The reliability of tomographic attenuation images depends critically
on the a priori assumptions of the inverse procedures. One important
assumption is that the radiation patterns can be represented by formulas
for dipole antennas located in a homogeneous full-space. Modeling
studies suggest that this assumption is valid for gas-filled boreholes,
but is violated for water-filled ones. To test experimentally these
numerical results we developed a special packer system. This allows
us to isolate the effects of the borehole filling by comparing data
gathers measured with (gas-filled hole) and without the packer system
(water-filled hole). Our experimental results demonstrate that the
radiative characteristics of borehole antennas do indeed differ substantially
for gas- and water-filled boreholes and that the dominant effects
are correctly predicted by simple modeling approaches.