The U.S. Environmental Protection Agency has proposed a drinking water standard for 222Rn in public water supplies. When promulgated, operators of public water supplies will be required to determine water 222Rn concentrations. Most likely, water samples will be collected and mailed to laboratories for analyses. Additionally, it is probable that homeowners will test private well water in a similar manner by collecting water samples and mailing them to a laboratory for 222Rn analyses. In anticipation of these actions, this study was conducted to evaluate two methods of water sample collection and to evaluate the potential losses of 222Rn from water samples mailed to a laboratory. Thirdly, temporal variations in 222Rn concentrations in several groundwater supplies in North Carolina were examined. Water supplies at four sites in eastern North Carolina and five sites in western North Carolina were sampled over a 7-mo period beginning in the fall of 1993 and continuing through the spring of 1994. Samples were analyzed using a liquid scintillation method. This study showed that a "slow-flow" method is not only suitable for sample collection, but may be the preferable method for water sample collection since slow-flow collection resulted in less 222Rn loss than was observed during syringe collection, i.e., the water 222Rn concentrations in samples collected by the slow-flow method were generally higher. Further, based on this study, mailing water samples to a laboratory for analysis should not have a substantial effect on the measured 222Rn concentrations. Consequently, water samples can be collected by the slow-flow method by water supply operators, as well as homeowners, and mailed to laboratories for 222Rn analyses with reasonable assurance that the samples have not suffered significant 222Rn loss. Temporal variations in water 222Rn concentrations were observed in this study. One factor complicating the study of temporal variations in 222Rn concentrations was the striking influence that inconstant water usage and inadequate well purging prior to sample collection had on 222Rn loss from the water.