Ebbinghaus Illusion

Same circle, different company

Two highlighted circles of exactly the same diameter sit in the figure above. One is surrounded by a ring of large circles, the other by a ring of small ones. The one among large neighbours looks noticeably smaller. Cover the rings, or measure the two circles, and the difference disappears.

Size is judged by comparison

The visual system rarely judges size as an absolute measurement. It compares a shape with what's around it, a strategy that works well in the real world, where a person next to a door or a coin in a hand is a quicker guide than any internal ruler. The Ebbinghaus figure isolates that comparison. Among large circles, the middle circle is small by comparison; among small circles, the same circle is large.

Distance matters too

The size of the surrounding circles isn't the only factor; so is how far away they are. Roberts and colleagues varied the two separately and found that both affect the strength of the illusion (Roberts, Harris & Yates, 2005). In the classic layout, used in the test above, the large circles sit further out and the small ones are packed in close, so the two factors work together.

A close relative, the Delboeuf illusion, uses a single ring instead of a set of circles: a circle with a ring drawn close around it looks bigger than the same circle with a ring further away. Both rely on the same principle, size judged in context, even though the geometry differs.

How the test above works

The test runs eight trials in random order. In three, you resize a circle surrounded by small neighbours to match one surrounded by large neighbours; in three more, it's the other way round, so a habit of always setting the circle too big or too small cancels out instead of masquerading as the illusion. The last two trials take the rings away altogether. They measure your ordinary eyeballing error, which gives the ring trials something to be compared against.

The adjustable circle starts at a random size each time, and the two figures sit at slightly different heights so you can't judge them by lining up their top edges. The score is measured by diameter: a result of 10% means the circle among large neighbours had to be 10% wider before the two looked the same. The rings follow the classic layout, with large circles spaced well out and small circles packed in close, so the score reflects both the size of the neighbours and their distance.

Why your score won't match everyone else's

People differ a lot in how strongly this illusion catches them, and some of that difference has been traced to the brain itself. Samuel Schwarzkopf, Chen Song and Geraint Rees measured the surface area of each participant's primary visual cortex (V1), which varies considerably from person to person, and found that people with a smaller V1 tended to experience a stronger Ebbinghaus illusion (Schwarzkopf, Song & Rees, 2011).

Culture plays a part too. When Jan de Fockert and colleagues tested the Himba, a remote semi-nomadic people in northern Namibia, they found the illusion was weaker than in English participants, so the Himba judged the circles' sizes more accurately. The researchers put this down to a habit of attending to local detail rather than to the scene as a whole (de Fockert et al., 2007). If you're good at ignoring the rings, you're in interesting company.

Do your hands see it?

The illusion is also a test case in a long-running debate about whether vision for perception and vision for action are separate. Salvatore Aglioti and colleagues had people pick up discs set in an Ebbinghaus arrangement and found that, although the discs looked different sizes, people opened their fingers to about the right width when reaching for them (Aglioti, DeSouza & Goodale, 1995). Volker Franz and colleagues then repeated the comparison with more careful controls and found that grip was fooled about as much as perception, once both were measured the same way (Franz et al., 2000). The question is still argued over, and the Ebbinghaus illusion remains the favourite test.

Ebbinghaus or Titchener?

The illusion is named after the German psychologist Hermann Ebbinghaus, better known for his pioneering research on memory and forgetting. The American psychologist Edward Titchener made the circles-of-circles version widely known through his 1901 laboratory manual (Titchener, 1901), which is why it's also called the Titchener circles. It belongs with the Müller-Lyer and Ponzo illusions, both of which have their own tests on this site.