It’s been an exciting week. Five years in the making, we’re very excited to release this new, updated, dynamic phylogeny of the world’s birds. By harnessing Open Tree of Life’s existing computing infrastructure to carefully curate and synthesize much of the published knowledge on the evolutionary relationships among birds, and uniting the result with an emerging unified global taxonomy and expert consensus on where previously phylogenetically unstudied species might plug in, Birds of the World has built a new tool that we believe will propel avian taxonomy and phylogeny towards increasing alignment and understanding. By uniting the new Phylogeny Explorer with Birds of the World content and illustrations, and by integrating with users’ personal eBird data, we have created a deeply immersive way to explore our planet’s avifauna, and our own relationship to it. It will, we believe, spark a newfound interest in the link between evolution and birdwatching. Before I dive into the new possibilities this tool unlocks, I begin with a short history, attempting to draw a direct line from the origins of human consciousness to Linnaeus to Darwin to the new Phylogeny Explorer, in a few brief paragraphs.

************************************************************************************************************
​​
The subject of taxonomy, if we think of it at all, tends to evoke images of quibbling scientists rattling off formidable sounding Latin names while pondering inscrutable insect genitalia through dissecting scopes. In truth, the concept of taxonomy has arisen in parallel with human consciousness, and anthropologists and linguists have uncovered broad similarities across folk taxonomies around the globe. A folk taxonomy is something everyone can relate to: fruits vs. vegetables, animals vs. plants. For millennia, this has worked for people who regularly communicated with one another, and while it didn’t necessarily translate across cultures and regions, it didn’t really matter either–people and their vegetables tended to stay put. Critically, these folk taxonomies tended to have an approachable number of “things” (say, a few hundred plants at most), because the entire point was to facilitate communication, and people can only remember so much. 

It wasn’t until the age of exploration, and the invention of movable type in the 15th century, that folks began working towards more comprehensive taxonomies that transcended local cultures and regions. The scale of the problem slowly began to emerge over the centuries that followed. In plants, for example, species were first lumped into genera, but as the number of genera began to accumulate, beyond that which even a specialist could easily remember, the taxonomic level of family became a more commonly used convention: when Tournefort published a list of plant genera (plural for genus) in 1700 there were 698 of them; when Linnaeus published Genera Plantarum in 1737 there were 935; in 1789 de Jussieu organized these genera (closer to 1800 at this point) into 100 easily remembered families; but, by 1920 this had expanded into around 300 recognized families. Remember though, these scientists were mostly working in the temperate zone. Little did they know what awaited them in the tropics. 

Exactly 100 years after Linnaeus published Genera Plantarum, Darwin was doodling the first rudimentary phylogenies in his private notebooks, and a decade or two later the concept of evolution finally emerged on the scene. We often get hung up on who divined this idea first out of thin air (the old Darwin vs. Wallace debate), but if you think about it, science had been building towards this for the last few centuries. Although the use of trees, or phylogenies, to illustrate the process awaited Darwin’s first sketches, people were drawing family trees (genealogies) long before this. Moreover, the very nature of taxonomies absolutely implied hierarchical clustering. So although Darwin, or Wallace, or your great aunt Martha, may have been the first to truly spell out the concept of evolution, science, led by taxonomy, had been slowly working its way there for centuries; someone was going to figure it out eventually. 

Birds, too, despite representing a mere fraction of all life on Earth, were always going to play an outsized role in our understanding of evolution. They’re colorful, obvious, and differences among them seem to inspire a never-ending array of questions. Galapagos finches and mockingbirds were a key inspiration for Darwin, while Wallace studied birds (and many other organisms) from the Amazon to the Malay Archipelago. The discovery of Archaeopteryx, soon thereafter, forced scientists to grapple with the links between modern birds and fossil dinosaurs. As with taxonomy, birds were at the forefront, charting a course towards an accelerating understanding of evolution and life’s mysteries.

In 1758, Linnaeus published the 10th edition of Systema Naturae, and in it he described 554 species of birds. Despite considering only a fraction of the globe’s species we know today (remember, biologists were only just exploring the tropics), Linnaeus’ organization was far from perfect. While many names have stood the test of time, he misclassified many others; the seven species he included in the genus Gracula, for example, represent what we now recognize as the members of no fewer than five separate taxonomic families. Others soon identified these flaws and, as new bird species were collected, sent home to museums, and named, the list and its many levels grew; the organization within the list improved, little by little, study by study, argument by heated argument. This is exactly how the process should work: as new information and new analysis is brought to bear on the questions of relationships within the tree of life, past assumptions can be questioned and revised.

Despite the clear conceptual links between evolution, phylogenies, and taxonomy, in practice these remained fairly disparate endeavors for another 130 years or so after the publication of Darwin’s Origin of the Species. Phylogeneticists worked out the details on how to construct trees from character matrices; taxonomists argued about homologous characters, and used these characters to refine avian taxonomy. Sure, this is an oversimplification, but it’s not that far from the truth. Certainly no one was trying to make a global phylogeny of all birds–scientists were still just working out how many species there were, even approximately. How would it ever be possible to make a phylogeny using variation in morphological characters if some groups lacked those characters entirely? Which characters would suffice to show us the true evolutionary relationships among all birds? Importantly, even once a study with clear bearing on taxonomy was published–for example, good evidence that a new species should be recognized, or an existing one reclassified–the actual revision required a formal taxonomic proposal and deliberation. That is, the taxonomic process was (and remains) distinct from the science that informed it.

