Experimental support for the steady state 15O inhalation technique, as used to measure cerebral blood flow (CBF) and oxygen utilisation (CMRO2), was obtained by describing the response of the cerebral vasculature to variations in arterial PCO2 in 6 anaesthetised dogs. Measurements were made using a positron emission tomography (ECAT II) and arterial blood sampling, during the sequential constant inhalation of C15O2 and 15O2. Values of CBF and CMRO2 were calculated for a mixture of white and grey matter, using the steady state tracer equations derived by Jones et al. (1976). The mean CMRO2 was 3.58 +/- 0.81 ml O2 . 100 ml-1 . min-1, whilst the mean CBF and OER (oxygen extraction ratio) values (for an arterial PCO2 of 40 mm Hg) were 39.9 ml . 100 ml-1 and 0.50 +/- 0.06, respectively. Arterial PCO2 was varied between 20 and 150 mm Hg. CBF was found to correlate closely with arterial PCO2, resulting in a mean slope (specific reactivity) of 1.52 +/- 0.38 ml . 10 ml-1. mm Hg-1. Pooling the flow data resulted in a linear relationship between CBF (% change) and arterial PCO2 in the range 20-70 mm Hg, with a slope (% reactivity) of 3.2% mm Hg-1 (2 P less than 0.001). The oxygen extraction ratio (OER) fell with increasing values of arterial PCO2 resulting in a stable CMRO2 throughout each study. There was no correlation between CMRO2 and artificially increased CBF. These results support and give confidence in the use of the 15O inhalation technique for measuring CBF, OER and CMRO2.