Event dates are an important aspect of hydrological analysis, for questions like: �Are the timing of high flows, low flows, or ice cover changing over time?� Circular statistical techniques should be used to conduct analysis of this type of data rather than the methods widely applied to continuous data. The nature and background of circular or directional data are described, and analogues for classical statistical analysis are demonstrated using R and three synthetic samples. Methods for examining single samples and comparing two samples are given and discussed. A real world application demonstrates how these techniques can be applied to detecting trends in the timing of mixed populations of annual maximum floods.