Thymidine incorporation into DNA is widely used to estimate rates of bacterial growth and secondary production in aquatic systems. The procedure requires the use of several conversion factors and assumptions to convert rates of thymidine uptake to rates of carbon production. Perhaps the most controversial of the conversion factors is that which converts rates of incorporation to the rate of cell production. During a year-long study in Lake Arlington, Texas, we empirically determined conversion factors from bacterial growth in filtered (1.0 mum porosity) and diluted (1:9) lake water. Bacterial growth rate constants determined from changes in cell abundance were compared to growth rate constants estimated from empirically derived (both instantaneous and annually averaged), theoretical, and modeled conversion factors. Single value conversion factors (i.e., theoretical, or average of 19 empirically determined conversion factors) did not yield estimates of bacterial growth rate constants that compared favorably to growth estimates from changes in bacterial abundance. Conversion factors, determined from a regression model based on empirically determined conversion factors, gave annual growth estimates that were similar to those obtained from changes in cell abundance.