An Adaptive Kernel Approach to Finding dSph Galaxies Around the Milky Way

Because of their low surface brightness, several dwarf spheroidal (dSph) companions of the Milky Way were discovered only recently. Because dSph's are partially resolved into individual stars, we can conduct a systematic search for them using lists of stars from scanned plates in the COSMOS/UKST survey of the southern sky. We thus constrain the population of dSph companions to the Galaxy. In principle, application of the technique to a deeper all-sky survey can place interesting constraints on the Local Group luminosity function. Using a variant of the adaptive kernel approach employed to search for galaxy clusters (Postman et al., AJ, 111, 615, 1996), we construct a two-dimensional histogram of the stellar surface density, filter it with a circularly symmetric filter parameterized by a scale factor, and search for maxima in the filtered histogram. Our technique easily identifies the two existing dSph galaxies in the region of sky examined. All of our unexplained detections are plate defects and mis-identified cluster galaxies. We find no unambiguous new objects with L > 1/8L(Sculptor), the luminosity of the Sculptor dSph. We verify this limit by testing the detectability of artificial dSph galaxies placed onto COSMOS/UKST fields. Several of the known dSph companions of the Milky Way are close to the detection limit of our technique, implying that the sample of dSphs is probably complete to the limits of current surveys. (C) 1997 American Astronomical Society.