Diophantine Tuples: 2000 Years in Search of the Fifth Element
At first glance, the set of four numbers {1, 3, 8, 120} seems unremarkable. But it turns out to conceal a problem two thousand years in the making — one that connects Diophantus, Fermat, Euler, and modern mathematicians, and touches on the theory of transcendental numbers, Pell equations, Fibonacci numbers, and elliptic curves. Let’s take it from the beginning.
How Diophantine Tuples Were Born: The Problem from the Third Century
It all started in the third century AD.
Diophantus was looking for a set of numbers {a₁, a₂, a₃, a₄} such that for any pair the following holds:
aᵢ · aⱼ + 1 = a perfect square
In other words, the product of any two numbers in the set, plus one, must be a perfect square. Such a set became known as a Diophantine quadruple, or D-quadruple. Diophantus himself found a solution in the form of rational fractions:
{1/16, 33/16, 17/4, 105/16}
Twelve centuries later, Fermat produced the first integer example — those very numbers {1, 3, 8, 120}. But that immediately raised a new question: just how many such sets of four numbers are there?
Euler’s Answer: Infinitely Many Diophantine Quadruples
The great Euler provided the answer. It turns out there are infinitely many such quadruples. Moreover, Euler gave a formula for constructing them.
If you have a pair {a, b} with ab + 1 = r², you can build a quadruple as follows:
{a, b, a + b + 2r, 4r(a + r)(b + r)}
Having settled the question of quadruples, a new question arose immediately: does a Diophantine quintuple exist? The answer came even later than the proof of Fermat’s Last Theorem.
Baker, Davenport, and the First Result on Diophantine Tuples
In 1969, Baker and Davenport found the answer for Fermat’s specific set: no fifth element exists. The triple {1, 3, 8} can be extended in only one way — to 120, and that’s it. The proof relied on Baker’s results from the theory of transcendental numbers. But Fermat’s quadruple is just one set out of infinitely many.
In 2004, Dujella pushed the answer further. He proved that Diophantine sextuples do not exist, and that there are only finitely many quintuples. That was already a major step forward: in principle, one could simply check all remaining candidates by computer. But for mathematicians, that wasn’t enough.
The Final Theorem: No Integer Diophantine Quintuple Exists
In 2016, He, Togbé, and Ziegler put the definitive end to the problem: no integer Diophantine quintuple exists at all. The problem reduces to a system of Pell equations of the form a · e + 1 = x², which simply has no solutions — the system is inconsistent. A fifth element cannot be added.
Open Questions in the Theory of Diophantine Tuples
But as so often happens in number theory, many questions remain open:
The regularity conjecture: is it true that every quadruple must satisfy the condition (a + b − c − d)² = 4(ab + 1)(cd + 1)?
How to extend Diophantus’s original rational quadruple to five elements (rational numbers turn out to be far more “flexible” than integers).
How quadruples behave for special values of n — that is, when you add +n instead of +1 to require a perfect square.
Whether a rational septuple exists (it turns out there are infinitely many rational sextuples).
Whether there exists a rational quadruple that generates an elliptic curve of rank greater than 10 (it turns out every quadruple can be associated with an elliptic curve).
Why Number Theory Never Runs Out of Surprises
This is exactly what makes number theory so captivating. An innocent-seeming question immediately leads into a new world — a world of numbers with their own structures and astonishing connections. A world we will be exploring for decades, centuries, and most likely thousands of years to come.
