1. The primary objective of this study was to examine in man, how induced changes in global cerebral blood flow (gCBF) affected a regional cerebral blood flow (rCBF) increase resulting from a neural activation task (opening of eyes). A secondary objective was to quantify how such induced changes in gCBF were distributed between representative regions of either predominantly grey matter or white matter. 2. Positron emission tomography with intravenous infusion of H2(15)O, was used to measure gCBF in six normal males. Concomitant measures of rCBF were obtained in three different regions of interest (ROI): a representative area of predominantly grey matter, a representative area of predominantly white matter and an area of visual cortex. 3. Cerebral blood flow was altered by establishing steady-state changes in PCO2 at a near constant ventilation of approximately 30 l min-1. The mean PET,CO2 (+/- S.D.) levels (mmHg) that resulted were: low, 21.8 +/- 1.8; normal, 39.8 +/- 1.0, and high, 54.8 +/- 1.2. The normal and high levels were obtained by adding appropriate amounts of CO2 to the inspirate. The corresponding mean gCBF levels across all six subjects with eyes closed were: low, 24.2 +/- 4.6; normal, 37.2 +/- 3.9 and high, 66.8 +/- 7.6 ml min-1 dl-1. 4. Blood flow in grey matter (insular cortex) and white matter (centrum semiovale) at normal levels of PCO2 averaged 56.8 +/- 10.1 and 20.3 +/- 3.4 ml min dl-1 respectively. As PCO2 rose, the increase in rCBF to grey matter was approximately three times greater than that to white matter. 5. An activation state of eyes open in a brightly lit room was compared to a baseline state of eyes closed in a darkened room at the three levels of PCO2 (and hence at three levels of gCBF). Over the whole gCBF range a significant (P = 0.028) effect of increasing rCBF in the visual cortex ROI was found in response to opening the eyes; the effect of this activation on rCBF was not significantly dependent (P = 0.34) on the PCO2 (and hence gCBF) level. The effect of the activation on the rCBF was apparently 'additive' to the rise of rCBF associated with PCO2-related gCBF increase. 6. The results confirm the need to normalize for changes in gCBF during studies of rCBF in response to an activation protocol. They also provide support for the use of an 'additive' model to achieve such normalization provided that other cortical areas behave in a similar manner to that of the visual cortex.