Balloon-borne observations of concentrations of OH, HO2, H2O, and O3 in the middle and upper stratosphere are used to test our understanding of HOx photochemistry. Assuming our photochemical model is complete, the measured [OH> and [HO2> above 38 km (where HOx partitioning is no longer dependent on [NO>) are modeled best by calculations that use a 25% reduction in the ratio of the reaction rate constants for O+HO2 and O+OH as well as either a 25% reduction of the rate constant of OH+HO2 (the primary HOx sink) or a 25% increase in HOx production. All of these changes are consistent with the uncertainties in the recommended rate constants. The kinetic parameters required to explain our observations of [OH> and [HO2> do not lead to a resolution of the long-standing ozone deficit problem above 45 km. ¿ 1998 American Geophysical Union