Measurement of local Cerebral Blood Flow (CBF) using the C15O2 continuous inhalation technique coupled with Position Emission Tomography (PET) rests on the assumption that the in vivo labelled water (H215O) of blood diffuses freely within the brain water pool. This requirement however, may be disputed: in the rhesus monkey, Eichling et al (1974) showed a linear relationship between the single-pass extraction of water (E) and CBF after intra-carotid H215O injection, such that E decreases as CBF increases. Such a limitation in water transport across the blood brain barrier has also been suggested to exist in man (Paulson, 1977). In order to verify its possible effects on measured CBF in baboons, simple computer analysis was performed comparing the data obtained with CBF values reported in similar experiments, but using a freely diffusible inert gas (133Xe). This allowed calculation of a water extraction index (IE), which was shown to decline as CBF increased. This result may be considered as a presumption of the part played by the limited water diffusibility on the final CBF values, but other factors must also be considered, such as the inescapable admixture of grey and white matter (partial volume effect), a limitation of present day PET technology.