During the 1984 spring program of the NASA Global Tropospheric Experiment, measurements of NO, NO2, O3, and CO were made during two aircraft flights that encountered tropopause fold events. The technique used to measure NOx (NO+NO2) is described. In the neighborhood of both tropopause fold events, CO and O3 were strongly anticorrelated <Danielsen et al., 1987; Hipskind et al., 1987>. In one fold, encountered over the eastern Pacific Ocean at night, NOx was positively correlated with O3, negatively correlated with the dew point, and negatively correlated with CO, consistent with the stratospheric origin of the air mass. In this fold at altitudes between 5.5 and 6.0 km, NOx reached mixing ratios near 150 parts per trillion by volume (pptv), considerably larger than the levels of 10--50 pptv in the air external to the fold. In contrast, in a daytime encounter with a fold over the southwestern United States at altitudes of 6.7--8.8 km, NOx did not correlate well with CO, O3, or dew point, but remained roughly constant at levels of 100--200 pptv. Outside of the fold, in a region of strong convective activity, NOx peaked to just over 500 pptv. It appears that mixing of tropospheric air containing elevated NOx weakened the expected trends for the fold over the continent compared to the fold over the ocean, and therefore NOx may not always be a good tracer of air of recent stratospheric origin. However, the observations emphasize that the stratospheric source of NOx acts to introduce NOx over a short period and through a large vertical region of the troposphere, not just near the tropopause, as is assumed in many tropospheric models. ¿ American Geophysical Union 1988