Dumbacher. 2016. Phylogeography of the Vermilion Flycatcher species complex: Multiple speciation events, shifts in migratory behavior, and an apparent extinction of a Galápagos-endemic bird species. Molecular Phylogenetics and Evolution 102: 152–173
Dickinson and Christidis (2014). 2014. The Howard and Moore Complete Checklist of the Birds of the World, fourth ed., vol. 2. Aves Press, Eastbourne, UK.
Ellison, K., Wolf, B.O., Jones, S.L., 2009. Vermilion Flycatcher (Pyrocephalus rubinus). In: Poole, A. (Ed.), The Birds of North America Online. Cornell Lab of Ornithology, Ithaca.
Farnsworth, A., Lebbin, D.J., 2004. Vermilion Flycatcher. In: del Hoyo, J., Elliott, A., Christie, D. (Eds.), Handbook of the Birds of the World, Cotingas to Pipits and Wagtails, vol. 9. Lynx Edicions, Barcelona, pp. 374–375.
Alvaro Jaramillo, April 2019
Note from Remsen: Voting structure is as follows:
818A. Split Galapagos nanus (including dubius) from widespread mainland taxa.
818B. Treat dubius as a separate species from nanus.
818C. Treat all mainland taxa as P. obscurus, as a separate species from nominate rubinus.
English names (if splits adopted):
818D: Use separate English names for each species rather than compound names, i.e. “Something Vermilion-Flycatcher.”
818A-E. Use “Brujo” as the “first name” for nanus
818B-E. Use “San Cristobal” as the “first name” for dubius
818C-E1. Retain “Vermilion” as the “first name” for widespread obscurus
818C-E2. Use “Ruby” as the “first name” for nominate rubinus
Comments from Areta:
“818A. YES to recognizing both Galapagos forms as a separate species from mainland taxa. The few vocalizations of nanus I´ve heard are clearly different from rubinus and obscurus, and the females differ notably from those elsewhere.
“818B. NO, until biological data such as vocalizations (if not extinct) or more thorough genetic work provides deeper information on the genetic architecture of dubius. The age of the split between nanus and dubius is not impressive, and given that we do not know what this would mean in terms of reproductive compatibilities, I prefer to recognize dubius as a subspecies of nanus. Also, note that the difference between the two nanus groups is quite deep in comparison to that between rubinus and obscurus, yet we are not discussing their treatment as different species based only on genetic distance.
“818C. YES. I was skeptical of this given the paraphyly of obscurus with respect to rubinus (this is something that I would have like explained in the paper itself). However, after checking all available recordings of songs of the different taxa, I agree with Alvaro in that the vocal differences between rubinus and obscurus (including from mexicanus to cocachacrae) are constant. The lack of answer between blatteus and rubinus, while mexicanus responds to cocachacrae seal the deal for me. There is ample room here to publish these playback experiments together with a thorough vocal analysis of these taxa. It is regrettable that the comments on vocalizations and playbacks were taken out of the ms.”
Comments from Claramunt:
“818A: YES. Very complicated case. I think it is fair to tentatively separate the Galapagos forms as a different species given their plumage, morphological, and song differences, and the fact that they form a separate mitochondrial lineage. So, YES to A.
“818B. NO. Elevating dubius to species mainly because of high levels of mtDNA "divergence" is not justified, in my opinion. Despite widespread belief, haplotypes with 2% “divergence” can perfectly coexist within a single species (see Benham & Cheviron 2019 Molecular Ecology 28:1765–1783). I would like to see more evidence regarding this potential split. So, NO to B.
“818C. NO. P. r. rubinus is somewhat distinctive in song and in the fact that it is an austral migrant but male plumage is barely differentiated and it is not a different lineage genetically: its mtDNA is part of the south American continental genealogy, and it shares nuclear alleles, particularly with P. r. saturatus. Regarding reproductive isolation, I don’t think that a female of a different subspecies will ignore a male of rubinus just because he sounds a little different. She may not react to the song in isolation, but visual cues seem important in this species. Therefore, NO to C.”
Comments from Robbins: “At a minimum, at least two species should be recognized, Galapagos and mainland based on the differences in vocalizations and genetics. Genetic data support a mainland split into at least two species, North and South America. Based on genetic data and the time axis, if one recognizes North and South America as different species, then one should also recognize Galapagos dubius as a species. Given the three options we have been presented, for now, I vote as follows:
“818B. YES, based on comments above.
Comments from Pacheco: “The multiple data available are compelling to separate the Galapagos taxa from those on the Continent. However, for the reasons listed by Nacho and Santiago, I also prefer to maintain nanus associated subspecifically with dubius. Agreeing with Nacho, I am particularly impressed by the constant vocal distinctions between the northern taxa and that southern (nominate) migratory taxon in the continental bloc. Therefore, my votes are: 818A – YES; 818B – NO; 818C – YES.”
Comments from Remsen: “A. YES. All lines of evidence point to species rank, as summarized in the proposal.”
Comments from Bonaccorso:
“818A. YES. All available evidence (molecular, song, plumage, osteology) point towards a distinct species.
“818B. NO. Agree with Santiago. Genetic distance and structure should not be the only criteria for species status.
“818C. Abstain. Neither tree topology nor nuclear networks show a different enough clade. Also, based on tree topology only, it will seem odd to call P. rubinus a species and lump all other subspecies into P. obscurus (then P. obscurus will be paraphyletic). On the other hand, migratory species are different in the way they speciate (they may cause paraphyly on the tree). So, it would be important to publish those song and playback records and do proper analyses, in order to make a better-informed decision.”