Effects of different electrode placements and indifferent electrodes were investigated for the vestibular evoked myogenic potential (VEMP) recorded from the sternocleidomastoid muscle (SCM). In 5 normal volunteers, the motor point of the left SCM was identified and an electrode placed there. A grid of 7 additional electrodes was laid out, along and across the SCM, based upon the location of the motor point. One reference electrode was placed over the sternoclavicular joint and another over C7. There were clear morphological changes with differing recording sites and for the two reference electrodes, but the earliest and largest responses were recorded from the motor point. The C7 reference affected the level of rectified EMG and was associated with an initial negativity in some electrodes. The latencies of the p13 potentials increased with distance from the motor point but the n23 latencies did not. Thus the p13 potential behaved as a travelling wave whereas the n23 behaved as a standing wave. The C7 reference may be contaminated by other evoked myogenic activity. Ideally recordings should be made with an active electrode over the motor point.