Consider the function:

$$f(a) = \frac{K\left(\frac{2 k(a) i}{g(a) + k(a) i}\right)}{\sqrt{g(a) + k(a) i}}$$

where $g(a)$ and $k(a)$ are smooth real-valued functions of a real parameter $a\in[0,1]$, with $g(a)$ positive on part of its domain, and $k(a)$ positive on its entire domain. $K$ is the complete elliptic integral of the first kind, and we are following the convention where:

$$K(m) = \int_0^{\pi/2}\frac{1}{\sqrt{1-m \sin(\theta)^2}}\,d\theta$$

When $g(a)\ge 0$, $f(a)$ will be real-valued. However, if $g(a)$ becomes negative at a point where:

$$\Re\left(\frac{2 k(a) i}{g(a) + k(a) i}\right) \gt 1$$

then we encounter a branch cut in $K$ that stretches from 1 to $\infty$ along the real axis, and $f$ switches discontinuously to an imaginary-valued function.

Is there any systematic way to continue the smooth, real-valued function defined on the portion of the domain where $g(a)\ge 0$, into a larger domain?

In the particular example of interest to me, $g(a)$ only changes sign at one point in the domain, and:

$$\frac{2 k(a) i}{g(a) + k(a) i}$$

traces out an approximate three-quarters of a unit circle centred on 1 in the complex plane, with a gap in the circle between $\pi$ and about $\frac{3\pi}{2}$.


The question is resolved by this detailed answer on the analytic continuation of the elliptic integral:



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.