Respiratory critical oxygen pressures (COPR) determined from O(2)-depletion rates in media bathing intact or excised roots are unreliable indicators of respiratory O(2)-dependency in O(2)-free media and wetlands. A mathematical model was used to help illustrate this, and more relevant polarographic methods for determining COPR in roots of intact plants are discussed. Cortical [O(2)] near the root apex was monitored indirectly (pea seedlings) from radial oxygen losses (ROL) using sleeving Pt electrodes, or directly (maize) using microelectrodes; [O(2)] in the root was controlled by manipulating [O(2)] around the shoots. Mathematical modelling of radial diffusive and respiratory properties of roots used Michaelis-Menten enzyme kinetics. Respiration declined only when the O(2) partial pressure (OPP) in the cortex of root tips fell below 0.5-4.5 kPa, values consistent with depressed respiration near the centre of the stele as confirmed by microelectrode measurements and mathematical modelling. Modelling predictions suggested that the OPP of a significant core at the centre of roots could be below the usual detection limits of O(2)-microelectrodes but still support some aerobic respiration. In O(2)-free media, as in wetlands, the COPR for roots is likely to be quite low, dependent upon the respiratory demands, dimensions and diffusion characteristics of the stele/stelar meristem and the enzyme kinetics of cytochrome oxidase. Roots of non-wetland plants may not differ greatly in their COPRs from those of wetland species. There is a possibility that trace amounts of O(2) may still be present in stelar 'anaerobic' cores where fermentation is induced at low cortical OPPs.