Legendre's Three-Square Theorem

In the vast landscape of human knowledge, pure mathematics and physical science often seem to occupy separate realms—one of abstract logic, the other of tangible reality. Yet, powerful and unexpected connections sometimes emerge, revealing a deep unity in the fabric of the universe. Legendre's three-square theorem is a prime example of such a link, a simple rule from number theory that has profound and surprising consequences for the physical world. The theorem addresses a seemingly simple question: which whole numbers can be written as the sum of three squared integers? The answer, as we will see, is a precise and elegant rule that creates a family of "forbidden numbers."

This article bridges the gap between abstract arithmetic and concrete physics by demonstrating how this mathematical constraint is not just a curiosity but a fundamental law that nature itself appears to obey. We will explore how a theorem formulated over two centuries ago dictates which energy levels are possible inside a quantum system, which structures can form in a crystal, and even what patterns we might observe on a cosmic scale.

This journey begins by exploring the theorem's core principles and mechanisms, uncovering the simple but powerful logic that governs the sums of three squares. From there, we will broaden our view to see how this abstract rule applies across a surprising range of applications and interdisciplinary connections, revealing a hidden mathematical blueprint for the physical world.

Imagine you are standing in an infinitely large, perfectly dark room. You have a special flashlight that can only create a single point of light at any location $(x,y,z)$ where the coordinates are all whole numbers — integers. Think of it as a vast, three-dimensional grid, a crystal lattice of possible light points stretching to infinity. Now, you switch on a machine that illuminates not just one point, but all the possible integer points that lie precisely on the surface of a sphere centered on you at the origin $(0,0,0)$.

As you change the radius $R$ of the sphere, you'd expect a fascinating, changing constellation of lights to flicker into existence. A tiny sphere might catch no points. A slightly larger one might catch six points, at $(\pm 1, 0, 0)$, $(0, \pm 1, 0)$, and $(0, 0, \pm 1)$. Increase the radius a bit more, and you might get twelve points, like $(\pm 1, \pm 1, 0)$ and its permutations. The core question, a beautiful puzzle of geometry and numbers, is this: for any given squared radius, $n = R^2$, can we always find some integer points on the sphere's surface? In other words, does the equation

always have a solution for any positive integer $n$?

The answer, astonishingly, is no. There's a hidden law, a cosmic selection rule, that forbids certain spheres from ever intersecting our integer grid, no matter how you orient them.

The rule was first discovered by the great French mathematician Adrien-Marie Legendre. It's as precise as it is mysterious. ​​Legendre's three-square theorem​​ states that a positive integer $n$ can be written as the sum of three integer squares if and only if it is ​​not​​ of the form:

where $a$ and $b$ are any non-negative integers ($0, 1, 2, \dots$).

What does this strange formula mean? Let's unpack it. The dangerous part is the $8b+7$ term. This just means any number that leaves a remainder of 7 when you divide it by 8. So, the numbers $7, 15, 23, 31, 39, \dots$ are all in this forbidden family from the start (this is the case $a=0$). The $4^a$ part adds another layer. It means you also have to exclude numbers that, after being divided by 4 as many times as possible, leave you with a number of the form $8b+7$. For instance, $28$ is forbidden because $28 = 4 \times 7$. And $112$ is forbidden because $112 = 4 \times 28 = 4^2 \times 7$. These are the "forbidden numbers." No sphere whose squared radius is one of these numbers will ever contain a single point from our integer lattice.

Why this peculiar rule involving 4s and 8s? It's not magic; it's arithmetic, and we can uncover the secret by playing a simple game with remainders—a game we call ​​modular arithmetic​​.

Let's look at any integer's square and see what its remainder is when we divide by 8.

• $0^2=0 \equiv 0 \pmod{8}$
• $1^2=1 \equiv 1 \pmod{8}$
• $2^2=4 \equiv 4 \pmod{8}$
• $3^2=9 \equiv 1 \pmod{8}$
• $4^2=16 \equiv 0 \pmod{8}$
• an even number $(2k)^2 = 4k^2$, which is $0 \pmod 8$ if $k$ is even, and $4 \pmod 8$ if $k$ is odd.
• an odd number $(2k+1)^2 = 4k^2+4k+1 = 4k(k+1)+1$. Since either $k$ or $k+1$ must be even, $k(k+1)$ is always even, so $4k(k+1)$ is a multiple of 8. Thus, any odd square is $1 \pmod 8$.

