On September 12, 1998, at 1655 central daylight time (cdt), a Bell 206B-3, N52BH, piloted by an airline transport pilot, was substantially damaged during a forced landing approach following a partial loss of power and rotor RPM shortly after initiating a climb following a fire-fighting water pickup over a lake. Visual meteorological conditions prevailed at the time of the accident. The 14 CFR Part 91 fire fighting flight was not operating on a flight plan. The pilot reported no injuries. The flight departed a lake near Hinkley, Minnesota, at 1655 cdt. According to the pilot, "I had just filled the bucket and started to climb out. Everything was normal. As the aircraft started to enter translational lift, the engine started losing power and the rotor RPM started to drop below limits." The pilot said he dumped the water from the bucket and "...nursed..." the helicopter to the shoreline. He said the rotor RPM would diminish whenever he added power. The pilot said there was not a suitable landing area on the shore. "As I brought the aircraft forward, the tail rotor struck a steel fence post and snapped the tail rotor drive shaft." He said the helicopter began to spin and he entered auto rotation, touching down shortly afterwards. The on-scene investigation revealed a separated tail rotor drive shaft, the tail rotor blades had separated from their mounts with several pieces striking the helicopter's rear fuselage. A section of the tail rotor drive shaft had exited the helicopter through the engine cowling. The engine and airframe logbooks did not reveal any related discrepancies that had been repaired which would relate to the events surrounding the loss of rotor rpm. The mechanics and pilots who routinely worked on and flew the helicopter said they did not recall any situations that related to the accident. During the on-scene examination of the engine no anomalies were found that would prevent it from producing power. The engine was removed from the helicopter and shipped to the manufacturer. The engine was test-run on November 10 and 12, 1998. During the engine test run the engine started and accelerated normally, "...but was N2 limited, RPM started to drop upon attaining torque of 64.0 (psig)." A second test run showed the same results. The condition was isolated to the power turbine governor. The governor was removed and a new one was installed. The second engine run showed the engine operated to manufacturer specifications. The governor from the engine was inspected and test run. The differential of regulated pheneumatic air (Pr) and the governor air to the fuel control (Pg) was found to be 5.7-percent below the minimum setting. Upon inspection of the governor's components it was revealed that the cam follower lever's spring had worn radii at both ends and the spring hook holes in the spring retainer and cam follower lever were also worn. The worn areas were examined under magnification. The worn surfaces appeared irregular and featureless. Reports from the engine and governor manufacturers are appended to this report. According to the engine manufacturer, the governors are an on- condition item regarding maintenance and inspection. The manufacturer's representative said the governors have not had this type of problem because engine health checks performed in flight will show component wear. These checks are performed During the annual inspection, 100-hour inspection or the Manufacturer's approved maintenance inspection program. The engine health test flight on N52BH was conducted as part of its 100-hour inspection about 89-hours before the accident flight. A check of the Federal Aviation Administration's (FAA)Service Difficulty Reports (SDR) showed 8 reports that involved the governor creating a power loss. Specific reasons for the loss of power were not provided. The FAA's accident/incident data showed 3 events where the governor was shown as causal. The listings are appended to this report.