Although modelling studies have predicted(1-7) that particulate and reactive gas-phase species in the exhaust plume of large rockets might cause significant local ozone depletion, the actual response of the stratosphere after rocket launches has never been directly determined. Here we report comprehensive measurements that follow the evolution of stratospheric ozone in the wake of two Titan IV rockets launched on 12 May and 20 December 1996. In both cases, ozone concentrations dropped to near-zero values in the plume wake, across regions four to eight kilometres wide, within 30 minutes after launch; intense ozone lass persisted for 30 minutes after which time concentrations recovered to ambient levels. Our data indicate that the number of ozone molecules lost in the plume regions significantly exceed the number of chlorine molecules deposited by the two rockets, This suggests that a catalytic cycle based on Cl2O2, other than Cl-2, and unique to solid rocket motor (SPM) plumes might be responsible for our observations. However, the limited spatial and temporal extent of the observed ozone losses implies that neither the catalytic Cl2O2 cycle nor other reactions involving exhaust compounds from large solid-fuelled rockets have a globally significant impact on stratospheric chemistry.