Migrating insects are not blown along. Radar shows them climbing to the altitude where the wind suits them and holding a heading that corrects for drift — and in most species the animal that arrives is not the animal that set out, because the journey takes more than one generation.
Two things make insect migration different from the bird version rather than a smaller copy of it. The first is scale in the wrong direction: it is invisible. Radar looking straight up finds insects crossing at hundreds of metres in numbers estimated in the trillions annually over a single country, moving biomass comparable to bird migration, and almost nobody has ever seen it happen. The second is generational. A monarch leaving Canada in autumn reaches Mexico, and the butterflies that arrive back in Canada the following summer are its descendants two or three generations on. That is not a smaller version of a bird flying south and back — it is a different problem, because whatever gets a butterfly to a Mexican grove it has never seen cannot be memory. What is inherited is a compass and a set of responses to conditions, and the clock-shift experiments that demonstrate the compass are among the cleanest in the field.
Developed record · 48% complete · reviewed 2026-08-10
What this page covers
A behaviour rather than a group. Covers butterflies, moths, hoverflies, dragonflies and locusts, with the quantitative evidence dominated by radar studies in northern Europe.
Often confused with: Dispersal, which is one-way movement without a return; Irruptions and outbreaks, which are irregular rather than seasonal
Quick facts
Not passive
Migrants select flight altitude and correct for wind drift
Monarchs
The round trip takes several generations
Compass
Time-compensated sun compass, shown by clock-shift experiments
Scale
Radar estimates trillions of insects crossing the UK annually
Not every insect that goes somewhere is migrating.
Insects move constantly, and most of that movement is not migration. Foraging is not migration. Dispersal — leaving where you were born and not coming back — is not migration either, though it moves individuals a long way. What earns the word is directed, seasonal movement between areas used at different times of year, undertaken in a distinct physiological state and usually with reproduction suspended.
That last part is a useful test and is easy to overlook. A migrating insect is generally not breeding while it travels: the reproductive system is held in check, fat is laid down, and flight behaviour changes from short hops between flowers to sustained travel at altitude. The switch is physiological and can be measured, which is what distinguishes a migrant from an insect that happens to be flying a long way.
Seasonal migrants: painted lady butterflies between Africa and northern Europe; silver Y moths on the same route.
Multi-generational migrants: monarchs, where the return leg is completed by descendants.
Altitudinal migrants: species moving up and down mountains with the seasons.
Wind-assisted long-haul: many moths and aphids, ascending to find a following wind.
Irruptive movers: desert locusts, whose mass movements are driven by rainfall and crowding rather than a seasonal calendar.
Locust swarms are often called migration and fit awkwardly. The movement is real and enormous, and it is triggered by density and rainfall rather than by season, so it behaves more like an irruption than a migration.
Words used here
Migratory syndrome
The package of changes — suspended reproduction, fat deposition, sustained flight — that distinguishes a migrating insect from one merely flying.
Dispersal
One-way movement away from where an animal was born, with no return leg.
Migrating insects are not blown along — they pick a wind and steer within it
Well supported
Good evidence backs this, though some details remain open.
Vertical-looking radar shows migrating insects ascending selectively to altitudes where wind direction favours their seasonal heading, and adopting common orientations that partially correct for wind drift. Migration occurs in enormous numbers, is associated with distinct physiological and reproductive states, and transports biomass and nutrients on a scale comparable to bird migration.
Who this applies to
high-altitude insect migrants measured by radar, principally in northern Europe
Studied in
Insecta, Autographa gamma, Vanessa cardui
You may have heard
“Insects are just carried by the wind”
Some are, and the migrants are not. Radar shows them climbing to the altitude where the wind suits them and then holding a heading that corrects for drift — which is a choice about transport rather than being transported. The scale is the other surprise: it happens hundreds of metres up, in numbers measured in trillions, entirely unseen from the ground.
Why we rate it this way, and what the caveats are
Well supportedModerate confidence
Radar measures heading, speed and altitude directly across millions of individuals. Confidence is moderate because radar identifies size classes rather than species, so behaviour is attributed to groups rather than to identified animals.
How far it can be extended
Altitude selection and common orientation have been recorded across several unrelated migrant taxa; tropical systems are far less sampled.
Caveats
Radar identifies size and heading, not species.
Quantitative work is concentrated in northern Europe.
Whether an individual completes a migration is usually unobservable, so generational structure is inferred.
Still unanswered
What compass do nocturnal insect migrants use, given that a sun compass is unavailable?
How much of tropical insect movement is migration in the same sense?
Radar-derived altitude selection, common orientation and the scale of the biomass moved.
Insect migration was studied for a century mainly by noticing arrivals — a wave of butterflies appearing in a country where they had not been the week before. Vertical-looking radar changed that by detecting insects in transit, hundreds of metres up, and measuring how fast they were going, in what direction, and how many of them there were.
