The physical meaning of the energy velocity in lossy Lorentz media is clarified. First, two expressions for the energy velocity, one by Brillouin and another by Diener, are examined. We show that, while Diener's is disqualified, Brillouin's is acceptable as energy velocity. Secondly, we show that the signal velocity defined by Brillouin and Baerwald is exactly identical with the Brillouin's energy velocity. Thirdly, by using triangle-modulated harmonic wave, we show that the superluminal group velocity plays its role as a revelator only after the arrival of the signal traveling at the subluminal energy velocity. In short, nothing moves at the group velocity, and every frequency component of a signal propagates at its own energy velocity.

Gaussian pulse has no beginning point, so has no Laplace transform and is non-physical. We propose sinnt pulse (referred to as pseudo-Gaussian pulse or PGP) as an approximation of the Gaussian pulse. PGP has the Laplace transform and approaches the Gaussian pulse as n→∞. The propagation of PGP-modulated wave packet in the highly anomalous dispersion band of a Lorentz medium is investigated by numerical inversion of Laplace transform. Our results are greatly different from the conventional results obtained by the saddle point method. Our results show that the velocity of a Gaussian wave packet cannot be explained only by the concept of the group velocity as has been done so far.