Hubble Space Telescope imaging of 11 fields in M31 were reduced to color-magnitude diagrams. The fields were chosen to sample all galactocentric radii to 50 kpc. Assuming that the bulk of the sampled stellar populations are older than a few Gyr, the colors of the red giants map to an abundance distribution with errors of order 0.1 dex in abundance. The radially sampled abundance distributions are all about the same width, but show a mild abundance gradient that flattens outside ~20 kpc. The various distributions were weighted and summed with the aid of new surface brightness profile fits to obtain an abundance distribution representative of the entirety of M31. M31 is a system near chemical maturity. This ``observed closed box'' is compared to analytical closed box models. M31 suffers from a lack of metal-poor stars and metal-rich stars relative to the simplest closed-box model in the same way as the solar neighborhood.Comparing to several simple chemical evolution models, neither complete mixing of gas at all times nor zero mixing, inhomogeneous models give the most convincing match to the data. As noted elsewhere, the outer disk of M31 is a factor of ten more metal-rich than the Milky Way halo, ten times more metal-rich than the dwarf spheroidals cospatial with it, and more metal-rich than most of the globular clusters at the same galactocentric radius. Difficulties of interpretation are greatly eased if we posit that the M31 disk dominates over the halo at all radii out to 50 kpc. In fact, scaling from current density models of the Milky Way, one should not expect to see halo stars dominating over disk stars until beyond our 50 kpc limit. A corollary conclusion is that most published studies of the M31 "halo" are actually studies of its disk.