Syncopation is a fundamental feature of rhythm in music. However, the relationship between theory and perception is currently not well understood. This thesis is concerned with characterising this relationship and identifying areas where the theory is incomplete. We start with a review of relevant musicological background and theory. Next, we use psychophysical data to characterise the perception of syncopation for simple rhythms. We then analyse the predictions of current theory using this data and identify strengths and weaknesses in the theory. We then introduce further psychophysical data which characterises the perception of syncopation for simple rhythms at different tempi. This leads to revised theory and a new model of syncopation that is tempo-dependent.