Numerous literature data indicate that the mean heat of adsorption of a monolayer of N 2 ( Q N 2 ) on ice or snow at 77.15 K, determined by volumetric methods, is highly variable, suggesting that ice surface properties strongly depend on its mode of formation and its thermal history. Less numerous data on CH 4 adsorption show smaller variations of Q CH 4 . If such variations are real, the extrapolation to atmospheric chemistry models of adsorption parameters measured on laboratory-made ice may be unwarranted. We have measured CH 4 adsorption on variable amounts of a crushed ice sample, to show that when the total surface area of the sample is below a threshold value, Q CH 4 decreases. We identify the cause of this artifact as an error in the molar budget, because the temperature gradient in the tube connecting the introduction and expansion volumes is not taken into account. Performing an adequate molar budget suppresses this artifact, except for ice samples with very small total surface areas, where the resolution of the manometer becomes a limiting factor and a further decrease in Q CH 4 is observed. Error in Q gas results in large errors in surface area, and we suggest that the value of Q gas obtained can be used to test the reliability of the surface area measurement.