The effects and relative impacts of environmental variables on the behaviour of pesticides, through the effect on pesticide-degrading microorganisms, was studied in a broad spectrum and covered the most relevant emission routes. It is shown that the effect of landscape geochemistry, which is a pre-set condition in an agricultural management, may be significant (chapter 2). Adjoining soil types, which occur within short distance in an agricultural unit, were characterized and tested on their pesticide leachability and potential risk to groundwater. The use of a 14C radio-labelled pesticide enabled accurate study of its movement in soil layers. Distinctly different pesticide behaviour in topsoil and subsoil layers were observed within this soil series sequence. An alternative of conventional (i.e. uniform) pesticide application rates, which generally results in over-treating some sites and under-treating others, is site specific treatment. In such managements, doses are tailored to individual soil types or soil properties. This may be beneficial from both an economical as an environmental point of view. A prerequisite in such pest managements is the accuracy and reliability in which the leaching potentials of pesticides in various soils are predicted. Of much concern is the fact that there appears to be little agreement between the existing pesticide leaching models in predicting the residence time in soil and subsequent leaching to groundwater (chapter 3). This may be especially valid in soils that display preferent flow patterns, in which solutes by-pass the soil matrix without being subjected to significant sorption and degradation. It was concluded that the introduction of a variety of conceptual descriptions, thus covering ranges in specific soil properties, may improve the applicability of such models. From an environmental point of view, it is necessary to approximate concentrations not only in terms of magnitude but also at the time and the duration at which these occur. With conventional application ...