In 1833 a naturalist looked at a hummingbird’s tongue, saw two narrow grooves, and concluded that nectar must rise up them the way water climbs a paper towel.
That answer stood for about 180 years. It was taught, it was modelled mathematically, and it was not directly tested with high-speed video of free-feeding birds. When somebody finally pointed a high-speed camera at the problem, the answer turned out to be wrong.
What replaced it is stranger and better. The tongue is not a tube. It is a pump, and the bird’s own beak is what powers it.
Quick Facts
What the tongue is: forked at the tip, with two long open-sided grooves running down each fork
What it is made of: the distal half is entirely keratin, the same protein as your fingernails and the bird’s feathers
How thin: finer than fishing line
How far it reaches: roughly a bill’s length beyond the bill tip
How fast: 15 to 17 licks a second in most species, with rates up to 20 recorded
Muscles inside the flower: none; the working end of the tongue has no muscle of its own
Current model: elastic micropump (a structure that stores elastic energy when flattened and uses that energy to draw fluid in when it springs open), established 2015, refined 2024
Previous models: capillary action from 1833 to 2011, then a fluid trap from 2011 to 2015
What the Tongue Actually Is
Start with the anatomy, because it is odd enough without embellishment.
A hummingbird’s tongue is forked. Each of the two forks is not a solid prong but an open-sided groove, a channel with a slit running down its length. Near the very tip, the edges of those grooves break up into fine hair-like fringes called lamellae.
The whole structure is transparent and thinner than fishing line. The outer half of it is pure keratin, which is the same material as a fingernail, a hoof or a feather. That matters for what comes later. Keratin is springy, and the mechanism depends on it.
Now the part that surprises people. The portions of the tongue that go inside a flower have no muscles at all. There is nothing in there to squeeze or pump with. Everything is driven from outside the bill.
The Apparatus That Drives It
The tongue is pushed and pulled by the hyoid apparatus, a set of slender bones and muscles that extends well beyond the mouth. In hummingbirds and woodpeckers these bones are enormous relative to the skull. They run back from the floor of the mouth, wrap around the back of the braincase, and anchor near the base of the upper bill close to the nostril.
When the muscles contract, the bones slide forward and shoot the tongue out of the bill. It is a mechanical linkage rather than a muscular one, which is how a bird with no muscle in its tongue tip can reciprocate it seventeen times a second.
You will sometimes read that a hummingbird can touch the back of its own head with its tongue. That is a garbled version of this. The bird does not lick its skull. The machinery that drives the tongue is wrapped around it.
The 180-Year Mistake
Capillary action is real, and it is easy to see why it looked like the answer. Dip a thin glass tube in water and the water climbs it, pulled up by its attraction to the walls. A hummingbird tongue has two thin channels. Nectar appears in them. Case closed, apparently, from 1833 onward.
The idea hardened in 1983, when Kingsolver and Daniel published a mathematical model of hummingbird feeding built on capillary rise. That model became the standard reference. It was used to predict which nectar concentrations hummingbirds ought to prefer, and it shaped decades of thinking about what hummingbirds eat and why.
Two things should have raised doubts.
Capillary rise is slow, and it gets slower. The further liquid climbs a narrow tube, the more the friction of the walls resists it. A tongue that filled by capillarity would take longer with every millimetre. Hummingbirds lick fifteen times a second and show no sign of that penalty.
The predicted diet was wrong. Capillary models put the ideal nectar for a hummingbird somewhere around 20 to 40 percent sugar. Offered a free choice, hummingbirds take solutions well above that, and well above what their own flowers produce. The theory and the birds disagreed, and the birds were not consulted.
Nobody had filmed the mechanism. The tongue is two centimetres of transparent tissue moving in a fraction of a blink, inside a flower, and until high-speed video became cheap there was no way to look.
2011: Not a Capillary Tube
Alejandro Rico-Guevara and Margaret Rubega, then at the University of Connecticut, built transparent artificial flowers and filmed wild hummingbirds drinking from them at more than 500 frames a second.
They recorded thirty birds across ten species, and examined twenty more specimens under the microscope. What the film showed was nothing like capillary rise.
The moment the tongue tip touches nectar, the fringes unfurl, the two forks separate, and the tip spreads outward across the liquid surface. As the tongue is withdrawn, the fringes curl back inward and the tips close, trapping nectar inside the furled structure. The tongue is not wicking liquid up. It is grabbing it.
Better still, the closing is passive. No nerve signal, no muscle, no effort. Pulling the tongue across the surface of a liquid is enough to make it shut. The researchers called it a fluid trap, and their 2011 paper in PNAS was titled accordingly.
That paper did two things. It ended the capillarity era, and it was still not the final answer.
