Human activities facilitate hybridization in Allen’s Hummingbird

Genetic analyses uncover a hybrid population in southern California.

In 1966, ornithologists recorded breeding Allen’s Hummingbirds (Selasphorus sasin) on the Palos Verdes Peninsula in Los Angeles County (California). Morphological measurements and analyses of ringing data suggested that these hummingbirds belong to the subspecies sedentarius, which can be found on the Channel Islands in southern California. The move to the mainland might have been facilitated by human activities, such as nectar feeders and ornamental plants in gardens. However, a recent study in the journal Conservation Genetics shows that the situation is more complex than the population expansion of one subspecies.

Three Clusters

Between 2004 and 2016, Braden Godwin and his colleagues collected samples from Allen’s Hummingbirds across California. Genetic analyses revealed three main clusters: the northern California mainland, the Channel Islands and the newly established urban population. The northern cluster contains members of the migratory species sasin, whereas the Channel Islands cluster corresponds to the sedentary subspecies sedentarius. But what about the urban population? It is clearly a separate genetic cluster (see figure below), but does it contain sedentarius individuals as suggested by previous studies?

The answer is not a straightforward yes or no. The population has genetic signatures from sedentarius, but also from the other subspecies. In other words, it is a hybrid population. Indeed, the researchers conclude that “our population genomic analyses indicate that S. sasin hummingbirds inhabiting mainland southern California are a hybrid population resulting from admixture between S. s. sasin and S. s. sedentarius.”

A principal component analysis reveals three genetic clusters that correspond to northern California (NC), the Channel Islands (CI) and the newly established urban population (SC). The latter one turned our to be a hybrid population. From Godwin et al. (2020) Conservation Genetics

Good or bad?

The formation of this hybrid population is a nice example of human-induced hybridization. The nectar feeders and ornamental flowers in Californian gardens attract hummingbirds from different subspecies that consequently interbreed. A few months ago, I published a review paper on this topic in the journal Evolutionary Applications. In that study I touched upon the benefits and dangers of human-induced hybridization: “While the interbreeding of different populations or species can have detrimental effects, such as genetic extinction, it can be beneficial in terms of adaptive introgression or an increase in genetic diversity.” From a conservation point of view, we are thus faced with a difficult dilemma: should we prevent potential genetic extinction with conservation measures (e.g., culling hybrids) or should we not intervene to provide the opportunity for adaptive introgression and an increase in genetic diversity?

The same question applies to the hummingbird situation. And although the authors mention the issues of genetic swamping and extinction, they focus on the positive side of this hybridization event.

The southern California hybrid zone could act as a conservation reservoir for S. s. sasin alleles in the face of potentially declining abundance and potential maladaptive alleles introduced by S. rufus [i.e. rufous hummingbird] or as a beneficial introduction of new alleles from S. s. sedentarius to potentially help the declining S. s. sasin subspecies. The expanding population of Allen’s Hummingbirds in southern California could be interpreted as a positive development as the overall population of the species appears to be increasing and alleles specific to S. s. sasin are remaining in the subspecies complex

Hybridization is not always bad news for conservation.


Godwin, B. L., LaCava, M. E., Mendelsohn, B., Gagne, R. B., Gustafson, K. D., Stowell, S. M. L., Engilis Jr., A., Tell, L. A. & Ernest, H. B. (2020). Novel hybrid finds a peri-urban niche: Allen’s Hummingbirds in southern California. Conservation Genetics21(6), 989-998.

Featured image: Allen’s Hummingbird (Selasphorus sasin) © M. Shattock | Wikimedia Commons

This paper has been added to the Apodiformes page.

A scoring scheme to assess the reliability of bird hybrids

I explain this new scoring system and apply it to Tinamous.

In 2015, I published my very first paper in the journal Ibis where I introduced the Avian Hybrids Project (the website you are reading right now). Apart from launching this website, I also estimated the percentage of bird hybrids. Using the Handbook of Avian Hybrids of the World by Eugene McCarthy and the online Serge Dumont Bird Hybrids Database, it turned out that ca. 16% of bird species has hybridized in the wild. Although these sources provide supporting references for the documented hybrids, the reliability of these references has not been systematically assessed. That is why I decided to develop a simple scoring scheme to determine the reliability of the data sources supporting hybrid records. This system was recently published in the journal Ornithology Research.

The scoring scheme is based on three criteria: (1) the observation of a putative hybrid with photographic evidence or a detailed description, (2) thorough morphological analyses in which the putative hybrid is compared with potential parental species, and (3) genetic analyses of the putative hybrid with reference material from potential parental species. The first criterion encompasses sightings in online databases, such as eBird, as well as scientific papers describing the observation of a presumed hybrid. The second criterion requires analyses in which the hybrid specimen is quantitatively and/or qualitatively compared with the probable parental species. And the third criterion includes genetic analyses that confirm the hybrid origin of an individual or report recent gene flow between hybridizing species.

