The aim was to assess the treatment efficiencies of experimental storm water detention (extended storage) systems based on the Atlantis Water Management Limited detention cells receiving concentrated runoff that has been primarily treated by filtration with different inert aggregates. Randomly collected gully pot liquor was used in stead of road runoff. To test for a 'worst case scenario', the experimental system received higher volumes and pollutant concentrations in comparison to real detention systems under real (frequently longer but diluted) runoff events. Gravel (6 and 20 mm), sand (1.5 mm), Ecosoil (inert 2 mm aggregate provided by Atlantis Water Management Limited), block paving and turf were tested in terms of their influence on the water quality. Concentrations of five-day at 20 degrees C ATU biochemical oxygen demand (BOD) in contrast to suspended solids (SS) were frequently reduced to below international secondary wastewater treatment standards. The denitrification process was not completed. This resulted in higher outflow than inflow nitrate-nitrogen concentrations. An analysis of variance indicated that some systems were similar in terms of most of their treatment performance variables including BOD and SS. It follows that there is no advantage in using additional aggregates with high adsorption capacities in the primary treatment stage.