Electrostatic double layers have been proposed as an acceleration mechanism in solar flares and other astrophysical objects. They have been extensively studied in the laboratory and by means of computer simulations. The theory of steady-state double layers implies several existence criteria, in particular the Bohm criteria, restricting the conditions under which double layers may form. In the present paper several already published theoretical models of different types of double layers are discussed. It is shown that the existence conditions often imply current-driven instabilities in the ambient plasma, at least for strong double layers, and it is argued that such conditions must be used with care when applied to real plasmas. Laboratory double layers, and by implication those arising in astrophysical plasmas often produce instabilities in the surrounding plasma and are generally time-dependent structures. Naturally occuring double layers should, therefore, be far more common than the restrictions deduced from idealised time-independent models would imply. In particular it is necessary to understand more fully the time-dependent behaviour of double layers. In the present paper the dynamics of weak double layers is discussed. Also a model for a moving strong double layer, where an associated potential minimum plays a significant role, is presented.