Flow-cytometry was used to measure cell volumes and DNA contents of single cells in cultures of soil bacteria during exponential growth and starvation conditions. DNA was measured after staining with mitramycin/ethidium bromide. The measurement of DNA was calibrated with rifampicin-treated cells of E. coli containing even numbers of genomes per cell. Cell volumes were assessed by scatter light measurements. Constant DNA to biovolume relations over a range of cell sizes were found for each of the bacteria at exponential growth, and DNA contents per cell varied over a range equivalent to 1-4 genomes per cell. At generation times of 1.0-1.5 h, two genomes were registered as a mean. After starvation of washed cells in a salt solution (24 hrs), a fraction of the cells in each culture had DNA contents equivalent to 1 genome, but significant fractions retained DNA contents equivalent to 2-4 genomes. Attempts to create cells with even numbers of genomes per cell by treatment with rifampicin was successful on an Acinetobacter sp. In contrast, the response to rifampicin was less clear for Pseudomonas fluorescens and P. chlororaphis, and unclear for the gram positive bacteria isolated from soil. The mean decrease in biovolume upon starvation was 4.1 times (range 1.3-8.1 times) and larger than the mean decrease in DNA content of 1.8 (range 1.3-2.7 times). Cell volume determinations by measurements of scatter light was compared with volume determinations by fluorescence microscopy. The amounts of scatter light per volumes was variable, not only did we find large differences between bacterial types, but also between starving and exponentially growing cells of the same isolate. In order to use light scatter as a measure of biovolume, internal standards has to be chosen of comparable size and surface properties as to soil bacteria.