It has been assumed that there exists an explicit "exercise stimulus" (i. e., a signal that is correlated to metabolic CO2 production and appropriately drives breathing). Alternatively, the behavior of the respiratory controller during exercise and inspiratory CO2 loading may be dictated by an optimal controller. The purpose of this paper is to examine the behavior of two alternative optimal controller configurations which do not require an explicit "exercise stimulus" for their operation. The first yields a multiplicative behavior between an exercise and airway CO2 load. The second yields an additive behavior between these two loads?a behavior consistent with that generally observed in man. Thus, an optimal controller structure may be part of the composite structure of respiratory control.