Skip to content
NatureHQ

Birdsspecies

Homing pigeons

Columba livia domestica

A century of experiments on one bird, and the finding is that there is no single mechanism: what a pigeon uses depends on whether it recognises where it is.

No single mechanism explains pigeon homing, and a century of experiments is why we know that. Over familiar ground a pigeon follows a learned route; from somewhere it has never been, taking away its sense of smell leaves it unable to choose a direction.

The homing pigeon has been the laboratory animal of navigation for a hundred years, and the main lesson from that literature is that the question "how does a pigeon find its way home" has no single answer — it has answers that depend on where the bird was released and how much it already knew. From an unfamiliar site a pigeon must work out where it is before it can choose a direction, and the evidence that this step involves smell is substantial and still contested: birds made unable to smell choose poor bearings from unfamiliar places, while their compass orientation and their homing over familiar ground are unaffected. Whether the odours form a genuine gradient map, and which compounds they are, remains open — the objection that has never been fully answered is that every method of removing a bird’s sense of smell is invasive and might impair something else. Over familiar ground the picture is different and clearer. GPS tracking shows that experienced pigeons stop flying the straight line and start flying their own idiosyncratic routes, repeatable to within a few metres, and clock-shifting them does not knock them off those routes — which a bird still steering by the sun would have been. Layered underneath both is the sun compass, whose time compensation is the most quantitatively exact result in the whole field.

Developed coverage · 51% complete · reviewed 2026-09-03

What this page covers

Domestic pigeons bred for homing. They are the most experimentally studied navigator in biology, and they are also a selectively bred domestic animal, which is the standing caveat on everything learned from them.

Often confused with: Wild rock doves and city pigeons, which are the same species and are not bred for this; Migratory birds, which solve a different problem — pigeons are homing, not migrating

Quick facts

From unfamiliar ground
Removing smell impairs the choice of direction
Over familiar ground
Learned, idiosyncratic routes — repeatable to a few metres
The compass
The sun, time-compensated; shift the clock and the heading rotates
The standing caveat
A domestic bird bred for exactly this behaviour

What homing pigeons can detect

Assembled from the evidence graph. Each entry is a link the corpus can justify, with the limits it was recorded under.

The forty-year argument about smell

It sounds implausible, the experiments keep working, and the objection to them has never been fully answered.

How we know

Taking away a pigeon’s sense of smell, and choosing where to release it

Smell seems an unlikely way to find a place a hundred kilometres away. Does removing it actually stop a pigeon?

Across four decades, pigeons were made unable to smell — by cutting the olfactory nerve, by anaesthetising the nasal cavity, or by washing the epithelium with zinc sulphate — and released alongside untreated birds. The crucial variable was not the treatment but the release site: some releases were from places within the birds’ familiar area, others from sites they had never been.

What happened

From unfamiliar sites, anosmic birds chose poor initial bearings and homed less successfully. From familiar sites they performed like untreated birds. Their compass orientation was unaffected either way.

What it shows

That olfactory information belongs to the map rather than to the compass, and specifically to the map used where the landscape is unknown. The familiar-site result is what makes the case: a bird impaired in every condition might simply be an unwell bird.

What it does not show

It does not identify a single compound, or show what the odours encode. Every method of removing smell is invasive and may impair more than smell — the standing objection to this whole literature. And homing pigeons are domestic birds bred for this, not wild migrants.

The controls — what makes this evidence rather than a story
  • Untreated pigeons released at the same site on the same day.
  • Several unrelated methods of removing smell, so the result does not depend on one procedure’s side effects.
  • Familiar and unfamiliar sites compared, which is what turns a general deficit into a specific one.

From Forty years of olfactory navigation in birds

Take away a pigeon’s sense of smell and it still homes from places it knows. Release it somewhere unfamiliar and it cannot choose a direction — which places smell in the map, not in the compass.

Well supported

Good evidence backs this, though some details remain open.

