The fate of the fungicide carbendazim (applied in the formulation Derosal) in soil was determined in Terrestrial Model Ecosystem (TME) tests and corresponding field-validation studies, which were performed in four different countries (United Kingdom, Germany, Portugal, and The Netherlands). The tests used different soil types, and lasted for 16 weeks. On three of the four sites, grassland soils were used while the fourth site had an arable soil. TMEs consisted of intact soil columns (diameter 17.5 cm; length 40 cm) and were taken from the site where the field study was performed. In the first series of TME tests, carbendazim was applied at four dosages ranging between 0.36 and 77.8 kg a.i./ha, while in the second series of TME tests and the field-validation studies six dosages between 0.36 and 87.5 kg a.i./ha were applied. DT50 values for the dissipation of carbendazim in the TME and field tests were in most cases not significantly affected by the dosage used and ranged between 3.1 and 13.9 weeks in the top 15 cm soil layers. Corresponding DT90 values ranged between 10.1 and 46.1 weeks. DT50 and DT90 values tended to be higher in the more acidic soils of Amsterdam and Flörsheim (pH-KCl 4.8-5.1 and 5.3-5.9, respectively) than in the less acidic soils of Bangor and Coimbra (pH-KCl 5.8-6.6 and 6.4-7.1, respectively). Fate of carbendazim in soil showed similar patterns in the two TME tests and the corresponding field-validation study performed at each site. The only exception was Flörsheim, where the compound was significantly more persistent in the field probably due to different climatic conditions. Carbendazim was not recovered from leachates produced in the TME tests, nor was the compound detected in soil layers deeper than 15 cm. This demonstrates that no significant leaching occurred. This study demonstrates the the TME tests were quite successful in predicting the fate of carbendazim under field conditions.