The duration of the late exteroceptive suppression period (ES2) of temporal muscle EMG activity has been reported to be reduced in patients suffering from chronic tension-type headache. Methods of recording and analysing ES2 have varied between centers and reproducibility of results within subjects, although insufficiently studied, has generally been poor. ES2 was investigated in 30 healthy subjects, using a computerized technique of recording, rectifying and averaging the EMG signals. Hour to hour and week to week variations of ES2 durations were calculated, and the influence of pain during a cold pressor test and of sustained muscle contraction on ES2 durations was investigated. The intra-individual variation of ES2 durations was 16.0% from hour to hour and 20.7% from week to week. The inter-individual variation was 36.7%. The present method for analysis of ES2 periods proved to be reliable, as the intra-observer variation was 4.2% and the inter-observer variation 4.6%. ES2 periods were significantly shorter on the first compared to the second day of examination (p = 0.006) and during experimental pain (p = 0.0005). We recommend the use of the computerized average technique in future studies and caution against the dependence of results upon factors such as conditioning and pain.