Current methods for estimating the concentrations of inorganic chlorine/bromine species (Cly/Bry) in the stratosphere due to decomposition of tropospheric source gases assume that the Cly/Bry concentration in the stratosphere is determined mainly by the balance between production from in situ oxidation of the source gases in the stratosphere and removal by transport of Cly/Bry out of the stratosphere. The rationale being that for source gases whose lifetimes are of the order of several months or longer the concentration of Cly/Bry in the troposphere is small because they are produced at a relatively slow rate and also removed efficiently by washout processes. As a result of the small concentration, the rate at which Cly/Bry is transported to the stratosphere is expected to be small compared to the in situ stratospheric production. Thus the transport of Cly/Bry from the troposphere contributes little to the stratospheric concentration. In contrast, the origin of stratospheric Cly/Bry from reactive source gases with tropospheric lifetimes comparable to the washout lifetime of Cly/Bry (of the order of 10--30 days) in the troposphere is distinctly different. The in situ source in the stratosphere is expected to be significantly smaller because only a small portion of the source gas is expected to survive the troposphere to be transported into this region. At the same time these short-lived source gases produce appreciable amounts of Cly/Bry in the troposphere such that transport to the stratosphere offers a larger source for stratospheric Cly/Bry than in situ production. Thus, for reactive source species, simple methods of estimating the concentration of stratospheric Cly/Bry that ignore the tropospheric contribution will seriously underestimate the loading. Therefore estimation of the stratospheric Cly/Bry loading requires not only measurements of tropospheric source gases but also measurements of Cly/Bry at the tropopause. This paper illustrates the mechanism by using results from a two-dimensional chemistry-transport model. However, in view of the importance of tropospheric transport on stratospheric loading the detailed values should be further evaluated using a three-dimensional model with appropriate treatment of convective transport. ¿ 1997 American Geophysical Union