Physics assigns to gravity the responsibility of forming structure on scales ranging from terrestrial to cosmological. An apparent threshold arises somewhere between scales characteristic of star clusters and those characteristic of galaxies. The internal dynamics of gravitationally bound structures smaller than a few parsecs, of which globular clusters are the largest examples, are reasonably well described in terms of standard gravity (i.e., Einstein's general theory and/or its Newtonian approximation) sourced by known substances. The internal dynamics of gravitationally bound structures larger than a few tens of parsecs, of which dwarf spheroidal (dSph) galaxies are the smallest examples, are not.
The ubiquity of dark matter on galactic and larger scales signifies new physics. Either there exists an otherwise unknown substance that contributes to dynamical mass but not to baryonic mass, or the standard dynamical framework requires modification (or both). The ‘substance’ hypothesis is not falsifiable, but in principle it can be confirmed with the detection of non-gravitational interactions involving dark matter particles. In any case, dSphs provide the most extreme examples of dark matter phenomenology, with dynamical mass-to-light ratios M / LV ≳ 10 [M / LV]⊙ even at their centers. This fact has made dSphs the focus of intense scrutiny in the effort to understand the nature of dark matter.
This article reviews the development of empirical constraints on the amount and distribution of dark matter within the Milky Way's dSph satellites. These results follow from the application of a rich variety of analyses applied to observations conducted by many individuals and groups. All analyses described here are formulated within the Newtonian dynamical framework. The reader is welcome to interpret ‘dark matter’ in terms of the substance hypothesis or more generally as a quantification of the discrepancy between dynamical and baryonic mass. The focus here is on the relationship between data and constraint, and one hopes that this information translates meaningfully into alternative dynamical frameworks (as formulated, e.g. by Bergmann, 1968, Milgrom, 1983, Bekenstein, 2004, Moffat, 2006).