Proposal (1006) to South American Classification Committee

Treat Amazona guatemalae as a separate species from A. farinosa (Mealy Parrot)

Note from Remsen: This is a NACC proposal (2023-A-16) that was rejected that I am posting here, with permissions) because it has the potential to affect the distribution and component taxa of what we currently classify as Amazona farinosa. You can see official comments from NACC voters here; the vote was 7 to 5 to reject.

Background:

The Mealy Parrot (Amazona farinosa) occurs in southern Mexico, through all of Central America, parts of northern South America and the Amazon Basin, and also has a disjunct population in the Atlantic Forest of Brazil. Most authorities currently recognize 3-5 subspecies of A. farinosa, which are often split into two groups (sensu Clements et al. 2021): the Northern Mealy Parrot (A. f. guatemalae from the Caribbean slope of southeastern Mexico to northwestern Honduras and A. f. virenticeps from the Sula Valley of Honduras to extreme w Panama) and the Southern Mealy Parrot (A. f. farinosa, which occurs east and south from Panama to Colombia, Peru, Bolivia, the Guianas, and disjunctly in the Atlantic Forest of southeastern Brazil). Although most authorities consider the Southern Mealy Parrot to be monotypic, it is sometimes split into three subspecies, including A. f. Inornata (Panama and Colombia), A. f. chapmani (SE Peru to NW Bolivia), and A. f. farinosa in the central Amazon Basin and Atlantic Forest.

Until recent HBW-Birdlife and IOC splits, the two putative species have been treated as conspecific. HBW-BL split A. farinosa into two species based on the following rationale.

Until recently, [guatemalae] was considered conspecific with A. farinosa, but differs in its yellow vs red lower carpal edge (2); blue-suffused (or blue) crown with broader, more heavily scaled nape feathers forming frequently or usually ruffled ruff or cape (3); blackish vs pale bill (2); black bristles on nares more extensive, and black shaft streaks on face (lores to below eye) (ns1); less powdery plumage (ns1); more oblong, less circular and slightly less broad white eye-patch (mensural score: allow 1). This split is supported by molecular analysis (Wenner, Russello & Wright 2012).

The IOC note on this issue is: "Northern Mealy Amazon is split from [Southern] Mealy Amazon (Wenner et al. 2012; HBW Alive)."

As suggested above, this proposed split is largely based on slight differences in plumage coloration between the Northern Mealy Parrot and Southern Mealy Parrot groups with support from population genetic data. The NACC and SACC have not yet considered these data in voting on species limits within the Mealy Parrot complex.

New Information:

Morphology:

This is not new information per se, but rather a synopsis of phenotypic differences between the Northern Mealy Parrot and Southern Mealy Parrot groups.

Ridgway (1916) determined that phenotypic variation between Central and South American lineages was clinal (Table 1). Although we assume that Ridgway was referring to the morphometric measurements that appear directly above his statement about intergradation, it’s not absolutely clear whether he was referring to morphometrics, color, or some other aspect of phenotype.

Table 1: From Ridgway (1916), morphometric measurements of A. f. inornata and statement beneath stating that intergradation occurs between A. f. farinosa and A. f. virenticeps. We are unclear based on the placement of this statement what characters Ridgway (1916) was referring to, but believe the statement was in reference to morphometric characters.

In contrast, Wetmore (1968) noted that farinosa (inornata from Panama) averaged larger in wing, tail, culmen (from cere), and tarsus length (Table 2). As mentioned above, the HBW split was based on the minor plumage differences summarized here. Bill color differs between the two, being pale in southern and blackish in northern. Crown color is blue or suffused blue in Northern Mealy Parrot while Southern Mealy Parrot lacks blue in the crown. Although Southern Mealy Parrots tend to show yellow in their crowns more often, some Northern Mealy Parrots also have yellow in their crowns. Northern Mealy Parrots also typically have more heavy scaling on their nape. Other differences include a yellow lower carpal edge in Northern Mealy Parrot, whereas this is red, yellow, or a combination of both in Southern Mealy Parrot. Also, Northern Mealy Parrots tend to have more extensive bristles on nares and shaft streaks on the face, less “powdery” plumage, and a more oblong and narrower white eye patch compared to Southern Mealy Parrots.

Table 2. Wetmore (1968) measurements

	wing	tail	Culmen from cere	tarsus
virenticeps male (n=9)	229.5 mm	122.9 mm	34.6 mm	28.1 mm
virenticeps female (n=8)	225.4 mm	123.7 mm	34.6 mm	28.1 mm
inornata male (n=10)	235 mm	131.7 mm	36.3 mm	29.3 mm
inornata female (n=8)	233.8 mm	132.7 mm	36.4 mm	28.6 mm

Below are Macaulay Library photos showing variation in some of these features, especially bill, crown, and eye ring.

