Move a young bird 500 km and it flies the same heading into the wrong country. Move an adult and it corrects.
Migration is a regular, seasonal, two-way journey. A first-year bird flies an inherited compass heading for an inherited length of time; an experienced bird knows where it is and can correct if moved. The compass is not a compass in the usual sense — it reads the angle of the magnetic field against the ground, not which end points north.
The word "instinct" does most of the damage here. It is used as though it explained the journey, when what it names is the thing needing explanation — and once you separate the parts, the answer turns out to be several different mechanisms doing different jobs. A young bird making its first migration inherits a direction and roughly how long to hold it. That is genuinely enough to arrive somewhere sensible, and it fails in a specific and testable way: displace a juvenile a few hundred kilometres and it flies the same heading from the new position into the wrong country, while an adult that has done the trip before corrects and arrives where it meant to. So there are two capabilities, not one: a compass, which is inherited, and a map, which is learned by going. The compass itself is really three — the sun, the stars and the magnetic field, cross-calibrated against each other. And the star compass is learned in the nest from watching which point the sky turns around, which is a better design than inherited constellations, since the pole star changes over millennia and the rule does not.
Developed coverage · 83% complete · reviewed 2026-09-03
What this page covers
A behaviour rather than a group. Covers migratory birds generally, with the experimental evidence concentrated in a handful of well-studied songbirds and seabirds.
Often confused with: Animal navigation, which is the mechanism for finding direction and place across all animals; Dispersal, which is a one-way movement without a return
Quick facts
Two capabilities
An inherited compass heading, and a learned sense of position
Magnetic compass
Reads field inclination, not polarity
Star compass
Learned from the rotation of the night sky
Fuel
Fat, laid down before departure and burned at roughly 1% of body mass per hour
Regular, seasonal and two-way — which excludes a good deal of bird movement.
Migration is a regular seasonal movement between a breeding area and a non-breeding area, with a return. That definition does real work: it excludes dispersal, where young birds leave and do not come back; it excludes irruptions, where a food failure pushes birds somewhere unusual in some years and not others; and it excludes nomadism, where movement follows resources with no fixed pattern.
Long-distance migration: swallows and warblers between Europe and sub-Saharan Africa; Arctic terns between polar regions.
Short-distance and partial migration: some individuals of a species move while others stay, often splitting by age, sex or condition.
Altitudinal migration: up a mountain to breed and down for the winter — a few hundred metres of vertical movement doing the work of a thousand kilometres.
Moult migration: a journey to a safe place to replace feathers, made separately from the main migration.
Partial migration is worth pausing on, because it undermines the idea of migration as a fixed species trait. In many populations some birds go and some stay, and which they do depends on age, sex, dominance and how the winter is going. It is a decision made under conditions rather than a property of being that kind of bird.
Until the nineteenth century the disappearance of birds each autumn was explained by hibernation in mud, transformation into other species, or flight to the moon. The question was not silly; the answer was simply not available until birds could be marked and found again.
Words used here
Partial migration
Some individuals of a population migrate and others do not. Extremely common, and evidence that migration is conditional rather than automatic.
Irruption
A mass movement in some years only, usually driven by a failure of a food crop. Not migration, because it is not regular.
One experiment separates two things the word "instinct" merges.
A first-year bird follows an inherited direction; an adult can correct for being moved
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Displacement experiments separate two navigational capabilities. Juvenile migrants on their first journey maintain the inherited compass heading from wherever they are released and arrive at a correspondingly displaced destination. Experienced adults compensate, changing heading to reach their usual wintering area, which requires knowing their position relative to a goal rather than only a direction.
Who this applies to
starlings and chaffinches displaced during autumn migration in western Europe
Studied in
Sturnus vulgaris, Fringilla coelebs
You may have heard
“Birds instinctively know where to go”
What is inherited is a direction and roughly how long to fly it, which is enough to reach a plausible destination and no help at all if the bird is moved. Knowing where you are is a separate ability, and it is learned by making the journey. Move a young bird and it flies the same heading into the wrong country.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
A large-scale displacement with more than eleven thousand marked birds and a clear, predicted difference between age classes recovered from independent reports.
