Barber Pole Illusion

Stripes that climb forever

A barber's pole is a cylinder wrapped in a diagonal red and white stripe, turning on its vertical axis. Every point on it moves in a horizontal circle. Yet what you see is stripes flowing steadily upwards (or downwards), as if the pole were feeding out stripe from the bottom forever.

The demo shows the same diagonal stripes moving behind two differently shaped windows. Through the tall one they seem to travel up; through the wide one, sideways. The stripes are identical: only the window changes.

The aperture problem

When a straight edge moves behind a small opening, you can't tell which way it's really going. Motion along the edge's own length leaves the view unchanged, so the only thing you can measure is how the edge moves at right angles to itself. A diagonal stripe sliding sideways and the same stripe sliding diagonally along its length produce exactly the same view through a small hole. Vision scientists call this the aperture problem, and it applies to every motion-detecting neuron, each of which only "sees" through a tiny window of its own.

To resolve the ambiguity, the visual system uses the places where the motion is not ambiguous: the ends of the stripes, where they meet the edges of the window. Hans Wallach showed in 1935 that these line endings largely decide the perceived direction (Wallach, 1935; Wuerger, Shapley & Rubin, 1996). On a tall, narrow pole, most of the stripe ends sit on the long left and right edges, and they slide up the pole as it turns, so the stripes seem to rise.

Change the window, change the direction

That's why the shape of the window, not the movement of the stripes, sets the direction. Show the same moving stripes through a short, wide window and most of the stripe ends sit on the top and bottom edges, sliding sideways, so the motion looks horizontal. If you could see the whole cylinder at once, from above, the true rotation would be obvious and there'd be no illusion.

The same problem turns up whenever moving patterns are seen through openings. The brain normally solves it by combining the ambiguous signals from many edges, a process Edward Adelson and Anthony Movshon studied using overlapping moving gratings that the eye sees either as one pattern moving in a new direction or as two patterns sliding over each other (Adelson & Movshon, 1982).

A sign older than the science

The red and white barber's pole dates back centuries, to when barbers also performed bloodletting and minor surgery. The explanation came much later: Hans Wallach's 1935 study of the perceived direction of moving lines set out the problem so clearly that it's still cited today. The Shepard tone is sometimes called the auditory barber pole, because it too turns a repeating cycle into endless movement in one direction. For still images that seem to move, see the peripheral drift illusion.