In vitro models simulating tissues were used to test the validity of three methods for determining the free concentration of drugs and endogenous compounds in the extracellular space by means of microdialysis. Theophylline was used as a model substance. Recovery is defined as the proportion of compounds extracted from the medium surrounding the probe. The water recovery method is the use of recovery, determined in a water solution, to estimate concentrations in other media. This method was shown to underestimate the surrounding concentration when the effective rate of diffusion is smaller in the other medium than in water. The difference method measures the net transport over the dialysis membrane at varying concentrations in the perfusion medium. The point of equilibrium, where no net transport takes place, is used to estimate the surrounding concentration. The perfusion rate method involves two phases. In the first phase (calibration), surrounding media with different diffusion characteristics were used as tissue models. The amount recovered at different perfusion rates was measured and a multivariate regression method was used to calculate a mathematical model. In the second phase, the mathematical model was used to predict the concentration in the surrounding medium in new experiments. The two latter methods gave satisfactory predictions of the surrounding concentrations. Protein binding did not affect the methods. It is concluded that the difference method and the perfusion rate method may be used to estimate the in vivo concentration of drugs and endogenous compounds.