We investigated 1) the regional distribution of cerebral blood flow (CBF), 2) the influence of end-tidal PCO2 (PETCO2) on CBF, and 3) the potential for an extracranial blood "steal" from the anterior brain region during passive hyperthermia. Nineteen (13 male) volunteers underwent supine passive heating until a steady-state esophageal temperature of 2 degrees C above resting was established. Measurements were obtained 1) during normothermia (Normo), 2) during poikilocapnic hyperthermia (Hyper), and 3) during hyperthermia with PETCO2 and end-tidal P-O2 clamped to Normo levels (Hyper-clamp). Blood flow in the internal carotid ((Q) over dot(ICA)), vertebral ((Q) over dot(VA)), and external carotid ((Q) over dot(ECA)) arteries (Duplex ultrasound), blood velocity of the middle cerebral (MCAv) and posterior cerebral (PCAv) arteries (transcranial Doppler), and cutaneous vascular conductance on the cheek (cheek CVC; Doppler velocimetry) were measured at each stage. During Hyper, PETCO2 was lowered by 7.0 +/- 5.2 mmHg, resulting in a reduction in (Q) over dotICA ( 18 17%), (Q) over dot(VA) ( 31 +/- 21%), MCAv ( 22 +/- 13%), and PCAv ( 18 +/- 10%) compared with Normo (P < 0.05). The reduction in (Q) over dot(VA) was greater than that in (Q) over dotICA (P = 0.017), MCAv (P = 0.047), and PCAv (P = 0.034). Blood flow/ velocity was completely restored in each intracranial vessel (ICA, VA, MCA, and PCA) during Hyper-clamp. Despite a 250% increase in (Q) over dot(ECA) and a subsequent increase in cheek CVC during Hyper compared with Normo, reductions in (Q) over dot(ICA) were unrelated to changes in (Q) over dot(ECA). These data provide three novel findings: 1) hyperthermia attenuates(Q) over dot(VA) to a greater extent than(Q) over dot(ICA), 2) reductions in CBF during hyperthermia are governed primarily by reductions in arterial PCO2, and 3) increased (Q) over dot(ECA) is unlikely to compromise (Q) over dot(ICA) during hyperthermia.