Notice a pattern? The square of any integer, when divided by 8, can only leave a remainder of ​​0, 1, or 4​​. That's it. Nothing else is possible.

Now, what about the sum of three squares, $x^2 + y^2 + z^2$? The remainder of this sum modulo 8 can only be the sum of three numbers chosen from the set $\{0, 1, 4\}$. Let's try to get a sum of 7:

• $4+1+1 = 6$
• $4+4+0 = 8 \equiv 0$
• $4+4+1 = 9 \equiv 1$
• $1+1+1 = 3$
• $4+4+4 = 12 \equiv 4$

No matter how you combine them, you can never get a sum of 7! This simple observation reveals the heart of the matter: if a number $n$ is of the form $8b+7$, it cannot possibly be the sum of three squares because the two sides of the equation $x^2+y^2+z^2 = n$ would have different remainders when divided by 8. This is called a ​​congruence obstruction​​. It’s a complete roadblock.

The $4^a$ part of the rule comes from a clever "descent." If $x^2+y^2+z^2$ is a multiple of 4, it turns out that $x, y,$ and $z$ must all be even. So, we can write $x=2x'$, $y=2y'$, $z=2z'$, and our equation becomes $4x'^2+4y'^2+4z'^2=n$. Dividing by 4 gives $x'^2+y'^2+z'^2 = n/4$. This means if $n$ is a sum of three squares, then so is $n/4$. We can keep dividing by 4 until we get a number that isn't divisible by 4. The theorem tells us that it is this "core" number that must pass the "not 7 modulo 8" test.

You might think this is just a curious bit of number theory. But nature, it seems, has a deep appreciation for it. The rules governing the universe at its smallest scales—in ​​quantum mechanics​​—play by these very same numerical laws.

Consider one of the first systems every physicist studies: a particle in a box. Imagine an electron trapped inside a perfect cube. Quantum mechanics tells us that the electron cannot have just any energy. Its energy is ​​quantized​​; it can only exist in specific, discrete energy levels. For a cubic box of side length $L$, these allowed energies are given by a simple formula:

Here, $h$ is Planck's constant, $m$ is the electron's mass, and the numbers $(n_x, n_y, n_z)$ are the ​​quantum numbers​​. They describe the state of the electron wave in each of the three dimensions.

Now comes the crucial part. The physical constraints of the box determine what kind of numbers $n_x, n_y, n_z$ can be.

• In a standard "hard-wall" box (known as an infinite potential well), the electron's wave must vanish at the walls. This forces the quantum numbers to be ​​positive integers​​: $n_i \in \{1, 2, 3, \dots\}$.
• In a crystal lattice, a more realistic model often uses ​​periodic boundary conditions​​, where the electron behaves as if the box is seamlessly repeated throughout space. In this case, the quantum numbers can be ​​any integers​​: $n_i \in \{\dots, -2, -1, 0, 1, 2, \dots\}$.

In both cases, up to a constant factor, the energy is determined by the "energy index" $N = n_x^2+n_y^2+n_z^2$. Suddenly, our purely mathematical question becomes a question of physics: ​​Which energy levels are possible?​​

Legendre's theorem gives an immediate and profound answer. For any type of box, an energy level corresponding to an index $N$ of the form $4^a(8b+7)$ is ​​absolutely forbidden​​. Nature simply cannot produce a particle with that energy in a 3D box, because no three integer squares can sum to it.

But the constraints matter! In the hard-wall box where quantum numbers must be positive, the rules are even stricter. Take $N=5$. Legendre's theorem says 5 is fine, because $5 = 2^2 + 1^2 + 0^2$. But in the hard-wall box, a quantum number of 0 is not allowed! And you can't find three positive integers whose squares sum to 5. So, the energy level $N=5$ is a "ghost": mathematically allowed for general integers, but physically impossible under these specific boundary conditions. This is a beautiful lesson: the underlying mathematical structure provides a landscape of possibilities, but the specific physical reality carves out its own allowed territory within that landscape.

