A laboratory-scale nutrient removal activated sludge system, based on the AAO configuration, was used to treat a synthetic wastewater from a milkpowder/butter factory. In this system, substrate is fed to anaerobic and anoxic selectors in series with an aerobic reactor. Sludge is returned to the anaerobic selector, and mixed liquor from the aerobic reactor is recycled to the anoxic selector. The overall system is operated at an HRT of 7 days and a nominal sludge age of 20 days. This system was prone to prolonged bulking periods, with filamentous bacteria Sphaerotilus natans, Type 0411 and Haliscomenobacter hydrossis being identified in the mixed liquors, although effective clarifier operation prevented loss of suspended solids. Theory suggests that selectors may be used to circumvent low F:M bulking, and to bring about enhanced biological nutrient removal. An investigation of the initial design revealed that relatively high nitrite levels were present in the system, and a larger anoxic selector with an HRT of 820 minutes was substituted for the original one with an HRT of 48 minutes. This resulted in a decrease in nitrite and a equivalent increase in nitrate in the system. Overall nitrogen removal remained unchanged at 66%, and SVI levels did not improve. On resetting the anoxic selector to its original size, the effect was not reversed. Phosphorus removal efficiency was detrimentally affected by the anoxic sizing experiment (49% to 20%), and this may be linked to the raised level of nitrate in the system. COD removal efficiency remained excellent throughout the trial at over 90% removal.