From the Autumn 2017 issue of Living Bird magazine. Subscribe now.

If you had opened a copy of the Sibley Guide to Birds when it was first published in the year 2000 and flipped to the section on wood-warblers, you would have found 13 pages devoted to members of a single genus: Dendroica, Latin for tree-dweller. Dendroica’s inhabitants included 21 colorful species—such as Magnolia, Blackburnian, and Cerulean Warblers—dear to the hearts of many birders.

Open a copy of the second edition of the Sibley Guide today, and Dendroica is nowhere to be found.

subscribe to Living Bird magazine

There hasn’t been a mass extinction in the intervening years. The wood-warbler species are all still there, but filed under a different genus name, Setophaga. Instead, there has been a major shift in how ornithologists sort and classify bird species, and the genus name Dendroica was a casualty.

Decisions about how North American bird species are classified and what is and is not considered a species are made every summer by a special committee of the American Ornithological Society. An AOS committee bases its judgments on the best available science. But the science is rapidly expanding. Like many other branches of biology, ornithologists are trying to make sense of a flood of new information flowing from the latest advances in genome mapping. Today, avian geneticists can dive deep into genomes to unveil the molecular differences underlying variation between birds.

But instead of providing clarity, all this new technology is adding complexity. It turns out there’s much more that separates bird species than meets the birder’s eye—or binocular lens.

The Bullock’s Oriole (left) and Baltimore Oriole (right) were once lumped together as one species, the Northern Oriole, because they can mate and produce viable offspring. Photos by Brian Small.

The definition of what constitutes a species has changed more than once in the past two centuries, as scientists have advanced their understanding of evolution and behavior. In the 1800s, ornithologists classified birds based mostly on morphology, or how they looked. Examining specimens harvested via shotgun, they split populations into smaller and smaller segments based on tiny variations in size, shape, and plumage. As legions of scientific expeditions roamed America’s unexplored wilds in the 19th century—collecting specimens and sending them back east to institutions such as the Smithsonian—there was a continuous increase in the number of recognized species through the turn of century. For example, ornithologists had divvied up juncos into 14 distinct species by the early 1900s.

But this species bonanza didn’t last. A pivotal point for ornithology, and for biology in general, was the publication of Ernst Mayr’s Systematics and the Origin of Species in 1942, which formalized the biological species concept. This new definition of species as “groups of interbreeding natural populations that are reproductively isolated from other such groups” sparked a trend of species lumping (or combining two or more species into one). Now, scientists were considering the behavior of living birds, not just how they looked.

There were nearly 19,000 scientifically recognized bird species in the world at the turn of the 20th century. But the introduction of the biological species concept caused a reshuffling that dropped that number by half. Graphic adapted from Handbook of the Birds of the World, Lynx Edicions 1997.There were nearly 19,000 scientifically recognized bird species in the world at the turn of the 20th century. But the introduction of the biological species concept caused a reshuffling that dropped that number by half. Graphic adapted from Handbook of the Birds of the World, Lynx Edicions 1997. View larger image.

Many such lumps occurred when scientists in the field documented that two species originally thought to be distinct were actually mating with each other and producing viable hybrid offspring. That’s how the number of scientifically recognized junco species shrank from 14 to three. Baltimore and Bullock’s Orioles were combined under the name “Northern Oriole,” and Myrtle and Audubon’s Warblers were consolidated under the single “Yellow-rumped Warbler.” From 1900 to the end of the 20th century, the number of scientifically recognized bird species worldwide plummeted from nearly 19,000 to around 10,000.

Debates about how best to define species didn’t end with Ernst Mayr, however. In the 1980s, a new idea known as the phylogenetic species concept began to gain prominence, partly owing to scientific advances in constructing the evolutionary trees (or phylogenies) that depict the evolutionary relationships among birds, from today’s species all the way back to the earliest avian origins. Proponents of the phylogenetic species approach argue that species should be classified according to their evolutionary past—a species is any unique, independently evolving lineage, regardless of its potential mating isolation (or not) from its current relatives. A bird’s taxonomic status, they contend, should be informed by its history.

magnolia Warbler by Tim HOpwood via BirdshareThe Magnolia Warbler was once a member of the genus Dendroica, but is now in the genus Setophaga. Photo by Tim J. Hopwood via Birdshare.

