A nonlinear control strategy is developed for competitive mixed-culture bioreactors in which two cell populations compete for a common growth limiting substrate. A stream is periodically removed from the reactor, and the two cell populations are separated using specific cell adhesion. The steady state corresponding to the desired population fraction is stabilized by discarding faster growing cells and recycling slower growing cells to the reactor. The recycle loop must be operated periodically to allow regeneration of the adhesion column after each separation. As a result, the manipulated input is chosen as the sampling interval during which material is removed from the reactor. The nonlinear controller is designed using a simplified dynamic model that assumes continuous separation of the cell populations. The controller is implemented by calculating the sampling interval that leads to the same amount of material being removed from the reactor as that computed from the continuous control law. A nonlinear, closed-loop observer is used to generate one-timedelay- ahead predictions of the measured cell concentrations and the unmeasured substrate concentration. The efficacy of the proposed control strategy is evaluated via simulation.