The aim of this study was to investigate the impact on numbers of using different media for the enumeration of Escherichia coli subjected to stress, and to evaluate the use of different resuscitation methods on bacterial numbers. E. coli was subjected to heat stress by exposure to 55 degrees C for 1h or to light-induced oxidative stress by exposure to artificial light for up to 8h in the presence of methylene blue. In both cases, the bacterial counts on selective media were below the limits of detection whereas on non-selective media colonies were still produced. After resuscitation in non-selective media, using a multi-well MPN resuscitation method or resuscitation on membrane filters, the bacterial counts on selective media matched those on non-selective media. Heat and light stress can affect the ability of E. coli to grow on selective media essential for the enumeration as indicator bacteria. A resuscitation method is essential for the recovery of these stressed bacteria in order to avoid underestimation of indicator bacteria numbers in water. There was no difference in resuscitation efficiency using the membrane filter and multi-well MPN methods. This study emphasises the need to use a resuscitation method if the numbers of indicator bacteria in water samples are not to be underestimated. False-negative results in the analysis of drinking water or natural bathing waters could have profound health effects.