It was phylogeny that led to the demise of Dendroica at the genus level. Irby Lovette, director of the Fuller Evolutionary Biology Program at the Cornell Lab of Ornithology, and his colleagues used analyses of the DNA of 107 warbler species to rewrite the warbler family tree. Lovette and his team sequenced more than 10,000 individual nucleotides from the birds’ nuclear and mitochondrial DNA and applied a series of statistical techniques to the results. They concluded, among other things, that all of the erstwhile Dendroica species fell within an evolutionary group that also included Setophaga, another genus that as of 2011 only included the American Redstart. So the Dendroica and Setophaga warblers belonged in one genus. Setophaga happened to be the older name, and the rules of scientific nomenclature meant it took precedence, so Dendroica was history.

The biological species concept remains the most widely accepted standard among ornithologists for classification decisions at the species level. But in the last five years, yet another revolution has rocked the world of avian classification. Gone are the days when ornithologists would labor for months to sequence just a few individual genes on their way to building an evolutionary tree. Today’s “high-throughput” or “next-generation” genome sequencing means a student in an evolutionary biology laboratory today can assemble an entire genome—all 1 billion or so units in a bird’s DNA—in just a few weeks.

Lovette describes it as a shift from “genetics” to “genomics.”

“Before, we did what I would call genetics, pulling out very, very small pieces of DNA,” he says. “With next-generation sequencing, we’ve moved from dealing with tiny fractions of the genome to all of it, and that’s expanded the amount of information we have by many orders of magnitude. It’s opened the door to all kinds of new analyses of how genetic variation is packaged through space and time in birds. It’s not that the underlying biology has changed, it’s just that our window into it has widened so dramatically.”

And birds, it turns out, make an ideal subject for these genomic studies.

“Birds have a modestly sized genome compared to a lot of other organisms,” says Lovette. “It’s still huge, but not as vast as in mammals, for example. And the organization of the avian genome is really similar across all birds, which makes comparing bird genomes with one another much more straightforward. We can do genomic-scale studies on birds relatively cost-effectively.”

All of this means that scientists have the ability to peer more deeply into the DNA of birds today than ever before. But in some ways the resulting picture for species classification isn’t getting clearer—rather, it’s getting blurrier. It seems that the more closely evolutionary biologists look into the genome, the more arbitrary the boundaries between some species appear to be. It’s a bit like stepping too close to a pointillist painting: instead of revealing tiny details on the picnickers’ faces, the whole thing dissolves into dots.

“On the one hand, this is the kind of information we’ve been wanting forever,” says Lovette. “We can see how different parts of the genome are, say, moving between species that are hybridizing, or evolving in different directions due to selection. In that sense, it’s a fantastic advance.

“But genomics, especially at the species level, is also starting to illustrate just how many pathways there are to becoming a separate species,” he continues. “Now that we can see these patterns in all of their wonderful complexity, how do we link these fairly complicated histories to the more simplistic models of what we expected to see among species when we didn’t have that great richness of information?”

Put another way, how does next-generation genomics—with all its high-tech analysis sifting through millions of genes at light speed—translate into the age-old discipline of taxonomy where birds are ordered like last names in a phone book?

Lovette was as sad as anyone to see Dendroica go. (“I once had a license plate that read DNDRCA,” he admits.) However, Lovette is quick to point out that questions about how to organize birds at the genus, family, and order level are distinct from questions about whether to lump or split individual species.

“The higher-level classification derives from the evolutionary trees, while the species-level divisions derive from what we think is happening among those species in an ongoing biological context,” he says. “They’re really two different things.”

In other words, higher-level taxonomy like the kind he tackled in the Dendroica paper is based on the past—the evolutionary relationships that underlie the diversity we see today. To understand things at the species level, we need to understand what’s happening now.

Some of the most recent genomic research to hit scientific journals is still being hotly debated among ornithologists. But if the latest findings find wider acceptance, they threaten to overturn—or restore—the classification of some of North America’s most beloved and sought-after songbird species.

Redpolls Photos (left to right) by Melissa Groo, Brian Small, David Stimac, David Speiser, David Stimac, and David StimacThe redpoll spectrum: The bird in the upper left (1) is a classic streaky Common Redpoll, while the bird on the lower right (6) is a snowy Hoary Redpoll. In the field, birders see flocks of redpolls with many variations in between these two extremes. Research from the Cornell Lab suggests that all these redpolls may be the same species. Photo credits: (1) Melissa Groo, (2) Brian Small, (3, 5, 6) David Stimac, (4) David Speiser.

If you’ve ever seen a Hoary Redpoll, chances are you were pretty cold at the time. American birders routinely travel to remote parts of northern Minnesota and Wisconsin in hope of adding this small, red-capped bird to their life lists. It breeds in the far reaches of the Arctic, and even in winter it only irregularly descends into the Lower 48 States.

