Solution of Erd s Problem 633

arXiv:2604.03609 · math.CO, math.MG · Submitted 2026-04-04 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: I'm Ada.

Emmy: And I'm Emmy, here with Kolya.

Ada: We're looking at the paper: "Solution of Erd s Problem 633".

Emmy: Here's the gist of the paper. It is well-known that any triangle can be tiled (cut) into any square number of congruent triangles similar to the original triangle <ref:2604.03609#pg2>.

Ada: Our work shows that, aside from isosceles triangles, this set of triangles is countable (up to similarity) <ref:2604.03609#pg2>.

Emmy: Classification of Non-Square Tilings The main result settles Erd˝os Problem 633 by classifying all triangles that can be tiled into a non-square number of congruent triangles <ref:2604.03609#pg2>.

Ada: Theorem 1 states that a triangle T admits a non-square tiling if and only if it satisfies one of eight specific conditions based on its angles (A, B, C) <ref:2604.03609#pg2>.

Paper summary: Ada: So we're looking at this paper, "Solution of Erd s Problem six hundred thirty-three" and what they've done there <ref:2604.03609#pg1>. Essentially, it settles the question about whether any triangle can be tiled by a non-square number of copies of another triangle similar to it.

Emmy: Right, the main thing is that they classify all triangles that can be tiled into a non-square number of congruent triangles, and aside from isosceles triangles, this set of triangles is countable up to similarity.

Kolya: That sounds like a lot to track, Ada. What's the core claim they are trying to prove about the tiling numbers themselves?

Ada: They show that if a triangle T has any tiling into congruent triangles that aren't similar to T, then every such tiling must have a non-square number of tiles, unless one of eight specific conditions based on its angles is met <ref:2604.03609#pg2>.

Emmy: So they start by defining what a square tiling is and what a reptiling is, and then they aim to find the exceptions where you can have any square number of tiles.

Kolya: I'm curious about those conditions. Are we talking about some simple angle relationships, or something more complex involving trigonometry?

Ada: It involves specific things like being isosceles, having certain right triangles with integer side ratios, and satisfying various trigonometric rationality constraints <ref:2604.03609#pg2>.

Emmy: That's where the algebraic geometry comes in; they connect these geometric constraints to number theory to prove that for most triangles, every square tiling must be a reptiling.

Ada: Exactly, and the paper shows that if you have a non-reptile tiling of T, the number of tiles cannot be a square unless T is isosceles or it satisfies one of those eight conditions <ref:2604.03609#pg2>.

Kolya: That sounds like they are using some kind of Diophantine equation analysis to show that the number of tiles N just can't be a perfect square under those circumstances.

Emmy: They use number theory on equations derived from the area ratio, showing that the ranks of these elliptic curves are zero, which limits solutions to torsion points <ref:2604.03609#pg3>.

Ada: So they're basically saying that if you want a square number of tiles, you'd need a specific kind of solution to an equation that just doesn't exist for most triangles.

Paper summary: Kolya: That means for the vast majority of triangles, any tiling into congruent pieces is either similar or it has a non-square number of tiles.

Ada: The forward direction they prove, showing necessity, relies on analyzing non-reptile tilings and applying theorems about group types, like Group one or Group two <ref:2604.03609#pg6>.

Emmy: And Lemma twelve shows that if you have a non-isosceles triangle T with a non-reptile tiling, one of four angle conditions must hold: C equals pi over three, B equals two A, C equals A over two plus B, or C equals two A plus B over two <ref:2604.03609#pg6>.

Kolya: Those angle conditions sound very specific. What does that imply for the actual physical tiling you're seeing?

Ada: Proposition thirteen then systematically relates the angles of T to the angles of the tile R based on those four cases through a complex table <ref:2604.03609#pg7>.

Emmy: It sounds like they are mapping out all the possible ways a non-similar tiling can happen based purely on the shape's geometry.

Ada: Then they move to the right-to-left direction, which is proving sufficiency, by actually exhibiting tilings for each of those eight cases listed in Theorem one <ref:2604.03609#pg10>.

