Abstract-Increasing the total volume of test medium resulted in decreased lag times (T L ) in biodegradability shake flask batch tests conducted with either surface water or with synthetic mineral medium inoculated with supernatant from settled activated sludge. Experiments were performed with test volumes ranging from 1.8 ml to 100 L using two 14 C-labeled model chemicals, 2,4-dichlorophenoxyacetic acid (2,4-D) and p-nitrophenol (PNP), both of which are known to be readily degradable after variable lag phases. The T L ranged from 2.1 to 30.4 d for PNP and from 16 to 37 d for 2,4-D. Decreasing the test volume tended to increase the lag time, even when a single test batch was redistributed into smaller flasks. With 5 ml supernatant added to different volumes of mineral medium, lag times for PNP were independent of the test volume in a range from 10 to 1,000 ml. At small volumes of 10 ml or less, degradation failed randomly. Our findings are partly explained by the hypotheses that a sufficient total amount as well as a sufficient concentration of specifically degrading microorganisms or consortia of bacteria must be present initially for biodegradation to get started, from which follows that with too small inoculations or with too small test volumes, biodegradation may fail randomly. A straightforward practical implication of the findings is that the test volume in biodegradability tests can significantly influence the lag time and thus sometimes be decisive for the outcome in biodegradation studies.