Dispersion processes occurring in flow injection (FI) systems were reviewed. The basic characteristics of dispersion and problems encountered in understanding dispersion as it applies to FI systems were considered. Obtaining reproducible dispersion is the basic principle underling FI systems. A uniformly acceptable mathematical description of the dispersion process, however, has not been developed. This has not diminished the utility of FI nor has it impeded its being applied to processes occurring in chromatographic injectors, detectors, connective tubing, and post column reactors. Predictive (theoretical) models that have attempted to explain dispersion phenomena in FI systems were discussed. These have been developed for nonreactive systems in which the sample does not react with the carrier and for kinetic (reactive) systems in which the sample does react with the carrier. Theoretical models for dispersion in nonreactive systems have attempted to describe dispersion in FI systems in terms of diffusive and convective processes, secondary flow processes, an axially dispersed plug flow, and as processes occurring in well stirred tanks (mixing chambers). Models developed for reactive FI systems have also modeled dispersion processes in terms of diffusion and convection, axially dispersed plug flows, and mixing chamber processes. These types of models have examined dispersion processes using typical chemical kinetic approaches and in terms of random walk and distributed and lumped parameter models. Purely descriptive approaches for describing FI systems were discussed. Descriptive approaches utilize information obtained from detector response curves to define or describe an FI system. Such approaches have been used to deconvolute detector responses, map response surfaces, analyze peak shapes, and to analyze errors.