The Reality of Birds and How to Deal with It
This question was driven home for me a couple of autumns ago when I had the nerve to suggest, not for the first time, that the traits of a bird posted online “favored” Alder. (I mentioned the grayish and greenish, rather than brownish, tones in the plumage; the rounded profile of the head; and the relatively long extension of the primaries.) Although a number of commenters agreed, one took the opportunity to release what was clearly an overstuffed backlog of resentment. Mocking laughter and a demand for “diagnostics” ultimately gave way to the conclusion: “I cannot take you seriously. I think you don’t actually go look at birds.”
The gauntlet had been thrown.
Well, I didn’t feel the need to defend the actuality of my looking at birds, but it did get me thinking about the different ways in which this person and I might be looking at birds. Alder vs. Willow is such a potent flashpoint because it activates some deep fault lines. One tracks between an older generation that came up when these two species were lumped vs. a younger generation eager to topple the received wisdom; the split was formalized in 1973, but further decades elapsed before visually distinguishing the species was considered anything but heresy. Another tracks between the conservative, methodical mindset of the bander and the more freewheeling, less accountable modus operandi of the field birder. These are caricatures, but banders are often found leading the charge of insistence that Traill’s Flycatchers should be left unidentified. And with good reason. After all, decades of study have yet to produce a single magic formula that can reliably put a given bird in the correct bin. Ornithological statisticians have analyzed as many as 14 different variables—but good luck applying those when a bird is in the hand, let alone 60 yards away across a marsh in the wind.
And yet, in flagrant defiance of this, an increasing number of top birders—and even decidedly-not-top birders such as myself—have come to believe that they can ID many individual Traill’s with confidence. How much confidence? Enough to go on record, that is, to mark the species down on a checklist. We’ll return to that later. But, for now, let’s simply agree that, in defending these calls, we cannot appeal to “diagnostics.” Empid ID in general, and Traill’s ID in particular, is instead a game of probabilities. Each field mark on its own is subtle, relative, and overlapping. But by noting which way each one tilts and toting up these weights, we can tip the scale decisively one way or the other. Such is the claim.
For some birders, this method will never suffice. In the absence of strict rules, the whole process has the taint of subjectivity. Without calipers and a colorimeter, you can’t define the shape of a flycatcher’s head, or where its mantle feathers sit on the brown–green spectrum, with any precision—and thus any attempt to compute probabilities amounts to “GIGO” (garbage in, garbage out). There are two things I’d like to submit to these skeptics. One is that to underrate our ability to make subtle distinctions is to do ourselves a colossal injustice. Statistical analysis may seem like the domain of supernerds, but it’s child’s play compared to what your human brain is already doing a thousand times a day.
Consider the fusiform gyrus. It is this region of the cerebral cortex—more specifically, a cluster of neurons known as the fusiform face area—that takes on one of the most challenging tasks to be attempted by any computer, whether carbon- or silicon-based, and handles it with astonishing aplomb. The software running here can take a snapshot of a human face, even at a distance, at a funky angle, in half-shadow, etc., and classify it within milliseconds as your niece or your next-door neighbor or, for that matter, Patrick Mahomes or Taylor Swift. Could we define the “diagnostics” that it uses? Not without terrific effort. By dissecting Mahomes’s or Swift’s face millimeter by millimeter, we might pick out some key variables—a slightly steeper-than-average angle in the distal eyebrow or what-have-you. But when you consider the number of variables involved, and the tremendous number of acquaintances and celebrities that we effortlessly recognize, the mind begins to boggle at the complexity of the computations.
How does the brain do it? We don’t know. Reverse-engineering this “program” has been one of AI’s holiest of grails, and considerable effort has leveraged a statistical technique known as principal component analysis. The basic idea is first to digest thousands of real faces, with countless variables to differentiate them, and pin down which combination of those variables packs the most useful information. It’s as though you were somehow holding a many-dimensional cloud of data points in your hands, and turning it until you find the angle that spreads them out for the best viewing. That key can then be used to translate the unwieldy dataset into a much smaller set of “pseudofaces,” serving as a sort of stylized gallery of human variation. Then, the next time a new image of a face appears, it can be quickly held up against each pseudoface and scored for similarity. The final positioning of the new face among the pseudofaces, and comparison of this result to a memory bank containing the scores of your niece, Patrick Mahomes, Taylor Swift, and so on, can result in a very accurate and efficient classification.
Whether the brain uses these specific tricks is ultimately beside the point. The proof is in the pudding, in our ability to function as social animals without constant embarrassment—not to mention fear of bodily harm. But beyond the implications for social standing and personal safety, what does any of this have to do with birding?
As a study by neuroscientist Isabel Gauthier and coauthors, published in 2000 inNature Neuroscience, showed, the fusiform face area is not just for faces. For this study, the authors recruited subjects with varying levels of expertise in identifying cars and birds, and showed them pictures of these while monitoring their brain activity via functional MRI.
The Gauthier team confirmed their suspicions with a remarkable finding: They could predict a subject’s level of bird ID expertise by watching how brightly their fusiform face area lit up when inspecting an image. In other words, the better we know birds, the more we come to recognize them as familiar faces.
