The present study deals with the genotoxicity assessment of vineyard pesticides in fish exposed in the field or in mesocosm conditions. Primary DNA damage was quantified as strand breaks using the single cell gel electrophoresis assay (Comet assay) applied to fish erythrocytes. In a first experiment, a significant genotoxic effect was observed following an upstream-downstream gradient in early life stages of brown trout (Salmo trutta fario) exposed in the Morcille River contaminated by a mixture of vineyard pesticides during three consecutive years. The pronounced response in terms of DNA damage reported in the present study could argue for a high sensitivity of fish early life stage and/or a high level of exposure to genotoxic compounds in the Morcille River. This stresses the interest in using trout larvae incubated in sediment bed to assess genotoxic compounds in the field. In a second experiment, adult European topminnow (Phoxinus phoxinus) were exposed in water running through artificial channels to a mixture of diuron and azoxystrobin, two of the main pesticides detected in the Morcille watershed. As compared with the unexposed channel, a 3-5-fold increase in the DNA damage was observed in fish exposed to chronic environmental pesticide concentrations (1-2 microg L(-1) for diuron and 0.5-1 microg L(-1) for axoxystrobin). A single 6h pulse of pesticide (14 microg L(-1) of diuron and 7 microg L(-1) of azoxystrobin) was applied to simulate transiently elevated chemical concentrations in the river following storm conditions. It did not increase genotoxicity. After a 1-month recovery period, DNA damage in exposed fish erythrocytes recovered to unexposed level, suggesting possible involvement of both repair mechanisms and cellular turnover in this transient response. This work highlights that vineyard treatment by pesticides and in particular diuron and azoxystrobin can represent a genotoxic threat to fish from contaminated watershed rivers.