This study examines the potential of removing arsenic from water by coprecipitation with naturally occurring iron. It focuses on the conditions found in the groundwater in Bangladesh where there is a need to develop a low-cost method of arsenic removal, which can be implemented at household level. The experimental study examined the sensitivity of removal of arsenic in response to manual mixing and prolonged settlement. It was demonstrated that up to 90% of the arsenic could be removed after 3 days' settlement, this being sufficient to attain the Bangladesh standard of arsenic concentration of 0.05 mg l−1 provided the Fe/As ratio and pH were favourable. The necessary iron (as Fe3+) to attain the standard was expressed by Fe=66 As1.75, where Fe and As are in mg l−1. The study also examined the effects of interference and found for typical levels of co-occurring solutes encountered in the groundwater in Bangladesh, that removal was enhanced. It was also apparent that the removal rate was sensitive to the form of the iron. There was evidence that adsorption may be the dominating trapping mechanism when Fe/As weight ratio was ≥10.