Inscribed Square Problem

Importance: Medium ✭✭
Author(s): Toeplitz
Subject: Topology
Recomm. for undergrads: no
Posted by: dlh12
on: April 10th, 2008
Conjecture   Does every Jordan curve have 4 points on it which form the vertices of a square?

A Jordan curve is a continuous function $ f $ from the closed interval $ [0,1] $ to the plane $ \mathbb{R}^{2} $ with the properties that $ f $ is injective on the half-open interval $ [0,1) $ (i.e., $ f $ is simple) and $ f(0)=f(1) $ (i.e., $ f $ is closed).


[M] Meyerson, M.D., Equilateral triangles and continuous curves, Fund. Math. 110, (1980), 1--9.

* indicates original appearance(s) of problem.

Already solved

This was proven in

inscription of squares in simple closed curves

There is a theorem that says:

in all simple closed curves there are 4n points that are vertex of n squares (inf = > n > =1)

Jorge Pasin.

in all simple closed curves there are 4n points

Would you clarify? An obtuse triangle has only one inscribed square, so this theorem is not true for n>=2. Do you have a reference to this theorem? Strashimir Popvassilev

Is the conjecture known to

Is the conjecture known to be true for C^1-smooth curves?


Yes. Walter Stromquist, Inscribed squares and square-like quadrilaterals in closed curves, Mathematika 36: 187-197 (1989).


If it were true for C^1 curves, then since a Jordan curve is compact, it may be weierstrass approximated by a series of C^1 curves (indeed by curves whose component functions are polynomials) such that the series converges uniformly to the given jordan curve. Then by assumption, each curve in the sequence contains 4 points forming a square, and the sequence of squares can be regarded as (eventually) a sequence in the (sequentially) compact space of the 4-fold product of any closed epsilon enlargement of the area bounded by the original jordan curve. It follows that the sequence of squares contains a convergent subsequence, which can be shown to be a square lying on the original jordan curve.

Thus, proving the C^1 case proves the general case.

This is flawed

The approximation argument is flawed: the squares on approximating curves may have sides decreasing to 0, in which case the limiting "square" degenerates to a point. In fact, Stromquist's theorem covers a much wider class of curves than C^1, but not all continuous curves.


Phrasing should be changed from "Does any..." to "Does every..."

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.