How we know
Pointing a radar at the sky to find out what is up there
Insects appear in a country where they were not the week before. Are they being blown in, or are they going somewhere?
Insect migration had been studied almost entirely from arrivals, which cannot distinguish transport from travel. Vertical-looking radar changed the observable: a beam pointed straight up detects individual insects passing overhead, and from the returned signal it recovers each one’s altitude, body size, ground speed and — because insects are not radially symmetric — the direction its body is pointing. Heading and track can therefore be measured separately, which is exactly what separates steering from drifting.
What happened
Migrants concentrated at altitudes where winds favoured their seasonal direction rather than spreading through the air column, and adopted common headings that partially corrected for lateral drift. Annual movement over a single country was estimated in the trillions of individuals, moving biomass comparable to bird migration.
What it shows
That insect migration is orientated behaviour rather than passive transport. The animals are choosing an airstream and steering within it, which is a navigational achievement rather than a meteorological accident — and it is happening at a scale nobody could see because it happens hundreds of metres up.
What it does not show
Radar identifies size classes and headings, not species, so behaviour is attributed to groups rather than to identified animals. It cannot follow an individual, so whether one insect completes a journey — and therefore how generational the migration is — has to be established by other means. And the quantitative record is overwhelmingly northern European, where seasonality is strong; the tropical picture may be quite different.
The controls — what makes this evidence rather than a story
Body orientation measured independently of ground track, so drift correction can be detected rather than assumed.
Continuous operation across seasons and years, distinguishing regular seasonal movement from one-off weather events.
Wind profiles recorded simultaneously at each altitude, so altitude selection is tested against what the wind was actually doing.
Size classes recorded, allowing large migrants to be separated from small drifting insects in the same air column.
The behaviour it revealed is not passive. Migrants ascend until they find a wind blowing roughly where they want to go, and they concentrate at that altitude rather than spreading through the air column. Within it they adopt common headings that partially correct for being pushed sideways — so the resulting track is closer to the seasonal direction than the wind alone would produce.
Using wind this way is not laziness; it is the only way an animal weighing a fraction of a gram covers a thousand kilometres. A moth flying under its own power against a headwind would exhaust itself in hours. The skill being exercised is selecting and exploiting a moving airstream, which is a navigational problem rather than a muscular one.
Words used here
Vertical-looking radar
A radar pointed straight up that detects and measures individual insects passing overhead, including their heading and size.
Common orientation
Many individuals independently adopting the same heading — the signature that distinguishes migration from drift.
The most famous insect migration, and the part of it that is usually got wrong.
No monarch butterfly makes the round trip; the return north takes several generations
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
The eastern North American monarch migration is completed by successive generations. A long-lived autumn generation flies south to overwintering sites in central Mexico and returns partway north in spring; subsequent shorter-lived generations complete the northward journey, so the butterflies arriving at the northern end of the range are the descendants of those that left it. Migrating monarchs orient using a time-compensated sun compass, demonstrated by clock-shift experiments in a flight simulator.
Who this applies to
the eastern North American migratory population
Studied in
Danaus plexippus
You may have heard
“Monarchs return to the same tree their grandparents left”
The overwintering groves really are used year after year, and no individual butterfly is returning to anything — the ones arriving are several generations removed from the ones that left. That makes the navigation harder to explain rather than easier, because whatever gets them there cannot be memory of having been.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
The generational structure is established from tagging, isotope analysis of wing tissue indicating natal origin, and reproductive state. The sun compass is demonstrated by a clock-shift manipulation with a quantitative prediction.
How far it can be extended
Western North American and non-migratory monarch populations behave differently, and the generational structure described here is specific to the eastern flyway.
Caveats
A tethered butterfly in a simulator is not flying anywhere, and its motivation cannot be assessed.
A sun compass gives direction only, and does not explain how monarchs locate specific overwintering groves.
The western North American population has a different and less well-characterised pattern.
Still unanswered
How do autumn monarchs locate specific overwintering sites they have never visited?
What triggers the switch to the long-lived migratory generation?
Places multi-generational migration within the broader insect migration literature.
In eastern North America, monarchs emerging in late summer are physiologically different from the generations before them. They do not breed, they live months rather than weeks, and they fly south — up to several thousand kilometres — to a small number of overwintering groves in the mountains of central Mexico, where they cluster in enormous numbers through the winter.
In spring that same generation flies partway north, breeds and dies. Its offspring continue north, breed and die. Two or three generations later, monarchs are back at the northern end of the range. The butterflies arriving in Canada are the great-grandchildren of the ones that left it.
How we know
Moving a butterfly’s clock and watching it turn
Migrating monarchs hold a consistent southwesterly heading. Is that a sun compass, and if so, is it corrected for the time of day?
