The fate of [14C]-amitrole herbicide was studied in eight soils having different capacities for amitrole mineralisation. Laboratory incubations were run combining different experimental conditions: temperature (4, 28 and 50°C), soil moisture (50, 100 and 150% of soil water holding capacity) and microbial activity (sterile and non-sterile conditions). During incubation, samples of the soils were periodically extracted with 0.5 M NH4OH and extracts were analysed by HPLC. The lengths of time needed for 50% dissipation of amitrole (DT50) in soils ranged from less than 1 day to more than 70 days. Amitrole mineralisation occurred only in non-sterile soils, showing that it is a biological process. Mineralisation was lower in soils with a coarse texture than in soils with a fine texture. Soil water content had little influence on the total amount of amitrole mineralised at the end of incubation. Temperature had a greater influence on mineralisation, although rates were still high at low and high temperatures. In non-sterile as in sterile soils, the major product detected in the extracts was amitrole. Additional non-identified radioactivity was occasionally extracted. However, it never represented more than 10% of initially applied amitrole. Non-extractable residues represented less than 15% of applied radioactivity in acidic soils and about 30% of applied radioactivity in alkaline and neutral soils. The amount of non-extractable radioactivity formed was enhanced in sterile as compared to non-sterile soils. Furthermore, in sterile soils, high temperature induced an increase of non-extractable residues, showing that amitrole is chemically quite reactive.