A biomass-based model for cohort analysis was examined for its performance in estimating total stock biomass over time. The model is based on the premise that a biomass-based analysis that incorporates growth can account for changes in population biomass resulting from changes in individual weight in addition to changes in number, and thereby give better estimates of population biomass over time. The biomass-based approach also facilitates the direct incorporation of data into the analysis in terms of biomass, bypassing the process of converting from biomass to numbers to do the calculations, and then converting the numbers back to biomass in order to make management decisions based on biomass. To incorporate growth into the analysis, we assumed that growth in weight occurs exponentially over the time increment of the analysis. Sensitivity of the analysis to the assumption of exponential growth was tested by examining its performance on three hypothetical data sets generated by simulations incorporating alternative representations of growth. The tests demonstrated that including growth enhanced the quality of the estimates. In addition, the model was extended to incorporate seasonal growth and seasonal catch. The model was applied to data on stocks of Alaska plaice Pleuronectes quadrituberculatus, walleye pollock Theragra chalcogramma, and yellowfin sole Limanda aspera from the eastern Bering Sea to demonstrate the features of the biomass-based analysis. Annual changes in growth rate were sometimes significant and influenced the results substantially, indicating that the inclusion of growth into this type of analysis may be important. The inclusion of seasonal growth and catch further improved the results of the analysis.