I doubt there's anything special about the 22 ft. As many have pointed
out, it's a deliberately unbalanced antenna which radiates from both
vertical and horizontal parts (and influenced by the towers, trees, and
buildings nearby), with a nice choke at some point to keep RF from coming
back in the shack on the outside of the coax.
...

If the choke is effective in removing the common mode current, then the
feedline will not radiate very much.

I believe that's not necessarily true.

If the distance between the main feedpoint, and the common-mode choke
is close to 1/4 wavelength at the frequency in question, then the
impedance "looking down" the outside feedline from the main feedpoint
will be relatively low, and a fair bit of RF current will flow on the
outside of the feedline. I think it's also likely that the choke
will have difficulty being very effective in this location, as it will
have to have a very high choking reactance.

If the distance from feedpoint to choke is close to a half-wavelength,
not much current will flow.

If the distance is somewhere in the middle, the outside-of-the-
feedline impedance will be rather reactive. This reactivity may
interact with the reactivity of the radiator... e.g. if the feedline
impedance is inductive, and the radiator's impedance is capacitive,
the two might tend to cancel out to some extent. This could have two
effects - a significant amount of RF current flow on the outside of
the feedline, and (perhaps) a better impedance match at the
feedpoint/balun and a lower SWR as seen by the rig.

It's possible that the 22' vertical-section length for the Carolina
Windom was selected, after experimentation, to provide the best
matching of the antenna/balun/choke system to a 50-ohm transmitter on
certain bands on which it would otherwise be difficult to match.

--
Dave Platt AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!