The paper examines the validity of two approaches frequently used to determine surface areas in activated carbons, namely the BET method and the use of immersion calorimetry. The study is based on 21 well characterized carbons, whose external and microporous surface areas, S e and S mi , have been determined by a variety of independent techniques. It appears clearly that S BET and the real surface area S mi + S e are in agreement only for carbons with average pore widths L o around 0.8-1.1 nm. Beyond, S BET increases rapidly and S BET À S e is practically the monolayer equivalent of the micropore volume W o . This confirms that a characterization of surface properties based on S BET is, a priori, not reliable. The study of the enthalpy of immersion of the carbons into benzene at 293 K, based on DubininÕs theory, shows that D i H consists of three contributions, namely from the interactions with the micropore walls (À0.136 J m À2 ), the external surface (À0.114 J m À2 ), and from the volume W Ã o of liquid found between the surface layers in the micropores (À141 J cm À3 ). It appears that for carbons where L o > 1 nm, the real surface area cannot be determined in a reliable way from the enthalpy of immersion and a specific heat of wetting alone.