Proving Legendre's Relation Holds for the Weierstrass Zeta Function

548 Views Asked by At

The Weierstrass Zeta Function $\zeta$ is a meromorphic function satisfying

  • $\zeta(z+\omega_1) = \zeta(z) + \eta_1$
  • $\zeta(z+\omega_2) = \zeta(z) + \eta_2$
  • The singularities of $\zeta$ are poles of residue 1 at the points $m\omega_1+n\omega_2$ for $m,n \in \mathbb{Z}$

Here $\omega_1, \omega_2, \eta_1, \eta_2$ are complex constants with $\frac{\omega_2}{\omega_1}$ not real. Use Cauchy's Residue Theorem to prove Legendre's relation $\omega_2\eta_1 - \omega_1\eta_2 = \pm 2\pi i$ and express the sign in terms of $\omega_1$ and $\omega_2$.

Take a parallelogram with vertex at the origin with sides $\omega_1$ and $\omega_2$ and shift it appropriately so the center is 0. I get that Cauchy's Residue Theorem gives us that the integral around the parallelogram is $2\pi i$.

Here's where I get lost. Take the case $\Im(\frac{\omega_2}{\omega_1}) > 0$. Then integrate along the sides. How is this done? Parametrize somehow? Don't seem to be able to understand the abstraction here.

1

There are 1 best solutions below

2
On

By the pseudo periodicity of the Weierstrass zeta functions, the integrals of $\zeta(z)\,dz$ over two opposite sides of the parallelogram almost cancel, but don't.

Remember that $\zeta(z+\omega_j)=\zeta(z)+\eta_j$.