During 4 diel cycles in a Danish coastal environment, bacterial growth rates were estimated by several methods. ³H-thymidine incorporation into DNA gave results of the same order of magnitude as data obtained from frequency of dividing cells (FDC). In contrast to minor diel changes in cell numbers and bacterial secondary production, marked diel changes were observed in the percentage of active bacteria determined from microautoradiography using ³H-thymidine. Between 20 and 80 % of the bacteria were active. Increases in percent activity were frequently but not always found in the morning. Attempts to estimate specific activity of thymidine nucleotide pools by isotope dilution procedure suggested that there was less intracellular dilution than expected from literature data; however, it seemed to be an underestimate and results were often difficult to evaluate. Results from biomass changes sometimes revealed appreciably higher production rates than those obtained from thymidine incorporation, whether or not the thymidine data were separately adjusted for presumed isotope dilution. Based on the assumption that all dividing bacteria were active, bacterial biomass and FDC were corrected according to the percent active bacteria determined from autoradiography. Corrected bacterial production rates were on average 28 % (range 4 to 169 %) higher than those obtained from standard FDC procedure.