The effects of pesticide application, hydrology, and chemical and physical properties on the occurrence of pesticides in surface water in the San Joaquin River Basin, California, were examined. The study of pesticide occurrence in the highly agricultural San Joaquin?Tulare Basins is part of the National Water-Quality Assessment Program of the U.S. Geological Survey. One hundred forty-three water samples were collected throughout 1993 from sites on the San Joaquin River and three of its tributaries: Orestimba Creek, Salt Slough, and the Merced River. Of the 83 pesticides selected for analysis in this study, 49 different compounds were detected in samples from the four sites and ranged in concentration from less than the detection limit to 20 micrograms per liter. All but one sample contained at least one pesticide, and more than 50 percent of the samples contained seven or more pesticides. Six compounds were detected in more than 50 percent of the samples: four herbicides (dacthal, EPTC, metolachlor, and simazine) and two insecticides (chlorpyrifos and diazinon). None of the measured concentrations exceeded U.S. Environmental Protection Agency drinking water criteria, and many of the measured concentrations were very low. The concentrations of seven pesticides exceeded criteria for the protection of freshwater aquatic life: azinphos-methyl, carbaryl, chlorpyrifos, diazinon, diuron, malathion, and trifluralin. Overall, some criteria for protection of aquatic life were exceeded in a total of 97 samples. Factors affecting the spatial patterns of occurrence of the pesticides in the different subbasins included the pattern of application and hydrology. Seventy percent of pesticides with known application were detected. Overall, 40 different pesticides were detected in Orestimba Creek, 33 in Salt Slough, and 26 in the Merced River. Samples from the Merced River had a relatively low number of detections, despite the high number (35) of pesticides applied, owing to the generally low percentage of irrigation return flow and contribution of pesticide-free streamflow from reservoir releases. Irrigation return flows in the Orestimba Creek and Salt Slough subbasins generally contained more pesticides at higher concentrations. In addition, the distribution of seven pesticides (alachlor, cyanazine, dacthal, fonofos, molinate, napropamide, and trifluralin) in the subbasins showed a direct spatial correspondence between occurrence and application rates. Temporal patterns of occurrence also were affected by patterns of application and hydrology. Most pesticides showed a clear correspondence between the times of their application and their occurrence. Fourteen pesticides had maximum application and concentrations during the summer irrigation season. However, several pesticides exhibited maximum concentrations during winter storms, although maximum application occurred at some other time of year?the result of differences in precipitation and streamflow between seasons. In some subbasins, precipitation runoff was more effective than irrigation return flows at transporting pesticides from the site of application to the stream. Also, during autumn, when there was neither precipitation nor irrigation, the transport of pesticides to streams was limited. The effect of chemical and physical properties on the occurrence of pesticides was examined for the San Joaquin River Basin as a whole. The runoff potential of each pesticide, calculated from the solubility, water-soil organic carbon partition coefficient Koc, and hydrolysis half-life, is generally consistent with the frequency of detection of pesticides in surface water in relation to the amount applied. These three properties each were generally, and weakly, correlated with the relative load of the pesticides in surface water. Pesticide occurrence and concentrations at the mouth of the basin (the San Joaquin River near Vernalis) was compared with pesticide occurrence and concentration in the three