The longitudinal dispersion coefficient describes the change in characteristics of a solute cloud, as it travels along the longitudinal axis of a flow. Within potable water networks, it is important to be able quantify this parameter, to predict the fate of solutes introduced into the network. Current water quality models assume steady, highly turbulent flow [Tzatchkov et al., 2009]. However, this assumption is not valid for the network's periphery, where water leaves the main network and comes to the point of consumption. Here, the flow can be both unsteady and turbulent, transitional or laminar [Buchberger et al., 2003; Blokker et al., 2008, 2010]. Taylor [1954] proposed a now classical expression to predict the longitudinal dispersion coefficient within steady, turbulent pipe flow. However, experimental data has shown significant deviation from his prediction for Re < 20000. Within the present work, new experimental data is presented for steady and unsteady flows for a range of discharges corresponding to 2000 < Re < 50000. From this experimental work, results describing the mixing processes through steady and unsteady turbulent and transitional pipe flow are presented, as well as an explana- tion as to why Taylor's theory fails to predict experimental data for Re < 20000. In addition, a simple numerical model is proposed for steady flow for 2000 < Re < 50000. The model extends Taylor's analysis to predict the longitudinal disper- sion coefficient in a manner more consistent with experimental data for Re < 20000. Furthermore, the model is extended for use within unsteady flow.