The relation between the mean all-shift concentration of xylene vapor in the atmosphere and the amount of conjugated toluic acids excreted in urine of exposed persons, was derived from the results of 38 experiments with pure xylene isomers and with a mixture of isomers at ratio 1:1:1. It was proved by statistical processing of experimental data that the decisiveness of the exposure test depends on the way of expressing the excreted amount of metabolites. The data of the concentration (mg metabolite/1 1 l of urine) show the largest dispersion of results. Least variable are the results obtained when using the data of the amount of the metabolite excreted within a defined period and recalculated to 1 kg weight of the exposed person. It is possible to use either the actual weights of the persons for the recalculation, or so-called ideal weights, the numeric value of which equals the body height (in cm) decreased by 100. The other methods (concentration data recalculated to standard density of urine 1.024 g/ml, data recalculated to 1 g of excreted creatinine, data on the amount of metabolite excreted within a defined period) yield results, the reliability of which lies between the margin values expressed in mg/l and in mg/weight. It has been proved by studies of the course of toluic acid excretion that short-term samples of urine from the last 2 h of the shift are not suitable for estimation of the level of exposure, if the exposure is interrupted or if the vapor concentration in the atmosphere changes. In such cases it is necessary to use all-shift, or better all-day urine samples. The biologic limits corresponding to an 8-h exposure to a concentration of xylene vapors 200 μg/l (= Czechoslovak MAC) were determined for various urine samples (short-term, all-shift, all-day) as well as for different data on the excreted amount of metabolite.