The predominantly unknown nature and magnitude of the total taphonomic biasing that affects fossil assemblages commonly limits the testing of paleoecological hypotheses by the quantitative comparison of fossil assemblages. Multiple regression analyses of a suite of taphonomic characteristics against a measure of the total taphonomic biasing of a fossil assemblage can be used to generate a taphonomic model that identifies and quantifies many of those taphonomic processes that are important in biasing fossil assemblages. By examining the parameters of regression-based quantitative taphonomic models, key elements of the taphonomic history of a fossil assemblage can be reconstructed and such histories compared between fossil assemblages. We discuss the methodology pertinent to the building of quantitative taphonomic models and illustrate the effectiveness of such an approach using a case study considering two samples (857 specimens) of an Oligocene mammalian fossil assemblage from the Scenic Member of the Brule Formation, South Dakota. It is possible to quantify all of the taphonomic biasing affecting bone abundances in each of the two sampled assemblages using a small number of taphonomic characteristics. The most important of these characteristics is bulk density. The control of biasing by density is attributed to a combination of the increased resistance to physical damage of dense bones as well as the decreased probability of transport. Element shape and surface area to volume ratio are shown to have a significant but lesser control on assemblage biasing. Comparison of the taphonomic models for each sampled assemblage using currently viable methods demonstrates that the two assemblages are not statistically isotaphonomic. This interpretation may change with the gathering of additional data.