Absorbance difference kinetics were measured on quinone-reduced membrane-bound wild type Rhodobacter sphaeroides reaction centers in the wavelength region from 690 to 1060 nm using 800 nm excitation. Global analysis of the data revealed five lifetimes of 0.18, 1.9, 5.1, and 22 ps and a long-lived component for the processes that underlie the spectral evolution of the system. The 0.18 ps component was ascribed to energy transfer from the excited state of the accessory bacteriochlorophyll (B*) to the primary donor (P*). The 1.9 ps component was associated with a state involving a BChl anion absorbing in the 1020 nm region. This led to the conclusion that primary electron transfer is best described by a model in which the electron is passed from P* to the acceptor bacteriopheophytin (HL) via the monomeric bacteriochlorophyll (BL), with the formation of the radical pair state . An analysis assuming partial direct charge separation from B* [Van Brederode, M. E., Jones, M. R., and Van Grondelle, R. (1997) Chem. Phys. Lett. 268, 143-149] was also consistent with the data. Within the framework of a five component model, the 5.1 and 22 ps lifetimes were associated with charge separation and relaxation of the radical pair state respectively, providing a description which adequately accounted for the complex kinetics of decay of P*. Alternatively, by assuming that the 5.1 and 22 ps components originate from a single component with a multi-exponential decay, a simpler analysis with only four components could be employed, resulting in only a small increase (7%) in the weighted root mean square error of the fit. In both descriptions part of the decay of P* proceeds with a lifetime of about 2 ps. The relative merits of these alternative descriptions of the primary events in light-driven electron transfer are discussed. Similar measurements on YM210H mutant reaction centers revealed four lifetimes of 0.2, 3.1, and 12 ps and a long-lived component. The 3.1 and 12 ps lifetimes are ascribed to multi-exponential decay of the P* state. The differences with the WT data are discussed.