To express the varying levels of confidence that each of these criteria provide, they will be weight differently in the final score for a putative hybrid, namely one point for an observation, two for a morphological analysis, and three for a genetic test. The final tally of these three criteria (ranging from 0 to 6 points) will indicate the level of confidence for a particular hybrid combination. Got it? Let’s apply it to some bird hybrids!

Tinamou Hybrids

This scoring system developed quite naturally while I was writing a short review on tinamou hybrids. To structure my thinking and keep track of the papers I was reading, I constructed a table with the three criteria explained above. The tinamous (Neotropical family Tinamidae) turned out to be an ideal group to test the scoring scheme. Only a handful of hybrids have been reported and the evidence supporting them is easy to interpret. My analysis revealed one well-documented case and three doubtful records that require further investigation. All information is summarized in the table below, providing a quick overview of the reliability of potential tinamou hybrids.

Crypturellus boucardi × cinnamomeus
In the Handbook of Avian Hybrids of the World, McCarthy (2006) mentioned one well-documented case of hybridization in tinamous, namely slaty-breasted tinamou (Crypturellus boucardi) and thicket tinamou (C. cinnamomeus) which interbreed in Honduras. Two hybrid specimens were briefly described in Monroe Jr. (1965).

Eudromia elegans × formosa
In his zoogeographic analysis of the South American Chaco avifauna, Short (1975) suggested that elegant crested tinamou (Eudromia elegans) and Quebracho crested tinamou (E. formosa) may be in contact but indicated that their interactions are unknown.

Nothura boraquira × Nothoprocta cinerascens
Short (1976) discussed the morphological resemblance between white-bellied nothura (Nothura boraquira) and brushland tinamou (Nothoprocta cinerascens), but offered no clear evidence for hybridization.

Rhynchotus rufescens × maculicollis
Finally, when describing the distribution of the red-winged tinamou (Rhynchotus rufescens), Short (1975) mentioned that the subspecies pallescens seems to intergrade with two other subspecies (rufescens and maculicollis). The latter one has recently been elevated to species rank based on a distinctive song. However, these species do not seem to overlap in distribution and occur on different elevation levels.

Table summarizing the evidence supporting different tinamou hybrids. From: Ottenburghs (2021) Ornithology Research.

Additional Evidence

In addition to the references supporting certain hybrid records, I also took into account the distribution of the parental species and their divergence time. If the putative parental species do not overlap in distribution, the hybrid record can be considered less reliable (although you have to consider escaped or introduced individuals). Similarly, an extremely old divergence time between the parental species can cast doubt on a proposed hybrid.

For the tinamou hybrids, the information on distribution and divergence times did not add much. With the exception of red-winged tinamou x huayco tinamou, all parental species overlapped in distribution (see maps below). And I could not find a reliable estimate of divergence time due to a lack of genetic studies. I had to rely on a maximum date that corresponds to the split between the subfamilies Nothurinae and Tinaminae (17 million years ago). This date is within the range of divergence times at which bird species still have the ability to hybridize (on average 21 million years).

Distribution maps of putative parental species for four hybrid combinations. From: Ottenburghs (2021) Ornithology Research.

Picking the Journal

A final note on the journal where this study appeared. I decided to submit my manuscript to Ornithology Research (previously known as Revista Brasileira de Ornitologia | Brazilian Journal of Ornithology), because they focus on Neotropical species. Moreover, it was also a small act of rebellion against the current academic climate. The scientific world is so focused on developing your cv by publishing in high-impact journals and collecting citations. Smaller journals are often neglected, even though they reach an important audience. I hope my paper will inspire local researchers to work on avian hybrids.


Ottenburghs, J. (2021) An evidence-based overview of hybridization in Tinamous. Ornithology Research. Early View.

Featured image: Pale-browed Tinamou (Crypturellus transfasciatus) © Nick Athanas | Flickr

A gene within a supergene: An estrogen receptor shapes the behavior of White-throated Sparrow morphs

Expression levels of the estrogen receptor determine aggressive behavior in these songbirds.

White-throated sparrows (Zonotrichia albicollis) come in two distinct morphs: the white-striped (WS) and the tan-striped morph (TS). These morphs do not only differ in their plumage patterns, but also in behavior, such as the degree of parental care that they provide (which I discussed in this blog post). The differences between these morphs have a solid genetic basis. Already in 1966, Thorneycroft identified a chromosomal rearrangement that explains the occurrence of the two white-throated sparrow morphs. Recent molecular work showed that this rearrangement is an inversion (i.e. a flipped section of DNA, more on inversions in this blog post), giving rise to a so-called supergene which links numerous genes that influence the morphology and behavior of these birds. Tan morphs have the same version of the supergene (i.e. they are homozygous) whereas white-striped morphs have two different versions (i.e. they are heterozygous).

