In this study, we have investigated for the first time the metabolism of E. coli grown on agar using C-13 metabolic flux analysis (C-13-MFA). To date, all C-13-MFA studies on microbes have been performed with cells grown in liquid culture. Here, we extend the scope of C-13-MFA to biological systems where cells are grown in dense microbial colonies. First, we identified new optimal C-13 tracers to quantify fluxes in systems where the acetate yield cannot be easily measured. We determined that three parallel labeling experiments with the tracers [1,2-C-13] glucose, [1,6-C-13] glucose, and [4,5,6-C-13] glucose permit precise estimation of not only intracellular fluxes, but also of the amount of acetate produced from glucose. Parallel labeling experiments were then performed with wild-type E. coli and E. coli.ackA grown in liquid culture and on agar plates. Initial attempts to fit the labeling data from wild-type E. coli grown on agar did not produce a statistically acceptable fit. To resolve this issue, we employed the recently developed co-culture C-13-MFA approach, where two E. coli subpopulations were defined in the model that engaged in metabolite cross-feeding. The flux results identified two distinct E. coli cell populations, a dominant cell population (92% of cells) that metabolized glucose via conventional metabolic pathways and secreted a large amount of acetate (similar to 40% of maximum theoretical yield), and a second smaller cell population (8% of cells) that consumed the secreted acetate without any glucose influx. These experimental results are in good agreement with recent theoretical simulations. Importantly, this study provides a solid foundation for future investigations of a wide range of problems involving microbial biofilms that are of great interest in biotechnology, ecology and medicine, where metabolite cross-feeding between cell populations is a core feature of the communities.