With no internal vibrational or rotational degrees of freedom, atomic solutes serve as the simplest possible probe of a condensed-phase environment's influence on solute electronic structure. Of the various atomic species that can be formed in solution, the quasi-one-electron alkali atoms in ether solvents have been the most widely studied experimentally, primarily due to the convenient location of their absorption spectra at visible wavelengths. The nature of solvated alkali atoms, however, remains controversial: the consensus view is that solvated alkali atoms exist as (Na + , e -) tight-contact pairs (TCPs), species in which the alkali valence electron is significantly displaced from the alkali nucleus and confined primarily by the first solvent shell. Thus, to shed light on the nature of alkali atoms in solution and to further our understanding of condensedphase effects on solutes' electronic structure, we have performed mixed quantum/classical molecular dynamics simulations of sodium atoms in liquid tetrahydrofuran (Na 0 /THF). Our interest in this particular system stems from recent pump-probe experiments in our group, which found that the rate at which this species is solvated depends on how it was created (Science 2008, 321, 1817); in other words, the solvation dynamics of this system do not obey linear response. Our simulations reproduce the experimental spectroscopy of this system and clearly indicate that neutral Na atoms exist as (Na + , e -) TCPs in solution. We find that the driving force for the displacement of sodium's valence electron is the formation of a tight solvation shell around the partially exposed Na + . On average, four THF oxygens coordinate the cation end of the TCP; however, we also observe fluctuations to other solvent coordination numbers. Furthermore, we find that species with different solvent coordination numbers have unique absorption spectra and that interconversion between species with different solvent coordination numbers requires surmounting a free energy barrier of several k B T. Taken together, our results suggest that the Na 0 /THF species with different solvent coordination numbers may be viewed as chemically distinct. Thus, we can explain the kinetics of Na TCP formation as being dictated by changes in the Na + solvent coordination number, and we can understand the dependence on initial conditions seen in the solvation dynamics of this system as resulting from the fact that the important solvent coordinate involves the motion of only a few molecules in the first solvation shell.