The atmosphere of Io has been detected at infrared, millimeter, and ultraviolet wavelengths. The observations indicate a predominantly SO2 tenuous atmosphere on both the leading and trailing daysides, probably variable but permanently detectable. The observed SO2 pressure is on the order of 1 nanobar. There are several indications that the atmosphere is patchy rather than spatially uniform. The observable atmosphere may be produced from sublimation of SO2 frost or directly from volcanic output, while surface sputtering may produce a more tenuous atmospheric component in specific regions. The large horizontal pressure gradients probably drive a supersonic meteorology that affects the pressure distribution on both local and large scales and the sublimation/condensation exchanges. SO appears to be a significant minor component along with SO2, but its spatial distribution (global or localized) remains uncertain. Fundamental issues are unresolved, as the detailed interpretation of some observations is still ambiguous. The surface pressure and areal extent of the atmosphere are only known to within factors of similar to 10 and 5, respectively. The vertical structure is extremely poorly characterized. We do not know for sure whether Io's atmosphere is in hydrostatic equilibrium or if the gas is rapidly ejected by volcanic venting. The gas temperature is very difficult to constrain from observations, although detailed models are available. The dominant source of Io's atmosphere remains uncertain. Current data sets would suggest a preference for the concept of an atmosphere in dynamical equilibrium with volcanic sources, but this conclusion must be viewed as tentative. The existence, composition, thickness, and extent of the atmosphere on the nightside are unknown. (C) 1996 Academic Press, Inc.