The main objective of this study was to evaluate the efficacy of four different wood modification systems (acetylation, furfurylation, ThermoWood and linseed oil) using an expanded version of the ENV 807 test with three different soil types, i.e. terrestrial microcosm (TMC) test. The sub-goals were to test: 1) the effect of different soil types, 2) the performance of the different modifications compared to reference treatments, 3) differences in performance after 24 and 40 weeks exposure, 4) the potential differences between the laboratories at SP Trätek and Norwegian Forest and Landscape Institute (NFLI) in duplicate set-ups. Generally compost soil (ENV 807) gave the highest mass loss. In this test the mass loss from the soil from Simlångsdalen test field and the forest soil differed slightly between the labs and between treatments. Even though the different soil types did give differences in degree of decay, the relative performance of the modified wood, compared to each other and compared to the references, was generally the same for all soil types. Both furfurylation, acetylation and ThemoWood D and D+ performed well. Acetylated wood at 23 and 30 WPG gave the best effect. ThermoWood D and D+ perfomed better than ThermoWood S+. Visorwood 25, 37 and 50 WPG had the same performance as ThermoWood D and D+. ThermoWood S+ was approximately in the same range as linseed oil. The modified wood performed at the same level, or better, than the reference CC and CCA preservatives in retentions for UC4 use. An interesting result is the poor performance of the CC preservative in this test. No great variation in relative performance was detected between 24 and 40 weeks of exposure. The duplicate set-ups gave very similar results for all soil types and all wood treatments except for untreated pine sapwood in forest soil, with a higher mass loss in Sweden than in Norway.