The study sought to determine whether, as implied by a trait interpretation of individual differences in non-specific electrodermal responding, the frequency of NSRs is a consistent characteristic of subjects across conditions. 40 male subjects completed two 1-hour sessions spaced at an interval of 2 weeks. In each session, three 12-min conditions were administered: relaxation, vigilance, and mental arithmetic. Continuous recordings were made throughout each condition of electrodermal and respiratory activity, and at the end of each condition subjects completed an Adjective Checklist (ACL). Counts were made of the total number of 10-sec intervals during each condition in which at least one NSR was recorded. Separate counts were made for large amplitude (equal to or greater than 1% of baseline) and for small amplitude (less than 1%) responses. Analyses of variance performed on the NSR data indicated that the major component of variance was that due to subjects, and that the interactions between subjects and conditions and subjects and sessions, which would contradict a trait interpretation, were negligible. Consistency in subject's responding was estimated by intraclass correlations across conditions and sessions. For NSR large, the coefficient was 0.58 and for NSR small 0.50. The data were interpreted as support for a trait interpretation of NSR frequency.