The stray capacitances to the rest of the world will be of the same order as
the required value of capacitance. The presence of stray capacitances, not
being accounted for in the design, will cause whatever it is to behave in an
unpredictable manner.

You may be right Reg, however the capacitive loading I need is not distributed
throughout the length on the antenna. Seems the stray capacitance would be
distributed. In particular I need -j2500 ohms at a specific point on the
antenna.
What I am attempting is to change the current distribution on an Extended
Double Zepp (1.25 lambda dipole) such that the gain and F/B ratio are more to
my liking. There is a published article on doing just this on a 40M EDZ.
-2500 ohms at 7.2 mhz is 9pf. I just scaled it up to 147mhz to .4pf. I have
tried different values of loading, and different positions on the antenna, but
this seems optimum.
What I have now is a 3 element beam with 17dbi forward gain with 12 db F/B
ratio, at least on paper.
Methinks you should re-engineer whatever it is to something which requires
a capacitance of something greater than 0.4 pF and is of a more realiseable
and predictable value.
Maybe back to the drawing board? Decrease loading and change position. What
size cap. would be resonable?

73 Gary N4AST