The surface metabolism theory posits that adaptive evolution initiated when autocatalytic chemical systems became spatially localized on mineral surfaces. We searched for such surface-limited metabolisms (SLiMes) using a chemical ecosystem selection paradigm. This involves creating a prebiotic microcosm containing mineral grains and a “soup,” rich in food and potential sources of chemical energy, and then serially transferring a subset of the grains to a new microcosm containing fresh soup and new grains. This repeated dilution should enrich for chemical systems that can self-propagate more rapidly than the rate of serial dilution, and such enrichment should be detectable based on changes in microcosm chemistry over the course of multiple transfers. We deployed chemical ecosystem selection on several different soups and minerals and identified a combination that appears to be conducive to the enrichment of a SLiMe. In these conditions, chemical changes were observed over the first 12–18 transfers, most notably a loss of both orthophosphate and organics (as detected by optical density) from the solution. This loss from the solution correlated with the appearance of fractal structures on the surface of the grains. The putative SLiMes show clear evidence of self-propagation ability and manifest basic ecological dynamics. Ongoing work is evaluating the systems’ evolutionary capacity.