A Hummingbird Does Not Put Hovering in Reverse
Backward flight looks like a simple reversal. High-speed video shows a subtler manoeuvre: the bird reorients its body, stroke plane and wing timing while keeping the same flapping system at work.
In short
What happened. A hummingbird leaves a flower without turning around. It appears to run its hovering machinery backward, but recordings show that the bird changes posture and wing motion rather than reversing a programme.
What it means. The wings still sweep and rotate to push air. Backward travel emerges when the bird redirects the average aerodynamic force by pitching its body, flattening the stroke plane and altering timing across the wingbeat.
Risks and impact. Familiar phrases such as “figure-eight wings” and “the only bird that flies backward” turn a flexible three-dimensional movement into a trick. They can hide both what experiments established and what remains uncertain.
What can be done. Read a slow-motion clip in layers: follow the body angle, the plane traced by the wingtips, the wing’s twist and the direction of travel. Do not infer the mechanism from one blurred frame.
What to watch. Ask whether a claim comes from free flight or a wind tunnel, how many birds were tested, and whether it concerns kinematics, aerodynamic force or metabolic cost. Those are related questions, not interchangeable answers.
Shown as a summary because of your reading settings.
What happened
A hummingbird approaching a flower typically moves forward, hovers to feed and retreats backward. The last movement is not a video played in reverse. During forward flight, air passes over a body and wings already producing lift and thrust. To move backward, the bird must keep supporting its weight while giving the combined aerodynamic force a rearward component.
Hummingbird wings are unusually well suited to redirecting that force. They sweep through a large arc and rotate around their long axes between half-strokes. Both halves contribute to weight support, although not equally. In a 2005 Nature experiment using particle-image velocimetry on hovering rufous hummingbirds, Douglas Warrick, Bret Tobalske and Donald Powers estimated that the downstroke supplied about 75% of weight support and the upstroke about 25%.
That result corrects two easy pictures at once. The bird is not suspended by an invisible cushion, and its upstroke is not simply dead time. Nor does one universal flat figure eight describe the wing. The path and angle of the wing vary with speed, direction and the individual manoeuvre.
What the evidence supports
The clearest direct test of backward flight was published by Nir Sapir and Robert Dudley in the Journal of Experimental Biology on 15 October 2012. They placed five male Anna’s hummingbirds, averaging 4.59 grams, in a wind tunnel. A feeder and respirometry mask let the researchers compare oxygen use and high-speed video during backward flight, forward flight and still-air hovering.
At an airspeed of 3 metres per second, sustained backward flight used oxygen at a rate similar to forward flight at the same speed and roughly 20% below hovering. Backward flight came with a steeper body angle, greater head flexion, a higher wingbeat frequency, a stroke plane flatter relative to the horizontal and altered upstroke-to-downstroke timing. The bird changed a coordinated set of variables; it did not flip a single reverse switch.
One bird was also tested from hovering through backward speeds up to 4.5 metres per second. Its metabolic curve resembled the forward-flight curve at equivalent speeds, but it could not feed while flying backward at 6 metres per second. That observation is intriguing, not a species-wide speed limit: the extended series involved one animal in an artificial airflow.
The whole study was small and constrained. A wind tunnel stabilises airspeed and gives cameras a repeatable view, but a bird withdrawing from a real flower may accelerate, brake, turn and react to gusts. The experiment supports a mechanism and a comparison under specified conditions. It does not reduce every backward departure to one number.
How the story is being framed
To a person watching a feeder, backward flight is a vanishing act: the bill clears the flower and the bird slides away before the eye can separate one wingbeat from the next. The useful observation is behavioural. A straight retreat lets a long bill leave a narrow corolla without requiring the bird to turn in a crowded patch.
To an aerodynamicist, the puzzle is force direction. Weight support must continue while the balance of horizontal force changes. Body angle, stroke-plane orientation, wing rotation and timing therefore matter more than the everyday labels “forward” and “backward.” The air responds to the wing’s velocity and angle, not to the bird’s intention.
To an evolutionary biologist, the manoeuvre belongs to a package: hovering, aerial nectar feeding, specialised shoulder motion and control of a relatively stiff wing. But it is safer to say hummingbirds are exceptional backward fliers than to declare them uniquely capable. The 2012 paper notes transient backward flight in other flower-visiting birds and several insects; what distinguishes hummingbirds is how routinely the manoeuvre is integrated into feeding.
