A migratory bird steers by the Earth’s magnetic field, and the field is far too weak to be felt.
That’s the part I keep circling. Not weak the way a whisper is weak — weak past the point where the ordinary rules of sensing should work at all. The field at the ground is about fifty-millionths of a tesla. A fridge magnet is a hundred times stronger and a bird flies through that field its whole life without it doing anything to its body. There’s a reason: the energy a field that faint can deposit in a single molecule is something like a million times smaller than the thermal noise that molecule is already drowning in. Everything at body temperature is jostling, colliding, getting knocked around by heat. A push that small can’t hold a molecule still, can’t bend it, can’t tip it one way. By the only arithmetic that should matter, the signal is buried under the noise by six orders of magnitude before it begins. The field cannot be the push. There is no push.
So the bird does something else. In a protein in the retina — cryptochrome, the leading candidate, though I’ll come back to how unfinished that word “candidate” still is — a photon of blue light knocks an electron loose and strands it one molecule over. Now there are two molecules each holding one unpaired electron: a radical pair. And the two lonely electrons remember each other. Their spins start out correlated, locked in step, and then they begin to drift in and out of phase, oscillating between two configurations the chemists call singlet and triplet, millions of times a second.
Here is the whole trick. That oscillation has no energy barrier. It’s not a thing the field has to shove over a hill; it’s a coherent precession, a turning, and a turning can be nudged by a vanishingly small torque if you just wait. The faint field can’t move a molecule, but it can change the timing of the spins’ drift between singlet and triplet. And the two states, when the electrons finally recombine, fall into different chemical products. Singlet makes one thing. Triplet makes another. So the angle of the magnetic field gets written, slowly, into the ratio of two chemicals piling up in the bird’s eye.
That’s what the bird reads. Not a force. A yield. A concentration. The field never pushes anything; it only biases which way a reaction was already going to fall — tips a coin that was already in the air. Evolution found the one place in the body where a signal too faint to do any work can still leave a mark, because in that place nothing has to be moved, only chosen between.
Two more things I didn’t know I’d find. The first: it isn’t a north-pointing compass at all. It reads the inclination — the angle the field lines dip into the ground — and it can’t tell the north magnetic pole from the south. It knows “toward the nearest pole” and “toward the equator,” nothing more. Flip the vertical part of the field in a lab and the bird turns around, confidently, the wrong way. The second: for the ratio to come out sharp enough to steer by, the two electrons have to stay quantum-coherent — in lockstep, undisturbed — for a few microseconds. That is an absurdly long time for a delicate quantum correlation to survive inside a warm, wet, rattling protein. By rights the noise should erase it almost instantly. It doesn’t, and a computational paper this past November is still working out exactly how the molecule holds the pair far enough apart, along the right electron-transfer pathway, to keep the lockstep alive long enough to matter.
And the honest end of it is that nobody has caught the bird in the act. Cryptochrome is the best guess, confirmed magnetically sensitive in a test tube, modeled in finer and finer detail — but the chain from that molecule in a dish to the warbler turning south on a cloudy night isn’t closed. The leading explanation for how an animal reads the planet is a force too faint to feel, registered as the faint preference of a coin that was already falling, and we are still not sure it’s the right story.
Sources: The quantum needle of the avian magnetic compass (PNAS); Night-migratory songbirds possess a magnetic compass in both eyes (PLOS One); The quantum compass mechanism in cryptochromes (arXiv, 2025); Princeton: computational support for radical-pair magnetoreception.