The NOAA Twin Otter conducted more than a dozen overflights of ground-level air quality monitoring stations during the 1995 Southern Oxidants Study (SOS) Nashville/Middle Tennessee Ozone Project Field Intensive. Surface and aircraft observations of ozone and ozone precursors were examined to identify systematic sampling errors, and to assess the degree to which surface measurements may be considered representative of the larger planetary boundary layer (PBL). Overall agreement between surface and aircraft trace gas measurements was excellent in the well developed mixed layer, especially in rural-regional background air and under stagnant conditions, where surface concentrations change only slowly. On July 2, surface level measurements were representative of the larger mixed layer over distances as far as 70 km in background air, and 30 km in the weakly advected urban plume. Vertical variations in trace gas concentrations were often minimal in the well-mixed PBL, and measurements at the surface always agreed well with aircraft observations up to 460 m above ground level. Under conditions of rapidly varying surface concentrations (e.g., during episodes of power plant plume fumigation and early morning boundary layer development), agreement between surface and aloft is dependent upon the spatial (aircraft) and temporal (ground) averaging intervals used in the comparison. Under these conditions, surface sites may be representative of the PBL only to within a few kilometers. Under clear skies in the well-mixed PBL, regression of aircraft trace gas data collected within 5 km of the ground sites against 15--20 min average surface concentrations centered on the times of the overflights yielded the following relationships: 3>aircraft=(3>surface¿0.9374)+4.86 ppbv(r2=0.9642); aircraft=(surface¿0.8914)+16.4 ppbv(r2=0.9673); 2>aircraft=(2>surface¿0.9414)-0.069 ppbv(r2=0.9945). Although O3, CO, and SO2 measurements at the surface and aloft generally agreed well, isolated examples of unexplained measurement discrepancies emerged, illustrating the need for side-by-side instrument comparisons. NOY measurements agreed poorly between surface and aircraft: Y>aircraft=(Y>surface¿0.9184)+4.56 ppbv(r2=0.9188).