SummaryThe persistence and movement of residues of alachlor, alrazine and metribuzin were measured in a mini‐lysimeter system in the field. This comprised a number of soil columns (11 cm diametert; 30 cm long), and permitted the vertical distribution of residues to be determined at. intervals alter application and the collection and analysis of leaehale water. Laboratory experiments were also performed to determine the degradation rates of the three herbicides and their strengths of adsorption by the test soil. The results showed an order of degradation rate of metribuzin>alachlor>atrazine and an order of adsorption of alacblor>atrazine>melribuzin. Movement of residues in the soil columns and concentrations in the leachate were inversely related to the strength of adsorption. Parameters derived from the laboratory data were used in conjunction with weather data for the period of the field experiment in three mathematical models of pesticide leaching: VARLEACH, LEACHP and PRZM2. In most instances, the models gave acceptable predictions of the distribution of residues in soil. This was particularly so for the less mobile compound alachlor. With the most mobile compound, metribuzin, residues were not well predicted at the later sampling dates. All three models gave accurate predictions of the volumes of drainage water, but none of them predicted the concentrations of herbicide in the leachate, presumably because they do not take account of preferential flow pathways of water and solute in the soi