The toxicity of fresh and weathered gasoline and diesel fuel to willow and poplar trees was studied using a tree transpiration toxicity test. Soils were taken from an abandoned filling station. Concentrations in the samples were measured as the sum of hydrocarbons from C5 to C10 (gasoline) and C12 to C28 (diesel). Concentrations ranged from 145 to 921 mg/kg gasoline and 143 to 18231 mg/kg diesel. The correlation between log soil concentration and toxicity to willows (Salix viminalis x schwerinii) was highly significant for the diesel fraction (r2=0.81, n=19) and for the sum of hydrocarbons (r2=0.84, n=19). The EC50 (50% inhibition of transpiration) for the sum of hydrocarbons was determined at 3910 mg/kg (95% C.I., 2900 to 5270 mg/kg) and followed a log-normally distributed sigmoidal curve. The EC10 was 810 mg/kg (95% C.I., 396 to 1660 mg/kg). The results were verified with artificially mixed diesel and gasoline contaminated soils, and two willow and one poplar species (S. viminalis, S. alba and Populus nigra). Fresh diesel at about 1000 mg/kg showed no effect on S. alba, although P. nigra was more sensitive. 10000 mg/kg seriously affected the transpiration of all species, silver willow (S. alba) being the least sensitive. Free phase diesel killed all trees within six weeks. Fresh gasoline at 1000 mg/kg was deadly for all trees, hence was more toxic than weathered gasoline. Survival of poplars and willows planted at the abandoned filling station was compared to the laboratory findings. There was some correlation, but in the field, trees also suffered from other stress factors than fuel pollution.