The adiabatic electron affinity (AEA) for the Watson-Crick adenine-thymine (AT) DNA base pair is predicted and contrasted to that of guanine-cytosine (GC) with a range of density functional methods with double- and triple-zeta plus polarization plus diffuse (DZP++ and TZ2P++) basis sets. An estimate of the true AEA is provided using a bracketing method that has been calibrated against a comprehensive tabulation of experimental electron affinities [Rienstra-Kiracofe, J. C.; Tschumper, G. S.; Schaefer, H.F.; Nandi, S.; Ellison, G. B. Chem. Rev. 2002, 102, 10163]. Optimized structures for AT and the AT anion are compared to the neutral and anionic forms of the individual bases as well as Rich's 1976 X-ray structure for the related sodium adenylyl-3',5'-uridine hexahydrate, ApU.6H(2)O. In contrast to the angular distortions (to nonplanarity) occurring in GC upon anion formation, the angular distortions for the AT anion are slight. However, in an analogous fashion to the GC anion, major changes in the AT anion hydrogen bond distances, from 0.27 to 0.32 Angstrom, are predicted relative to neutral AT. Natural population analysis (NPA) charge distributions are also seen to shift. Those of the AT anion also indicate that the unpaired electron is localized on the pyrimidine (thymine). Density functional theory consistently predicts a substantial positive adiabatic electron affinity for the AT pair (e.g., TZ2P++/B3LYP: +0.31 eV). This contrasts to second-order perturbation theory (MP2) treatments which predict unbound base pair anions. Despite the greater AEA of isolated T relative to C (+0.15 vs -0.02 eV), the AEA of the AT pair is slightly smaller than that of GC (0.31 vs 0.48 eV). This difference is attributed to the weaker solvating capacity of A (A.T-) relative to G (G.C-). The pairing (dissociation to A + T-) energy of AT(-) is determined to be 14.8 kcal/mol. This value is slightly greater than previous estimates for neutral AT from theory (12.4 kcal/mol) and experiment (13.0 kcal/mol).