2
$\begingroup$

Consider the Jacobi elliptic function $sn(\cdot,k)$ restricted to the interval $(0,2K)$, where $K=K(k)$ is complete elliptic integral of the first kind. If $0<k<1$, then it is well known the this function first increases on $(0,K)$, from 0 to 1, then it achieves its global maximum at $u=K$, and then decreases again from 1 to 0. The graph of $sn(\cdot,k)$ on $(0,2K)$ is similar to the graph of sine on $(0,\pi)$.

There is a strong numerical evidence that this properties holds true also for complex $k$ taking the absolute value of $sn(\cdot,k)$. More precisely, my question is as follows: Let $|k|\leq 1$, $k\neq\pm1$, is it true that the function $$u\mapsto |sn(u,k)|,$$ restricted on $(0,2K)$, achieves its global maximum at $u=K$?

Remarks:

  1. I can prove it for purely imaginary $k$ using Jacobi imaginary transformation for $sn(u,ik)$.

  2. Since it holds that $sn(2K-u,k)=sn(u,k)$, it would be sufficient to show the function in question in increasing on $(0,K)$.

  3. Note that $K$ is complex in general, so the meaning of $(0,2K)$ is just the line connecting points $0$ and $2K$ in $\mathbb{C}$.

  4. On the picture, you can see the plot of $|sn(Kx,k)|$, with $k=0.7+0.6i$ and $x\in(0,2)$. enter image description here

At the first glance, it seems that it should not be hard to answer this (very concrete) question. However, I was not able to do that trying to do so for some time. Thank you for your time!

$\endgroup$

1 Answer 1

1
$\begingroup$

It is not true.

Here is part of the plot (in Maple) of $|sn(tK,k)|$ in the case $k = 1.25$. A similar picture will occur for complex $k$ near $1.25$.

enter image description here

Note: Maple and Mathematica use different conventions for the second parameter of the elliptic functions. Mathematica will give the same picture as this using ${1.25}^2$ instead of $1.25$.

$\endgroup$
1
  • $\begingroup$ Yes, the question has to be slightly rephrased if $|k|>1$. <b>Please restrict yourself to $|k|\leq1$, $k\neq\pm1$!</b> (as indicated in Remark 1.). Sorry, for this. With the modulus within the unit circle, numerical plotting did not yield to a contradiction in my experiments. $\endgroup$
    – Twi
    Oct 8, 2015 at 8:28

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.