Integration of perennial filter strips (PFS) into the toeslopes of agricultural watersheds may decrease downstream nitrate (NO3) losses. However, long-term NO3 removal depends on the relative importance of several NO3 sinks in the PFS. Plant biomass and labile soil organic matter (SOM) are temporary NO3 sinks, while stable SOM is a long-term, but potentially finite, NO3 sink. In contrast, denitrification is a permanent NO3 sink. We investigated the relative importance of these NO3 sinks in PFS at the toeslope of row crop watersheds in Iowa. Using 25- × 30-cm in situ mesocosms, we added 15NO3 to PFS soils and quantified 15NO3–N recovery in plant biomass and SOM after one growing season. Further, we compared 15NO3–N recovery in particulate (relatively labile) and mineral-associated (relatively stable) SOM in mesocosms with and without growing perennial vegetation. To determine the potential importance of denitrification, we compared denitrification enzyme activity in soils from paired watersheds with and without PFS. Transfer of 15NO3–N into labile and stable SOM pools was rapid and initially independent of growing vegetation. However, SOM and plant biomass were both relatively minor NO3 sinks, accounting for 15NO3–N inputs. Denitrification enzyme activity data indicated that dissolved organic carbon derived from perennial vegetation increased potential denitrifier activity in PFS soils compared with row crop soils. Together, these results constrain SOM and plant biomass as NO3 sinks and indicate that denitrification was the most important NO3sink in perennial filter strips over one growing season.