We determine the degree of denitrification that occurred during the 1996--1997 Arctic winter using a technique that is based on balloon and aircraft borne measurements of NOy, N2O, and CH4. The NOy/N2O relation can undergo significant change due to isentropic mixing of subsided vortex air masses with extravortex air due to the high nonlinearity of the relation. These transport related reductions in NOy can be difficult to distinguish from the effects of denitrification caused by sedimentation of condensed HNO3. In this study, high-altitude balloon measurements are used to define the properties of air masses that later descend in the polar vortex to altitudes sampled by the ER-2 aircraft (i.e., ~20 km) and mix isentropically with extravortex air. Observed correlations of CH4 and N2O are used to quantify the degree of subsidence and mixing for individual air masses. On the basis of these results the expected mixing ratio of NOy resulting from subsidence and mixing, defined here as NOy **, is calculated and compared with the measured mixing ratio of NOy. Values of NOy and NOy ** agree well during most parts of the flights. A slight deficit of NOy versus NOy ** is found only for a limited region during the ER-2 flight on April 26, 1997. This deficit is interpreted as indication for weak denitrification (~2--3 ppbv) in that air mass. The small degree of denitrification is consistent with the general synoptic-scale temperature history of the sampled air masses, which did not encounter temperatures below the frostpoint and had relatively brief encounters with temperatures below the nitric acid trihydrate equilibrium temperature. Much larger degrees of denitrification would have been inferred if mixing effects had been ignored, which is the traditional approach to diagnose denitrification. Our analysis emphasizes the importance of using other correlations of conserved species to be able to accurately interpret changes in the NOy/N2O relation with respect to denitrification. ¿ 1999 American Geophysical Union