I am trying to solve the following quadratic diophantine equation in $\mathbb C[T]$, but I did not manage. I hope someone could give some hints or solutions to my problem. Here is the equation

$$\left[(t+1)X+tY-1\right]^2=24XY,$$ where the unknowns are $X,Y$. If $\deg(X)>\deg(Y)$ this is impossible. Same for $\deg(X)<\deg(Y)$ Hence, $\deg(X)=\deg(Y)$. So now I am stuck here.

Thanks in advance for any help.

  • 3
    $\begingroup$ Here is a hint. There is a solution where both $X$ and $Y$ are constant in $t$. You can use that solution to find many other solutions in the usual way. $\endgroup$ Commented Jul 19, 2015 at 13:09

1 Answer 1


Clearly $X$, $Y$ are coprime, hence both are squares of polynomials. Write $X=3f^2$, $y=2g^2$, then $3(t+1)f^2+2tg^2-1=\pm 12 fg$, without loss of generality sign is plus, else change sign of $g$. This rewrites as $$(3(t+1)f-6g)^2+6(t^2+t-6)g^2=3(t+1).$$ Denote $3(t+1)f-6g=h$, then $h^2+6(t^2+t-6)g^2=3(t+1)$. This is Pell-type equation $h^2-Ng^2=3(t+1)$, where $N=6(t^2+t-6)=6(t-2)(t+3)$. There is a solution $(h_0,g_0)$ of $h_0^2-Ng_0^2=1$ with linear $h_0$ and constant $g_0$. Namely, $h_0+1=2(t+3)/5$, $h_0-1=2(t-2)/5$, $g_0=2/5\sqrt{6}$. Then if $(h,g)$ is a solution of $h^2-Ng^2=3(t+1)$ so are $(hh_0\pm Ngg_0,hg_0\pm gh_0)$. Without loss of generality $\deg h=\deg g+1$, $h=ct^{d+1}+\dots$, $g=\sqrt{6}ct^d+\dots$. Then I claim that if $d>0$ then $(hh_0-Ngg_0,hg_0-gh_0)$ is a solution with lower value of degree. This may be checked by direct computation of two leading coefficients or by studying asymptotics of $h\pm \sqrt{N} g$ for large values of $t$. That is, any solution reduces to case $d=0$, corresponding to constant $f$, $g$.

  • $\begingroup$ Thanks you for this very precise answer. It is very clear and smart !! $\endgroup$
    – joaopa
    Commented Jul 19, 2015 at 20:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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