This thesis describes the utilisation of a little-known physiological phenomena (otoacoustic emissions) as a biometric. The role of the human auditory system as the sound to nerve impulse transducer is well known and extensively studied. Less well known is the fact that this system can also generate sounds; these are called 'otoacoustic emissions'. Several distinct types of emission are known: of these, one particular type, 'transient otoacoustic emissions' (TEOAEs) show potential as a biometric. TEOAEs are sounds recorded in the ear canal as the product of an active process in the inner ear which was stimulated. Otoacoustic emissions are analysed from the perspective of what constitutes the fundamental properties a characteristic must posses to be a good biometric and the underlying mathematics are described. Transient OAEs (TEOAEs) are evaluated against these criteria and using linear analysis demonstrate a predicted equal error rate (EER) of 0.182% with a dataset of 561 subjects, confirmed by a similar result on a smaller test dataset of 23 individuals. Permanence of 6 months is demonstrated (with difficulty due to the nature of the current OAE acquisition systems) however stability over several years is suggested by examples from medical literature. A novel acquisition method has been developed with a view to solving the two main problems with current TEOAE acquisition systems: The acceptability problem of having to insert a probe into the ear canal (which is perceived as unnatural and unhygienic) and the extreme sensitivity of the TEOAE to the fit of the probe in the ear canal. Finally a survey of biometrics places TEOAE in context within the field.