During the fall 1997 North Atlantic Regional Experiment (NARE 97), two separate intercomparisons of aircraft-based carbon monoxide measurement instrumentation were conducted. On September 2, CO measurements were simultaneously made aboard the National Oceanic and Atmospheric Administration (NOAA) WP-3 by vacuum ultraviolet (VUV) fluorescence and by tunable diode laser absorption spectroscopy (TDLAS). On September 18, an intercomparison flight was conducted between two separate instruments, both employing the VUV fluorescence method, on the NOAA WP-3 and the U.K. Meteorological Office C-130 Hercules. The results indicate that both of the VUV fluorescence instruments and the TDLAS system are capable of measuring ambient CO accurately and precisely with no apparent interferences in 5 s. The accuracy of the measurements, based upon three independent calibration systems, is indicated by the agreement to within 11% with systematic offsets of less than 1 ppbv. In addition, one of the groups participated in the Measurement of Air Pollution From Satellite (MAPS) intercomparison [Novelli et al., 1998] with a different measurement technique but very similar calibration system, and agreed with the accepted analysis to within 5%. The precision of the measurements is indicated by the variability of the ratio of simultaneous measurements from the separate instruments. This variability is consistent with the estimated precisions of 1.5 ppbv and 2.2 ppbv for the 5 s average results of the C-130 and the WP-3 instruments, respectively, and indicates a precision of approximately 3.6% for the TDLAS instrument. The excellent agreement of the instruments in both intercomparisons demonstrates that significant interferences in the measurements are absent in air masses that ranged from 7 km in the midtroposphere to boundary layer conditions including subtropical marine air and continental outflow with embedded urban plumes. The intercomparison of the two VUV instruments that differed widely in their design indicates that the VUV fluorescence technique for CO measurements is not particularly sensitive to the details of its implementation. These intercomparisons help to establish the reliability of ambient CO measurements by the VUV fluorescence technique.