christofire wrote:

All else being equal, a lossless 'ground plane' type antenna, be it a
monopole developed from a dipole or a discone developed from a biconical
dipole, over an infinite ground plane should exhibit 3 dB _more_ gain than
the symmetrical 'parent' form of antenna. This is because its radiation
pattern is limited to half the solid angle of the parent (e.g. only the
space above the ground plane). Then for a given number of watts fed into
the antenna, the power-flux density must be greater in the region where it
can radiate. In practice, the ground plane isn't infinite so there is some
'undercutting' of the vertical radiation pattern, but the gain should still
be somewhat greater.
. . .

This is entirely a fictional scenario, although it's the one used by
virtually all the currently available modeling programs. In those
programs you can choose "free space" or "ground plane", where the
infinite "ground plane" restricts the field, as Chris says, to a single
hemisphere while "free space" allows radiation in both hemispheres.

In real life, you can't have either one, except that outer space would
be a reasonable approximation of "free space". The only thing that
matters is whether the field reflects from a large surface like the
Earth on its way to the receiver. If it does, you potentially pick up
field strength from reinforcement of the direct and reflected rays --
but of course you can also lose field strength if the two rays cancel
rather than reinforcing. This is another way of describing the same
phenomenon of increased gain due to a ground plane. Even if you put an
antenna hundreds of wavelengths high, some of the signal will reflect
from the ground -- it's not truly in "free space". It doesn't matter
whether your antenna is a "ground plane" or a dipole -- if a reflection
occurs between the transmitter and receiver, you potentially get that
extra gain; if it doesn't, you don't.

Roy Lewallen, W7EL