How far it can be extended
The juvenile–adult split has since been found in several other migrant species, though the degree of correction varies.
Caveats
Recoveries depend on birds being found and reported, which varies by region and season.
Displacement is a severe manipulation with effects on condition and motivation that cannot be measured.
How adults determine position — the map component — is not answered by this result.
Still unanswered
What information does the adult map component actually use?
The displacement itself, and the divergence between juvenile and adult recoveries that separates compass from map.
How we know
Moving eleven thousand starlings to see what they would do
Is migration one ability or several? Specifically: does a bird know where it is, or only which way to fly?
Starlings were trapped in the Netherlands during autumn migration, ringed, transported several hundred kilometres south to Switzerland, and released — a place none of them had any business being. The crucial element is that recoveries were analysed separately by age. Juveniles had never migrated; adults had made the journey at least once. If migration were a single inherited programme, both groups should behave identically.
What happened
Juveniles continued on the original compass heading from the displaced release point and were recovered in Spain, well outside the species’ normal wintering range. Adults changed heading and were recovered in their usual wintering areas in northern France and Britain.
What it shows
That "migration instinct" is at least two capabilities. A young bird inherits a direction and roughly how long to hold it, which works from the place it was born and fails predictably from anywhere else. An experienced bird knows its position relative to a goal — true navigation — and that is learned rather than inherited.
What it does not show
It does not reveal what the adults use to work out where they are; the map component is demonstrated here and identified nowhere. Recovery data are also biased by where people are and how likely they are to report a ring, which distorts distributions in ways that cannot be fully corrected. And displacement is a severe manipulation — the birds arrive in poor condition in an unfamiliar place, and motivation cannot be measured.
The controls — what makes this evidence rather than a story
Age determined from plumage at ringing, so the two groups are separated before anyone knows what will happen.
A very large sample — more than eleven thousand birds — because recovery rates for ringed birds are low and the result depends on the shape of a distribution.
Birds trapped mid-migration rather than at the breeding site, so all were already in migratory condition.
Recoveries reported independently by finders across several countries, with no involvement from the experimenters in where birds turned up.
What a first-year bird inherits is a vector: fly this way, for about this long. That is a genuinely good solution — it is cheap, it needs no experience, and it puts a naive bird in approximately the right part of the world. Its failure mode is exactly what the displacement showed: it has no way to notice that the starting point has changed.
An adult that has made the journey has something more: a sense of where it is relative to where it is going. That is what "navigation" properly means, and it is far harder than direction-finding. What information the map uses — magnetic intensity and inclination gradients, smells, learned landmarks, or some combination — is still not settled, and is one of the genuinely open questions in the field.
A bird’s magnetic compass tells poleward from equatorward, not north from south
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
European robins tested in coil systems reoriented when the horizontal component of the magnetic field was turned, but reversing the horizontal component alone did not reverse their heading; reversing the vertical component did. The compass therefore responds to the inclination of the field lines relative to gravity rather than to magnetic polarity, distinguishing poleward from equatorward directions.
Who this applies to
night-migrating songbirds, established in the European robin
Studied in
Erithacus rubecula, Aves
You may have heard
“Birds follow the Earth’s magnetic field like a GPS”
Two errors in one sentence. A compass gives direction and GPS gives position — knowing which way is poleward tells a bird nothing about where it is. And the bird’s compass is not the kind in a hiker’s pocket: it reads the angle the field makes with the ground rather than which end points north.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
A direct manipulation with the diagnostic control: separating the horizontal and vertical components tells the two candidate mechanisms apart, and the result is unambiguous.
How far it can be extended
The inclination compass has since been demonstrated in many migratory bird species; it is not universal among animals with magnetoreception.
Caveats
Cage orientation measures intention rather than a flight path.
The receptor mechanism remains contested, with cryptochrome and magnetite hypotheses both unresolved.
An inclination compass gives no useful signal at the magnetic equator, where the field lines are horizontal.
Still unanswered
Which receptor actually detects the field, and where is it?
Places the behavioural compass alongside the unresolved question of which receptor detects the field.
