On Walsh and Weil's defense of the conventional method

As we [Lee and Papas, 1965] have pointed out, the steady-state, time-average power P emitted by an antenna in an anisotropic medium has two parts, an irreversible part Pirr and a reversible part Prev. That is, P = Pirr + Prev. It is our contention that true radiated power is given by Pirr and not necessarily by P. In special cases where Prev = 0, e.g., the case where the antenna is in a vacuum or the case where the antenna is in a homogeneous, isotropic, lossless medium, it makes no difference whether one uses P or Pirr as the measure of radiated power. However, if the medium surrounding the antenna is gyroelectric, there is no reason to believe that Prev = 0, and one must be careful to account for the difference between P and Pirr.