Water‐table altitude, a controlling factor for ground‐ water flow, was estimated from detailed topographic data by subtracting the estimated depth‐to‐water. Land‐surface altitude of the Coastal Plain in the south‐central United States varies from 0 to more than 800 feet above sea level. Predevelopment depth‐to‐water in 6,825 wells less than 150 feet deep averages 25.7 feet (standard deviation, 19.5 feet). Most water‐table‐altitude variation is due to variation in land‐surface altitude and not due to variation in depth‐to‐ water. Digital topographic data, from 1:250,000 scale maps for every 30 seconds of latitude and longitude are available for the continental United States. About 90 altitudes were averaged for each 25‐square‐mile block of a rectangular grid used for ground‐water flow modeling. Multiple linear regressions of predevelopment water‐level data and topographic data were used to derive empirical equations relating water‐table altitude to topography. The regression method was more consistent, efficient, and accurate than manually digitizing values from manually contoured water‐table maps. Water‐table maps usually are prepared from few data that are concentrated in topographically flat areas. Manually digitizing water‐table maps on a regional scale introduces additional error. About 35 percent of the water‐table altitudes obtained manually were greater than average land‐surface altitudes from topographic data. The mean difference between water‐table altitudes from the two methods was less than 10 feet, which indicates no systematic error was incorporated in the regression method. Copyright © 1989, Wiley Blackwell. All rights reserved