The stability and reproducibility of the dynamic susceptibility contrast (DSC) MRI method for sequential relative cerebral blood volume (relCBV) measurements was evaluated to validate the method for use in quantitative studies of cerebral hemodynamics in humans. A spin echo echo planar imaging protocol was used in conjunction with multiple bolus injections of the susceptibility contrast agent gadoteridol (GD). The effects of variation in interbolus interval (10 min to 4 h), the number of injections (two to four), and the effect of the cerebral vasodilating agent acetazolamide (ACZ) were evaluated in 44 experiments performed with 22 normal subjects. Two fundamental observations were made. First, with multiple injections of GD, the change in MR signal over time was not consistent from first to subsequent boluses. A second bolus administered 10 min to 2 h after an initial bolus resulted in signal change of greater amplitude and duration, resulting in artifactually elevated estimates of relCBV, consistent with a residual effect of GD. Second, a relative steady state could be reached with serial injections of GD, such that the profile of subsequent boluses closely paralleled those of previous ones. This facilitates the reliable measurement of relCBV during activation, as demonstrated by use of ACZ.