Detailed fast transient absorption measurements have been performed at low temperature on reaction centers from Rhodobacter sphaeroides strain R-26 and on a double mutant, [LH(L131) + LH-(M160)], in which the P/P+ oxidation potential is roughly 140 mV (1100 cm-1) above that of wild-type reaction centers. In both samples, the decay of the excited singlet state of the initial electron donor is not well described by a single-exponential decay term. This is particularly true for reaction centers from the double mutant where at least three exponential kinetic components are required to describe the decay, with time constants ranging from a few picoseconds to hundreds of picoseconds. However, singular value decomposition analysis of the time-dependent absorption change spectra indicates the presence of only two spectrally distinct states in reaction centers from both R-26 and the double mutant. Thus, the complex decay of P* at low temperature does not appear to be due to formation of either the state P+BA- as a distinct intermediate in electron transfer or P+BB- as an equilibrated side product of electron transfer. Instead, the decay kinetics are modeled by assuming dynamic solvation of the charge-separated state, as was done for the long-lived fluorescence decay in the accompanying paper [Peloquin, J. M., Williams, J. C., Lin, X., Alden, R. G., Taguchi, A. K. W., Allen, J.P., & Woodbury, N. W. (1994) Biochemistry 33, 8089-8100]. The results of assuming a static distribution of electron-transfer rates at early times followed by dynamic solvation of the charge-separated states on longer time scales are also presented. Regardless of which model is used to describe the early time kinetics of excited-state decay, the time-dependent excited-state population on the 100-ps or longer time scale is best described in terms of thermal repopulation of P* from the charge-separated state, even at 20 K. This results in a time- and temperature-dependent driving force estimated for initial electron transfer of less than 200 cm-1 on all time scales from picoseconds to nanoseconds. Assuming a nonzero internal reorganization energy associated with charge separation, the small driving force does not appear to be consistent with the lack of temperature dependence of electron transfer and the fact that a mutant with a P/P+ oxidation potential 140 mV (1100 cm-1) higher than wild type is still able to undergo electron transfer, even at low temperature.(ABSTRACT TRUNCATED AT 400 WORDS)