A European robin does not have a north-south compass. It has something stranger: an inclination compass. The distinction matters.
In 1972, Wolfgang and Roswitha Wiltschko ran a definitive test. Using Helmholtz coils, they could manipulate the magnetic field around a robin in a cage — redirect north, invert the vertical component, reverse the horizontal component, or flip both simultaneously. The results were diagnostic. Reversing only the vertical component: birds became disoriented. Reversing only the horizontal component: same. Reversing both together — which flips polarity while leaving the geometry of field lines intact — had no effect at all. The birds went on orienting as if nothing had changed.
What this means: a robin is blind to polarity. It does not detect which end of the field line points toward the geographic pole. Instead it reads the inclination — the angle the field lines make with gravity. In the Northern Hemisphere, field lines dip downward toward the pole. In the Southern Hemisphere, they dip downward toward the other pole. A polarity compass would require recalibration at the magnetic equator; an inclination compass does not. The robin that migrates from Europe to sub-Saharan Africa uses the same mechanism on both sides of the equator without any adjustment. The geometry is the signal.
This should have been a clue about the underlying mechanism. A magnetite-based compass — iron crystals physically rotating to align with the field — would detect polarity naturally. Something that detects inclination without polarity must be responding to field-line orientation, not direction. And it must be doing so chemically.
The mechanism is this: a photon of blue or green light (up to about 565 nm) is absorbed by a flavin chromophore in a protein called cryptochrome 4a, located in the outer segments of double-cone photoreceptors in the robin's retina. The photon excites the flavin. Within 0.4 picoseconds, an electron is transferred from a nearby tryptophan amino acid to the excited flavin. Then a second tryptophan donates to the first; then a third to the second; then a fourth to the third. A chain of four electron hops, taking about 200 picoseconds total, ending with two electrons separated by more than 18 angstroms — one on the flavin, one on the terminal tryptophan — each carrying an unpaired spin.
This is a radical pair. What happens next is quantum mechanics.
The two electrons are created in a singlet state: spins antiparallel, quantum mechanically correlated. Inside the molecule, each electron spin is coupled to nearby atomic nuclei — protons, nitrogen — through hyperfine interactions. These nuclear spins are random and incoherent, but their magnetic influence is precisely what drives the singlet state to oscillate coherently into the triplet state (spins parallel) at megahertz frequencies. Earth's magnetic field modulates this oscillation. By altering the relative precession rates of the two electron spins, the external field biases the ratio of time the pair spends in each spin state. When the pair recombines — which can only happen from the singlet state — the field-dependent yield of recombination products encodes the orientation of the molecule relative to Earth's field.
The number that should stop you: Earth's field is roughly 50 microtesla. The interaction energy between this field and an electron spin is about a million times smaller than kBT — thermal energy at body temperature. Classical physics says this should be lost in noise. Yet the experiment works. The reason it isn't lost is that the radical pair is not in thermal equilibrium: it was created in a specific quantum state by a photon, and the spin dynamics are coherent, not thermally driven. The magnetic field is acting on a non-equilibrium quantum system, and the relevant comparison is not field energy versus thermal energy but field-induced precession rate versus spin relaxation rate. This is the crux. Genuine quantum mechanics is not a description here — it's a requirement. Peter Hore showed in 2020 that semiclassical approximations to the spin dynamics yield directional predictions off by 15 to 30 degrees.
The compass only works within a functional window: roughly 20-25% of the local geomagnetic field strength in either direction. Too weak, too strong — orientation collapses. This is unintuitive for a polarity detector and makes perfect sense for a quantum chemical mechanism sensitive to the ratio of the magnetic interaction to hyperfine coupling strengths. At the wrong field strength, the coherence structure changes and the angle signal disappears.
The compass only works in certain wavelengths of light. Blue (424 nm): robins orient. Turquoise (501 nm): yes. Green (565 nm): yes. Yellow (590 nm): disoriented. Red: disoriented. The cutoff aligns precisely with cryptochrome's absorption spectrum, not rhodopsin. In 1978, Klaus Schulten had predicted a biochemical radical pair compass on theoretical grounds, with no specific molecule. In 2000, after learning about cryptochromes from a colleague, he proposed them specifically. The gap between mechanism and molecule was 22 years.
In 2021, Jingjing Xu, Lauren Jarocha, Henrik Mouritsen, Schulten, Hore, and Ilia Solov'yov's group published the first in vitro demonstration in Nature: purified European robin Cryptochrome 4 is magnetically sensitive under conditions relevant to Earth's field. They substituted each of the four tryptophans individually with phenylalanine to block each hop in the electron transfer chain, confirming all four are necessary. Crucially, robin CRY4 was more magnetically sensitive than CRY4 from chicken or pigeon — non-migratory species. This was the cover paper for Nature on June 23, 2021.
In 2014, Henrik Mouritsen published a result that is still striking. When his lab moved to Oldenburg in 2002, the birds stopped orienting. For three years they jumped in random directions. Eventually his graduate student Nils-Lasse Schneider suggested lining the huts with aluminum sheeting. The birds oriented immediately. Seven years of fully double-blinded experiments followed. The conclusion: anthropogenic electromagnetic noise in the 2 kHz to 5 MHz range — generated by ordinary computers, monitors, and fluorescent lights in the building — was sufficient to disrupt magnetic compass orientation completely. The fields involved were on the order of nanotesla. This is orders of magnitude below any known biological or thermal threshold. Only a radical pair mechanism would be sensitive to oscillating fields in this frequency range, because RF at the electronic Larmor frequency drives resonance transitions between singlet and triplet states. Ordinary electronics, at ordinary indoor levels, can blind a migratory bird to Earth's magnetic field.
A 2024 paper in Nature Communications added another piece: the quantum Zeno effect. In the standard model, the radical pair is well-separated and weakly coupled, allowing spin dynamics to proceed freely. But an alternative pair — flavin and superoxide — forms within cryptochrome and is tightly bound. Tight binding normally kills field sensitivity because strong exchange interactions arrest the spin dynamics. The 2024 result showed that if recombination is strongly asymmetric (one spin state recombines much faster than the other), the quantum Zeno effect — a phenomenon from the foundations of quantum measurement theory, where repeated "observation" of a quantum state inhibits its evolution — can actually restore magnetic sensitivity to the bound pair. A mechanism from quantum measurement theory, operating in bird retinas.
The visual hypothesis: because cryptochrome is in the retina and the information routes through Cluster N (a visual processing region in the forebrain), the current best model is that robins experience magnetic information as a modulation of their visual field — a pattern superimposed on what they see, symmetrical around the magnetic field axis, rotating as their head moves. Not a feeling of direction, but a visible feature of the visual scene. The robin may not sense north the way we sense pressure or temperature. It may see it.
The most precise test of the visual hypothesis requires finding what, exactly, the pattern looks like from inside. That experiment has not been done and probably cannot be done. What can be done — and was done, in the 2004 Ritz paper, and the 2014 Mouritsen paper, and the 2021 Nature paper — is ruling out every other plausible mechanism. The inclination compass in a migratory bird's eye is almost certainly a quantum chemical transducer, using correlated electron spins and the modulation of their coherent dynamics by Earth's 50-microtesla field, processed through the visual cortex, resulting in the perception — if perception is the right word for it — of something like a directional gradient in the light.
A robin navigating at night over the Mediterranean, oriented by a field a million times too weak to matter thermally, seeing north as a feature of the dark.