We use the yeast Saccharomyces cerevisiae, a unicellular eukaryote amenable to high-throughput molecular analyses, as a model organism for empirical validation of evolutionary theories of programmed or non-programmed aging and age-related death. We found that lithocholic bile acid (LCA) is a geroprotector which significantly delays the onset and reduces the rate of yeast chronological aging. Unlike mammals, yeasts do not synthesize bile acids. Our analysis of how LCA and other anti-aging compounds (including resveratrol, caffeine, and rapamycin) synthesized and released into the environment by one species of the organisms composing an ecosystem extend longevity of other species within this ecosystem suggests that these interspecies chemical signals may create xenohormetic, hormetic and cytostatic selective forces driving the evolution of longevity regulation mechanisms at the ecosystemic level. To test this hypothesis, we carried out the LCA-driven experimental evolution of long-lived yeast species. Our selection yielded three yeast strains with greatly extended lifespans. As an empirical test of the antagonistic pleiotropy and life history optimization theories of aging, we analyzed the trade-offs between early-life fitness and longevity by measuring the relative fitness of each of the laboratory-evolved long-lived yeast species in a direct competition assay with the parental wild-type strain. The assay was carried under caloric restriction (CR) conditions, which mimic the natural stressful environment of cyclical starvation, as well as under non-CR conditions in a more favorable environment. We found that, if cultured under field-like CR conditions, each of the three long-lived yeast species exhibits reduced fitness. Such a trade-off between early-life fitness and longevity was significantly less pronounced under laboratory-like non-CR conditions. Our studies of the mechanisms underlying the observed trade-off between early-life fitness and longevity revealed that 1) none of the laboratory-evolved long-lived yeast species has reduced fecundity; and 2) the decreased relative fitness of each of these species was not due to a reduction in growth rate early in life. Our monitoring of the age-dependent dynamics of changes in mitochondrial composition, morphology, and function implies that the longevity-associated fitness defect in each of the three laboratory-evolved long-lived yeast species is a dominant genetic trait attributed to specific lifespan-extending alterations in mitochondria. Based on our findings, we propose a hypothesis of the natural selection forces and underlying mechanisms that drive the evolution of yeast longevity and maintain a finite yeast lifespan within ecosystems.