Knowing that a supergene underlies the different morphs is only the first step. Now, we can zoom in on the contents of this supergene and determine how these linked genes work together in shaping the plumage and behavior of white-throated sparrows. A recent study in the journal PNAS performed some clever experiments to understand the role of one particular gene.

The different morphs of the White-throated Sparrow (A and B) prefer to mate with the opposite morph (see percentages in C). The differences between the morphs can be traced back to a super-gene (D). From: Campagna (2016) Current Biology.

Estrogen Receptor

As mentioned above, the white-striped and the tan-striped morphs behave differently. Studies in wild populations found that WS birds are more aggressive compared to TS birds when defending their territories. Given that territorial aggression in songbirds has been linked to steroid hormones, it makes sense to search for genes that are involved in the production or processing of these hormones. Interestingly, one of the genes (ESR1) in the supergene codes for an estrogen-receptor. Moreover, this gene comes in two different versions (ZAL2 and ZAL2m) that follow the genetic patterns underlying the two morphs. Tan morphs have the same version of gene (two times ZAL2) whereas white-striped morphs have two different versions (ZAL2 and ZAL2m). Sounds like the perfect candidate gene!

The researchers quantified the level of aggression of different birds in several behavioral trials. Next, they measured the expression levels of the different ESR1-versions in certain brain areas. They summarized their findings as follows: “the degree to which a bird engaged in territorial aggression, which was markedly higher in the WS birds than in the TS birds, was predicted by the relative expression of the ZAL2m allele.” In another experiment, the researchers knocked down the expression of the ESR1-gene in certain brain areas and assessed the aggression of the birds. This experiment revealed that the more aggressive birds became less aggressive when the ESR1-gene was turned off.

One allele (ZAL2m, in red) was more highly expressed in white-striped morphs, and correlated with aggressive behavior (measures as the number of songs per 10 minutes). These results support a central role for the estrogen receptor in shaping the behavior of the morphs. From: Merritt et al. (2020) PNAS.

Gene Network

These findings provides direct evidence that the estrogen-receptor plays a crucial role in determining the behavior of these morphs. However, it remains to be determined how it actually works. This receptor is a transcription factor that interactions with a large number of other proteins as well as with numerous regulatory elements. A previous study reported that ESR1 lies within an interconnected module of 157 genes that are differentially expressed between the morphs. Of these 157 genes, 115 are located in the supergene. More experimental work is needed to disentangle this complex web. Slowly but steadily we are getting closer to the genetic underpinnings of these intriguing morphs.


Merritt, J. R., Grogan, K. E., Zinzow-Kramer, W. M., Sun, D., Ortlund, E. A., Soojin, V. Y., & Maney, D. L. (2020). A supergene-linked estrogen receptor drives alternative phenotypes in a polymorphic songbird. Proceedings of the National Academy of Sciences117(35), 21673-21680.

Featured image: White-throated sparrows (Zonotrichia albicollis) © Cephas | Wikimedia Commons

The phylogeographic story of a Manakin and a Bamboo Tyrant

Ecology explains the genetic differences in two Atlantic Forest species.

One of my favorite science stories is the discovery of the neutrino by the Austrian physicist Wolfgang Pauli. During an experiment, he found that energy appeared not to have been conserved. Reluctant to give up the universal idea of conservation of energy, Pauli developed an explanation. He speculated that the missing energy was carried off by a new particle. Next, he developed a mathematical model to predict certain properties of this new particle, so that its existence could be verified. Twenty-five years later this new particle was found and is now a well-established member of particle physics, even if still hard to detect. This story illustrates the power of formulating explanations and hypotheses which can consequently be tested with new experiments and observations.

A recent study in the journal Molecular Phylogenetics and Evolution took a similar approach when examining the genetic population structure of of two species in the Montane Atlantic Forest: the blue manakin (Chiroxiphia caudata) and the drab-breasted bamboo tyrant (Hemitriccus diops). What explanations could account for the similarities and differences in population structure between these species?

Last Glacial Maximum

Reconstructing the past distributions of these species revealed that they responded similarly to the climatic conditions during the Last Glacial Maximum (about 20,000 years ago). At this time, the Montane Atlantic Forest covered a larger area of South America, allowing both species to expand their range. This finding was also supported by the genetic analyses where several statistics (Fu’s Fs and R2) indicated population expansion in both species.

Despite these similarities, additional genetic analyses of the mitochondrial ND2 gene revealed some striking differences. The blue manakin did not display a clear phylogeographic structure, whereas the drab-breasted bamboo tyrant showed a phylogeographic break near the Doce River. What could explain these differences?

