The impacts of the fungicide Opus(A (R)) (epoxiconazole) on marine phytoplankton communities were assessed in a 12-day field experiment using in situ microcosms maintained underwater at 6 m depth. Three community analysis methods were compared for their sensitivity threshold in fungicide impact detection. When phytoplankton communities were exposed to 1 mu g l(-1) of epoxiconazole, no effects could be demonstrated using TTGE (Temporal Temperature Gradient gel Electrophoresis), flow cytometry or HPLC. When exposed to 10 mu g l(-1), TTGE fingerprints from PCR amplified 18S rDNA of communities exhibited significant differences compared with controls (ANOSIM, P = 0.028). Neither flow cytometry counts, nor HPLC pigment profiles allowed to show significant differences in microcosms exposed to 10 mu g l(-1) of epoxiconazole. When exposed to 100 mu g l(-1), all three methods allowed to detect significant differences in treated microcosms, as compared to controls. The TTGE analysis appears in this study as the most sensitive method for fungicide impact assessment on eukaryote microbial communities.