Across four decades of experiments using nerve section, nasal anaesthesia and zinc-sulphate treatment, anosmic homing pigeons show impaired initial orientation and reduced homing success from unfamiliar release sites while retaining compass orientation and homing performance within familiar areas.

Who this applies to
Domestic homing pigeons, in a literature concentrated in Italy and Germany.Do not extend this beyond the taxa listed — the popular version over-reaches.
Studied in
Columba livia
Why we rate it this way, and what the caveats are
Well supportedModerate confidence

The behavioural effect is reproducible across laboratories and deprivation methods, which is the strongest argument for it. The standing objection — that olfactory deprivation is invasive and may impair more than smell — has never been fully answered.

How far it can be extended

Homing pigeons are selectively bred for this behaviour. Olfactory navigation in wild migrants is supported in some seabirds and remains open in most species.

Caveats

  • Every method of removing smell is invasive, and some also affect trigeminal sensation or general condition.
  • No compound has been identified as the one birds use; the map’s ingredients are unknown.
  • Effects are on the distribution of initial bearings, not on whether an individual bird eventually gets home.

Still unanswered

  • What the odours encode — a gradient, a set of learned associations, or something else — which is the live disagreement in this field.

Last reviewed 2026-09-03

The evidence (3 studies)

Diagram

Two releases, two different problems

Schematic. Track shapes are illustrative.

Take away smell, and it depends where you release the birdSchematic. Track shapes are illustrative.Familiar arealoftreleaseduntreatedcannot smell — much the sameUnfamiliar sitereleasedhomewardcannot smell — bearings scatterThe compass is intact in both cases. What is missing is knowing which way home is.
The same explanation in words

Two panels. In the first, a release from within the bird’s familiar area: an untreated pigeon and an anosmic pigeon both head home along a similar winding route, and the labels note that performance is comparable. In the second, a release from an unfamiliar site well outside that area: the untreated pigeon departs on a bearing close to the homeward direction, while the anosmic pigeon’s initial bearings scatter in all directions. Beneath, a note explains that the compass is intact in both cases — the deficit is in choosing which way home is, which is the map step rather than the compass step.

What the odours would have to be doing is the hard part, and it is where the disagreement lives. The proposal is that a bird at its loft learns how the ratios of atmospheric trace compounds vary with the wind direction they arrive on, and can later use that learned gradient to infer roughly where it has been taken. Atmospheric measurements suggest the required orderliness exists. No compound has been identified, and until one is, the map has a shape and no ingredients.

How it knows where to go

A homing pigeon

  • The sun

    Demonstrated

    A compass bearing, corrected by an internal clock

    Clock-shifted birds depart on a heading rotated by roughly the angle the sun moves during the shift.

  • Learned visual routes

    Demonstrated

    The way home over ground the bird already knows

    Repeated GPS-tracked releases produce idiosyncratic routes that survive clock-shifting.

  • Smell

    Supported

    The map step: which way home is from somewhere unfamiliar

    Anosmic birds choose poor bearings from unfamiliar sites and perform normally from familiar ones.

  • The magnetic field

    Contested

    Claimed as positional information, and much weaker here than in songbirds

    Asserted widely in popular accounts; the beak-magnetite receptor turned out to be immune cells.

Words used here
Anosmic
Unable to smell. In these experiments produced by nerve section, local anaesthesia or zinc sulphate — all invasive, which is the standing objection.
Initial bearing
The direction a bird takes in the first minutes after release, before it can use anything further along the route. It is the measurement most of this literature reports.

What experience changes

How we know

Releasing the same pigeon from the same field, twenty times

Does a pigeon that knows the way home fly the same way each time — and if so, why that way?

Pigeons carrying GPS loggers were released repeatedly from the same sites over many flights, and the similarity between each bird’s successive tracks was measured. Once routes had stabilised, some birds were clock-shifted before release, which rotates the heading of any animal still steering by the sun.

