Ozone depletion in the Arctic vortex has become an important issue. As part of a Canadian research program designed to study the effect of polar stratospheric clouds on the chemistry of the Arctic ozone layer, project CANOZE (Canadian Arctic Northern Ozone experiment) 4, a series of ozonesondes were flown during January and February 1989 from Alert at 82.5¿N. This time series of ozone profiles indicated a decrease of ozone during this period. We have applied a simple chemical model to estimate the possible ozone depletions that could have occurred during this period. The results suggest a chemical depletion of ~4% per week is possible, in approximate agreement with the observed depletion rate at 20 km of 5% per week. The major features of the analysis confirmed earlier suggestions that thermal dissociation of Cl2O2 could limit the amount of O3 destruction. For our simulation conditions of largely unperturbed O3 densities the ClO+O reaction is a major loss process for O3 above 19 km, while Cl2O2 photolysis is the major loss process below 19 km, with the BrO+ClO reaction contributing about 14% of the loss in both cases. The branch of the BrO+ClO reaction that results in the formation of BrCl also acts to limit the nightime O3 destruction via BrO+ClO since BrO is rapidly converted to BrCl at twilight. The photolytic release of NO and NO2 from HNO3 at the latitudes encountered by the air parcels driving these late winter/early spring conditions is sufficiently slow that it has no major impact on O3 destruction by Clx and Brx on the time scale of a ~1 month. ¿1991 American Geophysical Union