The blue manakin did not show any phylogeographic study, while the drab-breasted bamboo tyrant showed a clear phylogeographic break. From: da Silva Ribeiro et al. (2020) Molecular Phylogenetics and Evolution.

Ecological Differences

The researchers discuss several explanations for this phylogeographic incongruence. A first possibility is that there has been more gene flow between several blue manakin populations, preventing the build-up of genetic differences. This higher exchange of individuals (and genes) between populations could relate to the diet and mating system of this species. The blue manakin is a frugivore. Given that fruit is a more ephemeral resource in time and space, frugivorous species are expected to travel large distances to find food. In addition, the blue manakin exhibits lekking behavior in which females visit several locations where multiple males display (see video below). Again, lekking species are expected to disperse over larger distances during their visits. The drab-breasted bamboo tyrant, on the other hand, is a insectivore (plenty of those around) and has a territorial mating system (no need to travel far). These characteristics might lead to less gene flow between populations and the accumulation of genetic differentiation.

A second explanation concerns the different generation times of both species. Female manakins start breeding when they are 2–3 years old and have a generation time of 4.9 years. This is roughly three times larger than the average generation time of non-lekking passerine birds (ca. 1.7 years). Hence, the authors suggest that “the shorter generation time of H. diops could have favored the accumulation of differences between geographically isolated populations.”

The complex courtship ritual at a manakin lekking spot.

More Research

Similar to Pauli postulating the existence of a new particle, the formulation of all these explanations is just the first step in this scientific endeavor. Now, the researchers will need to dig deeper and test these explanations with other analytical tools and new datasets. Indeed, they mention that “the incongruent population structure pattern shown in our comparative study indicates that life history and ecological traits can be important in diversification processes. Further investigations of these traits are needed to clarify their micro-evolutionary role.” Whether you are a physicist or a biologist, further investigations are the way forward.


da Silva Ribeiro, T., Batalha-Filho, H., Silveira, L. F., Miyaki, C. Y., & Maldonado-Coelho, M. (2020). Life history and ecology might explain incongruent population structure in two co-distributed montane bird species of the Atlantic Forest. Molecular Phylogenetics and Evolution153, 106925.

Featured image: Blue Manakin (Chiroxiphia caudata) © Dave Curtis | Flickr

How the Common Starling conquered Australia

Genetic population structure is mainly shaped by distance effects.

European settlers introduced several animals to Australia for transportation or farming purposes, while others were brought in as pets or for hunting. From the 1850s onward, Common Starlings (Sturnus vulgaris) were released in several locations, such as Melbourne, Brisbane and Adelaide. Although they lost their ability to migrate, these birds managed to get a foothold and spread across the island. The population dynamics during this invasion shaped the current genetic make-up of Common Starlings in Australia.

It is possible that the introduced populations already differed genetically and diverged even more during the invasion. This divergence can follow a neutral path of isolation-by-distance where neighboring populations continue to exchange gene while distant ones accumulate genetic differences. Or perhaps different populations adapted to local conditions in distinct environments, resulting in isolation-by-environment. A recent study in the journal Molecular Ecology tested these patterns using 568 starling samples from 24 localities across Australia.

Two Clusters

The genetic analyses uncovered two main clusters that correspond to southern Australia (from Western Australia to Tasmania) and northern Australia (New South Wales and southern Queensland). These two groups are separated by an extremely arid region between Victoria and New South Wales. Interestingly, the Great Dividing Range – a mountain range in the east – did not have a big effect on the population genetic structure, suggesting that Common Starlings can cross this mountainous area. If you are not familiar with the geography of Australia, you can check out the map below.

A match between introduction site and genetic ancestry indicated that the genetic make-up of the founder populations partly influenced the present-day patterns. However, the main process shaping the genetic population structure of the Common Starling appears to be isolation-by-distance. As these birds spread across Australia, gene flow between distant populations diminished, giving rise to the correlation between genetic and geographical distance we see today.

Figure a given an overview of the introduction sites (blue dots) and sightings of Common Starlings (grey dots) on eBird. Figure b shows the genetic population structure of the Common Starling, revealing two main groups (red and blue outline). The Great Dividing Range is indicated with a dark blue line. From: Stuart et al. (2021) Molecular Ecology.

Candidate Genes

The authors reported that isolation-by-environment was not significantly correlated with genetic differentiation. This finding can be explained by phenotypic plasticity (i.e. one genotype produces more than one phenotype when exposed to different environments) or by local adaptation driven by a small number of genetic loci. The latter explanation is supported by further analyses in which the researchers tested for genetic variants under positive selection. This search uncovered “several hundred unique candidate loci (375) for local adaptation” and “a total of 25 proteins were identified”. These candidate loci were associated with differences in aridity, precipitation and temperature, suggesting that these environmental factors may be important in driving adaptive variation across the Common Starling’s invasive range.

