Mathematical models were utilized to study water pollution control programs in a river basin. Sensitivity analyses, with a steady state model, showed substantial variation of cost for sewage treatment, depending upon stream purification parameter selections. When actual parameters are less favorable than design values, quality standards may not be met; these effects are more serious with lower levels of treatment. An unsteady state model was developed to trace a time profile at any specified station in terms of flow and quality while up-stream discharge, water temperature, and solar radiation vary. The techniques assume that, for short reaches and/or times, steady state conditions apply without undue loss of accuracy. A new empirical procedure was developed to route unsteady stream flow. The time varying model was used to investigate the effectiveness of an assumed configuration of treatment plants when the stream's assimilative capacity varies with distance and time. Susceptibility to poorer conditions increases with higher BOD releases. Lower treatment levels also result in a greater range of river conditions than high levels. Sensitivity analyses of stream parameters were also made with the time varying model. (Author)