A sun compass makes a quantitative, falsifiable prediction that almost nothing else does: if an animal reads direction from the sun while correcting for the sun’s movement using an internal clock, then shifting that clock should rotate its chosen heading by a predictable angle. Monarchs were tethered in a flight simulator that let them fly continuously while their heading was recorded. Some were clock-shifted beforehand by several hours on an artificial light cycle; others were not.
What happened
Untreated monarchs held a consistent autumn southwesterly heading. Clock-shifted monarchs shifted their heading by approximately the predicted angle, in the predicted direction.
What it shows
That migrating monarchs use a time-compensated sun compass. The prediction is precise enough that passing it is strong evidence, and it establishes the mechanism as a compass rather than the inherited map often assumed in popular accounts of the monarch journey.
What it does not show
A compass gives direction, and the striking thing about the monarch migration is the destination: specific groves the arriving generation has never seen. Nothing here addresses that. A tethered butterfly is also not flying anywhere, so motivation and persistence cannot be assessed, and clock-shifting is a substantial physiological manipulation with effects beyond the compass.
The controls — what makes this evidence rather than a story
Unshifted monarchs flown in the same apparatus, establishing the natural heading and that tethering does not itself determine direction.
The shift magnitude predicts a specific angular displacement, so the result is quantitative rather than a change in the right general direction.
Shift direction varied, predicting rotations both ways — a general disorientation would not follow the sign.
Only monarchs in migratory condition tested, since a breeding butterfly has no migratory heading to measure.
The clock-shift result establishes what the butterflies are using to hold their direction: the sun, corrected for time of day by an internal clock. Shift the clock by six hours and the heading rotates by about ninety degrees, which is exactly what a time-compensated sun compass predicts and nothing else does.
What it does not explain is the destination. A compass gives direction, and the overwintering sites are specific groves on specific mountains that no butterfly in the arriving cohort has ever seen. How that resolution is achieved remains open, and the generational structure makes every memory-based explanation unavailable.
The western North American monarch population overwinters on the Californian coast and behaves differently. "Monarch migration" describes at least two systems, and the eastern one is the one in the photographs.
Words used here
Time-compensated sun compass
Using the sun for direction while correcting for its movement through the day, which requires an internal clock.
Overwintering site
Where a migratory population spends the non-breeding season. For eastern monarchs, a handful of high-altitude groves in Mexico.
Both groups migrate seasonally, both use compasses, and both exploit winds. The differences are where the interest is.
Two solutions to the same seasonal problem
Aspect
Birds
Insects
Who completes the journey
The individual, usually many times over a life
Often successive generations; the returning animal never made the outward trip
Learning
Adults acquire a map and correct for displacement
No opportunity to learn a route that takes generations
Wind
Used, but many species can make headway against it
Essential — most migrants cannot progress against a headwind at all
Compass
Magnetic inclination, stars and sun, cross-calibrated
Time-compensated sun compass; magnetic input demonstrated in some species
Altitude
Variable, often low over land
Frequently hundreds of metres up, selected for wind
The generational difference is the one that changes the science rather than the scale. A young bird that gets it wrong can learn and do better next year. A monarch has one journey and no ancestors to copy, so everything it needs must be inherited or read from the environment on the way — which makes it, in a sense, the harder navigational problem.
Each step here is a new way of detecting an insect that has left. Until you can follow one, or count them in transit, migration can only be inferred from the fact that they turned up somewhere.
1937
First observation
Monarch tagging begins
Fred Urquhart starts attaching tags to monarch wings, building over decades a network of volunteers reporting recoveries — the only way at the time to establish where the butterflies were going.
1975
Landmark experiment
The Mexican overwintering groves are located
After nearly forty years of tagging, the overwintering sites in the Sierra Madre are identified, confirming a journey of several thousand kilometres.
Changes how the 1937 result reads
Tagging established that monarchs travelled south. Finding the destination turned a direction into a specific journey, and immediately raised the question of how they locate a place none of them has been.
2002
Modern discovery
The sun compass is demonstrated
Monarchs clock-shifted and flown in a simulator shift their heading by the predicted angle, establishing a time-compensated sun compass rather than an inherited map.
Vertical-looking radar shows migrants selecting flight altitudes with favourable winds and adopting common headings that correct for drift, in numbers running to trillions annually.
Changes how the 1975 result reads
Monarchs had made insect migration a story about one charismatic species. Radar showed it as a continent-scale phenomenon involving moths, hoverflies and aphids in numbers that dwarf the butterflies — and showed the animals steering rather than drifting.
How do monarchs find specific overwintering groves?
Why it matters: A sun compass gives a direction, not an address, and the arriving generation has never been there. Whatever supplies the resolution is unidentified.
What compass do nocturnal insect migrants use?
Why it matters: Most moth migration happens at night, when a sun compass is unavailable. Magnetic and wind-derived cues are proposed and the evidence is thin.
How much tropical insect movement is migration?
Why it matters: The radar evidence is overwhelmingly north-temperate, and seasonality in the tropics works differently. The global picture may look quite unlike the European one.