antenna length and frequency dependence

Hi,
I have the following question. Suppose I take an antenna with air substrate. Therefore there are no dielectric losses. The only source of loss is conductor loss, but I can made an antenna from silver and minimize the losses.
Now, there is a theorem that states that I can match any impedance to any other impedance at a single frequency point with a lossless network composed of an inductor and a capacitor.
As a result I can match the antenna at a low frequency compared with its size. The directivity of such electrically small antenna is close to 0 dBi and since there no losses the gain is equal to directivity.
However, I've seen measurement results of electrically small patch antennas on air substrate and their gain was much lower than 0 dBi even at the point of perfect match.
So the question is: what's wrong with my reasoning?
Thanks,
fencer32

Are you simulating your antenna.. in which software ..? e.g if using HFSS it wud be better to use perfect conductor than aluminium or silver.. please elaborate ur prblm .. i mite get a clear picture ..

Hi,
Every Antenna can Transmit or Receive a particular frequency, How ever Smart Antennas are Adaptive in Nature, we can move the beam by phase shft technique using adaptive algirithm but still the frequency is unique for a particular antanna

Aperture is also important in antenna, theoritically u can reduce the size of antenna but you may loose gain by downsizing the antenna

_sree