Kolya: So they're not just proving that these conditions *must* hold if you have a non-square tiling, but they are showing that if you satisfy one of those conditions, a non-square tiling actually exists.

Emmy: For cases one through eight, they use methods like refining dissections into similar triangles or biquadratic tilings where the number of tiles N is expressed as a sum of two squares <ref:2604.03609#pg24>.

Ada: And for case seven, they show that if C equals A over two plus B and two sine of A over four is rational, then the number of tiles N equals two K squared minus M squared, which is a square only when two K squared minus M squared is itself a square <ref:2604.03609#pg18>.

Kolya: That part about N equaling two to the k squared minus M squared being a square only under specific circumstances makes sense from a number theory standpoint. What happens when we look at cases where the angles are incommensurable?

Ada: In those cases, they show that the number of tiles N cannot be a square because they analyze Diophantine equations leading to elliptic curves whose rank is zero <ref:2604.03609#pg12>.

Paper summary: Emmy: Proposition twenty-one demonstrates that for certain rational values of t, N cannot be a square because the corresponding elliptic curve has only the trivial solution of zero, zero <ref:2604.03609#pg13>.

Ada: So they rule out possibilities where angles are rational multiples of pi as well, which is important for understanding the general case.

Kolya: It sounds like they've built a really robust argument here by connecting the geometric setup to hard number theoretic constraints on those tiling numbers. What about the uniqueness of the tile itself?

Ada: Theorem thirty-two asserts that for non-isosceles triangles admitting a non-square, non-reptile tiling, the tile is uniquely determined by T unless T matches specific rows in that classification table <ref:2604.03609#pg22>.

Emmy: That uniqueness argument is strong because it shows that if you try to find another tile that fits, it would imply the angles are commensurable, which contradicts the starting point of their theorem <ref:2604.03609#pg22>.

Ada: So they're not just finding a tiling; they're characterizing the specific geometric relationship between T and R for these non-square tilings.

Kolya: Looking at the examples, like the seventy-seven-tiling with N equals seventy-seven or the twenty-eight-tiling with N equals twenty-eight those are members of what they call "triquadratic tilings" <ref:2604.03609#pg31>.

Ada: Those examples illustrate that these non-square tilings can exist even when the angles are incommensurable, which is a key point for this paper.

Emmy: And the complexity of these tilings is quite high; some numbers exceed one million tiles <ref:2604.03609#pg19>.

Ada: So to wrap up, the main result of this paper is that aside from isosceles triangles, the set of triangles that can be tiled into a non-square number of congruent triangles is countable up to similarity <ref:2604.03609#pg22>.

Kolya: It settles Erd˝os Problem six hundred thirty-three by characterizing the exceptions to generic square tilings <ref:2604.03609#pg2>, showing that for most triangles, every square tiling must be a reptiling <ref:2604.03609#pg22>.

Ada: It’s a deep dive into classifying those exceptions based on eight specific conditions derived from geometric and number-theoretic constraints <ref:2604.03609#pg2>.

Emmy: So the title, "Solution of Erd s Problem six hundred thirty-three" suggests they've managed to solve that problem by providing a complete classification of which triangles can have non-square tilings <ref:2604.03609#pg1>.

Ada: It’s a comprehensive picture of what makes a triangle special in terms of its tiling properties.

Conclusion: Ada: So we've seen how they classify these triangles that can be tiled in non-square numbers, and now we're looking at what this whole paper actually means for tiling theory.

Emmy: It’s about this paper titled "Solution of Erd s Problem six hundred thirty-three" and it tackles the question of which triangles allow for a tiling by a non-square number of congruent copies.

Kolya: Basically, they've got this classification system based on eight specific angle conditions that determine if you can have any square number of tiles or not.

Ada: It seems the main achievement here is providing a complete characterization—an "if and only if"—for these non-square tilings, which settles that problem for most cases.

Emmy: They show that aside from isosceles triangles, this set of triangles is countable up to similarity, which is pretty significant because it gives us a way to measure the complexity of these shapes.

