Aerobic growth of Escherichia coli and Paracoccus denitrificans has been studied in chemostat, fed batch, and recycling fermentor modes under carbon and energy limitation. Two abrupt drops or discontinuities in molar growth yield, Y, have been found that occur over relatively short ranges in the value of specific growth rate. Before the first discontinuity, Y is constant and maximal. After the first discontinuity, at a doubling time of 33 h, Y becomes constant again and independent of mu until the second discontinuity appears at a doubling time of about 50 h, corresponding to a mu of about 0.014. At this point, Y drops to a lower value that is constant at doubling times longer than 100 h, corresponding to a mu of about 0.007. The second discontinuity is associated in Paracoccus with elevated levels of guanosine tetraphosphate (ppGpp) that impose stringent regulation as has been found previously with Bacillus and Escherichia species. It is thus likely that the stringent response generally occurs in bacteria in vivo at a doubling time of about 50 h. The cause of the first discontinuity is unknown. All experiments indicate that Pirt-type calculations relating mu, Y, and maintenance energy demand are no longer valid. In chemostat experiments, the intercept of the relationship between specific substrate utilization and specific growth rate is defined as maintenance. However, this intercept most probably is caused by stringent regulation at low dilution rates. Three regions of bacterial growth rates are defined by this study, corresponding to doubling times of 0.5 to 15 h, 33 to 50 h, and greater than 100 h. Some growth behavior in each region is unique to that region.