The GILDES model has been used to perform the first theoretical mechanistic study of the atmospheric corrosion of zinc in a controlled environment. Detailed comparisons are made of model results with laboratory data for short-term (0-1000 min) and long-term (0-680 h) exposures of zinc to SO2, NO2 and O-3 at high relative humidities. The agreement of theory and experiment is excellent, indicating that the model has captured the principal processes (especially ligand-promoted dissolution) involved in the corrosion of zinc under these conditions. Copyright (C) 1996 Published by Elsevier Science Ltd