The U.S. Geological Survey (USGS) Chlorofluorocarbon (CFC) Laboratory in Reston, Va., has been measuring concentrations of CFCs in ground-water samples since 1989 to estimate the year that a water sample was recharged to a ground-water flow system. The water samples have been collected in flame-sealed borosilicate ampoules. Typically for each site, three samples were analyzed within days to a few months after collection, and additional samples were archived for extended periods of time (up to four years). The stability of CFC concentrations in the archived water samples from the USGS CFC Laboratory was investigated by analyzing the CFC concentrations in archived water samples and comparing them with the CFC concentrations that were obtained soon after the samples were collected. The archived samples selected for analysis were chosen from sites with a wide variety of hydrogeologic and geochemical conditions. For CFC-11 and CFC-12 concentrations, approximately 14% and 10.5%, respectively, of the archived samples were statistically different (both higher and lower) from the concentrations obtained from analyses conducted soon after the sample collection. Most of the extraneous values were attributed to natural variability of CFC concentrations originally in the water discharged from wells, rather than to microbial degradation within the ampoule on storage.