Considerable variation in the export of dissolved inorganic nitrogen (DIN) and dissolved organic nitrogen (DON) was observed among catchments located within an old-growth sugar maple forest in central Ontario. Although discharge was a strong predictor of N-export, rates of export were variable for each catchment, ranging from -50% to +50% from the catchment-average response for DIN and -25% to +25% from the catchment-average response for DON. Among the catchments, a unifying flushing behavior was apparent for NO3--N, the dominant form of DIN in the discharge waters, providing a basis for explaining the variation in the export of DIN. Flushing occurs when a water table rises to the soil surface with subsequent mobilization of nutrients stored near or at the soil surface to surface waters. Catchment-specific flushing behaviors were captured in flushing characteristic time constants, defined as the time interval required for a decline in N concentrations in discharge waters to e-1 (37%) of their initial concentration. Variation in flushing behavior was linked to variation in N export; catchments with short flushing times (interpreted as catchments with source areas that are less variable) were observed to export less N than catchments with long flushing times (source areas that are more variable). A hypothesis was formulated in which catchment topography and its influence on variable source area dynamics accounts for variation in flushing behavior, hence variation in the export of NO3--N among the catchments. The implication of this hypothesis is that to predict accurately the export of NO3--N from catchments within a landscape, we need first to consider the influence of the topographic complexity of the catchments. Our understanding of the mechanisms of processing and export of DON is not sufficient for accurate prediction at this point, highlighting the need for additional research on DON. ¿ 1998 American Geophysical Union