(Abridged) Many dwarf galaxies exhibit sub-Solar metallicities, with some star-to-star variation, despite often containing multiple generations of stars. The total metal content in these systems is much less than expected from the heavy element production of massive stars in each episode of star formation. Such a deficiency implies that a substantial fraction of the enriched material has been lost from these small galaxies. Mass ejection from dwarf galaxies may have important consequences for the evolution of the intergalactic medium and for the evolution of massive galaxies, which themselves may have formed via the merger of smaller systems. We report here the results of three-dimensional simulations of the evolution of supernova-enriched gas within dwarf spheroidal galaxies (dSph's), with the aim of determining the retention efficiency of supernova ejecta. We consider two galaxy models, selected to represent opposite ends of the dSph sequence. For each model galaxy we investigate a number of scenarios, ranging from a single supernova in smooth gas distributions to more complex multiple supernovae in highly disturbed gas distributions. The results of these investigations suggest that, for low star-formation efficiencies, it is difficult to completely expel the enriched material from the galaxy. Most of the enriched gas is, however, lost from the core of the galaxy following multiple supernovae, especially if the interstellar medium is already highly disturbed by processes such as photo-ionization and stellar winds. If subsequent star formation occurs predominantly within the core where most of the residual gas is concentrated, then these results could explain the poor self-enrichment efficiency observed in dwarf galaxies.

29 pages, 10 figures, to appear in Astrophysical Journal