The paper discusses a few of these candidate genes, such as CACNA1C which plays a role in behavioral plasticity (an important trait for an invading species). However, I always remain reluctant to tell “Just So Stories“. The uncovered genes are interesting starting points for further research, but not the final story. Moreover, the authors also write that “many of the loci were reported for only one of the exploratory methods, and all the annotated proteins identified were unique to their identification method.” More analyses are thus needed to understand the detailed genetic consequences of the Common Starling’s conquest across Australia.

Several genetic loci correlated with particular environmental factors, making them interesting candidates for further research. From: Stuart et al. (2021) Molecular Ecology.


Stuart, K. C., Cardilini, A. P., Cassey, P., Richardson, M. F., Sherwin, W. B., Rollins, L. A., & Sherman, C. D. (2021). Signatures of selection in a recent invasion reveal adaptive divergence in a highly vagile invasive species. Molecular Ecology30(6), 1419-1434.

Featured image: Common Starling (Sturnus vulgaris) © Marie-Lan Nguyen | Wikimedia Commons

The social network of hybrid quails

Recent study reveals a dynamic web of social interactions in a hybrid quail population.

In Southern California, Gambel’s Quail (Callipepla gambelii) and California Quail (C. californica) come into contact and interbreed. This hybrid zone has been extensively described from a morphological, behavioral and genetic perspective by Jennifer Gee in the early 2000s (see the Galliformes page for more details). Recently, David Zonana and his colleagues – including Jennifer Gee – provided another perspective on this hybrid zone by applying a social network approach. Using radio-frequency identification tags, they managed to obtain a detailed picture of the social interactions between individual birds. I have covered the results from a previous study in another blog post. The final paragraph of that post nicely summarizes the main findings:

Quails pair up with individuals that share particular plumage patterns with them. Outside the contact zone, this strategy works perfectly because they only encounter members of their own species. In the contact zone, however, they run into quails from another species. But instead of focusing in the species-specific differences, they keep using the shared plumage traits to pick their partner. And voila, hybrids!

A new study in the Journal of Animal Ecology builds on these results by adding another layer of complexity: genetics. What do these social interactions mean for patterns of gene flow? Do quails with similar genetic ancestry flock together or do they ignore relatives to avoid inbreeding?

Genetic Ancestry

Using genomic data, the researchers calculated the percentage of shared genetic ancestry between individual birds and incorporated this information in the network analyses. It turned out that social associations were structured randomly with respect to genetic ancestry in all networks. They concluded that these random patterns of assortment “provide further evidence that behavioral reproductive isolation is likely weak within the hybrid zone.”

However, the genetic data uncovered some hidden patterns in the social network. The genetic parentage analyses revealed offspring from three pairs that were not strongly connected within the networks. These offspring might be the outcome of extra-pair copulations that can happen quickly (and are thus not picked up by the radio tags). Such extra-pair copulations can affect the direction of gene flow between hybridizing species. Whether this is the case in Gambel’s and California Quail remains to be investigated.

The researchers incorporated information about genetic ancestry (figures a and b) into the social network (figure c). From: Zonana et al. (2020) Journal of Animal Ecology.

Network Rewiring

In addition to the genetic analyses, the researchers also studied the change in network structure over time. The rewiring of a network can occurs in two main ways: (1) changes in relationships or (2) changes in members of the network. The first way occurs when individuals associate with different birds at different times during the year, while the second way concerns members leaving (e.g., through migration or death) or new members joining the network. With regard to the quails, the rewiring of the network primarily involved changes in relationships between different members. It seems that there is a core breeding population with relatively stable membership but highly dynamic patterns of association.

Most social networks are thus a snapshot in time and can change dramatically throughout the year. This insight has important consequences for the study of hybrid zones, where mating is often considered random. However, changing social interactions can affect mating opportunities and potentially the incidence of hybridization. These dynamics will obviously differ between different hybrid zones, but need to be taken into account. More social network studies in hybrid zones, please.

The social network in the hybrid zone between Gambel’s and California Quail changed over time. From: Zonana et al. (2020) Journal of Animal Ecology.


Zonana, D. M., Gee, J. M., Breed, M. D., & Doak, D. F. (2021). Dynamic shifts in social network structure and composition within a breeding hybrid population. Journal of Animal Ecology90(1), 197-211.

Featured image: A Gambel’s Quail (Callipepla gambelii) © Alan D. Wilson | NaturesPicsOnline

This paper has been added to the Galliformes page.

Why are some parrots more colorful than others?

It turns out that size matters.

