A model is described for the complex fluorescence of tryptophan in proteins involving a reversible process between the excited singlet state of tryptophan and a "dark" (nonemissive) state of high energy. A specific hypothesis-for the nature of this dark state is given that involves rapid ionization of the excited tryptophan followed by indole nitrogen proton transfer on the nanosecond time scale. The general kinetic scheme involved predicts that there will be a change in amplitude of the decay components when an external quencher is added. This experiment is presented and shown to be consistent with the model. It is argued that this result is inconsistent with any static conformational heterogeneity hypothesis.