This insight does much to explain the communication gap we so often see between expert and novice. Many times I have watched an advanced birder patiently explain an “obvious” ID that the beginner being “helped” confesses to not understanding. There’s no gap in intelligence here; they’re just using different parts of their brains. The novice is following conscious rules, memorizing field marks and applying them one by one, while the expert has developed enough comfort over the years to reassign much of this work to the fusiform gyrus. There, advanced methods deliver a verdict that feels intuitive, and yet is grounded in a sophisticated analysis of a riot of variables.
As I can confirm from my share of missteps, with flycatchers and otherwise, the brain is not infallible. There are always more data points to gather, more new variables to notice and process, in order to refine the recognition model. There’s also no substitute for calibrating your model against others’, which is the beauty of ID debates on crowdsourced databases and social media.
The takeaway here is that there is value in trusting your “gut,” in cultivating your “intuition”—for these much-maligned words are really just our best available shorthand for some of the most complex and advanced tools we possess. If a bird labeled as Willow Flycatcher somehow looks “not quite right” for the species, that means there’s dissonance with the internal model of Willow Flycatcher that you’ve developed. Spend some time with that feeling, and see if you can pinpoint the parts of the bird that stretch against your expectations. Perhaps it’s something like the slope of the forehead—an underrated mark in empid ID. It could be that your model needs updating; or maybe your analysis, humming along constantly like an underground, cutting-edge research lab, has uncovered something that others have missed.
Another thing nags at me during arguments over empids and the like. It has to do with our standards of truth. A skeptic’s line of thinking goes like this: There is some overlap in every individual field mark we invoke in separating Willow and Alder flycatchers. Case in point: Cin-Ty Lee and Andrew Birch’s widely noticed 2025 paper in Western Birds, which reports that a primary (p) spacing formula, the ratio of p7 projection to p6 projection on the closed wing, did a good job of indicating species—but also that some 5% of birds defied the rule. So the skeptics gleefully declare that no combination of those field marks can deliver 100% certainty. Case closed, they say; we cannot identify Traill’s Flycatchers by sight.
But 100% certainty is a fool’s errand. To enforce it would be paralyzing. Each of us violates it multiple times a day—as when I identify a flyby American Crow by sight, disregarding the nonzero chance that it is a vagrant Fish Crow; or I identify a warbler by sound, disregarding the decidedly nonzero chance that a bird of one species has learned another’s song; or I mark down a titmouse on my checklist, dimly recalling it from earlier in the morning, and disregarding the nonzero chance that my muddled brain has pasted in that memory from yesterday.
In scenarios like the preceding, we might console ourselves with the belief that there is a truth of the matter, however ill-equipped we were to find it out in the moment. The crow, the warbler, and the titmouse I thought I observed either were there or weren’t. The Alder Flycatcher I thought I recognized on the internet either was or wasn’t. This isn’t Schrödinger’s cat.
In recent years, the emergence of DNA sequencing tools has given a shiny new veneer to the notion that a bird’s “true identity” can always be determined in principle. Before Cordilleran and Pacific-slope flycatchers were mercifully re-lumped in 2023, there was a brief period when some birders encountering them out of range went to heroic lengths to collect fecal samples for that purpose. But this has also opened a whole new can of worms. We now must wrestle with the prospect that this “true identity” might diverge from anything we’re able to evaluate in the field. We know, for example, that what we thought of as Wandering Albatross—one of the world’s most iconic species—is in fact a set of populations each genetically distinct enough to warrant species status. These populations may be segregated during the breeding season into colonies that are oceans apart, but they roam across thousands of miles in the interim, and for some of the new species we have no way of reliably distinguishing them without a DNA sample and a laboratory. The years ahead will surely reveal more such cryptic species, putting birders in an awkward unprecedented bind.
On the other side of the coin, the “answers” from DNA are often murky. The thought of the secret existence of Willow x Alder hybrids, a possibility we’ve swept under the rug until now, might fill you with terror. Now we have evidence that this unholy union is a reality. A study by postdoctoral scientist Jordan B. Bemmels and coauthors, published in 2021 in Molecular Ecology, took samples from both species where their breeding ranges overlap, and found some evidence of hanky-panky most everywhere they looked. In one corner of British Columbia, every Willow Flycatcher tested was at least a few percent Alder. Interestingly, the species’ overlap there is thought to be relatively new, postdating the retreat of the glaciers. In the northeastern U.S., where they’ve been rubbing shoulders far longer, they seem to have worked out a better—though still not perfect—system for keeping their romantic distance.
The ID implications make this case especially threatening, but it’s far from the only case. Already notorious among many birders is that of Blue-winged and Golden-winged warblers, which unlike the flycatchers are visually unmistakable, but have turned out to be hiding a mess of dirty secrets. The two species regularly hybridize, creating distinctive first-generation offspring. Either warbler species has the tricksy habit of sometimes singing the other warbler’s song. On top of all that, we’ve discovered in the past quarter-century just how thoroughly admixed these species’ genomes are.
