Investigation of the Contamination of the Gould (2003) Halo Sample

A recent astroph posting argued that the Gould (2003a) halo sample is substantially contaminated with thickdisk stars, which would then “wash out” any signature of granularity in the halo velocity distribution due to streams. If correct, this would imply that the limits placed by Gould (2003b) on streams are not valid. Here I investigate such contamination using six different indicators, 1) morphology of the underlying reduced proper motion diagram used to select halo stars, 2) comparison of kinematic and parallax-based distance scales 3) comparison of derived halo parameters for the Gould (2003a) sample with other determinations 4) a precision color-color diagram for a random subsample 5) the 3-dimensional velocity distribution of a random subsample 6) metallicity distribution versus kinematic cuts on a random subsample I estimate that the contamination is of order 2 percent. Thus, the upper limits on the density of nearby streams derived by Gould (2003b) remain valid. In particular, at 95% confidence, no more than 5% of local halo stars (within about 300pc) are in any one coherent stream. Determining whether or not this local measurement is consistent with CDM remains an outstanding question. Subject headings: stars:halo – galaxies:substructure