I've seen a couple of postings so far recommending a single balun at the
feedpoint. While this takes care of conducted common mode current, it
doesn't eliminate current induced on the feedline from mutual coupling.
All the single balun does is to insure that the current is near zero at
the single point where the balun is inserted. Current is still induced
on the feedline, and you end up with a current maximum a quarter
wavelength down the line -- the distribution looks just like it does on
an antenna (which the feedline has actually become), with the balun
location being the end of the "antenna". The balun can even make the
induced current worse if the effective common mode open circuit at the
balun insertion point results in a more nearly resonant feedline length.
The only way to really effectively reduce the induced current to a low
level is to break up the resonance of the feedline by inserting a second
balun about a quarter wavelength below the feedpoint as I recommended
earlier.

Roy Lewallen, W7EL