The quantum connection gets even richer when we consider ​​degeneracy​​—the situation where multiple distinct quantum states have the exact same energy. In our cubic box, the energy only depends on the sum $n_x^2+n_y^2+n_z^2$, not the individual values.

Because the box is a cube, the state with quantum numbers $(1, 2, 3)$ has the exact same energy as the states $(3, 2, 1)$, $(2, 1, 3)$, and so on. They are different states (the wave is oriented differently in space), but they are energetically degenerate. The number of such distinct permutations gives the degeneracy:

• If all three quantum numbers are different, like $\{a,b,c\}$, there are $3! = 6$ distinct states.
• If two are the same, like $\{a,a,b\}$, there are $3!/2! = 3$ distinct states.
• If all three are the same, like $\{a,a,a\}$, there is only 1 state.

Sometimes, the richness of the number system provides even more degeneracy. Consider the energy level $N=110$. A quick search reveals that it can be formed in multiple, entirely different ways:

• $10^2 + 3^2 + 1^2 = 100 + 9 + 1 = 110$
• $7^2 + 6^2 + 5^2 = 49 + 36 + 25 = 110$

Each of these unordered sets of numbers corresponds to a distinct family of physical states. Since all the numbers in each set are different, the first set provides 6 distinct quantum states, and the second set provides another 6. So, the energy level $N=110$ has a total degeneracy of at least 12. The number of allowed states for a given energy is a direct reflection of the number-theoretic properties of its energy index $N$.

For a long time, Legendre's rule $n \ne 4^a(8b+7)$ might have seemed like a clever but isolated trick. Modern mathematics, however, reveals it as a manifestation of a much deeper and more elegant idea: the ​​Hasse-Minkowski local-global principle​​.

The principle, in essence, states that for certain types of equations (including ours), a solution exists using rational numbers if and only if a solution exists in every possible number system we have. What does "every number system" mean? It means the familiar ​​real numbers​​ ($\mathbb{R}$) and a family of less familiar but equally important systems called the ​​p-adic numbers​​ ($\mathbb{Q}_p$), one for each prime number $p=2, 3, 5, \dots$. You can think of p-adic numbers as an alternative way of measuring "closeness"—in $\mathbb{Q}_7$, for example, the number 49 is "closer" to 0 than 7 is.

So, to ask if $x^2+y^2+z^2=n$ has a rational solution, we can ask it locally, in each of these number systems:

• ​​In the real numbers $\mathbb{R}$:​​ Is there a solution? Yes, as long as $n \ge 0$. This is a trivial condition.
• ​​In the p-adic numbers $\mathbb{Q}_p$ for odd primes $p$​​: It turns out there is always a solution. No obstruction here.
• ​​In the 2-adic numbers $\mathbb{Q}_2$​​: This is the only place, besides the reals, where a problem can arise. And what is the condition for a solution to exist in $\mathbb{Q}_2$? You guessed it. A solution exists if and only if $n$ is not of the form $4^a(8b+7)$.

This is a profound revelation. Legendre's seemingly arbitrary rule is not arbitrary at all. It is the precise condition required for the equation to be solvable in the 2-adic world. The local-global principle tells us that this single "local" obstruction is so powerful that it determines the fate of the equation globally, in the world of rational numbers and integers.

From a pattern of lights on a sphere, to the quantization of energy-levels in a box, to a deep principle unifying all number systems—the story of three squares is a perfect example of what makes science so beautiful. It shows how a simple question can lead us to discover hidden rules, and how those rules, in turn, reveal a stunning, interconnected mathematical structure that resonates from pure number theory to the very fabric of the physical world.