The colors of parrots never disappoint. Think of the red, yellow and blue plumage of the scarlet macaw (Ara macao), or the bright pink crest of the Major Mitchell’s Cockatoo (Cacatua leadbeateri). A quick overview of the ca. 400 species of parrots (order Psittaciformes) reveals a huge variety of color combinations. What explains this wealth of colors? And why are some species more colorful than others? One important factor is that parrots can synthesize their own pigments for the colors red and yellow (i.e. psittacofulvins), while other bird species need to extract such pigments from their diet (mostly carotenoids). This might allow parrots to deposit more of these pigments in their feathers. A second reason relates to the breeding behavior of parrots. Most species make their nests in cavities where they are safe from predators. There is no need for camouflage, so parrots can develop colorful plumage patterns.

These are just two possible explanations for the colorfulness of parrots. A recent study in the Journal of Evolutionary Biology dug a little deeper and uncovered several other factors that contribute to the amazing variety of colors in parrots.

Big vs. Small

Luisana Carballo and her colleagues quantified the color patterns of 398 parrot species using the Handbook of Birds of the World. Their analyses showed that phylogeny explained a large part of the variation. It seems that birds of a feather flock together on the evolutionary tree. After controlling for the shared evolutionary history, the remaining factors explained up to 15 percent of the variation in colors. Let’s have a look at the main patterns.

First, larger parrot species tend to have more elaborate colors. This finding probably relates to the lower predation pressure on larger species. Most predators are no threat to large parrots, so they can afford to draw the attention with their colorful plumage. In addition, larger parrot species form long-lasting pair bonds which involves mutual mate choice. Both partners need to be satisfied with their choice if they are going to spend a lifetime together (which can be up to 100 years…). So, we can expect both sexes to be equally ornamented due to this mutual mate choice, similar to other tropical species (see for example this study).

Interestingly, although smaller species sport less elaborate colors, they show larger color differences between males and females (i.e. higher levels of sexual dichromatism). These patterns suggest stronger sexual selection in the smaller species. This idea is supported by a study on sperm morphology that reported a positive correlation between the length of sperm cells (a proxy for the strength of sperm competition) and the degree of sexual dichromatism in parrots. High levels of sperm competition also suggest that females will mate with multiple partners, but the frequency of these extra-pair copulations remains to be determined.

An overview of the data on color patterns in parrots. The figures illustrate the variation in (a) dark vs. light, (b) green vs. red, and (c) blue vs. yellow. Figure (d) show the spread of colors for males and females. From: Carballo et al. (2020) Journal of Evolutionary Biology.

Gloger’s Rule

Another peculiar pattern is that parrots tend to be darker in wetter areas, a phenomenon known as Gloger’s Rule (which I also covered in this blog post on gulls). The exact mechanisms behind this biogeographical rule remain unclear. It could be that darker colors are more suitable for camouflage in wetter environments where more vegetation results in lower light conditions. However, this hypothesis conflicts with the elaborate colors of larger parrot species that do not rely on camouflage. A second possibility concerns degradation of feathers by bacteria. Wet conditions are conducive for the growth of feather-degrading bacteria. The deposition of the dark pigment melanin protects the plumage from these feather-munching menaces. In addition, parrots are redder in wetter environments. The red pigment psittacofulvin might also provide more protection against these bacteria.

Parrots and cockatoos with more elaborate colous have lower levels of sexual dichromatism. The colors of the phylogeny (from blue to red) indicate the degree of color elaboration, while the bars on the outside correspond to the level of sexual dichromatism. From: Carballo et al. (2020) Journal of Evolutionary Biology.

Unweaving the Rainbow

This study nicely illustrates the complex interplay of evolutionary and ecological forces in shaping the wonderful colors of parrots. It reminds me of a conversation between physicist Richard Feynman and an artist. The artist argued that science takes away the beauty of nature. Science is “unweaving the rainbow” as John Keats put it. Feynman disagrees and explains how science only add to the beauty, using the example of a flower. The same reasoning applies to the amazing wealth of colors in parrots.

I can appreciate the beauty of a flower. At the same time, I see much more about the flower than he sees. I could imagine the cells in there, the complicated actions inside, which also have a beauty. I mean it’s not just beauty at this dimension, at one centimeter; there’s also beauty at smaller dimensions, the inner structure, also the processes. The fact that the colors in the flower evolved in order to attract insects to pollinate it is interesting; it means that insects can see the color. It adds a question: does this aesthetic sense also exist in the lower forms? Why is it aesthetic? All kinds of interesting questions which the science knowledge only adds to the excitement, the mystery and the awe of a flower. It only adds. I don’t understand how it subtracts.


Carballo, L., Delhey, K., Valcu, M., & Kempenaers, B. (2020). Body size and climate as predictors of plumage colouration and sexual dichromatism in parrots. Journal of Evolutionary Biology33(11), 1543-1557.

