A Doppler-spread theory for the “saturation” of middle-atmosphere gravity–wave spectra (in vertical wave-number m) is presented in a companion paper. It includes a formula for the large-m limit that would be imposed by the onset of instability. At sufficiently great heights, however, this limit may not be attained because of dissipation imposed by molecular viscosity and conductivity. Here, the transition between the two regimes is taken to mark the turbopause—the level at which turbulence ceases—and relevant relations are obtained. These are shown to be consistent not only with observations at turbopause levels, but also, after extrapolation downward through five orders of magnitude in atmospheric density (with the use of the Doppler-spread theory), with similar observations in the middle stratosphere. In a second application, the concepts behind the Doppler-spread theory are applied to circumstances that would be found on individual occasions (rather than in statistical ensembles, which the basic theory treats). Horizontally stratified layers of intensified turbulence are then found to be expected, perhaps in the stratosphere and more probably in the mesosphere, as is observed. The layers are often observed by medium-frequency, partial-reflection radar techniques, including in particular spaced-receiver “drift” techniques that are often taken to represent the ambient winds. Such an interpretation is confirmed to be appropriate to the present model.