I still think that page 10 of the November 2003
issue of Physics Today does not offer evidence
for any kind of rainbow interference. But, as we
know, the term "interference" can be used to describe
other things than equidistant fringes, as in Young's
experiment. Colors seen in light reflected and
refracted from thin films fall under the category of
interference.
How small should droplets of water (in air) be to
produce colors due to interference? Would droplets
of that size still produce rainbows (rather than
scattering of light in all directions)? These questions
are triggered by a piece sent to me in private.