Streambank fencing is a best management practice that is targeted to reduce suspended sediment and nutrient inputs to streams by reducing direct inputs from animals, eliminating streambank trampling, and promoting streambank revegetalion. A paired basin study is being conducted in two small adjacent basins to determine the waler quality effects of streambank fencing. This article documents the 3-yr calibration period between control and treatment basins prior to fence installation. Approximately 70% of land adjacent to streambanks in the study area is used as pasture. Nutrient quantities applied as manure, benthic-macroinvertebrate communities, and the physical habitat of each stream were similar in both basins. Total N, P, and suspended sediment yields measured at the outlet of each basin averaged about 56, 2.8, and 2650 kg ha⁻¹ on an annual basis. For both basins, about 90% of the total N yield was attributable to dissolved NO₃-N and about 90% of the total N yield occurred during nonstormflow; conversely, about 90% of the total P yield was attributable to stormflow and 60 to 65% of the total P yield was suspended. Regression equations developed between both basins for low flow and stormflow samples for nutrients, suspended sediment, and discharge indicated a significant relation for most constituents. Pretreatment relation between basins for low flow and stormflow samples would need to change by 6 and 14% for total N concentrations and 24 and 9% for total P concentrations in order for streambank fencing to significantly affect water quality in the treatment basin.