During Polar Sunrise Experiment 1992 at Alert in the Canadian high Arctic, size‐fractionated observations of aerosol constituents (halogens, Na, V, As, Sb, Zn, Al, Ca, SO4=, Sm, K, Mn, and Mg) were made using a low‐pressure cascade impactor and a high‐volume virtual impactor (HVVI). Over 80% of the mass of V, Br, I, As, Sb, Zn, and SO4= was in particles <2.5 μm diameter. In contrast to SO4=, NO3− has more mass in larger particles. Similarly, Cl peaked in larger particles (1.72 to 6.0 μm) than Na (0.49 and 1.72 μr). Both of these effects are likely caused by reactions of sulphate acids with salts of nitrates and chloride in submicrometer particles and the volatilization of HNO3 and HCl. For fine particles (diameter <2.5 μm), observations of relatively, nonvolatile primary particulate elements (Na, V, Mn) with the two measurement devices agreed within analytical uncertainty. In contrast, they did not agree well for the more volatile halogens. The HVVI always recorded higher. Sea‐salt enrichment factors of Cl (EFs s) referenced to Na for the HVVI data showed enrichment (1.2 to 1.8) in both fine and coarse aerosols during January. Thereafter, EFss averaged 0.92 and 0.57 for coarse and fine fractions, respectively, and were well correlated with each other. In contrast, EFss on fine particles collected with the cascade impactor were considerably lower (0.18 to 0.44), indicating volatilization of Cl within the device. A principal component analysis of fine particle (<2.5 μm) constituents from the HVVI identified six aerosol components consisting of anthropogenic (Mn, As, Sb, V, Zn, Al), sea salt (Cl, Na), iodine‐bromine (I, Br), photochemical (Br, −O3), soil (Ca, Al), and one loaded solely by Sm. For the whole period, there was a strong anticorrelation between O3 and aerosol Br but not with the other halogens. In this study, iodine was significantly linearly correlated with the square root of the Br concentration. It showed a weak anticorrelation with O3 only after polar sunrise but not before.