A number of experiments in the past have demonstrated that rats and mice have shorter free-running circadian rhythms when they have access to a running wheel in their cage. Moreover, within groups of rats and hamsters, individuals making most use of their running wheels tend to have shorter circadian rhythms. However, these effects are not always evident. This article analyzes the results of four additional experiments on hamsters, some showing correlations between high activity and fast rhythms, and others not. It is suggested that failure to find this relationship occurs when there is an insufficient range of activity levels within a group. When present, correlations between locomotor activity and periodicity reflect causal links because shorter rhythms can be produced by providing a type of running wheel on which hamsters run more. The effects of possible changes in activity on circadian period should be considered when interpreting experiments on physiological manipulations of the circadian period.