A water-quality index (WQI) was developed from historical data (1986-1995) for streams in the Atlanta Region and augmented with 'new' and generally more comprehensive biweekly data on four small urban streams, representing an industrial area, a developed medium-density residential area and developing and developed low-density residential areas. Parameter WQIs were derived from percentile ranks of individual water-quality parameter values for each site by normalizing the constituent ranks for values from all sites in the area for a base period, i.e. 1990-1995. WQIs were developed primarily for nutrient-related parameters due to data availability. Site WQIs, which were computed by averaging the parameter WQIs, range from 0.2 (good quality) to 0.8 (poor quality), and increased downstream of known nutrient sources. Also, annual site WQI decreases from 1986 to 1995 at most long-term monitoring sites. Annual site WQI for individual parameters correlated with annual hydrological characteristics, particularly runoff, precipitation quantity, and water yield, reflecting the effect of dilution on parameter values. The WQIs of the four small urban streams were evaluated for the core-nutrient-related parameters, parameters for specific dissolved trace metal concentrations and sediment characteristics, and a species diversity index for the macro-invertebrate taxa. The site WQI for the core-nutrient-related parameters used in the retrospective analysis was, as expected, the worst for the industrial area and the best for the low-density residential areas. However, macro-invertebrate data indicate that although the species at the medium-density residential site were diverse, the taxa at the site were for species tolerant of degraded water quality. Furthermore, although a species-diversity index indicates no substantial difference between the two low-density residential areas, the number for macro-invertebrates for the developing area was much less than that for the developed area, consistent with observations of recent sediment problems probably associated with construction in the basin. However, sediment parameters were similar for the two sites suggesting that the routine biweekly measurements may not capture the short-term increases in sediment transport associated with rainstorms. The WQI technique is limited by the number and types of parameters included in it, the general conditions of those parameters for the range of conditions in area streams, and by the effects of external factors, such as hydrology, and therefore, should be used with caution.