Balloon-borne in situ measurements of total reactive nitrogen (NOy), nitric oxide (NO), nitric acid (HNO3), ozone (O3), and aerosols were conducted over Aire sur l'Adour, France (44 ¿N), between October 11 and 12, 1994. NOy and NO were measured by chemiluminescence NO detectors combined with a gold catalytic converter, and HNO3 was measured by ion mass spectrometry. Two independent box models (Goddard Space Flight Center and Atmospheric and Environment Inc.) run on isentropic trajectories were used to interpret the observed partitioning of reactive nitrogen. The NOy partitioning calculated by the two models agreed within 10% of each other. The model-calculated NO/NOy ratios increase by ~25% when the updated reaction rates of NO2+OH, HNO3+OH, and NO2+O <Portmann et al., 1999> were used instead of those given by DeMore et al. <1997>, leading to a better agreement with the observations by the same amount. At 21--25 km the calculated NO/NOy ratios depended sensitively on the estimate of aerosol surface area. Within the uncertainty of the estimate of aerosol surface area the NO/NOy and HNO3/NOy ratios calculated by the models reproduced those observed within 20%. The same trajectory models also reproduced the NO/NOy ratios measured below 25 km under high aerosol loading in October and November 1992 within ¿20%, which is the accuracy level of the measurements. ¿ 2000 American Geophysical Union