The biological elimination of phosphorus is achieved through enhanced biological phosphorus removal (EBPR), which uses the capacity of certain bacteria to accumulate polyphosphate intracellularly; these organisms are selected for during sequential anaerobic–aerobic phases [21]. The key players in this (sometimes unreliable) process are the polyphosphate-accumulating organisms (PAO) and the glycogen-accumulating organisms (GAO), which, like the PAO, are also selected for during anaerobic–aerobic cycling but are believed to compete with the PAO for carbon. Cultivation studies suggesting that Acinetobacter spp. were the PAO responsible for EBPR in activated sludge proved misguided [22], and several new PAO candidates have since emerged, including the uncultivated betaproteobacterium ‘Candidatus Accumulibacter phosphatis’ [23]. Considerable evidence suggests that this and other Rhodocyclus-related organisms are important PAO in both laboratory [24] and full-scale [25] EBPR systems. A series of FISH-MAR ( Box 1) experiments targeting these organisms as well as actinobacterial PAO and the GAO – the presumed cause of many EBPR breakdowns because of their competition with PAO – yielded valuable insights into the ecophysiology of EBPR-relevant microorganisms. Uncultivated gammaproteobacterial GAO (e.g. ‘Candidatus Competibacter phosphatis’ [26]) are widespread in EBPR plants, and FISH-MAR revealed that these organisms share similar physiologies with the Rhodocyclus-type PAO, no doubt contributing to the competition among these functional groups 27 and 28. By contrast, GAO physiology differs substantially from that of the actinobacterial PAO, at least with regard to the uptake spectrum and storage of organic carbon sources. Indeed, of all the substrates tested, only amino acids were taken up by both groups [28]. Interestingly, actinobacterial PAO often dominated in EBPR plants, where GAO and the other PAO were largely absent. These observations led to the intriguing suggestion that the problematic effects of GAO in full-scale EBPR plants could be removed by stimulating the actinobacterial PAO and thereby removing both GAO and Rhodocyclus-related PAO altogether [28]. These findings provide an eloquent demonstration of the power of molecular ecophysiology techniques from both basic and applied perspectives.