How we know
Reversing a magnetic field the wrong way round on purpose
Birds orient by the magnetic field in total darkness. Is their compass the kind in a hiker’s pocket?
Robins in migratory condition were tested in orientation cages surrounded by coil systems that could alter the magnetic field around them, with the sky excluded. The design turns on separating two components that a hand compass treats as one. The horizontal component was reversed alone; the vertical component was reversed alone; and both were reversed together. A polarity compass and an inclination compass predict different answers to those three manipulations, which is what makes the experiment decisive rather than merely suggestive.
What happened
Turning the horizontal component turned the birds’ headings, confirming magnetic orientation. Reversing the horizontal component alone did not reverse their direction; reversing the vertical component did. The compass responds to the angle between the field lines and gravity.
What it shows
That the avian magnetic compass reads inclination, distinguishing poleward from equatorward rather than north from south. It follows that the compass gives no useful signal at the magnetic equator, and that a bird carried across the equator must reinterpret the same field.
What it does not show
Nothing about the receptor: a behavioural result establishes what information is used, not what detects it, and the sensory mechanism remains contested more than fifty years later. It also measures caged intention rather than flight, and one species stands in for a large and varied group.
The controls — what makes this evidence rather than a story
The natural field reproduced by the coils as the baseline condition, so the apparatus itself is not what changes behaviour.
Horizontal and vertical components manipulated independently — the manipulation that distinguishes the two candidate mechanisms.
Sky excluded entirely, removing star and sun compasses from the available cues.
Repeated trials per bird across the migratory season, since individual scatter in funnel data is wide.
The magnetic compass is the one everybody has heard of and the one most often described wrongly. It does not read polarity. It reads the angle the field lines make with the ground, which distinguishes poleward from equatorward — a perfectly serviceable instrument that would give no signal at all at the magnetic equator, where the lines run horizontal.
Young birds learn which way is north by watching the sky turn
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Indigo buntings tested under a planetarium sky oriented by the pattern of stars around the axis of celestial rotation rather than by specific constellations, and rotating the artificial sky rotated their headings. Birds reared without exposure to a rotating sky failed to orient by stars; birds reared under a sky rotating about an arbitrary star treated that star as the pole.
Who this applies to
indigo buntings under planetarium conditions
Studied in
Passerina cyanea
You may have heard
“Birds are born knowing the stars”
What is inherited is a rule — find the point the sky turns around — rather than a chart. That is a better design than inherited constellations, because the pole star changes over thousands of years and the rule does not have to.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
A developmental manipulation with a clean prediction: rear a bird under a sky turning about the wrong star and it should adopt the wrong pole, which is what happened.
How far it can be extended
Star-compass calibration from rotation has been shown in several night-migrating species since, though the developmental window differs.
Caveats
A planetarium sky is brighter and simpler than a real one.
Captive rearing is a severe manipulation with effects beyond the visual environment.
A star compass supplies direction only, and is unavailable under cloud.
Still unanswered
How do birds arbitrate between star, magnetic and sun compasses when they disagree?
The orientation funnel that makes the measurement possible at all.
How we know
Taking the stars away one at a time
Night-migrating birds orient under a starry sky. Are they reading specific constellations, or something more general?
Buntings in orientation funnels — which record attempted direction as ink footprints on sloping paper — were placed under a planetarium sky that could be manipulated in ways no real sky permits. The whole sky could be rotated, individual constellations blanked out, and the axis of rotation moved to an arbitrary star. Crucially, some birds were reared from fledging without ever seeing a rotating sky at all.
What happened
Birds oriented by the region of sky around the axis of rotation rather than by any particular constellation, and rotating the sky rotated their headings. Birds reared without a rotating sky failed to orient by stars at all. Birds reared under a sky turning about an arbitrary star treated that star as the pole.
What it shows
That the star compass is learned in early life from watching the sky turn, and that what is inherited is a procedure — find the still point — rather than a chart. That is a better design than inherited constellations, because precession moves the pole star over millennia while the rule stays correct.