Things changed abruptly with DNA-DNA hybridization work led by Charles Sibley and Jon Ahlquist in the 1980s. While their work was by no means the first to use molecular methods to produce phylogenies, it was the first to do so broadly across an entire class of organisms, using consistent methods, and it quickly changed the avian taxonomy discussion. At the risk of offending a few centuries worth of systematists, the entire field pivoted nearly overnight. The question was not “what can these characters tell us about the phylogeny?”, but “what can the phylogeny tell us about the evolution of these characters?” In parallel with rapidly advancing molecular methods (including both DNA sequencing and tree construction algorithms), ornithologists started pumping out molecular phylogenies for increasingly large sets of taxa. As the rate of publication of these trees increased through the 1990s and early 2000s, our understanding of avian evolutionary relationships did as well. 

Crucially, while taxonomies and phylogenies share much in common, phylogenies can capture much more information in a structured, quantitative format. For example, while the paired taxonomy and phylogeny we present here both contain 11,167 species, there are a further 11,166 internal nodes in the phylogeny, and this provides an opportunity to convey significantly more knowledge about how all these species fit together. In contrast, there are  2,673 named higher-level taxa (genera, families, and orders) in the corresponding taxonomy. Moreover, phylogenies don’t simply nest taxa within taxa, they can tell us exactly how those taxa connect to one another. Even better, using a combination of fossils and molecular clocks, phylogenies can be time-dated, meaning that phylogenies can also tell us when these taxa diverged from one another (last shared a common ancestor; note that there is significant work to do still with refining these date estimates, both in general and specifically with respect to the tree shown in the Phylogeny Explorer). In short, phylogenies are a rich information source, but they’re nothing without taxonomy.

In the 30-year sprint to sequence all the things and pin them on a tree, taxonomy has often lagged behind. Taxonomic revisions still require formal proposals. Nomenclature is critical to the taxonomic process, but it has formal sets of rules (the ICZN “Code”) that, although vital to standardization, can sometimes be sand in the gears of fast-paced progress. Phylogenies frequently contradict one another, and judgement calls must be made: how different do two taxa need to be before a new genus is necessary, and how much evidence is needed before a taxon is elevated to species status? Taxonomy is painstaking, a mix between art and science. Although the link between taxonomy and phylogeny is much more formal than it was in Darwin’s time, it can be stronger. This new Phylogeny Explorer marks a significant upgrade in the strength of the relationship, and it should come as no surprise that it occurs in birds. 

Beyond inspiring over a century of evolutionary inquiry, birds have other characteristics that make them well suited to star in this story. There are about 11,000 species, enough for people to dive deep in their obsessions–there are an estimated 96 million birdwatchers in the US alone–but few enough that if you spent your whole life looking, you could potentially see almost all of them. Most species have already been described and most of those are fairly well known and, despite the slight lag between phylogenetic study and taxonomy, avian taxonomy is a shining star in the world of biodiversity informatics. There is, for example, no single accepted list of all the world’s valid plants. The most widely used, aptly named The Plant List and published by the Royal Botanic Gardens in Kew, lists 1,166,054 names, but of these, over 65% are thought to be synonyms (i.e., alternate names for species already on the list). Botanists also have no idea how many undescribed plant species there are, but they agree it’s a lot. Birds, in contrast, have no fewer than four complete global bird lists, all very similar to one another, and all describing approximately the same 11,000ish species of birds. Even more excitingly, these teams have now come together to create a single, unified bird list. 

This new list, AviList, can now be linked one-to-one with a phylogeny capturing everything we know to date about avian evolutionary relationships. It brings a cohesive evolutionary awareness to the eBird and Macaulay Library data resources that already exist, and is poised to power a whole new research domain that was difficult to synthesize before. There are some technical wrinkles to iron out still, and some published work has not yet been captured in the database, but we’ve built a system that will allow taxonomy and phylogeny to proceed in parallel going forward. We can use the phylogeny to communicate when and how taxonomic revisions are needed, and we can use the taxonomy to decide how to summarize and incorporate molecular phylogenies into our broader, synthesized understanding of avian evolutionary relationships. It’ll all update and keep pace with our ever-improving understanding of birds, and we expect the result to be a short-term burst in taxonomic revision followed by increased stability in avian nomenclature and taxonomy. We will also be able to leverage the phylogeny to pinpoint where clade-level phylogenetic study is particularly needed, and which species have never been molecularly studied. Excitingly, the Phylogeny Explorer opens a new door for communicating evolutionary concepts to birdwatchers. While we are only just beginning to dream about what’s on the other side of this door, below we expand on some of the ideas that excite us most.