Kolya: I mean, the fact that they link these geometric constraints directly to number theory—like those elliptic curve ranks being zero—that’s where it gets interesting for probability.

Ada: Exactly. It suggests that for most triangles, if you try to tile them into a square number of pieces, you're almost certainly looking at a reptiling instead.

Emmy: It gives us a clear map of the exceptions to the generic case, so we know exactly what kind of geometry you need to look for when hunting for those non-square tilings.

math.CO, math.MG

Submitted: 2026-04-04

Updated: 2026-11-10

Comments: 33 pages, 18 figures

Project page: https://www.lmfdb.org/EllipticCurve/Q/144/a/1

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 84/100

The gist: It is well-known that any triangle can be tiled (cut) into any square number of congruent triangles similar to the original triangle <ref:2604.03609#pg2>.

Key concepts

Non-square Tilings
This refers to tiling a triangle with congruent copies of itself such that the total number of tiles used is not a perfect square. The paper investigates which triangles allow for such tilings, moving beyond the known fact that any triangle can be tiled by an arbitrary number of similar triangles.
Reptile Tilings
A tiling is called 'reptile' if the resulting shape (the union of the tiles) is similar to the original triangle. The paper analyzes tilings that are *not* similar to the original triangle, linking this property to whether the number of tiles must be a square.
Elliptic Curves and Rank
The proof uses number theory by relating tiling parameters to Diophantine equations that define elliptic curves. Showing these curves have a rank of zero severely restricts possible solutions, proving that non-square tilings are rare and only occur under very specific, constrained conditions.
Angle Conditions
Theorem 1 provides eight specific geometric conditions based on the angles (A, B, C) of the triangle. These conditions dictate precisely when a triangle can be tiled into a non-square number of congruent triangles, involving relationships like C = π/3 or B = 2A.

Terminology

Summary

It is well-known that any triangle can be tiled (cut) into any square number of congruent triangles similar to the original triangle <ref:2604. p. 0

Where the paper says it

SOLUTION OF ERDOS PROBLEM 633 ˝ MICHAEL BEESON, MIKLOS LACZKOVICH, AND YAN X ZHANG ´ Abstract. It is well-known that any triangle can be tiled (cut) into any square number of congruent triangles similar to the original triangle.

03609#pg2>. Our work shows that, aside from isosceles triangles, this set of triangles is countable (up to similarity) <ref:2604. p. 0
Where the paper says it

We classify all triangles that can be tiled into a non-square number of congruent triangles, settling Erd˝os Problem 633. Our work shows that, aside from isosceles triangles, this set of triangles is countable (up to similarity).

03609#pg2>.

Classification of Non-Square Tilings

The main result settles Erd˝os Problem 633 by classifying all triangles that can be tiled into a non-square number of congruent triangles <ref:2604.03609#pg2>. Theorem 1 states that a triangle T admits a non-square tiling if and only if it satisfies one of eight specific conditions based on its angles (A, B, C) <ref:2604. p. 0

Where the paper says it

Our main result, which settles the problem, is the following. A triangle T admits a non-square tiling if and only if it satisfies one of the following conditions, where (A, B, C) are the angles of T in some order: (1) A = B, i.

03609#pg2>. These conditions include being isosceles, having specific right triangles with integer side ratios, or satisfying various trigonometric rationality constraints <ref:2604. p. 0
Where the paper says it

A triangle T admits a non-square tiling if and only if it satisfies one of the following conditions, where (A, B, C) are the angles of T in some order: (1) A = B, i. T is an isosceles triangle (including equilateral); (2) C = π/2 and the legs of the right triangle T are in integer ratio M/K, where M2 + K2 is not a square; (3) (A, B, C) = (π/6, π/2, π/3); (4) C = π/3, with √ 3 tan(A/2) rational; (5) B = 2A, with √ 3 tan(A/2) rational; (6) B = 2A, with sin(A/2) rational; (7) C = A/2 + B, with 2 sin(A/4) rational, equal to M/K, where 2K2 − M2 is not a square and M, K ∈ Z; Date: August 26, 2026.

