1. A method is described using trypsin/formic acid cleavage for unambiguously measuring occupancies of phosphorylation sites in rat heart pyruvate dehydrogenase [(32)P]phosphate complexes. 2. In mitochondria oxidizing 2-oxoglutarate+l-malate relative initial rates of phosphorylation were site 1>site 2>site 3. 3. Dephosphorylation and reactivation of fully phosphorylated complex was initiated in mitochondria by inhibiting the kinase reaction. Using dichloroacetate relative rates of dephosphorylation were site 2>(1=3). Using sodium dithionite or sodium pyruvate or uncouplers+sodium arsenite or steady state turnover ((31)P replacing (32)P in inactive complex) relative rates were site 2>site 1>site 3. With dithionite reactivation was faster than site 3 dephosphorylation, i.e. site 3 is apparently not inactivating. 4. The steady state proportion of inactive complex was varied (92-48%) in mitochondria oxidizing 2-oxoglutarate/l-malate by increasing extramitochondrial Ca(2+) (0-2.6mum). This action of Ca(2+) induced dephosphorylation (site 3>site 2>site 1). These experiments enable prediction of site occupancies in vivo for given steady state proportions of inactive complexes. 5. The proportion of inactive complex was related linearly to occupancy of site 1. 6. Sodium dithionite (10mm) and Ca(2+) (0.5mum) together resulted in faster dephosphorylations of each site than either agent alone; relative rates were site 2>(1=3). 7. Dephosphorylation and possibly phosphorylation of sites 1 and 2 was not purely sequential as shown by detection of complexes phosphorylated in site 2 but not in site 1. Estimates of the contribution of site 2 phosphorylation to inactivation ranged from 0.7 to 6.4%. 8. It is concluded that the primary function of site 1 phosphorylation is inactivation, phosphorylation of site 2 is not primarily concerned with inactivation and that phosphorylation of site 3 is non-inactivating.