To arrive at detailed nutrient balances for duckweed-covered wastewater treatment systems, five laboratory-scale experiments were carried out in shallow (3.3 cm), 11 batch systems to assess separately the contributions of duckweed itself, attached and suspended bacteria as well as algae to N- and P-removal in domestic wastewater. Depending on the initial concentrations, our duckweed-covered systems removed 120-590 mg N m(-2) d(-1) (73-97% of the initial Kjeldahl-nitrogen) and 14-74 mg P m(-2) d(-1) (63-99% of the initial total phosphorus) in 3 days. Duckweed (Lemna gibba L.) itself was directly responsible for 30-47% of the total N-loss by uptake of ammonium and, probably dependent on the initial P-concentrations, for up to 52% of the total P-loss. The indirect contribution of duckweed to the total nutrient removal was also considerable and included the uptake (and adsorption) of ammonium and ortho-phosphate by algae and bacteria in the attached biofilm and the removal of N through nitrification/denitrification by bacteria attached to the duckweed. Together these accounted for 35-46 and 31-71% of the total N- and P-loss, respectively. Therefore, approximately 3/4 of the total N- and P-loss could be attributed to the duckweed mat. The remaining quarter is due to non-duckweed related components: uptake and nitrification/denitrification by algae and bacteria attached to the walls and the sediment of the system (including sedimentation). Other processes, like NH3-volatilisation, N-fixation and nutrient uptake as well as nitrification/denitrification by suspended microorganisms did not influence the N- and P-balance of our systems, but could become important with increasing water depths and retention times. (C) 1998 Elsevier Science Ltd. All rights reserved.