Efficacy and environmental fate of residual herbicides greatly depend on factors affecting their degradation in soil. The individuation of the main factors influencing herbicide dissipation in soil and their spatial variability is consequently needed. The objective of the present study was to investigate the field-scale variability of metolachlor dissipation in soil and factors considered to effect this process: soil microbial biomass (SMB), soil organic matter (SOM) content and pH. The study was carried out in a 10,000-m(2) maize field located in the Po Valley (Bologna, Italy). Metolachlor persistence (DT50) in soil ranged from 24.9 to 43.7 days, with an average value of 35.1 days. The coefficient of variation of metolachlor DT50 was 26.2%, confirming data reported in the literature. Correlation analysis revealed that metolachlor persistence in soil only depended to a minor extent to SOM content and soil pH. Conversely, dissipation of the herbicide was negatively correlated to the soil microbial biomass (r = -0.84; P < 0.05). These findings suggest that, under the studied conditions, SMB better predicted metolachlor persistence in soil than the other studied soil properties. SMB and metolachlor DT50 did not exhibit any clear spatial structure. Semivariograms for SOM and pH showed a range of 52 and 61 m respectively.