Both sound (s-) and galvanic (g-) vestibular-evoked myogenic potential (VEMP) enable us to study the saccular pathways. However, the VEMP can be abnormal for non-vestibular factors, such as insufficient activation of the sterno-cleido-mastoid (SCM) muscle or a lesion that involves the accessory nucleus and/or nerve or the SCM muscle. These drawbacks do not affect another technique that evaluates the saccular function: the N3 potential. We recorded both the s- and the g-VEMP and the N3 potential in a group of 31 healthy subjects to establish a reference range. The N3 potential and the s-VEMP were recordable bilaterally from all the subjects, whereas the g-VEMP was undetectable uni- or bilaterally in 7 subjects. The latency and amplitude values of the s-VEMP did not differ from those of the g-VEMP. For all three techniques, the latency and amplitude values from the right and from the left recording and/or stimulation side were the same. We suggest using normative latency and amplitude values based on the mean and ratio of the right- and left-side values. The s-VEMP, the N3 potential and the auditory evoked response (ABR) were compared in 15 subjects suffering from multiple sclerosis. The three techniques detected a similar number of abnormalities, but these abnormalities were not correlated. This suggests that these different techniques should be regarded as complementary in evaluating saccular function.