The aim of the present study was to elucidate consistent patterns in chronic polycyclic aromatic compound (PAC) toxicity to soil and sediment inhabiting invertebrates. Therefore we examined our experimental dataset, consisting of twenty-one chronic effect concentrations for two soil invertebrates (Folsomia candida and Enchytraeus cripticus) and two sediment invertebrates (Lumbriculus variegatus and Chironomus riparius) exposed to six PACs (two homocyclic isomers, anthracene and phenanthrene; two azaarene isomers: acridine and phenanthridine; and two azaarene transformation products, acridone and phenanthridone). In order to determine if effect concentrations were accurately predicted by existing toxicity-K(ow) relationships describing narcosis, chronic pore water effect concentrations were plotted jointly against logK(ow). Fifteen of the twenty-one effect concentrations (71%) were above the lower limit for narcosis, showing that narcosis was the main mode of action for the majority of the tested homo- and heterocyclic PACs during chronic exposure. Toxicity of all tested compounds to soil organisms was accurately described by the toxicity-K(ow) relationship. However, for the sediment invertebrates exposed to some of the tested heterocyclic PACs deviations from narcosis were identified, related to specific physicochemical properties of the test compounds and/or species specific sensitivities. It is concluded that existing toxicity-K(ow) relationships describing narcosis in some cases underestimate chronic PAC toxicity to sediment inhabiting invertebrates.