Featured image: Scarlet Macaw (Ara macao) © Travis Isaacs | Wikimedia Commons

Gene flow is an integral part of speciation in Beringian birds

Eight bird lineages with a trans-Beringian distribution show clear signatures of past gene flow.

Most people know Beringia as the land bridge that allowed humans to travel from Siberia to North America during the Ice Ages. Indeed, throughout the Pleistocene (between 2.5 million to 11,000 years ago) cold snaps resulted in low sea levels, which led to the formation of land bridges between the two continents. When the climate warmed again, these land bridges disappeared under the rising sea levels. This cycle of exposure and inundation of central Beringia occurred at least nine times (and perhaps up to twenty times or more). Not only did this facilitate the spread of humans across the globe, it also affected the numerous bird species that can be found on both sides of the Bering Strait.

These climatic cycles have probably shaped the genetic make-up of several Beringian bird species. During periods of high sea levels, bird populations were geographically isolated and diverged genetically. When sea levels dropped and land bridges formed, some of these populations might have come into secondary contact and exchanged some genetic material. This raises the question how much gene flow (if any) occurred during these periods of contact. A recent study in the journal Molecular Ecology took a closer look at eight population pairs to answer this question.

A view of Beringia throughout time. During glacial maxima in the Pleistocene, a land bridge existed between the continents. This land bridge disappeared during warmer interglacials. From: Laughlin et al. (2020) Molecular Ecology.

Eight Pairs

Jessica McLaughlin and her colleagues focused on eight bird lineages that cover the taxonomic range from populations over subspecies to distinct species. They used ultraconserved elements (UCEs) to determine the level of genetic divergence and the amount of gene flow between the following pairs:

  • Long-tailed duck (Clangula hyemalis – populations)
  • Bluethroat (Luscinia svecica – populations)
  • Green-winged teal (Anas crecca crecca and A. c. carolinensis – subspecies)
  • Whimbrel (Numenius phaeopus variegatus and N. p. hudsonicus – subspecies)
  • Pine grosbeak (Pinicola enucleator kamschatkensis and P. e. flammula – subspecies)
  • Eurasian and American wigeon (Mareca penelope and M. americana – species)
  • Grey-tailed and wandering tattler (Tringa brevipes and T. incana – species)
  • Eurasian and black-billed magpie (Pica pica and P. hudsonia – species)

The researchers tested several demographic models to reconstruct the evolutionary history of these species. In each case, gene flow was an integral part of the divergence process. Three pairs (whimbrel, pine grosbeak and magpies) followed a scenario of divergence-with-gene flow, while the evolution of two other pairs (long-tailed duck and wigeons) was best captured by gene flow at secondary contact. For the remaining three pairs (bluethroat, green-winged teal and tattlers), the analyses could not discriminate between divergence-with-gene-flow or secondary contact.

An overview of the different models that were tested. All species pairs followed a scenario with gene flow, either divergence-with-gene-flow (c and d) or gene flow during secondary contact (e and f).

Divergence Continuum

Given that these eight pairs span the taxonomic range from populations to species, you might expect to see this reflected as a continuum of genetic divergence and gene flow. This was, however, not the case. First, the genetic divergence (measured as Fst) did not follow the taxonomic classification. Some distinct species – such as Eurasian and American wigeon – were genetically quite similar (Fst = 0.04), while some subspecies – such as whimbrel (Fst = 0.27) or pine grosbeak (Fst = 0.44) – were more genetically distinct. Taxonomy is thus not a good predictor of genetic divergence.

Second, there is no clear continuum when plotting the relationship between genetic divergence and gene flow. Instead, two distinct groups are visible (see figure below). This result follows recent theoretical work that considers speciation as a two-state system with most populations pairs clustering near the two ends of the continuum (either showing genetic small differences or full reproductive isolation). Diverging populations are moving towards the right end of this continuum, but can be pulled back to the left end when gene flow occurs. Once a certain threshold of reproductive isolation has been achieved, populations will remain on the right end of the spectrum. The Beringian birds nicely represent both sides of this continuum.

Out of curiosity, I returned to my recent paper on the evolution of taiga and tundra bean goose (which I covered in this blog post). Using whole genome re-sequencing data, we found low genetic divergence (Fst = 0.033) and high levels of gene flow (M = 13). These numbers clearly put the bean geese on the left side of the spectrum. How general these patterns are remains to be determined, but Beringia seems like the perfect place to start.

The relationship between genetic divergence (Fst) and gene flow (M) reveals two distinct groups that correspond to predictions from theoretical work. From: McLaughlin et al. (2020) Molecular Ecology.


McLaughlin, J. F., Faircloth, B. C., Glenn, T. C., & Winker, K. (2020). Divergence, gene flow, and speciation in eight lineages of trans‐Beringian birds. Molecular Ecology29(18), 3526-3542.

