Richard Fry wrote:
Unless some means is provided to prevent r-f current flow
on the outside of the coax and on the tower structure,
they will radiate/receive r-f energy.

Thus making the vertical antenna longer than 5/8WL.
Using the top of the tower for a ground simply
makes the tower part of the antenna system turning
the entire array into an off-center-fed vertical
dipole with the bottom end grounded. For instance,
a 1/4WL 20m monopole mounted on top of a 60 foot
tower using the tower as the coax shield ground
has a take-off-angle of 57 degrees. The highest
RF current is near the middle of the tower. :-(

To make matters even worse: I had a similar problem
with drooping 1/4WL radials DC insulated from the tower.
The drooping radials coupled RF into the tower and
turned it into a radiator which screwed, oops, I
mean skewed the radiation pattern upwards. It took
me a long time to figure out why my horizontal dipole
was magnitudes better than my 1/4WL vertical on top of
the 1.25WL tall tower which was grounded at the bottom
and floating at the top.
--
73, Cecil, IEEE, OOTC, http://www.w5dxp.com