A systematic study for separating seven triazine pesticides was undertaken on the basis of micellar electrokinetic chromatography. The effects of five experimental factors (voltage, temperature and concentrations of buffer, sodium dodecylsulfate concentration, and acetonitrile) on the migration time, peak width, plate number and peak capacity were studied. Optimization was based on screening and a response surface designs. The study has shown that especially the peak width strongly varies in dependence on the experimental conditions. Therefore, common chromatographic measures for calculating the plate number and the peak capacity cannot be directly related to the experimental variables. Quantitative modelling of objective criteria, such as plate number and peak capacity on experimental factors is only possible if individual models are computed for each solute with respect to its migration time and peak width. On the basis of individual solute migration models reliable plate numbers and peak capacities were derived and used for evaluation of the optimum separation conditions. (C) 1999 Elsevier Science B.V. All rights reserved.