Northern

https://macaulaylibrary.org/asset/433053151

https://macaulaylibrary.org/asset/439969671

https://macaulaylibrary.org/asset/465517171

https://macaulaylibrary.org/asset/432513681

https://macaulaylibrary.org/asset/417454551

Southern

https://macaulaylibrary.org/asset/364752721

https://macaulaylibrary.org/asset/422619441

https://macaulaylibrary.org/asset/364752631

https://macaulaylibrary.org/asset/406771601

Population genetics:

Wenner et al. (2012) sequenced two mtDNA gene regions (1,157 bp of Cyt b + COI combined) and two nuDNA introns (1,145 bp of TGFB2 + TROP combined) to examine phylogenetic structure among the five recognized subspecies of A. farinosa (Fig. 1). Hellmich et al. (2021) expanded on this study to include samples of the geographically disjunct Atlantic Forest population of the nominate A. f. farinosa. Aside from the addition of the Atlantic Forest population, the two data sets are identical. Although both nuDNA and mtDNA were included in these studies, the sampling matrix is incomplete such that multiple individuals are missing data from one or more loci or gene regions. Additionally, the number of parsimony-informative sites in the mtDNA data set (n = 96) was far more than the nuDNA data set (n = 5), such that the concatenated / combined phylogenetic data sets are largely driven by information contained in the mtDNA genome.

Wenner et al. (2012) recovered 16 cyt-b haplotypes with 28 sequence differences between Northern Mealy Parrot and Southern Mealy Parrot clades (Fig. 2). This corresponded to mtDNA distances of 3.5–5.4% between the two clades, which translates to an approximate divergence time of 1.75–2.7 mya during the late Pliocene to early Pleistocene.

Figure 1: Sampling localities and ranges of currently recognized subspecies. Red “X” indicates approximate locality of additional samples from the Atlantic Forest of Brazil that were included by Hellmich et al. (2021).

Figure 2: (Left panel) Median joining haplotype network based on CytB data. Size of each circle corresponds to the number of individuals sharing that haplotype and color to each clade. (Right panel) Median joining haplotype network based on 4 mitochondrial genes (CytB, ND2, 12S, 16S). Size of each circle corresponds to the number of individuals sharing that haplotype and color to each clade. Ticks on each branch represent the number of sequence differences between each haplotype.

Using their combined data set of mtDNA + nuDNA, Wenner et al. (2012) recovered reciprocal monophyly and a deep phylogenetic split between the Northern Mealy Parrot (virenticeps and guatemalae) and the Southern Mealy Parrot (farinosa, inornata, and chapmani). The nuDNA tree with the highest maximum likelihood score had very low bootstrap support for all of the nodes within the Mealy Parrot complex, essentially producing a polytomy (Fig. 3). Additional sampling of the Atlantic Forest population by Hellmich et al. (2021) recovered the same topology, and found that the Atlantic Forest population formed a monophyletic group (Fig. 4).

Figure 3: Phylogenies from combined mtDNA + nuDNA (left) and nuDNA alone (right) of Mealy Parrots from Wenner et al. (2012). Posterior probabilities are shown above each node while maximum likelihood bootstrap values are shown below, or to the side for the nuDNA alone phylogeny.

Figure 4: Maximum likelihood majority rule consensus tree (cladogram without branch lengths) based on combined nuDNA and mtDNA from Hellmich et al. (2021) with expanded Atlantic Forest sampling. Numbers to the left of each node are bootstrap consensus values. Top-left inset is a phylogram with branch lengths included that are proportional to sequence divergence.

Vocalizations:

Hellmich et al. (2021) used 150 samples of contact calls (Fig. 5) to investigate differences in call structure across the 5 subspecies and the Brazilian Atlantic forest populations. They found that variation within each subspecies was as great as between subspecies with substantial overlap in acoustic principal component space among clades (Fig. 6). They also found no correlation between genetic differentiation and vocal differentiation among clades.

Figure 5: Map of vocal and genetic sampling locations from Hellmich et al. (2021). Spectrograms of representative calls from each clade are shown at their corresponding recording location. Genetic samples from the Wenner et al. (2012) study are indicated by grey-outlined triangles on the map. The location of the new genetic samples included in Hellmich et al. (2021) is indicated by the purple-outlined triangle.

Figure 6: Acoustic variation in call data. Plots of acoustic variation in contact calls based on principal components analysis of 27 call measures (left) and a multidimensional scaling of spectrogram cross-correlation values (right). The points represent individual calls, and the polygons represent the total area occupied by each clade’s set of calls in acoustic space. This is Figure 4 from Hellmich et al. (2021).

Recommendation:

Phenotypic differences between these groups are very slight—being limited to a few plumage characters that are not diagnostic—and may be clinal through the Isthmus of Panama. Vocalizations are variable throughout the complex and do not differ consistently between northern and southern groups. Although there is substantial mtDNA divergence (3.5–5.4%), no shared haplotypes, and reciprocal monophyly between the northern and southern groups, there is still a lot of uncertainty regarding contact zone dynamics. Most authorities state that the northern and southern groups are allopatric, but the evidence for this is unclear, and the distance between them is also unknown. Based on eBird records, the two groups appear to be separated by a narrow gap (~50 km) in central Panama, but current sampling for genetic analyses from the putative contact zone is sparse, and these subspecies can be difficult to distinguish in the field. Thus, the contact zone remains largely uncharacterized, both in terms of phenotypic and genetic differentiation. Mealy Parrots have also been commonly held in captivity throughout the region, both currently and historically by indigenous communities, which has increased opportunities for escapees to come into contact.

