Event-related potentials (ERPs) were recorded for ignored tones presented during the retention interval of a delayed serial recall task. The mismatch negativity (MMN) and N1 ERP components were measured to discern spatiotemporal and functional properties of their generation. A nine-token sequence with nine different tone pitches was more disruptive than an oddball (two-token) sequence, yet this oddball sequence was no more disruptive than a single repeating tone (one-token). Tones of the nine-token sequence elicited augmented N1 amplitudes compared to identical tones delivered in the one-token sequence, yet deviants elicited an additional component (MMN) with distinct temporal properties and topography. These results suggested that MMN and N1 are separate, functionally distinct components. Implications are discussed for the N1 hypothesis and the changing-state hypothesis of the disruption of serial recall performance by auditory distraction.