We address the previously unresolved puzzle of nitric acid formation in the polar winter mid to upper stratosphere, first indicated by Limb Infrared Monitor of the Stratosphere observations in the Arctic winter of 1978--1979. Several theoretical studies over the past 2 decades have tried to reproduce these observations with varying success. More recently, the onset, altitude range, and duration of the formation process have been clarified by the Cryogenic Limb Array Etalon Spectrometer onboard UARS and by a series of ground-based observations taken at the South Pole during the Antarctic winters of 1993, 1995, and 1999. Using the Stony Brook-SSt. Petersburg two-dimensional photochemical model, we have reexplored HNO3 formation via both the ion cluster chemistry and heterogeneous chemistry on sulfate aerosols considered by earlier investigators. By including what we believe to be a realistic flux of NOy from the mesosphere, we find that the model can generate observed mixing ratios through a combination of ion-cluster-enhanced chemistry in the upper to mid stratosphere, augmented by heterogeneous chemistry on sulfate aerosol below ~40 km. Results are presented which clarify the relative role of various processes and assumed NOy fluxes. These also point up the need to incorporate more accurate downward NOy fluxes in models being used to simulate the polar stratosphere. Finally, we emphasize the need to consider the influence of repartitioning of NOy or NOx into HNO3 before observed variations in amounts of the former reaching the mid to lower stratosphere in winter and early spring can properly be used as tracers to reflect variations in thermospheric-mesospheric NOy production or transport. ¿ 2001 American Geophysical Union