This study (1) explored the feasibility of using electrodermal activity (EDA) to characterize the arousal and sensory reactivity of children with high functioning autism (HFA) and Asperger's Syndrome (AS), (2) determined the reliability of electrodermal measures and (3) described the variability of EDA in this sample. Forty children with HFA and AS participated. All participants received a diagnostic psychological assessment and a physiological evaluation. Fourteen participated in the retest study on the physiological measures. Results indicated psychophysiologic testing was feasible with this sample. Seventy-three percent of the variables had reliability coefficients greater than 0.33, with a median variable reliability of 0.45. No significant differences were detected between HFA and AS groups. Visual inspection of skin conductance level (SCL) suggested two patterns: (1) high SCL (high arousal), with higher EDA magnitudes, faster latencies and slower habituation and (2) low SCL (low arousal), with lower EDA magnitudes, slower latencies and faster habituation. The presence of two EDA patterns applied equally when medications were eliminated. The previous inconsistency in studies of EDA in children with HFA and AS may be due to the presence of a high arousal groups and a low arousal group. Hence, this population should not be assumed to be homogeneous. (Contains 2 tables and 2 figures.)