This review attempts to interpret the generalized scale invariance observed in common atmospheric variables -wind, temperature, humidity, ozone and some trace species -in terms of the computed emergence of ring currents (vortices) in simulations of populations of Maxwellian molecules subject to an anisotropy in the form of a flux. The data are taken from 'horizontal' tracks of research aircraft and from 'vertical' trajectories of research dropsondes. It is argued that any attempt to represent the energy distribution in the atmosphere quantitatively must have a proper basis in molecular physics, a prerequisite to accommodate the observed longtailed velocity probability distributions and the implied effects on radiative transfer, atmospheric chemistry, turbulent structure and the definition of temperature itself. The relationship between fluctuations and dissipation is discussed in a framework of non-equilibrium statistical mechanics, and a link between maximization of entropy production and scale invariance is hypothesized.