What it does not show
A star compass gives direction and nothing else; it says nothing about how a bird knows where it is. The planetarium sky is brighter and simpler than a real one, and rearing birds without a natural sky is a drastic manipulation with effects beyond the visual. Funnel data are also intentions rather than journeys.
The controls — what makes this evidence rather than a story
Rotating the artificial sky while the room and the magnetic field stayed constant, so a change in heading has one available cause.
Selective removal of individual constellations, distinguishing dependence on a specific pattern from dependence on the rotational axis.
Birds reared without sky rotation, testing whether the compass is inherited or learned.
Birds reared under a sky rotating about a star that is not the pole star, which predicts a specific wrong answer if the compass is learned.
The sun compass is the third, and it requires a clock: the sun moves, so using it for direction means knowing the time of day and compensating. That gives a clean experimental prediction, and clock-shifting a bird by holding it under an artificial light cycle shifts its heading by the corresponding angle.
Having three is not redundancy for its own sake. Cloud removes the stars, the sun is unavailable at night, and magnetic anomalies exist. Birds cross-calibrate them — recalibrating the magnetic compass against the sky around sunset is well documented — and which one wins when they disagree depends on species and circumstance.
Words used here
Inclination compass
A compass reading the angle of magnetic field lines relative to gravity. Tells poleward from equatorward rather than north from south.
Clock-shift
Holding an animal on an artificial day–night cycle to move its internal clock, which predictably rotates any heading based on the sun.
A small bird crossing the Sahara is solving an energy problem before it is solving a navigation problem. In the weeks before departure migrants enter hyperphagia — eating far beyond normal requirements — and lay down fat that can approach half their body mass. Some species shrink their digestive organs before departure to save weight and rebuild them on arrival.
Fat is the fuel: energy-dense, and burned at roughly 1% of body mass per hour of flapping flight.
Stopover sites are not rest stops but refuelling stations, and a migrant may spend far longer on the ground than in the air.
Most songbirds migrate at night: cooler, calmer air, fewer predators, and the day left free for feeding.
Wind decides a great deal. Migrants wait for favourable conditions, and a night with the wrong wind means no departure.
Altitude varies from a few hundred metres to several thousand, and higher usually means faster and drier.
Migration is also dangerous in a way that gets understated. Mortality during migration is substantially higher than during the stationary periods either side, and the losses fall hardest on first-year birds making the journey without experience. A bird that has migrated successfully once is a bird that has passed a severe filter.
Sleep is part of the cost. Frigatebirds flying continuously over the ocean sleep around 42 minutes a day against more than twelve hours on land, and several songbirds show migration-season sleep reduction without the performance collapse that would follow in a non-migrant.
Two mechanisms, in different species, and the difference decides what happens when a population loses its old birds.
A young warbler migrating alone in its first autumn has nobody to copy, so whatever it is using came with it. A young crane migrates in a flock containing birds that have done it before. Those are genuinely different systems, and the evidence for each is good — which means the popular question "is migration instinctive or learned?" has no single answer, and species that answer it differently respond differently to everything from a bad season to a reintroduction programme.
How we know
Cross-breeding two populations that fly in different directions
Is a migratory direction inherited — and if so, inherited as what?
Blackcaps from an Austrian population that migrates south-east and a German population that migrates south-west were bred together in captivity. Their offspring were hand-raised without ever seeing an adult migrate or experiencing a migration, then tested in orientation cages during the migratory season, and their preferred headings compared with those of hand-raised birds from each parent population.
What happened
Birds from each parent population oriented in their population’s direction. The hybrids oriented between the two, on an intermediate bearing that neither parent population uses.
What it shows
That migratory direction is inherited, and inherited as a quantitative bearing rather than as a choice between two routes. An intermediate heading is what you get from a continuously varying trait, and it is not what you would get if the birds inherited a route or copied a parent.
What it does not show
It does not show that the *route* is inherited. What the hybrids carry is a compass bearing, and in the wild that bearing would take them into the Alps rather than along either parent’s route — which is why the two populations are separated by a migratory divide in the first place. Nor does it establish anything about how these birds would behave with a season’s experience behind them.
The controls — what makes this evidence rather than a story
Hand-raising from the nest, so no bird could have learned a direction from another.