************************************************************************************************************

We’ve already built some of the functionality to link your eBird list to the phylogeny, but there are so many exciting prospects yet to be explored. Right now, for example, it’s easy to see which genera and families you’ve seen (and this is fantastic–finally an eBird family life list!), but what if you could get a checkmark (a “badge”) after you’ve seen ALL the members of a genus or a family? What if you could list all the families you haven’t seen yet? Or get a notification in real time when you first see a new taxonomic family? We know that dozens, maybe hundreds, of birders and software developers have considered these ideas before, so we expect suggestions to build up over this next year. Get yours in, and we’ll see what we can manage for the next revision. 

A key corollary of this idea is the potential to create a new form of listing. Instead of each species counting the same towards your county, year, or day list, what if different species had scores that reflected their evolutionary uniqueness? There are two species scores here that are worth describing, and could go a long way towards turning this into the hottest new form of listing. The first is the unique evolutionary history each species represents–how different is a species from its nearest living relative? Here, species sitting at the end of their own long branches stand out. Often, that means non-passerines, but some really interesting passerines like Sapayoa (Sapayoa aenigma) stand out as well. The second score is, per-species, its average evolutionary distance away from all others. Here, non-passerines really steal the show; there aren’t all that many extant non-passerines in the world, when compared to passerines, and most of the non-passerines are fairly distantly related to one another. Maybe we’d use both scores, or maybe one is the clear winner here. Maybe there’s another, totally unexplored option we haven’t considered. There is also the potential to link this type of birding with geography in novel and exciting ways. What metrics would you want to tally here? No matter how the rules of the phylolisting game ultimately get written, we’re excited about the opportunities it presents to push our understanding evolutionary concepts into the mainstream. 

Evolution is a tricky subject to teach. Misconceptions abound. It’s hard to think about things that usually happen slowly, over thousands or millions of years. Hard to understand how, when individuals themselves don’t evolve, whole populations still manage to. Hard to imagine how, when mutations are random, remarkably elaborate things can still emerge. But phylogenies can help here. The mind-boggling displays of manakins make a little more sense when we zoom out, and understand they’re part of a larger clade, including cotingas and becards, many of which have shifted towards a diet of fruit; it is thought that the surfeit of fruit in tropical forests has, in turn, given space for these energy-intensive displays to flourish. We can use a phylogeny to capture these ideas, in graphical form, and suddenly it all makes a little more sense.

One of the most exciting opportunities unlocked by having this phylogeny integrated in the Lab of Ornithology’s computing infrastructure is the potential to use it to refine and improve search and download tools. For example, what if you could click on an internal node in the phylogeny, such as the base of a family, and it would pop up download options? You might be able to bundle trait data from Birds of the World with distribution data from eBird with multimedia from the Macaulay Library, and have a perfectly matching phylogeny to tie the analysis together. One big beautiful download. Even better, because these are all now part of a single system, taxonomic alignment between these datasets will remain consistent as new data is added, and our understanding of evolution and taxonomy improves over time. 

Releasing this phylogeny is a huge step, and one that has taken us five years to get to. It’s worth taking a moment to pause and appreciate that, to peruse the new Phylogeny Explorer and all the centuries of accumulated ornithological knowledge it contains. It synthesizes hundreds of recently published phylogenies into a unified representation that contains every single known, extant bird species on Earth; there is nothing else like it available. We’ve built a software package to directly integrate with the phylogeny that will be of huge benefit to researchers. But it’s also exciting to dream about what comes next. Let us know what you think so far. If you find issues, there is a feedback form you can use to suggest targeted improvements (missing data, improved taxon addition statements, etc.). And importantly, let us know what you want to see next!

Birds, with a few whacky exceptions (Dumbacher et al. 2008), are not poisonous to eat. I feel like I should cite this paper (Cott & Benson 1969) now too, because I'll probably never get another chance. And, aside from another weird exception (Londoño et al. 2015), birds don't usually look much like anything poisonous. It makes sense then that, unlike butterflies, there's very little Batesian or Müllerian mimicry driving the way birds are colored and patterned. Yet, there are some pretty remarkable instances of what seem to be plumage mimicry in birds. People have talked about these patterns for over 150 years (Wallace 1863), but whether or not these patterns represent socially-mediated convergence, convergent responses to similar environmental pressures, or simply shared evolutionary history has never adequately been resolved (Stresemann 1914, Murray 1976), although recent work has strongly supported the existence of the pattern (Jønsson et al. 2016). Moreover, what the selective pressures might be that would drive such remarkable convergence has been debated (Moynihan 1968, Cody 1969, Barnard 1979, Diamond 1982, Prum 2014). The result has been shockingly little discussion in the literature about what is, if fully corroborated, one of the coolest patterns in birds (me). 

[Full disclosure: I proposed doing my dissertation on this subject, but since I didn't have any good related research ideas at the time, that didn't happen].

