A note on the efficiency of residual-based a posteriori error estimators for some mixed finite element methods.