A nested surface water monitoring network was designed and tested to measure variability in pesticide concentrations in the San Joaquin River and selected tributaries during the irrigation season. The network design an d sampling frequency necessary for determining the variability and distribution in pesticide concentrations were tested in a prototype study. The San Joaquin River Basin, California, was sampled from April to August 1992, a period during the irrigation season where there was no rainfall. Orestimba Creek, which drains a part of the western San Joaquin Valley, was sampled three times per week for 6 weeks, followed by a once per week sampling for 6 weeks, and the three times per week sampling for 6 weeks. A site on the San Joaquin River near the mouth of the basin, and an irrigation drain of the eastern San Joaquin Valley, were sampled weekly during the entire sampling period. Pesticides were most often detected in samples collected from Orestimba Creek. This suggests that the western valley was the principal source of pesticides to the San Joaquin River during the irrigation season. Irrigation drainage water was the source of pesticides to Orestimba Creek. Pesticide concentrations of Orestimba Creek showed greater temporal variability when sampled three times per week than when sampled once a week, due to variations in field management and irrigation. The implication for the San Joaquin River basin (an irrigation-dominated agricultural setting) is that frequent sampling of tributary sites is necessary to describe the variability in pesticides transported to the San Joaquin River.