The purpose of this study was to investigate the dependence of contrast-to-noise ratio (CNR) on the dose and rate of sprodiamide injection in magnetic resonance relative cerebral blood volume (rCBV) imaging. rCBV maps for 35 normal volunteers were constructed from dynamic MR image sets acquired with echo-planar spin-echo imaging after intravenous injection of sprodiamide. Doses of .1, .2, and .3 mmol/kg, at rates of 2 ml/second and 5 ml/second, were tested. CNRs and blood/volume ratios of gray to white matter were computed. CNR depended on dose (P < .0001) but was independent of injection rate (P < .69). rCBV ratios of gray to white matter were dose independent (P < .38) and rate independent (P < .97). The dependence of CNR on dose, but not injection rate, has practical implications in optimal protocol design. The independence of gray/white ratios supports the theory underlying the generation of rCBV maps.