We have measured the effects of changing perilymphatic K+ by perfusing scala tympani in guinea pigs with salt solutions high or low in K+, while monitoring the distortion product otoacoustic emissions (DPOAEs) in the ear canal (a measure of mechanical vibration of the organ of Corti), the summating potential (SP) evoked by high-frequency tone-bursts (taken to be a measure of pre-synaptic electrical activity of the inner hair cells) and the compound action potential (CAP) of the auditory nerve (taken to be a measure of post-synaptic neural activity). We have attempted to investigate the osmotic effects of our perfusates by comparison with simple hyperosmotic sucrose perfusates and iso-osmotic versions of perfusates, and for the effects of changes in other ions (e.g. Na+ and Cl−) by keeping these constant in some perfusates while elevating K+. We have found that changing the K+ concentration over the range 0–30 mM elevated the SP and CAP thresholds almost equally in normal animals, and not at all in animals devoid of outer hair cells (OHCs), showing that OHCs are sensitive to the perfusates we have used, but the inner hair cells (IHCs) and the type I afferent dendrites are not, presumably because IHCs are shielded from perilymph by supporting cells, and the membranes of the afferent dendrite membranes exposed directly to our perfusates are dominated by Cl− permeability, rather than by K+ permeability. This view is supported by experiments in which the perilymphatic Cl− concentration was reduced, producing a large elevation in CAP threshold, but a much smaller elevation of SP threshold, suggesting disruption of action potential initiation. The view that threshold elevations with changes in perilymphatic K+ are due almost solely to a disruption of OHC function and a consequent change in the mechanical sensitivity of the organ of Corti was supported by measurements of amplitude of the 2f1-f2 distortion product otoacoustic emission. During elevations in K+, DPOAEs followed a similar time-course to that for SP and CAP, although the changes were less for DPOAEs. The lack of a 1:1 relationship between DPOAEs and SP and CAP is probably because the iso-input DPOAE measure used is a more complex indicator of mechanical sensitivity than the iso-output measure used by others. Taken together, these results suggest that changes in K+ in pathological conditions probably produce a hearing loss by disrupting OHCs rather than IHCs or neurones, and that OHC disruption in our experiments was due to a mixture of osmotic, K+ and possibly Cl− effects.