In human, the amplitudes of specific event-related potential (ERP) components can increase or decrease in response to increasing stimulus intensity depending on the location of the recording site. Large increases characterize components presumably generated by modality-specific sites, while smaller increases or even decreases are associated with those originating in associational areas. Comparable data from non-human primates, which would permit invasive studies of the neural substrates underlying these intensity-amplitude differences, are limited. To more fully characterize these relationships, auditory ERPs were recorded from chronically implanted epidural electrodes in 5 squirrel monkeys (Saimiri sciureus) in response to tones (500 Hz, 300 msec duration) of varying intensities (50, 60, 70, 80 dB SPL). Squirrel monkey ERPs recorded at Fz exhibited 3 peaks during the 200 msec post-stimulus interval. These peaks included a positivity (P1), followed by a negativity (N1), and then another positivity (P2). At posterior sites, the frontal P1-N1 configuration was recorded as an N1-P1 complex. At these sites, a small negativity (N2) preceded the last positive peak (P2). Changes in polarity were independent of reference site and posterior N1-P1 peaks exhibited latencies similar to those of the frontal P1-N1 components. Amplitudes at Fz, Cz, and Pz increased substantially with increasing stimulus intensity ('augmenting'). In contrast, only small increases or even decreases in amplitude ('reducing') were evident at T3 and T4. On the other hand, peak latencies decreased with higher stimulus intensities at most sites. The site-specific amplitude responses exhibited considerable temporal stability. In one subject, for example, similar 'augmenting' profiles were recorded at Fz in 8 sessions over a 6-month period. The topography of monkey intensity-amplitude response profiles, their temporal stability, and peak latency shifts resemble observations made in humans. The data show that 'augmenting' characterizes monkey vertex potentials, which, like the analogous human potentials, may originate in primary auditory cortex. In contrast, potentials recorded over temporal cortex, which may originate in auditory association cortex, exhibit 'reducing.' Thus, the data support the hypothesis that differences in amplitude with increasing intensity may reflect differences in cortical origin.