Conversion factors to relate the rate of 3H-thymidine (3HTdr) incorporation into soil bacteria to the corresponding rate of formation of cells and biomass were determined with pure cultures of bacteria and with a mixed soil bacterial population colonizing wheat seedling roots. One week after inoculation, sand with seedlings was dried out, and inorganic N was added to half the pots. After 12 h the sand was rewetted which resulted in a flush of bacterial growth for 1 day. Yeast and most Pseudomonas strains did not incorporate 3HTdr. Strains of Corynebacterium, Acinetobacter, Nocardia and one Pseudomonas incorporated thymidine with a resultant mean conversion factor of 2.4 × 10¹⁸ cells mol^-1 Tdr or 175 kg C mol^-1 Tdr. On wheat seedlings, before and 1 day after the flush, the mean formation of 2.1 × 10¹⁸ cells mol^-1 Tdr, and 101 kg C mo1^-1 Tdr for both rhizosphere and bulk sand was observed. Although the cultured isolates of soil bacteria were 20 times larger in cellular biovolume as a mean compared to the mixed soil population, mean conversion factors corresponded. Cell formation mol^-1 Tdr was up to one order of magnitude higher during the flush following rewetting of sand, and this shows that bacterial production in flush situations may only be measured by 3HTdr incorporation with very high uncertainty because of low efficiency of DNA labelling. Alternatively bacterial production may be based on increase in biovolume determined by microscopy because bacterial death rate will be negligible in this situation.