The stability of light gases in various hypothetical Titan atmospheres is examined. Only tiny amounts are free of blowoff unless the mean mass is greater than 4 atomic units. In a mixture of H2 (or He) with a heavier gas, a diffusive upward flux, due to buoyancy, must exist. Its value is insensitive to any parameters except the mixing ratio f1 of light to heavy gas, and is of order 2×1011 cm−2 sec−1 for fl=0.1. Normally, fl is independent of height except at great altitudes, and the extent of mechanical mixing of the atmosphere is therefore nearly irrelevant. The structure of the high atmosphere adjusts itself to accommodate the buoyant flux; the exospheric temperature affects this structure, but does not control the escape rate. The heavy gas is most probably N2 or CH4, (or both), and several existing models of this type are discussed. All seem acceptable as long as there is a large source of H2 near or below the surface. Photolysis of NH3 seems adequate, but only if solar ultraviolet penetrates... Abstract The stability of light gases in various hypothetical Titan atmospheres is examined. Only tiny amounts are free of blowoff unless the mean mass is greater than 4 atomic units. In a mixture of H2 (or He) with a heavier gas, a diffusive upward flux, due to buoyancy, must exist. Its value is insensitive to any parameters except the mixing ratio f1 of light to heavy gas, and is of order 2×1011 cm−2 sec−1 for fl=0.1. Normally, fl is independent of height except at great altitudes, and the extent of mechanical mixing of the atmosphere is therefore nearly irrelevant. The structure of the high atmosphere adjusts itself to accommodate the buoyant flux; the exospheric temperature affects this structure, but does not control the escape rate. The heavy gas is most probably N2 or CH4, (or both), and several existing models of this type are discussed. All seem acceptable as long as there is a large source of H2 near or below the surface. Photolysis of NH3 seems adequate, but only if solar ultraviolet penetrates...