Dissolved humic substances make up -50% or more of the dissolved organic carbon (DOC) in freshwater ecosystems, yet their trophodynamic roles remain unresolved. We separated DOC from two freshwater environments, a lake and a blackwater marsh, into two fractions, humic DOC and nonhumic DOC, using an XAD-8 resin to selectively adsorb the humic substances. Results of microcosm studies revealed that the humic fraction of DOC was used by natural bacterial assemblages from the lake and marsh as a C and energy source, as indicated both by increases in bacterial biovolume and rates of bacterial incorporation of [3H]thymidine. Humic substances supported fourfold less bacterial secondary production per unit of initial C, however, than did nonhumic substances from the same environment.Bacterial utilization of humic compounds accounted for a significant fraction of the total bacterial production on DOC, measured as increases in bacterial biovolume; humic substances supported an average of 22% of total growth on DOC from the lake and 53% of the total growth on DOC from the marsh. The relative bioavailability of both the humic and nonhumic fractions of DOC differed between the lake and blackwater marsh, with less bacterial production per unit of initial C occurring on marsh-derived dissolved compounds.In aquatic ecosystems, turnover of dissolved organic carbon (DOC) is almost exclusively the domain of bacteria as a result of their high numbers, large surface-to-volume ratios, and transport systems efficient at low substrate concentrations.Studies of bacterial utilization of DOC have indicated the existence of two distinct pools of dissolved compounds, one labile and one refractory, in most aquatic systems (Ogura 1975). Although the labile pool generally accounts for < 20% of the total DOC (Ogura 1975;Allen 1976), it turns over rapidly, on the order of hours to days, and is thought to support the bulk of DOC-based bacterial secondary production (Wright and Hobbie 1966;Allen 1976). The refractory pool is larger but turns over more slowly, on the order of weeks to months (Geller 1986) and therefore may be relatively unimportant as a substrate for bacterial growth. Neither of these two fractions of the DOC pool has been well characterized with respect to Acknowledgments We thank L. Pittman, G. Zeigler, E. Sheppard, and A. Brewer for technical assistance and J. Meyer, C. Strojan, and R. Wicks for helpful comments on this manuscript.