Abstract Both the sub-cooled liquid vapor pressure ( P L) and the octanol–air partition coefficient ( K OA) are used to describe the partitioning of non-polar organic compounds between the gas phase and a variety of natural organic substrates in soil, atmospheric particles and foliage. Whether the former is preferable over the latter depends on whether the interaction of the organic compound with the organic matter (OM) resembles more those in the pure liquid than those in liquid octanol. The activity coefficient in octanol ( γ Oct) is a quantitative measure of the difference between these two interactions. An analysis of P L and K OA values for several sets of non-polar and semi-volatile organic compounds (chlorobenzenes, PCBs, PCNs, PCDD/Fs, PBDEs), and of the γ Oct values derived from these, reveals that γ Oct tends to range from 1 to 10 suggesting that P L and K OA are very highly correlated. Furthermore, the estimated standard deviation of γ Oct tends to be so large that P L and K OA are virtually indistinguishable within the measurement uncertainty. Whether γ Oct within a group of related compounds increases, decreases or stays the same with increasing molecular mass depends on the specific K OA and P L data set used in the calculation of γ Oct. This implies that with the current precision of K OA, P L and partition coefficients involving OM it is impossible to judge one parameter better than the other.