I am trying to show that there are infinitely many solutions to the following diophantine equation:

$$x^2 - 3y^2 = 1$$

But I don't really know where to start. I hear there are numerical ways to solve it, but it isn't the right direction. I also tried to use Minkowski's convex body theorem, but I can't use is for many disjoint sets since they have to be symmetric w.r. to 0

I'd appreciate any ideas

  • 1
    $\begingroup$ This is the Pell's equation,this might help www-personal.umich.edu/~hlm/math475/pell.pdf $\endgroup$ – kingW3 Dec 10 '14 at 20:48
  • 2
    $\begingroup$ There exist infinitely many units in $\mathbb{Z}[\sqrt{3}]$. Take all the powers of $2+\sqrt{3}$. $\endgroup$ – Crostul Dec 10 '14 at 20:48
  • $\begingroup$ thanks, I actually aimed to prove the existence of infinitely many units in this ring. But now I know both algebraic and number-theory solutions $\endgroup$ – Jytug Dec 10 '14 at 21:10

Assume $x^2-3y^2=1$ for some integers x and y, then:

$$(2x + 3y)^2 - 3(x+2y)^2 = 4x^2 + 6xy + 9y^2 - 3x^2 - 6xy -12y^2=x^2 - 3y^2 = 1$$

And thus $(2x+3y, x+2y)$ is a new solution.

Since $(1, 0)$ is an integer solution and the above gives a new integer solution with larger x and y for each solution you find, there are infinitely many solutions.

A handy tool (which I also used to find the above) is Dario Alpern's generic two integer variable equation solver.

  • $\begingroup$ This corresponds, in sardoj's answer, to the observation that if $x+y\sqrt 3$ has norm $1$, then $(x+y\sqrt 3)(2+\sqrt 3)$ also has norm $1$. $\endgroup$ – Mariano Suárez-Álvarez Dec 10 '14 at 21:14

Recall that in the ring $\mathbb{Z}[\sqrt{d}]$ where $d\in\mathbb{Z}$ is square-free, the norm of an element $x$ is given by $N(x)=x\bar{x}$ where $\bar{x}$ is the conjugate of $x$.

If we factor $x^2-3y^2=1$ we obtain


This is essentially showing us that the norm of $x+\sqrt{3}y$ is 1 (a unit). Also recall that the norm is multiplicative. That is, $$N(ab)=N(a)N(b).$$ Thus, if we can find one set of solutions to $x^2-3y^2=1$, say, $(x_0,y_0)$, then the pairs we get from computing $(x_0+\sqrt{3}y_0)^n, n\in\mathbb{N}$ will also be solutions. By inspection we can see that $(2,1)$ is a solution. Therefore, the pairs we get from $$(2+\sqrt{3})^n$$ will also be solutions. This shows that there are infinitely many solutions.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.