OBJECTIVE The central nervous system controls the muscles of mastication and may dictate muscle outputs according to a biologically important objective. This study tested the hypotheses that (a) the effective sagittal TMJ eminence morphology, and (b) the outputs of the masticatory muscles during static biting, are consistent with minimisation of joint loads or minimisation of muscle effort. METHODS Numerical modelling predicted effective eminence morphology (from sagittal plane directions of TMJ force for centred loading over a range from molar to incisor biting) and TMJ and muscle forces during static unilateral biting in seven subjects. In vivo effective eminence morphology was measured from jaw tracking recorded from each subject. Muscle activities during biting tasks on first molar and incisor teeth were measured by electromyography using surface or indwelling electrodes. RESULTS Subject-specific predicted effective eminence morphology correlated with in vivo data (0.85< or =R2< or =0.99). Mixed and random coefficient analysis of covariance indicated good agreement between predicted and measured muscle outputs for all muscles of mastication investigated. Individual linear regression analysis showed that modelled muscle outputs accurately predicted EMG data, with average errors of 8% for molar and 15% for incisor biting. CONCLUSIONS Effective sagittal eminence morphology was consistent with minimisation of joint loads for all subjects. Masticatory muscle outputs during unilateral biting were consistent with minimisation of joint loads or minimisation of muscle effort, or both, depending on the subject. These results are believed to be the first to test model predictions of muscle output during biting for all muscles of mastication.