Each spring a layer of small particles forms between 20 and 30 km altitude in the polar regions. We present the first self-consistent explanation of the observed CN layer from a 2D microphysical model of sulfate aerosol. Our theory relies on photolysis of H2SO4 and SO3, consistent with recent laboratory measurements, to produce SO2 in the upper stratosphere and mesosphere. An additional source of SO2 may be required. Nucleation throughout the polar winter extends the top of the aerosol layer to higher altitudes, despite strong downward transport of ambient air. This may affect heterogeneous chemistry at the top of the aerosol layer in polar winter and spring. ¿ 1999 American Geophysical Union