What happened

Each bird converged on its own winding, highly repeatable route that was not the straight line home and not the same as its neighbours’. Clock-shifted birds followed their established routes rather than departing on the rotated bearing the sun compass would have produced.

What it shows

That homing changes with experience: a bird that has learned the ground stops computing a bearing and starts following a remembered path over the landscape. The clock-shift is the argument — a bird still using the sun would have set off wrong, and these did not.

What it does not show

GPS records where a bird went, not what it was attending to; "following landmarks" is inferred from the route and from the clock-shift result rather than observed. The landscape was southern England, full of hedgerows and roads, and route following may be far weaker over uniform terrain. It says nothing about the first flight from an unfamiliar place.

The controls — what makes this evidence rather than a story
  • Repeated releases of the same individuals, so route repeatability is measured within a bird.
  • Several birds from the same loft, so idiosyncrasy can be distinguished from a shared best path.
  • The clock-shift, which makes two mechanisms predict visibly different behaviour.

From Familiar route loyalty implies visual pilotage in the homing pigeon

Released from the same place again and again, a pigeon stops flying the straight line and starts flying its own route — the same wiggles every time, and not the same as the bird in the next loft.

Established

Specialists would state this without hedging. Multiple independent lines of evidence agree.

GPS-tracked homing pigeons released repeatedly from fixed sites developed individually distinctive and highly repeatable routes that deviated from the direct line home, and clock-shifted birds followed their established routes rather than departing on the deflected bearing, indicating a shift from compass orientation to learned visual route following with experience.

Who this applies to
Homing pigeons over familiar terrain in southern England, released repeatedly from the same sites.
Studied in
Columba livia
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

The clock-shift control converts a description of route shape into a mechanistic claim: a bird still steering by the sun would have set off wrong, and these birds did not.

How far it can be extended

Route loyalty has since been reported in other birds, but the demonstration is from domestic pigeons in a landscape rich in linear features.

Caveats

  • GPS records where a bird went, not what it was looking at; visual pilotage is inferred from route shape and from the clock-shift result.
  • A landscape of hedgerows, roads and rivers may support route following better than open terrain does.
  • Says nothing about how a bird navigates the first time, from a place it has never been.

Still unanswered

  • What features birds actually use to define a route, and how a route is stored well enough to be repeated to within a few metres.

Last reviewed 2026-09-03

The evidence (1 study)

The routes are the surprising part. They are not the straight line, they are not shared between birds from the same loft, and they persist: a pigeon released for the twentieth time flies its own particular set of bends, following features it has apparently decided to follow. That is a different kind of navigation from computing a bearing — it is closer to knowing the way than to knowing where you are — and it is what a bird does once the ground below has become familiar.

The sun moves, so using it as a compass means knowing the time. Shift an animal’s internal clock by six hours and it sets off about ninety degrees wrong — confidently.

Established

Specialists would state this without hedging. Multiple independent lines of evidence agree.

Sun-compass orientation requires time compensation for solar azimuth change. Animals held under light–dark cycles shifted from local time depart on headings rotated by approximately the angle the sun traverses during the shift, in the predicted direction.

Who this applies to
Demonstrated in homing pigeons and in birds in orientation cages, and separately in insects including monarchs and bees.
Studied in
Columba livia, Aves, Insecta
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

The prediction is quantitative — a specific angle for a specific shift — and it is met. Few results in behavioural biology are that sharp.

How far it can be extended

Clock-shift experiments produce the predicted rotation in birds and in insects, in laboratories on several continents.

Caveats

  • Clock-shift measures the initial heading; animals frequently correct later using other cues, so a wrong departure is not a failure to get home.
  • Homing pigeons are domestic birds selected for this behaviour, and the strength of their sun-compass reliance may not be typical.

Still unanswered

  • How the compass output is weighted against magnetic and landmark information when they disagree.

Last reviewed 2026-09-03

The evidence (3 studies)

How we know

Putting a pigeon’s clock six hours wrong

If an animal steers by the sun, it must be correcting for the sun moving. What happens if you break the correction?