Taken together, we feel that although there is considerable evidence for cryptic speciation based on mtDNA divergence, the small amount of nuDNA is largely uninformative and does not recover the same pattern of deep reciprocal monophyly between Northern and Southern Mealy Parrots. Furthermore, the phenotypic differences are slight compared to other Amazona sister species, and the potential for hybridization in the contact zone remains unstudied. Acting conservatively, we therefore feel that Northern and Southern Mealy Parrots should not be split.

We recommend a NO vote on this proposal.

Literature Cited:

Clements, J., T. Schulenberg, S. Billerman, T. Fredericks, J. Gerbracht, D. Lepage, B. Sullivan, and C. Wood (2021). The eBird/Clements checklist of Birds of the World: v2021. Cornell Lab of Ornithology.

Hellmich, D. L., A. B. S. Saidenberg, and T. F. Wright (2021). Genetic, but not behavioral, evidence supports the distinctiveness of the Mealy Amazon Parrot in the Brazilian Atlantic Forest. Diversity 13:273.

Ridgway, R (1916). The birds of North and Middle America. Part 7. Bulletin U.S National Museum 50.

Wenner, T. J., M. A. Russello, and T. F. Wright (2012). Cryptic species in a Neotropical parrot: genetic variation within the Amazona farinosa species complex and its conservation implications. Conservation Genetics 13:1427–1432.

Wetmore, A. (1968). The Birds of the Republic of Panama. Part 2. Columbidae (Pigeons) to Picidae (Woodpeckers). Smithsonian Miscellaneous Collections 150. Smithsonian Institution, Washington, D.C.

David Vander Pluym and Nicholas A. Mason, June 2024

_______________________________________________________________________________________________________________________________________________

Comments from Remsen: “NO, as I voted in the NACC proposal, and for all the reasons given in the proposal. There’s a potential contact zone in Panama, and that needs to be adequately characterized.”

Comments from Areta: “NO: the Panama area should be rigorously screened, phenotypic distinctions are minor, the discordance between mt and nuc DNA suggests caution, and the vocalizations (that tend to be well-marked in Amazona species) appear as a single cluster in the quantitative vocal analyses.”

Comments from Robbins: “NO, for all the reasons detailed in the proposal.”

Comments from Jaramillo: “YES – I will be the outlier. Back when I used to see elements of this group on tours, I was always surprised that the Central American birds just seemed quite different to me from the South American birds. That is not all that useful, but obviously part of why I am more permissive with this topic. The DNA data tells a partial story, but there is a signal there. I realize that most will likely vote this down, but I am persuaded. I do wish that sometimes we had some information from the parrot breeder people, I bet they see similarities and differences between populations that we do not, and with the right context, these could be useful in making some of these decisions.”

Comments from Stiles: “NO. As noted by Van and others, the central Panama area would be a contact zone critical for separating the northern and South American groups: here, genetic and phenotypic information are needed: are they parapatric, or do they intergrade?”

Comments from Lane: “NO to splitting the two groups of A. farinosa. The evidence provided in the proposal suggests that they are not sufficiently distinctive with respect to one another to warrant a split. Further, I am not sure what to do with Amazona voices. In some situations, particularly when in syntopy, they are clearly very important characters for species recognition. But when dealing with populations at several points within a species’ or taxon group’s range, I have been seeing some serious issues that (what I assume is) dialect formation can introduce. Obviously, parrots can and do learn their vocabulary, but we largely accept that there are common vocalizations (flight calls, particularly) that are fairly uniform over the majority of the distribution of any given species (using the term loosely here). But, for example, I first learned the voice of Amazona autumnalis from birds in Tamaulipas, Mexico, which sound like this (https://xeno-canto.org/28759; a very distinctive “CHEE-colek”), and then encountered them in Belize, where they sounded different (https://xeno-canto.org/28345; a shorter “TOE-tick” that still has a distinctly bisyllabic structure), but I thought I still detected enough similarity that it didn’t really faze me. Then, a few years ago, I encountered the species on the Gulf slope of Oaxaca (intermediate between the previous two sites), and was blown away by how different they sounded there (https://macaulaylibrary.org/asset/79933521, which lacked any bisyllabic construction at all)! In fact, I was misidentifying those birds as A. farinosa guatemalae until I was able to see the source! I should note that all three populations are considered to fall under the same, nominate, subspecies of A. autumnalis. I realize that this proposal is about A. farinosa and not A. autumnalis, but since these are congeners, I think it may be useful to infer from that species that the fact that the voices within the populations of the A. farinosa complex don’t match genetic populations may not be at all surprising here.”

Comments from Claramunt: “NO. Only mitochondrial DNA suggests the split; other types of data are ambiguous or not diagnostic.”