2015: An Elastic Micropump
Four years later the same group, now joined by the mechanical engineer Tai-Hsi Fan, published the mechanism that is current today. They had expanded the filming to eighteen species, covering seven of the nine main hummingbird groups across the Americas.
Here is what actually happens, in order.
On the way out. As the tongue is pushed forward through the bill, the nearly closed bill tips squeeze it flat. The two grooves collapse. Squeezing a springy keratin channel flat stores energy in its walls, exactly as bending a paperclip does, and the walls strain to return to their relaxed shape.
On contact. The flattened tongue touches nectar. The grooves spring open. Because the walls are recovering stored elastic energy rather than relying on surface attraction, they pull liquid in fast, and they fill completely rather than stopping at a meniscus.
On the way back. The tongue is drawn into the bill, which squeezes the nectar off it and simultaneously flattens the grooves again, loading the next stroke.
That is a pump. The energy for each fill was stored on the previous withdrawal, and the compression stroke is performed by the beak. It fills roughly an order of magnitude faster than capillarity could manage, which is what makes seventeen licks a second physically possible.
It also resolves the diet mismatch. A pump does not care about viscosity the way a capillary tube does, so thick nectar costs a hummingbird far less than the old model implied.
On the micropump numbers they should be able to exploit nectar in the 45 to 65 percent range, far above the 20 to 40 percent capillarity allowed. That prediction sits much closer to what choice experiments actually find. It is one reason the standard sugar water recipe works as well as it does.
Note the order carefully, because it is usually described wrong. The compression and the withdrawal are not two steps. They are the same step.
2024: The Bill Bends Too
The most recent chapter is barely a year old and has not yet reached most popular accounts.
A hummingbird’s bill looks like a rigid needle. It is not. In November 2024 Rico-Guevara and colleagues published micro-CT scans and video shot at 1,000 frames a second. They showed that the upper bill has a flexible zone in the middle, with stiffer regions ahead of it and behind it. The bird can bend it deliberately while drinking.
They filmed six species, among them Anna’s Hummingbird, at transparent feeders in Colombia, Ecuador and the United States.
The sequence they describe is a three-part choreography repeated ten to fifteen times a second:
- The bill tip opens to let the tongue out.
- The tip closes behind the returning tongue and scrapes the nectar off it, like a squeegee.
- The middle of the bill stays sealed while the base opens slightly, enlarging the space inside the mouth and drawing the nectar back toward the throat.
The consequence is that a hummingbird uses two different kinds of fluid flow at once. Shear-driven flow, the kind a dog produces when it laps, moves nectar on the tongue’s retraction. Pressure-driven flow, the kind a straw produces, moves it along the bill. Most animals use one or the other. As Rico-Guevara put it, hummingbirds are a rare example of using both.
So the tongue was never the whole story. It is a tongue-and-bill system, and the bill is doing more work than anyone credited.
Why the Speed Matters
Seventeen licks a second is a nice statistic. It is also the reason the rest of the bird can exist.
A hummingbird runs the highest metabolic rate of any bird, and it has to refuel through the day at intervals of roughly ten to fifteen minutes. Hovering at a flower is expensive, so every second spent at the blossom has to return more energy than it costs. A slow tongue would make hovering a losing proposition and the whole way of life impossible.
The pump is what closes that equation. Full grooves, emptied and refilled seventeen times a second, deliver enough sugar in a few seconds of hovering to justify the flight.
The same arithmetic runs in the other direction overnight. With no feeding possible in darkness, the bird drops into torpor and lets its body temperature fall rather than burn reserves it cannot replace. Fast feeding by day and shutdown by night are two halves of one strategy, and both follow from the same energy problem. It is among the clearest examples of what sets hummingbirds apart from other birds.
The feeding rate also explains the pace of autumn. A bird putting on fat before migration has only so many daylight hours, and the tongue is the bottleneck it has evolved around.
Engineers Are Copying It
This is the part of the story that is still being written.
The hummingbird tongue solves a problem that matters in medical devices: moving a very small, precisely controlled volume of liquid without a motor, a battery or a pressure source. Most laboratory microfluidics does this with pumps, centrifuges or absorbent paper. All three have drawbacks at small scale.
In May 2025 a team in Brussels published a design in Nature Communications for grooved elastic sheets modelled directly on the hummingbird tongue. Dipped into liquid, the grooves first fill by capillary rise, then the whole sheet curls shut and traps more liquid in the core. Closing the structure more than doubles the volume captured compared with rigid open grooves.
The authors propose it for point-of-care diagnostics and blood sampling, and they demonstrated it with treated blood. It has a second trick as well. The time the structure takes to close is directly proportional to the liquid’s viscosity, so the same device can measure the fluid it is collecting.
