Precise estimation of toxicity is essential in the risk assessment of hydrophobic organic compounds (HOCs). Toxicity testing of HOCs in aquatic bioassays is particularly challenging due to loss of substance from the water phase during the test. To obtain a causal relationship between exposure and toxicological effects, the exposure concentration need to be stable during the test period. A continuous addition of substance is possible by using a flow-through system. In static tests, like algal assays, flow-through is however inconvenient to use. Another method capable of maintaining stable concentrations is passive dosing. By introducing a solid phase that acts as a reservoir of test substance, exposure concentration in the water phase can be calculated from the partition coefficient for each test substance. Passive dosing also has the advantage of excluding the need for a carrier solvent. This study aimed to investigate the performance of passive dosing compared with conventional dosing in a Skeletonema costetum assay. Polydimethylsiloxane (PDMS) was chosen as the solid-phase and was loaded into each well in a 96-well polystyrene microtiter plate. The coating of silicone was easy to perform and had good reproducibility. The polymer was spiked with tree separate HOCs with log KOW values ranging from 3.79 to 4.88. The release of test substance was rapid and equilibrium was reached within 40 minutes for all compounds. The buffering capacity of silicone film was however insufficient for the two least hydrophobic substances. When comparing concentration causing 50% growth inhibition (EC50) of S. constatum, an underestimation of the toxicity was observed for the substances in the conventional test. EC50 values were 2.6 times higher for oxybenzone and 6.6 times higher for pyrene in the conventional method compared with passive dosing. EC50 values were not obtained for phenanthrene within the water solubility.