FtsZ is the major cytoskeletal protein operating in bacterial cell division. FtsZ assembles into protofilaments in vitro, and there has been some controversy over whether the assembly is isodesmic or cooperative. Assembly has been assayed previously by sedimentation and light scattering. However, these techniques will under-report small polymers. We have now produced a mutant of Escherichia coli FtsZ, L68W, which gives a 250% increase in tryptophan fluorescence upon polymerization. This provides a real-time assay of polymer that is directly proportional to the concentration of subunit interfaces. FtsZ-L68W is functional for cell division, and should therefore be a valid model for studying the thermodynamics and kinetics of FtsZ assembly. We assayed assembly at pH 7.7 and pH 6.5, in 2.5 mM EDTA. EDTA blocks GTP hydrolysis and should give an assembly reaction that is not complicated by the irreversible hydrolysis step. Assembly kinetics was determined with a stopped-flow device for a range of FtsZ concentrations. When assembly was initiated by adding 0.2 mM GTP, fluorescence increase showed a lag, followed by nucleation, elongation, and a plateau. The assembly curves were fit to a cooperative mechanism that included a monomer activation step, a weak dimer nucleus, and elongation. Fragmentation was absent in the model, another characteristic of cooperative assembly. We are left with an enigma: how can the FtsZ protofilament, which appears to be one-subunit thick, assemble with apparent cooperativity?