Eukaryote-metabolic inhibitors have been applied in published incubation experiments to inhibit protozoans and infer their impacts on bacteria. This method has been used with water-column, sedimentary, biofilm, biofilter, soil, groundwater, wastewater, and sludge samples. The approach rests on assumptions that rarely have been tested: 1) the inhibitor is completely effective against all protozoans; 2) the inhibitor has no effects on other eukaryotes in the system (e.g., microalgae, metazoans) that could influence bacterial activity; and 3) the inhibitor has no effect on bacterial activity. We tested these assumptions for 10 inhibitors used in published studies (amphotericin B, anisomycin, colchicine, cycloheximide, cytochalasin B, fumagillin, griseofulvin, neutral red, nystatin, and thiram) and one combination of inhibitors (colchicine + cycloheximide) using marine sedimentary organisms. Only thiram was completely effective against the natural assemblages of ciliates and flagellates. With some inhibitors, partial inhibition of ciliates was followed by regrowth within 6-24 h or a bloom of a resistant species. Most inhibitors also killed infaunal invertebrates. Five inhibitors were further tested for direct effects on bacteria at concentrations that inhibited ciliates. Each inhibitor altered either esterase activity, denitrification, or both. All inhibitors thus violated one or more assumptions of the method. Conclusions in published work that did not demonstrate the validity of all assumptions underlying this method should not be accepted. Any future attempt to use a eukaryotic inhibitor against protozoans must confirm all assumptions of this method for the particular habitat, species, and processes to which the inhibitor will be applied.