I'll give you three quick examples. Downy and Hairy Woodpeckers, common backyard denizens and frequent feeder visitors across North America, are not particularly closely related--they are each more closely related to dissimilar looking species--yet they somehow have evolved to look nearly identical to one another in plumage (Weibel & Moore 2005). Ramphastos toucans can be split into two major clades: croakers and yelpers. Across the Americas one can frequently find two sympatric species, one from each clade. And, in approximate parallel across the Americas, they track one another in plumage. White-throated and Channel-billed Toucan in the Amazon, Chestnut-mandibled and Keel-billed in Central America, etc (Weckstein 2005). Then, on either side of the Andes, you have three suboscine genera--Lipaugus, Rhytipterna, and Laniocera--each with a "single" (it's not quite that simple) species on either side. West of the Andes, up into Central America, these are all rufous colored; East of the Andes, these are all gray. Ridgely and Tudor talked about this somewhere in Birds of South America (1994), which was my first exposure to the idea, and Diamond (1982) and Prum (2014) have also talked about it. Anyhow--what? What could possibly be the selective pressures that would drive that? Why would three similar-shaped, somewhat ecologically similar species all "decide" to turn gray in sympatry in the Amazon, and rufous in sympatry in Central America? It certainly doesn't seem like substrate matching. 

Well, as it turns out, there are at least 7 proposed hypotheses for why this could happen:

1) The larger species would suck to try and kill and eat, and the smaller species mimicked the model to make predators think twice before giving it a go (Wallace 1863). I'm going to add some slight details to this idea, since a paper we're working on now will be able to shed some light on it: the larger (model) species is outside the preferred prey size range of smaller sympatric raptors, and the mimic enjoys reduced predation pressure by looking like the model. 

2) All species involved converged on the phenotype to facilitate flocking behavior and thereby reduce predation pressure (Barnard 1979). 

3) All species involved converged on the phenotype to facilitate flocking and encourage stereotyped behavioral interactions between flock members (Moynihan 1968).

4) All species involved converged on the phenotype to facilitate interspecific territorial interactions (Cody 1969).

5) The larger species converged on the model (the smaller species) so as to fool predators into believing it to not be a worthwhile target (the skin and bones hypothesis). I don't know a citation for this idea, but it's out there.

6) The smaller species converged on the model so as to confuse the model and experience reduced aggression from the model (Diamond 1982, Prum & Samuelson 2012).

7) The smaller species converged on the model so as to confuse third parties and hold a more dominant position around contested resources than it otherwise would (Wallace 1863, Diamond 1982). 

NB: in our recent paper (below), we attributed this last idea solely to Wallace, a view likewise stated by Prum and Samuelson (2012) and Prum (2014). However, I have recently more carefully read those original Wallace papers, and Jønsson et al. (2016) and Diamond (1982) are more accurate in stating Wallace's idea had to do with hypothesis 1, above. To be fair, Wallace was sort of agnostic about the precise driver, and his brief passages could be interpreted as giving rise to hypothesis 7, but I now think hypothesis 7 is best attributed to Diamond (1982). I guess the moral of the story is that arguing 155 years later over which evolutionary driver Wallace was invoking four years after the publication of the Origin of the Species is a fool's errand. Wallace seems like a fine choice. 

So what is the evidence for and against each of these hypotheses?

1) Sorry, but you'll have to keep your eyes out for our third installment from Project FeederWatch, which looks at the predation data that participants have submitted. Otherwise, no experimental evidence in birds that I know of. Hypothesis 1 remains standing. 

2) Seems unlikely to me that species would converge on visually obvious phenotypes when they could just as well blend into the background, but I guess that doesn't kill this hypothesis entirely. To invoke sexual selection as as a driver of bold plumages, and then suggest this group-level selection somehow drives near perfect plumage convergence seems fraught, but we'll leave it as a hypothesis someone with a lot of time on their hands might want to look into. 

3) Community selection? Not a lot of support for that. Here are Moynihan's 5 predictions from his paper.

A) Related species which are sympatric will usually be more similar in overall appearance, on the average, than equally closely related species of similar habits which are allopatric.
B) The closer and more frequent are the contacts between individuals of sym- patric species, the closer will be the resem- blances between them.
C) Resemblances among sympatric spe- cies should be positively correlated with age of sympatry. Species which have been sympatric for a long time (without diverg- ing ecologically to such an extent as to prevent contacts between individuals of the different species) will be more similar in overall appearance, on the average, than comparable species which have been sym- patric for a short time.
D) "Old" faunas (which have evolved for some appreciable length of time with- out being invaded by species from other areas) should include larger numbers of similar-appearing species than otherwise comparable "young" faunas.
E) The strength and extent of resem- blances among sympatric species should be positively correlated with the number of species involved. The more species overlap and come into contact with one another frequently, the more likely it is that they will evolve convergent resem- blances for social mimicry. (This, of course, is one of the reasons why social mimicry is more conspicuous, and pre- sumably more common, in tropical conti- nental areas than in other regions.)

There is strong evidence against most of these ideas, e.g. (Martin et al. 2015, Pigot & Tobias 2013). So, I'd say we can scratch hypothesis 3 off the list. 

