We analyze HST Goddard High-Resolution Spectrograph observations of the 1216, 2600, and 2800 angstrom spectral regions for the spectroscopic binary system Capella, obtained at orbital phase 0.26 with 3.27-3.57 km s-1 resolution and high signal-to-noise ratio. We infer the column densities of H I, D I, Mg II, and Fe II for the local interstellar medium along this 12.5 pc line of sight, together with estimates of the temperature and turbulent velocity. If we assume that the intrinsic Lyalpha lines of the component stars in the Capella system can be approximated as scaled solar lines (with self-reversals), which is consistent with the observed Lyalpha profile of the high radial velocity star delta Lep, then the interstellar neutral hydrogen column density toward Capella, N(HI) = 1.80(+/- 0.1) x 10(18) cm-2. This corresponds to an average hydrogen number density, n(HI) = 0.047 cm-3. To account for a wider range of intrinsic line profiles that may characterize the more active Capella stars, we increase the allowed range of N(HI) to 1.8(+ 0.3, - 0.1) x 10(18) cm-2. The deuterium column density and line width parameter are found to be N(DI) = 2.97( + 0.13, - 0.05) x 10(13) cm-2 and b(D) = 7.81(+ 0.23, - 0.03) km s-1. The widths of the interstellar D I, Mg II, and Fe II lines indicate that the broadening has both thermal and turbulent components with the temperature T = 7000 +/- 200 K, and the turbulent velocity xi = 1.66 +/- 0.03 km s-1. These parameters indicate that b(HI) = 10.9 km s-1. We infer that the atomic deuterium/hydrogen ratio by number is (D/H)LISM = 1.65(+ 0.07, - 0.18) x 10(-5) for this line of sight. Our value of the D/H ratio lies near the mean of many earlier but less certain values for the Capella line of sight and toward other stars located as far as 1 kpc from the Sun. We present evidence that a constant value for (D/H)LISM in the nearby Galactic disk should be adopted as the best available working hypothesis, but this hypothesis must be tested by future HST observations of Capella at phase 0.75 and of other stars. Galactic evolution calculations indicate that the primordial D/H ratio, (D/H)p, probably lies in the range of (1.5-3) x (D/H)LISM. Standard big bang nucleosynthesis models for (D/H)p = 2.2-2.5 x 10(-5) imply that OMEGA(B)h50(2) = 0.06-0.08, where OMEGA(B) is the baryonic density in units of the Einstein-de Sitter closure density, and h50 is the Hubble constant in units of 50 km s-1 Mpc-1. If H-0 = 80 km s-1 Mpc-1 as recent evidence suggests, then OMEGA(B) = 0.023-0.031. Thus the universe will expand forever, unless nonbaryonic matter greatly exceeds the amount of baryonic matter.