03609#pg2>.

Analysis of Non-Reptile Tilings

The paper establishes conditions under which a tiling of triangle T that is not similar to T must have a non-square number of tiles <ref:2604.03609#pg3>. Theorem 3 states that if T has a non-reptile tiling, the number of tiles cannot be a square unless specific conditions are met, such as T being isosceles or satisfying certain angle relationships <ref:2604. p. 1

Where the paper says it

Then the number of tiles cannot be a square unless one of the following is true: (1) T is isosceles; (2) T has angles (A, B, C) with C = A/2 + B and 2 sin(A/4) rational, equal to M/K, where 2K2 − M2 is a square. After giving some preliminaries in Section 2, we prove the left-to-right direction of Theorem 1 in Section 3, showing that one of the given conditions must hold if we have a non-square tiling.

03609#pg3>. The proof involves applying number theory to Diophantine equations derived from the area ratio N = Area(T)/Area(R), showing that the ranks of these elliptic curves are zero, limiting solutions to torsion points <ref:2604. p. 1
Where the paper says it

These equations reduce to elliptic curves. As it turns out, the ranks of all these elliptic curves are zero, so it is enough to find their torsion points, and to check that those points do not correspond to actual tilings.

03609#pg3>.

Derivation of Necessary Conditions

The forward direction of Theorem 1 is proven by analyzing non-reptile tilings and applying theorems regarding group types (Group 1 or Group 2) for such tilings <ref:2604.03609#pg6>. Lemma 12 shows that if a non-isosceles triangle T has a non-reptile tiling, one of four angle conditions must hold: C = π/3, B = 2A, C = A/2 + B, or C = 2A + B/2 <ref:2604. p. 4

Where the paper says it

□ We say that a tiling is Group 1 or Group 2 if it is of the corresponding type in the statement of Theorem 11. If a non-isosceles triangle T has a non-reptile tiling, then one of the following four cases must hold for some permutation of the angles (A, B, C) of T: • C = π/3; • B = 2A; • C = A/2 + B; • C = 2A + B/2.

03609#pg6>. Proposition 13 then systematically relates the angles of T to the angles of the tile R based on these four cases through a complex table <ref:2604. p. 5
Where the paper says it

6 BEESON, LACZKOVICH, AND ZHANG (α, β, γ), and satisfy rationality conditions, according to the other columns of the table. Angles of T Rationality T in terms of (α, β, γ) Relations C = π/3 √ 3 tan(A/4) ∈ Q (2α, 2β, α + β) γ = 2π/3 C = π/3 √ 3 tan(A/2) ∈ Q (α, α + 2β, α + β) γ = 2π/3 B = 2A √ 3 tan(A/2) ∈ Q (α, 2α, 3β) γ = 2π/3 B = 2A sin(A/2) ∈ Q (α, 2α, 2β) 3α + 2β = π C = A/2 + B sin(A/4) ∈ Q (2α, β, α + β) 3α + 2β = π C = 2A + B/2 √ 3 tan(A/2) ∈ Q (α, 2β, 2α + β) γ = 2π/3 The proof of this proposition will be given by breaking it into the next four propositions.

03609#pg7>.

Proof of Sufficiency and Existence

The right-to-left direction is proven by exhibiting actual tilings for each case listed in Theorem 1 <ref:2604.03609#pg10>. For cases (1) through (8), the authors provide constructions, often using methods like refining dissections into similar triangles or biquadratic tilings <ref:2604.03609#pg14>. The proof for case (7) involves showing that if C = A/2 + B and 2 sin(A/4) is rational, the number of tiles N equals 2K squared - M squared, which is a square if and only if 2K squared - M squared is a square <ref:2604. p. 1

Where the paper says it

Then the number of tiles cannot be a square unless one of the following is true: (1) T is isosceles; (2) T has angles (A, B, C) with C = A/2 + B and 2 sin(A/4) rational, equal to M/K, where 2K2 − M2 is a square. After giving some preliminaries in Section 2, we prove the left-to-right direction of Theorem 1 in Section 3, showing that one of the given conditions must hold if we have a non-square tiling.

