The change in cerebral blood flow was determined after a step decrease in the Pco 2 of arterial blood from 40 to 25 mm Hg in awake man. Subjects monitored their own end-tidal Pcoo (infrared analyzer) and adjusted their voluntary ventilation to produce the step change, which they maintained for at least 1 hour. Cerebral blood flow relative to control was determined from the arterial-jugular venous oxygen saturation differences. After the step change, arterial Pco 2 fell in less than 30 sec to a plateau, cerebral blood flow fell with a time constant (to 1/e) of 0.3 min to a plateau of 68% of control, while jugular venous Pco 2 fell with a time constant for the fast component of 3.5 min. Base excess rose 1.2 mEq/liter within 1 min and remained at that level. It is concluded that CO 2 affects cerebral blood flow by direct diffusion into arteriolar walls, rather than by its effect on brain tissue Pco 2 or pH. It is postulated that the pH of the extracellular fluid of arteriolar smooth muscle is the common controlled variable through which CO 2 , and possibly hypoxia and blood pressure, determine vascular tone. This investigation was supported in part by Public Health Service Research Grants HE 08285 from the National Heart Institute and 5T1-GM-63 from the Institute of General Medicine. Dr. Severinghaus is a recipient of a Research Career Award (no. 5-K6-HE-19) from the National Heart Institute.Accepted for publication December 19, 1966. of cerebral vasomotion by CO 2 , and the experiments to be described, which may be regarded as an extension of theirs (although completed before their publications appeared), have led us to the opposite conclusion; namely, that CO2 exerts its action directly upon the arteriolar wall, flow being largely independent of brain tissue Pco2. However, as will be discussed later, their findings in hypocapnia are similar to ours, and both sites may contribute under certain circumstances.Method We have induced a step decrease in the Pco 2 of arterial blood (Paco 2 ) and utilized the succeeding washout period to observe the jugular venous Pco 2 and the arterio-venous (A-V) O 2 saturation difference, the former as an index of tissue Pco 2 , the latter as an index of cerebral blood flow. The justification of the assumptions required to calculate changes in blood flow from A-V O 2 saturation differences have been carefully discussed by Shapiro, Wasserman and Patterson (1). We studied seven healthy adult male subjects in semisupine position, with about 20° head-up tilt. The jugular bulb was cannulated percutaneously with an 18-20 gauge, 2. brachial artery was catheterized by the Seldinger technique. The sampling catheter of an infrared CO 2 analyzer was inserted into a nostril extension (10 cm long, 1 cm diam), the mouth, or a mouthpiece. The subjects were taught to monitor their own end-tidal Pco 2 (PACO 2 ) using the panel meter, and to hyperventilate suddenly, sufficiently to reduce their PACO 2 from control to 20 mm Hg with two or three fast deep breaths, and then to hold it at this l...