The rates of the primary electron-transfer processes in Rhodobacter sphaeroides reaction centers have been examined in detail by using 150-fs excitation flashes at 870 nm. At room temperature the apparent time constants for both initial charge separation (P* -> P+BPhj) and subsequent electron transfer (P+BPhj -+ P+QA) are found to encompass a range of values ('1.3-4 ps and "100-320 ps, respectively), depending on the wavelength at which the kinetics are followed.We suggest this reflects a distribution of reaction centers (or a few conformers), having differences in factors such as distances or orientations between the cofactors, hydrogen bonding, or other pigment-protein interactions. We also suggest that the time constants observed at cryogenic temperatures (""1.3 and "100 ps, respectively, with much smaller or negble variation with detection wavelength) do not reflect an actual increase in the rates with decreasing temperature but rather derive from a shift in the distribution of reaction centers toward those in which electron transfer inherently occurs with the faster rates.The temperature dependence of the rates of electron transfer in bacterial reaction centers (RCs) has received considerable experimental and theoretical attention. For the most part the temperature dependence of the initial charge separation step (observed to occur in ""3 ps at 295 K and 1 ps at 10 K) has been followed via the decay of the broad stimulated emission band of P*, which has a maximum near 920 nm in Rhodobacter sphaeroides (6, 7). The temperature dependence of electron transfer from BPhj to QA in Rb.sphaeroides (r 200 ps at 295 K and 100 ps at 77 K and below) similarly has been followed primarily in a single region, namely via decay of the broad BPhL anion band centered near 665 nm (5). Kinetics also have been measured at a few wavelengths near 800 nm in the recent studies of the P*-P+BPh-process, with the primary focus being a search for evidence for the possible transient reduction of the monomeric pigment BChlL (6,7,11,12). During the course of our previous investigations of electron transfer from BPhjL to QA, we examined the near-infrared (730-830 nm) region also, with the expectation of finding time constants that agreed with those measured via decay of the BPhL anion band. This proved not to be uniformly the case, however. At room temperature, for example, we found that at certain wavelengths the observed time constant was somewhat less than 200 ps, and at other wavelengths it was somewhat greater than 200 ps (5). A major limitation of these measurements was that the signal-to-noise ratio prevented a complete study in the near-infrared region in that accurate time constants at room temperature could be measured only within two narrow (-5 nm) wavelength regions centered near the peak ground state absorptions of the BPhs and the monomeric BChls.We have reexamined these findings by using instrumentation that gives a significantly better signal-to-noise ratio in AA and, in addition, utilizes 150-fs flashes (compared to...