How Monarch Butterflies Navigate 3000 Miles: The Strange Multigenerational Memory of Migration

A monarch butterfly that hatches in Ontario in late August will fly 3000 miles south to a specific cluster of fir trees in the Sierra Madre Occidental of central Mexico, overwinter there with hundreds of millions of conspecifics on a few hectares of forest, and fly partway back north in the spring before laying eggs and dying. The next generation continues north, lays eggs, and dies. The generation after that does the same. By the time the descendants reach Canada in the summer, they are four or five generations removed from the butterfly that left Mexico the previous spring. No individual butterfly experiences the round trip. The migrating generation has never been to Mexico and has never seen a parent that was there. And yet they go, and they arrive at the same dozen mountain colonies their great-great-grandparents left the previous year.

The puzzle is not just that monarchs migrate. Many species do. The puzzle is that the navigation information is encoded across generations in a way the species cannot teach. Salmon return to a stream they were born in because they imprinted on its chemistry as juveniles. Cuckoos navigate to wintering grounds they have never seen but were raised by birds that flew there too. Monarchs navigate to a place no individual in the migrating cohort has ever been, and they do it on the first try. The mechanism is one of the strangest cases of inherited navigation in biology.

The basic facts

Eastern North American monarchs (Danaus plexippus) follow an annual cycle of three to five generations. Late summer monarchs that hatch in southern Canada and the northern United States are physiologically different from their parents and grandparents. They enter reproductive diapause — sexually immature, with much longer lifespans (8-9 months instead of 4-6 weeks) and elevated fat reserves. This migratory generation flies south, navigating to roosting forests in Mexico's transvolcanic belt, primarily in Michoacan and Mexico State. They overwinter clustered on Oyamel fir trees at elevations between 2900 and 3300 meters.

In late February and March, they begin the return flight. They fly north into Texas and the southern United States, break diapause, mate, and lay eggs on milkweed before dying. The eggs hatch, the caterpillars pupate, and a new generation of breeding-form butterflies emerges. That generation flies further north and repeats the cycle. By July or August, the monarchs in Canada are several generations removed from the ones that left Mexico in March.

The wintering colonies are remarkably specific. There are about a dozen well-documented sites in Mexico, occupying a total area of a few hectares in any given year, and the monarch population concentrates in this tiny region from a continental breeding range. The colonies were unknown to science until 1975, when Fred Urquhart's network of citizen-scientist taggers led him to the discovery via Kenneth Brugger, a textile engineer living in Mexico. Until then, the destination of the eastern North American monarch migration was a genuine scientific mystery.

The sun-compass mechanism

The first piece of the navigation puzzle solved was the compass mechanism. Steven Reppert's lab at the University of Massachusetts established in a series of papers from 2002 onward that monarchs use a time-compensated sun compass: they read the position of the sun and adjust for the time of day using a circadian clock, producing a consistent bearing toward Mexico regardless of when in the day they are flying.

The mechanism is anatomically traceable. Monarchs have photoreceptors in their compound eyes that detect polarized light patterns in the sky, which encode the sun's azimuth even when the sun itself is obscured. The information is processed in a region of the central complex of the brain called the protocerebral bridge, which integrates the visual input with a circadian clock running in the antennae. Cut the antennae off and the monarchs still orient to the sun, but they fly in circles relative to the ground because the time compensation is broken.

The circadian clock in the antennae is unusual — most animals have a single brain-centered circadian clock — and it is what makes the sun compass work as a navigation tool rather than just a visual orientation. The clock provides the time-of-day input that lets the system compute "in this season at this time of day, the sun is in the south-southeast, so to go south-southwest I should fly with the sun on my left side."

The magnetic compass backup

A sun compass fails on cloudy days. Patrick Guerra and Reppert's 2014 Nature Communications paper showed that monarchs have a backup: a magnetic compass based on cryptochrome proteins in the same eye photoreceptors that handle polarized light. The mechanism is the radical-pair model that European robins and other migratory birds use — quantum-mechanical spin states in flavoprotein radicals are sensitive to Earth's magnetic field at the strength found at the surface (around 50 microtesla).

