The effects of metals and petroleum hydrocarbons on benthic marine diatom communities were assessed using a manipulative field experiment at Casey Station, Antarctica. Uncontaminated, metal contaminated, and petroleum hydrocarbon contaminated sediments were deployed for 11 weeks during the 1999 austral summer. The treatments were deployed at three different locations: Brown Bay, which has elevated levels of anthropogenic contaminants, and two uncontaminated reference locations, O'Brien Bay and Sparkes Bay, the latter of which has naturally occurring high concentrations of some heavy metals. At each location, significant differences between the composition of diatom communities recruiting to control and petroleum hydrocarbon contaminated treatments were observed. Navicula directa (Smith) Ralfs occurred at lower abundances in the petroleum hydrocarbon contaminated treatments than in the control treatments. In contrast, Navicula cancellata Donkin occurred at higher relative abundances in both contaminated treatments relative to the control treatment. Interactions between treatment and location were also observed for several species, including Navicula glaciei Van Heurck. Significant differences in the overall community composition of diatom communities between control and metal contaminated treatments and metal contaminated and petroleum hydrocarbon contaminated treatments were only observed within Brown Bay. The location of deployment also had a significant influence on the composition of diatom communities. Brown Bay had higher abundances of Achnanthes brevipes Agardh but lower abundances of Navicula aff. cincta (Ehrenberg) Van Heurck than either reference locations. This experiment demonstrated that benthic diatom communities are sensitive to sediment contamination and would be suitable for future monitoring work within this and other areas of Antarctica.