Fluxes of dissolved and particulate nitrogen (N) and phosphorus (P) from three adjacent watersheds were quantified with a high-resolution sampling program over a five-year period. The watersheds vary by an order of magnitude in area (12,875, 7968 and 1206 ha), and in all three watersheds intensive agriculture comprises > 90% of land. Annual fluxes of dissolved N and P per unit watershed area (export coefficients) varied ∼2X among watersheds, and patterns were not directly related to watershed size. Over the five-year period, mean annual flux of soluble reactive P (SRP) was 0.583 kg P· ha-1· yr-1 from the smallest watershed and 0.295 kg P· ha-1· yr-1 from the intermediate-sized watershed, which had the lowest SRP flux. Mean annual flux of nitrate was 20.53 kg N· ha-1· yr-1 in the smallest watershed and 44.77 kg N· ha-1· yr-1 in the intermediate-sized watershed, which had the highest nitrate flux. As a consequence, the export ratio of dissolved inorganic N to SRP varied from 80 (molar) in the smallest watershed to 335 in the intermediate-sized watershed. Because most N was exported as nitrate, differences among watersheds in total N flux were similar to those for nitrate. Hence, the total N:P export ratio was 42 (molar) for the smallest watershed and 109 for the intermediate-sized watershed. In contrast, there were no clear differences among watersheds in the export coefficients of particulate N, P, or carbon, even though > 50% of total P was exported as particulate P in all watersheds. All nutrient fractions were exported at higher rates in wet years than in dry years, but precipitation-driven variability in export coefficients was greater for particulate fractions than for dissolved fractions. Examination of hydrological regimes showed that, for all nutrient fractions, most export occurred during stormflow. However, the proportion of nitrate flux exported as baseflow was much greater than the proportion of SRP flux exported as baseflow, for all three watersheds (25-37% of nitrate exported as baseflow vs. 3-13% of SRP exported as baseflow). In addition, baseflow comprised a greater proportion of total discharge in the intermediate-sized watershed (43.7% of total discharge) than the other two watersheds (29.3 and 30.1%). Thus, higher nitrate export coefficients in the intermediate-sized watershed may have resulted from the greater contribution of baseflow in this watershed. Other factors potentially contributing to higher nitrate export coefficients in this watershed may be a thicker layer of loess soils and a lower proportion of riparian forest than the other watersheds. The among-watershed variability in SRP concentrations and export coefficients remains largely unexplained, and might represent the minimum expected variation among similar agricultural watersheds.