Featured image: Whimbrel (Numenius phaeopus) © Andreas Trepte | Avi-Fauna

Genetic patterns along the South American dry diagonal: The case of the Narrow-billed Woodcreeper

What environmental factors determine the genetic population structure of this species?

South America is known for its lush rainforests and high Andean mountain peaks. Tropical and subtropical rainforests can be found along the Amazon river and on the Atlantic coast. These regions are separated by a broad corridor of open vegetation. This so-called dry diagonal is often seen as a formidable barrier for avian rainforest species that occur on either side of it (a topic I covered in this blog post). For other bird species, however, the dry diagonal is not an obstacle but an important habitat. Indeed, numerous birds live in the several biomes that make up this open vegetation landscape, such as the Chaco, Cerrado and Caatinga.

Given that birds in open environments tend to show high dispersal capacity, we can expect gene flow between neighboring populations. Only when the distances become too large will the levels of gene flow start to decrease. This phenomenon is known as isolation-by-distance, and can easily be tested by correlating genetic divergence with geographical distance. A recent study in the Journal of Avian Biology tested this prediction in the Narrow-billed Woodcreeper (Lepidocolaptes angustirostris), a species that be found throughout the dry diagonal. Will this species show isolation-by-distance or do other factors play a role?

The expected pattern of isolation-by-distance: a positive relationship between genetic divergence and geographical distance. Here illustrated by a landrace of Thai rice. From: Pusadee et al. (2009) PNAS.

Genetic Lineages

Amanda Rocha and her colleagues collected 63 individuals of the Narrow-billed Woodcreeper, covering the main biomes within the open vegetation corridor (Chaco, Cerrado and Caatinga). The sampling effort included representatives from the two main morphological groups within this species: group angustirostris, with brown back and heavily streaked pattern in the chest and vent (containing the subspecies angustirostris, praedatus, certhiolus and hellmayri), and group bivittatus, with more rufescent in the back and unstreaked pattern below (containing the subspecies bivittatus, griseiceps, coronatus and bahiae).

Genetic analyses of the mitochondrial gene ND2 revealed five main lineages: one in Caatinga, three in Cerrado (NE_Cerrado, E_Cerrado and W_Cerrado) and one in Chaco. Hence, each genetic lineages corresponds to a particular biome in the dry diagonal. Detailed analyses of each genetic lineages indicated strong population structure and no signs of isolation-by-distance. It thus seems that other environmental factors have shaped the genetic make-up of the Narrow-billed Woodcreeper populations.

The phylogenetic tree of the Narrow-billed Woodcreeper show that different genetic lineages correspond to particular biomes within the dry diagonal. Notice that the morphological groupings and subspecies are not supported by the genetic data. From: Rocha et al. (2020) Journal of Avian Biology.

Isolated Biomes

To determine which environmental factors influenced the evolution of the Narrow-billed Woodcreeper, the researchers reconstructed the past habitat of this species during the Pleistocene (between 2.5 million and 11,000 years ago). These analyses showed that the open biomes contracted during warmer interglacial periods and expanded during cold and
dry glacial periods. These dynamics suggest that certain populations became isolated during the interglacial periods and diverged genetically. When these biomes – and the populations within them – expanded again, the heterogeneity of the landscape prevented the populations from completely mixing again, giving rise to the genetic patterns we see today. The authors summarized their findings succinctly:

All genetic lineages identified here are historically associated with one stable climatic area suggesting that diversification within L. angustirostris has been primarily influenced by Pleistocene climatic oscillations, which promoted allopatric diversification during interglacial periods. Secondarily, the heterogeneous landscape along the dry diagonal may have being limiting gene flow among the genetic lineages after contact was re-established among them.

Present and past models of geographic distributions of the Narrow-billed Woodcreeper show that the different biomes were isolated in the past. Over time, the populations within these biomes diverged genetically. From: Rocha et al. (2020) Journal of Avian Biology.

Some Taxonomy

Finally, this study also has some taxonomic implications. The phylogenetic analyses indicated that none of the genetic lineages correspond to the recognized subspecies (which are based on morphological data). This result is in line with a recent review on the plumage variation within the Narrow-billed Woodcreeper, which can thus be seen as a single species with strong genetic structure.


Rocha, A. V., Cabanne, G. S., Aleixo, A., Silveira, L. F., Tubaro, P., & Caparroz, R. (2020). Pleistocene climatic oscillations associated with landscape heterogeneity of the South American dry diagonal explains the phylogeographic structure of the narrow‐billed woodcreeper (Lepidocolaptes angustirostris, Dendrocolaptidae). Journal of Avian Biology51(9).

Featured image: Narrow-billed Woodcreeper (Lepidocolaptes angustirostris) © Evaldo Resende | Wikimedia Commons