Constructed wetlands (CWs) in the agricultural landscape reduce non-point source pollution through removal of nutrients and particles. The objective of this study was to evaluate if measurements of natural abundance of (15)NO(3)(-) can be used to determine the fate of NO(3)(-) in different types of small CWs treating agricultural runoff. Nitrogen removal was studied in wetland trenches filled with different filter materials (T1--sand and gravel; T3--mixture of peat, shell sand and light-weight aggregates; T8--barley straw) and a trench formed as a shallow pond (T4). The removal was highest during summer and lowest during autumn and winter. Trench T8 had the highest N removal during summer. Measurements of the natural abundance of (15)N in NO(3)(-) showed that denitrification was not significant during autumn/winter, while it was present in all trenches during summer, but only important for nitrogen removal in trench T8. The (15)N enrichment factors of NO(3)(-) in this study ranged from -2.5 to -5.9 per thousand (T3 and T8, summer), thus smaller than enrichment factors found in laboratory tests of isotope discrimination in denitrification, but similar to factors found for denitrification in groundwater and a large CW. The low enrichment factors compared to laboratory studies was attributed to assimilation in plants/microbes as well as diffusion effect. Based on a modified version of the method presented by Lund et al. [Lund LJ, Horne AJ, Williams AE, Estimating denitrification in a large constructed wetland using stable nitrogen isotope ratios. Ecol Engineer 2000; 14: 67-76], denitrification and assimilation were estimated to account for 53 to 99 and 1 to 47%, respectively, of the total N removal during summer. This method is, however, based on a number of assumptions, and there is thus a need for a better knowledge of the effect of plant uptake, microbial assimilation as well as nitrification on N isotopic fractionation before this method can be used to evaluate the contribution of dinitrification in CWs.