Two Identical Triangles. Position Alone Makes Them Different Greys.
Which patch is lighter?
You are looking at the Benary cross illusion, described by Gestalt psychologist Wilhelm Benary in 1924. A black cross (think plus-sign) sits on a white background. Two identical grey triangles are placed: one in the concave angle formed where the cross meets the white (nestled into an inside corner); the other against the edge of one of the cross’s arms (sitting out in the white background against the black edge). The two triangles have, depending on exact geometry, roughly the same amount of black-vs-white neighbourhood · but their perceived brightness differs visibly. One looks lighter, the other darker.
What you are about to learn. What the Benary cross actually is, why it is one of the earliest illusions to argue that brightness depends on perceptual grouping (not local contrast), how it challenged lateral-inhibition theory, the geometric variation you can do to make or break the effect, and its relationship to Adelson, Koffka, and White.
What the Illusion Looks Like
Draw a black plus-sign (cross) on a white background. Now place two identical medium-grey triangles.
- Triangle A sits outside the cross, in the white background, against one arm’s edge. Most of its neighbourhood is white; it touches black only along one straight edge.
- Triangle B sits inside one of the concave corners formed where two arms meet. It is wedged into the black, touching black along two straight edges; most of its neighbourhood is… also white (because the concave corner is filled with white).
Here is the punchline: Triangle A and Triangle B have nearly identical amounts of neighbouring black and neighbouring white. By local-contrast theory, they should look equally grey. But they do not · Triangle B, the one wedged inside the concave corner, looks lighter than Triangle A.
The minimal recipe. A black figure on a white background with two concavities. Identical grey triangles placed so that one is visually “part of” the black cross (wedged into its corner) and the other is visually “part of” the white background (against its edge). The one grouped with the black reads as lighter, because it is compared against black. The one grouped with the white reads as darker, because it is compared against white.
Why It Works
The Benary illusion is a grouping-based brightness effect · closely related to the Koffka ring and White’s illusion. The mechanism:
Your visual system parses the figure into objects. The black cross is a single object. The white background is a single region. Triangle A visually belongs to the white region (its edges fit the background’s geometry). Triangle B visually belongs to the black cross (it fits into the cross’s concavity).
Brightness is computed per perceptual group. Each triangle is assigned a brightness relative to the object it belongs to · not relative to its raw local pixel neighbourhood.
Triangle A is relative-to-white: reads darker. Triangle B is relative-to-black: reads lighter. Same pixels, different reference frames.
This is grouping-based brightness, at the highest resolution. A local-contrast theory (lateral inhibition only) predicts no difference between the two triangles · their total immediate neighbourhoods are nearly balanced. The fact that we see a robust difference means perceptual grouping is doing real work. Benary’s 1924 paper was one of the first to document this, ahead of Koffka (1935) and well ahead of White (1979). The Gestalt psychologists had the right hypothesis decades before mainstream perception science caught up.
The Triangle-Orientation Variation
What happens if you rotate one of the triangles so that it no longer “fits” into the cross’s concavity? The effect weakens dramatically.
Test it. In your mind, take the “inside the cross” triangle and flip it so its hypotenuse faces the wrong way · so it no longer wedges perfectly into the concave corner but instead sits at an awkward angle, no longer grouped with the cross. The illusion collapses: both triangles now read as similar grey. The brightness difference was entirely a by-product of perceptual grouping. Change the grouping, change the perception.
Why Classical Contrast Theory Fails
Pre-Gestalt brightness theory held that a grey patch’s perceived brightness was determined by the local luminance around it · specifically, by the difference between the patch luminance and the average luminance in an immediately-surrounding annular region. This is the standard lateral-inhibition account.
Common misconception: “there is more black near one triangle than the other.” Measure it. In any sensible definition of local neighbourhood, the two triangles have nearly equal black:white ratios. If your computer-vision model only uses local pixel statistics to predict brightness, it will predict the two triangles are identical · and it will be wrong. The perceptual system is reading global object structure, not local pixel averages.
The Gestalt Heritage
Benary was a student of Max Wertheimer, one of the founders of Gestalt psychology. His 1924 paper was part of the broader Gestalt project of demonstrating that perception cannot be reduced to local pixel-level operations. The Koffka ring (1935), the Kanizsa triangle (1955), and eventually White’s illusion (1979) all fall in the same lineage · they are all evidence that the whole is not the sum of the parts in visual perception.
The Gestalt canon. If you were building a course on brightness perception and wanted the three best demonstrations of grouping-over-local-contrast, you would choose: Benary cross (position within a figure), Koffka ring (splitting a figure disrupts grouping), and White’s illusion (figures embedded in patterns). All three run on the same engine. Benary is the oldest of the trio and arguably the cleanest · it requires only one black cross and two grey triangles to produce the effect.
A Harder Variant
Below is a Benary cross figure at difficulty 3, with more aggressive triangle placement. The two triangles are always identical in pixel value.
Which patch is lighter?
The paper-covered proof. Take two small pieces of white paper and cover everything in the figure except the two triangles. Now you can see they are the same grey. Lift the papers to reveal the cross and the background, and the grouping reactivates, splitting the two triangles apart in perceived brightness. This is the fastest proof that the perceptual grouping is the active ingredient · remove the context, the grouping disappears, and the triangles equalise.
Where the Benary Appears
- Iconography and signage. A grey symbol placed into a concave area of a logo reads as a different brightness than the same symbol placed against a straight edge of the logo. Icon designers working with complex black-on-white figures use the Benary mechanism to adjust perceived element weights.
- Typography. Letterforms with complex interior spaces (serifs, counters) produce Benary-style brightness shifts on adjacent glyphs. Typographers have been silently correcting for this for centuries.
- Architecture. Benches, statues, and signage placed into concave spaces of building facades (recessed niches, alcoves) are perceived with different brightnesses than the same objects placed against flat walls. The Benary effect is a small but real contributor to how recessed features “feel” to the eye.
- Graphic design. The layout relation between a small grey element and a large black figure changes the element’s apparent lightness · a designer placing a grey dot inside a black shape’s concavity will find it reads as noticeably lighter than the same dot placed on the shape’s straight edge.
Test Yourself on 50 More Illusions
The Benary cross is one of more than 50 classical illusions on PlayMemorize. Each round draws a deterministic SVG scene and asks one grounded question: which is larger, which is brighter, which is actually parallel. The reveal overlay shows the true geometry plus a one-line “why it works” caption.
- Keep playing Benary Cross → · the standalone game, pinned to this one figure with fresh seeds each round
- Play Illusions → · spot the tricks across size, colour, orientation, and impossible figures
- Play Spatial → · train mental rotation and area estimation
- Play Matrix → · abstract pattern reasoning under time pressure
The takeaway. The Benary cross is early, clean evidence that brightness perception is not a pixel-level operation · it is an object-level operation, running on perceptually-grouped representations. Two triangles with identical local neighbourhoods can look radically different depending on whether your visual system has grouped them with a black figure or with a white background. Wilhelm Benary demonstrated this in 1924, the Gestalt movement made it a signature result, and modern computational vision keeps rediscovering the same lesson: perception happens on objects, not on pixels.
Illusions
Your eyes lie - the math knows the truth. Spot equal lengths, identical greys, and truly parallel lines across 57 classic optical illusions
Gioca ora - è gratisNessun account necessario. Funziona su qualsiasi dispositivo.
