The sources of error involved in the measurement of regional cerebral blood flow (rCBF) with radioactive tracers in man have been studied. This report shows that one of the main sources is the weakness of the energy emitted by the tracers used. Clearance curves obtained with scalp detectors are subject to interference: the radiation measured originates not only in the region studied but also, by scattering, from the rest of the brain. Because of the weak energy levels involved, the scattered radiation contributes significantly to the total measured radiation, and for the same reason it is impossible to distinguish clearly the scattered radiation from that arriving directly. These systematic errors in rCBF measurements were studied quantitatively by in vitro tests and computer simulation, and limits were established for resulting alterations of all varialbes determined (grey and white matter flow, mean flow, compartment weight). It was concluded that the biologic origin of a modification of rCBF can be affirmed only in the case of changes of at least 10%.