How to arrive at Bessel function solution to 1D polynomial potential

[Peeter]

My quantum text, leading up to the connection formulas for WKB and the Bohr-Sommerfeld quantization condition states that for

$$\begin{align}u'' + c x^n u = 0 \end{align}$$

one finds that one solution is

$$\begin{align}u &= A \sqrt{\eta k} J_{\pm m}(\eta) \\ m &= \frac{1}{{n + 2}} \\ k^2 &= c x^n \\ \eta &= \int_0^x dx' k(x')\end{align}$$

I'd like to know how this was arrived at. Could somebody outline what set of change of variables would one make to put the differential equation above into the standard Bessel differential form?

 

[dextercioby]

Try first a change of variables x^n/2 = t. Compute the ODE in terms of u(t) then.