Abstract
The Villars-Weisskopf-Wheelon theory describing turbulent mixing of an established gradient is shown to contain a contradiction which necessitates its being discarded. To fill the gap thus created, the theory of isotropic mixing is extended to account for the presence of a gradient. The results indicate that "mixing-ingradient" cannot be employed to explain the wavelength dependence characteristic of much of the radio data. On the other hand, it is shown that experimentally determined spectra of refractive index fluctuations lend strong support to the mixing theory herein set forth. Hence, the conclusion is reached that scatter theory, as it is currently based on atmospheric turbulence, can provide, at best, an incomplete description of transhorizon propagation.