Five control configurations for continuous bioreactors are analyzed using a simple model. A controller-independent measure, the partial disturbance gain, is used for evaluating their controllability with respect to disturbance rejection. The computation of the partial disturbance gain is generalized to the case with several uncontrolled outputs. At substrate-limited growth conditions, the concentration turbidostat using the feed substrate concentration as the manipulated variable is the best control configuration. The conventional turbidostat should be avoided. When the cell growth is not substrate limited, all these control configurations are effective except the concentration nutristat which is not feasible regardless if the growth is substrate limited or not.