Previous studies have shown that the vestibular short-latency-evoked potential (VsEP) in response to the brief head acceleration stimulus is a compound action potential of neurons innervating the otolith organs. However, due to the lack of direct evidence, it is currently unclear whether the VsEP is primarily generated by the activity of utricular or saccular afferent neurons, or some mixture of the two. Here, we investigated the origin of the VsEP evoked by brief bone-conducted vibration pulses in guinea pigs, using selective destruction of the cochlea, semicircular canals (SCCs), saccule, or utricle, along with neural blockade with tetrodotoxin (TTX) application, and mechanical displacements of the surgically exposed utricular macula. To access each end organ, either a dorsal or a ventral surgical approach was used. TTX application abolished the VsEP, supporting the neurogenic origin of the response. Selective cochlear, SCCs, or saccular destruction had no significant effect on VsEP amplitude, whereas utricular destruction abolished the VsEP completely. Displacement of the utricular membrane changed the VsEP amplitude in a non-monotonic fashion. These results suggest that the VsEP evoked by BCV in guinea pigs represents almost entirely a utricular response. Furthermore, it suggests that displacements of the utricular macula may alter its response to bone-conduction stimuli.