The logistics, expense and inefficiency of doing a pilot EEG/sleep study at a remote site not equipped with formal EEG/sleep facilities are significantly improved by using ambulatory EEG equipment and, under specific circumstances, the Internet Systems to transfer the data for printout and analysis. The task was to determine the sleep onset times of 6 subjects under 3 different conditions; 0.0 (placebo), 0.3, and 1.0 mg of melatonin. Recording for 2 h after going to bed was accomplished using an 18-channel ambulatory EEG system modified to record the various sleep parameters. 13 Mbytes of data were generated per hour per subject. To condense the study, 3 subjects were tested per day, twice per week. At the end of each recording session, the data were transferred (< 5 min) from the individual recording systems onto the local network computer (MIT). The data were transferred over the Internet from MIT to the BIH (< 20 min). The data were then replayed (< 60 min) onto a laser printer at 30 sec/page for formal reading. During the 18 recording sessions, no data were lost and we were able to efficiently complete a pilot study using existing resources at remote sites without the expense of duplication and awkward logistics.