4) Again, there is not much empirical evidence out there showing that natural selection operates at these sorts of community levels. Species that hold mutually interspecific territories appear fully capable of distinguishing one another based on numerous characteristics such as vocalizations--they don't need to look like one another to make those distinctions. The idea here is that looking similar means that interactions between individuals are "definite". We look like each other so we know we're fighting here, no beating around the trunk trying to figure out if we want to fight at all. Many species that are agonistic towards one another over large portions of their range look nothing alike, and many species that look alike do not seem to hold interspecific territories. Given that, plus all the other selective pressures on plumages, I find this hypothesis really unlikely. Let's scratch it off the list. 

5) Maybe, but unlikely in my opinion. We should hopefully be able to indirectly address it in our forthcoming FeederWatch predation paper. A prediction would be that both species would suffer low predation risk. I'm not saying it's impossible, but it would be much like a monarch mimicking a tasty skipper butterfly.

6) I see Hairy Woodpeckers occasionally displace Downy Woodpeckers in the forest in natural settings. Kilham (1972) saw a Campephilus woodpecker displace its mimic Dryocopus in Panama. Most birds seem able to recognize specific individuals of their own species. I have always found the idea that the purported model species somehow chooses not to attack the smaller mimic, when it would likely attack an unknown individual of its own species, difficult to believe. Fortunately, we don't have to rely on my feelings for whether to strike this hypothesis from the list. In a recent paper (Leigton et al. 2018) we tested the idea, and found strong evidence against it. Instead of experiencing reduced aggression, Downy Woodpeckers (the purported mimics) receive ample aggression from Hairy Woodpeckers--much more than would be expected based on the species' relative abundances. There is a slight caveat here: our study took place around feeders, which could be argued are artificial boxing rings. Yet, the evidence came down so hard against this idea, I'm striking it from my own list of reasonable hypotheses. Show me some experimental evidence and I'll change my tune. Instead of supporting this hypothesis, we found some support for the fact that...

7) It seems that Downy Woodpeckers hold slightly higher positions in the dominance hierarchy than would be expected based on their body mass and phylogenetic position. What do I mean by that? Well, we recently showed that woodpeckers tend to be dominant in general, and that body mass does a good job of predicting species' dominance ranks (Miller et al. 2016). So, even after accounting for that boost in dominance, Downy Woodpeckers are more dominant than expected based on their body mass. Indeed, they are predicted to be less dominant than House Sparrows, Song Sparrows, White-throated Sparrows, Yellow-rumped Warblers, and maybe even White-breasted Nuthatches (Leighton et al. 2018). So, is the fact that they look like Hairy Woodpeckers the thing that gives Downy the boost in the fight club rankings? We think it's likely, but the jury's still out.

Literature cited:

Barnard, C. J. (1979). Predation and the evolution of social mimicry in birds. The American Naturalist, 113(4), 613-618.

Cody, M. L. (1969). Convergent characteristics in sympatric species: a possible relation to interspecific competition and aggression. Condor, 71(3), 223-239.

Cott, H. B., & Benson, C. W. (1969). The palatability of birds, mainly based upon observations of a tasting panel in Zambia. Ostrich, 40(S1), 357-384.

Diamond, J. M. (1982). Mimicry of friarbirds by orioles. The Auk, 99, 187-196.

Dumbacher, J. P., Deiner, K., Thompson, L., & Fleischer, R. C. (2008). Phylogeny of the avian genus Pitohui and the evolution of toxicity in birds. Molecular Phylogenetics and Evolution, 49(3), 774-781.

Jønsson, K. A., Delhey, K., Sangster, G., Ericson, P. G., & Irestedt, M. (2016). The evolution of mimicry of friarbirds by orioles (Aves: Passeriformes) in Australo-Pacific archipelagos. Proceedings of the Royal Society B: Biological Sciences, 283, 20160409. http://dx.doi.org/10.1098/rspb.2016.0409

Kilham, L. (1972). Habits of the Crimson-crested Woodpecker in Panama. The Wilson Bulletin, 84(1), 28-47.

Leighton, G. M., Lees, A. C., & Miller, E. T. (2018). The Hairy-Downy game revisited: an empirical test of the Interspecific Social Dominance Mimicry Hypothesis. Animal Behaviour, 137, 141-148. https://doi.org/10.1016/j.anbehav.2018.01.012

Londoño, G. A., García, D. A., & Martínez, M. A. S. (2015). Morphological and behavioral evidence of Batesian mimicry in nestlings of a lowland Amazonian bird. The American Naturalist, 185(1), 135-141.

Martin, P. R., Montgomerie, R., & Lougheed, S. C. (2015). Color patterns of closely related bird species are more divergent at intermediate levels of breeding-range sympatry. The American Naturalist, 185(4), 443-451. https://doi.org/10.1086/680206

Miller, E. T., Bonter, D. N., Eldermire, C., Freeman, B. G., Greig, E. I., Harmon, L. J., … Hochachka, W. M. (2017). Fighting over food unites the birds of North America in a continental dominance hierarchy. Behavioral Ecology, Online Early. https://doi.org/10.1093/beheco/arx108

Moynihan, M. (1968). Social mimicry; character convergence versus character displacement. Evolution, 22(2), 315-331.

