The real refractive index (nr) of dry aerosols measured at an Antarctic site was calculated using two methods. Number size distributions were applied in a Mie model where nr was varied until the measured and modeled scattering coefficients matched. Ion concentrations analyzed from impactor samples were used for calculating size-dependent nr. Generally the two methods were consistent with each other. The average iteration-derived nr varied mainly between ~1.4 and ~1.5 but also unreasonably low values, nr < 1.3, were obtained in air masses from the ocean. Possible explanations, such as inaccuracies in the method, water in the particles, non-analyzed biogenic particles, or non-spherical particles being responsible for the low values are discussed.