An airborne instrument for in situ measurements of tropospheric nitrogen dioxide (NO2) was developed using the photolytic conversion technique followed by chemiluminescence detection of NO. This instrument was used for the measurements of NO2 on board the NASA P-3B aircraft during the Transport and Chemical Evolution Over the Pacific (TRACE-P) campaign. Comparison in the laboratory indicated less than 10% difference between our NO2 instrument and two independent laser-induced fluorescence instruments in the NO2 range of 30 parts per trillion by volume (pptv) to 50 ppbv. The magnitudes of potential errors in airborne tropospheric NO2 measurements were further assessed using the TRACE-P data set. The systematic errors estimated for the median NO2 mixing ratios, 70 pptv at 0--2 km (30 pptv at 2--8 km), were 19% (39%). The random errors for a 10 s integration time were estimated to be 5--10%, depending on altitude. The observed NO2 mixing ratios were compared to those calculated by a photochemical box model. Overall, the calculated NO2 values correlated very well with those observed (r2 = 0.97), although the calculations were systematically higher than the observations by about 30%, except for the highest flight levels. The calculated/observed NO2 ratio remained nearly constant, having values close to 1.3 at 0--4 km, and decreased with altitude. The difference between the observed and model-calculated values, however, was within the combined uncertainty in the measurement and model calculation. The underlying causes for this difference are to be determined in future studies.