In this study we tested the hypothesis that claustrophobia is composed of two elements: fear of suffocation and fear of restriction. A self-report claustrophobia questionnaire, interview questions, and behavioural exposure tests were administered to an unselected sample of 179 university students. Results from each method of assessment supported the hypothesis, and suffocation and restriction fears were found to be moderately correlated. The basis of the correlation is considered and a number of explanations are set out. It is also suggested that these fears may be necessary but probably not sufficient for claustrophobia to occur, and that other factors, such as anxiety sensitivity, may play an important role.