Two annular denuder samplers were evaluated for the determi-nation of atmospheric nitric acid, SO2, and ” ne particulate sulfate and nitrate. One, the University Research Glassware (URG) sam-pler, used an elutriator-impactor to remove coarse particles prior to sampling the aerosol through the denuder sections for the re-moval of acid gases. The second, the Rupprecht and Patashnick (R&P) ChemSpec, removes the acid gases in diffusion denuder sec-tions and then removes coarse particles prior to the collection of ” ne particles. Possible sampling artifacts examined included the removal of HNO3(g) by the inlet of the URG sampler before the collection of this gas-phase species by the diffusion denuder, re-sulting in the underestimation of HNO3(g) and the deposition of some coarse particles in the R&P denuders prior to the removal of these larger particles, and the collection of ” ne particles on a ” lter, resulting in the overestimation of HNO3 (g) and SO2(g). Samples were collected in Riverside, CA, Bakers” eld, CA, and Provo, UT. Both sampling artifacts were found to be present. However, the ef-fect on the determination of HNO3 (g) by the URG sampler and on the determination of particulate-phase sulfate and nitrate by the R&P ChemSpec were both small, on the order of a few tenths of a microgram per cubic meter.a 1% w/w NaCl/glycerine coated ” rst denuder is used in the ChemSpec. Nitric acid can also be determined by the absolute differential method according to Equation (2) when there is no initialNaCl/glycerinecoated denuderand 2 NaH2CO3/glycerine coated denuders are used in the ChemSpec. The differential method is used in this case because a small fraction of inter-fering gases (e.g., PAN, NO2) can be absorbed by the NaHCO3/glycerine coated denuders. Concentrations of nitric acid deter-mined by the 2 methods agreed, suggesting that the differential method is valid. Regression analysisof nitric acid concentrations determined by the 2 methods were comparable by the t-test, gaveBakers” eld, CA, however, the calculated slope and ?% were poor in the Bakers” eld study, as indicated in Table 1, because the con-centrations of HNO3 were very low. Because the URG sampler was only used without an initial NaCl/glycerin coated denuder, the comparison of ChemSpec and URG annular denuder sam-plers will be based on the ChemSpec samples collected without the ”rstNaCl/glycerincoateddenuder.Concentrations of HNO3 were higher than those for partic-ulate nitrate in Riverside, CA. The average ratio of particulate nitrate to nitric acid was 1:2.4. However, HNO3 concentrations determined by the URG annular denuder sampler (which uses a Te‘ on coated elutriator-impactorinletbefore the denuder) by the absolute differential method (ADM) were 15% lower than com-parable data obtained by the ChemSpec sampler, which places the cyclone inlet after the denuders (Figure 4 and Table 1). This difference was statistically signi” cant only at the 90% level by the t-test. This difference might be caused by the URG inlet ad-sorption of HNO3 or coarse particle impact on the denuder of the ChemSpec. The ” rst reason is the more plausible because there was little coarse particulate (>2.5 1m) nitrate at Riverside, CA, during the fall (Chow et al. 1994), being only 2% of the total PM10 nitrate.