Having journeyed through the elegant logic of Legendre's three-square theorem, we now arrive at a viewpoint from which we can appreciate its true power. A theorem of pure mathematics, born from the abstract world of integers, might seem a distant curiosity. Yet, as we are about to see, this simple rule about which numbers can be formed by summing three squares is, quite astonishingly, a rule that nature herself follows. It is a subtle but profound constraint woven into the fabric of the physical world, its signature appearing in the structure of matter, the laws of quantum mechanics, the behavior of fields, and even in our theories about the cosmos. In this chapter, we will explore these remarkable connections, seeing how one beautiful idea can illuminate a vast and diverse landscape of scientific inquiry.

Our first stop is the world of solid-state physics and chemistry, the study of the orderly arrangement of atoms in crystals. A perfect crystal is a marvel of regularity, a repeating grid of atoms stretching out in three dimensions. Physicists and chemists probe this hidden architecture using tools like X-ray diffraction. When a beam of X-rays strikes a crystal, it scatters off the planes of atoms, creating a characteristic pattern of bright spots or rings on a detector. Bragg's law tells us that the angle of each bright ring is related to the spacing, $d$, between the atomic planes that produced it.

For the simplest case, a simple cubic crystal, the spacing between a set of planes identified by integer indices $(h, k, l)$ is given by a wonderfully straightforward formula: $d_{hkl} = \frac{a}{\sqrt{h^2+k^2+l^2}}$, where $a$ is the lattice constant, the side length of the cubic unit cell. Notice the sum of three squares! This immediately implies that the possible diffraction patterns are not arbitrary. The spectrum of observable plane spacings is directly governed by number theory.

Imagine conducting such an experiment. You would find rings corresponding to $h^2+k^2+l^2=1$, $2$, $3$, $4$, $5$, $6$... but then, inexplicably, you would find no ring for $7$. And no ring for $15$. Nor for $23$. These are not experimental errors or imperfections in the crystal. They are fundamental, systematic absences dictated by Legendre's theorem. The numbers $7$, $15$, and $23$ are of the form $8m+7$, and therefore can never be the sum of three integer squares. Nature is forbidden from creating crystal planes whose spacing corresponds to these numbers. This number-theoretic rule manifests as literal gaps in a physical measurement.

This same principle appears when we consider the crystal's "reciprocal lattice," a concept fundamental to understanding waves (like electrons or phonons) traveling through the crystal. The reciprocal lattice is essentially a map in momentum space, and its points are defined by vectors whose squared lengths are proportional to $h^2+k^2+l^2$. When we count the allowed points in this space, we find that there are spherical shells on which no reciprocal lattice point can ever exist. It is as if the architect of the crystal lattice was given a blueprint that included the footnote: "No structures corresponding to integers of the form $4^k(8m+7)$."

The connection deepens as we move from the classical picture of a crystal to the quantum world. One of the first systems a student of quantum mechanics encounters is the "particle in a box"—a single particle confined within a three-dimensional cubic volume. The Schrödinger equation dictates that the particle can only exist in states with specific, quantized energy levels. For a cubic box of side length $L$, these allowed energies are given by:

$E = \frac{h^2}{8mL^2} (n_x^2 + n_y^2 + n_z^2)$

where $n_x, n_y, n_z$ are positive integers. Once again, we find the sum of three squares at the heart of a fundamental physical law.

This has a startling consequence: the energy spectrum of this simple quantum system is riddled with gaps defined by Legendre's theorem. A particle in a cubic box can have an energy proportional to $6$ (from state $(2,1,1)$) and $9$ (from $(2,2,1)$), but it can never have an energy proportional to $7$. The availability of a level like $8 = 2^2+2^2+0^2$ depends on the boundary conditions: it is allowed with periodic boundaries where integer quantum numbers can be zero, but forbidden in a hard-wall box where they must be positive. There is a "silent note" in the quantum harmonics of the box.

This is not a minor curiosity. It's a key feature that distinguishes highly symmetric, "integrable" systems like the cubic box from "chaotic" ones. If you were to slightly deform the box, making its sides incommensurate, these neat degeneracies and number-theoretic gaps would dissolve into a seemingly random mess of energy levels. The gaps are a fingerprint of the cube's perfect symmetry.

