The relationship between specific growth rate (mu) and steady-state glucose concentration was investigated for Escherichia coli ML30 in carbon-limited chemostat culture. This was made possible by the development of a method for measuring reducing sugars in culture media in the microgram.1-1-range. Cells initially cultivated in batch culture at high glucose concentrations required long-term adaptation to nutrient-limited growth conditions in the chemostat (between 100-200 volume changes at D = 0.6 h-1) until steady-state with respect to residual glucose concentration was reached; for adapted cells, however, new steady-state glucose concentrations were usually obtained within less than 10 volume changes. A statistical evaluation of different kinetic models showed that between 0.2 h-1 < D < 0.8 h-1 the three models proposed by Monod (1942), Shehata and Marr (1971), and Westerhoff et al. (1982) described the data equally well and the applicability of the different models is discussed. Depending on the model used, calculated glucose concentrations supporting half maximum growth rate (Ks) were in the range of 40-88 micrograms.1-1. The data strongly suggest that the large differences in Ks constants reported in the literature (ranging from 40 micrograms.1-1 up to 99 mg.1-1) are due to the use of E. coli cells adapted to different degrees to nutrient-limited growth conditions. This indicates that it is probably not possible to describe the kinetic properties of a bacterium with a single set of kinetic 'constants'.