Birders who brave the cold for their lifer Hoary Redpoll may have a hard time distinguishing it from its more southerly cousin, the Common Redpoll. As their name suggests, Hoary Redpolls tend to have whiter plumage than streaky brownish Common Redpolls. But what looks obvious in a field guide often isn’t quite so clear in the field.

“You really want to see a Hoary Redpoll, but then when you do you’re like, am I sure what that is? There’s a lot of variation,” says University of Colorado scientist Scott Taylor. Taylor saw his lifer Hoary among a redpoll flock in Cortland, New York, when he was a postdoctoral researcher at the Cornell Lab. As he scanned the flock, he noticed that the birds varied along a spectrum in their appearance. Some were very streaky (clearly Common), some had a snow-white breast (clearly Hoary), but there were lots of redpolls with many degrees of streakiness and snowiness in between.

Genome Mapping Explained

DNA mapping, Illustration by Virginia Greene, Bartels Science Illustration InternIllustration by Virginia Greene, Bartels Science Illustration Intern.

How exactly do scientists map a bird’s genome? In short: they color-code it.

First, scientists unzip the DNA’s double helix, exposing the sequence of chemical “bases” (A, T, C, and G) that are combined to make a genome. Then, they chop the long strands of DNA into smaller, more manageable fragments, and place the DNA fragments onto a tiny plate about the size of a quarter. (Think of a microscope slide.)

Next, scientists add a fluorescent dye that makes the bases light up. For example, A’s turn yellow, T’s turn green, C’s turn red, and G’s turn blue. A high-speed computing program then reads the color-coded bases and outputs the fragment sequences. A genome contains more than 1.2 billion base pairs, so stitching all of these short sequences together requires big-data computer processing.

In the early 2000s, the Human Genome Project cost hundreds of millions of dollars. But today, the genome of a bird can be sequenced for much less—thousands of dollars. Bird genomes from several species can even be sequenced side by side, which makes it possible to compare differences. That’s how evolutionary biologists at the Cornell Lab of Ornithology compared Golden-winged and Blue-winged Warbler genomes and found that they differ in just 0.03 percent of their bases.

—Alison Haigh, Branegan Science Writing Fellow

In 2015 Taylor and his colleague Nick Mason (a graduate student at Cornell) published a genomic study that surveyed the genetic variation that might underpin this redpoll plumage variation. Altogether, they analyzed more than 236,000 individual points in the redpoll genome—still less than 1 percent of the whole thing, but previous studies of redpoll genetics had never sequenced more than a dozen regions at a time.

Even with this much larger sample size, Taylor and Mason couldn’t find a single DNA difference where they could reliably distinguish between Hoary and Common Redpolls, or between those two species and their European cousin the Lesser Redpoll.

“We didn’t see the kinds of genomic differences that we’d expect to see between birds that have been [reproductively] isolated for long periods of time,” says Taylor.

Instead, says Mason, the world’s three redpoll species seem to be “functioning as members of a single gene pool that wraps around the top of the globe.” Indeed, Hoary, Common, and Lesser Redpolls all have overlapping breeding ranges around the Arctic.

Instead of different genes, the morphological variation—or differences in how various redpolls look—seemed to stem from genetic expression. It’s kind of like how two humans might have the same gene for brown hair, but one person’s hair might be lighter than the other’s—that gene is being expressed differently. In the same way, Hoary and Common Redpolls have remarkably similar sets of genes, but those genes are expressed differently, causing plumage differences.

Mason and Taylor used their research to produce an official proposal for the AOS to lump Hoary, Common, and Lesser Redpolls into a single species, based on the genetic evidence. But while genomics research has put the Hoary Redpoll’s species status into question, it also put two beloved species lost to a past lumping on the path to a potential comeback via a split.

Many older birders never stopped using the names “Myrtle Warbler” and “Audubon’s Warbler” for the eastern and western forms of the Yellow-rumped Warbler, after they were lumped in 1973 based on the fact that these forms meet and hybridize along a narrow band in western North America. New research led by David Toews, another Cornell postdoctoral researcher, and his colleagues suggests that maybe the species division between Myrtle and Audubon’s Warblers was correct all along.

