One of the fundamental tasks of the Air Infiltration Centre has been to undertake a program to validate mathematical models of air infiltration. The principal objectives of this study have been to use experimental data to assess the reliability and full range of applicability of air infiltration models and to identify the key parameters that must be accurately specified to ensure reliable results. Three numerical data sets, representing 70 sets of air-infiltration measurements made in dwellings in Canada, Switzerland, and the United Kingdom, have been prepared. These are being used in eight air-infiltration models, of which two have been fully assessed. Excellent agreement between calculated and measured infiltration rates has been obtained for a wide range of terrain, shielding, and climatic conditions. The most difficult model parameter to quantify was found to be the wind-pressure coefficient, and it is concluded that much improved results would be possible by performing a systematic wind-tunnel study incorporating fixed degrees of shielding and a small range of building shapes.