Response of a species to an environmental variable may be modeled and predicted using a wide spectrum of different functions. Contrary to other functions (Gaussian, polynomial etc), all parameters of the β-function are interpretable in ecological terms. However, computational difficulties in the determination of the β-function parameters initiated controversial debates on the applicability and usefulness of this function in vegetation modelling and gradient analysis. We propose a simple algorithm for fitting the β-function to observed data. Analytic properties of the algorithm (its ability to recover the known species responses along gradients) are tested using a series of simulated data. In most cases the algorithm correctly estimated parameters of the simulated responses.