The fractal dimension of soil surface (Ds) is used to characterize soil geometric heterogeneity and therefore, it may characterize soil sorption activity. The range of its values is 2 ≤ Ds ≤ 3. The higher (nearer to 3) the Ds value, the more convoluted the surface is, and the Ds value close to 2 characterizes practically flat soil surface. Here, we attempted to evaluate the fractal dimensions of soils typical for Middle Volga region, Russia based on sorption data, and to establish quantitative relationships between Ds values and soil composition. Several methods of Ds determination in adsorption experiments have been proposed. One group of these methods requires the interpretation of only one isotherm in multilayer sorption and capillary condensation regions. Using such approach, Slawinski et al. (2002) did not find any reliable correlation between Ds values and soil organic matter (SOM) content of peat soils. Additionally, the same authors (Sokolowska and Sokolowski, 1999) observed no correlation between Ds values of kaolin samples modified with humic acids and their amounts. Furthermore, several significant limitations for the adequacy of this approach are known (Neimark, 1990). In particular, Farin and Avnir (1989) demonstrated that for some sorption systems the values of Ds < 2 and Ds > 3 are possible. In the present work, we used the Pfeifer-Avnir fractal method - the most flawless in terms of fractal theory (Neimark, 1990). This method is based on the comparison of adsorption isotherms of compounds (with molecules of different size) in the monolayer region. The most exact results are obtained if homologous sorbates (e.g. aliphatic alcohols) are used (Avnir et al., 1984). For the series of sorbates with similar nature, Ds value is given by n ~ σ ^-Ds/2, where n is the monolayer coverage (moles of sorbate per gram of sorbent) and σ is the molecule area. From this relation we can obtain lnV_m = alnV₀ + const, where V_m is sorbate monolayer volume, V₀ is the molar sorbate volume and a is the slope tangent. Next we immediately obtain Ds =3 (1 - а). Because the interaction between vaporous sorbates and SOM can be realized as diffusion (at least partially) it can be assumed that a portion of sorbate molecules penetrates into the SOM phase. Hence, for some soils the expected Ds values would be higher then 3. Such overstating of Ds values has been previously noted by Gorbatchuk et al. (1997) for solid proteins via sorption data of organic compounds. With the help of the static gas-chromatographic headspace analysis method we estimated fractalities of oven-dried soils (leached chernozem, gray forest and soddy podzolic soils) from sorption isotherms of aliphatic alcohols (methanol, ethanol, propanol-2 and 2-methyl-propanol-2). For soddy podzolic soil the Ds value was less than 3 indicating the predominant sorption mechanism as adsorption onto soil surface. On the contrary, for leached chernozem Ds value was much higher than 3, due to two additional mechanisms - SOM partitioning or mineral micropore filling. To distinguish between these two mechanisms, we determined Ds values (Ds = 8.39, 5.82, 5.69 and 4.06) for dark-gray forest soil samples from different layers (4-10, 10-20, 20-30 and 35-45 cm) with similar clay but significantly different SOM contents (7.5, 4.4, 3.1 and 2.2%). As a result, for all examined different soil samples, we have found close direct correlation between Ds values and SOM contents. Here all Ds values (with the exception of Ds for soddy podzolic soil) exceeded 3. This result strongly indicated the organic sorbate interaction with SOM under oven-dry conditions as sorbate diffusion into the soil organic phase. If Ds value overstating was caused by the microporous structure of soil, then the Ds - C_org correlation would be inverse due to the blocking of active mineral centers (i.e. micropores) by SOM. Thus, we conclude that Pfeifer-Avnir method does not allow estimating "true" values of soil fractality based on the sorption data of aliphatic alcohols. The obtained Ds values do not adequately describe the soil surface character. However, they allow to estimate the mechanism of interaction between SOM and probe molecules and its contribution into the sorption binding of organic sorbates by soils. The work was supported by ISTC, Project #2419; RFBR, Project 06-04-49097a; and Grant of President of Russia MK-2439.2005.5.