Our previous study showed that an activated-sludge process broke down at the phenol-loading rate of 1.5 g l(-1) day(-1), when non-flocculating bacteria (called R6T and R10) overgrew the sludge, resulting in a sludge washout. In this study, we attempted to circumvent this breakdown problem by reclaiming the consortium structure. Activated sludge was fed phenol, and the phenol-loading rate was increased stepwise from 0.5 g l(-1) day(-1) to 1.0 g l(-1) day(-1) and then to 1.5 g l(-1) day(-1) Either galactose or glucose (at 0.5 g l(-1) day(-1)) was also supplied to the activated sludge from the phenol-loading rate of 1.0 g l(-1) day(-1). Pure culture experiments have suggested galactose to be a preferential substrate for a floc-forming bacterium (R6F) that predominantly degrades phenol under low phenol-loading conditions. Supplying galactose allowed sustainment of the R6F population and suppression of the overgrowth of R6T and R10 at the phenol-loading rate of 1.5 g l(-1) day(-1) This measure allowed the activated-sludge process to treat phenol at a phenol-loading rate up to 1.5 g l(-1) day(-1), although it broke down at 2.0 g l(-1) day(-1). In contrast, supplying glucose reduced the R6F population and allowed the activated-sludge process to break down at the phenol-loading rate of 1.0 g l(-1) day(-1) This study demonstrated that reclamation of the activated-sludge consortium by selective biostimulation of the hoc-forming population improved the phenol-treating ability of the process.