Polarised light vision

Two things have to be true for this to work as a compass, and both are. The pattern has to be predictable from the sun’s position, which it is — the geometry of scattering fixes it. And the animal has to be able to measure the direction of vibration, which needs receptors that are not equally sensitive in all directions. Insect photoreceptors are built from stacked membrane tubes, the microvilli, which all lie parallel within one receptor; that alignment makes the receptor respond most strongly to light vibrating along it. Put several such receptors at different angles side by side and the comparison between them gives the direction.

Human photoreceptors are not built that way, and the pigment in them is oriented such that the response is essentially the same whatever the direction of vibration. There is one small exception — Haidinger’s brush, a faint yellowish figure some people can learn to see when looking at a uniform polarised field — which is real, and which is roughly as far from an insect’s sky compass as noticing a draught is from having a barometer.

A sky compass on its own gives direction and not time, and the sun moves. Animals using it therefore need a clock as well: a bee that has learned a food direction in the morning and flies to it in the afternoon must compensate for the sun having travelled, and it does. That is why the polarisation compass is best understood as one input to a system rather than a sense that solves navigation by itself.

• Bees use the pattern to fix the direction they communicate in the waggle dance, which is why the dance still works under partial cloud.
• Desert ants combine it with a step counter, giving a direction to walk and a distance to walk it — enough to return in a straight line after a wandering search.
• Dung beetles use it at night, orienting to the polarisation pattern the Moon produces, which is the same physics at about a millionth of the intensity.
• Cephalopods and many crustaceans have polarisation-sensitive eyes underwater, where the pattern is produced by scattering in the water rather than in the sky.

1. 1844

   First observation

   A faint figure in a uniform sky

   Wilhelm Haidinger described a small yellowish brush-shaped figure visible to some people when looking at polarised light. It established that human vision is not entirely blind to polarisation, and — because the effect is so weak — it did nothing to suggest any animal might navigate by it.

2. 1949

   Landmark experiment

   Bees turn when the patch of sky above them is rotated

   Karl von Frisch found that dancing bees, shown only a patch of blue sky, changed the direction of their dances when a polarising filter above them was rotated — and that the dances stayed correct under cloud so long as some clear sky remained visible. The sky, not just the sun, was carrying the direction.

3. 1969

   Modern discovery

   The part of the eye that does it

   Anatomical and electrophysiological work identified a specialised region at the dorsal rim of the insect eye whose receptors are arranged at different angles and are strongly polarisation-sensitive, while the rest of the eye is not. A behavioural capacity acquired a specific piece of hardware.

4. 1997

   Reinterpretation

   The compass is one input, not the navigation system

   Work on desert ants established that the sky compass supplies direction into a path-integration system that separately tracks distance, and that ants deprived of one still use the other. The compass became a component rather than an explanation.

   Changes how the 1949 result reads

   The 1949 result showed that bees take a direction from the sky. It was often read as showing how insects navigate; what it actually identified was one input into a system that also counts steps, keeps time and stores landmarks.

5. 2006

   Landmark experiment

   The same solution across ants, bees and beetles

   A synthesis of the celestial compass work across insect groups set out how the dorsal rim, the sun compass and the internal clock are combined, and how similar the arrangement is in animals that are not closely related.

   Polarization vision

• Do any vertebrates genuinely use polarisation?

  Why it matters: It is claimed for several fish and for some birds, and vertebrate photoreceptors lack the structural feature that makes insect ones sensitive — so a positive result would need a different mechanism.

  What would settle it: Behavioural tests with intensity and wavelength controlled as carefully as the polarisation is, which is where most of the existing work is criticised.

• What do polarisation-sensitive animals use it for besides orientation?

  Why it matters: Cephalopods have polarisation-sensitive vision and body patterns that vary in polarisation, which suggests a signalling channel invisible to most of their predators. The evidence is suggestive rather than settled.

  What would settle it: Behavioural tests in which the polarisation of a displayed pattern is manipulated independently of everything else about it.

• How is the pattern used when most of the sky is obscured?

  Why it matters: The compass is often described as working under cloud, and how small a patch of clear sky is enough — and how the animal handles a patch giving an ambiguous reading — is not well characterised.

  What would settle it: Systematic occlusion experiments, which have been done in a few species and not across the range of animals that use the compass.