A simulation model is used to predict the fate of carbon derived from lignocellulose in the food web of a marsh community in the Okefenokee Swamp, Georgia. The model, constructed from empirical work spanning a 6—yr period, predicts the role of lignocellulose—derived detritus in supporting secondary production in this ecosystem. Although Carex walteriana lignocellulose is the dominant product of primary production in this marsh and the dominant source of detritus, the model predicts that it may not be quantitatively important in the support of tropic groups other than bacteria and protozoa. The efficiency of transfer lignocellulosic carbon through the microbial loop is low, and overall only 8% of the carbon used by bacteria and fungi is predicted to be ingested subsequently by metazoans. An average of 32% of the bacterial biomass in the marsh water and sediments is predicted by the model to be produced at the expense of lignocellulosic carbon, whereas <10% of the biomass of larger detritivorous animals can be traced to carbon from lignocellulose. Thus in this marsh ecosystem, detritus food webs based on other, less abundant, sources of detrital carbon, or perhaps classical plant—herbivore food webs, are more likely to be the major source of carbon for metazoan production.