Uptake and cycling of nitrogen (N) in the littoral zone of a lake-receiving nutrient-rich mine water located in Boliden, northern Sweden, was investigated. Stable isotope tracer solutions of 15 N as NH 4? (NAM mesocosm) or NO 3 -(NOX mesocosm) were added to mesocosms enclosing plants of common reed (Phragmites australis). The 15 N abundance in various plant parts was measured at pre-defined time intervals over an experimental period of 22 days. During the course of the experiment, plant parts from the NAM mesocosms were significantly more enriched in 15 N than plant parts from the NOX mesocosms. On day 13, Dd 15 N values of the fine roots from the NAM mesocosms had reached ?8220 %, while the maximum Dd 15 N value in NOX roots was considerably lower at ?4430 %. Using 15 N values in macrophyte tissues present at the end of the experiment enabled calculations of uptake rates and % of tracer N recovered in the plant (%tracerNrecov). Maximum tracer uptake rates were higher for the NAM mesocosms (1.4 lg g -1 min -1 or 48 mg N m -2 d -1 ) compared to the NOX mesocosms (0.23 lg g -1 min -1 or 8.5 mg N m -2 d -1 ). Calculations of %tracerNrecov indicated that 1-8 and 25-44 % of added N was assimilated by plants in the NOX and NAM mesocosms, respectively. Hence, P. australis was more effective in assimilating NH 4 ? , and a larger portion of the tracer N accumulated in the roots compared to the other plant parts. Consequently, macrophyte N removal is most effective for cold-climate aquatic systems receiving mine water dominated by NH 4? . For permanent removal of N, the whole plant (including the roots) should be harvested.