In a sandy loam soil cropped to barley bacterial production in the rhizosphere was compared to the results of a parallel investigation on rhizodeposition. Bacterial production was stimulated in the rhizosphere as revealed by an increased biomass of bacteria (643–883 µg C g^-1 soil) and protozoa (7.2–15 × 10⁴ cells g^-1 soil) as well as elevated thymidine incorporation (9.7–12 pmol g^-1 soil) in rhizosphere soil compared to bulk soil. Rhizodeposition, as determined by several pulse labellings with ¹⁴CO₂, was estimated to be 412 µg C g^-1 dry wt soil in the 0–15 cm layer. Bacterial production, as determined by incorporation of ³H-labelled thymidine converted to bacterial C, revealed a plant-induced formation of 1348 µg bacterial C g^-1 soil in the 0–15 cm layer. This is probably the first estimate for bacterial production based on thymidine incorporation which has been compared to an estimate of C availability in a terrestrial ecosystem. The discrepancy between the rhizodeposition data and the estimate of bacterial production suggests that conversion factors between [³H]thymidine incorporation and bacterial production obtained in laboratory incubations are not generally applicable to the field situation. It is proposed that estimates of bacterial production should be based on measurements of actual conversion factors to improve the accuracy of the method.