An axisymmetric, 110 N class, rocket configured with a free expansion between the rocket nozzle and a surrounding duct was tested in an altitude simulation facility. The propellants were gaseous hydrogen and gaseous oxygen and the hardware consisted of a heat sink type copper rocket firing through copper ducts of various diameters and lengths. A secondary flow of nitrogen was introduced at the blind end of the duct to mix with the primary rocket mass flow in the duct. This flow was in the range of 0 to 10% of the primary massflow and its effect on nozzle performance was measured. The random measurement errors on thrust and massflow were within ]/-1%. One dimensional equilibrium calculations were used to establish the possible theoretical performance of these rocket-in-a-duct nozzles. Although the scale of these tests was small, they simulated the relevant flow expansion physics at a modest experimental cost. Test results indicated that lower performance was obtained at higher free expansion area ratios and longer ducts, while, higher performance was obtained with the addition of secondary flow. There was a discernable peak in specific impulse efficiency at 4% secondary flow. The small scale of these tests resulted in low performance efficiencies, but prior numerical modeling of larger rocket-in-a-duct engines predicted performance that was comparable to that of optimized rocket nozzles. This remains to be proven in large-scale, rocket-in-a-duct tests.Schneider, Steven J. and Reed, Brian D.Glenn Research CenterROCKET NOZZLES; ROCKET-BASED COMBINED-CYCLE ENGINES; MATHEMATICAL MODELS; RELIABILITY ANALYSIS; DUCTS; GASEOUS ROCKET PROPELLANTS; ALTITUDE SIMULATION; HYDROGEN; OXYGEN; SECONDARY FLOW; DATA REDUCTION; SPECIFIC IMPULSE; ROCKET THRUST; NOZZLE FLOW; PROPULSIVE EFFICIENCY