On June 15, 1999, about 1004 eastern daylight time, a White Velocity XL/RG, N19DW, registered to a private individual, operating as a Title 14 Part 91 personal flight, crashed landed on an interstate highway following a loss of engine power while attempting a forced landing at the St. Lucie County International Airport, Fort Pierce, Florida. Visual meteorological conditions prevailed and no flight plan was filed. The aircraft sustained substantial damage and the private-rated pilot, the sole occupant, was not injured. The flight originated from Hollywood, Florida, about 34 minutes before the accident. According to the pilot/builder, he was in cruise flight at 1,500 feet msl, about 10 miles southwest of the St. Lucie County Airport. He was maneuvering to attain a visual on his destination, a private airstrip. Without warning, he experienced a rapid decrease of engine power that activation of the electric fuel boost pump appeared to remedy. One or 2 minutes later, the engine again lost power and he noticed the fuel pressure gage indicating zero. He turned toward St. Lucie County Airport, but when it became obvious to him that engine power available would not sustain flight to the airport, he chose the highway for a forced landing. The aircraft collided with pole-mounted power transmission lines on approach to the highway about 4 miles west of the St. Lucie County Airport, that caused a hard touchdown to the pavement. He estimated that his fuel distribution at the beginning of the flight to be 5 to 10 gallons in the right tank, 10 to 12 gallons in the left tank, and 3 gallons in the sump tank. He further stated that he had recently replaced the factory-provided wing tank fuel caps with a different set of caps about a week before the accident. The new caps were not vented, and in retrospect, he stated the caps did not feel as tightly sealed to the filler pipe flange as the previous ones, but thought no more of it. He added that he began to have fuel balance problems that he had not experienced before, but did not make the connection. According to St. Lucie County FAA Control Tower personnel, at 1002 the pilot radioed that he was a possible emergency due to low fuel pressure and shortly thereafter, advised that he could not make the field due to low power and was landing on I-95. Subsequent examination of the wreckage and interview of the pilot/builder by FAA personnel revealed that the aircraft is configured with a 33-gallon fuel tank in each wing that gravity feeds into a fuselage mounted 3-gallon sump tank and thence to the aft-mounted engine via an electric driven boost pump. There is no fuel selector valve on the aircraft. The fuel quantity indication system consisted of graduations on a left and right sight-gage made of clear flex-tubing spliced inline with fuel feed lines between the wing tanks and the sump tank and routed into the cabin. Eight to nine gallons of uncontaminated 100LL aviation fuel was contained in the left wing/sump tank combination and no fuel was found in the right tank, postcrash. The wreckage was subsequently moved to county property where operational testing of the fuel system and engine was conducted. The engine started with no problems and exhaust sound was relatively smooth and consistent. High end rpm was not tested due to propeller damage and imbalance. The new fuel tank caps are shipped from the kit factory with adjustment instructions attached, but the pilot does not remember seeing them, (a copy of the instructions is included in this report). All three tanks are vented to the outside atmosphere, but inspection of the total vent system revealed some plumbing component's inside diameter measurements to be as small as .18 inch. Pressurizing either wing tank with air revealed that the fuel vent system could not adequately equalize the pressure, and air bypass at the loosest fitting fuel cap occurred. It is reasoned that the opposite would also take effect, i.e. a negative pressure as occurs above the wing during flight would also cause air bypass, (siphoning of air) out of the wing tank through the loose fitting cap. The net effect would be a low pressure or vacuum created in the wing fuel tank that would interfere with gravity feed into the sump tank and eventual fuel starvation. The factory provides, as an option to the basic airplane kit at additional cost, a "low fuel" warning light for the sump tank, but this airplane was not so equipped. As a result of this occurrence, flight testing was performed at the Velocity Aircraft Factory, Sebastian, Florida, using loose-fitting fuel tank caps in an attempt to duplicate the conditions present during the accident. It was revealed that the wing tanks did, in fact, stop their gravity feed and the sump tank quantity started decreasing when using the loose cap. As a result of the testing, the factory is experimenting with replacing the 1/4 inch fuel vent tubing with 3/8 inch tubing in the wing tanks. Additionally, the factory eliminated the sump fuel vent in favor of an air bleed-off valve or "burp" valve for use during pilot's preflight walk-around inspection. The factory has also issued an alert via newsletter to all kit owners concerning the hazards of using loose-fitting wing tank fuel caps. Applicable excerpts from that newsletter are included in this report.