Chemical tracers can be used to assess the simulated circulation in ocean models. Tracers that have been used in this context include tritium, chlorofluorocarbons, natural and bomb-produced radiocarbon, and to a lesser extent, oxygen, silicate, phosphate, isotopes of organic and inorganic carbon compounds, and certain noble gases (e.g., helium and argon). This paper reviews the use of chemical tracers in assessing the circulation and flow patterns in global and regional ocean models. It will be shown that crucial information can be derived from chemical tracers that cannot be obtained from temperature-salinity (T-S) alone. In fact, it turns out that a model with a good representation of T-S can have significant errors in simulated circulation, so checking a model's ability to capture chemical tracer patterns is vital. Natural chemical tracers such as isotopes of carbon, argon, and oxygen are useful for examining the model representation of old water masses, such as North Pacific and Circumpolar Deep Water. Anthropogenic or transient tracers, such as tritium, chlorofluorocarbons, and bomb-produced 14C, are best suited for analyzing model circulation over decadal timescales, such as thermocline ventilation, the renewal of Antarctic Intermediate Water, and the ventilation pathways of North Atlantic Deep Water and Antarctic Bottom Water. Tracer model studies have helped to reveal inadequacies in the model representation of certain water mass formation processes, for example, convection, downslope flows, and deep ocean currents. They show how coarse models can chronically exaggerate the spatial scales of open-ocean convection and deep currents while underestimating deep flow rates and diffusing downslope flows with excessive lateral mixing. Higher-resolution models typically only resolve thermocline ventilation because of shorter integration times, and most resort to high-latitude T-S restoring to simulate reasonable interior water mass characteristics. This can be seen to result in spuriously weak chemical tracer uptake at high latitudes due to suppressed convective overturn and vertical motion. Overall, the simulation of chemical tracers is strongly recommended in model assessment studies and as a tool for analyzing water mass mixing and transformation in ocean models. We argue that a cost-effective approach is to simulate natural radiocarbon to assess long-timescale processes, and CFCs for decadal to interdecadal ocean ventilation. ¿ 2001 American Geophysical Union