Skip to content
NatureHQ

A guided journey · 4 stops

Navigation without a map

An ant crosses a salt pan with no landmarks, finds food a hundred metres from home, and runs straight back. Pick it up and move it, and it runs the same line into empty ground.

The route

  1. Path integration
  2. Polarised light vision
  3. Ants
  4. Animal navigation

This route is the smallest navigation problem in the corpus and the most completely solved, which is exactly why it is worth following. Everything about a desert ant's homeward run has been measured — the direction, the distance, the error, and the sense supplying each part.

It also answers a question the bigger animals cannot. A bird that gets home might be using anything. An ant on a featureless salt pan has almost nothing available, and the experiment that reveals the mechanism is a displacement so simple it could be done with a pair of forceps: move the animal, and watch it run a course that is now wrong in a completely specific way.

The order goes from the computation, to the sense that supplies its direction, to the animal that does it best.

  1. Stop 1 of 4. Senses and abilities

    The running sum, and the displacement that reveals it

    Begin with the computation itself, because the experiment is what makes it undeniable. An animal retracing its route and an animal integrating one behave identically until you move it — then the first is lost and the second runs a straight line it has never walked. The systematic error in that line is the fingerprint of an approximation rather than a memory.

    Read Path integrationPick up an ant that has just found food, put it down a hundred metres away, and it will run the exact course home was from where you took it — into empty ground.

  2. Stop 2 of 4. Senses and abilities

    Where the direction comes from

    A vector needs a direction, and on a salt pan under a high sun there is nothing obvious to take one from. The answer is a property of light humans cannot see: scattered sunlight is polarised in a pattern fixed by where the sun is, and a strip at the top of the insect eye reads it. Rotate a filter over an ant by forty degrees and it walks off forty degrees wrong.

    Read Polarised light visionRotate a polarising filter over an ant by forty degrees and the ant walks home forty degrees wrong. There is nothing else it could be responding to.

Where this leaves you

You end with a mechanism described precisely enough to model: a running vector, direction from the polarisation pattern of the sky, distance from stride count, and an error that accumulates and is corrected by a systematic search rather than by better arithmetic.

You also end with the limit. Path integration alone never quite gets an animal home — it gets it close, and something else finishes. That is why it is always the first stage of a return and never the last, and why an ant that has run out of vector begins looping rather than continuing.

Last reviewed 2026-09-03. All guided journeys