One of the classical tests of the early dynamical evolution of the Milky Way is the prediction of the monolithic collapse model (e.g., Eggen, Lynden-Bell, & Sandage 1962) of a decline in the mean stellar abundance of the halo population as one proceeds to stars at greater distances, or equivalently, with higher local space velocities in the solar neighborhood. This stands in contrast to the prediction of galaxy formation within the Cold Dark Matter paradigm, where assembly from multiple fragments (e.g., Searle & Zinn 1978) would not be expected to produce an abundance gradient in the halo. Although several attempts have been made in the past to test this idea, all such efforts have been limited by small sample sizes, concerns about selection biases, or both. We are presently analyzing a very large sample of over 24,000 stars selected as calibration objects (used for providing checks on the spectrophotometric flux and reddening corrections) from SDSS-I DR-5. These stars are primarily F (and early G-type) turnoff stars in the thick-disk and halo populations of the Galaxy. The color-based selection ensures that an adequate number (several thousand) of very low-metallicity ([Fe/H] < -2.0) stars exist in order to search for the presence (or not) of a halo abundance gradient. Accurate estimates of radial velocity, metallicity, temperature, surface gravity, and distance are obtained for all of these stars by application of the (still evolving) SDSS/SEGUE spectroscopic analysis pipeline discussed in other contributions at this meeting. This information is combined with proper motions derived from the recalibrated USNOB-2 catalog, as discussed by Munn et al. (2004), in order to obtain estimates of their full space motions. Results on the search for a halo abundance gradient, based on these data, will be reported.