03609#pg18>.

Number Theoretic Constraints

The final part of the proof involves showing that the number of tiles N cannot be a square for the cases where T has incommensurable angles <ref:2604.03609#pg22>. This is achieved by analyzing Diophantine equations derived from the tiling parameters, such as those leading to elliptic curves whose rank is zero <ref:2604. p. 1

Where the paper says it

For the number of tiles to be a square, we would have integer solutions to Diophantine equations of the form N = Z 2 . These equations reduce to elliptic curves.

03609#pg12>. Proposition 21 demonstrates that for certain rational values of t, N cannot be a square because the corresponding elliptic curve has only the trivial rational solution (0, 0) <ref:2604.03609#pg13>. The analysis also confirms that incommensurability constraints rule out possibilities where angles are rational multiples of π <ref:2604. p. 1
Where the paper says it

We end with examples of our tilings in Section 7. Preliminaries We say that a triangle has commensurable angles if all of its angles are rational multiples of π and incommensurable angles otherwise.

03609#pg22>.

Examples and Tile Uniqueness

The paper concludes by providing examples of the tilings, such as the 77-tiling and 28-tiling, which are members of the triquadratic tilings family <ref:2604.03609#pg28>. Theorem 32 asserts that for non-isosceles triangles admitting a non-square, non-reptile tiling, the tile is uniquely determined by T unless T matches specific rows in the classification table <ref:2604. p. 1

Where the paper says it

Then the number of tiles cannot be a square unless one of the following is true: (1) T is isosceles; (2) T has angles (A, B, C) with C = A/2 + B and 2 sin(A/4) rational, equal to M/K, where 2K2 − M2 is a square. After giving some preliminaries in Section 2, we prove the left-to-right direction of Theorem 1 in Section 3, showing that one of the given conditions must hold if we have a non-square tiling.

03609#pg22>. This uniqueness is established by showing that any other potential matching would imply commensurable angles, contradicting the initial hypothesis <ref:2604. p. 9
Where the paper says it

Then b = n 2 − m2 and we may substitute a and b into c 2 = a 2 + b 2 + ab to get (7) c 2 = n 4 − m2n 2 + m4 . 638], the only positive solutions to (7) are of the form (m, n, c) = (t, t, t2 ), which would force b = 0, contradicting the hypothesis that b is positive.

03609#pg22>. The existence of tilings for cases (4) through (8) is confirmed by citing previous works and providing explicit constructions <ref:2604. p. 8
Where the paper says it

Some number theory In order to prove the “if” part of Theorem 1 we will show that (1) tilings with the parameters given in the table of Proposition 13 do exist; (2) there are rational parametrizations for the number of tiles required, per case of tiling type; (3) those rational functions do not take square values. The first two parts will be accomplished by carefully combining previous results.

p. 8
Where the paper says it

Some number theory In order to prove the “if” part of Theorem 1 we will show that (1) tilings with the parameters given in the table of Proposition 13 do exist; (2) there are rational parametrizations for the number of tiles required, per case of tiling type; (3) those rational functions do not take square values. The first two parts will be accomplished by carefully combining previous results.

03609#pg26>.

Conclusion on Tile Families

The work demonstrates that usually only one tile is possible, and at most two triangles T can match more than one row in the classification table <ref:2604.03609#pg22>. The existence of tilings for cases (4) through (8) is shown by citing explicit constructions <ref:2604. p. 4

Where the paper says it

Therefore, by Lemma 4, α, β, γ are not commensurable either, and we can apply [10], Theorem 4. That theorem has more cases in the conclusion than shown here, but does not have our hypothesis that T is not isosceles.

03609#pg26>. The final result confirms that the number of tiles N is not a square in all non-trivial cases, except when specific conditions related to the tile's geometry are met <ref:2604.03609#pg18>. This addresses Erd˝os Problem 633 by classifying triangles that can only be tiled by square tilings <ref:2604. p. 0
Where the paper says it

It is well-known that any triangle can be tiled (cut) into any square number of congruent triangles similar to the original triangle. We classify all triangles that can be tiled into a non-square number of congruent triangles, settling Erd˝os Problem 633.

