The behavior of herbicides in surface waters is a function of many variables, including scale of the watershed, physical and chemical properties of the herbicide, physical and chemical properties of the soil, rainfall intensity, and time of year. In this study, the transport of 6 herbicides and 12 herbicide degradates was examined during the 2004 growing season in an intermediate-scale agricultural watershed (146 ha) that is drained by a first-order intermittent stream, and the mass load for each herbicide in the stream was estimated. The herbicide load during the first week of storm events after application ranged from 17% of annual load for trifluralin to 84% of annual load for acetochlor. The maximum weekly herbicide load in the stream was generally within the first 3 weeks after application for those compounds that were applied within the watershed during 2004, and later for herbicides not applied within the watershed during 2004 but still detected in the stream. The apparent dominant mode of herbicide transport in the stream-determined by analysis amongst herbicide and conservative ion concentrations at different points in the hydrograph and in base flow samples-was either overland runoff or shallow subsurface flow, depending on the elapsed time after application and type of herbicide. The load as a percentage of use (LAPU) for the parent compounds in this study was similar to literature values for those compounds applied by the farmer within the watershed, but smaller for those herbicides that had rainfall as their only source within the watershed.