How animals find their way

Imagine being blindfolded, driven to a town you have never seen and told to walk home. A compass would show you which way north is, but not whether home lies to the north or to the east. You would first have to work out where you are relative to home, and only then would the compass help.

In 1953 the German biologist Gustav Kramer proposed that homing birds face the same two steps. In his map-and-compass model, the map step gives the animal its position relative to its goal, and so the direction in which the goal lies. The compass step then finds that direction out in the world, using cues such as the sun, the stars or the Earth's magnetic field. The two can come apart: an animal may have a working compass and still no map.

Which parts of this equipment is a bird born with? The Dutch biologist Albert Perdeck tested it on a grand scale. Between 1948 and 1957, more than eleven thousand starlings caught on autumn migration in the Netherlands were ringed, flown about 600 kilometres to Switzerland and released there. Rings found later showed where they ended up.

Young birds on their first migration kept flying in the direction they had been travelling, and wintered in France and Spain, far from their usual winter quarters. Adults that had made the journey before turned north-west, back towards those quarters. A 2024 re-analysis found that the young birds' routes could not be explained by their following other starlings.

So a young starling carries an inborn compass heading but no map. The map, which let the adults correct for being moved, is learned from experience.

Pigeon keepers knew long ago that the map has to be learned. In ninth-century Basra, al-Jāḥiẓ wrote in his Book of Animals that a homing pigeon comes back from a distant release point only after training in stages, released farther from home each time, and that keepers would not risk an untried bird on a long flight.

Modern experiments agree, and add a deadline. Researchers in Italy kept young pigeons in an aviary screened from the wind for their first three months after fledging, then let the wind in for three more. Released far from home about six months after fledging, these birds mostly failed to head for home or get back, while birds that had felt the wind from the start homed well. The team thinks young pigeons build their map partly from smells carried on the wind. The experience counts most in an early window, a sensitive period.

The sun was the first bird compass to be shown by experiment. In the migration season a caged migrant grows restless and faces the way it would fly. Around 1950 Gustav Kramer put starlings in a round cage with six windows and recorded which way they faced. Then he fitted mirrors that made the sunlight seem to come from about 90 degrees to one side, and the birds changed the direction they faced by about the same angle.

The starlings were steering by the sun's position: a sun compass. Al-Jāḥiẓ had observed in the ninth century that a pigeon unsure of its way works it out from the wind and the sun's place in the sky. A sun compass is not simple, because the sun does not stay put. Seen from Europe, it rises in the east, stands in the south at midday and sets in the west.

To steer by a moving sun, an animal must allow for the time of day. The sun's direction swings round at 15 degrees an hour on average, a full circle in 24 hours. A bird that knows the time can work out where south is from where the sun stands. The time comes from its internal clock, the same daily clock that times sleep and waking, and light resets it each day.

Resetting it makes the proof. In 1958 Klaus Schmidt-Koenig, a student of Kramer's, kept homing pigeons for several days in a room whose lights came on and went off six hours out of step with the real day. Released far from their loft on sunny days, the shifted birds set off about 90 degrees away from the home direction, as six hours at 15 degrees an hour predicts. They read the sun right, for the wrong time.

What if clouds hide the sun? In the late 1940s Karl von Frisch found that honeybees can still keep their bearings as long as they see a patch of blue sky. Sunlight is scattered by the molecules of the air, and scattered light from the sky is polarised: its waves vibrate more in one direction than in others. Across the sky those directions form a pattern arranged around the sun, strongest 90 degrees away from it.

Bees can see this pattern, which is invisible to us, and read the sun's position from it. Many insects do the same, desert ants among them. Because the pattern moves with the sun through the day, it too is a compass that needs the internal clock.

Many songbirds migrate at night, steering by the stars. Because the Earth spins, the night sky seems to turn around one point, which in the northern sky lies close to the Pole Star and marks north. Are young birds born knowing the constellations? Stephen Emlen tested this with indigo buntings he raised by hand.

