The effects of sample geometry and enzyme concentration on the H2-evolving activity of hydrogenase from Thiocapsa roseopersicina was measured. The specific activity increased linearly with increasing interface area between the liquid and the gas phase. Enzyme concentration was varied over four orders of magnitude and within this range the apparent specific activity depended on hydrogenase concentration. The experimental findings have been interpreted by a mathematical model involving competing H2 consumption reactions. The observed phenomena interfere with the widely used hydrogenase assay so that most of the previously published specific activity values are underestimated and should be corrected. The systematic error due to these hitherto unspecified parameters can easily exceed 10 000%; therefore, a thorough standardization of the assay procedure is necessary in order to make the data from various laboratories comparable.