03609#pg2>.

Examples of Tiling Numbers

The paper provides specific numerical examples, such as the 77-tiling with N = 77 and the 28-tiling with N = 28, which are members of the triquadratic tilings family <ref:2604.03609#pg31>. The number of tiles for these specific examples is calculated using formulas derived from Proposition 19 and Proposition 21 <ref:2604. p. 8

Where the paper says it

That completes the proof of (7). Some number theory In order to prove the “if” part of Theorem 1 we will show that (1) tilings with the parameters given in the table of Proposition 13 do exist; (2) there are rational parametrizations for the number of tiles required, per case of tiling type; (3) those rational functions do not take square values.

03609#pg16>. These examples illustrate the existence of non-square tilings for triangles whose angles are incommensurable <ref:2604. p. 1
Where the paper says it

We end with examples of our tilings in Section 7. Preliminaries We say that a triangle has commensurable angles if all of its angles are rational multiples of π and incommensurable angles otherwise.

03609#pg28>. The complexity of these tilings is noted, with some numbers exceeding 1 million tiles <ref:2604. p. 8
Where the paper says it

That completes the proof of (7). Some number theory In order to prove the “if” part of Theorem 1 we will show that (1) tilings with the parameters given in the table of Proposition 13 do exist; (2) there are rational parametrizations for the number of tiles required, per case of tiling type; (3) those rational functions do not take square values.

03609#pg19>.

General Tiling Construction Methods

The existence part relies on methods such as refining a dissection into similar triangles or using biquadratic tilings where the number of tiles N is a sum of two squares <ref:2604.03609#pg24>. These constructions allow for the creation of tilings for various triangle types, including those with commensurable angles <ref:2604. p. 2

Where the paper says it

Combining these two theorems, we obtain: Corollary 7. If T has commensurable angles, then T admits a non-square tiling if and only if one of the following is true: (1) T is isosceles (including equilateral); (2) T is a right triangle with legs in integral ratio M/K where M2 + K2 is not a square; (3) T is the (π/2, π/3, π/6) right triangle.

03609#pg19>. The paper shows that different methods can lead to different numbers of tiles, highlighting the dependence on the specific construction chosen <ref:2604. p. 0
Where the paper says it

Our work shows that, aside from isosceles triangles, this set of triangles is countable (up to similarity). We also show that (with a few exceptions) if a triangle T has any tiling into congruent triangles not similar to T, then every such tiling has a non-square number of tiles.

03609#pg26>. The existence of tilings for cases (4) through (8) is confirmed by citing previous works and providing explicit constructions <ref:2604.03609#pg19>.

Summary of Key Findings

The classification shows that aside from isosceles triangles, the set of triangles that can be tiled into a non-square number of congruent triangles is countable up to similarity <ref:2604.03609#pg22>. The paper proves Theorem 1 by showing necessity through analysis of tiling groups and sufficiency through explicit constructions <ref:2604. p. 1

Where the paper says it

Then the number of tiles cannot be a square unless one of the following is true: (1) T is isosceles; (2) T has angles (A, B, C) with C = A/2 + B and 2 sin(A/4) rational, equal to M/K, where 2K2 − M2 is a square. After giving some preliminaries in Section 2, we prove the left-to-right direction of Theorem 1 in Section 3, showing that one of the given conditions must hold if we have a non-square tiling.

p. 1
Where the paper says it

Then the number of tiles cannot be a square unless one of the following is true: (1) T is isosceles; (2) T has angles (A, B, C) with C = A/2 + B and 2 sin(A/4) rational, equal to M/K, where 2K2 − M2 is a square. After giving some preliminaries in Section 2, we prove the left-to-right direction of Theorem 1 in Section 3, showing that one of the given conditions must hold if we have a non-square tiling.

