Enhanced biological phosphorus removal was performed in a continuous laboratory-scale two-reactor system with sludge recirculation over a 75-day period. Influent wastewater was a synthetic medium based on acetate, and the sludge age was kept at 12 days. The adapted sludge stored poly-beta-hydroxyalkanoic acids (PHA) in the anaerobic reactor with a conversion ratio of 1.45 PHA/acetic acid (based on chemical O-2 demand : COD/COD) and gave ratio of a phosphate-P release to acetic acid uptake of 0.51 P/CH3COOH (w/w). Fractionation of anaerobic and aerobic sludges showed that the main part of phosphorus taken up, was eluted in the trichloroacetic acid fraction indicating that it was polyphosphate. A total of 60% of the phosphorus in the aerobic sludge was solubilized in the trichloroacetic acid fraction, whereas this fraction accounted for only 32% of the phosphorus in the anaerobic sludge. Only 4% of the total phosphorus in the aerobic sludge and 2% in the anaerobic sludge was found in the EDTA fraction, indicating low amounts of metal-bound phosphates. Isolation on acetate-based agar medium showed that Acinetobacter strains were present in the sludge. However, a more complete analysis of the bacterial community of the sludge was obtained by creating a clone library based on the 16S rRNA gene. A total of 51 partial clone sequences were phylogenetically evaluated. The predominating group was found in the high-(G + C) (mol%) gram-positive bacterial subphylum (31% of the sequenced clones), while the gamma proteobacteria only constituted 9.8% of the clones.