This thesis takes a fresh approach to cochlear mechanics. Over the last quarter of a century, we have learnt that the cochlea is active and highly tuned, observations suggesting that something may be resonating. Rather than accepting the standard traveling wave interpretation, here I investigate whether a resonance theory of some kind can be applied to this remarkable behaviour.¶ A historical survey of resonance theories is first conducted, and advantages and drawbacks examined. A corresponding look at the traveling wave theory includes a listing of its short-comings.¶ A new model of the cochlea is put forward that exhibits inherently high tuning. The surface acoustic wave (SAW) model suggests that the three rows of outer hair cells (OHCs) interact in a similar way to the interdigital transducers of an electronic SAW device. Analytic equations are developed to describe the conjectured interactions between rows of active OHCs in which each cell is treated as a point source of expanding wavefronts. Motion of a cell launches a wave that is sensed by the stereocilia of neighbouring cells, producing positive feedback. Numerical calculations confirm that this arrangement provides sharp tuning when the feedback gain is set just below oscillation threshold.¶ A major requirement of the SAW model is that the waves carrying the feedback have slow speed (5-200 mm/s) and high dispersion. A wave type with the required properties is identified - a symmetric Lloyd-Redwood wave (or squirting wave) - and the physical properties of the organ of Corti are shown to well match those required by theory.¶ The squirting wave mechanism may provide a second filter for a primary traveling wave stimulus, or stand-alone tuning in a pure resonance model. In both, cyclic activity of squirting waves leads to standing waves, and this provides a physical rendering of the cochlear amplifier. In keeping with pure resonance, this thesis proposes that OHCs react to the fast pressure wave rather than to bending of stereocilia induced by a traveling wave. Investigation of literature on OHC ultrastructure reveals anatomical features consistent with them being pressure detectors: they possess a cuticular pore (a small compliant spot in an otherwise rigid cell body) and a spherical body within (Hensens body) that could be compressible. I conclude that OHCs are dual detectors, sensing displacement at high intensities and pressure at low. Thus, the conventional traveling wave could operate at high levels and resonance at levels dominated by the cochlear amplifier. ¶ The latter picture accords with the description due to Gold (1987) that the cochlea is an ‘underwater piano’ - a bank of strings that are highly tuned despite immersion in liquid.¶ An autocorrelation analysis of the distinctive outer hair cell geometry shows trends that support the SAW model. In particular, it explains why maximum distortion occurs at a ratio of the two primaries of about 1.2. This ratio also produces near-integer ratios in certain hair-cell alignments, suggesting that music may have a cochlear basis.¶ The thesis concludes with an evaluation and proposals to experimentally test its validity.