Murray, B. G. (1976). A critique of interspecific territoriality and character convergence. The Condor, 78(4), 518-525.

Pigot, A. L., & Tobias, J. A. (2013). Species interactions constrain geographic range expansion over evolutionary time. Ecology Letters, 16(3), 330-338. https://doi.org/10.1111/ele.12043

Prum, R. O. (2014). Interspecific social dominance mimicry in birds. Zoological Journal of the Linnean Society, 172(4), 910-941.

Prum, R. O., & Samuelson, L. (2012). The Hairy-Downy Game: A model of interspecific social dominance mimicry. Journal of Theoretical Biology, 313, 42-60.

Prum, R. O., & Samuelson, L. (2016). Mimicry cycles, traps, and chains: the coevolution of toucan and kiskadee mimicry. The American Naturalist, 187(6), 753-764.

Ridgely, R. S., & Tudor, G. (1994). Birds of South America Volume II, The Suboscine Passerines: Ovenbirds and Woodcreepers, Antbirds, Gnateaters, and Tapaculos, Tyrant Flycatchers, Manakins and Cotingas. Austin: The University of Texas.

Wallace, A. R. (1863). List of birds collected in the island of Bouru (one of the Moluccas), with descriptions of new species. Proceedings of the Zoological Society of London, 1863, 18-28.

Weckstein, J. D. (2005). Molecular phylogenetics of the Ramphastos toucans: implications for the evolution of morphology, vocalizations, and coloration. The Auk, 122(4), 1191-1209.

Weibel, A. C., & Moore, W. S. (2005). Plumage convergence in Picoides woodpeckers based on a molecular phylogeny, with emphasis on convergence in downy and hairy woodpeckers. The Condor, 107(4), 797-809.

​Birders are a curious bunch. There are many millions of them in the world, and they range from mildly to severely obsessed with all things birds. Few things inspire such fevered participation by the fervent, and such bemused wondering by the uninitiated. Who else gets up on purpose at 4 AM on a non-workday? Birding can be a pretty minimalist hobby--you don't need much to watch birds, though a pair of binoculars is nice. Indeed, you don’t even need to see to be a good birder. [Birdwatching does have an unfortunate diversity skew--here's a funny post on that]. Despite that, there is no shortage of ways birders can spend their money. And spend it they do--a 2013 US Fish & Wildlife study estimated birders spend $107 billion/year on their pursuit of our planet's feathered denizens. Pretty cool if you ask me.

Like I said, birders vary in the severity of their avian afflictions. There is a subset, not insignificant in absolute numbers, of people who know which birds they've seen, which birds they haven't seen, where those birds live, when to find them, and how to find them...in short, there are at least a couple hundred thousand people in the US who when I say, "there's a Rustic Bunting four hours away", respond, "get in the car."

Given that mass of educated, passionate, motivated birders, one might think that there are no new ways to consider which birds one has or has not seen. But, after some discussion with various folks recently who have also had these ideas, it occurred to me that one aspect of birding has not been well explored outside of a few science nerd types. 

What I'm proposing here is a phylogenetic life list. In this context, evolutionarily unique species are "worth more". You might have been excited this spring when you ticked your first Golden-winged Warbler, but let me temper that enthusiasm for you. What you saw was, in an evolutionary sense, barely distinguishable from a Blue-winged Warbler (Toews et al. 2016). Come talk to me when you've seen a Kagu in the mountains of New Caledonia. Playing this sort of listing game, all new species (branches on the evolutionary tree) count towards your phylogenetic life list, but evolutionarily unique species count more. Put differently, seeing a Golden-winged Warbler (or a Kagu) ticks off one of the ~10,000 bird species to see globally under a traditional counting strategy (0.01% of all species to see in the world). But, under a phylogenetic life list strategy, seeing a Golden-winged Warbler ticks off just 0.003% of global unique evolutionary diversity, while seeing a Kagu adds 0.13%! 

I should digress and say taxonomy is hard. Converting the taxonomic concepts in my eBird life list to the taxonomy used by the Jetz et al. (2012) world bird tree took me a week's work. After that though, the rest was pretty easy. 

For the purposes of this post, there are two ways to consider evolutionary uniqueness. The length of the pendant edge of the species in question, and the mean relatedness of that species to all others. The pendant edge describes the amount of time that branch of the tree of life has been evolving on its own path (all caveats about extinction aside). Take the Kagu. It's evolutionarily unique according to this measure because its closest extant relative is the Sunbittern. They last shared a common ancestor 63 million years ago. Seeing a Kagu is kind of like seeing a velociraptor. In contrast, the mean relatedness describes the average evolutionary distance of a species to all others. In this context, species like tinamous and ostriches are even more evolutionarily unique than the Kagu, because tinamous and such are distantly related to the vast majority of other birds. 