We can even ask a statistical question: in the limit of very high energies, what fraction of all possible integer energy levels are forbidden? The answer, derived by summing the densities of all forbidden numbers of the form $4^a(8b+7)$, is a stunningly simple and elegant $1/6$. One-sixth of the energy landscape is simply off-limits, a direct consequence of a 200-year-old theorem.

This number-theoretic discreteness has profound implications for statistical mechanics. The classical behavior of a gas in a box, for instance, emerges at high temperatures by averaging over these jagged quantum energy levels. The partition function, a central quantity in thermodynamics, is a sum over all states, weighted by their energy. This sum can be written in terms of the degeneracy $g(N)$, the number of ways an integer $N$ can be written as a sum of three squares. This function $g(N)$ is a wild, arithmetic beast—it's zero for forbidden $N$, and it fluctuates erratically for allowed $N$. To recover the smooth, predictable laws of classical thermodynamics, one must approximate this sum with an integral, effectively smoothing over the quantum fluctuations and arithmetic details. The classical world is, in a sense, a blurred-out version of a sharp, number-theoretic quantum reality.

The theorem's reach extends beyond discrete particles to the continuous world of fields. Consider a "Beltrami field," a vector field that is everywhere parallel to its own curl: $\nabla \times \mathbf{X} = \lambda \mathbf{X}$. These fields are not just mathematical toys; they describe stable, self-sustaining structures in fluid dynamics (like long-lived vortices) and are crucial in plasma physics and astrophysics for modeling force-free magnetic fields.

If we study such fields within a contained space, like a 3-torus (a cube with periodic boundaries, a common model for plasmas or turbulent fluids), we find that only certain values of the proportionality factor $\lambda$ are allowed. These are the eigenvalues of the curl operator. A careful analysis reveals that the possible squared values of these eigenvalues, when properly scaled, must be integers that are sums of three squares, $| \mathbf{n} |^2 = n_x^2 + n_y^2 + n_z^2$.

The implication is immediate and striking: a vortex or a force-free magnetic field on a torus cannot possess just any amount of "twist" or "helicity." The spectrum of the curl operator—a fundamental object in differential geometry and physics—has gaps. The smallest integer value that the squared eigenvalue term can never take is $7$. The very geometry of fields, when confined, is constrained by the same arithmetic rule we found in crystals and quantum boxes.

For our final and perhaps most mind-bending application, we turn to cosmology. While our universe appears flat on the largest scales, its global topology remains an open question. It might be simply infinite, or it could be finite but unbounded, like the surface of a sphere. One intriguing possibility considered by cosmologists is that our universe has the topology of a giant 3-torus. If this were true, traveling in a straight line for long enough would bring you right back to where you started.

In such a universe, we could see "ghost images" of our own Milky Way galaxy. Light from our galaxy could travel across the universe, wrap around the torus, and return to us from a different direction, appearing as a distant replica of ourselves. These ghost images would be located at positions on a vast cosmic grid, with comoving coordinates $(n_x L, n_y L, n_z L)$, where $L$ is the side length of the universe's toroidal cell. The distance to such a ghost would be $d = L \sqrt{n_x^2+n_y^2+n_z^2}$.

Of course, we can only see objects that lie within our "particle horizon," the maximum distance light could have traveled since the Big Bang. This sets a spherical boundary on our observable universe. The grand question then becomes: what is the pattern of observable ghost galaxies? The answer would again be constrained by the sums of three squares. There would be spherical shells of specific radii within our observable universe where a ghost image could never appear, simply because the required squared distance corresponds to a forbidden integer like $7$ or $15$. While the 3-torus model is purely hypothetical, it provides a breathtaking illustration of how a deep truth of number theory could be inscribed on the largest possible canvas—the very structure of our cosmos.

From the atomic precision of a crystal, to the discrete energies of a quantum particle, to the stable forms of a swirling vortex, and to the hypothetical pattern of cosmic mirages, Legendre's three-square theorem has emerged again and again. It is a unifying thread, a single piece of pure logic that helps explain the structure we find in the world at every scale. This is the profound beauty that science and mathematics seek: not just a collection of facts, but the discovery of simple, elegant principles that reveal the deep and unexpected unity of the universe.