Here's something,word for word, I wrote on rainbows two years back for physics-zone,a yahoo group(hope it's useful):-
I've read a wonderful review of a book on rainbows and
I'll share what I've read.There is a primary (rain)bow
below a secondary bow---in between there is a dark
band called Alexander's band.The order of colours in
the primary bow is reverse that of the order of the
colours in the secondary bow.Just below the top of the
primary rainbow are pastel fringes called
supernumerary bows.The primary and secondary bows are
easily explainable by geometrical optics--reflection
and refraction (inside and at the surface of a water
droplet).
Supernumerary bows are explained by considering the
wave theory of light.It turns out the primary bow is
the first interfernce maximum,2nd,3rd etc. maxima
correspond to the supernumerary bows---try to figure
this out considering a wavefront incident on a droplet
and following its evolution inside and out of the
droplet.
The transition to the dark bank is not an abrupt one
due to diffraction effects.All this was wonderfully
worked out by Airy.Airy's function is the solution of
the equation
y'' + xy = 0
The solution is an oscillating function with
decreasing
amplitudes of maxima--the first corresponds to the
primary bow and subsequent ones to the supernumerary
bows.The transition to the dark band also turns out to
be a smooth one.
Even Airy's theory is an approximate one and can not
explain some of the laboratory generated rainbows.Mie
worked out this problem by considering scattering of
light off homogeneous spheres.Peter Debye also worked
on the problem.The scattering amplitudes are expressed
as an infinite sum(partial wave expansion).The no. of
terms that must be retained in the sum(before the
series can be truncated turns out to be quite high---a
lot of computer time would be required to to arrive at
a numerical solution!
A complete analysis of the problem requires
introducing a complex angular momentum variable---the
contributions to the partial wave series are then
redistributed to a set of points in the complex
plane(called Regge poles in particle physics).
There are many more aspects in this complex analysis
like 'collision of two real saddle points' resulting
in 2 complex saddle points being born!As far as the
physics is concerned,Airy's theory suffices.
Jagmeet Singh