The cryptochrome magnetic-compass mechanism requires light to work, specifically blue light. The monarchs in Guerra's experiments lost their magnetic orientation when blue light was filtered out, consistent with the cryptochrome requirement. The magnetic compass gives the inclination of the field (the angle relative to the surface), which monarchs use to distinguish "polar" from "equatorial" — useful for a north-south migrator.

Together, the sun compass and magnetic compass give monarchs a consistent bearing that works in all weather conditions a migrating butterfly is likely to encounter. The two systems agree under normal conditions and provide redundancy when one is unavailable.

The destination problem

A compass tells you which direction is south. It does not tell you to stop in central Mexico. The puzzle of how monarchs know where to land is harder than the puzzle of how they know which way to fly, and the answer is still not fully resolved.

The current best theory combines several factors: monarchs continue south until they encounter a particular microclimate signature (specifically, the cool moist conditions at high elevation in the transvolcanic belt of Mexico), they aggregate when they encounter other monarchs (positive feedback that amplifies small initial preferences into population concentration), and they have some inherited preference for the Oyamel fir habitat that genetically encodes the destination.

The genetic component is supported by Marcus Kronforst's lab work at the University of Chicago. Monarchs from non-migratory populations (Hawaii, Caribbean) have lost the migratory phenotype, and the genetic differences between migratory and non-migratory populations are concentrated in a small number of loci, several of which affect flight muscle metabolism. The implication is that migration is genetically simple to switch on or off, but the destination encoding is more diffuse and harder to localize.

The crisis

Eastern North American monarch populations have declined by approximately 80% since the mid-1990s. The Mexican wintering area, measured as occupied forest hectares, has gone from a peak of 21 hectares in 1996-1997 to lows around 1 hectare in 2013-2014. The decline is attributed to a combination of breeding-range habitat loss (loss of milkweed across the American midwest due to herbicide-tolerant corn and soybean cultivation), illegal logging at the Mexican overwintering sites, and climate disruption of the timing cues that drive the migration.

The conservation response includes the Monarch Joint Venture in the United States, the World Wildlife Fund Mexico's reforestation work at the colony sites, and the Migratory Bird Treaty Act consideration that has placed the monarch in an active candidate status for federal Endangered Species Act protection. The conservation case is unusual because the monarch is not rare in any individual generation — there are still hundreds of millions of monarchs — but the migratory phenomenon itself is at risk in a way the species per se is not. Hawaiian non-migratory monarchs are abundant. The thing that might disappear is the migration, not the butterfly.

What is still mysterious

The compass mechanism is well-understood. The destination problem is partly understood. The thing that remains genuinely mysterious is how the destination preference is genetically encoded such that a monarch never raised in Mexico knows to go to a forest type it has never seen. The genetic-architecture work shows that a small number of loci differ between migratory and non-migratory populations, but those loci affect flight physiology more than navigation per se.

The current best speculation is that the destination is not encoded as "Mexico" but as a target microclimate (cool, moist, evergreen-forested, high-elevation) that happens to be reliably found in the transvolcanic belt at the end of a south-southwest migration. The monarch is genetically encoded to fly southwest from a Canadian summer until it encounters that habitat, and the population aggregates at the sites that meet the criteria because positive feedback amplifies any initial small concentration.

This is plausible but unconfirmed. The colonies remain spatially specific in a way that pure habitat-targeting does not fully explain. Other Mexican mountains have similar microclimates and are not used. Something about the specific colonies — possibly the genetic memory of historical site fidelity, possibly some unidentified chemical or geographical cue — is still hidden.

The deeper observation

The monarch migration is one of the closest things in nature to inherited geographic memory across generations of individuals who do not share experience. The mechanism turns out to be a combination of well-characterized compass systems and less-well-characterized destination preferences, and the result is a continent-spanning annual cycle that has been operating in roughly the current form for at least 10,000 years since the most recent glacial retreat. Watching a butterfly the size of a postage stamp arrive at a forest it has never seen and where its great-great-grandparents departed is to watch genetics doing something that looks more like culture than biology usually permits. The fact that we still do not fully understand the destination mechanism is a useful reminder that biology has been solving navigation problems for hundreds of millions of years and the inventory of mechanisms is still being catalogued.