The background
Most birds change flight mode by changing more than flapping frequency. A three-dimensional study published in 2007 filmed five rufous hummingbirds with synchronized cameras at 500 frames per second across airspeeds from zero to 12 metres per second. As speed rose, body and stroke-plane angles shifted substantially, while wingbeat frequency did not change significantly. Many geometric details did.
The researchers described the hummingbird hand-wing as kinematically rigid relative to that of many other birds, with an average wingtip-span ratio of 93%. Flexibility therefore comes strongly from rotation and movement at the shoulder and from reorienting the whole stroke, not from folding the wing like a conventional bird’s recovery stroke.
Later work has pushed the question from motion toward control. A 2022 musculoskeletal modelling study proposed that the primary flight muscles do more than power the wing: they can contribute to deviation and pitching torques, while smaller muscles tune the movement. Because this is a functional model, its torque estimates are hypotheses generated from anatomy and dynamics, not direct recordings of every muscle command in a freely flying bird.
The history of the evidence matters. First, observers saw an apparently impossible retreat. Then cameras separated posture and wing timing. Flow measurements showed how both half-strokes contribute in hover. Metabolic measurements tested whether the manoeuvre carries a hidden energetic penalty. Each tool answered a different part of the same wonder.
Who it touches
Picture an Anna’s hummingbird facing a tubular flower. Its bill enters while its body hangs steeply and its wings beat too quickly for an unaided eye to resolve. When feeding ends, the bird does not need to rotate through 180 degrees in the flower’s doorway. It raises its bill, adjusts its body and redirects the wingbeat forces while continuing to carry its own weight.
From the observer’s viewpoint, the body moves first and the wings become two translucent commas. From the bird’s viewpoint, there is no category called “reverse gear.” There is only continuous control: enough upward force not to fall, enough backward component to clear the flower and enough stability to decide where to go next.
That distinction is the source of the wonder. A manoeuvre that looks like a party trick is an ordinary solution performed dozens of times in a feeding day.
The deeper story
You can make the mechanism more visible without specialised equipment. Find a slow-motion recording with a side view and play it at quarter speed. First ignore the wings and mark the body’s long axis. Does the bird become more upright as it retreats? Then imagine a plane joining the extreme forward and backward positions of the wingtips. Is that stroke plane level with the horizon or tilted with the body?
Next watch a single wing near the end of each half-stroke. The surface rotates; it does not remain at one fixed angle. Finally, follow the bird’s centre rather than its head or tail. A flexed neck can keep the bill aligned with the feeder even while the body changes orientation.
This four-layer reading—body, stroke plane, wing rotation, centre of motion—guards against the most common mistake: treating a two-dimensional silhouette as the aerodynamic explanation. A wing can sweep one way while its angle redirects force another way. A bird can hold its wingbeat frequency nearly steady while reorganising the geometry of the beat.
Use the same discipline when a natural-history claim carries a spectacular superlative. Ask what was measured, in which species, under what conditions and with what comparison. “Backward flight costs less than hovering” is accurate for five Anna’s hummingbirds at the tested speed. It is not a law for every bird, wind or escape manoeuvre.
Something to sit with
- When you picture hovering, do you imagine stillness—or constant correction hidden by speed?
- Which observation would best distinguish reversed motion from redirected force?
- What other animal “tricks” might become more interesting once the mechanism replaces the slogan?
Sources
- Journal of Experimental Biology — Backward flight in hummingbirds employs unique kinematic adjustments and entails low metabolic cost — https://journals.biologists.com/jeb/article/215/20/3603/11039/Backw...
- Nature — Aerodynamics of the hovering hummingbird — https://pubmed.ncbi.nlm.nih.gov/15973407/
- Journal of Experimental Biology — Three-dimensional kinematics of hummingbird flight — https://journals.biologists.com/jeb/article/210/13/2368/16900/Three...
- Proceedings of the Royal Society B — Hummingbird flight control is based on a multi-scale architecture of distributed biomechanical modules — https://pmc.ncbi.nlm.nih.gov/articles/PMC9727662/
We report facts from the sources above in our own words and link to the originals. Interpretation is ours, not theirs.
At 3 metres per second in the 2012 wind-tunnel study, how did the oxygen cost of sustained backward flight compare with the other tested modes?
Five male Anna’s hummingbirds showed similar oxygen uptake in backward and forward flight at 3 m/s, while backward flight was about 20% less costly than hovering. The small wind-tunnel study describes those conditions, not every species or natural manoeuvre.
♻︎ Free to republish
Copy this HTML into your CMS. Credit line and licence are included. Republish our work — free
Every headline has a deeper story. This is ours.
What we are doing here