The rationale for the present study was to investigate several aspects of P300 topography in relation to aging and neuropsychological measures. We administered an auditory oddball ERP task to 72 participants aged 21.8 to 94.7 years, 36 males and 36 females, in addition to the Wechsler Abbreviated Scales of Intelligence (WASI) and digit span from the Wechsler Adult Intelligence Scales--Revised (WAIS-R). The relationship between age and P300 latency and amplitude at different electrodes was investigated, as well as the changes in the correlational pattern between P300 latency and amplitude with increasing age. A formal test of curvilinear relationships for P300 latency/amplitude and age was performed. Principal component factor analyses were performed for P300 latency and amplitude separately in order to check for possible superordinate structures in the distribution of the electrical activity measured at the scalp. In addition, each of the electrodes and each of the factors were correlated with the different neuropsychological measures, and the contribution of age to the observed relationships is discussed. The main conclusion drawn is that the activity generated from different brain areas change at different rates with age. While the posterior area shows a clear reduction of P300 amplitude and a delay of P300 latency with age, the amplitude does not decrease at the same rate in the fronto-central areas, and there is at the same time a marked hemispheric asymmetry in the age dependent change of activation. Based on our data, it may be concluded that a curvilinear expression generally does not explain the aging effect on the ERP component P300. Correlations between neuropsychological measures and P300 did, as expected, vary with area of activation, and the strongest correlations were generally found between matrices, block design and digit span, and the midline and left fronto-temporal electrodes. These relationships were in turn mediated by age. Implications of the findings are discussed.