Both parent populations raised and tested identically alongside the hybrids.
Testing in cages that exclude landscape cues, leaving the birds’ own preference as the measurement.
A young bird inherits a direction and roughly how long to fly, not a route. Cross two populations that head different ways and the offspring head between them.
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Migratory direction and the amount of migratory restlessness are heritable quantitative traits: hand-raised birds express their population’s direction without exposure to it, hybrids of populations with divergent headings orient intermediately, and both traits respond rapidly to artificial selection.
Who this applies to
Established in blackcaps and related warblers; the inherited-programme model is applied more widely than it has been tested.
Studied in
Sylvia atricapilla, Sylviidae
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
Hand-rearing removes the possibility of learning the direction, and the intermediate hybrid headings are a quantitative prediction that was met.
How far it can be extended
The cage-orientation and cross-breeding results are from a few species, but consistent with the age-related displacement results across many.
Caveats
What is inherited is a direction and a duration, which is not a route: the same programme carries a displaced bird into the wrong continent.
Cage restlessness is a proxy for migration, and the hybrids were never tracked on a real journey.
Species that learn routes socially do not fit this model at all.
Still unanswered
How the inherited programme and later route experience interact in a bird making its second migration.
The consequence in the wild: a first-year bird carrying an inherited heading and displaced sideways flies it anyway.
What the hybrids inherit is worth being exact about, because it is not a route. It is a bearing — a direction to hold and roughly how long to hold it — and in the wild that intermediate bearing would carry a bird into the Alps rather than along either parent’s path. An inherited programme is a rule, and a rule can be wrong for the place an animal finds itself in.
Diagram
One experienced bird beats any number of inexperienced ones
Schematic. Deviation is drawn to show the pattern rather than measured values.
The same explanation in words
A chart with deviation from the direct route on the vertical axis and successive years on the horizontal. A line starts high at a bird’s first migration and falls steadily year on year, flattening as the bird gains experience. A panel beside it records the separate finding: a group containing one older bird performed better than a larger group without one, and group size mattered far less than the presence of an experienced individual. A closing note states that this was shown in a reintroduced population where every bird’s age and travelling companions were known.
How we know
Following young cranes for eight years to see who taught them
When a migratory route improves with age, is the bird getting better on its own or learning from older birds?
A reintroduced whooping crane population provided something wild populations never do: every bird’s exact age was known, and so was the composition of the group it travelled with each year. Eight years of migration tracks were analysed for how far each bird strayed from the direct route, against its own age and against the ages of its companions.
What happened
Deviation from the direct route fell steadily with age. A group containing an older bird performed far better than a group of the same size without one, and the presence of a single experienced individual mattered more than any amount of extra company.
What it shows
That in this species the route is a piece of social knowledge: the improvement travels from experienced birds to inexperienced ones, rather than each bird slowly working it out alone.
What it does not show
It does not generalise to migratory birds at large — most songbirds migrate alone in their first year and cannot be learning from anybody. And this population learned its very first route by following a microlight flown by people, which is not how a wild population acquires one. It is observational: nobody assigned birds to fly with or without an elder.
The controls — what makes this evidence rather than a story
Known individual ages, which separates maturation from experience-of-others.
Group size analysed separately from group age composition, so that "more birds" and "older birds" do not get confused.
Multiple years per individual, so that improvement can be measured within a bird as well as across birds.
In some species the route is not inherited at all — it is learned from older birds. A single experienced individual in a flock improves the whole group’s route more than any number of young ones.
Well supported
Good evidence backs this, though some details remain open.
In a reintroduced whooping crane population with known individual ages and group composition, deviation from the direct migratory route declined with age, and the presence of older individuals in a travelling group improved performance substantially more than group size did.
Who this applies to
One reintroduced population of a long-lived, socially migrating bird whose first route was human-led.Do not extend this beyond the taxa listed — the popular version over-reaches.
Studied in
Grus americana
Why we rate it this way, and what the caveats are
Well supportedHigh confidence
Eight years of tracks with known ages and companions — a record wild populations do not provide — and an effect size that separates social learning from simple maturation.