In 1970 he reported on three groups. Birds that never saw a night sky before their first autumn could not pick a direction under the dome of a planetarium. Birds shown a normal, turning planetarium sky headed south. A third group saw a sky that turned around Betelgeuse, a bright star in Orion; in autumn they headed away from Betelgeuse, as if it marked north. So the star compass is learned. A young bird is not born knowing the stars: it finds north at the sky's centre of rotation, then learns the star patterns nearby.

The Earth's magnetic field gives a compass that works by day or night, under any sky. Its field lines are not level everywhere. At the magnetic equator they run parallel to the ground; towards either pole they tilt more and more steeply, until at the magnetic poles they stand upright. This angle between the field lines and the ground is the field's inclination, and it grows with latitude.

In 1972 Wolfgang and Roswitha Wiltschko tested caged European robins inside coils that could reshape the field around them. Reversing the whole field, north for south, made no difference to the birds. Reversing only the tilt sent them the opposite way. A robin's compass does not read which end is north, as a needle does. It reads which way the lines slope into the ground: towards the nearer pole, or towards the equator.

Every sense has to turn something in the world into nerve signals, a step called transduction. For the magnetic sense, nobody yet knows where or how that happens. Two main ideas compete.

One looks to the eye. The robin's magnetic compass depends on light, and a light-sensitive protein in the retina, cryptochrome, may be the sensor: light could start a chemical reaction in it whose outcome depends on how the protein lies in the field. In 2021 the robin's form of this protein was shown to respond to magnetic fields in the laboratory. The other idea is tiny crystals of magnetite, a magnetic iron mineral. Iron-rich cells in pigeons' beaks, long the favourite candidate, proved in 2012 to be immune cells. Yet in 2026 magnetic iron particles turned up in immune cells of the pigeon's liver, and pigeons without those cells got lost under cloud. The search goes on.

The magnetic field can serve as a map as well as a compass. Its inclination changes mostly with latitude, and its strength varies across the globe too, so most places have their own pairing of the two: a magnetic signature.

Loggerhead sea turtles appear to use it. Hatchlings from Florida spend years circling the North Atlantic in a great loop of currents. In 2001 Kenneth Lohmann's team put hatchlings in tanks where the field copied that of three distant points on the loop; at each, they swam in the direction that would keep them on course. They had never been to sea, so this first, simple map seems to be inborn. The signature of home, by contrast, seems to be learned: grown females nest near the beach where they hatched, and over nineteen years, as the field drifted, their nesting shifted along the Florida coast in step with it.

Salmon hatch in a stream, spend years at sea, then swim back to spawn in the stream where they began. In 1954 Warren Wisby and Arthur Hasler trapped coho salmon that had just chosen one of the two branches of Issaquah Creek, near Seattle. They plugged the noses of about half the fish, carried all of them back below the fork and let them go. Fish that could smell mostly chose their home branch again; fish with plugged noses often took the wrong one.

Young salmon learn their stream's smell before they leave, a lasting early learning called imprinting. In a 1976 test, young coho exposed to a particular chemical and released into Lake Michigan returned 18 months later, mostly to the stream scented with it. Smell only works near the river, though; out in the open ocean, salmon may find the right stretch of coast by its magnetic signature.

A desert ant leaves its nest and zigzags across open sand looking for food. Once it finds some, it runs almost straight back towards the nest. It can do this because it keeps a running total of each direction it has walked and each distance it has covered, so it always knows its displacement from the nest, the straight line home. This is path integration. Direction comes from the ant's sky compass, which reads polarised light.

How does it judge distance? In 2006 Matthias Wittlinger and colleagues trained ants to walk 10 metres to a feeder, then changed their legs there: glued-on pig bristles lengthened some, and others' legs were shortened. Heading home, ants on stilts searched for the nest about 15 metres out, and ants on the shortest stumps at under 6. The ants judge distance, in effect, by counting their steps.