The postulation that hypoxia increases local cerebral blood flow (lCBF) mainly by perfusing more capillaries (the capillary recruitment hypothesis) was tested in awake adult male Sprague-Dawley rats exposed to 10% O2 and control rats. The [14C]iodoantipyrine technique was used to measure lCBF. Local cerebral blood volume was determined by measuring plasma and red cell distribution spaces within the brain parenchyma with 125I-labeled serum albumin (RISA) and 55Fe-labeled red cells (RBC), respectively. Tissue radioactivity in 44 brain areas was estimated by quantitative autoradiography. Hypoxia raised lCBF by 25-90% in all brain areas. In about one-quarter of the brain areas, the rise in blood flow was associated with a small increase in microvascular plasma and blood volumes. This change in blood volume, which could be the result of perfusing more parenchymal microvessels and/or increasing parenchymal microvessel diameter, is not sufficient to account for the observed rise in lCBF. In the remaining areas the RISA, RBC, and blood spaces were either unchanged or only marginally increased by hypoxia. For this hypoxic perturbation, the major mechanism of raising blood flow appears to be increased velocity of microvessel perfusion and not perfusion of more capillaries. These findings provide only limited support for the capillary recruitment hypothesis.