The researchers checked their design against the real thing. Running the hummingbird tongue’s own dimensions through their model puts the bird comfortably inside the regime where the closing mechanism works. That is a tidy piece of confirmation from an unexpected direction.
What This Does Not Explain
Two things readers often attach to the tongue belong elsewhere.
Tongue and mouth problems in hummingbirds are nearly always candidiasis, a yeast infection that swells the tongue until the bird cannot feed. It is a feeder-hygiene issue rather than an anatomical one, and it is covered in full on the hummingbird health and diseases page.
Flowers and hummingbirds shaping each other is coevolution, and the pressures now acting on that relationship are habitat loss and climate shift rather than anything to do with tongue mechanics. Those sit with the other threats to hummingbirds.
How to See It Yourself
You will not see the grooves with the naked eye. You can see the tongue.
A hummingbird at a feeder flicks its tongue out past the bill tip repeatedly, and at close range in good light it shows as a fine dark thread. Phone cameras at high frame rates now catch it reliably, which was not true a decade ago.
Clear feeder ports help, and so does a perch, because a settled bird feeds for longer than a hovering one. The rest is proximity, which is a question of attracting hummingbirds in the first place and of having the feeder out at the right time of year. It is worth knowing which species you are likely to get, since bill length varies and the longer-billed birds give you more tongue to look at.
Frequently Asked Questions
How do hummingbirds drink nectar?
Not by suction and not by capillary action. The bill squeezes the tongue’s two grooves flat as it goes out, storing elastic energy in the groove walls. On contact with nectar the grooves spring open and draw liquid in, then the bill scrapes it off on the way back. The tongue works as an elastic pump powered by the beak.
How fast can a hummingbird lick?
Most species manage 15 to 17 licks per second, and rates up to 20 have been recorded. A Ruby-throated Hummingbird runs at around 13 a second.
How long is a hummingbird’s tongue?
It can be extended roughly a bill’s length beyond the bill tip. The figure varies a great deal with species, since bill length across the family ranges from stubby to longer than the bird’s body. The extremes of hummingbird size give some sense of the range involved.
Is a hummingbird’s tongue forked?
Yes. It splits into two prongs, each an open-sided groove, with fine fringes called lamellae near the tip. The fork is not decorative. It roughly doubles the area that contacts nectar.
Do hummingbirds suck nectar through their beaks like a straw?
No, though the 2024 research shows the bill does generate some straw-like pressure flow internally. The nectar is collected by the tongue and then moved back toward the throat by the bill opening and closing in sequence.
Can hummingbirds taste?
Yes, and unusually. They lost the standard vertebrate sweet receptor and repurposed their savoury receptor to detect sugars instead. That is a separate piece of sensory rewiring from how they see colour, and both help them judge a flower before committing to the hover.
Conclusion
Three models in two centuries, and the current one is only a decade old.
Capillary action was a reasonable guess that nobody checked for 178 years. The fluid trap of 2011 was right and incomplete. The elastic micropump of 2015 is the mechanism as we now understand it. The bill-bending work of 2024 showed that the beak is a working part of that pump rather than a passive sheath.
It is worth remembering how recent all of this is. Millions of people watch this bird in their gardens. It sits in every field guide and on every feeder, and until 2015 nobody had correctly described how it drinks. There is more of that left than we tend to assume. Some of it is at the end of a garden, happening seventeen times a second.
Quiz: How Hummingbird Tongues Work
The questions in this quiz are grounded in the published research and reference sources listed below:
The paper that ended the capillarity era. Thirty birds across ten species filmed at more than 500 frames a second, showing the lamellae unfurling on contact and furling again on withdrawal, with no muscular effort involved.
The current model. Eighteen species from seven of the nine main hummingbird groups, the elastic energy stored in the groove walls as the bill squeezes the tongue flat, and filling roughly an order of magnitude faster than capillary action allows.
Micro-CT scanning and video at 1,000 frames a second across six species, revealing a flexible zone in the upper bill and the three-part opening and closing sequence that moves nectar toward the throat.
Review article covering the keratin composition of the outer tongue, the absence of muscle in the parts that enter a flower, the Kingsolver and Daniel capillarity model of 1983, and the nectar concentrations hummingbirds actually prefer.
Biomimetic engineering published in Nature Communications. Grooved elastic sheets that curl shut on withdrawal, capture more than twice the liquid of rigid grooves, and double as a viscosity probe for point-of-care diagnostics.
Plain-language account of the 2024 bill-bending study, including the filming locations in Colombia, Ecuador and the United States, and the comparison of the two fluid flow regimes a hummingbird uses at once.
Describes the hyoid apparatus in hummingbirds and woodpeckers, the paired bones that wrap around the back of the skull and anchor near the nostril, and the licking rate recorded in Ruby-throated Hummingbirds.