While the in vitro reactivity of cerebral conducting vessels following subarachnoid hemorrhage has been extensively studied, in vivo cerebrovascular CO2 reactivity has not been systematically investigated. We tested the hypothesis that, in the canine model of subarachnoid hemorrhage, the rise in cerebral blood flow normally seen with hypercapnia is blunted during delayed vasospasm. Four groups of animals were studied: one received two 4-ml subarachnoid injections of nonheparinized arterial blood into the cisterna magna (n = 8), one received three subarachnoid injections of 5 ml blood (n = 5), one received two subarachnoid injections of 4 ml saline (n = 5), and a control group (n = 5) had no subarachnoid injections or angiography. Basilar artery diameter was measured from baseline and follow-up angiography. We determined CO2 reactivity by randomly varying the concentration of inspired CO2 and measuring regional cerebral blood flow with radiolabeled microspheres. Basilar artery diameter was not affected by saline injection and was reduced by 26 +/- 2.9% in the two-hemorrhage group and 55 +/- 1.9% in the three-hemorrhage group. Baseline cerebral blood flow and CO2 reactivity were similar in all four groups. We conclude that, in this model of delayed vasospasm, regional cerebral vascular CO2 reactivity is intact and extrapolation of in vitro data regarding basilar artery diameter and reactivity to cerebral blood flow must be done cautiously.