The present study was undertaken to assess the utility of the theory of planned behaviour, using separate measures of the two variables that are considered to comprise the notion of perceived behavioural control, namely, beliefs concerning the controllability of the behaviour and efficacy expectancies. The study was concerned with the prediction of intentions to engage in regular exercise (for at least 20 minutes, three times a week for a fortnight) and actual exercise behaviour. A sample of 146 undergraduate subjects participated in the study. It was prospective in design; measures of attitudes, norms, intentions, perceived control and self-efficacy were obtained at the first wave of data collection, while actual behaviour was assessed two weeks later. The results of the study revealed support for the view that separate measures of self-efficacy and perceived behavioural control should be employed in the theory of planned behaviour. In the first instance, confirmatory factor analysis revealed that the two variables could be empirically distinguished. Second, the effects of perceived behavioural control and self-efficacy on behavioural intentions and actual behaviour differed. As predicted, efficacy expectancies influenced behavioural intentions, but not actual behaviour. In contrast, levels of perceived behavioural control had no effect on behavioural intentions, but emerged as a significant (positive) predictor of actual behaviour (there was also evidence that the effects of intentions on behaviour were moderated by the level of perceived behavioural control).