Quark-hadron duality and its potential applications are discussed. We focus on theoretical efforts to model duality.1 What is Duality?In general, duality implies a situation in which two different languages give an accurate description of Nature. While one may be more convenient than the other in certain situations, both are correct. If we are interested in hadronic reactions, the two relevant pictures are the quark-gluon picture and the hadronic picture. In principle, we can describe any hadronic reaction in terms of quarks and gluons, by solving Quantum Chromodynamics (QCD). While this statement is obvious, it rarely has practical value, since in most cases we can neither perform nor interpret a full QCD calculation. In general, we also cannot perform a complete hadronic calculation. We will refer to the statement that, if one could perform and interpret the calculations, it would not matter at all which set of states -hadronic states or quark-gluon states -was used, as "degrees of freedom" duality.However, there are cases where another, more practical form of duality applies: for some reactions, in a certain kinematic regime, properly averaged hadronic observables can be described by perturbative QCD (pQCD). This statement is much more practical than the "degrees of freedom" duality introduced above. In contrast to full QCD, pQCD calculations can be performed, and in this way, duality can be exploited and applied to many different reactions.Duality in the latter form was first found by Bloom and Gilman in 1970 in inclusive, inelastic electron scattering 1 . Duality in this reaction is therefore commonly referred to as Bloom-Gilman duality. Recently, it was impressively confirmed to high accuracy in measurements carried out at Jefferson Lab 2 .Duality also appears in the semileptonic decay of heavy quarks 3,4 , in the reaction e + e − → hadrons 5 , in dilepton production in heavy ion reactions 6 ,