Homing pigeons were kept for several days in a room whose lights came on and went off on a schedule shifted by a known number of hours from the world outside — so that the birds’ internal clocks were running fast or slow by that amount. They were then taken away from the loft and released, and the compass bearing they took in the first minutes was recorded and compared with untreated birds released at the same place.

What happened

The shifted birds set off on a heading rotated from the correct one by roughly the angle the sun travels in the period of the shift — about fifteen degrees per hour — and in the direction the shift predicts. Controls departed correctly.

What it shows

That the sun compass is time-compensated, and that the compensation runs off an internal clock rather than off anything in the sky. The quantitative match is what makes it decisive: a confused bird would scatter, and these birds were confident and specific and wrong by the predicted amount.

What it does not show

It does not show that the pigeon fails to get home — most clock-shifted birds correct later, once other information becomes available, which is itself a clue that the sun compass is one input among several. And it says nothing about the map: a bird that departs on a rotated bearing was told which way to fly by the sun, and told where home was by something this experiment never touches.

The controls — what makes this evidence rather than a story
  • Untreated pigeons released at the same sites on the same days.
  • A known, deliberate shift, which turns the prediction into a specific angle rather than a vague expectation of confusion.
  • Initial bearings taken before the birds could use any landmark information from further along the route.

From Der Einfluß experimentell veränderter Zeitschätzung auf das Heimfindevermögen bei Brieftauben

Related

What is still unsettled

  • What are the odours?

    Why it matters: The behavioural effect has been reproduced for forty years and no compound has been identified. Until one is, the olfactory map is an inference from deficits.

    What would settle it: Identifying compounds with the right spatial structure and showing that manipulating them alone changes behaviour.

  • Do pigeons use magnetic information for position?

    Why it matters: It is asserted constantly in popular accounts. The behavioural evidence in pigeons is much weaker than the equivalent evidence in migratory songbirds.

    What would settle it: Virtual magnetic displacement of the kind run on reed warblers, applied to pigeons.

  • How much of this transfers to wild birds?

    Why it matters: Homing pigeons are a domestic animal selected for this behaviour over centuries, and homing is not migration.

    What would settle it: Equivalent manipulations in wild birds, which are far harder to run and correspondingly rare.

Claims about this, checked

Things people have heard, and what the evidence actually supports.

The research behind this page

3 studies, newest first. Each one has a page explaining what it found and what it could not show.

2013Journal of Experimental Biology

Forty years of olfactory navigation in birds

Anosmic pigeons released at unfamiliar sites are impaired in choosing a homeward bearing, while their compass orientation and their homing from familiar areas remain intact; the effect has been reproduced across laboratories and methods of olfactory deprivation.

2004Animal Behaviour

Avian olfactory navigation: its empirical foundation and conceptual state

Measured atmospheric trace-gas ratios do vary spatially in a sufficiently orderly way to be usable in principle, and the behavioural record is consistent with birds learning such gradients at the home loft.

2004Proceedings of the National Academy of Sciences

Familiar route loyalty implies visual pilotage in the homing pigeon

Individual birds converged on idiosyncratic, highly repeatable routes that were not the straight line home, and clock-shifted birds followed their established routes rather than departing at the deflected bearing.

Where to go from here

Each of these follows from something on this page — a relationship in the evidence, a claim people ask about, or the next mechanism along.

How complete this page is, and what it is still missing

NatureHQ publishes its own gaps. This page is at 51% completeness against what we would call a finished subject, and was last reviewed on 2026-09-03. It carries 2 claims and answers 4 mapped search questions.

  • no research from the last few years is attached — check for newer work
  • The magnetic side of pigeon navigation is treated as an open question rather than reviewed, and there is a substantial literature arguing both ways.
  • Pigeon racing, and what it has contributed to and cost the science, is not covered.
  • The familiar-area map — what a pigeon knows about its own landscape — is described through route loyalty rather than in its own right.