A long-term program to monitor aquatic ecosystems requires a long-term commitment of significant resources. To maximize the effectiveness of such a program, monitoring sites must be carefully selected, addressing a complex array of relevant concerns in the process. Monitoring sites for the TIME Project (Temporally Integrated Monitoring of Ecosystems), scheduled to begin in 1991, will be selected for chemical and biological sampling in order to meet several distinct objectives including (1) establishing an 'early warning' network to detect changes in aquatic resource acidification or recovery, (2) providing periodic regional assessments of surface water acidification status, (3) providing data for validation of watershed models, and (4) relating surface water responses to atmospheric deposition. The paper focuses on a method for selecting sites to meet the first objective for lakes. The approach involves a cluster analysis to stratify the NSWS sample and ordination to identify lakes that are likely to respond to changes in acid inputs.