A theoretical framework for analysing control implementation structures and processes is discussed. The framework is applied to a comparative study of a control strategy implemented in Germany and one implemented in Sweden. The differences between these control strategies are explained. In Germany, legislation was used to combat single-source, measurable emissions. Environmental problems that are identifiable and measurable lend themselves to such detailed legislation. However, this is not the case with newly defined environmental problems such as diffuse emissions. Thus, in Sweden, openly formulated directives were used to combat diffuse, hard-to-measure emissions. This illustrates the recent tendency in Swedish environmental control, which is a change from centralized control through regulation to decentralization of environmental control through reforms.