Recall that, for a complex number $\tau$ with positive imaginary part, the $j$-invariant is given by $j(\tau)=1728 \frac{g_2(\tau)^3}{g_2(\tau)^3-27g_3(\tau)^2}$ where

$g_2(\tau)=60 \sum_{(m,n)\neq(0,0)}(m+n\tau)^{-4}$

and $g_3$ is unimportant. Clearly the zeros of $j$ are the same as the zeros of $g_2$ (and, if we are only considering the zeros, we can forget about the 60 and just consider the sum.)

Is there a way to directly see that $g_2(\rho)=0$, where $\rho=exp(\frac{2}{3} \pi i)$? (e.g. is there an algebraic trick to see that the terms in the sum cancel each other out?)

If not is there a reasonably simple indirect proof?


1 Answer 1


I have since worked out the answer to this. So in case anybody comes across the question, I will sketch it below.

Observe $\rho$ has the minimal polynomial $x^2+x+1=0$, so if $\Lambda$ is the lattice $\mathbb{Z}+\rho\mathbb{Z}$, we have $\rho \Lambda = \rho\mathbb{Z}+\rho^2\mathbb{Z}=\rho\mathbb{Z}-\mathbb{Z}-\rho\mathbb{Z}=\Lambda$.

But clearly $g_2(a z)=a^4 g_2(z)$, so $g_2(\rho)=\rho^4 g_2(\rho)$, which gives $g_2(\rho)=0$.

Of course, this doesn't tell us that these are all the zeros (upto $SL_2(\mathbb{Z})$-invariance), but it does at least find the zero.

  • $\begingroup$ Heh, $-(1+\rho)\Bbb Z$ isn't the same as $-\Bbb Z-\rho\Bbb Z$ but your equality still holds. :-) $\endgroup$
    – whacka
    Apr 14, 2015 at 0:17

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.