Microbial transformation rate data and theoretical considerations were analyzed for selected organic chemicals with respect to the general utility of mathematical models for predicting microbial transformation rates for risk assessment and regulatory purposes. By recognizing the unique problems associated with predicting microbial transformation rates within specific substrate concentration ranges, [S], the research, development, and testing of predictive mathematical models for environmental exposure assessment can be better focused. Lacking site-specific data, such an approach may yield useful interim models to meet our current needs as our understanding of environmental processes continues towards developing models more capable of accurately predicting transformation rates over broader ranges of conditions. One range of [S] considered for separate treatment was a sub-maintenance range (⩽0.05 μM), where microbial transformation rates may depart from pseudo-first-order kinetics (first-order in [S] with constant biomass) as a result of an insufficient substrate concentration for cell biomass maintenance. In another range of [S] considered for separate treatment, the toxic range (⩾5 mM), organic substrates may be sufficiently toxic to the general metabolic activities of the organism transforming the substrate that transformation rates do not follow Michaelis-Menten kinetics. Without site-specific data, mirobial transformation rates in these ranges of [S] should probably either be disregarded or assumed to occur at the maximum possible rate, depending, respectively, on whether the environmental concern lies in the disappearance of a toxic parent compound or in the formation of a toxic metabolite. Between these ranges lies an area characterized by multisystem kinetics, a region of [S] where microbial transformation rates appear to be reasonably predictable, albeit at a wide range of rates depending on the particular microorganisms dominating the transformation process, their concentrations, and the associated environmental conditions. For this region of [S], we suggest that there are important considerations to be made whenever a set of apparent kinetic parameters determined by conventional plots of kinetics data at low [S] are used to predict microbial transformation rates at much highest [S], and vice versa. To accomodate the possible effects of multiple systems operating over a wide range of [S], either the Michaelis - Menten equation with K m and V max values experimentally determined at moderately high [S] (approximately 0.1–1 mM), or an empirical equation for extrapolating pseudo-first transformation rate coefficients and half-life values to higher ranges of [S], should be used.