The activation and recovery phases of the murine rod photo-response were determined from corneal electroretinograms (ERGs) obtained in response to pairs of full-field flashes producing 50-10(5) photoisomerized rhodopsins (R*) per rod. The a-wave component of the ERG in response to the initial flash provided a well established measure of the activation phase of the rod response. The amplitude of the a-wave response to an intense second flash (45,000 R*) delivered 0.2-5 seconds (s) after the first flash was used to reconstruct the recovery phase of the response. For 160-3000 R* rod-1, recovery curves were isomorphic, translating on the time axis such that each e-fold increase in R* produced an incremental recovery delay of tau c = 210 +/- 50 ms (mean +/- SD). For initial flashes producing > 3000 R*, recovery curves lost their initial isomorphism and half-times had intensity dependence exceeding 1 s per e-fold increase in R*. We conclude that for flashes producing < 3000 R*, the effective lifetime of these R* is not > 210 ms. Two extant and non-mutually exclusive hypotheses are discussed that can account for the sharp increase in recovery times from flashes producing > 3000 R*. They are as follows: (1) approximately 0.03% of R* have a lifetime exceeding 1 s; and (2) the gamma subunit of phosphodiesterase (PDE gamma) serves as a GTPase-activating factor, and 3000 R* produce sufficient activated G-protein (G*) to exceed the total quantity of PDE gamma subunits such that excess G* must wait for unoccupied PDE gamma to inactivate via GTP hydrolysis.