The paper describes the use of selftuning concepts in the classical problem of extremum control. In particular, the situation is considered whereby a recursive parameter estimator is used to determine online the parameters of a performance index. The estimates are then used to decide the directions and step lengths employed in an online extremum-seeking function. The paper considers a self-tuning extremum controller for a specific one-dimensional performance index, and demonstrates its behaviour when used in simulation and in a practical situation drawn from the automotive industry.