A low-speed wind tunnel study and flight tests were performed to examine the effects of a wing fence on the T-38A. Wind tunnel results were based upon force and moment data collected with a six-component balance and flow visualization at Reynolds numbers up to 0.3 x 106, based on mean aerodynamic chord. The model did not include the last 7.79 feet of the aircraft, and the engine and exhaust were modeled as through-holes. Five fence geometries, placed at wing station 125 ( 0.825 semispan), were compared. The best performer of these designs, based on drag polar, was the fence that wrapped the leading edge and extended 84.6 percent of the local chord length along the wing's upper surface. Wind tunnel data showed that this fence increased the lift coefficient by up to 6.3 0.6 percent and reduced spanwise and separated flow outboard the fence. The flight-tested fence was based on the best performing fence design from the wind tunnel study. The results were based on aircraft instrumentation and flow visualization at Reynolds numbers up to 9.98 x 106. It was inconclusive whether the fence caused an increase in lift coefficient. The fence reduced the roll-off tendency and wing rock during approaches to stall. Tuft visualization on the aircraft wing suggested that the fence reduced spanwise and separated flow outboard the fence, which agreed with the wind tunnel results.
