This was studied in 37 patients suffering from myofascial pain-dysfunction. Bipolar surface electrodes were used to record tonic, resting EMG activity for 2 consecutive trials. This protocol was repeated at a second recording session 2 weeks later. A custom-made, plastic template was constructed for each subject during the first visit in order to relocate the electrodes accurately at the second recording session. Correlation coefficients were calculated for each muscle area. Within-session (same day) r values ranged from 0.7620 to 0.8884 for the masseteric area and from 0.8686 to 0.9109 for the anterior temporal area. Across-session (different day) r values ranged from 0.5645 to 0.6503 for the masseteric area and 0.3309 to 0.4844 for the anterior temporal area. The lower correlation between different day recordings could reflect a methodological shortcoming in relocating the electrodes, particularly on the anterior temporal area. The greater variability recorded from the anterior temporal area could also reflect the dynamic role of the temporalis muscle in maintaining the postural rest position of the mandible.