These researchers sequenced around 37,000 regions of the warblers’ genomes (again, a tiny fraction of their total DNA, but far more than any past studies). Their results, published last year (see Goodbye Yellow-Rump?), show that about 60 of those regions differ significantly between “Myrtle” and “Audubon’s” Warblers, and there may be active selection at work on the genes in those particular areas maintaining the two populations’ distinctiveness. The proof, says Toews, lies in the hybrid zone between “Myrtle” and “Audubon’s” Warblers—the narrow 80-mile sliver in their ranges where the two forms meet and hybridize in western Canada, the entire reason they were lumped into Yellow-rumped in the first place.

“Audubon’s” Yellow-rumped Warbler by Marie Read“Audubon’s” Yellow-rumped Warbler. Photo by Marie Read.
"Myrtle” Yellow-rumped Warbler by Marie Read"Myrtle” Yellow-rumped Warbler. Photo by Marie Read.

Toews notes that the Myrtle and Audubon’s hybrid zone hasn’t budged in the last 50 years. In other words, the hybrids aren’t thriving reproductively; they seem to be stuck in neutral. Toews thinks there must be some sort of detriment in the hybrids that keeps them from surviving and carrying their genetic material farther afield. The checklist committee used that very same reasoning to re-split Northern Oriole back into Baltimore and Bullock’s in 1995, because upon further review the level of interbreeding where the two ranges overlapped was not substantial enough to justify the lumping.

This year Toews submitted a proposal to re-split Yellow-rumped Warblers as well.

Toews, Taylor, and Lovette were all involved in an effort to compare the entire genomes of yet another pair of hybridizing species: the Golden-winged Warbler and the Blue-winged Warbler, two species that look distinctly different and usually sing different songs, but that hybridize frequently.

Speciation in Action, Illustration by Virginia Greene, Bartels Science Illustration Intern“When you capture that process of speciation in action, you’re naturally going to find in-between cases that are in this gray zone," explains Lovette, "where by some criteria they’re separate species and by other criteria they’re not.” Illustration by Virginia Greene, Bartels Science Illustration Intern

In contrast to the 60 or more regions of differentiation between Audubon’s and Myrtle forms of Yellow-rumped Warblers, the Cornell team discovered that Golden-winged and Blue-winged Warblers differ in only six regions—or just 0.03 percent—of their entire genomes (see Golden-winged and Blue-winged Warblers Are 99.97 Percent Alike Genetically). While these two warblers look different, the few genes that determine those plumage patterns are among the only things that set the two species apart genomically.

The complexities of reconciling species status with genome mapping are enough to set a birder’s head spinning. Two birds can look similar (as in Hoary and Common Redpolls) or quite a bit different (as in Golden-winged and Blue-winged Warblers) and still be nearly genetically identical, whereas birds that look only modestly different (Audubon’s and Myrtle forms of Yellow-rumped Warblers) can be quite distinct genomically.

When it comes to understanding what makes a species, what birders can see through their binoculars is only part of the story.

Hoary Redpoll by Pierre Casavant via Birdshare.

So are birders everywhere about to lose their coveted Hoary Redpoll checkmarks from their life lists? Not so fast, says the American Ornithological Society’s North American Classification Committee. A document updated every July by this 11-member group of ornithologists—the official Checklist of North and Middle American Birds—is the ultimate authority on what is and is not considered a species when it comes to our continent’s birds.

Every year, the committee invites proposals for official taxonomic changes. They don’t only handle lumps and splits, but also issues such as moving species among genera, changing birds’ official English common names, and adding new species that have expanded into the checklist’s geographic area.

“I compile all the proposals into a set and send them out to the committee,” says Terry Chesser, a research zoologist with the USGS Patuxent Wildlife Research Center and the AOS committee’s current chair. “People send back votes and comments on the proposals, and there’s usually some discussion back and forth about some of them. The species-limits ones, the lumps and splits, tend to be more controversial.”

The committee leans conservative in its decisions, preferring to wait for further data in cases where the evidence is deemed intriguing but not yet conclusive. The group generally uses the biological species concept as its guiding principle, though its members recognize the longstanding differences of opinion within the ornithological community on the best way of defining a species.

“I don’t want to do things in haste when there’s no compelling reason why you can’t wait a year to make a decision,” says Chesser. “We try to make considered decisions based on as much evidence as we can, and if for some reason I feel that that hasn’t happened, then I’d rather just put it off for a year and allow us to do a good job. The last thing you want to do is get into a position where you’re reversing yourself.”

Red Crossbill (by Wandering Sagebrush via birdshare) and Cassia Crossbill (by Craig Benkman)Is the Cassia Crossbill (left) a separate species from the Red Crossbill (right) even though they look the same? Cassia Crossbill by Craig Benkman, Red Crossbill by Wandering Sagebrush via Birdshare.