I've made some visualizations to go along with this idea. ​This first one shows the world bird tree, pruned to the bird species that I've seen (according to eBird and as converted to world bird tree taxonomy). The evolutionary uniqueness of these species, according to the pendant edge method, is color-coded on there. I've also stretched branches out as a function of the pendant edge, so that evolutionarily unique species are colored in yellow, and are situated on long spines out from the tree. A friend affectionately called this an urchin plot. The plot is too low-resolution to zoom in, so I'll narrate it for you. At 3 o'clock are the Ostrich and Emu, and at 4 o'clock are the Sunbittern and Oilbird. Over at 8 o'clock is the Sapayoa. [I'm 100% convinced that there is a tight correlation between one-word bird names and evolutionary uniqueness, much like the tight correlation between one-word soccer player names and soccer skill]. 

This second one shows the evolutionary uniqueness of all the species I've seen according to the mean distance measure. Here, tinamous and the like are really unique, whereas passerines, which comprise the bulk of avian diversity are not. 

Last up is the entire world bird tree, with species I have and have not seen color-coded according to their (pendant edge) evolutionary uniqueness. Gray = not unique, I have seen it. Black = unique, I have seen it. Blue = not unique, I have not seen it. Yellow = unique, I have not seen it. Because it's such a large file, and I'm just adding a jpg of the zoomable pdf, I've added a few screenshots here to get a better sense of what's going on.

I see some obvious extensions of this. First, to return the hordes of interested birders, it would be pretty amazing to be able to generate these sorts of plots for any user of eBird, and serve these up in a more interactive setting. I think it presents some great opportunities for helping people to think about evolution and practice tree-reading skills. I'm thinking OneZoom-ish. Second, it would be really interesting to be able to put the phylogenetic life list in a geographic context. For instance, where on Earth should I go to see the most novel phylogenetic diversity in the most geographically constrained area? Or, if I visit my aunt in southern California, what would be the most evolutionarily unique species I might see?


Literature cited: 

David P.L. Toews, Scott A. Taylor, Rachel Vallender, Alan Brelsford, Bronwyn G. Butcher, Philipp W. Messer, Irby J. Lovette. 2016. Plumage Genes and Little Else Distinguish the Genomes of Hybridizing Warblers. Current Biology 26:2313-2318. DOI: http://dx.doi.org/10.1016/j.cub.2016.06.034

W. Jetz, G. H. Thomas, J. B. Joy, K. Hartmann, A. O. Mooers. 2012. The global diversity of birds in space and time. Nature 491: 444-448. doi:10.1038/nature11631

I'm excited to report that our first paper using citizen scientist-collected behavioral observations has been published in Behavioral Ecology (Miller et al. 2017).  Well over a thousand FeederWatchers participated in the project, collecting data on aggressive interactions at feeders around North America. We assembled these interactions into a large directed network, and calculated some metrics of dominance. Ultimately, we settled on a modified version of the Bradley-Terry model (Bradley and Terry 1952) to calculate a species' dominance--other approaches yielded results that didn't make any biological sense. 

In and of itself, this was a neat product--a continental dominance hierarchy of feeder birds! But what determined species' positions in the hierarchy? We found that, unsurprisingly, body mass explained a lot of the variance in species' positions. However, certain lineages (e.g., families) of birds tended to be more or less dominant than expected based on their body mass. Presumably, the members of these lineages share certain traits that make them more or less aggressive. Take woodpeckers for instance: long, heavy bills they wield against unwielding substrates every day. Or warblers: little, feather-clad bundles of anger. Ok, maybe that's not really a trait, but you get the point.

We also looked for the existence of rock-paper-scissors relationships in the data. These sorts of relationships are interesting, as they are thought to promote local species coexistence (Levine et al. 2017); no species is able to gain a competitive advantage over all others. We found few such instances of these relationships, but those that we did find tended to involve invasive species. This seems like something worth following up in the future!

Stay tuned for an upcoming blog about the paper on the FeederWatch page. Soon thereafter, keep your eyes peeled for a glossy article in the Living Bird, complete with original illustrations by Jillian Ditner! (http://jillianditnerstudios.com/).

Literature cited:

Bradley, R.A & M.E. Terry. 1952. Rank analysis of incomplete block designs: I. The method of paired comparisons. Biometrika. 39:324–345.

Levine, J.M., J. Bascompte, P.B. Adler & S. Allesina. 2017. Beyond pairwise mechanisms of species coexistence in complex communities. Nature. 546:56-64. doi:10.1038/nature22898

Miller, E.T., D.N. Bonter, C. Eldermire, B.G. Freeman, E.I. Greig, L.J. Harmon, C. Lisle & W.M. Hochachka. 2017. Fighting over food unites the birds of North America in a continental dominance hierarchy. Behavioral Ecology. Online Early. doi:10.1093/beheco/arx108