At one study site in Pennsylvania where both warblers breed, ornithologist Frank Gill reported in a 1997 paper in Evolution that every sampled bird presenting as Golden-winged turned out to have mitochondrial DNA that would classify it as Blue-winged! What does this mean? A bird’s mitochondrial DNA is inherited separately from its “ordinary” (nuclear) DNA, and passed down solely from its mother. This tells us that somewhere on each of these birds’ maternal lines can be found a female Blue-winged ancestor. Does that qualify it as a hybrid? How far back would the interbreeding event need to be pushed for the descendant to be considered “pure”? If a bird is likely smuggling genes from another species, but those genes don’t manifest in any visible way, can I count it on my life list? Do “Blue-winged” and “Golden-winged” warblers really even exist?
These questions have no clear answers—and when pushed too far, to my ears, they carry echoes of racist reasoning. To venture down the road of enforcing purity, or of establishing arbitrary thresholds for qualification as a species, is to overstep our bounds in imposing order on nature. Better to accept that a species simply is not what we were raised to believe. It’s not a box, inside which an individual either is or isn’t. Instead, it’s more like a cloud—a drifting assemblage that is recognizable for a while as an entity, but with an ever-changing shape and fuzzy edges liable to blur into its neighbor.
Birds generally fall neatly into visual—and aural—bins that permit us to ignore this truth most of the time. Botanists studying oaks, or mycologists studying fungi, do not have the same luxury. And even a birder, if not already shaken by the cases presented above, should be quickly cured of species purism by a visit to a place like Daphne Major. That’s the parched islet in the Galápagos archipelago where, thanks to the dogged efforts of evolutionary biologists Peter and Rosemary Grant, we’ve laid bare some of the innermost workings of natural selection. Trying to publish a definitive field guide to the complex “finch” fauna—they’re actually tanagers, we now know—of Daphne Major would be laughably futile. But the situation in stable North America differs only in degree. All we can hope to do is capture a limited snapshot of the diversity that surrounds us, labeled with categories that map loosely—for now—onto the turbulent flow of genes.
So where does all of this leave us as birders? We might feel discouraged, knowing that the truth about a bird’s identity is hopelessly out of reach. But I find it oddly liberating. The traditional life list has lost its hold on me, although it’s still a fun game. More fun now is the practice of cataloging all the identifiable phenotypes I can find and distinguish. That includes formally recognized species or subspecies, as well as hybrid variations within tangled complexes like that of Blue-winged and Golden-winged warblers, and the beings that don’t yet have a name or a status. Like the “doublet” vs. “triplet” dialects sung by White-throated Sparrows, an astounding example of rapid cultural coevolution reported in a 2020 paper in Current Biology by Ken A. Otter and coauthors. Or the mysterious, promiscuous variation in American Herring Gulls and related species. Or even—cover your ears—the slightly different flavors of Alder Flycatchers that I suspect can be picked out across their range. The list goes on.
The types we can pin down in the field won’t always jibe with the boundaries outlined by genetic analysis. But it’s exciting to look forward to how these practices will inform each other more deeply over the years ahead. It’s humbling, too, to be reminded by the DNA that the birds themselves always know things we don’t—and that what we think defines a species can change right under our noses.
If a “truth” composed of fuzzy shades, rather than bold lines, makes you anxious, take heart: We evolved to make sense of this world, and all of us have an extraordinary toolkit in our brains for doing so. What you notice in a glance about a Willow Flycatcher, however conscious or unconscious it may be, could be as penetrating and significant as any formula honed by the experts. So take your field guide as just that: a set of guidelines, rather than ironclad rules. Know that “reality” will always be more complicated, more interesting, more resistant to our feeble attempts to organize. Pay attention to what catches your eye. Defend it against the skeptics. They might learn something from your way of looking.
Works Cited
Bemmels, J. B., A. C. Bramwell, S. A. S. Anderson, V. E. Luzuriaga-Aveiga, E. K. Mikkelsen, and J. T. Weir. 2021. Geographic contact drives increased reproductive isolation in two cryptic Empidonax flycatchers. Molecular Ecology 30: 4833–4844.
Gauthier, I., P. Skudlarski, J. C. Gore, and A. W. Anderson. 2000. Expertise for cars and birds recruits brain areas involved in face recognition. Nature Neuroscience 3: 191–197.
Gill, F. B. 1997. Local cytonuclear extinction of the Golden-winged Warbler. Evolution 51: 519–525.
Lee, C.-T., and A. Birch. 2025. Identification of Willow and Alder flycatchers by primary-tip spacing: The p6:7 ratio. Western Birds 56: 21–44.
Otter, K. A., A. Mckenna, S. E. LaZerte, and S. M. Ramsay. 2020. Continent-wide shifts in song dialects of White-throated Sparrows. Current Biology 30: 3231–3235.
Matt Smith is a software developer at the Cornell Lab of Ornithology, as well as a lifelong birder and writer. He has published two children’s books and various essays on science and nature, and he created the web game “Fantasy Birding.” Matt lives near the Blue Ridge Mountains in Virginia.
