A diffusion denuder, total sulfur chemiluminescence detector instrument for the measurement of SO2 was tested as the second part of the Gas-Phase Sulfur Intercomparison Experiment (GASIE 2). The SO2 at mixing ratios between 27 and 182 parts per trillion by volume (pptv) was provided by a dynamic dilution apparatus. The data were kept blind from the other party and analyzed by two independent referees. The following was concluded: (1) The independent calibrations of each system are within a few percent. (2) The precision on any one day is better than day-to-day variability. (3) Runs in dry air show a small but significant nonzero intercept in correlation plots. (4) No effects of adding NO2+O3 or CO2+CH4+CO+dimethylsulfide are distinguishable. (5) A small but significant effect due to added H2O is evident in both slope and intercept, but the source could not be discerned. (6) On any one day the systems can distinguish among 0, 20, and 40 pptv, but because of day-to-day variability, they can only distinguish among 0, 30, and 60 pptv on different days. ¿ 2000 American Geophysical Union