Experience regarding the testing of models of masonry buildings on simple shaking tables is discussed. It is stressed that because scale effects affect some phenomena and influence the dynamic response, only overall seismic behaviour of structural systems can be studied by testing reduced size models, and not structural details. The basic principles of physical modelling are explained and options for preparing different types of models presented. The test facility, needed for simulation of earthquake ground motion, is described and pointers are given for the selection of simulated earthquakes. Some practical cases of model shaking table studies are also presented.