A description of the EIDI III experiment is given. The impedance of a short electric dipole was measured in the ionosphere in conditions where, at variance with the preceding EIDI experiments, the plasma anisotropy was non-negligible (operating frequencies ∼1.6 and 2.7 MHz, ƒo F2 ∼ 3 MHz, ƒH ∼ 1 MHz). The simultaneous in-situ measurement of the relevant parameters, as well as the numerous cross-checks and experimental configurations (antenna length, dc bias, orientation, etc.), ensured a reliable comparison with theory. For the theoretical calculations, the plasma is described by the hydrodynamic equations with tensor pressure, and a vacuum sheath (the dimensions of which are calculated by using the measured plasma parameters) surrounds the antenna. The agreement is good in the part of the trajectory where the excitation frequency is above the plasma frequency; otherwise, the measured resistance is much higher than the calculated one. It is shown that the use of a more rigorous description (kinetic theory) of the homogeneous plasma cannot remove this discrepancy and it is suggested that taking into account the inhomogeneous plasma in the sheath may greatly increase the losses and explain the measured resistance, as occurred in an isotropic plasma (EIDI I experiment).