Between the interplanetary nledium, lilled by winds, magnetic slruclures, etc., and the interior of stars, opaque, and dominated heavily by the gravitational spherical field, Ihe slellar atmosphere is a place where the true physical equilibrium, on the inside, sufficiently described by the parameters/.~, ..~/, 'R., and the chemical composition ,:V. y, Z, is progressively changing into a situation far tor equilibrium, which needs many more paramelers m be properly described.
The assumption that the equilibrium situation was dominating in the atmosphere has been generally accepled during the first half of this century. Since 1950 or so, we progressively learnt thai the du~rmodvnamical equilihrium (TE), and even the 'local' Ihermodynamical equilibrium (El'E), are far from being actually in existence, that the radiative equilibrium (RE) is not actually perfect, convection, diffl.lsion, magnetism, dissipation processes.., playing a non-negligible part in lhe energy transport, that the hydrostatic equilibrium (HE) is only an approximation, as the convection and the magnclism are affecting the atmospheric layers, lllat neither the .whericiO of an'no.v~heric laver,s (plane-parallel h)7)othesis: PP) is achieved, nor the homogeneity of slellar iso-r layers. I)uring the 1950s and following decades, we began to suspecl these difficuhies and their consequences. In this paper, we Itlrn towards a new consequence of the last-mentioned effect: the influence of non-sphericity and inhomogcneity upon the stellar (and solar i?crhaps) abundances of elements.