Very little manganese is imported into Escherichia coli under routine growth conditions: the import system is weakly expressed, the manganese content is low, and a manganese-dependent enzyme is not correctly metallated. Mutants that lack MntH, the importer, grow at wild-type rates, indicating that manganese plays no critical role. However, MntH supports the growth of iron-deficient cells, suggesting that manganese can substitute for iron in activating at least some metalloenzymes. MntH is also strongly induced when cells are stressed by hydrogen peroxide. This adaptation is essential, as E. coli cannot tolerate peroxide stress if mntH is deleted. Other workers have observed that manganese improves the ability of a variety of microbes to tolerate oxidative stress, and the prevailing hypothesis is that manganese does so by chemically scavenging hydrogen peroxide and/or superoxide. We found that manganese does not protect peroxide-stressed cells by scavenging peroxide. Instead, the beneficial effects of manganese correlate with its ability to metallate mononuclear enzymes. Because iron-loaded enzymes are vulnerable to the Fenton reaction, the substitution of manganese may prevent protein damage. Accordingly, during H2O2 stress, mutants that cannot import manganese and/or are unable to sequester iron suffer high rates of protein oxidation.