Abstract A disc infiltrometer with a disc diameter of 80 mm has been operated in the field at − 0.2 kPa a few days after the spring of 2003 spraying of a mixture of bentazone and isoproturon on three dried tilled topsoils differing in texture but under the same climatic and vegetation conditions. In the first part of this series of papers, we characterized the transient water flow, allowing us to study the complementary transient herbicide transport in the studied field-dried topsoils. Proportions of initial herbicide which remained in the studied topsoils significantly decreased from the silt loam soil, where we suggested predominant capillary flow, to the sandy loam and silty clay soils, where we suggested a combination of capillary flow and gravity-driven transient flow. However, we attempted to use the capillary invasion theory to infer the retardation factors in the initially unsaturated soil aggregates, on the assumption of predominant transport of herbicide residues by capillary flow. We showed that the more clayey the soil, the more retarded was the invading herbicide against the invading water, the invading isoproturon being significantly more retarded than the invading bentazone, whatever the soil was, as it was expected from their significant difference of solubility in water. Moreover, the batch k d (or k f) sorption coefficient values were larger than their field capillary flow k d value counterparts. This discrepancy increased when the amount of organo-mineral clay increased and was also much more pronounced in this study, based on assumed capillary invasion of soil herbicide residues, than in laboratory studies carried out on mostly saturated soil columns under continuous flow conditions, but more experimental and conceptual work is required.