On September 28, 1997, at 1730 central daylight time, a Piper PA-28-181, N35652, lost engine power and executed an emergency landing in Fulton, Mississippi. The airplane was operated by the pilot under the provisions of Title 14 CFR Part 91 and visual flight rules. A visual flight rules flight plan was filed, but not activated, for this personal flight. The private pilot and his passenger both received minor injuries, and the airplane was substantially damaged. The flight originated for Waycross, Georgia at 1405 eastern daylight time. According to the pilot, he planned a long flight from Waycross, Georgia to Hamilton, Alabama. Although the manual said the engine would consume 8.8 gallons per hour (gph), the pilot planned his fuel consumption using 10 gph, allowing for a margin of error. The airplane had full tanks upon departing from Waycross. Although the pilot stated he tried to open his flight plan, the pilot was unable to establish communication with the flight service station. Upon arriving at Hamilton, the pilot found that no one was there to fuel the airplane. After considering his fuel remaining, the pilot attempted to fly to Tupelo, Mississippi to refuel the airplane. It was an additional 30 miles. About 10 minutes after leaving Hamilton, the engine sputtered and quit. The pilot used the restart procedure and was able to restart the airplane, only to have it sputter and quit again. He made an emergency landing, approximately 18 miles outside of Tupelo. The FAA inspector indicated that he confirmed that the airplane had full tanks upon departure from Waycross. He also indicated that after completing fuel consumption calculations the airplane would not have run out of fuel. After a cursory examination of the engine, the carburetor was sent to Precision Airmotive Corporation, Everett, Washington, for inspection. It was determined during a teardown that the carburetor was sending a rich fuel/air mixture to the engine. The economizer screw was one full turn rich. There was also a bench test done on the carburetor in which five different settings were simulated. By changing the airflow and manifold pressure through the carburetor, it was possible to simulate five throttle settings from idle to full throttle. While bench testing the carburetor, it was noted that when the airflow was set at 400, 700, and 900 lbs/hr, the carburetor was above the rich limits. (see attached report) According to Precision Airmotive Corporation, these airflow settings coincide with the cruise settings in most airplanes.