Migratory birds carry the light-sensitive protein cryptochrome (CRY4) in retinal photoreceptor cells. Absorbing a blue/green photon (~420–565 nm) kicks an electron across a chain of tryptophan residues, leaving a correlated radical pair whose two unpaired electron spins coherently interconvert between quantum singlet and triplet states. Earth's weak field (~25–65 µT) tips this interconversion rate depending on the angle between the field and the molecule — a genuinely quantum-mechanical compass. Longer wavelengths (>590 nm, orange/red) or darkness don't supply enough photon energy to trigger the reaction, and caged/disoriented-bird experiments confirm birds lose magnetic orientation under red light exactly as modeled here.
Because the radical-pair signal only depends on the axis the field makes with the molecule (mod 180°), not which way the field vector "points," birds read inclination — the dip angle between field and horizontal — not polarity like a compass needle. The classic Wiltschko & Wiltschko (1972) proof: reverse the entire field vector (both horizontal and vertical components) and a real bird is unaffected, because the axial line is identical; but reverse only one component and the bird gets confused or reverses, because that changes the axis. This sim's dipole field also gets more intense toward either pole (B ∝ √(1+3sin²λ), polarity-independent) — a second cue a bird can combine with the inclination axis to tell "poleward" from "equatorward" without ever needing polarity.
- Latitude — sets the local dipole inclination via the real relation tan(I) = 2 tan(λ).
- Polarity toggle — flips the true field vector. The inclination-compass bird's course is unchanged; the naive polarity-compass bird's course flips 180°.
- Light — gates the radical-pair reaction; no clear signal in the dark or under red light.