This summer, the committee considered proposals to lump redpolls and split Yellow-rumped Warblers. Both failed to pass, though the vote tallies were very close. The redpoll vote was split five to five. Another proposed split—breaking out the new Cassia Crossbill of southern Idaho from the widespread Red Crossbill—passed on its second time through the AOS committee. A similar crossbill proposal failed in 2009 due to insufficient evidence, but this time the split succeeded due in part to new genomic analyses that underscored the Cassia Crossbill’s distinctness, and also field evidence that hybridization between these and other crossbills is rare.

Louisiana State University ornithologist Van Remsen is a member of the AOS Classification Committee, and he says his votes are guided by the biological species concept. Remsen voted against both the Yellow-rumped and redpoll proposals, though he emphasizes that the research itself was excellent in both cases. But he contends that scientists just don’t know enough about how frequently Hoary and Common Redpolls interbreed in their remote Arctic nesting grounds. Similarly, despite the new genetic data on Yellow-rumped Warblers, he thinks that the ease with which they intermingle in the wild demonstrates that Audubon’s and Myrtle Warblers view each other as members of the same species.

“The biological species concept is intuitively pleasing, because it allows the birds to express their opinions, so to speak,” he says. “If they don’t tell each other apart, why should we?”

Remsen sees the use of genomics for species-level taxonomy as still in its infancy.

“We don’t have any context into which to put the data,” he says. He wants to see more genomic studies done of pairs of birds whose species status is not in question to provide a basis for comparison.

Golden-winged and Blue-winged warblers by David SpeiserAre the Golden-winged Warbler (left) and Blue-winged Warbler (right) one species just because their genes are the same? Photos by David Speiser.

The “one species or two?” questions, as with “Audubon’s” and “Myrtle” Warblers, probably don’t have a single answer. The reason, as more and more members of the ornithology community acknowledge, is that speciation is an ongoing process that plays out over many thousands of years. When scientists study these kinds of situations, they catch species in the midst of this long process.

Lovette is a strong proponent of this viewpoint: “We’re all programmed to think of bird diversity like a page in a field guide, where everything is in clean categories and every species is separated. We don’t often stop to consider the fact that all of those species on a page of warblers or sparrows evolved from the same common ancestor, and they’ve all become diverse through a process of speciation that’s played out over time.”

It usually takes about a million years, he says, to go from one species to two or more.

When you capture that process of speciation in action, you’re naturally going to find in-between cases that are in this gray zone, where by some criteria they’re separate species and by other criteria they’re not.”

Lovette likens it to looking at just one section of a fork in a waterway and trying to determine whether it’s one river or two. You can’t know just by looking at the water at this one point; you need to also know what the river looks like upstream and downstream.

Likewise, when scientists study birds in the process of evolution at a single point in time, what is and is not a species is—occasionally—a judgment call.

“The most wonderful aspect of all of this,” says Lovette, “is that this process is where biodiversity comes from. So instead of being frustrated by our inability to define clear breakpoints between all sets of related species, I’m excited about the chance to explore these many ways that species come to be.”

Birdwatchers. Photos from GBBCPeople watch birds all over the world and keep track of the species they see, but the numbers of species they see may change as official stances on the number of species in existence change. Photos from the Great Backyard Bird Count.

Why, ultimately, do we even need an official checklist of bird species? Is the whole concept eventually going to become obsolete?

“Our names for birds are partly motivated by our need to categorize the world in ways that are useful for humans, and sometimes that’s more simplistic than the actual biological reality that we’re trying to classify,” Lovette admits.

But the checklist isn’t going the way of Dendroica just yet. The AOS’s Chesser puts it succinctly: “People need a standardized taxonomy in order to communicate.”

“We need to be able to give names to things. Whether it’s for the Endangered Species Act or for ecological studies of how many species of birds exist in a certain place, you have to have that organizing principle. You really can’t function without it,” says Lovette, who is also on the AOS Classification Committee.

“I personally voted in favor of both the redpoll lump and Yellow-rumped Warbler split, but I’m not upset that I was in the minority,” he says. “In ornithology, we are very fortunate to have official groups like the AOS committee that give us some order and standardization. While I might not agree with the outcome of every AOS vote, it’s even more important to me that we retain an agreed-upon system for classifying birds.

“And besides, every recreational bird watcher knows that keeping lists is super fun.”

Only one thing’s for sure: With all the possible taxonomic changes being debated, field guide publishers won’t go out of business any time soon.

Rebecca Heisman is a freelance science writer and communications assistant for ornithology journals The Auk and The Condor.