How far it can be extended
Most migratory birds travel alone on their first migration and cannot be learning the route from anybody. Social route learning is a real mechanism in some species, not the general case.
Caveats
A reintroduced population taught its first route behind an aircraft, which is not how a wild population acquires one.
Observational: nobody assigned birds to travel with or without an elder.
Says nothing about species that migrate alone in their first year, which is most songbirds.
Still unanswered
Whether wild populations of socially migrating species show the same dependence on a few experienced individuals — and what happens to a route when those individuals are lost.
A boundary between neighbouring populations that migrate in different directions. Birds from either side that interbreed produce offspring heading between the two, often into unsuitable terrain.
Two of the most-quoted numbers in natural history, and both changed when somebody measured rather than estimated.
Bar-tailed godwits fly from Alaska to New Zealand in one go — more than ten thousand kilometres over open ocean, without landing, eating or drinking, for something like eight or nine days.
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Satellite tracking of bar-tailed godwits of the baueri subspecies documented continuous southbound transoceanic flights from Alaska to New Zealand and eastern Australia exceeding 10,000 km, undertaken without stopover and timed to depart with favourable synoptic wind conditions.
Who this applies to
One subspecies on the southbound leg of one flyway. The northbound return is not nonstop.
Studied in
Limosa lapponica baueri
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
Direct satellite tracking of individual birds, with continuous positions across the ocean — the flight is observed rather than inferred.
How far it can be extended
Other bar-tailed godwit populations migrate differently, and the record applies to this route.
Caveats
Small samples: transmitter mass restricts which individuals can be tagged.
The southbound leg is nonstop; the northbound journey is staged through the Yellow Sea, so "never stops" is wrong as a description of the annual cycle.
Later tracking has recorded longer individual flights; the record figure is a moving one and NatureHQ states the study it comes from.
Still unanswered
How the birds sleep, if they do, across eight days of continuous flapping flight.
Gill et al., 2009 · Proceedings of the Royal Society B: Biological Sciences
The tracked transpacific flights themselves.
Before satellite tags, a nonstop Pacific crossing was an inference from absence: nobody had seen the birds anywhere in between. That is weak evidence and it turned out to be right. What tracking added was not only confirmation but the detail that matters — the birds are choosing their weather, departing on the systems that will push them south, rather than leaving on a date.
Arctic terns travel further in a year than any known animal — but not in a straight line. The tracked route wanders, pauses in the mid-Atlantic for weeks, and is far longer than the gap between its ends.
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Light-level geolocator tracking of Arctic terns from Greenland and Iceland recorded annual return journeys substantially exceeding the straight-line pole-to-pole distance, following S-shaped Atlantic routes with a prolonged mid-Atlantic stopover and prevailing-wind-following southbound legs.
Who this applies to
Birds from two North Atlantic colonies, tracked over one annual cycle.
Studied in
Sterna paradisaea
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
Direct tracking replaced an estimate. The distances are computed from recorded positions rather than from a map measurement between endpoints.
How far it can be extended
Subsequent tracking of other colonies has found comparably long and similarly indirect routes.
Caveats
Geolocators are recovered only from birds that survive and return, which biases the sample.
Light-level positions are accurate to roughly a hundred kilometres and fail near the equinoxes, so route length carries real uncertainty.
Distances differ between colonies and between individuals; a single headline figure hides that.
Still unanswered
How much of the route is chosen and how much is imposed by wind fields the birds are exploiting.
The general lesson that direct tracking overturns estimates made from indirect data.
How we know
Waiting a year to find out where the bird went
How far does an Arctic tern actually travel in a year — measured, rather than estimated?
Arctic terns at colonies in Greenland and Iceland were fitted with light-level geolocators weighing about a gram. The device records nothing but ambient light against time and transmits nothing at all; the birds were left to migrate, and the following season the same individuals were caught again at the same colonies and the devices removed and read. Sunrise and sunset times give day length and local noon, from which latitude and longitude are reconstructed.
What happened
The birds followed winding Atlantic routes with a prolonged mid-ocean stopover, tracked south along either the African or the South American side, and covered annual distances far greater than the straight-line distance between the poles.