Sarah Wagner and I spent many months between 2009 and 2012 in Australia chasing honeyeaters around. We visited some mind-bendingly cool places during that time, but it wasn't just for fun. We were chasing and watching honeyeaters, recording what they were eating and how they were getting it. We eventually were able to do this for almost all of the species (we never were able to find sufficient individuals of the tricky Gray Honeyeater), and that dataset is now available here. We complemented the foraging dataset with a museum-based morphometric dataset, also freely available here. It took a few years, but we've finally managed to synthesize those datasets into a paper on the honeyeater ecomorphology, recently published at The American Naturalist. You can find a nice press release for the paper here. Put briefly, we found that morphology largely predicts ecology, but that arid-adapted honeyeaters, which come from a restricted subset of lineages, use their phylogenetically conserved morphology in novel ways. We suggest that certain lineages have managed to invade the recently (15 million years ago recent) created deserts, and that the species from these lineages are exploiting the ecological opportunity afforded by these new habitats by shaking what their mama gave them, even if it's not quite suited to the task--desert honeyeaters do more with less. You can see some clips I shot of honeyeater foraging on YouTube (embedded below), including some short clips of the notably arid-adapted Gibberbird and Orange Chat...although the shot of the Gibberbird doesn't show it doing much at all, much less "doing more with less". Check out that Orange Chat though! Looks more like an Orange Sandpiper-Honeyeater to me. 

A few months ago, walking in the woods near our house, I came up with what I thought would be the phylogenetic equivalent of the trait-based metric functional dispersion (FDis, Laliberte & Legendre 2010). The metric would be the mean distance of a sample of taxa to their most recent common ancestor (MRCA). Almost as soon as the idea crossed my mind, I realized this new measure, let's call it distMRCA, was not exactly analogous to FDis. Given that in an ultrametric tree, a sample of species are all equally distant from their MRCA, distMRCA would be the same for one species or when averaged across all species in a sample. Kind of boring. Also, in many empirical cases, given a sample of taxa, it's likely their MRCA is the root (or near it) of the phylogeny in question. So I let the idea walk to the other side of my mind and forgot about it.

A few weeks ago I noticed that two different folks (one, two) have implemented (with attribution--thanks!) a mean root distance (MRD) function from my now defunct U. of Missouri, St. Louis webpage. Since it fits within the framework introduced in our recent paper, I decided to include MRD in the metricTester package.

While updating the MRD code, a variant of the distMRCA function occurred to me. Rather than simply finding the distance of the set of taxa from their MRCA, perhaps the mean of all distances between all pairwise MRCAs and the root would provide a better measurement of subtle variation in the arrangement of present taxa across a phylogeny. For instance, given three species (s1, s2, and s3), the original idea would have simply found the distance of those taxa from their MRCA. The updated idea would be to find the distance of the MRCA of each taxon-pair (s1-s2, s1-s3, s2-s3) from the root. Sounded promising. Maybe you already see where this is going. It took me coding the whole thing out and testing how it behaved until I realized I'd re-invented the wheel (admission: my version is extra heavy and slow rolling). Specifically, I'd come up with a new way of calculating phylogenetic species variability (Helmus et al. 2007), which was really just a new flavor of mean pairwise phylogenetic distance (Webb 2000), which was just a new way of calculating Δ+ (Warwick and Clarke 1995). And so on and so forth. Maybe there are no new phylogenetic community structure metrics under the sun.

In retrospect, it's quite obvious why these things should be equivalent. Given an ultrametric tree, the pairwise distance between two taxa is related to their distance from the root. By that, I mean that if the total tree height is 10 mya, and two taxa are separated by a "pairwise" distance of 4 mya, then we know they shared a MRCA 8 mya (10 - 4/2). Obviously, taking the averages of these distMRCA and these MPD values are going to be closely correlated.

On the plus side, I thought of a way of abundance-weighting MRD. That is incorporated into metricTester now too. Given my experience with re-inventing the MPD wheel, I figured it would be related to IAC (Cadotte et al. 2010), but an initial test showed that not to be the case. Since it's just a node-based measure, it seems likely it's not of great use to empiricists in today's age of dated phylogenies, but it's there in case someone would like to experiment with it.

References:

Cadotte, M. W., T. Jonathan Davies, J. Regetz, S. W. Kembel, E. Cleland, and T. H. Oakley. 2010. Phylogenetic diversity metrics for ecological communities: integrating species richness, abundance and evolutionary history. Ecology Letters 13:96–105. 10.1111/j.1461-0248.2009.01405.x

Helmus, M. R., T. J. Bland, C. K. Williams, and A. R. Ives. 2007. Phylogenetic measures of biodiversity. The American Naturalist 169:E68–E83. 10.1086/511334

Laliberté, E., and P. Legendre. 2010. A distance-based framework for measuring functional diversity from multiple traits. Ecology 91:299–305. 10.1890/08-2244.1

Miller, E. T., D. R. Farine, and C. H. Trisos. 2016. Phylogenetic community structure metrics and null models: a review with new methods and software. Ecography:In press. 10.1111/ecog.02070

Warwick, R. M., and K. R. Clarke. 1995. New "biodiversity’ measures reveal a decrease in taxonomic distinctness with increasing stress. Marine Ecology Progress Series 129:301–305. 10.3354/meps129301

Webb, C. O. 2000. Exploring the phylogenetic structure of ecological communities: an example for rain forest trees. The American Naturalist 156:145–155. 10.1086/303378