The nature of the diffusion velocity for ordinary diffusion is carefully examined. It is shown that its physical character is different to the forced diffusion velocity in that it is not necessary to associate diffusion stresses and kinetic energy with ordinary diffusion. As a consequence, momentum and energy conservation equations based on coexisting-continua models found in the literature are shown to be of doubtful utility, and are often incorrectly interpreted. In addition, a widely seen form of the thermal energy equation is shown to be in error. A single continuum model is proposed for non-ionized gas mixtures that is adequate for practical use. An appropriate form of the coexisting-continua model conservation equations is given for plasma when forced diffusion is the dominant diffusion mechanism.