The immediate Solar neighborhood should be a fair sample of the local Galaxy. However, the chemical abundance distribution of long-lived disk stars very near the Sun contains a factor of 5 to 10 more metal-poor stars, -1 less than or similar to [Fe/H] less than or similar to - 0.4 dex, than is consistent with modem star-count models of larger scale Galactic structure. The metallicity distribution of complete samples of long-lived stars has long been recognized as providing unique constraints on the early stages of chemical evolution of the Galaxy, so that one would like to resolve this anomaly. We present a new derivation of the local G-dwarf metallicity distribution, based on the Third Gliese catalog combined with Olsen's (A&AS, 54, 55, 1983) Stromgren photometry. Kinematic data for these same stars, as well as for a high-precision sample studied by Edvardsson et al. (A&A, 275, 101, 1993), provide clear evidence that the abundance distribution below [Fe/H] similar to - 0.4 contains two over-lapping distributions, the thick disk and the thin disk. However these samples in isolation do not allow a reliable deconvolution of the relative numbers in each population. We achieve this by comparing the local metallicity distribution with a recent determination (Gilmore, Wyse, & Jones, AJ, 109, 1095, 1995) of the metallicity distribution of stars, selected with the same evolutionary criteria as applied to our nearby star sample, but found in situ some 1500 pc from the Sun. The gravitational sieve of the Galactic potential acts on this second sample to segregate the low velocity dispersion, thin-disk, component of the local sample, leaving predominantly the second, higher velocity dispersion component. Thus the two samples are complementary, with the local sample providing accurate data primarily for the thin disk, but weak thin-disk/thick-disk discrimination, and the distant sample providing excellent thick-disk/thin-disk segregation. We are thus able to break the degeneracy between overlapping phase-space distributions. That is, the combination of these two datasets allows us to determine the source of the local paradox: there is a substantial tail of the thin disk (defined kinematically) metallicity distribution, which extends below [Fe/H] approximate to - 0.4 dex. This is a robust conclusion, being consistent with the sum of star count, stellar spatial density distribution, and both local and distant kinematic and chemical abundance data. Using the sum of this information, we deconvolve the local abundance distribution function into thin disk and thick disk components, and show these overlap considerably. This overlap has implications for some dynamical merger models of the formation of the thick disk. The observed scatter in the thin disk age-metallicity relation however obscures any simple interpretation in terms of thick disk formation models. The chemical abundance distributions derived in this paper, which are based on the sum of local and distant data, and so are directly integrated over the Solar Cylinder, improve long-standing constraints on Galactic disk evolution. When combined with age and element ratio data, comprehensive constraints on the evolution of the disk will be available. (C) 1995 American Astronomical Society.
