HISTORY OF FLIGHT On December 7, 1996, about 0152 hours Pacific standard time, a Beech 95-C55, N66MB, operated by Arnold and Arnold of San Diego, California, was destroyed during a landing approach at March Air Force Base, Riverside, California. The pilot received serious injuries. The personal flight originated at San Diego at 0125 after a round trip to Oakland, California. The pilot was reporting for weekend National Guard duty at March AFB. The pilot contacted the Southern California Terminal Radar Approach Control (SOCAL) for VFR advisories en route to March AFB. Approach control contacted the March Air Traffic Control Tower (ATCT) and advised them that Baron N66MB was inbound from 25 miles southeast, and requested ATCT verify that runway 32 was the active, and requested the weather. SOCAL also reported that the pilot had the airport in sight and wanted to continue the approach to check the weather. ATCT advised SOCAL that the weather was 300 feet overcast with 1/8 mile visibility in fog. The pilot was handed off to March ATCT. He contacted the ATCT and stated that he had the airport in sight, but admitted that he could not see the control tower. The pilot was given a visual clearance to land. The ATCT had no visual contact with the aircraft, but did have the aircraft on the Digital Bright Radar Indicator (D-Brite) to a point 3/4 mile east of the runway. The pilot requested the runway visual range (RVR), which was not available at that time. The airport surveillance radar (ASR-9) was out of service. All lighting systems were reported to be operational. According to the audio tape obtained from the ATCT, the pilot questioned the controller regarding the legality of the approach under the conditions reported. The controller responded/replied that he had no problem with the approach and landing. The last radio contact with the pilot was 1 mile out on final approach for runway 32. The ATCT made several calls to the aircraft with no response. When the aircraft failed to arrive at the airport or declare a missed approach or a go-around, a search was started. The aircraft was not found until 0647 due to the dense ground fog. The wreckage was located between taxi way "Charlie" and taxi way "Bravo" about 200 feet northeast beyond the approach end of runway 32. The pilot was interviewed by telephone shortly after the accident. He reported that during the go-around/missed approach both engines quit within seconds of one another. He further reported that when they quit he changed tanks and hit the boost pumps. A weather observation taken by the control tower about an hour before the accident reported 1/8 mile in fog with less than 1,000 feet of runway visual range. PERSONNEL INFORMATION The pilot was FAA certificated as an airline transport pilot and reported a total flight time of 15,481 flight hours with 1,575 hours in the make and model of the accident aircraft. He was type certificated in the Boeing 707 and 720 aircraft as a military pilot. During a postaccident telephone conversation, the pilot stated that his day started about 16 hours before the accident, about 1000 on the morning of December 6, 1997. He flew from San Diego to Oakland, attended a dinner, flew back to San Diego to drop off his passengers, and then flew to Riverside. AIRCRAFT INFORMATION According to maintenance records, the last annual inspection was accomplished on February 19, 1996, at a total airframe time of 3,467.7 hours. On August 11, 1994, both engines were replaced with factory remanufactured engines. At the time of the accident both engines accumulated about 217 hours since remanufacture, and the airframe had accumulated 3,569.8 total hours. The main fuel tanks hold 40 gallons each with 37 usable. The wing auxiliary tanks hold 31 gallons each with 31 usable, for a combined total of 136 gallons of usable fuel. The fuel system is placarded "use aux tanks and cross feed in level flight only." The pilot provided a recap of the flight times, as best he could recall, from Oakland to San Diego to Riverside. The flight included two takeoffs, and a total flight time of 2 hours 24 minutes. The aircraft departed Oakland with full fuel. He verbally stated that at 65 percent power he would typically burn 27 gallons per hour, or 65 gallons for that trip. According to data obtained by the Safety Board from the GPS that was installed in the aircraft, the flight had taken about 2 hours 46 minutes and used about 75 gallons of fuel using the pilot's consumption figures. METEOROLOGICAL INFORMATION At the time of the accident the Metar surface weather observation system was reporting: wind at 060 degrees at 2 knots, visibility 1/8 statute mile in fog; runway 32 visual range less than 1,000 feet; vertical visibility 200 feet; temperature 32 degrees Fahrenheit; dew point 32 degrees Fahrenheit; and the altimeter was 30.12 inHg. WRECKAGE AND IMPACT INFORMATION The wreckage was not examined at the accident site by the Safety Board. An FAA inspector viewed the accident site and wreckage prior to it's removal to a storage facility. According to the FAA inspector, the main fuel tank selectors were found selected to the main tank positions. The fuel quantity selector switch was found selected to the auxiliary fuel tank position. According to the aircraft retriever, during disassembly it was noted that the auxiliary fuel tanks were found to be empty. The right main fuel tank had an estimated 22 gallons of fuel that the retriever removed during disassembly. The left fuel tank was damaged and near empty. The FAA inspector did not check the fuel tanks, however, he noted that there was a substantial fuel spill on the ground. On March 12, 1997, the fuel system was examined in detail at the storage facility. The fuel system plumbing, venting, and cell interior were examined on each tank. The main fuel cells were reservoir equipped. The left reservoir zipper was found open. Both reservoir flapper doors worked properly. The plumbing and venting was intact and clear of obstructions. TESTS AND RESEARCH INFORMATION Both engines were shipped to Continental Motors at Mobile, Alabama, for examination and testing under the supervision of the Safety Board. On February 27, 1997, the engines were installed into test cells for running and recorded in a Continental engine test log. The engines were dye-checked at the crankshaft flange radius prior to running. Each had a fracture in the flange radius. The right engine was run for 15 minutes to a maximum of 2,680 rpm. The left engine was run for 20 minutes to a maximum of 2,716 rpm. Each engine met the Continental test requirements for proper fuel flow, ignition, and general operation. The magneto switch wiring was traced and examined from each engine's magneto to the switch panel. There were no abrasions or grounding of the separate magneto wire systems found. ADDITIONAL INFORMATION On March 19, 1997, the wreckage was released to the insurance company representative.