Weakly coupled superconducting layers are described by the anisotropic 3D XY model. A lowtemperature layer decoupling due to a proliferation of fluxons between planes, as proposed by Friedel, does not occur. The same is true for a periodic superlattice of high and low T c layers, although the interplane coherence can become extremely weak. On the other hand a true layer decoupling is found for a random stack. [S0031-9007(96)01487-1] PACS numbers: 74.20.De, 74.80.Dm The Berezinskii-Kosterlitz-Thouless (BKT) transition [1,2], mediated by the unbinding of vortex pairs, has been clearly observed in superfluid films [3]. More recently, nonlinear transport experiments in layered hightemperature superconductors [4] have also shown typical signatures of vortex unbinding slightly below the critical temperature T c . This is surprising, since the Josephson coupling between the layers renders the system three dimensional, in particular close to T c . Specific heat experiments for YBCO single crystals indeed give evidence for three-dimensional critical behavior [5].Some time ago Friedel [6] argued that the interlayer coupling could be effectively suppressed by a proliferation of Josephson vortex loops or fluxons between the layers. A simple estimate for the fluxon energy suggests that the layers are decoupled at a temperature T ء , T c , so that there would be a two-dimensional regime for T ء , T , T c , with a BKT transition at T c . Unfortunately, a closer look at the problem [7] indicates that this "Friedel transition" does not occur below T c [8]. Therefore the question arises why nevertheless in nonlinear transport experiments BKT signatures are clearly observed. A way out of this dilemma has been offered by Jensen and Minnhagen [9], who realized that the Lorentz force acting on the vortices can overcome the interlayer confinement of vortex pairs.In this Letter we study again the possible existence of a two-dimensional regime below the critical temperature, starting from the classical XY model with strong intralayer coupling J k and weak interlayer coupling J Ќ . A simple criterion for the decoupling transition together with Monte Carlo simulations shows that a layer decoupling below T c does not occur, in agreement with previous studies [8]. We attribute this negative result to a strong increase of the fluxon energy as a function of temperature. We turn then to question whether a decoupling can be excluded on general grounds, by considering a superlattice of high and low T c layers [10]. For an ordered array a decoupling seems to occur around the low T c , but a closer look shows that this apparent transition is in reality a crossover from strong to weak interplane coherence. On the other hand, in a disordered array we do find a true layer decoupling.Incidentally, the problem is of more general relevance, as the question of long-range coherence arises also in other quantum systems. One can, for instance, ask whether a disordered layered system of electrons, with a large difference between the masses for the mot...