Do Hummingbirds Remember Humans? What the Research Shows

Search this and you get a page of confident yes answers. We went looking for the study behind them and did not find one.

What we did find is stranger, better measured, and more useful if you actually keep water or a feeder out for these birds.

The short answer

What is demonstrated

Hummingbirds remember individual flowers to within centimeters, remember which ones they have already emptied, and in some tests remember roughly how long a flower takes to refill

What is not demonstrated

That a hummingbird can tell one person from another. We could not find a controlled study on it

Why the gap matters

The two get blended together constantly. A bird with a precise memory for a place is not the same thing as a bird with a memory for your face

What the popular pages are citing

Mostly each other. The memory research they gesture at is real, and it is about flowers

What is actually measured

Susan Healy and Andrew Hurly have spent years testing wild rufous hummingbirds in the field rather than birds in a laboratory. Their review of that work gives numbers, which is more than any of the recognition pages do.

The test

A bird is trained to a rewarded artificial flower. The flower is then removed and the researchers record where the bird goes

With larger flowers

The birds returned to within about 60 cm horizontally and 20 cm vertically of where it had been

With smaller flowers

Accuracy improved to roughly 20 cm horizontally and around 5 cm vertically

How fast

One trial. The review describes birds remembering a rewarded location from a single visit lasting only a few seconds

Position or color

When locations stayed put, the birds went to the place rather than matching the color of the flower

That last line is worth pausing on if you have ever been told hummingbirds come for the color red. In these tests, when the two cues disagreed, the birds went by where the flower had been.

These were wild birds on their own territory

This is the part that separates the hummingbird work from everything else in this series, and it cuts both ways.

The other studies we have written up

Captive ravens in aviaries, captive bred birds in a trading task, city pigeons in a park, mockingbirds at nests on a university campus

This work

Free living rufous hummingbirds tested at artificial flowers inside their own territories, with no cage and no training program beyond the task itself

What that buys you

None of the results can be waved away as an artifact of captivity. The bird chose to be there and could leave at any point

What it costs you

Far less control. Nobody knows what a wild bird did between trials, and the sample sizes are small. Three males in the refill experiment, eighteen in the four flower study

Small samples of wild animals doing a natural thing is a different kind of evidence from large samples of captive animals doing an imposed one. Neither is automatically better. It is worth knowing which one you are being handed.

They track time too, but not as well

Two results sit next to each other here and they do not say quite the same thing. We are going to give you both rather than pick the tidier one.

In one experiment described in the review, three territorial males learned within a single day that some flowers refilled every ten minutes and others every twenty. They came back to the ten minute flowers at around ten minutes and the twenty minute flowers at around twenty.

In a separate study run by Healy with colleagues at Lethbridge and Edinburgh, eighteen wild male rufous hummingbirds were given an array of four flowers, one of them rewarded, in the morning and again in the afternoon. The birds handled the what and the where well. Healy's own summary of that work was that the time component was the hardest part for them to learn.

What nobody appears to have tested

We went looking for the hummingbird equivalent of the experiments we have written up for other species, and came back empty handed.

The crow design

A rubber mask worn by different people, so the person underneath cannot be the explanation. Covered in the crow article

The pigeon design

Two matched researchers, one friendly and one hostile, who then swap lab coats. Covered in the pigeon article

The mockingbird design

The same nest approached the same way on successive days, with the person swapped on the last day. Covered in the mockingbird article

The hummingbird equivalent

We could not find one. Not a mask study, not a swap study, not a field trial with matched observers

That is not proof that hummingbirds cannot do it. It is an absence of evidence, and the honest thing is to call it that rather than fill the hole with a story.

Why the belief is so persuasive anyway

Because the thing they demonstrably do looks identical from a kitchen window.

A bird that holds a location to within a few centimeters, remembers whether that location is currently worth visiting, and has some grip on the schedule will appear at the feeder shortly after you refill it. From where you are standing that reads as a bird that knows you. What the research supports is a bird that knows the place, the state of the place, and something about when the place pays out.

You are usually part of that scene rather than the subject of it.

What this changes about your setup

1. Do not move things. The precision numbers cut both ways. A feeder or a basin shifted a meter is, to the bird, gone. Expect hovering at the old spot before it finds the new one.
2. Keep it in the same state. Emptied flowers get skipped. A feeder that is dry half the time is teaching the same lesson an emptied flower teaches.
3. Stop leaning on red. Color is a signal, but in these tests position beat color when the two were put in conflict. A red fixture in a new place is still a new place.
4. Be boring about timing. The evidence on time is the weakest of the three, but it is there. Topping up at roughly the same time is free and it is the only lever you have that the research actually speaks to.

Where water fits

The Cornell Lab lists the diet as nectar from red or orange tubular flowers such as trumpet creeper, cardinal flower and honeysuckle, plus feeders and tree sap, along with insects like mosquitoes, gnats and fruit flies taken in flight or off spider webs.

Water is not on that list, and a conventional basin is the wrong shape for a bird that weighs about as much as a coin. What works is a shallow film or a moving mist, and we have written both of those up already in the hummingbird bird bath guide and the guide to making a hummingbird mister bath.

The finding above does not change what to buy. It changes where to put it and how often to move it, which is never.

Cornell gives a sense of the scale involved. The nest is about two inches across, built from thistle and dandelion down bound together with spider silk, usually 10 to 40 feet up on a slender branch, holding one to three eggs with an incubation of 12 to 14 days. An animal operating at that size is not going to use a basin built for a robin, and the memory numbers above start to look less like a curiosity and more like a requirement.

What we could not check

What we read in full

The Healy and Hurly review of their field work, the EarthSky report of the four flower study, and the Cornell Lab account for the Ruby-throated Hummingbird

What we did not read

The four flower paper itself. The numbers and the quote from it come from the report, and we have said so where we used them

One claim we went looking for and could not stand up

That individual hummingbirds return to the same yard or the same feeder in later years. It is repeated everywhere. The Cornell life history page we read does not say it, and we are not going to assert it on the strength of repetition

Frequently asked questions

Does it remember my feeder from last year

We could not confirm this from anything we could open. The Cornell life history page is silent on individuals returning to the same yard, and the memory research above is about flowers within a season. Widely repeated is not the same as demonstrated.

It comes over when I walk outside. Is that not recognition

It might be. Nothing we found has tested it. The alternative explanation, that the bird has learned a place and a rough schedule and you happen to be in the frame, is the one with measurements behind it.

Do they really remember which flowers they already drained

Yes. That one is directly measured and it is the least surprising of these findings to anyone who has watched a bird skip a feeder port.

Does the feeder have to be red

Not according to these tests. When position and color were put in conflict, position won. Red helps a bird find something new. It does not override a location it already knows.

Will a hummingbird hold a grudge the way a crow does

There is no study we could find either way. The corvid results are covered in the raven article, and we would not carry a corvid number across to a bird this different.

References

1. Reference: What hummingbirds can tell us about cognition in the wild | Healy and Hurly, Comparative Cognition and Behavior Reviews, volume 8, 2013
2. Reference: Hummingbird memory | EarthSky, reporting the four flower study led by Susan Healy at the University of St Andrews with colleagues at Lethbridge and Edinburgh
3. Reference: Ruby-throated Hummingbird Life History | The Cornell Lab of Ornithology, All About Birds