The C major scale was used either as frequent or as infrequent stimulus in the oddball auditory evoked field measurement where the other stimulus was constructed by removing one the tones in the scale. Multivariate statistical analysis was employed to judge whether there was a significant difference between the responses to complete and incomplete scales in each subject for each `target' tone which was removed in an incomplete scale. Incomplete scales lacking, especially E, or B caused responses in both of the two oddball schemes but less significantly when used as frequent stimuli indicating that the complete major scale stored in the long term memory retained its influence as `reference' stimulus even when presented with a smaller probability.