What it shows
The longest annual migration measured in any animal, and — more usefully — its actual shape, which is nothing like the straight line the older estimates assumed.
What it does not show
The method reports only from birds that survived and came back to the same colony, so every route here is a survivor’s. Light-level positions are good to roughly a hundred kilometres and unusable near the equinoxes, so route length carries real uncertainty. Two colonies do not describe a species.
The controls — what makes this evidence rather than a story
Recapture of the same marked individuals, so each track belongs to a known bird.
Two colonies, giving some check on whether a route is a local peculiarity.
Known accuracy limits applied when reconstructing positions, including the equinox periods where longitude cannot be derived.
The Arctic tern figure is the better cautionary tale of the two. The old number was a straight line between the ends of the journey; the measured route wanders, pauses in the mid-Atlantic for weeks and runs down one side of the ocean or the other, and it is substantially longer. A record quoted without its method is a record that can quietly be about something else.
This history is unusually clean, because each step is a new way of *knowing where a bird went*. The explanations changed when the observations became possible, not when the arguments improved.
1703
First observation
Birds are proposed to winter on the moon
An anonymous English pamphlet argues that migrating birds fly to the moon and back. It is easy to mock and worth taking seriously as a symptom: no method existed for following a bird beyond the horizon.
1822
Landmark experiment
A stork arrives in Germany carrying a Central African spear
The Pfeilstorch — a white stork found with a spear of African manufacture through its neck — is the first hard physical evidence that European birds travel to Africa and return.
Changes how the 1703 result reads
One bird settled a question centuries of argument had not, because it carried an object that could only have come from a specific place. Evidence beat plausibility.
1899
Landmark experiment
Systematic ringing begins
Hans Christian Cornelius Mortensen begins marking starlings with numbered metal rings, turning migration from speculation into a data set built from recoveries.
1958
Landmark experiment
Displacement separates compass from map
Eleven thousand starlings are moved several hundred kilometres and released. Juveniles keep the inherited heading and land in Spain; adults correct and reach their usual wintering grounds.
Changes how the 1899 result reads
Ringing showed where birds went. Displacement showed what they were using to get there, by breaking it — and revealed that "migration instinct" is two different capabilities.
The Emlen funnel records the direction a caged migrant tries to go as ink footprints on sloping paper, allowing the sky, the magnetic field and the bird’s clock to be manipulated experimentally.
The magnetic compass turns out to read inclination
Robins in altered fields reorient when the vertical component is reversed but not when the horizontal component alone is — an inclination compass, not a polarity compass.
Changes how the 1966 result reads
The funnel made the question answerable; this is the answer, and it is not the compass anyone assumed. Poleward-versus-equatorward is a different instrument from north-versus-south.
Miniature devices on free-flying birds record brain activity, position and altitude for days at a time, showing frigatebirds sleeping in flight over open ocean.
Why it matters: The compass is well characterised and the map is not. Magnetic gradients, olfactory landscapes and learned landmarks are all candidates, and no experiment has cleanly isolated one.
Which receptor detects the magnetic field?
Why it matters: A behaviour demonstrated in 1972 still has no agreed sensory basis. Cryptochrome in the eye and magnetite in the beak are both proposed and neither is established.
How fast can migration timing shift as climate changes?
Why it matters: Departure is cued partly by day length, which does not change, while the food peak at the destination moves. The mismatch is measurable and the flexibility available is not.
How do migrating songbirds cope with reduced sleep?
Why it matters: Non-migratory birds deprived of sleep perform worse; migrants appear not to, which suggests something about the state changes seasonally.
How complete this page is, and what it is still missing
NatureHQ publishes its own gaps. This page is at 83% completeness against what we would call a finished subject, and was last reviewed on 2026-09-03. It carries 14 claims and answers 25 mapped search questions.
45 high-priority search intent(s) not yet covered
no research from the last few years is attached — check for newer work
Insect migration — monarchs, hoverflies, dragonflies — is a large parallel literature not covered here.
Specific flyways and species accounts are deliberately out of scope; this is the mechanism, not a field guide.
Conservation of migratory species is named but not treated.