Inscribed Square Problem: Why Is It Still Unsolved?

Quick answer: The inscribed square problem, or square peg problem, asks whether every simple closed curve in the plane contains four points that are the corners of a square. Otto Toeplitz posed it in 1911. It is proved for convex curves, smooth curves and polygons, but open for general continuous curves. In 2020, Joshua Greene and Andrew Lobb proved every smooth closed curve contains rectangles of every aspect ratio.

Draw any closed loop on paper, as wiggly as you like, as long as it does not cross itself. Can you always find four points on it that form a perfect square? Try it on a circle, an ellipse or a kidney shape and you will find one. Yet after more than 110 years, nobody can prove it works for every loop.

The Problem

A Jordan curve is a closed loop in the plane that does not cross itself. A square is inscribed in the curve if all four of its corners lie on the curve (the sides may cross the curve). Otto Toeplitz asked in 1911: does every Jordan curve have an inscribed square?

For a circle, every inscribed square has side \(r\sqrt2\), from the Pythagorean theorem applied to its diagonal \(2r\). For an equilateral triangle of side 1, the inscribed square resting on one side (two corners on that side, one on each other side) has side

\[ s = \frac{\sqrt3}{2 + \sqrt3} = 2\sqrt3 – 3 \approx 0.464. \]

What Is Proved

Type of curveInscribed square?Proved by
Convex curvesYesEarly work by Arnold Emch (1913); later general proofs
Smooth curvesYesLev Schnirelmann (1929), and others
PolygonsYesKnown (piecewise linear case)
Locally monotone curvesYesWalter Stromquist (1989)
Curves satisfying a quantified closeness and winding condition relative to a smooth curveYesGregory Chambers (2022 preprint, published 2025)
Every continuous Jordan curveOpenNo proof

The difficulty is wild curves, such as fractal-like loops, where squares might shrink to a point in the limit and disappear.

Rectangles: The Möbius Strip Trick

A weaker question has a beautiful answer: every Jordan curve contains an inscribed rectangle. The proof, credited to Herbert Vaughan, pairs up points of the curve. Two pairs with the same midpoint and the same distance form a rectangle. Map each unordered pair to a point in 3D: above its midpoint, at a height equal to the distance between the two points. The space of unordered pairs is a Möbius strip whose edge lands on the curve itself, in the plane. If no two pairs landed on the same point, gluing the flat region inside the curve onto that edge would place a real projective plane in 3D space without self-intersection, which is impossible. So two different pairs share a midpoint and a length, and they form a rectangle. 3Blue1Brown’s video below shows this argument.

Joshua Greene and Andrew Lobb went further in 2020 (published in the Annals of Mathematics in 2021): every smooth Jordan curve contains a rectangle similar to any prescribed rectangle, that is, of every aspect ratio, using symplectic geometry. A 2026 preprint by Asano and Ike claims the same for all finite-length Jordan curves.

Related 2026 Developments

In October 2026 (revised November 4), Lorenzo Valerio posted a preprint resolving a variant raised by Terence Tao: four simple closed polygonal curves on a cylinder, each winding once, jointly inscribe a square (possibly degenerate) whenever an alternating signed-area condition holds. The proof uses Lagrangian Floer homology. It does not resolve Toeplitz’s original question. As of November 2026 the general inscribed square problem remains open.

Common Misconceptions

“The square has to fit inside the curve.” No. Only the four corners must lie on the curve.

“It is solved because smooth curves work.” Approximating a rough curve by smooth ones gives squares that might shrink to a single point, so that argument fails.

Watch: The Inscribed Square Problem in Video

If you learn better by watching, these videos are worth your time.

This open problem taught me what topology is (3Blue1Brown). The Möbius strip proof for rectangles, beautifully animated.
Attempt of Toeplitz Conjecture Proof (Alex Khmil). A visual exploration of why the problem is hard (not an accepted proof).

Check Your Understanding

Five quick questions to see how much of this has stuck. Pick an answer and press Submit; you will see a short explanation either way.

Answer each question and press Submit to see your result right away.

1. Who posed the inscribed square problem in 1911?
Found a bug? Let us know!
2. What must lie on the curve for a square to be inscribed?
Found a bug? Let us know!
3. For which curves is the problem still open?
Found a bug? Let us know!
4. Which shape helps prove every Jordan curve has an inscribed rectangle?
Found a bug? Let us know!
5. What did Greene and Lobb prove in 2020?
Found a bug? Let us know!

Frequently Asked Questions

What is the inscribed square problem?

It asks whether every simple closed curve in the plane contains four points forming the corners of a square. It is also called the square peg problem or Toeplitz conjecture.

Has the inscribed square problem been solved?

Not in general. It is proved for convex, smooth and polygonal curves and several other classes, but open for arbitrary continuous curves as of November 2026.

Does every closed curve contain a rectangle?

Yes. Every Jordan curve has an inscribed rectangle, by Vaughan’s topological argument using the Möbius strip and the projective plane. Greene and Lobb (2020) showed smooth curves have rectangles of every aspect ratio.

Why is the problem so hard?

For rough curves, squares found on smooth approximations can shrink to a point in the limit, so the standard approximation argument breaks down.

Keep Exploring

Sources and Further Reading

Related to This Article

What people say about "Inscribed Square Problem: Why Is It Still Unsolved? - Effortless Math"?

No one replied yet.

Leave a Reply