03609#pg10>. The core finding is that for most triangles, every square tiling must be a reptiling <ref:2604. p. 1
Where the paper says it

Apart from the isosceles triangles, the number of similarity classes of triangles having a non-square tiling is countable. We also prove that for most triangles every square tiling must be a reptiling.

p. 1
Where the paper says it

Apart from the isosceles triangles, the number of similarity classes of triangles having a non-square tiling is countable. We also prove that for most triangles every square tiling must be a reptiling.

03609#pg22>. This settles Erd˝os Problem 633 by characterizing the exceptions to generic square tilings <ref:2604. p. 0
Where the paper says it

Every triangle has “generic” square reptilings into N 2 tiles for any N (see Figure 1), so the following ($25) Erd˝os Problem from [21, p. 48] (indexed as Problem 633 in [5]) can be seen as trying to understand which triangles T can be “non-generically” tiled.

p. 0
Where the paper says it

Every triangle has “generic” square reptilings into N 2 tiles for any N (see Figure 1), so the following ($25) Erd˝os Problem from [21, p. 48] (indexed as Problem 633 in [5]) can be seen as trying to understand which triangles T can be “non-generically” tiled.

03609#pg2>. The classification is based on eight conditions derived from geometric and number-theoretic constraints <ref:2604.03609#pg2>.

Final Summary Statement

Theorem 1 states that a triangle T admits a non-square tiling if and only if it satisfies one of the eight listed conditions, which includes being isosceles or meeting specific trigonometric rationality requirements <ref:2604.03609#pg10>. This comprehensive classification settles Erd˝os Problem 633 by characterizing the exceptions to generic square tilings <ref:2604. p. 0

Where the paper says it

Every triangle has “generic” square reptilings into N 2 tiles for any N (see Figure 1), so the following ($25) Erd˝os Problem from [21, p. 48] (indexed as Problem 633 in [5]) can be seen as trying to understand which triangles T can be “non-generically” tiled.

03609#pg2>. The paper proves that for most triangles every square tiling must be a reptiling <ref:2604. p. 1
Where the paper says it

Apart from the isosceles triangles, the number of similarity classes of triangles having a non-square tiling is countable. We also prove that for most triangles every square tiling must be a reptiling.

03609#pg22>.

The classification shows that aside from isosceles triangles, the set of triangles that can be tiled into a non-square number of congruent triangles is countable up to similarity <ref:2604.03609#pg22>. The paper proves Theorem 1 by showing necessity through analysis of tiling groups and sufficiency through explicit constructions <ref:2604.03609#pg10>. The core finding is that for most triangles, every square tiling must be a reptiling <ref:2604.03609#pg22>. This settles Erd˝os Problem 633 by characterizing the exceptions to generic square tilings <ref:2604.03609#pg2>. The classification is based on eight conditions derived from geometric and number-theoretic constraints <ref:2604.03609#pg2>.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Automated classification of triangles for non-square tilings can be performed by implementing a decision tree based on Theorem 1, allowing rapid determination of which conditions (1) through (8) are satisfied for any given triangle's angles.

  2. The system can check if a tiling is a reptiling or not using the criteria derived from Theorem 5, effectively filtering out non-square tilings that are also reptilings based on the number of tiles.

  3. A module can be developed to analyze Diophantine equations arising from the number of tiles, specifically checking if they reduce to elliptic curves with zero rank to confirm whether a non-square tiling is possible for a given configuration.

  4. The system can verify the rationality conditions required for certain triangle types, such as checking if sin(A/2) or sin(A/4) are rational when the angles fall under specific cases like (5), (6), (7), or (8).

  5. The system can use a database of elliptic curves and their ranks to computationally prove that the rank is zero for the Diophantine equations associated with each tiling case, providing rigorous proof for the non-square nature of N.

  6. The system can be trained to identify which rows in Proposition 13 a given triangle matches, enabling it to suggest potential tile shapes and tilings based on pre-calculated results like those in Proposition 31 or Theorem 32.

Abstract

We classify triangles that can be tiled only into a square number of congruent triangles, settling Erdős Problem 633.

Sources

Related papers